ic
/
magnum
mirror of https://github.com/mosra/magnum.git
6
0
Fork 0
Code Issues Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
9564 Commits
21 Branches
19 Tags
70 MiB
Tree: abdbada144
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'abdbada144'
${ noResults }
magnum/src/Magnum/Trade/Test/AbstractImporterTest.cpp

7670 lines
283 KiB
Raw Normal View History Unescape

Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
/*
This file is part of Magnum.
Updated copyright year.
6 years ago
Copyright © 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019,
Updated copyright year.
4 years ago
2020, 2021, 2022 Vladimír Vondruš <mosra@centrum.cz>
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
Permission is hereby granted, free of charge, to any person obtaining a
copy of this software and associated documentation files (the "Software"),
to deal in the Software without restriction, including without limitation
the rights to use, copy, modify, merge, publish, distribute, sublicense,
and/or sell copies of the Software, and to permit persons to whom the
Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included
in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
DEALINGS IN THE SOFTWARE.
*/
Trade: debug output for AbstractImporter::Feature and Features.
8 years ago
#include <sstream>
Adapted to Corrade changes.
11 years ago
#include <Corrade/Containers/ArrayView.h>
Removed std::optional backwards compatibility. This was a dirty mess. Good riddance. Also removing the include from the header now, as it has a forward declaration and thus is not needed.
7 years ago
#include <Corrade/Containers/Optional.h>
Trade: port to the new Utility::Path. For file opening there's no longer an unatomic pair of exists() + read(), but since Path::read() now returns an Optional, it means we can reliably distinguish between empty files and failures. While at it, also added TODOs for removal of the StringStl.h header that's needed in various places for compatibility with APIs still using STL strings.
4 years ago
#include <Corrade/Containers/String.h>
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
#include <Corrade/Containers/StringStl.h> /** @todo remove once Debug is stream-free */
Making includes absolute, part 4: using absolute includes for Corrade. Also adapted to some Corrade renames.
12 years ago
#include <Corrade/TestSuite/Tester.h>
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
#include <Corrade/TestSuite/Compare/Container.h>
Trade: return Optional instead of Pointer from *Importer::material(). The plugin interface version got bumped to avoid ABI issues when loading plugins that weren't updated for the change, but apart from that this shouldn't be a breaking change, as the API returns a type that can be both an Optional and a Pointer.
6 years ago
#include <Corrade/TestSuite/Compare/Numeric.h>
Trade: port to the new Utility::Path. For file opening there's no longer an unatomic pair of exists() + read(), but since Path::read() now returns an Optional, it means we can reliably distinguish between empty files and failures. While at it, also added TODOs for removal of the StringStl.h header that's needed in various places for compatibility with APIs still using STL strings.
4 years ago
#include <Corrade/TestSuite/Compare/String.h>
Adapted to Corrade changes. Mostly missing includes now.
7 years ago
#include <Corrade/Utility/DebugStl.h>
Trade: port to the new Utility::Path. For file opening there's no longer an unatomic pair of exists() + read(), but since Path::read() now returns an Optional, it means we can reliably distinguish between empty files and failures. While at it, also added TODOs for removal of the StringStl.h header that's needed in various places for compatibility with APIs still using STL strings.
4 years ago
#include <Corrade/Utility/Path.h>
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
Reserve zero [Compressed]PixelFormat for an invalid value. Catches accidents better -- just look at the diff. Might catch accidents in your code now as well, so beware!
7 years ago
#include "Magnum/PixelFormat.h"
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
#include "Magnum/FileCallback.h"
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#include "Magnum/Math/Matrix3.h"
#include "Magnum/Math/Matrix4.h"
Making includes absolute, part 5: absolute includes everywhere. The only places where they aren't absolute are: - when header is included from corresponding source file - when including headers which are not part of final installation (e.g. test-specific configuration, headers from Implementation/)
12 years ago
#include "Magnum/Trade/AbstractImporter.h"
Trade: animation import interface in AbstractImporter.
8 years ago
#include "Magnum/Trade/AnimationData.h"
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
#include "Magnum/Trade/ArrayAllocator.h"
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
#include "Magnum/Trade/CameraData.h"
#include "Magnum/Trade/ImageData.h"
#include "Magnum/Trade/LightData.h"
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
#include "Magnum/Trade/MeshData.h"
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
#include "Magnum/Trade/PhongMaterialData.h"
#include "Magnum/Trade/SceneData.h"
Trade: skin import in AbstractImporter.
6 years ago
#include "Magnum/Trade/SkinData.h"
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
#include "Magnum/Trade/TextureData.h"
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#include <Corrade/Containers/ArrayTuple.h>
#include <Corrade/TestSuite/Compare/Container.h>
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
#include <Corrade/Utility/Algorithms.h>
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
Trade, MeshTools: deprecate MeshDataXD and everything that uses it.
6 years ago
#define _MAGNUM_NO_DEPRECATED_MESHDATA /* So it doesn't yell here */
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#define _MAGNUM_NO_DEPRECATED_OBJECTDATA /* So it doesn't yell here */
Trade, MeshTools: deprecate MeshDataXD and everything that uses it.
6 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#include "Magnum/Trade/MeshData2D.h"
#include "Magnum/Trade/MeshData3D.h"
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#include "Magnum/Trade/MeshObjectData2D.h"
#include "Magnum/Trade/MeshObjectData3D.h"
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#endif
Making includes absolute, part 6: better filenames for internal files. As we are now using absolute includes, there is no need to prefix everything with "magnum<Namespace>" etc. All generated configuration files are renamed to configure.h and their path is included _before_ everything else to avoid accidental collisions.
12 years ago
#include "configure.h"
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
Wrap all tests in unnamed namespaces. This makes the compiler warn about test cases that aren't ever used, which is a good thing.
7 years ago
namespace Magnum { namespace Trade { namespace Test { namespace {
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct AbstractImporterTest: TestSuite::Tester {
explicit AbstractImporterTest();
void construct();
void constructWithPluginManagerReference();
Trade: add flags to AbstractImporter and AbstractImageConverter.
6 years ago
void setFlags();
void setFlagsFileOpened();
void setFlagsNotImplemented();
Trade: test also plain AbstractImporter::openData().
7 years ago
void openData();
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
void openDataDeprecatedFallback();
#endif
Trade: ability to open non-temporary memory. Using openMemory() instead of openData() allows the implementation to assume the data will stay in scope for as long as needed, which can prevent unnecessary copies in some plugin implementations. It warranted a new flag, DataFlag::ExternallyOwned, to describe this kind of memory. I couldn't reuse Owned as that's used for allocations owned by the instance, which is too little for certain future use cases. For example returning *Data instances referencing an Owned memory would mean the user has to assume the memory is gone when the importer instance is gone, and that's generally not true for memory passed to openMemory(). Originally I thought I would do this later, but then realized the existing plugin implementations would need to get all updated again to be aware of the new flag, with some being forgotten, and it's just easier to do the whole thing in a single step.
5 years ago
void openMemory();
Trade: AbstractImporter::openFile() and openState() had no tests?! I wonder what the reason was. Too simple to be tested? Coverage from testing actual plugins was enough??
4 years ago
void openFile();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void openFileAsData();
void openFileAsDataNotFound();
Trade: AbstractImporter::openFile() and openState() had no tests?! I wonder what the reason was. Too simple to be tested? Coverage from testing actual plugins was enough??
4 years ago
void openState();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void openFileNotImplemented();
void openDataNotSupported();
void openDataNotImplemented();
void openStateNotSupported();
void openStateNotImplemented();
void setFileCallback();
void setFileCallbackTemplate();
void setFileCallbackTemplateNull();
Trade: accept const references in AbstractImporter file callbacks.
7 years ago
void setFileCallbackTemplateConst();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void setFileCallbackFileOpened();
void setFileCallbackNotImplemented();
void setFileCallbackNotSupported();
void setFileCallbackOpenFileDirectly();
void setFileCallbackOpenFileThroughBaseImplementation();
void setFileCallbackOpenFileThroughBaseImplementationFailed();
void setFileCallbackOpenFileAsData();
void setFileCallbackOpenFileAsDataFailed();
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
void thingCountNotImplemented();
void thingCountNoFile();
void thingForNameNotImplemented();
void thingForNameNoFile();
Trade: print a message when named thing is not found. Otherwise it's hard to distinguish whether it's a particular importer fault (failing without a message) or if the name is not found.
6 years ago
void thingByNameNotFound();
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
void thingNameNoFile();
void thingNoFile();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void defaultScene();
void defaultSceneNotImplemented();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void defaultSceneOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void scene();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void object();
#ifdef MAGNUM_BUILD_DEPRECATED
void sceneDeprecatedFallback2D();
void sceneDeprecatedFallback3D();
void sceneDeprecatedFallbackParentless2D();
void sceneDeprecatedFallbackParentless3D();
void sceneDeprecatedFallbackTransformless2D();
void sceneDeprecatedFallbackTransformless3D();
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
void sceneDeprecatedFallbackMultiFunctionObjects2D();
void sceneDeprecatedFallbackMultiFunctionObjects3D();
Trade: report at least some *D object count if everything fails. As the comment says -- before the user code expected that if the scene hierarchy is broken, particular objects will fail to import. Now the whole scene fails to import, so we don't even get to know the actual (expanded) deprecated 2D/3D object count. To reduce the suffering, return at least the dimension-less object count there. It won't include the duplicates from the single-function-object conversion but better than reporting 0.
5 years ago
void sceneDeprecatedFallbackObjectCountNoScenes();
void sceneDeprecatedFallbackObjectCountAllSceneImportFailed();
Trade: another case for the deprecated path not being 100% compatible. Honestly I don't care much, this is just that the original PrimitiveImporter tests started to fail because they expected object3DForName() to return -1 for a 2D name, but that's no longer the case. It would be possible to fix this, but I doubt anyone ever relied on such behavior for 2D scenes, so just add a test that acknowledges this new behavior.
5 years ago
void sceneDeprecatedFallbackBoth2DAnd3DScene();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void sceneNameNotImplemented();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void objectNameNotImplemented();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void sceneForNameOutOfRange();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void objectForNameOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void sceneNameOutOfRange();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void objectNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void sceneNameCustomDeleter();
void objectNameCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void sceneNotImplemented();
void sceneOutOfRange();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void sceneNonOwningDeleters();
void sceneCustomDataDeleter();
void sceneCustomFieldDataDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
void sceneFieldName();
void sceneFieldNameNotImplemented();
void sceneFieldNameNotCustom();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void sceneFieldNameCustomDeleter();
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void animation();
void animationNameNotImplemented();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void animationForNameOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void animationNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void animationNameCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void animationNotImplemented();
void animationOutOfRange();
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void animationNonOwningDeleters();
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void animationGrowableDeleters();
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
void animationCustomDataDeleter();
void animationCustomTrackDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void light();
void lightNameNotImplemented();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void lightForNameOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void lightNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void lightNameCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void lightNotImplemented();
void lightOutOfRange();
void camera();
void cameraNameNotImplemented();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void cameraForNameOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void cameraNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void cameraNameCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void cameraNotImplemented();
void cameraOutOfRange();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void object2D();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void object2DCountNotImplemented();
void object2DCountNoFile();
void object2DForNameNotImplemented();
void object2DForNameNoFile();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void object2DForNameOutOfRange();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void object2DByNameNotFound();
void object2DNameNotImplemented();
void object2DNameNoFile();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void object2DNameOutOfRange();
void object2DNotImplemented();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void object2DNoFile();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void object2DOutOfRange();
void object3D();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void object3DCountNotImplemented();
void object3DCountNoFile();
void object3DForNameNotImplemented();
void object3DForNameNoFile();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void object3DForNameOutOfRange();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void object3DByNameNotFound();
void object3DNameNotImplemented();
void object3DNameNoFile();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void object3DNameOutOfRange();
void object3DNotImplemented();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void object3DNoFile();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void object3DOutOfRange();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
Trade: skin import in AbstractImporter.
6 years ago
void skin2D();
void skin2DNameNotImplemented();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void skin2DForNameOutOfRange();
Trade: skin import in AbstractImporter.
6 years ago
void skin2DNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void skin2DNameCustomDeleter();
Trade: skin import in AbstractImporter.
6 years ago
void skin2DNotImplemented();
void skin2DOutOfRange();
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
void skin2DNonOwningDeleters();
void skin2DCustomJointDataDeleter();
void skin2DCustomInverseBindMatrixDataDeleter();
Trade: skin import in AbstractImporter.
6 years ago
void skin3D();
void skin3DNameNotImplemented();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void skin3DForNameOutOfRange();
Trade: skin import in AbstractImporter.
6 years ago
void skin3DNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void skin3DNameCustomDeleter();
Trade: skin import in AbstractImporter.
6 years ago
void skin3DNotImplemented();
void skin3DOutOfRange();
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
void skin3DNonOwningDeleters();
void skin3DCustomJointDataDeleter();
void skin3DCustomInverseBindMatrixDataDeleter();
Trade: skin import in AbstractImporter.
6 years ago
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
void mesh();
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
void meshDeprecatedFallback();
#endif
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
void meshLevelCountNotImplemented();
void meshLevelCountOutOfRange();
void meshLevelCountZero();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void meshForNameOutOfRange();
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
void meshNameNotImplemented();
void meshNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void meshNameCustomDeleter();
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
void meshNotImplemented();
void meshOutOfRange();
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
void meshLevelOutOfRange();
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void meshNonOwningDeleters();
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void meshGrowableDeleters();
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
void meshCustomIndexDataDeleter();
void meshCustomVertexDataDeleter();
void meshCustomAttributesDeleter();
void meshAttributeName();
void meshAttributeNameNotImplemented();
void meshAttributeNameNotCustom();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void meshAttributeNameCustomDeleter();
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void mesh2D();
void mesh2DCountNotImplemented();
void mesh2DCountNoFile();
void mesh2DForNameNotImplemented();
void mesh2DForNameNoFile();
void mesh2DNameNotImplemented();
void mesh2DNameNoFile();
void mesh2DNameOutOfRange();
void mesh2DNotImplemented();
void mesh2DNoFile();
void mesh2DOutOfRange();
void mesh3D();
void mesh3DCountNotImplemented();
void mesh3DCountNoFile();
void mesh3DForNameNotImplemented();
void mesh3DForNameNoFile();
void mesh3DNameNotImplemented();
void mesh3DNameNoFile();
void mesh3DNameOutOfRange();
void mesh3DNotImplemented();
void mesh3DNoFile();
void mesh3DOutOfRange();
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void material();
Trade: return Optional instead of Pointer from *Importer::material(). The plugin interface version got bumped to avoid ABI issues when loading plugins that weren't updated for the change, but apart from that this shouldn't be a breaking change, as the API returns a type that can be both an Optional and a Pointer.
6 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
void materialDeprecatedFallback();
#endif
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void materialForNameOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void materialNameNotImplemented();
void materialNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void materialNameCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void materialNotImplemented();
void materialOutOfRange();
Trade: check for custom MaterialData deleters in AbstractImporter. And adapt MaterialData and the deprecated PhongMaterialData constructor to follow these rules. Basically the same as done for ImageData and MeshData.
6 years ago
void materialNonOwningDeleters();
void materialCustomAttributeDataDeleter();
void materialCustomLayerDataDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void texture();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void textureForNameOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void textureNameNotImplemented();
void textureNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void textureNameCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void textureNotImplemented();
void textureOutOfRange();
void image1D();
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void image1DLevelCountNotImplemented();
void image1DLevelCountOutOfRange();
void image1DLevelCountZero();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void image1DForNameOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void image1DNameNotImplemented();
void image1DNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void image1DNameCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void image1DNotImplemented();
void image1DOutOfRange();
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void image1DLevelOutOfRange();
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void image1DNonOwningDeleter();
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void image1DGrowableDeleter();
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
void image1DCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void image2D();
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void image2DLevelCountNotImplemented();
void image2DLevelCountOutOfRange();
void image2DLevelCountZero();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void image2DForNameOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void image2DNameNotImplemented();
void image2DNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void image2DNameCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void image2DNotImplemented();
void image2DOutOfRange();
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void image2DLevelOutOfRange();
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void image2DNonOwningDeleter();
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void image2DGrowableDeleter();
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
void image2DCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void image3D();
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void image3DLevelCountNotImplemented();
void image3DLevelCountOutOfRange();
void image3DLevelCountZero();
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void image3DForNameOutOfRange();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void image3DNameNotImplemented();
void image3DNameOutOfRange();
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void image3DNameCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void image3DNotImplemented();
void image3DOutOfRange();
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void image3DLevelOutOfRange();
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void image3DNonOwningDeleter();
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void image3DGrowableDeleter();
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
void image3DCustomDeleter();
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
void importerState();
void importerStateNotImplemented();
void importerStateNoFile();
void debugFeature();
void debugFeatures();
Trade: add flags to AbstractImporter and AbstractImageConverter.
6 years ago
void debugFlag();
void debugFlags();
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
};
Trade: print a message when named thing is not found. Otherwise it's hard to distinguish whether it's a particular importer fault (failing without a message) or if the name is not found.
6 years ago
constexpr struct {
const char* name;
bool checkMessage;
} ThingByNameData[] {
{"check it's not an assert", false},
{"verify the message", true}
};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
using namespace Math::Literals;
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
AbstractImporterTest::AbstractImporterTest() {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
addTests({&AbstractImporterTest::construct,
&AbstractImporterTest::constructWithPluginManagerReference,
Trade: add flags to AbstractImporter and AbstractImageConverter.
6 years ago
&AbstractImporterTest::setFlags,
&AbstractImporterTest::setFlagsFileOpened,
&AbstractImporterTest::setFlagsNotImplemented,
Trade: test also plain AbstractImporter::openData().
7 years ago
&AbstractImporterTest::openData,
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
&AbstractImporterTest::openDataDeprecatedFallback,
#endif
Trade: ability to open non-temporary memory. Using openMemory() instead of openData() allows the implementation to assume the data will stay in scope for as long as needed, which can prevent unnecessary copies in some plugin implementations. It warranted a new flag, DataFlag::ExternallyOwned, to describe this kind of memory. I couldn't reuse Owned as that's used for allocations owned by the instance, which is too little for certain future use cases. For example returning *Data instances referencing an Owned memory would mean the user has to assume the memory is gone when the importer instance is gone, and that's generally not true for memory passed to openMemory(). Originally I thought I would do this later, but then realized the existing plugin implementations would need to get all updated again to be aware of the new flag, with some being forgotten, and it's just easier to do the whole thing in a single step.
5 years ago
&AbstractImporterTest::openMemory,
Trade: AbstractImporter::openFile() and openState() had no tests?! I wonder what the reason was. Too simple to be tested? Coverage from testing actual plugins was enough??
4 years ago
&AbstractImporterTest::openFile,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::openFileAsData,
&AbstractImporterTest::openFileAsDataNotFound,
Trade: AbstractImporter::openFile() and openState() had no tests?! I wonder what the reason was. Too simple to be tested? Coverage from testing actual plugins was enough??
4 years ago
&AbstractImporterTest::openState,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::openFileNotImplemented,
&AbstractImporterTest::openDataNotSupported,
&AbstractImporterTest::openDataNotImplemented,
&AbstractImporterTest::openStateNotSupported,
&AbstractImporterTest::openStateNotImplemented,
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
&AbstractImporterTest::setFileCallback,
&AbstractImporterTest::setFileCallbackTemplate,
&AbstractImporterTest::setFileCallbackTemplateNull,
Trade: accept const references in AbstractImporter file callbacks.
7 years ago
&AbstractImporterTest::setFileCallbackTemplateConst,
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
&AbstractImporterTest::setFileCallbackFileOpened,
&AbstractImporterTest::setFileCallbackNotImplemented,
&AbstractImporterTest::setFileCallbackNotSupported,
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
&AbstractImporterTest::setFileCallbackOpenFileDirectly,
&AbstractImporterTest::setFileCallbackOpenFileThroughBaseImplementation,
&AbstractImporterTest::setFileCallbackOpenFileThroughBaseImplementationFailed,
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
&AbstractImporterTest::setFileCallbackOpenFileAsData,
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
&AbstractImporterTest::setFileCallbackOpenFileAsDataFailed,
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
&AbstractImporterTest::thingCountNotImplemented,
&AbstractImporterTest::thingCountNoFile,
&AbstractImporterTest::thingForNameNotImplemented,
Trade: print a message when named thing is not found. Otherwise it's hard to distinguish whether it's a particular importer fault (failing without a message) or if the name is not found.
6 years ago
&AbstractImporterTest::thingForNameNoFile});
addInstancedTests({&AbstractImporterTest::thingByNameNotFound},
Containers::arraySize(ThingByNameData));
addTests({&AbstractImporterTest::thingNameNoFile,
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
&AbstractImporterTest::thingNoFile,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::defaultScene,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::defaultSceneOutOfRange,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::defaultSceneNotImplemented,
&AbstractImporterTest::scene,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
&AbstractImporterTest::object,
#ifdef MAGNUM_BUILD_DEPRECATED
&AbstractImporterTest::sceneDeprecatedFallback2D,
&AbstractImporterTest::sceneDeprecatedFallback3D,
&AbstractImporterTest::sceneDeprecatedFallbackParentless2D,
&AbstractImporterTest::sceneDeprecatedFallbackParentless3D,
&AbstractImporterTest::sceneDeprecatedFallbackTransformless2D,
&AbstractImporterTest::sceneDeprecatedFallbackTransformless3D,
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
&AbstractImporterTest::sceneDeprecatedFallbackMultiFunctionObjects2D,
&AbstractImporterTest::sceneDeprecatedFallbackMultiFunctionObjects3D,
Trade: report at least some *D object count if everything fails. As the comment says -- before the user code expected that if the scene hierarchy is broken, particular objects will fail to import. Now the whole scene fails to import, so we don't even get to know the actual (expanded) deprecated 2D/3D object count. To reduce the suffering, return at least the dimension-less object count there. It won't include the duplicates from the single-function-object conversion but better than reporting 0.
5 years ago
&AbstractImporterTest::sceneDeprecatedFallbackObjectCountNoScenes,
&AbstractImporterTest::sceneDeprecatedFallbackObjectCountAllSceneImportFailed,
Trade: another case for the deprecated path not being 100% compatible. Honestly I don't care much, this is just that the original PrimitiveImporter tests started to fail because they expected object3DForName() to return -1 for a 2D name, but that's no longer the case. It would be possible to fix this, but I doubt anyone ever relied on such behavior for 2D scenes, so just add a test that acknowledges this new behavior.
5 years ago
&AbstractImporterTest::sceneDeprecatedFallbackBoth2DAnd3DScene,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#endif
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::sceneForNameOutOfRange,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
&AbstractImporterTest::objectForNameOutOfRange,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::sceneNameNotImplemented,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
&AbstractImporterTest::objectNameNotImplemented,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::sceneNameOutOfRange,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
&AbstractImporterTest::objectNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::sceneNameCustomDeleter,
&AbstractImporterTest::objectNameCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::sceneNotImplemented,
&AbstractImporterTest::sceneOutOfRange,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
&AbstractImporterTest::sceneNonOwningDeleters,
&AbstractImporterTest::sceneCustomDataDeleter,
&AbstractImporterTest::sceneCustomFieldDataDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
&AbstractImporterTest::sceneFieldName,
&AbstractImporterTest::sceneFieldNameNotImplemented,
&AbstractImporterTest::sceneFieldNameNotCustom,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::sceneFieldNameCustomDeleter,
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Trade: animation import interface in AbstractImporter.
