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.
11214 Commits
21 Branches
19 Tags
70 MiB
Tree: 9b4d394a14
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '9b4d394a14'
${ noResults }
magnum/src/Magnum/GL/AbstractTexture.h

851 lines
48 KiB
Raw Normal View History Unescape

Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
#ifndef Magnum_GL_AbstractTexture_h
#define Magnum_GL_AbstractTexture_h
Splitting non-templated things from Texture to AbstractTexture.
15 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,
It's October already?!
2 years ago
2020, 2021, 2022, 2023, 2024
Vladimír Vondruš <mosra@centrum.cz>
Relicensing to MIT/Expat license, part 2: headers.
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.
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
*/
/** @file
Split the OpenGL layer out, pt 4: new Magnum::GL namespace. At the moment just the GL library itself w/o the tests, and without backwards compatibility aliases. The following types were left in the root namespace, despite being in the GL/ directory, as they will get moved back soon: * Image, CompressedImage and their dimensional typedefs * ImageView, CompressedImageView and their dimensional typedefs * PixelStorage Not PixelFormat etc., that one will stay in the GL namespace and a completely new PixelFormat enum will be provided in the root namespace.
8 years ago
* @brief Class @ref Magnum::GL::AbstractTexture
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
*/
Use Utility::move() and swap() instead of STL almost everywhere. Partially needed to avoid build breakages because Corrade itself switched as well, partially because a cleanup is always good. Done except for (STL-heavy) code that's deprecated or SceneGraph-related APIs that are still quite full of STL as well.
3 years ago
#include <initializer_list>
#include <Corrade/Utility/Move.h>
Allocation-free overloads for *::setLabel(). In most cases the label is set directly from code, e.g.: texture.setLabel("diffuse-duck"); Avoiding conversion to std::string and passing char(&)[size] directly will avoid one allocation and deallocation. Better solution would be to use std::string_view everywhere, but we're not in C++17 yet.
12 years ago
Support for ARB_get_texture_sub_image. Only thing missing are compressed texture queries, but that needs a lot more work.
11 years ago
#include "Magnum/DimensionTraits.h"
GL: fill appropriate ImageFlags when using Texture::image() and friends. Array texture and cubemap (array) subimage queries get ImageFlag*D::Array, cubemap whole image ImageFlag3D::CubeMap, cubemap array whole image get ImageFlag3D::CubeMap and ImageFlag3D::Array. No flags are checked when uploading images or when downloading images to views -- using an array texture as a cube map and vice versa is a valid case, and others are probably also, so there's no point. The flag restrictions will come into place when ImageFlag*D::YUp / YDown etc. are a thing.
4 years ago
#include "Magnum/ImageFlags.h"
Ability to instantiate OpenGL objects without creating them. Similarly to what's now done with NoInit tags for Containers::Array and all math types such as Vector, there's now NoCreate tag for creating wrappers without actually creating the underlying OpenGL object. The instance is then equivalent to moved-from state. Useful to avoid needless creation/deletion of OpenGL object in case you would overwrite the instance later anyway: Mesh mesh{NoCreate}; std::unique_ptr<Buffer> indices, vertices; std::tie(mesh, indices, vertices) = MeshTools:compile(...);
11 years ago
#include "Magnum/Tags.h"
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
#include "Magnum/GL/AbstractObject.h"
#include "Magnum/GL/GL.h"
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
GL: port label() / setLabel() away from std::string. All the tests were updated to explicitly check that non-null-terminated strings get handled properly (the GL label APIs have an explicit size, so it *should*, but just in case). Also, because various subclasses override the setter to return correct type for method chaining and the override has to be deinlined to avoid relying on a StringView include, the tests are now explicitly done for each leaf class, instead of the subclass The <string> being removed from the base class for all GL objects may affect downstream projects which relied on it being included. In case of Magnum, the breakages were already fixed in the previous commit. Compile time improvement for the MagnumGL library alone is 0.2 second or 4% (6.1 seconds before, 5.9 after). Not bad, given that there's three more files to compile and strings are still heavily used in other GL debug output APIs and all shader stuff. For a build of just the GL library and all tests, it goes down from 28.9 seconds to 28.1. Most tests also still rely quite heavily on std::stringstream for debug output testing, so the numbers still could go further.