8 years ago
&AbstractImporterTest::animation,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::animationForNameOutOfRange,
Trade: animation import interface in AbstractImporter.
8 years ago
&AbstractImporterTest::animationNameNotImplemented,
&AbstractImporterTest::animationNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::animationNameCustomDeleter,
Trade: animation import interface in AbstractImporter.
8 years ago
&AbstractImporterTest::animationNotImplemented,
&AbstractImporterTest::animationOutOfRange,
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
&AbstractImporterTest::animationNonOwningDeleters,
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
&AbstractImporterTest::animationGrowableDeleters,
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
&AbstractImporterTest::animationCustomDataDeleter,
&AbstractImporterTest::animationCustomTrackDeleter,
Trade: animation import interface in AbstractImporter.
8 years ago
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::light,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::lightForNameOutOfRange,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::lightNameNotImplemented,
&AbstractImporterTest::lightNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::lightNameCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::lightNotImplemented,
&AbstractImporterTest::lightOutOfRange,
&AbstractImporterTest::camera,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::cameraForNameOutOfRange,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::cameraNameNotImplemented,
&AbstractImporterTest::cameraNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::cameraNameCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::cameraNotImplemented,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
&AbstractImporterTest::cameraOutOfRange});
#ifdef MAGNUM_BUILD_DEPRECATED
addTests({&AbstractImporterTest::object2D,
&AbstractImporterTest::object2DCountNotImplemented,
&AbstractImporterTest::object2DCountNoFile,
&AbstractImporterTest::object2DForNameNotImplemented,
&AbstractImporterTest::object2DForNameNoFile,
&AbstractImporterTest::object2DForNameOutOfRange});
addInstancedTests({&AbstractImporterTest::object2DByNameNotFound},
Containers::arraySize(ThingByNameData));
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
addTests({&AbstractImporterTest::object2DNameNotImplemented,
&AbstractImporterTest::object2DNameNoFile,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::object2DNameOutOfRange,
&AbstractImporterTest::object2DNotImplemented,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
&AbstractImporterTest::object2DNoFile,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::object2DOutOfRange,
&AbstractImporterTest::object3D,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
&AbstractImporterTest::object3DCountNotImplemented,
&AbstractImporterTest::object3DCountNoFile,
&AbstractImporterTest::object3DForNameNotImplemented,
&AbstractImporterTest::object3DForNameNoFile,
&AbstractImporterTest::object3DForNameOutOfRange});
addInstancedTests({&AbstractImporterTest::object3DByNameNotFound},
Containers::arraySize(ThingByNameData));
addTests({&AbstractImporterTest::object3DNameNotImplemented,
&AbstractImporterTest::object3DNameNoFile,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::object3DNameOutOfRange,
&AbstractImporterTest::object3DNotImplemented,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
&AbstractImporterTest::object3DNoFile,
&AbstractImporterTest::object3DOutOfRange});
#endif
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
addTests({&AbstractImporterTest::skin2D,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::skin2DForNameOutOfRange,
Trade: skin import in AbstractImporter.
6 years ago
&AbstractImporterTest::skin2DNameNotImplemented,
&AbstractImporterTest::skin2DNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::skin2DNameCustomDeleter,
Trade: skin import in AbstractImporter.
6 years ago
&AbstractImporterTest::skin2DNotImplemented,
&AbstractImporterTest::skin2DOutOfRange,
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
&AbstractImporterTest::skin2DNonOwningDeleters,
&AbstractImporterTest::skin2DCustomJointDataDeleter,
&AbstractImporterTest::skin2DCustomInverseBindMatrixDataDeleter,
Trade: skin import in AbstractImporter.
6 years ago
&AbstractImporterTest::skin3D,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::skin3DForNameOutOfRange,
Trade: skin import in AbstractImporter.
6 years ago
&AbstractImporterTest::skin3DNameNotImplemented,
&AbstractImporterTest::skin3DNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::skin3DNameCustomDeleter,
Trade: skin import in AbstractImporter.
6 years ago
&AbstractImporterTest::skin3DNotImplemented,
&AbstractImporterTest::skin3DOutOfRange,
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
&AbstractImporterTest::skin3DNonOwningDeleters,
&AbstractImporterTest::skin3DCustomJointDataDeleter,
&AbstractImporterTest::skin3DCustomInverseBindMatrixDataDeleter,
Trade: skin import in AbstractImporter.
6 years ago
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
&AbstractImporterTest::mesh,
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
&AbstractImporterTest::meshDeprecatedFallback,
#endif
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
&AbstractImporterTest::meshLevelCountNotImplemented,
&AbstractImporterTest::meshLevelCountOutOfRange,
&AbstractImporterTest::meshLevelCountZero,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::meshForNameOutOfRange,
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
&AbstractImporterTest::meshNameNotImplemented,
&AbstractImporterTest::meshNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::meshNameCustomDeleter,
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
&AbstractImporterTest::meshNotImplemented,
&AbstractImporterTest::meshOutOfRange,
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
&AbstractImporterTest::meshLevelOutOfRange,
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
&AbstractImporterTest::meshNonOwningDeleters,
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
&AbstractImporterTest::meshGrowableDeleters,
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
&AbstractImporterTest::meshCustomIndexDataDeleter,
&AbstractImporterTest::meshCustomVertexDataDeleter,
&AbstractImporterTest::meshCustomAttributesDeleter,
&AbstractImporterTest::meshAttributeName,
&AbstractImporterTest::meshAttributeNameNotImplemented,
&AbstractImporterTest::meshAttributeNameNotCustom,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::meshAttributeNameCustomDeleter,
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::mesh2D,
&AbstractImporterTest::mesh2DCountNotImplemented,
&AbstractImporterTest::mesh2DCountNoFile,
&AbstractImporterTest::mesh2DForNameNotImplemented,
&AbstractImporterTest::mesh2DForNameNoFile,
&AbstractImporterTest::mesh2DNameNotImplemented,
&AbstractImporterTest::mesh2DNameNoFile,
&AbstractImporterTest::mesh2DNameOutOfRange,
&AbstractImporterTest::mesh2DNotImplemented,
&AbstractImporterTest::mesh2DNoFile,
&AbstractImporterTest::mesh2DOutOfRange,
&AbstractImporterTest::mesh3D,
&AbstractImporterTest::mesh3DCountNotImplemented,
&AbstractImporterTest::mesh3DCountNoFile,
&AbstractImporterTest::mesh3DForNameNotImplemented,
&AbstractImporterTest::mesh3DForNameNoFile,
&AbstractImporterTest::mesh3DNameNotImplemented,
&AbstractImporterTest::mesh3DNameNoFile,
&AbstractImporterTest::mesh3DNameOutOfRange,
&AbstractImporterTest::mesh3DNotImplemented,
&AbstractImporterTest::mesh3DNoFile,
&AbstractImporterTest::mesh3DOutOfRange,
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#endif
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::material,
Trade: return Optional instead of Pointer from *Importer::material(). The plugin interface version got bumped to avoid ABI issues when loading plugins that weren't updated for the change, but apart from that this shouldn't be a breaking change, as the API returns a type that can be both an Optional and a Pointer.
6 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
&AbstractImporterTest::materialDeprecatedFallback,
#endif
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::materialForNameOutOfRange,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::materialNameNotImplemented,
&AbstractImporterTest::materialNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::materialNameCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::materialNotImplemented,
&AbstractImporterTest::materialOutOfRange,
Trade: check for custom MaterialData deleters in AbstractImporter. And adapt MaterialData and the deprecated PhongMaterialData constructor to follow these rules. Basically the same as done for ImageData and MeshData.
6 years ago
&AbstractImporterTest::materialNonOwningDeleters,
&AbstractImporterTest::materialCustomAttributeDataDeleter,
&AbstractImporterTest::materialCustomLayerDataDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::texture,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::textureForNameOutOfRange,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::textureNameNotImplemented,
&AbstractImporterTest::textureNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::textureNameCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::textureNotImplemented,
&AbstractImporterTest::textureOutOfRange,
&AbstractImporterTest::image1D,
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
&AbstractImporterTest::image1DLevelCountNotImplemented,
&AbstractImporterTest::image1DLevelCountOutOfRange,
&AbstractImporterTest::image1DLevelCountZero,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::image1DForNameOutOfRange,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::image1DNameNotImplemented,
&AbstractImporterTest::image1DNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::image1DNameCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::image1DNotImplemented,
&AbstractImporterTest::image1DOutOfRange,
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
&AbstractImporterTest::image1DLevelOutOfRange,
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
&AbstractImporterTest::image1DNonOwningDeleter,
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
&AbstractImporterTest::image1DGrowableDeleter,
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
&AbstractImporterTest::image1DCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::image2D,
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
&AbstractImporterTest::image2DLevelCountNotImplemented,
&AbstractImporterTest::image2DLevelCountOutOfRange,
&AbstractImporterTest::image2DLevelCountZero,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::image2DForNameOutOfRange,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::image2DNameNotImplemented,
&AbstractImporterTest::image2DNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::image2DNameCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::image2DNotImplemented,
&AbstractImporterTest::image2DOutOfRange,
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
&AbstractImporterTest::image2DLevelOutOfRange,
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
&AbstractImporterTest::image2DNonOwningDeleter,
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
&AbstractImporterTest::image2DGrowableDeleter,
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
&AbstractImporterTest::image2DCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::image3D,
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
&AbstractImporterTest::image3DLevelCountNotImplemented,
&AbstractImporterTest::image3DLevelCountOutOfRange,
&AbstractImporterTest::image3DLevelCountZero,
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
&AbstractImporterTest::image3DForNameOutOfRange,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::image3DNameNotImplemented,
&AbstractImporterTest::image3DNameOutOfRange,
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
&AbstractImporterTest::image3DNameCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::image3DNotImplemented,
&AbstractImporterTest::image3DOutOfRange,
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
&AbstractImporterTest::image3DLevelOutOfRange,
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
&AbstractImporterTest::image3DNonOwningDeleter,
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
&AbstractImporterTest::image3DGrowableDeleter,
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
&AbstractImporterTest::image3DCustomDeleter,
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
&AbstractImporterTest::importerState,
&AbstractImporterTest::importerStateNotImplemented,
&AbstractImporterTest::importerStateNoFile,
Trade: debug output for AbstractImporter::Feature and Features.
8 years ago
&AbstractImporterTest::debugFeature,
Trade: add flags to AbstractImporter and AbstractImageConverter.
6 years ago
&AbstractImporterTest::debugFeatures,
&AbstractImporterTest::debugFlag,
&AbstractImporterTest::debugFlags});
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::construct() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
CORRADE_COMPARE(importer.features(), ImporterFeatures{});
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_VERIFY(!importer.isOpened());
importer.close();
CORRADE_VERIFY(!importer.isOpened());
}
void AbstractImporterTest::constructWithPluginManagerReference() {
PluginManager::Manager<AbstractImporter> manager;
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct Importer: AbstractImporter {
explicit Importer(PluginManager::Manager<AbstractImporter>& manager): AbstractImporter{manager} {}
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer{manager};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_VERIFY(!importer.isOpened());
}
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
Trade: add flags to AbstractImporter and AbstractImageConverter.
6 years ago
void AbstractImporterTest::setFlags() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
void doSetFlags(ImporterFlags flags) override {
_flags = flags;
}
bool doIsOpened() const override { return false; }
void doClose() override {}
ImporterFlags _flags;
} importer;
CORRADE_COMPARE(importer.flags(), ImporterFlags{});
CORRADE_COMPARE(importer._flags, ImporterFlags{});
Trade,ShaderTools: add addFlags() / clearFlags() to all plugin interfaces. There will be Flag::FlipY for images at some point, enabled by default for compatibility with existing GL code, and so it makes sense to start discouraging setFlags() as early as possible to avoid people resetting the default by accident. Also update the imageconverter, sceneconverter and shaderconverter utils to use these instead of setFlags().
5 years ago
Trade: add flags to AbstractImporter and AbstractImageConverter.
6 years ago
importer.setFlags(ImporterFlag::Verbose);
CORRADE_COMPARE(importer.flags(), ImporterFlag::Verbose);
CORRADE_COMPARE(importer._flags, ImporterFlag::Verbose);
Trade,ShaderTools: add addFlags() / clearFlags() to all plugin interfaces. There will be Flag::FlipY for images at some point, enabled by default for compatibility with existing GL code, and so it makes sense to start discouraging setFlags() as early as possible to avoid people resetting the default by accident. Also update the imageconverter, sceneconverter and shaderconverter utils to use these instead of setFlags().
5 years ago
/** @todo use a real flag when we have more than one */
importer.addFlags(ImporterFlag(4));
CORRADE_COMPARE(importer.flags(), ImporterFlag::Verbose|ImporterFlag(4));
CORRADE_COMPARE(importer._flags, ImporterFlag::Verbose|ImporterFlag(4));
importer.clearFlags(ImporterFlag::Verbose);
CORRADE_COMPARE(importer.flags(), ImporterFlag(4));
CORRADE_COMPARE(importer._flags, ImporterFlag(4));
Trade: add flags to AbstractImporter and AbstractImageConverter.
6 years ago
}
void AbstractImporterTest::setFlagsFileOpened() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.setFlags(ImporterFlag::Verbose);
Trade,ShaderTools: add addFlags() / clearFlags() to all plugin interfaces. There will be Flag::FlipY for images at some point, enabled by default for compatibility with existing GL code, and so it makes sense to start discouraging setFlags() as early as possible to avoid people resetting the default by accident. Also update the imageconverter, sceneconverter and shaderconverter utils to use these instead of setFlags().
5 years ago
importer.addFlags(ImporterFlag::Verbose);
importer.clearFlags(ImporterFlag::Verbose);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::setFlags(): can't be set while a file is opened\n"
/* These all call into setFlags(), so the same assert is reused */
"Trade::AbstractImporter::setFlags(): can't be set while a file is opened\n"
"Trade::AbstractImporter::setFlags(): can't be set while a file is opened\n");
Trade: add flags to AbstractImporter and AbstractImageConverter.
6 years ago
}
Trade: match test implementation order with run order.
5 years ago
void AbstractImporterTest::setFlagsNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
CORRADE_COMPARE(importer.flags(), ImporterFlags{});
importer.setFlags(ImporterFlag::Verbose);
CORRADE_COMPARE(importer.flags(), ImporterFlag::Verbose);
/* Should just work, no need to implement the function */
}
Trade: test also plain AbstractImporter::openData().
7 years ago
void AbstractImporterTest::openData() {
Trade: minor cleanup.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData; }
Trade: test also plain AbstractImporter::openData().
7 years ago
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
void doOpenData(Containers::Array<char>&& data, DataFlags dataFlags) override {
CORRADE_COMPARE_AS(data,
Containers::arrayView({'\xa5'}),
TestSuite::Compare::Container);
CORRADE_COMPARE(dataFlags, DataFlags{});
/* The array should have a custom no-op deleter */
CORRADE_VERIFY(data.deleter());
_opened = true;
}
bool _opened = false;
} importer;
CORRADE_VERIFY(!importer.isOpened());
const char a5 = '\xa5';
CORRADE_VERIFY(importer.openData({&a5, 1}));
CORRADE_VERIFY(importer.isOpened());
importer.close();
CORRADE_VERIFY(!importer.isOpened());
}
#ifdef MAGNUM_BUILD_DEPRECATED
void AbstractImporterTest::openDataDeprecatedFallback() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData; }
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
Trade: test also plain AbstractImporter::openData().
7 years ago
void doOpenData(Containers::ArrayView<const char> data) override {
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
CORRADE_COMPARE_AS(data,
Containers::arrayView({'\xa5'}),
TestSuite::Compare::Container);
_opened = true;
Trade: test also plain AbstractImporter::openData().
7 years ago
}
bool _opened = false;
} importer;
CORRADE_VERIFY(!importer.isOpened());
const char a5 = '\xa5';
CORRADE_VERIFY(importer.openData({&a5, 1}));
CORRADE_VERIFY(importer.isOpened());
importer.close();
CORRADE_VERIFY(!importer.isOpened());
}
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
#endif
Trade: test also plain AbstractImporter::openData().
7 years ago
Trade: ability to open non-temporary memory. Using openMemory() instead of openData() allows the implementation to assume the data will stay in scope for as long as needed, which can prevent unnecessary copies in some plugin implementations. It warranted a new flag, DataFlag::ExternallyOwned, to describe this kind of memory. I couldn't reuse Owned as that's used for allocations owned by the instance, which is too little for certain future use cases. For example returning *Data instances referencing an Owned memory would mean the user has to assume the memory is gone when the importer instance is gone, and that's generally not true for memory passed to openMemory(). Originally I thought I would do this later, but then realized the existing plugin implementations would need to get all updated again to be aware of the new flag, with some being forgotten, and it's just easier to do the whole thing in a single step.
5 years ago
void AbstractImporterTest::openMemory() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData; }
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
void doOpenData(Containers::Array<char>&& data, DataFlags dataFlags) override {
CORRADE_COMPARE_AS(data,
Containers::arrayView({'\xa5'}),
TestSuite::Compare::Container);
CORRADE_COMPARE(dataFlags, DataFlag::ExternallyOwned);
/* The array should have a custom no-op deleter */
CORRADE_VERIFY(data.deleter());
_opened = true;
}
bool _opened = false;
} importer;
CORRADE_VERIFY(!importer.isOpened());
const char a5 = '\xa5';
CORRADE_VERIFY(importer.openMemory({&a5, 1}));
CORRADE_VERIFY(importer.isOpened());
importer.close();
CORRADE_VERIFY(!importer.isOpened());
}
Trade: AbstractImporter::openFile() and openState() had no tests?! I wonder what the reason was. Too simple to be tested? Coverage from testing actual plugins was enough??
4 years ago
void AbstractImporterTest::openFile() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
void doOpenFile(Containers::StringView filename) override {
CORRADE_COMPARE(filename, "yello.foo");
_opened = true;
}
bool _opened = false;
} importer;
CORRADE_VERIFY(!importer.isOpened());
CORRADE_VERIFY(importer.openFile("yello.foo"));
CORRADE_VERIFY(importer.isOpened());
importer.close();
CORRADE_VERIFY(!importer.isOpened());
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::openFileAsData() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
void doOpenData(Containers::Array<char>&& data, DataFlags dataFlags) override {
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
CORRADE_COMPARE_AS(data,
Containers::arrayView({'\xa5'}),
TestSuite::Compare::Container);
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
CORRADE_COMPARE(dataFlags, DataFlag::Owned|DataFlag::Mutable);
/* I.e., we can take over the array, it's not just a view */
CORRADE_VERIFY(!data.deleter());
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
_opened = true;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
bool _opened = false;
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
/* doOpenFile() should call doOpenData() */
CORRADE_VERIFY(!importer.isOpened());
Trade: port to the new Utility::Path. For file opening there's no longer an unatomic pair of exists() + read(), but since Path::read() now returns an Optional, it means we can reliably distinguish between empty files and failures. While at it, also added TODOs for removal of the StringStl.h header that's needed in various places for compatibility with APIs still using STL strings.
4 years ago
CORRADE_VERIFY(importer.openFile(Utility::Path::join(TRADE_TEST_DIR, "file.bin")));
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
CORRADE_VERIFY(importer.isOpened());
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.close();
CORRADE_VERIFY(!importer.isOpened());
}
void AbstractImporterTest::openFileAsDataNotFound() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct Importer: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
void doOpenData(Containers::Array<char>&&, DataFlags) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
_opened = true;
}
bool _opened = false;
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
CORRADE_VERIFY(!importer.openFile("nonexistent.bin"));
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_VERIFY(!importer.isOpened());
Trade: port to the new Utility::Path. For file opening there's no longer an unatomic pair of exists() + read(), but since Path::read() now returns an Optional, it means we can reliably distinguish between empty files and failures. While at it, also added TODOs for removal of the StringStl.h header that's needed in various places for compatibility with APIs still using STL strings.
4 years ago
/* There's an error message from Path::read() before */
CORRADE_COMPARE_AS(out.str(),
"\nTrade::AbstractImporter::openFile(): cannot open file nonexistent.bin\n",
TestSuite::Compare::StringHasSuffix);
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: AbstractImporter::openFile() and openState() had no tests?! I wonder what the reason was. Too simple to be tested? Coverage from testing actual plugins was enough??
4 years ago
void AbstractImporterTest::openState() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override {
return ImporterFeature::OpenState;
}
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
void doOpenState(const void* state, Containers::StringView filePath) override {
CORRADE_COMPARE(state, reinterpret_cast<const void*>(0xbadcafe));
CORRADE_COMPARE(filePath, "yello/foo/");
_opened = true;
}
bool _opened = false;
} importer;
CORRADE_VERIFY(!importer.isOpened());
CORRADE_VERIFY(importer.openState(reinterpret_cast<const void*>(0xbadcafe), "yello/foo/"));
CORRADE_VERIFY(importer.isOpened());
importer.close();
CORRADE_VERIFY(!importer.isOpened());
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::openFileNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct Importer: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
CORRADE_VERIFY(!importer.openFile("file.dat"));
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::openFile(): not implemented\n");
}
void AbstractImporterTest::openDataNotSupported() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
CORRADE_VERIFY(!importer.openData(nullptr));
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::openData(): feature not supported\n");
}
void AbstractImporterTest::openDataNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
CORRADE_VERIFY(!importer.openData(nullptr));
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::openData(): feature advertised but not implemented\n");
}
void AbstractImporterTest::openStateNotSupported() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
CORRADE_VERIFY(!importer.openState(nullptr));
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::openState(): feature not supported\n");
}
void AbstractImporterTest::openStateNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenState; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
CORRADE_VERIFY(!importer.openState(nullptr));
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::openState(): feature advertised but not implemented\n");
}
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
void AbstractImporterTest::setFileCallback() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData|ImporterFeature::FileCallback; }
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
void doSetFileCallback(Containers::Optional<Containers::ArrayView<const char>>(*)(const std::string&, InputFileCallbackPolicy, void*), void* userData) override {
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
*static_cast<int*>(userData) = 1337;
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
int a = 0;
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
auto lambda = [](const std::string&, InputFileCallbackPolicy, void*) {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return Containers::Optional<Containers::ArrayView<const char>>{};
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
};
importer.setFileCallback(lambda, &a);
CORRADE_COMPARE(importer.fileCallback(), lambda);
CORRADE_COMPARE(importer.fileCallbackUserData(), &a);
CORRADE_COMPARE(a, 1337);
}
void AbstractImporterTest::setFileCallbackTemplate() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData|ImporterFeature::FileCallback; }
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
void doSetFileCallback(Containers::Optional<Containers::ArrayView<const char>>(*)(const std::string&, InputFileCallbackPolicy, void*), void*) override {
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
called = true;
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
bool called = false;
} importer;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
int a = 0;
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
auto lambda = [](const std::string&, InputFileCallbackPolicy, int&) {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return Containers::Optional<Containers::ArrayView<const char>>{};
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
};
importer.setFileCallback(lambda, a);
CORRADE_VERIFY(importer.fileCallback());
CORRADE_VERIFY(importer.fileCallbackUserData());
CORRADE_VERIFY(importer.called);
/* The data pointers should be wrapped, thus not the same */
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
CORRADE_VERIFY(reinterpret_cast<void(*)()>(importer.fileCallback()) != reinterpret_cast<void(*)()>(static_cast<Containers::Optional<Containers::ArrayView<const char>>(*)(const std::string&, InputFileCallbackPolicy, int&)>(lambda)));
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
CORRADE_VERIFY(importer.fileCallbackUserData() != &a);
}
void AbstractImporterTest::setFileCallbackTemplateNull() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData|ImporterFeature::FileCallback; }
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
void doSetFileCallback(Containers::Optional<Containers::ArrayView<const char>>(*callback)(const std::string&, InputFileCallbackPolicy, void*), void* userData) override {
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
CORRADE_VERIFY(!callback);
CORRADE_VERIFY(!userData);
called = true;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
bool called = false;
} importer;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
CORRADE_VERIFY(true); /* to record correct function name */
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
int a = 0;
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
importer.setFileCallback(static_cast<Containers::Optional<Containers::ArrayView<const char>>(*)(const std::string&, InputFileCallbackPolicy, int&)>(nullptr), a);
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
CORRADE_VERIFY(!importer.fileCallback());
CORRADE_VERIFY(!importer.fileCallbackUserData());
CORRADE_VERIFY(importer.called);
}
Trade: accept const references in AbstractImporter file callbacks.
7 years ago
void AbstractImporterTest::setFileCallbackTemplateConst() {
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData|ImporterFeature::FileCallback; }
Trade: accept const references in AbstractImporter file callbacks.
7 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
void doSetFileCallback(Containers::Optional<Containers::ArrayView<const char>>(*)(const std::string&, InputFileCallbackPolicy, void*), void*) override {
called = true;
}
bool called = false;
} importer;
/* Just verify we can have const parameters */
const int a = 0;
auto lambda = [](const std::string&, InputFileCallbackPolicy, const int&) {
return Containers::Optional<Containers::ArrayView<const char>>{};
};
importer.setFileCallback(lambda, a);
CORRADE_VERIFY(importer.fileCallback());
CORRADE_VERIFY(importer.fileCallbackUserData());
CORRADE_VERIFY(importer.called);
}
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
void AbstractImporterTest::setFileCallbackFileOpened() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
importer.setFileCallback([](const std::string&, InputFileCallbackPolicy, void*) {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return Containers::Optional<Containers::ArrayView<const char>>{};
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
});
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::setFileCallback(): can't be set while a file is opened\n");
}
void AbstractImporterTest::setFileCallbackNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::FileCallback; }
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
int a;
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
auto lambda = [](const std::string&, InputFileCallbackPolicy, void*) {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return Containers::Optional<Containers::ArrayView<const char>>{};
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
};
importer.setFileCallback(lambda, &a);
CORRADE_COMPARE(importer.fileCallback(), lambda);
CORRADE_COMPARE(importer.fileCallbackUserData(), &a);
/* Should just work, no need to implement the function */
}
void AbstractImporterTest::setFileCallbackNotSupported() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
std::ostringstream out;
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
Error redirectError{&out};
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
int a;
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
importer.setFileCallback([](const std::string&, InputFileCallbackPolicy, void*) {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return Containers::Optional<Containers::ArrayView<const char>>{};
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
}, &a);
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::setFileCallback(): importer supports neither loading from data nor via callbacks, callbacks can't be used\n");
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
}
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
void AbstractImporterTest::setFileCallbackOpenFileDirectly() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::FileCallback|ImporterFeature::OpenData; }
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void doOpenFile(Containers::StringView filename) override {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
/* Called because FileCallback is supported */
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
CORRADE_COMPARE(filename, "file.dat");
CORRADE_VERIFY(fileCallback());
CORRADE_VERIFY(fileCallbackUserData());
_opened = true;
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
}
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
void doOpenData(Containers::Array<char>&&, DataFlags) override {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
/* Shouldn't be called because FileCallback is supported */
openDataCalledNotSureWhy = true;
}
bool _opened = false;
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
bool openDataCalledNotSureWhy = false;
} importer;
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
bool calledNotSureWhy = false;
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
importer.setFileCallback([](const std::string&, InputFileCallbackPolicy, bool& calledNotSureWhy) -> Containers::Optional<Containers::ArrayView<const char>> {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
calledNotSureWhy = true;
return {};
}, calledNotSureWhy);
CORRADE_VERIFY(importer.openFile("file.dat"));
CORRADE_VERIFY(!calledNotSureWhy);
CORRADE_VERIFY(!importer.openDataCalledNotSureWhy);
}
void AbstractImporterTest::setFileCallbackOpenFileThroughBaseImplementation() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::FileCallback|ImporterFeature::OpenData; }
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void doOpenFile(Containers::StringView filename) override {
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
CORRADE_COMPARE(filename, "file.dat");
CORRADE_VERIFY(fileCallback());
CORRADE_VERIFY(fileCallbackUserData());
openFileCalled = true;
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
AbstractImporter::doOpenFile(filename);
}
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
void doOpenData(Containers::Array<char>&& data, DataFlags dataFlags) override {
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
CORRADE_COMPARE_AS(data,
Containers::arrayView({'\xb0'}),
TestSuite::Compare::Container);
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
CORRADE_COMPARE(dataFlags, DataFlags{});
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
_opened = true;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
}
bool _opened = false;
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
bool openFileCalled = false;
} importer;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
struct State {
const char data = '\xb0';
bool loaded = false;
bool closed = false;
bool calledNotSureWhy = false;
} state;
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
importer.setFileCallback([](const std::string& filename, InputFileCallbackPolicy policy, State& state) -> Containers::Optional<Containers::ArrayView<const char>> {
if(filename == "file.dat" && policy == InputFileCallbackPolicy::LoadTemporary) {
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
state.loaded = true;
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return Containers::arrayView(&state.data, 1);
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
}
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
if(filename == "file.dat" && policy == InputFileCallbackPolicy::Close) {
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
state.closed = true;
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return {};
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
}
state.calledNotSureWhy = true;
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return {};
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
}, state);
CORRADE_VERIFY(importer.openFile("file.dat"));
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
CORRADE_VERIFY(importer.openFileCalled);
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
CORRADE_VERIFY(state.loaded);
CORRADE_VERIFY(state.closed);
CORRADE_VERIFY(!state.calledNotSureWhy);
}
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
void AbstractImporterTest::setFileCallbackOpenFileThroughBaseImplementationFailed() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::FileCallback|ImporterFeature::OpenData; }
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void doOpenFile(Containers::StringView filename) override {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
openFileCalled = true;
AbstractImporter::doOpenFile(filename);
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
bool openFileCalled = false;
} importer;
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
importer.setFileCallback([](const std::string&, InputFileCallbackPolicy, void*) -> Containers::Optional<Containers::ArrayView<const char>> {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return {};
});
std::ostringstream out;
Error redirectError{&out};
CORRADE_VERIFY(!importer.openFile("file.dat"));
CORRADE_VERIFY(importer.openFileCalled);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::openFile(): cannot open file file.dat\n");
}
void AbstractImporterTest::setFileCallbackOpenFileAsData() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData; }
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
bool doIsOpened() const override { return _opened; }
void doClose() override { _opened = false; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void doOpenFile(Containers::StringView) override {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
openFileCalled = true;
}
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
void doOpenData(Containers::Array<char>&& data, DataFlags dataFlags) override {
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
CORRADE_COMPARE_AS(data,
Containers::arrayView({'\xb0'}),
TestSuite::Compare::Container);
Trade: AbstractImporter::doOpenData() now takes a rvalue Array. This allows to better describe memory ownership and transfer it instead of forcing the plugins to allocate their own local copy if the import happens in-place on the imported data. Right now that's mainly for the openFile() use case, which implicitly allocated an Array with file contents only to pass it to openData() which then made a copy because it could not make any assumption about data scope. In other words, certain plugins (TgaImporter, KtxImporter, DdsImporter, CgltfImporter and possibly others) will now have their peak memory usage *halved*.
5 years ago
CORRADE_COMPARE(dataFlags, DataFlags{});
Trade: use a better-debuggable way to test passed parameters. The original tests were implemented back when it was only possible to use the verification macros from within test cases.
5 years ago
_opened = true;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
}
bool _opened = false;
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
bool openFileCalled = false;
} importer;
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
struct State {
const char data = '\xb0';
bool loaded = false;
bool closed = false;
bool calledNotSureWhy = false;
} state;
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
importer.setFileCallback([](const std::string& filename, InputFileCallbackPolicy policy, State& state) -> Containers::Optional<Containers::ArrayView<const char>> {
if(filename == "file.dat" && policy == InputFileCallbackPolicy::LoadTemporary) {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
state.loaded = true;
return Containers::arrayView(&state.data, 1);
}
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
if(filename == "file.dat" && policy == InputFileCallbackPolicy::Close) {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
state.closed = true;
return {};
}
state.calledNotSureWhy = true;
return {};
}, state);
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
CORRADE_VERIFY(importer.openFile("file.dat"));
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
CORRADE_VERIFY(!importer.openFileCalled);
CORRADE_VERIFY(state.loaded);
CORRADE_VERIFY(state.closed);
CORRADE_VERIFY(!state.calledNotSureWhy);
}
void AbstractImporterTest::setFileCallbackOpenFileAsDataFailed() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return ImporterFeature::OpenData; }
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void doOpenFile(Containers::StringView) override {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
openFileCalled = true;
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
bool openFileCalled = false;
} importer;
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
Move Trade::ImporterFileCallbackPolicy to InputFileCallbackPolicy. It'll get used outside of the root namespace and since the callbacks tend to be quite complex, it would be silly to require users to implement one callback for Trade, one for Text and one for Audio, for example.
7 years ago
importer.setFileCallback([](const std::string&, InputFileCallbackPolicy, void*) {
Trade: redesign importer file callbacks. The original implementation had a few problems: - If a file callback was set, openFile() was unconditionally calling right into doOpenData(), making it impossible for the importer to know the original path for correctly supplying paths to additional files. Now, if the importer supports Feature::FileCallback, doOpenFile() is always called. It's also possible for the importer to save the path and then just delegate to the base doOpenFile() implementation -- it will handle the file callbacks correctly too. - If the importer supported neither FileCallback nor OpenData and callbacks were set, the original doOpenFile() implementation was called without any warning or anything, doing silently a bad thing. Now in this case setFileCallbacks() asserts -- programmer has to check for feature support first. - It was not possible for the file callback to indicate file opening failure -- in general, empty files are valid, so a nullptr ArrayView is also a valid file. Now the callback return an Optional instead.
8 years ago
return Containers::Optional<Containers::ArrayView<const char>>{};
});
std::ostringstream out;
Error redirectError{&out};
CORRADE_VERIFY(!importer.openFile("file.dat"));
CORRADE_VERIFY(!importer.openFileCalled);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::openFile(): cannot open file file.dat\n");
Trade: support for file loading callbacks in AbstractImporter.
8 years ago
}
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
void AbstractImporterTest::thingCountNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
CORRADE_COMPARE(importer.sceneCount(), 0);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(importer.objectCount(), 0);
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
CORRADE_COMPARE(importer.animationCount(), 0);
CORRADE_COMPARE(importer.lightCount(), 0);
CORRADE_COMPARE(importer.cameraCount(), 0);
Trade: skin import in AbstractImporter.
6 years ago
CORRADE_COMPARE(importer.skin2DCount(), 0);
CORRADE_COMPARE(importer.skin3DCount(), 0);
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
CORRADE_COMPARE(importer.meshCount(), 0);
CORRADE_COMPARE(importer.materialCount(), 0);
CORRADE_COMPARE(importer.textureCount(), 0);
CORRADE_COMPARE(importer.image1DCount(), 0);
CORRADE_COMPARE(importer.image2DCount(), 0);
CORRADE_COMPARE(importer.image3DCount(), 0);
}
void AbstractImporterTest::thingCountNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
std::ostringstream out;
Error redirectError{&out};
importer.sceneCount();
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
importer.objectCount();
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
importer.animationCount();
importer.lightCount();
importer.cameraCount();
Trade: skin import in AbstractImporter.
6 years ago
importer.skin2DCount();
importer.skin3DCount();
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
importer.meshCount();
importer.meshLevelCount(7);
importer.materialCount();
importer.textureCount();
importer.image1DCount();
importer.image1DLevelCount(7);
importer.image2DCount();
importer.image2DLevelCount(7);
importer.image3DCount();
importer.image3DLevelCount(7);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::sceneCount(): no file opened\n"
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
"Trade::AbstractImporter::objectCount(): no file opened\n"
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
"Trade::AbstractImporter::animationCount(): no file opened\n"
"Trade::AbstractImporter::lightCount(): no file opened\n"
"Trade::AbstractImporter::cameraCount(): no file opened\n"
Trade: skin import in AbstractImporter.
6 years ago
"Trade::AbstractImporter::skin2DCount(): no file opened\n"
"Trade::AbstractImporter::skin3DCount(): no file opened\n"
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
"Trade::AbstractImporter::meshCount(): no file opened\n"
"Trade::AbstractImporter::meshLevelCount(): no file opened\n"
"Trade::AbstractImporter::materialCount(): no file opened\n"
"Trade::AbstractImporter::textureCount(): no file opened\n"
"Trade::AbstractImporter::image1DCount(): no file opened\n"
"Trade::AbstractImporter::image1DLevelCount(): no file opened\n"
"Trade::AbstractImporter::image2DCount(): no file opened\n"
"Trade::AbstractImporter::image2DLevelCount(): no file opened\n"
"Trade::AbstractImporter::image3DCount(): no file opened\n"
"Trade::AbstractImporter::image3DLevelCount(): no file opened\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
void AbstractImporterTest::thingForNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
CORRADE_COMPARE(importer.sceneForName(""), -1);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(importer.objectForName(""), -1);
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
CORRADE_COMPARE(importer.animationForName(""), -1);
CORRADE_COMPARE(importer.lightForName(""), -1);
CORRADE_COMPARE(importer.cameraForName(""), -1);
Trade: skin import in AbstractImporter.
6 years ago
CORRADE_COMPARE(importer.skin2DForName(""), -1);
CORRADE_COMPARE(importer.skin3DForName(""), -1);
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
CORRADE_COMPARE(importer.meshForName(""), -1);
CORRADE_COMPARE(importer.materialForName(""), -1);
CORRADE_COMPARE(importer.textureForName(""), -1);
CORRADE_COMPARE(importer.image1DForName(""), -1);
CORRADE_COMPARE(importer.image2DForName(""), -1);
CORRADE_COMPARE(importer.image3DForName(""), -1);
}
void AbstractImporterTest::thingForNameNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.sceneForName("");
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
importer.objectForName("");
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
importer.animationForName("");
importer.lightForName("");
importer.cameraForName("");
Trade: skin import in AbstractImporter.
6 years ago
importer.skin2DForName("");
importer.skin3DForName("");
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
importer.meshForName("");
importer.materialForName("");
importer.textureForName("");
importer.image1DForName("");
importer.image2DForName("");
importer.image3DForName("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::sceneForName(): no file opened\n"
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
"Trade::AbstractImporter::objectForName(): no file opened\n"
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
"Trade::AbstractImporter::animationForName(): no file opened\n"
"Trade::AbstractImporter::lightForName(): no file opened\n"
"Trade::AbstractImporter::cameraForName(): no file opened\n"
Trade: skin import in AbstractImporter.
6 years ago
"Trade::AbstractImporter::skin2DForName(): no file opened\n"
"Trade::AbstractImporter::skin3DForName(): no file opened\n"
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
"Trade::AbstractImporter::meshForName(): no file opened\n"
"Trade::AbstractImporter::materialForName(): no file opened\n"
"Trade::AbstractImporter::textureForName(): no file opened\n"
"Trade::AbstractImporter::image1DForName(): no file opened\n"
"Trade::AbstractImporter::image2DForName(): no file opened\n"
"Trade::AbstractImporter::image3DForName(): no file opened\n");
}
Trade: print a message when named thing is not found. Otherwise it's hard to distinguish whether it's a particular importer fault (failing without a message) or if the name is not found.
6 years ago
void AbstractImporterTest::thingByNameNotFound() {
auto&& data = ThingByNameData[testCaseInstanceId()];
setTestCaseDescription(data.name);
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: make importer by-name-lookup failure messages a bit more helpful. For example if we're looking for a 2D image named "cubemap", the importer will tell us there's 0 2D images, making a potential mistake (2D vs 3D) more obvious.
5 years ago
UnsignedInt doSceneCount() const override { return 1; }
UnsignedInt doAnimationCount() const override { return 2; }
UnsignedInt doLightCount() const override { return 3; }
UnsignedInt doCameraCount() const override { return 4; }
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedInt doSkin2DCount() const override { return 5; }
UnsignedInt doSkin3DCount() const override { return 6; }
Trade: make importer by-name-lookup failure messages a bit more helpful. For example if we're looking for a 2D image named "cubemap", the importer will tell us there's 0 2D images, making a potential mistake (2D vs 3D) more obvious.
5 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedInt doMeshCount() const override { return 7; }
UnsignedInt doMaterialCount() const override { return 8; }
UnsignedInt doTextureCount() const override { return 9; }
Trade: make importer by-name-lookup failure messages a bit more helpful. For example if we're looking for a 2D image named "cubemap", the importer will tell us there's 0 2D images, making a potential mistake (2D vs 3D) more obvious.
5 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedInt doImage1DCount() const override { return 10; }
UnsignedInt doImage2DCount() const override { return 11; }
UnsignedInt doImage3DCount() const override { return 12; }
Trade: print a message when named thing is not found. Otherwise it's hard to distinguish whether it's a particular importer fault (failing without a message) or if the name is not found.
6 years ago
} importer;
std::ostringstream out;
{
Containers::Optional<Error> redirectError;
if(data.checkMessage) redirectError.emplace(&out);
CORRADE_VERIFY(!importer.scene("foobar"));
CORRADE_VERIFY(!importer.animation("foobar"));
CORRADE_VERIFY(!importer.light("foobar"));
CORRADE_VERIFY(!importer.camera("foobar"));
Trade: skin import in AbstractImporter.
6 years ago
CORRADE_VERIFY(!importer.skin2D("foobar"));
CORRADE_VERIFY(!importer.skin3D("foobar"));
Trade: print a message when named thing is not found. Otherwise it's hard to distinguish whether it's a particular importer fault (failing without a message) or if the name is not found.
6 years ago
CORRADE_VERIFY(!importer.mesh("foobar"));
CORRADE_VERIFY(!importer.material("foobar"));
CORRADE_VERIFY(!importer.texture("foobar"));
CORRADE_VERIFY(!importer.image1D("foobar"));
CORRADE_VERIFY(!importer.image2D("foobar"));
CORRADE_VERIFY(!importer.image3D("foobar"));
}
if(data.checkMessage) {
CORRADE_COMPARE(out.str(),
Trade: make importer by-name-lookup failure messages a bit more helpful. For example if we're looking for a 2D image named "cubemap", the importer will tell us there's 0 2D images, making a potential mistake (2D vs 3D) more obvious.
5 years ago
"Trade::AbstractImporter::scene(): scene foobar not found among 1 entries\n"
"Trade::AbstractImporter::animation(): animation foobar not found among 2 entries\n"
"Trade::AbstractImporter::light(): light foobar not found among 3 entries\n"
"Trade::AbstractImporter::camera(): camera foobar not found among 4 entries\n"
Trade: print a message when named thing is not found. Otherwise it's hard to distinguish whether it's a particular importer fault (failing without a message) or if the name is not found.
6 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
"Trade::AbstractImporter::skin2D(): skin foobar not found among 5 entries\n"
"Trade::AbstractImporter::skin3D(): skin foobar not found among 6 entries\n"
Trade: print a message when named thing is not found. Otherwise it's hard to distinguish whether it's a particular importer fault (failing without a message) or if the name is not found.
6 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
"Trade::AbstractImporter::mesh(): mesh foobar not found among 7 entries\n"
"Trade::AbstractImporter::material(): material foobar not found among 8 entries\n"
"Trade::AbstractImporter::texture(): texture foobar not found among 9 entries\n"
Trade: skin import in AbstractImporter.
6 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
"Trade::AbstractImporter::image1D(): image foobar not found among 10 entries\n"
"Trade::AbstractImporter::image2D(): image foobar not found among 11 entries\n"
"Trade::AbstractImporter::image3D(): image foobar not found among 12 entries\n");
Trade: print a message when named thing is not found. Otherwise it's hard to distinguish whether it's a particular importer fault (failing without a message) or if the name is not found.
6 years ago
}
}
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
void AbstractImporterTest::thingNameNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.sceneName(42);
importer.animationName(42);
importer.lightName(42);
importer.cameraName(42);
Trade: skin import in AbstractImporter.
6 years ago
importer.skin2DName(42);
importer.skin3DName(42);
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
importer.meshName(42);
importer.materialName(42);
importer.textureName(42);
importer.image1DName(42);
importer.image2DName(42);
importer.image3DName(42);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::sceneName(): no file opened\n"
"Trade::AbstractImporter::animationName(): no file opened\n"
"Trade::AbstractImporter::lightName(): no file opened\n"
"Trade::AbstractImporter::cameraName(): no file opened\n"
Trade: skin import in AbstractImporter.
6 years ago
"Trade::AbstractImporter::skin2DName(): no file opened\n"
"Trade::AbstractImporter::skin3DName(): no file opened\n"
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
"Trade::AbstractImporter::meshName(): no file opened\n"
"Trade::AbstractImporter::materialName(): no file opened\n"
"Trade::AbstractImporter::textureName(): no file opened\n"
"Trade::AbstractImporter::image1DName(): no file opened\n"
"Trade::AbstractImporter::image2DName(): no file opened\n"
"Trade::AbstractImporter::image3DName(): no file opened\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
void AbstractImporterTest::thingNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
importer.defaultScene();
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
importer.scene(42);
importer.scene("foo");
importer.animation(42);
importer.animation("foo");
importer.light(42);
importer.light("foo");
importer.camera(42);
importer.camera("foo");
Trade: skin import in AbstractImporter.
6 years ago
importer.skin2D(42);
importer.skin2D("foo");
importer.skin3D(42);
importer.skin3D("foo");
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
importer.mesh(42);
importer.mesh("foo");
importer.material(42);
importer.material("foo");
importer.texture(42);
importer.texture("foo");
importer.image1D(42);
importer.image1D("foo");
importer.image2D(42);
importer.image2D("foo");
importer.image3D(42);
importer.image3D("foo");
Trade: AbstractImporter::importerState() expects a file to be opened. The restriction was there always, but was undocumented and untested.
6 years ago
importer.importerState();
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::defaultScene(): no file opened\n"
"Trade::AbstractImporter::scene(): no file opened\n"
"Trade::AbstractImporter::scene(): no file opened\n"
"Trade::AbstractImporter::animation(): no file opened\n"
"Trade::AbstractImporter::animation(): no file opened\n"
"Trade::AbstractImporter::light(): no file opened\n"
"Trade::AbstractImporter::light(): no file opened\n"
"Trade::AbstractImporter::camera(): no file opened\n"
"Trade::AbstractImporter::camera(): no file opened\n"
Trade: skin import in AbstractImporter.
6 years ago
"Trade::AbstractImporter::skin2D(): no file opened\n"
"Trade::AbstractImporter::skin2D(): no file opened\n"
"Trade::AbstractImporter::skin3D(): no file opened\n"
"Trade::AbstractImporter::skin3D(): no file opened\n"
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
"Trade::AbstractImporter::mesh(): no file opened\n"
"Trade::AbstractImporter::mesh(): no file opened\n"
"Trade::AbstractImporter::material(): no file opened\n"
"Trade::AbstractImporter::material(): no file opened\n"
"Trade::AbstractImporter::texture(): no file opened\n"
"Trade::AbstractImporter::texture(): no file opened\n"
"Trade::AbstractImporter::image1D(): no file opened\n"
"Trade::AbstractImporter::image1D(): no file opened\n"
"Trade::AbstractImporter::image2D(): no file opened\n"
"Trade::AbstractImporter::image2D(): no file opened\n"
"Trade::AbstractImporter::image3D(): no file opened\n"
Trade: AbstractImporter::importerState() expects a file to be opened. The restriction was there always, but was undocumented and untested.