5 years ago
#ifdef MAGNUM_BUILD_DEPRECATED
/* For label() / setLabel(), which used to be a std::string. Not ideal for the
return type, but at least something. */
#include <Corrade/Containers/StringStl.h>
#endif
Split the OpenGL layer out, pt 4: new Magnum::GL namespace. At the moment just the GL library itself w/o the tests, and without backwards compatibility aliases. The following types were left in the root namespace, despite being in the GL/ directory, as they will get moved back soon: * Image, CompressedImage and their dimensional typedefs * ImageView, CompressedImageView and their dimensional typedefs * PixelStorage Not PixelFormat etc., that one will stay in the GL namespace and a completely new PixelFormat enum will be provided in the root namespace.
8 years ago
namespace Magnum { namespace GL {
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
Implemented ARB_texture_swizzle.
12 years ago
namespace Implementation {
struct TextureState;
#ifndef MAGNUM_TARGET_GLES2
template<char> struct TextureSwizzle;
template<> struct TextureSwizzle<'r'> { enum: GLint { Value = GL_RED }; };
template<> struct TextureSwizzle<'g'> { enum: GLint { Value = GL_GREEN }; };
template<> struct TextureSwizzle<'b'> { enum: GLint { Value = GL_BLUE }; };
template<> struct TextureSwizzle<'a'> { enum: GLint { Value = GL_ALPHA }; };
template<> struct TextureSwizzle<'0'> { enum: GLint { Value = GL_ZERO }; };
template<> struct TextureSwizzle<'1'> { enum: GLint { Value = GL_ONE }; };
#endif
}
Moved global data from texture classes to per-context state.
12 years ago
Revert "Documentation: organizing the classes into "modules"." It only caused another maintenance burden and was confusing to users. Now when scene graph is in SceneGraph namespace there is no need for another grouping. Namespaces are (and should be) sufficient. This reverts commit 79945ab6fce4bf0b53faab1d2f694f4cd934d562. Conflicts: src/BufferedImage.h src/BufferedTexture.h src/Framebuffer.h src/Query.h src/SceneGraph/Scene.h
14 years ago
/**
Texture documentation reorganization. The part similar for all AbstractTexture subclasses should be in AbstractTexture itself, not anywhere else.
14 years ago
@brief Base for textures
GL: minor doc improvements for textures. Not great, not terrible.
5 years ago
Wraps an OpenGL texture object. Meant to be only used through
subclasses, see @ref Texture, @ref TextureArray, @ref CubeMapTexture,
@ref CubeMapTextureArray, @ref RectangleTexture, @ref BufferTexture and
@ref MultisampleTexture documentation for more information and usage examples.
Texture documentation reorganization. The part similar for all AbstractTexture subclasses should be in AbstractTexture itself, not anywhere else.
14 years ago
Split the OpenGL layer out, pt 4: new Magnum::GL namespace. At the moment just the GL library itself w/o the tests, and without backwards compatibility aliases. The following types were left in the root namespace, despite being in the GL/ directory, as they will get moved back soon: * Image, CompressedImage and their dimensional typedefs * ImageView, CompressedImageView and their dimensional typedefs * PixelStorage Not PixelFormat etc., that one will stay in the GL namespace and a completely new PixelFormat enum will be provided in the root namespace.
8 years ago
@section GL-AbstractTexture-webgl-restrictions WebGL restrictions
doc: document all WebGL restrictions.
12 years ago
@ref MAGNUM_TARGET_WEBGL "WebGL" puts some restrictions on type of data
submitted to @ref Texture::setSubImage() "*Texture::setSubImage()", see its
documentation for details.
Split the OpenGL layer out, pt 4: new Magnum::GL namespace. At the moment just the GL library itself w/o the tests, and without backwards compatibility aliases. The following types were left in the root namespace, despite being in the GL/ directory, as they will get moved back soon: * Image, CompressedImage and their dimensional typedefs * ImageView, CompressedImageView and their dimensional typedefs * PixelStorage Not PixelFormat etc., that one will stay in the GL namespace and a completely new PixelFormat enum will be provided in the root namespace.