6 years ago
"Trade::AbstractImporter::image3D(): no file opened\n"
"Trade::AbstractImporter::importerState(): no file opened\n");
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
}
void AbstractImporterTest::defaultScene() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
UnsignedInt doSceneCount() const override { return 43; }
Trade: make AbstractImporter::defaultScene() const. Makes more sense as the function isn't expected to fail (and thus any kind of lazy population is not possible as it would be too late for error checks anyway). *Not* updating interface strings even though this is an ABI break because we're doing that right after the skin import interface bump.
6 years ago
Int doDefaultScene() const override { return 42; }
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
} importer;
CORRADE_COMPARE(importer.defaultScene(), 42);
}
void AbstractImporterTest::defaultSceneNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
} importer;
CORRADE_COMPARE(importer.defaultScene(), -1);
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::defaultSceneOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 8; }
Int doDefaultScene() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.defaultScene();
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::defaultScene(): implementation-returned index 8 out of range for 8 entries\n");
}
Wrap all tests in unnamed namespaces. This makes the compiler warn about test cases that aren't ever used, which is a good thing.
7 years ago
int state;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::scene() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSceneForName(Containers::StringView name) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doSceneName(UnsignedInt id) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Containers::Optional<SceneData> doScene(UnsignedInt id) override {
Trade: initial redesign of the SceneData class. No backwards compatibility in place yet, will get added back in the followup commits, together with proper documentation.
5 years ago
if(id == 7)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedByte, 0, nullptr, {}, &state};
return SceneData{SceneMappingType::UnsignedByte, 0, nullptr, {}};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.sceneCount(), 8);
CORRADE_COMPARE(importer.sceneForName("eighth"), 7);
CORRADE_COMPARE(importer.sceneName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.scene(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.scene("eighth");
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void AbstractImporterTest::object() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedLong doObjectCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Long doObjectForName(Containers::StringView name) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(name == "eighth") return 7;
return -1;
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doObjectName(UnsignedLong id) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(id == 7) return "eighth";
return {};
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(importer.objectCount(), 8);
CORRADE_COMPARE(importer.objectForName("eighth"), 7);
CORRADE_COMPARE(importer.objectName(7), "eighth");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void AbstractImporterTest::sceneForNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
bool doIsOpened() const override { return true; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void doClose() override {}
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
UnsignedInt doSceneCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSceneForName(Containers::StringView) override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
importer.sceneForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::sceneForName(): implementation-returned index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void AbstractImporterTest::objectForNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedLong doObjectCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Long doObjectForName(Containers::StringView) override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
importer.objectForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::objectForName(): implementation-returned index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
void AbstractImporterTest::sceneDeprecatedFallback2D() {
/* Need to test the following combinations:
- few objects in the root
- an object with one child, with more than one, with none
- an object with a mesh and a material
- an object with a mesh and without a material
- an object with a mesh and a skin
- an object with a skin but no mesh
- an object with a camera
- an object with TRS transformation
- an object with TRS transformation and a mesh
- an object with nothing except parent / transformation
*/
struct Transform {
UnsignedInt object;
Int parent;
Matrix3 transformation;
};
struct Trs {
UnsignedInt object;
Vector2 translation;
Complex rotation;
Vector2 scaling;
};
struct Mesh {
UnsignedInt object;
UnsignedShort mesh;
Short meshMaterial;
};
struct Index {
UnsignedInt object;
UnsignedInt id;
};
struct ImporterState {
UnsignedInt object;
const void* importerState;
};
Containers::StridedArrayView1D<Transform> transformations;
Containers::StridedArrayView1D<Trs> trs;
Containers::StridedArrayView1D<Mesh> meshes;
Containers::StridedArrayView1D<Index> cameras;
Containers::StridedArrayView1D<Index> skins;
Containers::StridedArrayView1D<ImporterState> importerState;
Containers::Array<char> data = Containers::ArrayTuple{
{NoInit, 6, transformations},
{NoInit, 2, trs},
{NoInit, 2, meshes},
{NoInit, 1, cameras},
{NoInit, 2, skins},
{NoInit, 3, importerState}
};
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
int a, b, c;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
/* Object 3 is in the root, has a camera attached, TRS and children 5 + 4.
Because of the TRS, the actual transformation gets ignored. Has importer
state. */
transformations[0] = {3, -1, Matrix3::rotation(75.0_degf)};
trs[0] = {3, {0.0f, 3.0f}, Complex::rotation(15.0_degf), Vector2{1.0f}};
cameras[0] = {3, 15};
importerState[0] = {3, &a};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
/* Object 5 is a child of object 3, has a skin (which gets ignored by the
legacy interface) */
transformations[1] = {5, 3, Matrix3::rotation(-15.0_degf)};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
skins[0] = {5, 226};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
/* Object 1 is a child of object 2 */
transformations[2] = {1, 2, Matrix3::translation({1.0f, 0.5f})*Matrix3::rotation(15.0_degf)};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
/* Object 2 is in the root, has object 1 as a child but nothing else */
transformations[3] = {2, -1, {}};
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
/* Object 0 is in the root, has a mesh without a material and no children */
transformations[4] = {0, -1, Matrix3::rotation(30.0_degf)};
meshes[0] = {0, 33, -1};
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
/* Object 4 has TRS also, a mesh with a material and a skin and is a child
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
of object 3. The transformation gets ignored again. Has importer
state. */
transformations[5] = {4, 3, Matrix3::translation(Vector2::xAxis(5.0f))};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
trs[1] = {4, {}, {}, {1.5f, -0.5f}};
meshes[1] = {4, 27, 46};
skins[1] = {4, 72};
importerState[1] = {4, &b};
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
/* Object 6 has neither a transformation nor a parent, only an importer
state. It should get ignored. */
importerState[2] = {6, &c};
struct Importer: AbstractImporter {
explicit Importer(SceneData&& data): _data{std::move(data)} {}
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
bool doIsOpened() const override { return true; }
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
void doClose() override {}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedInt doSceneCount() const override { return 3; }
UnsignedLong doObjectCount() const override { return 7; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Long doObjectForName(Containers::StringView name) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(name == "sixth") return 5;
return -1;
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doObjectName(UnsignedLong id) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(id == 5) return "sixth";
return {};
}
Containers::Optional<SceneData> doScene(UnsignedInt id) override {
/* This one has seven objects, but no fields for them so it should
get skipped */
if(id == 0)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedByte, 7, nullptr, {}};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
/* This one has no objects, so it should get skipped as well
without even querying any fieldFor() API (as those would
assert) */
if(id == 1)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedShort, 0, nullptr, {}};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
/* This one is the one */
if(id == 2)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 7, {}, _data.data(), sceneFieldDataNonOwningArray(_data.fieldData())};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_INTERNAL_ASSERT_UNREACHABLE();
}
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
private:
SceneData _data;
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
} importer{SceneData{SceneMappingType::UnsignedInt, 7, std::move(data), {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
SceneFieldData{SceneField::Parent,
transformations.slice(&Transform::object),
transformations.slice(&Transform::parent)},
SceneFieldData{SceneField::Transformation,
transformations.slice(&Transform::object),
transformations.slice(&Transform::transformation)},
SceneFieldData{SceneField::Translation,
trs.slice(&Trs::object),
trs.slice(&Trs::translation)},
SceneFieldData{SceneField::Rotation,
trs.slice(&Trs::object),
trs.slice(&Trs::rotation)},
SceneFieldData{SceneField::Scaling,
trs.slice(&Trs::object),
trs.slice(&Trs::scaling)},
SceneFieldData{SceneField::Mesh,
meshes.slice(&Mesh::object),
meshes.slice(&Mesh::mesh)},
SceneFieldData{SceneField::MeshMaterial,
meshes.slice(&Mesh::object),
meshes.slice(&Mesh::meshMaterial)},
SceneFieldData{SceneField::Camera,
cameras.slice(&Index::object),
cameras.slice(&Index::id)},
SceneFieldData{SceneField::Skin,
skins.slice(&Index::object),
skins.slice(&Index::id)},
SceneFieldData{SceneField::ImporterState,
importerState.slice(&ImporterState::object), importerState.slice(&ImporterState::importerState)}
}}};
CORRADE_COMPARE(importer.sceneCount(), 3);
Containers::Optional<SceneData> scene = importer.scene(2);
CORRADE_VERIFY(scene);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(scene->children2D(),
(std::vector<UnsignedInt>{3, 2, 0}),
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(scene->children3D(),
(std::vector<UnsignedInt>{}),
TestSuite::Compare::Container);
CORRADE_COMPARE(importer.object2DCount(), 7);
CORRADE_COMPARE(importer.object2DForName("sixth"), 5);
CORRADE_COMPARE(importer.object2DName(5), "sixth");
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(importer.object3DCount(), 0);
CORRADE_COMPARE(importer.object3DForName("sixth"), -1);
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
{
Containers::Pointer<ObjectData2D> o = importer.object2D(0);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 33);
CORRADE_COMPARE(o->flags(), ObjectFlags2D{});
CORRADE_COMPARE(o->transformation(), Matrix3::rotation(30.0_degf));
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), -1);
CORRADE_COMPARE(mo.skin(), -1);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(1);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Empty);
CORRADE_COMPARE(o->instance(), -1);
CORRADE_COMPARE(o->flags(), ObjectFlags2D{});
CORRADE_COMPARE(o->transformation(), Matrix3::translation({1.0f, 0.5f})*Matrix3::rotation(15.0_degf));
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(2);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Empty);
CORRADE_COMPARE(o->instance(), -1);
CORRADE_COMPARE(o->flags(), ObjectFlags2D{});
CORRADE_COMPARE(o->transformation(), Matrix3{});
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{1},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(3);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), &a);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Camera);
CORRADE_COMPARE(o->instance(), 15);
CORRADE_COMPARE(o->flags(), ObjectFlag2D::HasTranslationRotationScaling);
CORRADE_COMPARE(o->transformation(), Matrix3::translation({0.0f, 3.0f})*Matrix3::rotation(15.0_degf));
CORRADE_COMPARE(o->translation(), (Vector2{0.0f, 3.0f}));
CORRADE_COMPARE(o->rotation(), Complex::rotation(15.0_degf));
CORRADE_COMPARE(o->scaling(), (Vector2{1.0f}));
CORRADE_COMPARE_AS(o->children(),
(std::vector<UnsignedInt>{5, 4}),
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(4);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), &b);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 27);
CORRADE_COMPARE(o->flags(), ObjectFlag2D::HasTranslationRotationScaling);
CORRADE_COMPARE(o->transformation(), Matrix3::scaling({1.5f, -0.5f}));
CORRADE_COMPARE(o->translation(), Vector2{});
CORRADE_COMPARE(o->rotation(), Complex{});
CORRADE_COMPARE(o->scaling(), (Vector2{1.5, -0.5f}));
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), 46);
CORRADE_COMPARE(mo.skin(), 72);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D("sixth");
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Empty);
CORRADE_COMPARE(o->instance(), -1);
CORRADE_COMPARE(o->flags(), ObjectFlags2D{});
CORRADE_COMPARE(o->transformation(), Matrix3::rotation(-15.0_degf));
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
} {
/* This one is not contained in any parent hierarchy, so it fails to
import */
std::ostringstream out;
Error redirectError{&out};
CORRADE_VERIFY(!importer.object2D(6));
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object2D(): object 6 not found in any 2D scene hierarchy\n");
}
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::sceneDeprecatedFallback3D() {
/* Need to test the following combinations:
- few objects in the root
- an object with one child, with more than one, with none
- an object with a mesh and a material
- an object with a mesh and without a material
- an object with a mesh and a skin
- an object with a skin but no mesh
- an object with a camera
- an object with a light
- an object with TRS transformation
- an object with TRS transformation and a mesh
- an object with nothing except parent / transformation
*/
struct Transform {
UnsignedInt object;
Int parent;
Matrix4 transformation;
};
struct Trs {
UnsignedInt object;
Vector3 translation;
Quaternion rotation;
Vector3 scaling;
};
struct Mesh {
UnsignedInt object;
UnsignedShort mesh;
Short meshMaterial;
};
struct Index {
UnsignedInt object;
UnsignedInt id;
};
struct ImporterState {
UnsignedInt object;
const void* importerState;
};
Containers::StridedArrayView1D<Transform> transformations;
Containers::StridedArrayView1D<Trs> trs;
Containers::StridedArrayView1D<Mesh> meshes;
Containers::StridedArrayView1D<Index> cameras;
Containers::StridedArrayView1D<Index> lights;
Containers::StridedArrayView1D<Index> skins;
Containers::StridedArrayView1D<ImporterState> importerState;
Containers::Array<char> data = Containers::ArrayTuple{
{NoInit, 6, transformations},
{NoInit, 2, trs},
{NoInit, 2, meshes},
{NoInit, 1, cameras},
{NoInit, 1, lights},
{NoInit, 2, skins},
{NoInit, 3, importerState}
};
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
int a, b, c;
/* Object 3 is in the root, has a camera attached, TRS and children 5 + 4.
Because of the TRS, the actual transformation gets ignored. Has importer
state. */
transformations[0] = {3, -1, Matrix4::rotationX(75.0_degf)};
trs[0] = {3, {0.0f, 0.0f, 3.0f}, Quaternion::rotation(15.0_degf, Vector3::xAxis()), Vector3{1.0f}};
cameras[0] = {3, 15};
importerState[0] = {3, &a};
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
/* Object 5 is a child of object 3, has a skin (which gets ignored by the
legacy interface) */
transformations[1] = {5, 3, Matrix4::rotationY(-15.0_degf)};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
skins[0] = {5, 226};
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
/* Object 1 is a child of object 2, has a light. */
transformations[2] = {1, 2, Matrix4::translation({1.0f, 0.0f, 1.0f})*Matrix4::rotationZ(15.0_degf)};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
lights[0] = {1, 113};
/* Object 2 is in the root, has object 1 as a child but nothing else */
transformations[3] = {2, -1, {}};
/* Object 0 is in the root, has a mesh without a material and no children */
transformations[4] = {0, -1, Matrix4::rotationX(30.0_degf)};
meshes[0] = {0, 33, -1};
/* Object 4 has TRS also, a mesh with a material and a skin and is a child
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
of object 3. The transformation gets ignored again. Has importer
state. */
transformations[5] = {4, 3, Matrix4::translation(Vector3::xAxis(5.0f))};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
trs[1] = {4, {}, {}, {1.5f, 3.0f, -0.5f}};
meshes[1] = {4, 27, 46};
skins[1] = {4, 72};
importerState[1] = {4, &b};
/* Object 6 has neither a transformation nor a parent, only an importer
state. It should get ignored. */
importerState[2] = {6, &c};
struct Importer: AbstractImporter {
explicit Importer(SceneData&& data): _data{std::move(data)} {}
Trade: add SceneField string mapping APIs to AbstractImporter.
5 years ago
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: animation import interface in AbstractImporter.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedInt doSceneCount() const override { return 3; }
UnsignedLong doObjectCount() const override { return 7; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Long doObjectForName(Containers::StringView name) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(name == "sixth") return 5;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: animation import interface in AbstractImporter.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doObjectName(UnsignedLong id) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(id == 5) return "sixth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: animation import interface in AbstractImporter.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
Containers::Optional<SceneData> doScene(UnsignedInt id) override {
/* This one has seven objects, but no fields for them so it should
get skipped */
if(id == 0)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedByte, 7, nullptr, {}};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
/* This one has no objects, so it should get skipped as well
without even querying any fieldFor() API (as those would
assert) */
if(id == 1)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedShort, 0, nullptr, {}};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
/* This one is the one */
if(id == 2)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 7, {}, _data.data(), sceneFieldDataNonOwningArray(_data.fieldData())};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_INTERNAL_ASSERT_UNREACHABLE();
Trade: animation import interface in AbstractImporter.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
private:
SceneData _data;
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
} importer{SceneData{SceneMappingType::UnsignedInt, 7, std::move(data), {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
SceneFieldData{SceneField::Parent,
transformations.slice(&Transform::object),
transformations.slice(&Transform::parent)},
SceneFieldData{SceneField::Transformation,
transformations.slice(&Transform::object),
transformations.slice(&Transform::transformation)},
SceneFieldData{SceneField::Translation,
trs.slice(&Trs::object),
trs.slice(&Trs::translation)},
SceneFieldData{SceneField::Rotation,
trs.slice(&Trs::object),
trs.slice(&Trs::rotation)},
SceneFieldData{SceneField::Scaling,
trs.slice(&Trs::object),
trs.slice(&Trs::scaling)},
SceneFieldData{SceneField::Mesh,
meshes.slice(&Mesh::object),
meshes.slice(&Mesh::mesh)},
SceneFieldData{SceneField::MeshMaterial,
meshes.slice(&Mesh::object),
meshes.slice(&Mesh::meshMaterial)},
SceneFieldData{SceneField::Camera,
cameras.slice(&Index::object),
cameras.slice(&Index::id)},
SceneFieldData{SceneField::Light,
lights.slice(&Index::object),
lights.slice(&Index::id)},
SceneFieldData{SceneField::Skin,
skins.slice(&Index::object),
skins.slice(&Index::id)},
SceneFieldData{SceneField::ImporterState,
importerState.slice(&ImporterState::object), importerState.slice(&ImporterState::importerState)}
}}};
CORRADE_COMPARE(importer.sceneCount(), 3);
Containers::Optional<SceneData> scene = importer.scene(2);
CORRADE_VERIFY(scene);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(scene->children2D(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(scene->children3D(),
(std::vector<UnsignedInt>{3, 2, 0}),
TestSuite::Compare::Container);
CORRADE_COMPARE(importer.object2DCount(), 0);
CORRADE_COMPARE(importer.object2DForName("sixth"), -1);
CORRADE_COMPARE(importer.object3DCount(), 7);
CORRADE_COMPARE(importer.object3DForName("sixth"), 5);
CORRADE_COMPARE(importer.object3DName(5), "sixth");
Trade: animation import interface in AbstractImporter.
8 years ago
Trade: allow retrieving importer data directly by name.
6 years ago
{
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
Containers::Pointer<ObjectData3D> o = importer.object3D(0);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 33);
CORRADE_COMPARE(o->flags(), ObjectFlags3D{});
CORRADE_COMPARE(o->transformation(), Matrix4::rotationX(30.0_degf));
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), -1);
CORRADE_COMPARE(mo.skin(), -1);
Trade: allow retrieving importer data directly by name.
6 years ago
} {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
Containers::Pointer<ObjectData3D> o = importer.object3D(1);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Light);
CORRADE_COMPARE(o->instance(), 113);
CORRADE_COMPARE(o->flags(), ObjectFlags3D{});
CORRADE_COMPARE(o->transformation(), Matrix4::translation({1.0f, 0.0f, 1.0f})*Matrix4::rotationZ(15.0_degf));
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(2);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Empty);
CORRADE_COMPARE(o->instance(), -1);
CORRADE_COMPARE(o->flags(), ObjectFlags3D{});
CORRADE_COMPARE(o->transformation(), Matrix4{});
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{1},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(3);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), &a);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Camera);
CORRADE_COMPARE(o->instance(), 15);
CORRADE_COMPARE(o->flags(), ObjectFlag3D::HasTranslationRotationScaling);
CORRADE_COMPARE(o->transformation(), Matrix4::translation({0.0f, 0.0f, 3.0f})*Matrix4::rotationX(15.0_degf));
CORRADE_COMPARE(o->translation(), (Vector3{0.0f, 0.0f, 3.0f}));
CORRADE_COMPARE(o->rotation(), Quaternion::rotation(15.0_degf, Vector3::xAxis()));
CORRADE_COMPARE(o->scaling(), (Vector3{1.0f}));
CORRADE_COMPARE_AS(o->children(),
(std::vector<UnsignedInt>{5, 4}),
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(4);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), &b);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 27);
CORRADE_COMPARE(o->flags(), ObjectFlag3D::HasTranslationRotationScaling);
CORRADE_COMPARE(o->transformation(), Matrix4::scaling({1.5f, 3.0f, -0.5f}));
CORRADE_COMPARE(o->translation(), Vector3{});
CORRADE_COMPARE(o->rotation(), Quaternion{});
CORRADE_COMPARE(o->scaling(), (Vector3{1.5, 3.0f, -0.5f}));
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), 46);
CORRADE_COMPARE(mo.skin(), 72);
} {
Containers::Pointer<ObjectData3D> o = importer.object3D("sixth");
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Empty);
CORRADE_COMPARE(o->instance(), -1);
CORRADE_COMPARE(o->flags(), ObjectFlags3D{});
CORRADE_COMPARE(o->transformation(), Matrix4::rotationY(-15.0_degf));
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
} {
/* This one is not contained in any parent hierarchy, so it fails to
import */
std::ostringstream out;
Error redirectError{&out};
CORRADE_VERIFY(!importer.object3D(6));
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object3D(): object 6 not found in any 3D scene hierarchy\n");
Trade: allow retrieving importer data directly by name.