8 years ago
@section GL-AbstractTexture-performance-optimization Performance optimizations and security
Documentation and example usage of textures. Mentions some not-yet-done features such as Vector2i or proper texture formats.
14 years ago
Implemented ARB_shader_image_load_store.
10 years ago
The engine tracks currently bound textures and images in all available texture
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
units to avoid unnecessary calls to @fn_gl_keyword{ActiveTexture},
@fn_gl_keyword{BindTexture} and @fn_gl_keyword{BindImageTexture}. Texture
configuration functions use dedicated highest available texture unit to not
affect active bindings in user units. Texture limits and implementation-defined
values (such as @ref maxColorSamples()) are cached, so repeated queries don't
result in repeated @fn_gl{Get} calls. See also @ref Context::resetState() and
@ref Context::State::Textures.
Tracking texture state - bindings and limits.
14 years ago
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
If @gl_extension{ARB,direct_state_access} (part of OpenGL 4.5) is available,
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
@ref bind(Int) and @ref unbind(Int) use @fn_gl{BindTextureUnit}. Otherwise, if
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
@gl_extension{ARB,multi_bind} (part of OpenGL 4.4) is available, @ref bind(Int)
GL: drop support for EXT_direct_state_access.
7 years ago
and @ref unbind() uses @fn_gl{BindTextures}.
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
GL: drop support for EXT_direct_state_access.
7 years ago
In addition, if @gl_extension{ARB,direct_state_access} (part of OpenGL 4.5) is
available, also all texture configuration and data updating functions use DSA
functions to avoid unnecessary calls to @fn_gl{ActiveTexture} and
@fn_gl{BindTexture}. See respective function documentation for more
information.
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
If @gl_extension{ARB,multi_bind} (part of OpenGL 4.5) is available,
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
@ref bind(Int, std::initializer_list<AbstractTexture*>) and @ref unbind(Int, std::size_t)
use @fn_gl{BindTextures} to avoid unnecessary calls to @fn_gl{ActiveTexture}.
Otherwise the feature is emulated with sequence of @ref bind(Int)/@ref unbind(Int)
calls.
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
If either @gl_extension{ARB,direct_state_access} (part of OpenGL 4.5) or
@gl_extension{ARB,robustness} desktop extension is available, image reading
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
operations (such as @ref Texture::image()) are protected from buffer overflow.
Ability to retrieve texture images.
13 years ago
Split the OpenGL layer out, pt 8: move out GL-specific PixelStorage parts. The general part stays in the root namespace, while the GL-specific stuff goes to the GL namespace. Also all GL-specific documentation was moved to relevant APIs in the GL namespace. This finally allows me to build PixelStorage.cpp as part of the root namespace. The PixelStorage::pixelSize() function and PixelStorage::dataProperties() taking a pair of GL::PixelFormat / GL::PixelType enums is deprecated, use GL::pixelSize() and dataProperties() taking pixel size directly instead. A lot of code is still using these, including images; the deprecated aliases are inlined in the header to avoid a compile-time dependency on the GL library.
8 years ago
Pixel storage mode defined by @ref PixelStorage and @ref CompressedPixelStorage
is applied either right before doing image upload (such as various
@ref Texture::setImage() "setImage()", @ref Texture::setSubImage() "setSubImage()",
@ref Texture::setCompressedImage() "setCompressedImage()" or
@ref Texture::setCompressedSubImage() "setCompressedSubImage()" functions)
using @fn_gl_keyword{PixelStore} with @def_gl{UNPACK_*} parameters or right
before doing image download (such as various @ref Texture::image() "image()",
@ref Texture::subImage() "subImage()", @ref Texture::compressedImage() "compressedImage()",
or @ref Texture::compressedSubImage() "compressedSubImage()" functions) using
@fn_gl{PixelStore} with @def_gl{PACK_*}. The engine tracks currently used pixel
pack/unpack parameters to avoid unnecessary calls to @fn_gl{PixelStore}. See
also @ref Context::resetState() and @ref Context::State::PixelStorage.
Tracking texture state - bindings and limits.
14 years ago
To achieve least state changes, fully configure each texture in one run --
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
method chaining comes in handy --- and try to have often used textures in
Follow OpenGL terminology for texture binding units. Until now the textures were bound to layers, which was rather confusing, especially when binding layered textures to layers (gaah). Also the wording might have implied that each texture must be in some layer in order to make it usable in shader. This is no longer the case with (yet unimplemented) bindless texture, so another reason to remove the confusion. All occurences of texture layers were replaced texture binding units to follow OpenGL naming. It was mostly in the docs, except for already-deprecated *Layer enums in shaders, but they will be removed soon anyway.
12 years ago
dedicated units, not occupied by other textures. First configure the texture
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
and *then* set the data, so OpenGL can optimize them to match the settings. To
avoid redundant consistency checks and memory reallocations when updating
texture data, set texture storage at once using @ref Texture::setStorage() "setStorage()"
and then set data using @ref Texture::setSubImage() "setSubImage()".
Tracking texture state - bindings and limits.
14 years ago
Fallback for texture storage feature. It's now possible to call *Texture::setStorage() even if OpenGL 4.2, ARB_texture_storage, OpenGL ES 3.0 or EXT_texture_storage in ES 2.0 is not available, the function will internally use sequence of setImage() calls as fallback. It seems to me that this behavior is better than forcing the user to always check for extension presence and then implementing this functionality again and again. The huge switches for setting proper image format and type are ugly though, but according to specs they must be set even if we are sending no data (which is weird).
13 years ago
Function @ref Texture::setStorage() "setStorage()" creates immutable texture
storage, removing the need for additional consistency checks and memory
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
reallocations when updating the data later. If OpenGL 4.2, @gl_extension{ARB,texture_storage},
OpenGL ES 3.0 or @gl_extension{EXT,texture_storage} in OpenGL ES 2.0 is not
Fallback for texture storage feature. It's now possible to call *Texture::setStorage() even if OpenGL 4.2, ARB_texture_storage, OpenGL ES 3.0 or EXT_texture_storage in ES 2.0 is not available, the function will internally use sequence of setImage() calls as fallback. It seems to me that this behavior is better than forcing the user to always check for extension presence and then implementing this functionality again and again. The huge switches for setting proper image format and type are ugly though, but according to specs they must be set even if we are sending no data (which is weird).
13 years ago
available, the feature is emulated with sequence of @ref Texture::setImage() "setImage()"
calls.
Doc++
12 years ago
You can use functions @ref Texture::invalidateImage() "invalidateImage()" and
Shotgun documentation updates. Fixed old references to Buffer::Usage, fixed parameter order in Image constructor, fixed some function references. Added explicit references to linkable stuff.
13 years ago
@ref Texture::invalidateSubImage() "invalidateSubImage()" if you don't need
texture data anymore to avoid unnecessary memory operations performed by OpenGL
in order to preserve the data. If running on OpenGL ES or extension
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
@gl_extension{ARB,invalidate_subdata} (part of OpenGL 4.3) is not available, these
doc: mention also OpenGL version for optionally used extensions.
12 years ago
functions do nothing.
Extension-aware Buffer and *Texture data invalidation. If ARB_invalidate_subdata is not available, these functions do nothing, instead of crashing on null pointer dereference. It results in more convenient usage, enabling users to call them whenever they want, improving performance on implementations which supports that.
14 years ago
Added TODOs.
13 years ago
@todo all texture [level] parameters, global texture parameters
@todo++
14 years ago
@todo Texture copying
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
@todo `GL_MAX_SAMPLE_MASK_WORDS` when @gl_extension{ARB,texture_multisample} is done
@todo Query for immutable levels (@gl_extension{ARB,ES3_compatibility})
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
*/
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
class MAGNUM_GL_EXPORT AbstractTexture: public AbstractObject {
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
public:
Implemented texture LOD parameters.
12 years ago
#ifndef MAGNUM_TARGET_GLES2
/**
* @brief Max level-of-detail bias
*
* The result is cached, repeated queries don't result in repeated
* OpenGL calls.