6 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::sceneDeprecatedFallbackParentless2D() {
/* As the Parent field is currently used to distinguish which objects
belong to which scene, its absence means the objects are advertised, but
aren't listed as children of any scene, and retrieving them will fail */
/** @todo adapt when there's a dedicated way to distinguish which objects
belong to which scene */
struct Field {
UnsignedInt object;
Matrix3 transformation;
} fields[]{
{5, Matrix3{}},
{2, Matrix3{}},
};
Trade: animation import interface in AbstractImporter.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
Containers::StridedArrayView1D<Field> view = fields;
struct Importer: AbstractImporter {
explicit Importer(SceneData&& data): _data{std::move(data)} {}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedInt doSceneCount() const override { return 1; }
UnsignedLong doObjectCount() const override { return 6; }
Containers::Optional<SceneData> doScene(UnsignedInt) override {
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 6, {}, _data.data(), sceneFieldDataNonOwningArray(_data.fieldData())};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
}
private:
SceneData _data;
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
} importer{SceneData{SceneMappingType::UnsignedInt, 6, {}, fields, {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
SceneFieldData{SceneField::Transformation,
view.slice(&Field::object),
view.slice(&Field::transformation)}
}}};
CORRADE_COMPARE(importer.sceneCount(), 1);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DCount(), 6);
CORRADE_COMPARE(importer.object3DCount(), 0);
CORRADE_IGNORE_DEPRECATED_POP
Containers::Optional<SceneData> scene = importer.scene(0);
CORRADE_VERIFY(scene);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(scene->children2D(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(scene->children3D(),
(std::vector<UnsignedInt>{}),
TestSuite::Compare::Container);
CORRADE_IGNORE_DEPRECATED_POP
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_VERIFY(!importer.object2D(0));
CORRADE_VERIFY(!importer.object2D(1));
CORRADE_VERIFY(!importer.object2D(2));
CORRADE_VERIFY(!importer.object2D(3));
CORRADE_VERIFY(!importer.object2D(4));
CORRADE_VERIFY(!importer.object2D(5));
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::object2D(): object 0 not found in any 2D scene hierarchy\n"
"Trade::AbstractImporter::object2D(): object 1 not found in any 2D scene hierarchy\n"
"Trade::AbstractImporter::object2D(): object 2 not found in any 2D scene hierarchy\n"
"Trade::AbstractImporter::object2D(): object 3 not found in any 2D scene hierarchy\n"
"Trade::AbstractImporter::object2D(): object 4 not found in any 2D scene hierarchy\n"
"Trade::AbstractImporter::object2D(): object 5 not found in any 2D scene hierarchy\n");
}
void AbstractImporterTest::sceneDeprecatedFallbackParentless3D() {
/* As the Parent field is currently used to distinguish which objects
belong to which scene, its absence means the objects are advertised, but
aren't listed as children of any scene, and retrieving them will fail */
/** @todo adapt when there's a dedicated way to distinguish which objects
belong to which scene */
struct Field {
UnsignedInt object;
Matrix4 transformation;
} fields[]{
{5, Matrix4{}},
{2, Matrix4{}},
};
Containers::StridedArrayView1D<Field> view = fields;
struct Importer: AbstractImporter {
explicit Importer(SceneData&& data): _data{std::move(data)} {}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: animation import interface in AbstractImporter.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedInt doSceneCount() const override { return 1; }
UnsignedLong doObjectCount() const override { return 6; }
Containers::Optional<SceneData> doScene(UnsignedInt) override {
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 6, {}, _data.data(), sceneFieldDataNonOwningArray(_data.fieldData())};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
}
private:
SceneData _data;
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
} importer{SceneData{SceneMappingType::UnsignedInt, 6, {}, fields, {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
SceneFieldData{SceneField::Transformation,
view.slice(&Field::object),
view.slice(&Field::transformation)}
}}};
CORRADE_COMPARE(importer.sceneCount(), 1);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DCount(), 0);
CORRADE_COMPARE(importer.object3DCount(), 6);
CORRADE_IGNORE_DEPRECATED_POP
Containers::Optional<SceneData> scene = importer.scene(0);
CORRADE_VERIFY(scene);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(scene->children2D(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(scene->children3D(),
(std::vector<UnsignedInt>{}),
TestSuite::Compare::Container);
CORRADE_IGNORE_DEPRECATED_POP
std::ostringstream out;
Error redirectError{&out};
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_VERIFY(!importer.object3D(0));
CORRADE_VERIFY(!importer.object3D(1));
CORRADE_VERIFY(!importer.object3D(2));
CORRADE_VERIFY(!importer.object3D(3));
CORRADE_VERIFY(!importer.object3D(4));
CORRADE_VERIFY(!importer.object3D(5));
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::object3D(): object 0 not found in any 3D scene hierarchy\n"
"Trade::AbstractImporter::object3D(): object 1 not found in any 3D scene hierarchy\n"
"Trade::AbstractImporter::object3D(): object 2 not found in any 3D scene hierarchy\n"
"Trade::AbstractImporter::object3D(): object 3 not found in any 3D scene hierarchy\n"
"Trade::AbstractImporter::object3D(): object 4 not found in any 3D scene hierarchy\n"
"Trade::AbstractImporter::object3D(): object 5 not found in any 3D scene hierarchy\n");
}
void AbstractImporterTest::sceneDeprecatedFallbackTransformless2D() {
/* If no transformation field is present, for backwards compatibility we
assume the objects are 3D -- the only plugin that has a 2D scene is
PrimitiveImporter and it has the transformation field. */
struct Field {
UnsignedInt object;
Int parent;
} fields[]{
{5, -1},
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
{2, 5},
{3, 5},
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
{1, -1}
};
Containers::StridedArrayView1D<Field> view = fields;
struct Importer: AbstractImporter {
explicit Importer(SceneData&& data): _data{std::move(data)} {}
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 1; }
UnsignedLong doObjectCount() const override { return 6; }
Containers::Optional<SceneData> doScene(UnsignedInt) override {
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 6, {}, _data.data(), sceneFieldDataNonOwningArray(_data.fieldData())};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
}
private:
SceneData _data;
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
} importer{SceneData{SceneMappingType::UnsignedInt, 6, {}, fields, {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
SceneFieldData{SceneField::Parent,
view.slice(&Field::object),
view.slice(&Field::parent)},
/* Required in order to have the scene recognized as 2D */
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneFieldData{SceneField::Transformation, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Matrix3x3, nullptr}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
}}};
CORRADE_COMPARE(importer.sceneCount(), 1);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DCount(), 6);
CORRADE_COMPARE(importer.object3DCount(), 0);
CORRADE_IGNORE_DEPRECATED_POP
Containers::Optional<SceneData> scene = importer.scene(0);
CORRADE_VERIFY(scene);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(scene->children2D(),
(std::vector<UnsignedInt>{5, 1}),
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(scene->children3D(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
/* If we have neither a matrix nor a TRS, having an identity TRS is better
as it's more flexible compared to a matrix */
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
{
Containers::Pointer<ObjectData2D> o = importer.object2D(5);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Empty);
CORRADE_COMPARE(o->instance(), -1);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
CORRADE_COMPARE(o->flags(), ObjectFlag2D::HasTranslationRotationScaling);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(o->transformation(), Matrix3{});
CORRADE_COMPARE_AS(o->children(),
(std::vector<UnsignedInt>{2, 3}),
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(2);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Empty);
CORRADE_COMPARE(o->instance(), -1);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
CORRADE_COMPARE(o->flags(), ObjectFlag2D::HasTranslationRotationScaling);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(o->transformation(), Matrix3{});
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(3);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Empty);
CORRADE_COMPARE(o->instance(), -1);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
CORRADE_COMPARE(o->flags(), ObjectFlag2D::HasTranslationRotationScaling);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(o->transformation(), Matrix3{});
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(1);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Empty);
CORRADE_COMPARE(o->instance(), -1);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
CORRADE_COMPARE(o->flags(), ObjectFlag2D::HasTranslationRotationScaling);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(o->transformation(), Matrix3{});
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
}
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::sceneDeprecatedFallbackTransformless3D() {
/* If no transformation field is present, for backwards compatibility we
assume the objects are 3D -- the only plugin that has a 2D scene is
PrimitiveImporter and it has the transformation field. */
struct Field {
UnsignedInt object;
Int parent;
} fields[]{
{5, -1},
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
{2, 5},
{3, 5},
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
{1, -1}
};
Containers::StridedArrayView1D<Field> view = fields;
struct Importer: AbstractImporter {
explicit Importer(SceneData&& data): _data{std::move(data)} {}
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 1; }
UnsignedLong doObjectCount() const override { return 6; }
Containers::Optional<SceneData> doScene(UnsignedInt) override {
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 6, {}, _data.data(), sceneFieldDataNonOwningArray(_data.fieldData())};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
}
private:
SceneData _data;
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
} importer{SceneData{SceneMappingType::UnsignedInt, 6, {}, fields, {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
SceneFieldData{SceneField::Parent,
view.slice(&Field::object),
view.slice(&Field::parent)},
/* Required in order to have the scene recognized as 3D */
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneFieldData{SceneField::Transformation, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Matrix4x4, nullptr}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
}}};
CORRADE_COMPARE(importer.sceneCount(), 1);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DCount(), 0);
CORRADE_COMPARE(importer.object3DCount(), 6);
CORRADE_IGNORE_DEPRECATED_POP
Containers::Optional<SceneData> scene = importer.scene(0);
CORRADE_VERIFY(scene);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(scene->children2D(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(scene->children3D(),
(std::vector<UnsignedInt>{5, 1}),
TestSuite::Compare::Container);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
/* If we have neither a matrix nor a TRS, having an identity TRS is better
as it's more flexible compared to a matrix */
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
{
Containers::Pointer<ObjectData3D> o = importer.object3D(5);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Empty);
CORRADE_COMPARE(o->instance(), -1);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
CORRADE_COMPARE(o->flags(), ObjectFlag3D::HasTranslationRotationScaling);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(o->transformation(), Matrix4{});
CORRADE_COMPARE_AS(o->children(),
(std::vector<UnsignedInt>{2, 3}),
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(2);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Empty);
CORRADE_COMPARE(o->instance(), -1);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
CORRADE_COMPARE(o->flags(), ObjectFlag3D::HasTranslationRotationScaling);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(o->transformation(), Matrix4{});
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(3);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Empty);
CORRADE_COMPARE(o->instance(), -1);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
CORRADE_COMPARE(o->flags(), ObjectFlag3D::HasTranslationRotationScaling);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(o->transformation(), Matrix4{});
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(1);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->importerState(), nullptr);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Empty);
CORRADE_COMPARE(o->instance(), -1);
Trade: the compatibility default for transformless scenes should be TRS. That's what plugins implemented with the legacy API did, as it's more flexible than a matrix.
5 years ago
CORRADE_COMPARE(o->flags(), ObjectFlag3D::HasTranslationRotationScaling);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_COMPARE(o->transformation(), Matrix4{});
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
TestSuite::Compare::Container);
}
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::sceneDeprecatedFallbackMultiFunctionObjects2D() {
/* Mostly just a copy of SceneToolsTest::convertToSingleFunctionObjects()
except that here we can't use the convenience combining tool so it's
done by hand */
struct Parent {
UnsignedInt object;
Int parent;
};
struct Mesh {
UnsignedInt object;
UnsignedInt mesh;
Int meshMaterial;
};
struct Camera {
UnsignedInt object;
UnsignedInt camera;
};
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
struct Skin {
UnsignedInt object;
UnsignedInt skin;
};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Containers::StridedArrayView1D<Parent> parents;
Containers::StridedArrayView1D<Mesh> meshes;
Containers::StridedArrayView1D<Camera> cameras;
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
Containers::StridedArrayView1D<Skin> skins;
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Containers::Array<char> dataData = Containers::ArrayTuple{
{NoInit, 5, parents},
{NoInit, 7, meshes},
{NoInit, 2, cameras},
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
{NoInit, 2, skins},
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
};
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
Utility::copy({{15, -1}, {21, -1}, {22, 21}, {23, 22}, {1, -1}}, parents);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Utility::copy({
{15, 6, 4},
{23, 1, 0},
{23, 2, 3},
{23, 4, 2},
{1, 7, 2},
{15, 3, 1},
{21, 5, -1}
}, meshes);
Utility::copy({{22, 1}, {1, 5}}, cameras);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
Utility::copy({{15, 9}, {21, 10}}, skins);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
/* Second scene that also has a duplicate, to verify the newly added object
IDs don't conflict with each other. A potential downside is that
multi-primitive nodes shared by multiple scenes get duplicated, but
that's a smaller problem than two unrelated nodes sharing the same ID
(and thus having a wrong name, etc). */
Containers::StridedArrayView1D<Parent> parentsSecondary;
Containers::StridedArrayView1D<Mesh> meshesSecondary;
Containers::Array<char> dataDataSecondary = Containers::ArrayTuple{
{NoInit, 1, parentsSecondary},
{NoInit, 2, meshesSecondary}
};
Utility::copy({{30, -1}}, parentsSecondary);
Utility::copy({
{30, 6, 2},
{30, 1, -1}
}, meshesSecondary);
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneData data{SceneMappingType::UnsignedInt, 32, std::move(dataData), {
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
SceneFieldData{SceneField::Parent, parents.slice(&Parent::object), parents.slice(&Parent::parent)},
SceneFieldData{SceneField::Mesh, meshes.slice(&Mesh::object), meshes.slice(&Mesh::mesh)},
SceneFieldData{SceneField::MeshMaterial, meshes.slice(&Mesh::object), meshes.slice(&Mesh::meshMaterial)},
SceneFieldData{SceneField::Camera, cameras.slice(&Camera::object), cameras.slice(&Camera::camera)},
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
SceneFieldData{SceneField::Skin, skins.slice(&Skin::object), skins.slice(&Skin::skin)},
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
/* Just to disambiguate this as a 2D scene */
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneFieldData{SceneField::Transformation, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Matrix3x3, nullptr},
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
}};
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneData dataSecondary{SceneMappingType::UnsignedInt, 31, std::move(dataDataSecondary), {
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
SceneFieldData{SceneField::Parent, parentsSecondary.slice(&Parent::object), parentsSecondary.slice(&Parent::parent)},
SceneFieldData{SceneField::Mesh, meshesSecondary.slice(&Mesh::object), meshesSecondary.slice(&Mesh::mesh)},
SceneFieldData{SceneField::MeshMaterial, meshesSecondary.slice(&Mesh::object), meshesSecondary.slice(&Mesh::meshMaterial)},
/* Just to disambiguate this as a 2D scene */
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneFieldData{SceneField::Transformation, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Matrix3x3, nullptr},
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
}};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
struct Importer: AbstractImporter {
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
explicit Importer(SceneData&& data, SceneData&& dataSecondary): _data{std::move(data)}, _dataSecondary{std::move(dataSecondary)} {}
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
UnsignedInt doSceneCount() const override { return 4; }
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
UnsignedLong doObjectCount() const override { return 63; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doObjectName(UnsignedLong id) override {
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
if(id == 1) return "object 1";
if(id == 15) return "object 15";
if(id == 23) return "object 23";
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
if(id == 30) return "object 30 from secondary scene";
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
if(id == 62) return "last";
CORRADE_INTERNAL_ASSERT_UNREACHABLE();
}
Containers::Optional<SceneData> doScene(UnsignedInt id) override {
/* This scene should get skipped when querying names as it's not
2D */
if(id == 0)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedByte, 32, nullptr, {}};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
/* This scene should get skipped when querying names as it has too
little objects */
if(id == 1)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedByte, 32, nullptr, {
SceneFieldData{SceneField::Transformation, SceneMappingType::UnsignedByte, nullptr, SceneFieldType::Matrix3x3, nullptr}
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
}};
if(id == 2)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 32, {}, _data.data(), sceneFieldDataNonOwningArray(_data.fieldData())};
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
/* A secondary scene, which should have non-overlapping IDs for the
newly added objects */
if(id == 3)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 31, {}, _dataSecondary.data(), sceneFieldDataNonOwningArray(_dataSecondary.fieldData())};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
CORRADE_INTERNAL_ASSERT_UNREACHABLE();
}
private:
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
SceneData _data, _dataSecondary;
} importer{std::move(data), std::move(dataSecondary)};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
CORRADE_COMPARE(importer.sceneCount(), 4);
CORRADE_IGNORE_DEPRECATED_PUSH
/* Total object count reported by the importer plus four new added for the
first and one for the second scene */
CORRADE_COMPARE(importer.object2DCount(), 63 + 4 + 1);
CORRADE_COMPARE(importer.object3DCount(), 0);
/* Object name should return parent names for the additional objects */
CORRADE_COMPARE(importer.object2DName(62), "last");
CORRADE_COMPARE(importer.object2DName(63), "object 23");
CORRADE_COMPARE(importer.object2DName(64), "object 23");
CORRADE_COMPARE(importer.object2DName(65), "object 15");
CORRADE_COMPARE(importer.object2DName(66), "object 1");
CORRADE_COMPARE(importer.object2DName(67), "object 30 from secondary scene");
CORRADE_IGNORE_DEPRECATED_POP
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Containers::Optional<SceneData> scene = importer.scene(2);
CORRADE_VERIFY(scene);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(scene->children2D(),
(std::vector<UnsignedInt>{15, 21, 1}),
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(scene->children3D(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
/* Only 9 objects should exist in total, go in order. Usually the object
IDs will be contiguous so no such mess as this happens. */
{
Containers::Pointer<ObjectData2D> o = importer.object2D(1);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 7);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{66},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), 2);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), -1);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(15);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 6);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{65},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), 4);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), 9);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(21);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 5);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{22},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), -1);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), 10);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(22);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Camera);
CORRADE_COMPARE(o->instance(), 1);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{23},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(23);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 1);
CORRADE_COMPARE_AS(o->children(),
(std::vector<UnsignedInt>{63, 64}),
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), 0);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), -1);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(63);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 2);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), 3);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), -1);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(64);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 4);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), 2);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), -1);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(65);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 3);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), 1);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), 9);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(66);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Camera);
CORRADE_COMPARE(o->instance(), 5);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
}
CORRADE_IGNORE_DEPRECATED_POP
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
Containers::Optional<SceneData> sceneSecondary = importer.scene(3);
CORRADE_VERIFY(sceneSecondary);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(sceneSecondary->children2D(),
std::vector<UnsignedInt>{30},
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(sceneSecondary->children3D(),
(std::vector<UnsignedInt>{}),
TestSuite::Compare::Container);
/* One additional duplicated object here */
{
Containers::Pointer<ObjectData2D> o = importer.object2D(30);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 6);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{67},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), 2);
} {
Containers::Pointer<ObjectData2D> o = importer.object2D(67);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType2D::Mesh);
CORRADE_COMPARE(o->instance(), 1);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData2D& mo = static_cast<MeshObjectData2D&>(*o);
CORRADE_COMPARE(mo.material(), -1);
}
CORRADE_IGNORE_DEPRECATED_POP
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
}
void AbstractImporterTest::sceneDeprecatedFallbackMultiFunctionObjects3D() {
/* Mostly just a copy of SceneToolsTest::convertToSingleFunctionObjects()
except that here we can't use the convenience combining tool so it's
done by hand */
struct Parent {
UnsignedInt object;
Int parent;
};
struct Mesh {
UnsignedInt object;
UnsignedInt mesh;
Int meshMaterial;
};
struct Camera {
UnsignedInt object;
UnsignedInt camera;
};
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
struct Skin {
UnsignedInt object;
UnsignedInt skin;
};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Containers::StridedArrayView1D<Parent> parents;
Containers::StridedArrayView1D<Mesh> meshes;
Containers::StridedArrayView1D<Camera> cameras;
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
Containers::StridedArrayView1D<Skin> skins;
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Containers::Array<char> dataData = Containers::ArrayTuple{
{NoInit, 5, parents},
{NoInit, 7, meshes},
{NoInit, 2, cameras},
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
{NoInit, 2, skins},
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
};
Trade: remove the useless indirection from SceneField::Parent. This got originally added as some sort of a kludge to make it easy to go to the parent transformation, assuming Parent and Transformation share the same object mapping: parentTransformation = transformations[parents[i]] But after some ACTUAL REAL WORLD use, I realized that there's often a set of objects that have a Parent defined, and then another, completely disjoint, set of objects that have a transformation (for example certain nodes having no transformation at all because it's an identity). And so this parent indirection is not only useless, but in fact an additional complication. Let's say we make a map of the transformations, where transformationMap[i] is a transformation for object i: transformationMap = {} for j in range(len(transformations)): transformationMap[transformationObject[j]] = transformation[j] Then, with *no* assumptions about shared object mapping, the indirection would cause parent transformation retrieval to look like this: parentTransformation = transformationMap[parentObjects[parents[i]] While *without* the indirection, it'd be just parentTransformation = transformationMap[parents[i]]
5 years ago
Utility::copy({{15, -1}, {21, -1}, {22, 21}, {23, 22}, {1, -1}}, parents);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Utility::copy({
{15, 6, 4},
{23, 1, 0},
{23, 2, 3},
{23, 4, 2},
{1, 7, 2},
{15, 3, 1},
{21, 5, -1}
}, meshes);
Utility::copy({{22, 1}, {1, 5}}, cameras);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
Utility::copy({{15, 9}, {21, 10}}, skins);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
/* Second scene that also has a duplicate, to verify the newly added object
IDs don't conflict with each other. A potential downside is that
multi-primitive nodes shared by multiple scenes get duplicated, but
that's a smaller problem than two unrelated nodes sharing the same ID
(and thus having a wrong name, etc). */
Containers::StridedArrayView1D<Parent> parentsSecondary;
Containers::StridedArrayView1D<Mesh> meshesSecondary;
Containers::Array<char> dataDataSecondary = Containers::ArrayTuple{
{NoInit, 1, parentsSecondary},
{NoInit, 2, meshesSecondary}
};
Utility::copy({{30, -1}}, parentsSecondary);
Utility::copy({
{30, 6, 2},
{30, 1, -1}
}, meshesSecondary);
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneData data{SceneMappingType::UnsignedInt, 32, std::move(dataData), {
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
SceneFieldData{SceneField::Parent, parents.slice(&Parent::object), parents.slice(&Parent::parent)},
SceneFieldData{SceneField::Mesh, meshes.slice(&Mesh::object), meshes.slice(&Mesh::mesh)},
SceneFieldData{SceneField::MeshMaterial, meshes.slice(&Mesh::object), meshes.slice(&Mesh::meshMaterial)},
SceneFieldData{SceneField::Camera, cameras.slice(&Camera::object), cameras.slice(&Camera::camera)},
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
SceneFieldData{SceneField::Skin, skins.slice(&Skin::object), skins.slice(&Skin::skin)},
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
/* Just to disambiguate this as a 3D scene */
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneFieldData{SceneField::Transformation, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Matrix4x4, nullptr},
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
}};
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneData dataSecondary{SceneMappingType::UnsignedInt, 31, std::move(dataDataSecondary), {
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
SceneFieldData{SceneField::Parent, parentsSecondary.slice(&Parent::object), parentsSecondary.slice(&Parent::parent)},
SceneFieldData{SceneField::Mesh, meshesSecondary.slice(&Mesh::object), meshesSecondary.slice(&Mesh::mesh)},
SceneFieldData{SceneField::MeshMaterial, meshesSecondary.slice(&Mesh::object), meshesSecondary.slice(&Mesh::meshMaterial)},
/* Just to disambiguate this as a 3D scene */
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneFieldData{SceneField::Transformation, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Matrix4x4, nullptr},
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
}};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
struct Importer: AbstractImporter {
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
explicit Importer(SceneData&& data, SceneData&& dataSecondary): _data{std::move(data)}, _dataSecondary{std::move(dataSecondary)} {}
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
UnsignedInt doSceneCount() const override { return 4; }
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
UnsignedLong doObjectCount() const override { return 63; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doObjectName(UnsignedLong id) override {
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
if(id == 1) return "object 1";
if(id == 15) return "object 15";
if(id == 23) return "object 23";
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
if(id == 30) return "object 30 from secondary scene";
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
if(id == 62) return "last";
CORRADE_INTERNAL_ASSERT_UNREACHABLE();
}
Containers::Optional<SceneData> doScene(UnsignedInt id) override {
/* This scene should get skipped when querying names as it's not
2D */
if(id == 0)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedByte, 32, nullptr, {}};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
/* This scene should get skipped when querying names as it has too
little objects */
if(id == 1)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedByte, 32, nullptr, {
SceneFieldData{SceneField::Transformation, SceneMappingType::UnsignedByte, nullptr, SceneFieldType::Matrix4x4, nullptr}
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
}};
if(id == 2)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 32, {}, _data.data(), sceneFieldDataNonOwningArray(_data.fieldData())};
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
/* A secondary scene, which should have non-overlapping IDs for the
newly added objects */
if(id == 3)
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 31, {}, _dataSecondary.data(), sceneFieldDataNonOwningArray(_dataSecondary.fieldData())};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
CORRADE_INTERNAL_ASSERT_UNREACHABLE();
}
private:
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
SceneData _data, _dataSecondary;
} importer{std::move(data), std::move(dataSecondary)};
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
CORRADE_COMPARE(importer.sceneCount(), 4);
CORRADE_IGNORE_DEPRECATED_PUSH
/* Total object count reported by the importer plus four new added for the
first and one for the second scene */
CORRADE_COMPARE(importer.object2DCount(), 0);
CORRADE_COMPARE(importer.object3DCount(), 63 + 4 + 1);
/* Object name should return parent names for the additional objects */
CORRADE_COMPARE(importer.object3DName(62), "last");
CORRADE_COMPARE(importer.object3DName(63), "object 23");
CORRADE_COMPARE(importer.object3DName(64), "object 23");
CORRADE_COMPARE(importer.object3DName(65), "object 15");
CORRADE_COMPARE(importer.object3DName(66), "object 1");
CORRADE_COMPARE(importer.object3DName(67), "object 30 from secondary scene");
CORRADE_IGNORE_DEPRECATED_POP
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
Containers::Optional<SceneData> scene = importer.scene(2);
CORRADE_VERIFY(scene);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(scene->children2D(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(scene->children3D(),
(std::vector<UnsignedInt>{15, 21, 1}),
TestSuite::Compare::Container);
/* Only 9 objects should exist in total, go in order. Usually the object
IDs will be contiguous so no such mess as this happens. */
{
Containers::Pointer<ObjectData3D> o = importer.object3D(1);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 7);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{66},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), 2);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), -1);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(15);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 6);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{65},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), 4);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), 9);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(21);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 5);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{22},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), -1);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), 10);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(22);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Camera);
CORRADE_COMPARE(o->instance(), 1);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{23},
TestSuite::Compare::Container);
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(23);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 1);
CORRADE_COMPARE_AS(o->children(),
(std::vector<UnsignedInt>{63, 64}),
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), 0);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), -1);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(63);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 2);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), 3);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), -1);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(64);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 4);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), 2);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), -1);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(65);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 3);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), 1);
Trade: correctly copy skins alongside duplicated multi-mesh objects. This is implemented in a generic way internally, to allow copying of arbitrary fields, not just skins for meshes, in the upcoming SceneTools utility.
4 years ago
CORRADE_COMPARE(mo.skin(), 9);
Trade: process SceneData to single-function-objects for backwards compat. Now the code has all needed backwards compatibility in place.
5 years ago
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(66);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Camera);
CORRADE_COMPARE(o->instance(), 5);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
TestSuite::Compare::Container);
}
CORRADE_IGNORE_DEPRECATED_POP
Trade: ensure newly added object IDs don't overlap each other.
4 years ago
Containers::Optional<SceneData> sceneSecondary = importer.scene(3);
CORRADE_VERIFY(sceneSecondary);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE_AS(sceneSecondary->children2D(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
CORRADE_COMPARE_AS(sceneSecondary->children3D(),
(std::vector<UnsignedInt>{30}),
TestSuite::Compare::Container);
/* One additional duplicated object here */
{
Containers::Pointer<ObjectData3D> o = importer.object3D(30);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 6);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{67},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), 2);
} {
Containers::Pointer<ObjectData3D> o = importer.object3D(67);
CORRADE_VERIFY(o);
CORRADE_COMPARE(o->instanceType(), ObjectInstanceType3D::Mesh);
CORRADE_COMPARE(o->instance(), 1);
CORRADE_COMPARE_AS(o->children(),
std::vector<UnsignedInt>{},
TestSuite::Compare::Container);
MeshObjectData3D& mo = static_cast<MeshObjectData3D&>(*o);
CORRADE_COMPARE(mo.material(), -1);
}
CORRADE_IGNORE_DEPRECATED_POP
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
}
Trade: report at least some *D object count if everything fails. As the comment says -- before the user code expected that if the scene hierarchy is broken, particular objects will fail to import. Now the whole scene fails to import, so we don't even get to know the actual (expanded) deprecated 2D/3D object count. To reduce the suffering, return at least the dimension-less object count there. It won't include the duplicates from the single-function-object conversion but better than reporting 0.
5 years ago
void AbstractImporterTest::sceneDeprecatedFallbackObjectCountNoScenes() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 0; }
UnsignedLong doObjectCount() const override { return 27; }
} importer;
/* There's no scenes to get data or hierarchy from, so there are no
2D/3D objects reported even though objectCount() says 27 */
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DCount(), 0);
CORRADE_COMPARE(importer.object3DCount(), 0);
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::sceneDeprecatedFallbackObjectCountAllSceneImportFailed() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 1; }
UnsignedLong doObjectCount() const override { return 27; }
Containers::Optional<SceneData> doScene(UnsignedInt) override {
return {};
}
} importer;
/* There's a scene but it failed to import, assume it was 3D and proxy the
objectCount() */
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DCount(), 0);
CORRADE_COMPARE(importer.object3DCount(), 27);
CORRADE_IGNORE_DEPRECATED_POP
}
Trade: another case for the deprecated path not being 100% compatible. Honestly I don't care much, this is just that the original PrimitiveImporter tests started to fail because they expected object3DForName() to return -1 for a 2D name, but that's no longer the case. It would be possible to fix this, but I doubt anyone ever relied on such behavior for 2D scenes, so just add a test that acknowledges this new behavior.
5 years ago
void AbstractImporterTest::sceneDeprecatedFallbackBoth2DAnd3DScene() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 2; }
UnsignedLong doObjectCount() const override { return 7; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Long doObjectForName(Containers::StringView name) override {
Trade: another case for the deprecated path not being 100% compatible. Honestly I don't care much, this is just that the original PrimitiveImporter tests started to fail because they expected object3DForName() to return -1 for a 2D name, but that's no longer the case. It would be possible to fix this, but I doubt anyone ever relied on such behavior for 2D scenes, so just add a test that acknowledges this new behavior.
5 years ago
if(name == "sixth") return 5;
return -1;
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doObjectName(UnsignedLong id) override {
Trade: another case for the deprecated path not being 100% compatible. Honestly I don't care much, this is just that the original PrimitiveImporter tests started to fail because they expected object3DForName() to return -1 for a 2D name, but that's no longer the case. It would be possible to fix this, but I doubt anyone ever relied on such behavior for 2D scenes, so just add a test that acknowledges this new behavior.
5 years ago
if(id == 5) return "sixth";
return {};
}
Containers::Optional<SceneData> doScene(UnsignedInt id) override {
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
if(id == 0) return SceneData{SceneMappingType::UnsignedInt, 7, nullptr, {
SceneFieldData{SceneField::Parent, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Int, nullptr},
SceneFieldData{SceneField::Translation, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Vector2, nullptr},
Trade: another case for the deprecated path not being 100% compatible. Honestly I don't care much, this is just that the original PrimitiveImporter tests started to fail because they expected object3DForName() to return -1 for a 2D name, but that's no longer the case. It would be possible to fix this, but I doubt anyone ever relied on such behavior for 2D scenes, so just add a test that acknowledges this new behavior.
5 years ago
}};
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
if(id == 1) return SceneData{SceneMappingType::UnsignedInt, 7, nullptr, {
SceneFieldData{SceneField::Parent, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Int, nullptr},
SceneFieldData{SceneField::Translation, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Vector3, nullptr},
Trade: another case for the deprecated path not being 100% compatible. Honestly I don't care much, this is just that the original PrimitiveImporter tests started to fail because they expected object3DForName() to return -1 for a 2D name, but that's no longer the case. It would be possible to fix this, but I doubt anyone ever relied on such behavior for 2D scenes, so just add a test that acknowledges this new behavior.
5 years ago
}};
CORRADE_INTERNAL_ASSERT_UNREACHABLE();
}
} importer;
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DCount(), 7);
CORRADE_COMPARE(importer.object3DCount(), 7);
CORRADE_IGNORE_DEPRECATED_POP
{
CORRADE_EXPECT_FAIL("No check for whether given object is 2D or 3D is done, so the names are reported for both 2D and 3D objects.");
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DForName("sixth"), -1);
CORRADE_COMPARE(importer.object2DName(5), "");
CORRADE_COMPARE(importer.object3DForName("sixth"), -1);
CORRADE_COMPARE(importer.object3DName(5), "");
CORRADE_IGNORE_DEPRECATED_POP
} {
/* Just to be sure, verify that the names get really reported for both
instead of some other weird shit happening */
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DForName("sixth"), 5);
CORRADE_COMPARE(importer.object2DName(5), "sixth");
CORRADE_COMPARE(importer.object3DForName("sixth"), 5);
CORRADE_COMPARE(importer.object3DName(5), "sixth");
CORRADE_IGNORE_DEPRECATED_POP
}
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#endif
void AbstractImporterTest::sceneNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 8; }
} importer;
CORRADE_COMPARE(importer.sceneName(7), "");
}
void AbstractImporterTest::objectNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedLong doObjectCount() const override { return 8; }
} importer;
CORRADE_COMPARE(importer.objectName(7), "");
}
void AbstractImporterTest::sceneNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.sceneName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::sceneName(): index 8 out of range for 8 entries\n");
}
void AbstractImporterTest::objectNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedLong doObjectCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.objectName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::objectName(): index 8 out of range for 8 entries\n");
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::sceneNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 1; }
Containers::String doSceneName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.sceneName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::sceneName(): implementation is not allowed to use a custom String deleter\n");
}
void AbstractImporterTest::objectNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedLong doObjectCount() const override { return 1; }
Containers::String doObjectName(UnsignedLong) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.objectName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::objectName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void AbstractImporterTest::sceneNotImplemented() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.scene(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::scene(): not implemented\n");
}
void AbstractImporterTest::sceneOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.scene(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::scene(): index 8 out of range for 8 entries\n");
}
void AbstractImporterTest::sceneNonOwningDeleters() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 1; }
Containers::Optional<SceneData> doScene(UnsignedInt) override {
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 0,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
Containers::Array<char>{data, 1, Implementation::nonOwnedArrayDeleter},
sceneFieldDataNonOwningArray(fields)};
}
char data[1];
SceneFieldData fields[1]{
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneFieldData{SceneField::Parent, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Int, nullptr}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
};
} importer;
auto data = importer.scene(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(static_cast<const void*>(data->data()), importer.data);
CORRADE_COMPARE(static_cast<const void*>(data->fieldData()), importer.fields);
}
void AbstractImporterTest::sceneCustomDataDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSceneForName(Containers::StringView) override { return 0; }
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
Containers::Optional<SceneData> doScene(UnsignedInt) override {
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 0,
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
Containers::Array<char>{data, 1, [](char*, std::size_t) {}},
{}};
}
char data[1];
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.scene(0);
importer.scene("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::scene(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::scene(): implementation is not allowed to use a custom Array deleter\n");
}
void AbstractImporterTest::sceneCustomFieldDataDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSceneCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSceneForName(Containers::StringView) override { return 0; }
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
Containers::Optional<SceneData> doScene(UnsignedInt) override {
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
return SceneData{SceneMappingType::UnsignedInt, 0, nullptr, Containers::Array<SceneFieldData>{&parents, 1, [](SceneFieldData*, std::size_t) {}}};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
}
Trade: rename SceneData "objects" to "mapping". Now it's a field and its corresponding object mapping, instead of field and "objects": - Goes better with the concept that there's not really any materialized "object" anywhere, just fields mapped to them. - No more weird singular/plural difference between field() and objects(), it's field() and mapping() now. - The objectCount() that actually wasn't really object count is now a mappingBound(), an upper bound for object IDs contained in the object mapping views. Which is quite self-explanatory without having to mention every time that the range may be sparse.
4 years ago
SceneFieldData parents{SceneField::Parent, SceneMappingType::UnsignedInt, nullptr, SceneFieldType::Int, nullptr};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.scene(0);
importer.scene("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::scene(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::scene(): implementation is not allowed to use a custom Array deleter\n"
);
}
void AbstractImporterTest::sceneFieldName() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
SceneField doSceneFieldForName(Containers::StringView name) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(name == "OctreeCell") return sceneFieldCustom(100037);
return SceneField{};
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doSceneFieldName(UnsignedInt id) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(id == 100037) return "OctreeCell";
return "";
}
} importer;
CORRADE_COMPARE(importer.sceneFieldForName("OctreeCell"), sceneFieldCustom(100037));
CORRADE_COMPARE(importer.sceneFieldName(sceneFieldCustom(100037)), "OctreeCell");
}
void AbstractImporterTest::sceneFieldNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
CORRADE_COMPARE(importer.sceneFieldForName(""), SceneField{});
CORRADE_COMPARE(importer.sceneFieldName(sceneFieldCustom(100037)), "");
}
void AbstractImporterTest::sceneFieldNameNotCustom() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
SceneField doSceneFieldForName(Containers::StringView) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
return SceneField::Translation;
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.sceneFieldForName("OctreeCell");
importer.sceneFieldName(SceneField::Translation);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::sceneFieldForName(): implementation-returned Trade::SceneField::Translation is neither custom nor invalid\n"
"Trade::AbstractImporter::sceneFieldName(): Trade::SceneField::Translation is not custom\n");
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::sceneFieldNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
Containers::String doSceneFieldName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.sceneFieldName(sceneFieldCustom(0));
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::sceneFieldName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
void AbstractImporterTest::animation() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doAnimationCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doAnimationForName(Containers::StringView name) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(name == "eighth") return 7;
return -1;
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doAnimationName(UnsignedInt id) override {
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
if(id == 7) return "eighth";
return {};
}
Containers::Optional<AnimationData> doAnimation(UnsignedInt id) override {
/* Verify that initializer list is converted to an array with
the default deleter and not something disallowed */
if(id == 7) return AnimationData{nullptr, {
AnimationTrackData{AnimationTrackType::Vector3,
AnimationTrackTargetType::Scaling3D, 0, {}}
}, &state};
return AnimationData{{}, {}};
}
} importer;
CORRADE_COMPARE(importer.animationCount(), 8);
CORRADE_COMPARE(importer.animationForName("eighth"), 7);
CORRADE_COMPARE(importer.animationName(7), "eighth");
{
auto data = importer.animation(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.animation("eighth");
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
}
void AbstractImporterTest::animationForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doAnimationCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doAnimationForName(Containers::StringView) override { return 8; }
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.animationForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::animationForName(): implementation-returned index 8 out of range for 8 entries\n");
}
void AbstractImporterTest::animationNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doAnimationCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: animation import interface in AbstractImporter.
8 years ago
CORRADE_COMPARE(importer.animationName(7), "");
}
void AbstractImporterTest::animationNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: animation import interface in AbstractImporter.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doAnimationCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: animation import interface in AbstractImporter.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.animationName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::animationName(): index 8 out of range for 8 entries\n");
Trade: animation import interface in AbstractImporter.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::animationNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doAnimationCount() const override { return 1; }
Containers::String doAnimationName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.animationName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::animationName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: animation import interface in AbstractImporter.
8 years ago
void AbstractImporterTest::animationNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: animation import interface in AbstractImporter.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doAnimationCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: animation import interface in AbstractImporter.
8 years ago
std::ostringstream out;
Error redirectError{&out};
importer.animation(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::animation(): not implemented\n");
}
void AbstractImporterTest::animationOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: animation import interface in AbstractImporter.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doAnimationCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: animation import interface in AbstractImporter.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.animation(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::animation(): index 8 out of range for 8 entries\n");
Trade: animation import interface in AbstractImporter.
8 years ago
}
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void AbstractImporterTest::animationNonOwningDeleters() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doAnimationCount() const override { return 1; }
Containers::Optional<AnimationData> doAnimation(UnsignedInt) override {
return AnimationData{Containers::Array<char>{data, 1, Implementation::nonOwnedArrayDeleter},
Containers::Array<AnimationTrackData>{&track, 1,
Trade: don't reinterpret no-op function pointers. It was a clever harmless trick. Well, it was way more harmless than it was clever, but even then it caused UBSan to complain. And that's Not A Good Thing for various reasons, so let's just comply. The main bad effect of this change is a *slightly* larger list of exported symbols but until we actually get rid of the major bloats like <iostream>, <string> and the like, this is not going to have any measurable impact.
4 years ago
Implementation::nonOwnedArrayDeleter}};
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
}
char data[1];
AnimationTrackData track;
} importer;
auto data = importer.animation(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(static_cast<const void*>(data->data()), importer.data);
}
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void AbstractImporterTest::animationGrowableDeleters() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doAnimationCount() const override { return 1; }
Containers::Optional<AnimationData> doAnimation(UnsignedInt) override {
Containers::Array<char> data;
Containers::arrayAppend<ArrayAllocator>(data, '\x37');
return AnimationData{std::move(data), {AnimationTrackData{}}};
}
} importer;
auto data = importer.animation(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->data()[0], '\x37');
}
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
void AbstractImporterTest::animationCustomDataDeleter() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doAnimationCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doAnimationForName(Containers::StringView) override { return 0; }
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
Containers::Optional<AnimationData> doAnimation(UnsignedInt) override {
return AnimationData{Containers::Array<char>{nullptr, 0, [](char*, std::size_t) {}}, nullptr};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.animation(0);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.animation("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::animation(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::animation(): implementation is not allowed to use a custom Array deleter\n");
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
}
void AbstractImporterTest::animationCustomTrackDeleter() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doAnimationCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doAnimationForName(Containers::StringView) override { return 0; }
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
Containers::Optional<AnimationData> doAnimation(UnsignedInt) override {
return AnimationData{nullptr, Containers::Array<AnimationTrackData>{nullptr, 0, [](AnimationTrackData*, std::size_t) {}}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.animation(0);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.animation("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::animation(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::animation(): implementation is not allowed to use a custom Array deleter\n");
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::light() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doLightCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doLightForName(Containers::StringView name) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doLightName(UnsignedInt id) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Containers::Optional<LightData> doLight(UnsignedInt id) override {
if(id == 7) return LightData{{}, {}, {}, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.lightCount(), 8);
CORRADE_COMPARE(importer.lightForName("eighth"), 7);
CORRADE_COMPARE(importer.lightName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.light(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.light("eighth");
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::lightForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doLightCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doLightForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.lightForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::lightForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::lightNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doLightCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.lightName(7), "");
}
void AbstractImporterTest::lightNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doLightCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.lightName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::lightName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::lightNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doLightCount() const override { return 1; }
Containers::String doLightName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.lightName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::lightName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::lightNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doLightCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
importer.light(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::light(): not implemented\n");
}
void AbstractImporterTest::lightOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doLightCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.light(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::light(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::camera() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doCameraCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doCameraForName(Containers::StringView name) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doCameraName(UnsignedInt id) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Containers::Optional<CameraData> doCamera(UnsignedInt id) override {
Trade: camera type, projection size and aspect ratio support.
8 years ago
if(id == 7) return CameraData{{}, Vector2{}, {}, {}, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.cameraCount(), 8);
CORRADE_COMPARE(importer.cameraForName("eighth"), 7);
CORRADE_COMPARE(importer.cameraName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.camera(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.camera("eighth");
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::cameraForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doCameraCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doCameraForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.cameraForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::cameraForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::cameraNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doCameraCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.cameraName(7), "");
}
void AbstractImporterTest::cameraNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doCameraCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.cameraName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::cameraName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::cameraNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doCameraCount() const override { return 1; }
Containers::String doCameraName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.cameraName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::cameraName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::cameraNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doCameraCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
importer.camera(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::camera(): not implemented\n");
}
void AbstractImporterTest::cameraOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doCameraCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.camera(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::camera(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::object2D() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doObject2DCount() const override { return 8; }
Int doObject2DForName(const std::string& name) override {
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
std::string doObject2DName(UnsignedInt id) override {
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Get rid of std::unique_ptr.
7 years ago
Containers::Pointer<ObjectData2D> doObject2D(UnsignedInt id) override {
if(id == 7) return Containers::pointer(new ObjectData2D{{}, {}, &state});
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.object2DCount(), 8);
CORRADE_COMPARE(importer.object2DForName("eighth"), 7);
CORRADE_COMPARE(importer.object2DName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.object2D(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.object2D("eighth");
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::object2DCountNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
} importer;
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DCount(), 0);
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::object2DCountNoFile() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
CORRADE_IGNORE_DEPRECATED_PUSH
importer.object2DCount();
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object2DCount(): no file opened\n");
}
void AbstractImporterTest::object2DForNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
} importer;
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object2DForName(""), -1);
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::object2DForNameNoFile() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
CORRADE_IGNORE_DEPRECATED_PUSH
importer.object2DForName("");
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object2DForName(): no file opened\n");
}
void AbstractImporterTest::object2DByNameNotFound() {
auto&& data = ThingByNameData[testCaseInstanceId()];
setTestCaseDescription(data.name);
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doObject2DCount() const override { return 5; }
} importer;
std::ostringstream out;
{
Containers::Optional<Error> redirectError;
if(data.checkMessage) redirectError.emplace(&out);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_VERIFY(!importer.object2D("foobar"));
CORRADE_IGNORE_DEPRECATED_POP
}
if(data.checkMessage) CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::object2D(): object foobar not found among 5 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::object2DForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doObject2DCount() const override { return 8; }
Int doObject2DForName(const std::string&) override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
importer.object2DForName("");
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object2DForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::object2DNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doObject2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.object2DName(7), "");
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::object2DNameNoFile() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
CORRADE_IGNORE_DEPRECATED_PUSH
importer.object2DName(42);
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object2DName(): no file opened\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::object2DNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doObject2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.object2DName(8);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object2DName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::object2DNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedInt doSceneCount() const override { return 1; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
UnsignedInt doObject2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.object2D(7);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
/* It delegates to scene(), but since the assert is graceful and returns a
null optional, it errors out immediately after */
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::scene(): not implemented\n"
"Trade::AbstractImporter::object2D(): object 7 not found in any 2D scene hierarchy\n");
}
void AbstractImporterTest::object2DNoFile() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
CORRADE_IGNORE_DEPRECATED_PUSH
importer.object2D(42);
importer.object2D("foo");
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::object2D(): no file opened\n"
"Trade::AbstractImporter::object2D(): no file opened\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::object2DOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doObject2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.object2D(8);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object2D(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::object3D() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doObject3DCount() const override { return 8; }
Int doObject3DForName(const std::string& name) override {
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
std::string doObject3DName(UnsignedInt id) override {
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Get rid of std::unique_ptr.
7 years ago
Containers::Pointer<ObjectData3D> doObject3D(UnsignedInt id) override {
if(id == 7) return Containers::pointer(new ObjectData3D{{}, {}, &state});
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.object3DCount(), 8);
CORRADE_COMPARE(importer.object3DForName("eighth"), 7);
CORRADE_COMPARE(importer.object3DName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.object3D(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.object3D("eighth");
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::object3DCountNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
} importer;
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object3DCount(), 0);
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::object3DCountNoFile() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
CORRADE_IGNORE_DEPRECATED_PUSH
importer.object3DCount();
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object3DCount(): no file opened\n");
}
void AbstractImporterTest::object3DForNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
} importer;
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_COMPARE(importer.object3DForName(""), -1);
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::object3DForNameNoFile() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
CORRADE_IGNORE_DEPRECATED_PUSH
importer.object2DForName("");
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object2DForName(): no file opened\n");
}
void AbstractImporterTest::object3DByNameNotFound() {
auto&& data = ThingByNameData[testCaseInstanceId()];
setTestCaseDescription(data.name);
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doObject3DCount() const override { return 6; }
} importer;
std::ostringstream out;
{
Containers::Optional<Error> redirectError;
if(data.checkMessage) redirectError.emplace(&out);
CORRADE_IGNORE_DEPRECATED_PUSH
CORRADE_VERIFY(!importer.object3D("foobar"));
CORRADE_IGNORE_DEPRECATED_POP
}
if(data.checkMessage) CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::object3D(): object foobar not found among 6 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::object3DForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doObject3DCount() const override { return 8; }
Int doObject3DForName(const std::string&) override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
importer.object3DForName("");
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object3DForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::object3DNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doObject3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.object3DName(7), "");
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
}
void AbstractImporterTest::object3DNameNoFile() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
CORRADE_IGNORE_DEPRECATED_PUSH
importer.object3DName(42);
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object3DName(): no file opened\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::object3DNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doObject3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.object3DName(8);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object3DName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::object3DNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
UnsignedInt doSceneCount() const override { return 1; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
UnsignedInt doObject3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.object3D(7);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
/* It delegates to scene(), but since the assert is graceful and returns a
null optional, it errors out immediately after */
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::scene(): not implemented\n"
"Trade::AbstractImporter::object3D(): object 7 not found in any 3D scene hierarchy\n");
}
void AbstractImporterTest::object3DNoFile() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
std::ostringstream out;
Error redirectError{&out};
CORRADE_IGNORE_DEPRECATED_PUSH
importer.object3D(42);
importer.object3D("foo");
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::object3D(): no file opened\n"
"Trade::AbstractImporter::object3D(): no file opened\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::object3DOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doObject3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.object3D(8);
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::object3D(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: adapt AbstractImporter for the new SceneData workflow. Same as with MeshData2D/3D, the original ObjectData API and plugin interfaces are preserved to keep existing code as well as existing importer implementations working. As Magnum's own importers will get updated to the new SceneData workflow, a backward compatibility layer provided that translates it to the subset that the legacy ObjectData understands. With this commit, both existing plugin code can build (and test against) the new workflow, and any ports to the new workflow can test against the legacy interfaces. Except that for now the compatibility layer doesn't deal with objects that have more than one mesh or for example a light and a camera attached, this will be done in a separate step.