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @see @fn_gl{Get} with @def_gl_keyword{MAX_TEXTURE_LOD_BIAS}
Removed autolink-preventing % noise and unneeded HTML from root namespace.
12 years ago
* @requires_gles30 Texture LOD bias doesn't have
Implemented texture LOD parameters.
12 years ago
* implementation-defined range in OpenGL ES 2.0.
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
* @requires_webgl20 Texture LOD bias doesn't have
* implementation-defined range in WebGL 1.0.
Implemented texture LOD parameters.
12 years ago
*/
static Float maxLodBias();
#endif
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
Texture-related limit queries. Renamed AbstractTexture::maxSupportedLayerCount() to maxLayers(), which is in fact alias to Shader::maxCombinedTextureImageUnits(). Also renamed Samples::maxSupportedAnisotropy() to maxAnisotropy(). It now has slightly confusing naming, will fix that later. Both original functions are now alias to new ones to retain source compatibility, will be removed in future releases. Also printing the values in magnum-info.
13 years ago
/**
* @brief Max supported color sample count
*
* The result is cached, repeated queries don't result in repeated
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* OpenGL calls. If neither extension @gl_extension{ARB,texture_multisample}
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* (part of OpenGL 3.2) nor OpenGL ES 3.1 is available, returns
* @cpp 0 @ce.
* @see @fn_gl{Get} with @def_gl_keyword{MAX_COLOR_TEXTURE_SAMPLES}
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
* @requires_gles30 Not defined in OpenGL ES 2.0.
* @requires_gles Multisample textures are not available in WebGL.
Texture-related limit queries. Renamed AbstractTexture::maxSupportedLayerCount() to maxLayers(), which is in fact alias to Shader::maxCombinedTextureImageUnits(). Also renamed Samples::maxSupportedAnisotropy() to maxAnisotropy(). It now has slightly confusing naming, will fix that later. Both original functions are now alias to new ones to retain source compatibility, will be removed in future releases. Also printing the values in magnum-info.
13 years ago
*/
static Int maxColorSamples();
/**
* @brief Max supported depth sample count
*
* The result is cached, repeated queries don't result in repeated
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* OpenGL calls. If neither extension @gl_extension{ARB,texture_multisample}
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* (part of OpenGL 3.2) nor OpenGL ES 3.1 is available, returns
* @cpp 0 @ce.
* @see @fn_gl{Get} with @def_gl_keyword{MAX_DEPTH_TEXTURE_SAMPLES}
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
* @requires_gles30 Not defined in OpenGL ES 2.0.
* @requires_gles Multisample textures are not available in WebGL.
Texture-related limit queries. Renamed AbstractTexture::maxSupportedLayerCount() to maxLayers(), which is in fact alias to Shader::maxCombinedTextureImageUnits(). Also renamed Samples::maxSupportedAnisotropy() to maxAnisotropy(). It now has slightly confusing naming, will fix that later. Both original functions are now alias to new ones to retain source compatibility, will be removed in future releases. Also printing the values in magnum-info.
13 years ago
*/
static Int maxDepthSamples();
/**
* @brief Max supported integer sample count
*
* The result is cached, repeated queries don't result in repeated
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* OpenGL calls. If neither extension @gl_extension{ARB,texture_multisample}
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* (part of OpenGL 3.2) nor OpenGL ES 3.1 is available, returns
* @cpp 0 @ce.
* @see @fn_gl{Get} with @def_gl_keyword{MAX_INTEGER_SAMPLES}
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
* @requires_gles30 Not defined in OpenGL ES 2.0.
* @requires_gles Multisample textures are not available in WebGL.
Texture-related limit queries. Renamed AbstractTexture::maxSupportedLayerCount() to maxLayers(), which is in fact alias to Shader::maxCombinedTextureImageUnits(). Also renamed Samples::maxSupportedAnisotropy() to maxAnisotropy(). It now has slightly confusing naming, will fix that later. Both original functions are now alias to new ones to retain source compatibility, will be removed in future releases. Also printing the values in magnum-info.