5 years ago
#endif
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: skin import in AbstractImporter.
6 years ago
void AbstractImporterTest::skin2D() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin2DForName(Containers::StringView name) override {
Trade: skin import in AbstractImporter.
6 years ago
if(name == "eighth") return 7;
return -1;
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doSkin2DName(UnsignedInt id) override {
Trade: skin import in AbstractImporter.
6 years ago
if(id == 7) return "eighth";
return {};
}
Containers::Optional<SkinData2D> doSkin2D(UnsignedInt id) override {
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
/* Verify that initializer list is converted to an array with
the default deleter and not something disallowed */
if(id == 7) return SkinData2D{{1}, {{}}, &state};
Trade: skin import in AbstractImporter.
6 years ago
return {};
}
} importer;
CORRADE_COMPARE(importer.skin2DCount(), 8);
CORRADE_COMPARE(importer.skin2DForName("eighth"), 7);
CORRADE_COMPARE(importer.skin2DName(7), "eighth");
{
auto data = importer.skin2D(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.skin2D("eighth");
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::skin2DForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin2DForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin2DForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin2DForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: skin import in AbstractImporter.
6 years ago
void AbstractImporterTest::skin2DNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 8; }
} importer;
CORRADE_COMPARE(importer.skin2DName(7), "");
}
void AbstractImporterTest::skin2DNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin2DName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin2DName(): index 8 out of range for 8 entries\n");
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::skin2DNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 1; }
Containers::String doSkin2DName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin2DName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin2DName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: skin import in AbstractImporter.
6 years ago
void AbstractImporterTest::skin2DNotImplemented() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin2D(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin2D(): not implemented\n");
}
void AbstractImporterTest::skin2DOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin2D(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin2D(): index 8 out of range for 8 entries\n");
}
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
void AbstractImporterTest::skin2DNonOwningDeleters() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin2DForName(Containers::StringView) override { return 0; }
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
Containers::Optional<SkinData2D> doSkin2D(UnsignedInt) override {
return SkinData2D{{}, jointData, {}, inverseBindMatrixData};
}
UnsignedInt jointData[1]{};
Matrix3 inverseBindMatrixData[1]{};
} importer;
std::ostringstream out;
Error redirectError{&out};
auto data = importer.skin2D(0);
CORRADE_COMPARE(data->joints(), importer.jointData);
CORRADE_COMPARE(data->inverseBindMatrices(), importer.inverseBindMatrixData);
}
void AbstractImporterTest::skin2DCustomJointDataDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin2DForName(Containers::StringView) override { return 0; }
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
Containers::Optional<SkinData2D> doSkin2D(UnsignedInt) override {
return SkinData2D{Containers::Array<UnsignedInt>{jointData, 1, [](UnsignedInt*, std::size_t){}}, Containers::Array<Matrix3>{1}};
}
UnsignedInt jointData[1]{};
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin2D(0);
importer.skin2D("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::skin2D(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::skin2D(): implementation is not allowed to use a custom Array deleter\n");
}
void AbstractImporterTest::skin2DCustomInverseBindMatrixDataDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin2DCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin2DForName(Containers::StringView) override { return 0; }
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
Containers::Optional<SkinData2D> doSkin2D(UnsignedInt) override {
return SkinData2D{Containers::Array<UnsignedInt>{1}, Containers::Array<Matrix3>{inverseBindMatrixData, 1, [](Matrix3*, std::size_t){}}};
}
Matrix3 inverseBindMatrixData[1]{};
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin2D(0);
importer.skin2D("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::skin2D(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::skin2D(): implementation is not allowed to use a custom Array deleter\n");
}
Trade: skin import in AbstractImporter.
6 years ago
void AbstractImporterTest::skin3D() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin3DForName(Containers::StringView name) override {
Trade: skin import in AbstractImporter.
6 years ago
if(name == "eighth") return 7;
return -1;
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doSkin3DName(UnsignedInt id) override {
Trade: skin import in AbstractImporter.
6 years ago
if(id == 7) return "eighth";
return {};
}
Containers::Optional<SkinData3D> doSkin3D(UnsignedInt id) override {
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
/* Verify that initializer list is converted to an array with
the default deleter and not something disallowed */
if(id == 7) return SkinData3D{{1}, {{}}, &state};
Trade: skin import in AbstractImporter.
6 years ago
return {};
}
} importer;
CORRADE_COMPARE(importer.skin3DCount(), 8);
CORRADE_COMPARE(importer.skin3DForName("eighth"), 7);
CORRADE_COMPARE(importer.skin3DName(7), "eighth");
{
auto data = importer.skin3D(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.skin3D("eighth");
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::skin3DForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin3DForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin3DForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin3DForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: skin import in AbstractImporter.
6 years ago
void AbstractImporterTest::skin3DNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 8; }
} importer;
CORRADE_COMPARE(importer.skin3DName(7), "");
}
void AbstractImporterTest::skin3DNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin3DName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin3DName(): index 8 out of range for 8 entries\n");
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::skin3DNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 1; }
Containers::String doSkin3DName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin3DName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin3DName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: skin import in AbstractImporter.
6 years ago
void AbstractImporterTest::skin3DNotImplemented() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin3D(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin3D(): not implemented\n");
}
void AbstractImporterTest::skin3DOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin3D(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::skin3D(): index 8 out of range for 8 entries\n");
}
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
void AbstractImporterTest::skin3DNonOwningDeleters() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin3DForName(Containers::StringView) override { return 0; }
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
Containers::Optional<SkinData3D> doSkin3D(UnsignedInt) override {
return SkinData3D{{}, jointData, {}, inverseBindMatrixData};
}
UnsignedInt jointData[1]{};
Matrix4 inverseBindMatrixData[1]{};
} importer;
std::ostringstream out;
Error redirectError{&out};
auto data = importer.skin3D(0);
CORRADE_COMPARE(data->joints(), importer.jointData);
CORRADE_COMPARE(data->inverseBindMatrices(), importer.inverseBindMatrixData);
}
void AbstractImporterTest::skin3DCustomJointDataDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin3DForName(Containers::StringView) override { return 0; }
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
Containers::Optional<SkinData3D> doSkin3D(UnsignedInt) override {
return SkinData3D{Containers::Array<UnsignedInt>{jointData, 1, [](UnsignedInt*, std::size_t){}}, Containers::Array<Matrix4>{1}};
}
UnsignedInt jointData[1]{};
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin3D(0);
importer.skin3D("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::skin3D(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::skin3D(): implementation is not allowed to use a custom Array deleter\n");
}
void AbstractImporterTest::skin3DCustomInverseBindMatrixDataDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doSkin3DCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doSkin3DForName(Containers::StringView) override { return 0; }
Trade: disallow custom deleters in AbstractImporter::skinXD().
6 years ago
Containers::Optional<SkinData3D> doSkin3D(UnsignedInt) override {
return SkinData3D{Containers::Array<UnsignedInt>{1}, Containers::Array<Matrix4>{inverseBindMatrixData, 1, [](Matrix4*, std::size_t){}}};
}
Matrix4 inverseBindMatrixData[1]{};
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.skin3D(0);
importer.skin3D("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::skin3D(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::skin3D(): implementation is not allowed to use a custom Array deleter\n");
}
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
void AbstractImporterTest::mesh() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
UnsignedInt doMeshLevelCount(UnsignedInt id) override {
if(id == 7) return 3;
Trade: minor test cleanup.
6 years ago
return {};
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMeshForName(Containers::StringView name) override {
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doMeshName(UnsignedInt id) override {
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
}
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
Containers::Optional<MeshData> doMesh(UnsignedInt id, UnsignedInt level) override {
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
/* Verify that initializer list is converted to an array with
the default deleter and not something disallowed */
Trade: allow specifying explicit vertex count on MeshData construction. Until now, except for an attribute-less index-less mesh, the vertex count was only implicitly taken from passed attributes, but it was severely limiting: - There was no way to set vertex count for an attribute-less indexed mesh, which didn't make sense - All code that made non-owning MeshData instances referencing another MeshData had to explicitly handle the attribute-less corner case to avoid vertex count getting lost - Offset-only attributes couldn't be used to specify static layout of meshes with dynamic vertex count, causing unnecessary extra allocations especially in the Primitives library.
6 years ago
if(id == 7 && level == 2) return MeshData{MeshPrimitive::Points, nullptr, {MeshAttributeData{MeshAttribute::Position, VertexFormat::Vector3, nullptr}}, MeshData::ImplicitVertexCount, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
}
} importer;
CORRADE_COMPARE(importer.meshCount(), 8);
CORRADE_COMPARE(importer.meshForName("eighth"), 7);
CORRADE_COMPARE(importer.meshName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
auto data = importer.mesh(7, 2);
Trade: allow retrieving importer data directly by name.
6 years ago
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
auto data = importer.mesh("eighth", 2);
Trade: allow retrieving importer data directly by name.
6 years ago
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
}
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
void AbstractImporterTest::meshDeprecatedFallback() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMeshForName(Containers::StringView name) override {
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doMeshName(UnsignedInt id) override {
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
}
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
Containers::Optional<MeshData> doMesh(UnsignedInt id, UnsignedInt level) override {
Trade: allow specifying explicit vertex count on MeshData construction. Until now, except for an attribute-less index-less mesh, the vertex count was only implicitly taken from passed attributes, but it was severely limiting: - There was no way to set vertex count for an attribute-less indexed mesh, which didn't make sense - All code that made non-owning MeshData instances referencing another MeshData had to explicitly handle the attribute-less corner case to avoid vertex count getting lost - Offset-only attributes couldn't be used to specify static layout of meshes with dynamic vertex count, causing unnecessary extra allocations especially in the Primitives library.
6 years ago
if(id == 7 && level == 0) return MeshData{MeshPrimitive::Points, nullptr, {MeshAttributeData{MeshAttribute::Position, VertexFormat::Vector3, nullptr}}, MeshData::ImplicitVertexCount, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
}
} importer;
CORRADE_IGNORE_DEPRECATED_PUSH
/* Nothing done for 2D as there were no known importers for these */
CORRADE_COMPARE(importer.mesh2DCount(), 0);
CORRADE_COMPARE(importer.mesh2DForName("eighth"), -1);
/* For 3D it's called through */
CORRADE_COMPARE(importer.mesh3DCount(), 8);
CORRADE_COMPARE(importer.mesh3DForName("eighth"), 7);
CORRADE_COMPARE(importer.mesh3DName(7), "eighth");
auto data = importer.mesh3D(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
CORRADE_IGNORE_DEPRECATED_POP
}
#endif
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
void AbstractImporterTest::meshLevelCountNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
} importer;
CORRADE_COMPARE(importer.meshLevelCount(7), 1);
}
void AbstractImporterTest::meshLevelCountOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.meshLevelCount(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::meshLevelCount(): index 8 out of range for 8 entries\n");
}
void AbstractImporterTest::meshLevelCountZero() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMeshForName(Containers::StringView) override { return 0; }
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
UnsignedInt doMeshLevelCount(UnsignedInt) override { return 0; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.meshLevelCount(7);
/* This should print a similar message instead of a confusing
"level 1 out of range for 0 entries" */
importer.mesh(7, 1);
importer.mesh("", 1);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::meshLevelCount(): implementation reported zero levels\n"
"Trade::AbstractImporter::mesh(): implementation reported zero levels\n"
"Trade::AbstractImporter::mesh(): implementation reported zero levels\n");
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::meshForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMeshForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.meshForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::meshForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
void AbstractImporterTest::meshNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
} importer;
CORRADE_COMPARE(importer.meshName(7), "");
}
void AbstractImporterTest::meshNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.meshName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::meshName(): index 8 out of range for 8 entries\n");
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::meshNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 1; }
Containers::String doMeshName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.meshName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::meshName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
void AbstractImporterTest::meshNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.mesh(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh(): not implemented\n");
}
void AbstractImporterTest::meshOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.mesh(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh(): index 8 out of range for 8 entries\n");
}
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
void AbstractImporterTest::meshLevelOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMeshForName(Containers::StringView) override { return 0; }
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
UnsignedInt doMeshLevelCount(UnsignedInt) override { return 3; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.mesh(7, 3);
importer.mesh("", 3);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::mesh(): level 3 out of range for 3 entries\n"
"Trade::AbstractImporter::mesh(): level 3 out of range for 3 entries\n");
}
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void AbstractImporterTest::meshNonOwningDeleters() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 1; }
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
Containers::Optional<MeshData> doMesh(UnsignedInt, UnsignedInt) override {
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
return MeshData{MeshPrimitive::Triangles,
Containers::Array<char>{indexData, 1, Implementation::nonOwnedArrayDeleter}, MeshIndexData{MeshIndexType::UnsignedByte, indexData},
Containers::Array<char>{nullptr, 0, Implementation::nonOwnedArrayDeleter},
meshAttributeDataNonOwningArray(attributes)};
}
char indexData[1];
MeshAttributeData attributes[1]{
MeshAttributeData{MeshAttribute::Position, VertexFormat::Vector3, nullptr}
};
} importer;
auto data = importer.mesh(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(static_cast<const void*>(data->indexData()), importer.indexData);
}
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void AbstractImporterTest::meshGrowableDeleters() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 1; }
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
Containers::Optional<MeshData> doMesh(UnsignedInt, UnsignedInt) override {
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
Containers::Array<char> indexData;
Containers::arrayAppend<ArrayAllocator>(indexData, '\xab');
Containers::Array<Vector3> vertexData;
Containers::arrayAppend<ArrayAllocator>(vertexData, Vector3{});
MeshIndexData indices{MeshIndexType::UnsignedByte, indexData};
MeshAttributeData positions{MeshAttribute::Position, Containers::arrayView(vertexData)};
return MeshData{MeshPrimitive::Triangles,
std::move(indexData), indices,
Containers::arrayAllocatorCast<char, ArrayAllocator>(std::move(vertexData)), {positions}};
}
} importer;
auto data = importer.mesh(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->indexData()[0], '\xab');
CORRADE_COMPARE(data->vertexData().size(), 12);
}
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
void AbstractImporterTest::meshCustomIndexDataDeleter() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMeshForName(Containers::StringView) override { return 0; }
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
Containers::Optional<MeshData> doMesh(UnsignedInt, UnsignedInt) override {
Trade: allow specifying explicit vertex count on MeshData construction. Until now, except for an attribute-less index-less mesh, the vertex count was only implicitly taken from passed attributes, but it was severely limiting: - There was no way to set vertex count for an attribute-less indexed mesh, which didn't make sense - All code that made non-owning MeshData instances referencing another MeshData had to explicitly handle the attribute-less corner case to avoid vertex count getting lost - Offset-only attributes couldn't be used to specify static layout of meshes with dynamic vertex count, causing unnecessary extra allocations especially in the Primitives library.
6 years ago
return MeshData{MeshPrimitive::Triangles, Containers::Array<char>{data, 1, [](char*, std::size_t) {}}, MeshIndexData{MeshIndexType::UnsignedByte, data}, 1};
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
}
char data[1];
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.mesh(0);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.mesh("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::mesh(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::mesh(): implementation is not allowed to use a custom Array deleter\n");
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
}
void AbstractImporterTest::meshCustomVertexDataDeleter() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMeshForName(Containers::StringView) override { return 0; }
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
Containers::Optional<MeshData> doMesh(UnsignedInt, UnsignedInt) override {
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
return MeshData{MeshPrimitive::Triangles, Containers::Array<char>{nullptr, 0, [](char*, std::size_t) {}}, {MeshAttributeData{MeshAttribute::Position, VertexFormat::Vector3, nullptr}}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.mesh(0);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.mesh("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::mesh(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::mesh(): implementation is not allowed to use a custom Array deleter\n");
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
}
void AbstractImporterTest::meshCustomAttributesDeleter() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMeshCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMeshForName(Containers::StringView) override { return 0; }
Trade: support mesh level import in AbstractImporter. Similar to image mip level import, but this is largely left to be importer-specific. For example PLY defines per-face data and sometimes one might want to import them as-is, without them being turned into a per-vertex property.
6 years ago
Containers::Optional<MeshData> doMesh(UnsignedInt, UnsignedInt) override {
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
return MeshData{MeshPrimitive::Triangles, nullptr, Containers::Array<MeshAttributeData>{&positions, 1, [](MeshAttributeData*, std::size_t) {}}};
}
MeshAttributeData positions{MeshAttribute::Position, VertexFormat::Vector3, nullptr};
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.mesh(0);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.mesh("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::mesh(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::mesh(): implementation is not allowed to use a custom Array deleter\n"
);
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
}
void AbstractImporterTest::meshAttributeName() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
MeshAttribute doMeshAttributeForName(Containers::StringView name) override {
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
if(name == "SMOOTH_GROUP_ID") return meshAttributeCustom(37);
return MeshAttribute{};
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doMeshAttributeName(UnsignedShort id) override {
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
if(id == 37) return "SMOOTH_GROUP_ID";
return "";
}
} importer;
CORRADE_COMPARE(importer.meshAttributeForName("SMOOTH_GROUP_ID"), meshAttributeCustom(37));
CORRADE_COMPARE(importer.meshAttributeName(meshAttributeCustom(37)), "SMOOTH_GROUP_ID");
}
void AbstractImporterTest::meshAttributeNameNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
} importer;
CORRADE_COMPARE(importer.meshAttributeForName(""), MeshAttribute{});
CORRADE_COMPARE(importer.meshAttributeName(meshAttributeCustom(37)), "");
}
void AbstractImporterTest::meshAttributeNameNotCustom() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
MeshAttribute doMeshAttributeForName(Containers::StringView) override {
Trade: implement Importer interfaces for the new MeshData. Deprecating of the old ones comes later.
6 years ago
return MeshAttribute::Position;
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.meshAttributeForName("SMOOTH_GROUP_ID");
importer.meshAttributeName(MeshAttribute::Position);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::meshAttributeForName(): implementation-returned Trade::MeshAttribute::Position is neither custom nor invalid\n"
"Trade::AbstractImporter::meshAttributeName(): Trade::MeshAttribute::Position is not custom\n");
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::meshAttributeNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
Containers::String doMeshAttributeName(UnsignedShort) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.meshAttributeName(meshAttributeCustom(0));
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::meshAttributeName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::mesh2D() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMesh2DCount() const override { return 8; }
Int doMesh2DForName(const std::string& name) override {
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
std::string doMesh2DName(UnsignedInt id) override {
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Containers::Optional<MeshData2D> doMesh2D(UnsignedInt id) override {
if(id == 7) return MeshData2D{{}, {}, {{}}, {}, {}, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.mesh2DCount(), 8);
CORRADE_COMPARE(importer.mesh2DForName("eighth"), 7);
CORRADE_COMPARE(importer.mesh2DName(7), "eighth");
auto data = importer.mesh2D(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh2DCountNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.mesh2DCount(), 0);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh2DCountNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh2DCount();
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh2DCount(): no file opened\n");
}
void AbstractImporterTest::mesh2DForNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.mesh2DForName(""), -1);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh2DForNameNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh2DForName("");
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh2DForName(): no file opened\n");
}
void AbstractImporterTest::mesh2DNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMesh2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.mesh2DName(7), "");
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh2DNameNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh2DName(42);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh2DName(): no file opened\n");
}
void AbstractImporterTest::mesh2DNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doMesh2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.mesh2DName(8);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh2DName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh2DNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMesh2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh2D(7);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh2D(): not implemented\n");
}
void AbstractImporterTest::mesh2DNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh2D(42);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh2D(): no file opened\n");
}
void AbstractImporterTest::mesh2DOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doMesh2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.mesh2D(8);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh2D(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh3D() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMesh3DCount() const override { return 8; }
Int doMesh3DForName(const std::string& name) override {
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
std::string doMesh3DName(UnsignedInt id) override {
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Containers::Optional<MeshData3D> doMesh3D(UnsignedInt id) override {
if(id == 7) return MeshData3D{{}, {}, {{}}, {}, {}, {}, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.mesh3DCount(), 8);
CORRADE_COMPARE(importer.mesh3DForName("eighth"), 7);
CORRADE_COMPARE(importer.mesh3DName(7), "eighth");
auto data = importer.mesh3D(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh3DCountNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.mesh3DCount(), 0);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh3DCountNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh3DCount();
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh3DCount(): no file opened\n");
}
void AbstractImporterTest::mesh3DForNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.mesh3DForName(""), -1);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh3DForNameNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh3DForName("");
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh3DForName(): no file opened\n");
}
void AbstractImporterTest::mesh3DNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMesh3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.mesh3DName(7), "");
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh3DNameNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh3DName(42);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh3DName(): no file opened\n");
}
void AbstractImporterTest::mesh3DNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doMesh3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.mesh3DName(8);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh3DName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh3DNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMesh3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh3D(7);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
/* Not mesh3D() because this one delegates into mesh() for backwards
compatibility */
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh(): not implemented\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::mesh3DNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
importer.mesh3D(42);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh3D(): no file opened\n");
}
void AbstractImporterTest::mesh3DOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doMesh3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_PUSH
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.mesh3D(8);
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
CORRADE_IGNORE_DEPRECATED_POP
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::mesh3D(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: deprecate AbstractImporter interfaces for MeshDataXD. For backwards compatibility these will delegate to the new MeshData interfaces for 3D (and nothing for 2D, because so far there were no 2D scene importers).
7 years ago
#endif
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::material() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMaterialCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMaterialForName(Containers::StringView name) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doMaterialName(UnsignedInt id) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: return Optional instead of Pointer from *Importer::material(). The plugin interface version got bumped to avoid ABI issues when loading plugins that weren't updated for the change, but apart from that this shouldn't be a breaking change, as the API returns a type that can be both an Optional and a Pointer.
6 years ago
Containers::Optional<MaterialData> doMaterial(UnsignedInt id) override {
if(id == 7) return Containers::optional<MaterialData>(MaterialTypes{}, nullptr, &state);
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.materialCount(), 8);
CORRADE_COMPARE(importer.materialForName("eighth"), 7);
CORRADE_COMPARE(importer.materialName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.material(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.material("eighth");
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: return Optional instead of Pointer from *Importer::material(). The plugin interface version got bumped to avoid ABI issues when loading plugins that weren't updated for the change, but apart from that this shouldn't be a breaking change, as the API returns a type that can be both an Optional and a Pointer.