13 years ago
*/
static Int maxIntegerSamples();
#endif
Reworked binding of textures to shaders. According to OpenGL specification the sampler uniform takes only texture layer and not any particular texture. In OpenGL 4.2 it is possible to explicitly specify desired texture layer in the uniform, so I suppose it's desired to have shader with fixed texture layers and on the other hand be able to bind texture to any layer, not only one -- the exact opposite of how it was in Magnum before. The shader should now specify in its public API which texture is on which layer and explicitly set the layer uniforms upon construction. All setUniform() functions taking textures as argument were thus removed, as they confuse things a lot. The textures now have bind() function which takes layer as argument, so it's now sufficient to only bind the texture before drawing the mesh without messing with sampler uniform.
14 years ago
Added *Texture::unbind(). Will be used in fallback implementation of multi-bind.
12 years ago
/**
* @brief Unbind any texture from given texture unit
*
GL: drop support for EXT_direct_state_access.
7 years ago
* If neither @gl_extension{ARB,direct_state_access} (part of OpenGL 4.5)
* nor @gl_extension{ARB,multi_bind} (part of OpenGL 4.4) is available,
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
* the texture unit is made active before unbinding the texture.
Added *Texture::unbind(). Will be used in fallback implementation of multi-bind.
12 years ago
* @note This function is meant to be used only internally from
* @ref AbstractShaderProgram subclasses. See its documentation
* for more information.
* @see @ref bind(), @ref Shader::maxCombinedTextureImageUnits(),
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @fn_gl_keyword{BindTextureUnit}, @fn_gl_keyword{BindTextures},
* eventually @fn_gl_keyword{ActiveTexture} and
* @fn_gl_keyword{BindTexture}
Added *Texture::unbind(). Will be used in fallback implementation of multi-bind.
12 years ago
*/
static void unbind(Int textureUnit);
Added multi-unbind function to AbstractTexture. More efficient (no allocation) and also more convenient than calling e.g. unbind(7, {nullptr, nullptr, nullptr}).
12 years ago
/**
* @brief Unbind textures in given range of texture units
*
Implemented ARB_shader_image_load_store.
10 years ago
* Unbinds all textures in the range @f$ [ firstTextureUnit ; firstTextureUnit + count ) @f$.
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* If @gl_extension{ARB,multi_bind} (part of OpenGL 4.4) is not available,
Added multi-unbind function to AbstractTexture. More efficient (no allocation) and also more convenient than calling e.g. unbind(7, {nullptr, nullptr, nullptr}).
12 years ago
* the feature is emulated with sequence of @ref unbind(Int) calls.
* @note This function is meant to be used only internally from
* @ref AbstractShaderProgram subclasses. See its documentation
* for more information.
Implemented ARB_shader_image_load_store.
10 years ago
* @see @ref bind(), @ref Shader::maxCombinedTextureImageUnits(),
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @fn_gl_keyword{BindTextures}
Added multi-unbind function to AbstractTexture. More efficient (no allocation) and also more convenient than calling e.g. unbind(7, {nullptr, nullptr, nullptr}).
12 years ago
*/
static void unbind(Int firstTextureUnit, std::size_t count);
Support ARB_multi_bind for textures.
12 years ago
/**
* @brief Bind textures to given range of texture units
doc: change all "latest" additions/deprecations to 2020.06.
6 years ago
* @m_since{2020,06}
Support ARB_multi_bind for textures.
12 years ago
*
* Binds first texture in the list to @p firstTextureUnit, second to
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* `firstTextureUnit + 1` etc. If any texture is @cpp nullptr @ce,
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* given texture unit is unbound. If @gl_extension{ARB,multi_bind} (part
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* of OpenGL 4.4) is not available, the feature is emulated with
* sequence of @ref bind(Int) / @ref unbind(Int) calls.
Doc++, minor cleanup.
12 years ago
* @note This function is meant to be used only internally from
* @ref AbstractShaderProgram subclasses. See its documentation
* for more information.
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @see @ref Shader::maxCombinedTextureImageUnits(),
* @fn_gl_keyword{BindTextures}
Support ARB_multi_bind for textures.
12 years ago
*/
GL: Add ArrayView<AbstractTexture*> overrides to bind and bindImages Signed-off-by: Squareys <squareys@googlemail.com>
7 years ago
static void bind(Int firstTextureUnit, Containers::ArrayView<AbstractTexture* const> textures);
/** @overload */
GL: remove unneded ArrayView.h include from headers. It was needed in the times before StringView, where the setLabel() APIs had overloads taking a const char(&)[size] to avoid allocating a std::string. But that got all removed and cleaned up in bc884428f8c1d8f66ee0216c9a6a338eff31fca4 (two years ago!), yet somehow I forgot to remove these includes as well. In some cases the ArrayView was still used inside the header for delegating from std::initializer_list overloads. Those are now deinlined to not need the include anymore. The only remaining place that *needs* an ArrayView include is Buffer.h, where it accepts an initializer_list<T>. But there it's fine I think, the class is dealing with memory as well.
3 years ago
static void bind(Int firstTextureUnit, std::initializer_list<AbstractTexture*> textures);
Support ARB_multi_bind for textures.
12 years ago
Implemented ARB_shader_image_load_store.
10 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
/**
* @brief Unbind any image from given image unit
*
* @note This function is meant to be used only internally from
* @ref AbstractShaderProgram subclasses. See its documentation
* for more information.
* @see @ref Texture::bindImage() "*Texture::bindImage()",
* @ref Texture::bindImageLayered() "*Texture::bindImageLayered()",
* @ref unbindImages(), @ref bindImages(),
* @ref AbstractShaderProgram::maxImageUnits(),
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @fn_gl_keyword{BindImageTexture}
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* @requires_gl42 Extension @gl_extension{ARB,shader_image_load_store}
Implemented ARB_shader_image_load_store.
10 years ago
* @requires_gles31 Shader image load/store is not available in OpenGL
* ES 3.0 and older.
* @requires_gles Shader image load/store is not available in WebGL.
*/
static void unbindImage(Int imageUnit);
#endif
#ifndef MAGNUM_TARGET_GLES
/**
* @brief Unbind images in given range of image units
*
* Unbinds all texture in the range @f$ [ firstImageUnit ; firstImageUnit + count ) @f$.
* @note This function is meant to be used only internally from
* @ref AbstractShaderProgram subclasses. See its documentation
* for more information.
* @see @ref Texture::bindImage() "*Texture::bindImage()",
* @ref Texture::bindImageLayered() "*Texture::bindImageLayered()",
* @ref unbindImage(), @ref bindImages(),
* @ref AbstractShaderProgram::maxImageUnits(),
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @fn_gl_keyword{BindImageTextures}
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* @requires_gl42 Extension @gl_extension{ARB,shader_image_load_store}
* @requires_gl44 Extension @gl_extension{ARB,multi_bind}
Implemented ARB_shader_image_load_store.
10 years ago
* @requires_gl Multi bind is not available in OpenGL ES and WebGL.
*/
GL: remove unneded ArrayView.h include from headers. It was needed in the times before StringView, where the setLabel() APIs had overloads taking a const char(&)[size] to avoid allocating a std::string. But that got all removed and cleaned up in bc884428f8c1d8f66ee0216c9a6a338eff31fca4 (two years ago!), yet somehow I forgot to remove these includes as well. In some cases the ArrayView was still used inside the header for delegating from std::initializer_list overloads. Those are now deinlined to not need the include anymore. The only remaining place that *needs* an ArrayView include is Buffer.h, where it accepts an initializer_list<T>. But there it's fine I think, the class is dealing with memory as well.
3 years ago
static void unbindImages(Int firstImageUnit, std::size_t count);
Implemented ARB_shader_image_load_store.
10 years ago
/**
* @brief Bind textures to given range of texture units
doc: change all "latest" additions/deprecations to 2020.06.
6 years ago
* @m_since{2020,06}
Implemented ARB_shader_image_load_store.
10 years ago
*
* Binds first level of given texture in the list to @p firstImageUnit,
* second to `firstTextureUnit + 1` etc. 3D, cube map and array
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* textures are bound as layered targets. If any texture is
* @cpp nullptr @ce, given image unit is unbound.