6 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
void AbstractImporterTest::materialDeprecatedFallback() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMaterialCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMaterialForName(Containers::StringView) override { return 0; }
Trade: return Optional instead of Pointer from *Importer::material(). The plugin interface version got bumped to avoid ABI issues when loading plugins that weren't updated for the change, but apart from that this shouldn't be a breaking change, as the API returns a type that can be both an Optional and a Pointer.
6 years ago
/* Using a deprecated PhongMaterialData constructor to verify that
propagating such instance works as well (array deleters etc.) */
Containers::Optional<MaterialData> doMaterial(UnsignedInt) override {
CORRADE_IGNORE_DEPRECATED_PUSH
return Containers::Optional<MaterialData>{std::move(PhongMaterialData{{},
{}, {},
{}, {},
{}, {},
{},
{}, {}, {}, {}, &state
})};
CORRADE_IGNORE_DEPRECATED_POP
}
} importer;
{
CORRADE_IGNORE_DEPRECATED_PUSH
Containers::Pointer<MaterialData> data = importer.material(0);
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_VERIFY(data);
CORRADE_COMPARE_AS(data->attributeCount(), 0, TestSuite::Compare::Greater);
CORRADE_COMPARE(data->importerState(), &state);
} {
CORRADE_IGNORE_DEPRECATED_PUSH
Containers::Pointer<MaterialData> data = importer.material("");
CORRADE_IGNORE_DEPRECATED_POP
CORRADE_VERIFY(data);
CORRADE_COMPARE_AS(data->attributeCount(), 0, TestSuite::Compare::Greater);
CORRADE_COMPARE(data->importerState(), &state);
}
}
#endif
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::materialForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMaterialCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMaterialForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.materialForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::materialForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::materialNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMaterialCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.materialName(7), "");
}
void AbstractImporterTest::materialNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doMaterialCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.materialName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::materialName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::materialNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMaterialCount() const override { return 1; }
Containers::String doMaterialName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.materialName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::materialName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::materialNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMaterialCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
importer.material(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::material(): not implemented\n");
}
void AbstractImporterTest::materialOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doMaterialCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.material(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::material(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: check for custom MaterialData deleters in AbstractImporter. And adapt MaterialData and the deprecated PhongMaterialData constructor to follow these rules. Basically the same as done for ImageData and MeshData.
6 years ago
void AbstractImporterTest::materialNonOwningDeleters() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMaterialCount() const override { return 1; }
Containers::Optional<MaterialData> doMaterial(UnsignedInt) override {
return MaterialData{{}, {}, attributeData, {}, layerData};
}
UnsignedInt layerData[1]{};
MaterialAttributeData attributeData[1]{
{MaterialAttribute::DiffuseColor, Color4{}}
};
} importer;
auto data = importer.material(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(static_cast<const void*>(data->attributeData()), importer.attributeData);
CORRADE_COMPARE(static_cast<const void*>(data->layerData()), importer.layerData);
}
void AbstractImporterTest::materialCustomAttributeDataDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMaterialCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMaterialForName(Containers::StringView) override { return 0; }
Trade: check for custom MaterialData deleters in AbstractImporter. And adapt MaterialData and the deprecated PhongMaterialData constructor to follow these rules. Basically the same as done for ImageData and MeshData.
6 years ago
Containers::Optional<MaterialData> doMaterial(UnsignedInt) override {
return MaterialData{{}, Containers::Array<MaterialAttributeData>{attributeData, 1, [](MaterialAttributeData*, std::size_t) {}}};
}
MaterialAttributeData attributeData[1]{
{MaterialAttribute::DiffuseColor, Color4{}}
};
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.material(0);
importer.material("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::material(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::material(): implementation is not allowed to use a custom Array deleter\n");
}
void AbstractImporterTest::materialCustomLayerDataDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doMaterialCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doMaterialForName(Containers::StringView) override { return 0; }
Trade: check for custom MaterialData deleters in AbstractImporter. And adapt MaterialData and the deprecated PhongMaterialData constructor to follow these rules. Basically the same as done for ImageData and MeshData.
6 years ago
Containers::Optional<MaterialData> doMaterial(UnsignedInt) override {
return MaterialData{{}, nullptr, Containers::Array<UnsignedInt>{layerData, 1, [](UnsignedInt*, std::size_t) {}}};
}
UnsignedInt layerData[1]{};
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.material(0);
importer.material("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::material(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::material(): implementation is not allowed to use a custom Array deleter\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::texture() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doTextureCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doTextureForName(Containers::StringView name) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doTextureName(UnsignedInt id) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Containers::Optional<TextureData> doTexture(UnsignedInt id) override {
if(id == 7) return TextureData{{}, {}, {}, {}, {}, {}, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.textureCount(), 8);
CORRADE_COMPARE(importer.textureForName("eighth"), 7);
CORRADE_COMPARE(importer.textureName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.texture(7);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.texture("eighth");
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::textureForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doTextureCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doTextureForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.textureForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::textureForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::textureNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doTextureCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.textureName(7), "");
}
void AbstractImporterTest::textureNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doTextureCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
importer.textureName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::textureName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::textureNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doTextureCount() const override { return 1; }
Containers::String doTextureName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.textureName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::textureName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
void AbstractImporterTest::textureNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
bool doIsOpened() const override { return true; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void doClose() override {}
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
UnsignedInt doTextureCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: reduce needless duplication in AbstractImporter test. Should have done this since the beginning, would save my carpal tunnels a bit.
6 years ago
importer.texture(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::texture(): not implemented\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::textureOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doTextureCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.texture(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::texture(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
void AbstractImporterTest::image1D() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 8; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
UnsignedInt doImage1DLevelCount(UnsignedInt id) override {
if(id == 7) return 3;
Trade: minor test cleanup.
6 years ago
return {};
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage1DForName(Containers::StringView name) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doImage1DName(UnsignedInt id) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
Containers::Optional<ImageData1D> doImage1D(UnsignedInt id, UnsignedInt level) override {
if(id == 7 && level == 2) return ImageData1D{PixelFormat::RGBA8Unorm, {}, {}, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.image1DCount(), 8);
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
CORRADE_COMPARE(importer.image1DLevelCount(7), 3);
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.image1DForName("eighth"), 7);
CORRADE_COMPARE(importer.image1DName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.image1D(7, 2);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.image1D("eighth", 2);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void AbstractImporterTest::image1DLevelCountNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 8; }
} importer;
CORRADE_COMPARE(importer.image1DLevelCount(7), 1);
}
void AbstractImporterTest::image1DLevelCountOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image1DLevelCount(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image1DLevelCount(): index 8 out of range for 8 entries\n");
}
void AbstractImporterTest::image1DLevelCountZero() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage1DForName(Containers::StringView) override { return 0; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
UnsignedInt doImage1DLevelCount(UnsignedInt) override { return 0; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image1DLevelCount(7);
/* This should print a similar message instead of a confusing
"level 1 out of range for 0 entries" */
importer.image1D(7, 1);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.image1D("", 1);
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::image1DLevelCount(): implementation reported zero levels\n"
Trade: allow retrieving importer data directly by name.
6 years ago
"Trade::AbstractImporter::image1D(): implementation reported zero levels\n"
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
"Trade::AbstractImporter::image1D(): implementation reported zero levels\n");
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::image1DForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage1DForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image1DForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image1DForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::image1DNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.image1DName(7), "");
}
void AbstractImporterTest::image1DNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doImage1DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.image1DName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image1DName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::image1DNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 1; }
Containers::String doImage1DName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image1DName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image1DName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::image1DNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
importer.image1D(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image1D(): not implemented\n");
}
void AbstractImporterTest::image1DOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doImage1DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.image1D(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image1D(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void AbstractImporterTest::image1DLevelOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage1DForName(Containers::StringView) override { return 0; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
UnsignedInt doImage1DLevelCount(UnsignedInt) override { return 3; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image1D(7, 3);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.image1D("", 3);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::image1D(): level 3 out of range for 3 entries\n"
"Trade::AbstractImporter::image1D(): level 3 out of range for 3 entries\n");
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
}
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void AbstractImporterTest::image1DNonOwningDeleter() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 1; }
Containers::Optional<ImageData1D> doImage1D(UnsignedInt, UnsignedInt) override {
return ImageData1D{PixelFormat::RGBA8Unorm, {}, Containers::Array<char>{data, 1, Implementation::nonOwnedArrayDeleter}};
}
char data[1];
} importer;
auto data = importer.image1D(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(static_cast<const void*>(data->data()), importer.data);
}
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void AbstractImporterTest::image1DGrowableDeleter() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 1; }
Containers::Optional<ImageData1D> doImage1D(UnsignedInt, UnsignedInt) override {
Containers::Array<char> data;
Containers::arrayAppend<ArrayAllocator>(data, '\xff');
return ImageData1D{PixelFormat::RGBA8Unorm, {}, std::move(data)};
}
} importer;
auto data = importer.image1D(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->data()[0], '\xff');
}
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
void AbstractImporterTest::image1DCustomDeleter() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage1DCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage1DForName(Containers::StringView) override { return 0; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
Containers::Optional<ImageData1D> doImage1D(UnsignedInt, UnsignedInt) override {
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
return ImageData1D{PixelFormat::RGBA8Unorm, {}, Containers::Array<char>{nullptr, 0, [](char*, std::size_t) {}}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image1D(0);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.image1D("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::image1D(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::image1D(): implementation is not allowed to use a custom Array deleter\n");
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::image2D() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 8; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
UnsignedInt doImage2DLevelCount(UnsignedInt id) override {
if(id == 7) return 3;
Trade: minor test cleanup.
6 years ago
return {};
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage2DForName(Containers::StringView name) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doImage2DName(UnsignedInt id) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
Containers::Optional<ImageData2D> doImage2D(UnsignedInt id, UnsignedInt level) override {
if(id == 7 && level == 2) return ImageData2D{PixelFormat::RGBA8Unorm, {}, {}, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.image2DCount(), 8);
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
CORRADE_COMPARE(importer.image2DLevelCount(7), 3);
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.image2DForName("eighth"), 7);
CORRADE_COMPARE(importer.image2DName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.image2D(7, 2);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.image2D("eighth", 2);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void AbstractImporterTest::image2DLevelCountNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 8; }
} importer;
CORRADE_COMPARE(importer.image2DLevelCount(7), 1);
}
void AbstractImporterTest::image2DLevelCountOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image2DLevelCount(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image2DLevelCount(): index 8 out of range for 8 entries\n");
}
void AbstractImporterTest::image2DLevelCountZero() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage2DForName(Containers::StringView) override { return 0; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
UnsignedInt doImage2DLevelCount(UnsignedInt) override { return 0; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image2DLevelCount(7);
/* This should print a similar message instead of a confusing
"level 1 out of range for 0 entries" */
importer.image2D(7, 1);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.image2D(7, 1);
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::image2DLevelCount(): implementation reported zero levels\n"
Trade: allow retrieving importer data directly by name.
6 years ago
"Trade::AbstractImporter::image2D(): implementation reported zero levels\n"
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
"Trade::AbstractImporter::image2D(): implementation reported zero levels\n");
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::image2DForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage2DForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image2DForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image2DForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::image2DNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.image2DName(7), "");
}
void AbstractImporterTest::image2DNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doImage2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.image2DName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image2DName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::image2DNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 1; }
Containers::String doImage2DName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image2DName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image2DName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::image2DNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
importer.image2D(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image2D(): not implemented\n");
}
void AbstractImporterTest::image2DOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doImage2DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.image2D(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image2D(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void AbstractImporterTest::image2DLevelOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage2DForName(Containers::StringView) override { return 0; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
UnsignedInt doImage2DLevelCount(UnsignedInt) override { return 3; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image2D(7, 3);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.image2D("", 3);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::image2D(): level 3 out of range for 3 entries\n"
"Trade::AbstractImporter::image2D(): level 3 out of range for 3 entries\n");
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
}
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void AbstractImporterTest::image2DNonOwningDeleter() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 1; }
Containers::Optional<ImageData2D> doImage2D(UnsignedInt, UnsignedInt) override {
return ImageData2D{PixelFormat::RGBA8Unorm, {}, Containers::Array<char>{data, 1, Implementation::nonOwnedArrayDeleter}};
}
char data[1];
} importer;
auto data = importer.image2D(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(static_cast<const void*>(data->data()), importer.data);
}
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void AbstractImporterTest::image2DGrowableDeleter() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 1; }
Containers::Optional<ImageData2D> doImage2D(UnsignedInt, UnsignedInt) override {
Containers::Array<char> data;
Containers::arrayAppend<ArrayAllocator>(data, '\xff');
return ImageData2D{PixelFormat::RGBA8Unorm, {}, std::move(data)};
}
} importer;
auto data = importer.image2D(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->data()[0], '\xff');
}
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
void AbstractImporterTest::image2DCustomDeleter() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage2DCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage2DForName(Containers::StringView) override { return 0; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
Containers::Optional<ImageData2D> doImage2D(UnsignedInt, UnsignedInt) override {
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
return ImageData2D{PixelFormat::RGBA8Unorm, {}, Containers::Array<char>{nullptr, 0, [](char*, std::size_t) {}}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image2D(0);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.image2D("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::image2D(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::image2D(): implementation is not allowed to use a custom Array deleter\n");
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::image3D() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 8; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
UnsignedInt doImage3DLevelCount(UnsignedInt id) override {
if(id == 7) return 3;
Trade: minor test cleanup.
6 years ago
return {};
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage3DForName(Containers::StringView name) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(name == "eighth") return 7;
Trade: minor test cleanup.
6 years ago
return -1;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Containers::String doImage3DName(UnsignedInt id) override {
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
if(id == 7) return "eighth";
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
Containers::Optional<ImageData3D> doImage3D(UnsignedInt id, UnsignedInt level) override {
if(id == 7 && level == 2) return ImageData3D{PixelFormat::RGBA8Unorm, {}, {}, &state};
Trade: minor test cleanup.
6 years ago
return {};
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.image3DCount(), 8);
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
CORRADE_COMPARE(importer.image3DLevelCount(7), 3);
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.image3DForName("eighth"), 7);
CORRADE_COMPARE(importer.image3DName(7), "eighth");
Trade: allow retrieving importer data directly by name.
6 years ago
{
auto data = importer.image3D(7, 2);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
} {
auto data = importer.image3D("eighth", 2);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->importerState(), &state);
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
void AbstractImporterTest::image3DForNameOutOfRange() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage3DForName(Containers::StringView) override { return 8; }
Trade: ensure AbstractImporter::fooForName() returns indices in bounds. I just accidentally returned a wrong index from a SceneData backwards compatibility wrapper and realized it'd be good to have that checked always.
5 years ago
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image3DForName("");
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image3DForName(): implementation-returned index 8 out of range for 8 entries\n");
}
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void AbstractImporterTest::image3DLevelCountNotImplemented() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 8; }
} importer;
CORRADE_COMPARE(importer.image3DLevelCount(7), 1);
}
void AbstractImporterTest::image3DLevelCountOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 8; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image3DLevelCount(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image3DLevelCount(): index 8 out of range for 8 entries\n");
}
void AbstractImporterTest::image3DLevelCountZero() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage3DForName(Containers::StringView) override { return 0; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
UnsignedInt doImage3DLevelCount(UnsignedInt) override { return 0; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image3DLevelCount(7);
/* This should print a similar message instead of a confusing
"level 1 out of range for 0 entries" */
importer.image3D(7, 1);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.image3D("", 1);
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::image3DLevelCount(): implementation reported zero levels\n"
Trade: allow retrieving importer data directly by name.
6 years ago
"Trade::AbstractImporter::image3D(): implementation reported zero levels\n"
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
"Trade::AbstractImporter::image3D(): implementation reported zero levels\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::image3DNameNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.image3DName(7), "");
}
void AbstractImporterTest::image3DNameOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doImage3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.image3DName(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image3DName(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
void AbstractImporterTest::image3DNameCustomDeleter() {
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 1; }
Containers::String doImage3DName(UnsignedInt) override {
return Containers::String{"a", 1, [](char*, std::size_t) {}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image3DName(0);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image3DName(): implementation is not allowed to use a custom String deleter\n");
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::image3DNotImplemented() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
importer.image3D(7);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image3D(): not implemented\n");
}
void AbstractImporterTest::image3DOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
UnsignedInt doImage3DCount() const override { return 8; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
Trade: provide less useless OOB assert messages. If this would be there yesterday already, it would made my debugging session of `image2D(-1)` immediately obvious (instead of wasting 20 minutes).
7 years ago
importer.image3D(8);
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::image3D(): index 8 out of range for 8 entries\n");
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
}
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
void AbstractImporterTest::image3DLevelOutOfRange() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 8; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage3DForName(Containers::StringView) override { return 0; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
UnsignedInt doImage3DLevelCount(UnsignedInt) override { return 3; }
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image3D(7, 3);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.image3D("", 3);
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::image3D(): level 3 out of range for 3 entries\n"
"Trade::AbstractImporter::image3D(): level 3 out of range for 3 entries\n");
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
}
Trade: allow non-owning aray deleters passed through Importer APIs.
7 years ago
void AbstractImporterTest::image3DNonOwningDeleter() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 1; }
Containers::Optional<ImageData3D> doImage3D(UnsignedInt, UnsignedInt) override {
return ImageData3D{PixelFormat::RGBA8Unorm, {}, Containers::Array<char>{data, 1, Implementation::nonOwnedArrayDeleter}};
}
char data[1];
} importer;
auto data = importer.image3D(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(static_cast<const void*>(data->data()), importer.data);
}
Trade: whitelist (exported) growable array deleters in imported data.
7 years ago
void AbstractImporterTest::image3DGrowableDeleter() {
struct: AbstractImporter {
ImporterFeatures doFeatures() const override { return {}; }
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 1; }
Containers::Optional<ImageData3D> doImage3D(UnsignedInt, UnsignedInt) override {
Containers::Array<char> data;
Containers::arrayAppend<ArrayAllocator>(data, '\xff');
return ImageData3D{PixelFormat::RGBA8Unorm, {}, std::move(data)};
}
} importer;
auto data = importer.image3D(0);
CORRADE_VERIFY(data);
CORRADE_COMPARE(data->data()[0], '\xff');
}
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
void AbstractImporterTest::image3DCustomDeleter() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
UnsignedInt doImage3DCount() const override { return 1; }
Trade: port AbstractImporter away from std::string. Except for file callbacks, for these I have another change planned for zero-copy import and it would be unwise to break stuff twice, providing two sets of backwards compatibility wrappers. The image / scene converter plugins went through a similar change earlier already and the shader converters were made sane since the very beginning. OTOH audio importers and text stuff are scheduled for merging with Trade or a larger rework anyway, so I didn't see any point in updating those. It's mostly a trivial change, except that returned String instances are now also checked for non-default deleters same as Arrays because yes, wow such flexibility compared to STL strings. Same was done for ShaderTools::AbstractConverter already anyway, so nothing unheard-of either. The importer plugin interface version is bumped as this likely breaks ABI in a nasty way that would lead to crashes.
4 years ago
Int doImage3DForName(Containers::StringView) override { return 0; }
Trade: support mip levels in image import. There's a new level option in each imageND() API, plus an imageNDLevelCount() query to get image level count.
6 years ago
Containers::Optional<ImageData3D> doImage3D(UnsignedInt, UnsignedInt) override {
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
return ImageData3D{PixelFormat::RGBA8Unorm, {}, Containers::Array<char>{nullptr, 0, [](char*, std::size_t) {}}};
}
} importer;
std::ostringstream out;
Error redirectError{&out};
importer.image3D(0);
Trade: allow retrieving importer data directly by name.
6 years ago
importer.image3D("");
CORRADE_COMPARE(out.str(),
"Trade::AbstractImporter::image3D(): implementation is not allowed to use a custom Array deleter\n"
"Trade::AbstractImporter::image3D(): implementation is not allowed to use a custom Array deleter\n");
Trade, Audio: explicitly disallow custom deleters returned from plugins.
7 years ago
}
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
void AbstractImporterTest::importerState() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
const void* doImporterState() const override { return &state; }
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.importerState(), &state);
}
void AbstractImporterTest::importerStateNotImplemented() {
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return true; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
CORRADE_COMPARE(importer.importerState(), nullptr);
}
void AbstractImporterTest::importerStateNoFile() {
Make this compile and test cleanly with CORRADE_NO_ASSERT defined.
6 years ago
#ifdef CORRADE_NO_ASSERT
CORRADE_SKIP("CORRADE_NO_ASSERT defined, can't test assertions");
#endif
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
struct: AbstractImporter {
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
ImporterFeatures doFeatures() const override { return {}; }
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
bool doIsOpened() const override { return false; }
void doClose() override {}
Trade: clean up unnecessary cruft in the AbstractImporter test.
7 years ago
} importer;
Trade: exhaustive test for AbstractImporter. Also fixed typos and copypaste errors in assertion messages; added assertions to functions that should never get called.
8 years ago
std::ostringstream out;
Error redirectError{&out};
importer.importerState();
CORRADE_COMPARE(out.str(), "Trade::AbstractImporter::importerState(): no file opened\n");
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
}
Trade: debug output for AbstractImporter::Feature and Features.
8 years ago
void AbstractImporterTest::debugFeature() {
std::ostringstream out;
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
Debug{&out} << ImporterFeature::OpenData << ImporterFeature(0xf0);
CORRADE_COMPARE(out.str(), "Trade::ImporterFeature::OpenData Trade::ImporterFeature(0xf0)\n");
Trade: debug output for AbstractImporter::Feature and Features.
8 years ago
}
void AbstractImporterTest::debugFeatures() {
std::ostringstream out;
Move Feature[s] enums from inside plugin classes outside. Too annoyingly long to type. Also added a debug output operator for the FontConverterFeature enums, for some reason there were none.
8 years ago
Debug{&out} << (ImporterFeature::OpenData|ImporterFeature::OpenState) << ImporterFeatures{};
CORRADE_COMPARE(out.str(), "Trade::ImporterFeature::OpenData|Trade::ImporterFeature::OpenState Trade::ImporterFeatures{}\n");
Trade: debug output for AbstractImporter::Feature and Features.
8 years ago
}
Trade: add flags to AbstractImporter and AbstractImageConverter.
6 years ago
void AbstractImporterTest::debugFlag() {
std::ostringstream out;
Debug{&out} << ImporterFlag::Verbose << ImporterFlag(0xf0);
CORRADE_COMPARE(out.str(), "Trade::ImporterFlag::Verbose Trade::ImporterFlag(0xf0)\n");
}
void AbstractImporterTest::debugFlags() {
std::ostringstream out;
Debug{&out} << (ImporterFlag::Verbose|ImporterFlag(0xf0)) << ImporterFlags{};
CORRADE_COMPARE(out.str(), "Trade::ImporterFlag::Verbose|Trade::ImporterFlag(0xf0) Trade::ImporterFlags{}\n");
}
Wrap all tests in unnamed namespaces. This makes the compiler warn about test cases that aren't ever used, which is a good thing.
7 years ago
}}}}
Trade: reworked AbstractImporter plugin interface. Implementation is moved into private virtual `do*()` functions and the public interface does additional checks aroung them to simplify plugin development. Opening files is by default done by the base implementation, which then calls function for opening raw data with file contents. Added test for file opening, bumped plugin interface version to 0.3. Y# Please enter the commit message for your changes. Lines starting
13 years ago
CORRADE_TEST_MAIN(Magnum::Trade::Test::AbstractImporterTest)