Implemented ARB_shader_image_load_store.
10 years ago
* @note This function is meant to be used only internally from
* @ref AbstractShaderProgram subclasses. See its documentation
* for more information.
* @see @ref Texture::bindImage() "*Texture::bindImage()",
* @ref Texture::bindImageLayered() "*Texture::bindImageLayered()",
* @ref unbindImages(), @ref unbindImage(),
* @ref AbstractShaderProgram::maxImageUnits(),
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @fn_gl_keyword{BindImageTextures}
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* @requires_gl42 Extension @gl_extension{ARB,shader_image_load_store}
* @requires_gl44 Extension @gl_extension{ARB,multi_bind}
Implemented ARB_shader_image_load_store.
10 years ago
* @requires_gl Multi bind is not available in OpenGL ES and WebGL.
*/
GL: Add ArrayView<AbstractTexture*> overrides to bind and bindImages Signed-off-by: Squareys <squareys@googlemail.com>
7 years ago
static void bindImages(Int firstImageUnit, Containers::ArrayView<AbstractTexture* const> textures);
/** @overload */
GL: remove unneded ArrayView.h include from headers. It was needed in the times before StringView, where the setLabel() APIs had overloads taking a const char(&)[size] to avoid allocating a std::string. But that got all removed and cleaned up in bc884428f8c1d8f66ee0216c9a6a338eff31fca4 (two years ago!), yet somehow I forgot to remove these includes as well. In some cases the ArrayView was still used inside the header for delegating from std::initializer_list overloads. Those are now deinlined to not need the include anymore. The only remaining place that *needs* an ArrayView include is Buffer.h, where it accepts an initializer_list<T>. But there it's fine I think, the class is dealing with memory as well.
3 years ago
static void bindImages(Int firstImageUnit, std::initializer_list<AbstractTexture*> textures);
Implemented ARB_shader_image_load_store.
10 years ago
#endif
Minor cleanup.
12 years ago
/** @brief Copying is not allowed */
AbstractTexture(const AbstractTexture&) = delete;
/** @brief Move constructor */
AbstractTexture(AbstractTexture&& other) noexcept;
Document disallowed copy and allowed move better. Moreover when the deleted function was private, the compilation failed with (rather cryptic) error that the function is private instead of noticing that the function is deleted.
13 years ago
/** @brief Copying is not allowed */
AbstractTexture& operator=(const AbstractTexture&) = delete;
Move constructor and move assignment for textures.
14 years ago
/** @brief Move assignment */
Simplified and tested Texture move constructor/assignment. Move constructor and move assignment is now noexcept and it just swaps the data of the two instances instead of calling GL API, thus it can be inline. Also removed unneded limitations in BufferTexture.
13 years ago
AbstractTexture& operator=(AbstractTexture&& other) noexcept;
Move constructor and move assignment for textures.
14 years ago
Minor code reorganization.
11 years ago
/** @brief OpenGL texture ID */
GLuint id() const { return _id; }
GL: add *Texture::target(). Useful when interacting with raw GL code.
6 years ago
/**
* @brief OpenGL texture target
* @m_since_latest
*/
GLenum target() const { return _target; }
Ability to wrap and release existing OpenGL objects. Allows for better interaction with third-party libraries. I should probably write a new documentation page about OpenGL wrapping.
11 years ago
/**
* @brief Release OpenGL object
*
* Releases ownership of OpenGL texture object and returns its ID so it
* is not deleted on destruction. The internal state is then equivalent
* to moved-from state.
* @see @ref BufferTexture::wrap(), @ref CubeMapTexture::wrap(),
* @ref CubeMapTextureArray::wrap(),
* @ref MultisampleTexture::wrap(), @ref RectangleTexture::wrap(),
* @ref Texture::wrap(), @ref TextureArray::wrap()
*/
GLuint release();
First-class WebGL support, part 2: no debug output.
11 years ago
#ifndef MAGNUM_TARGET_WEBGL
Support for object labels from KHR_debug and EXT_debug_label. Can't test EXT_debug_label, as that is apparently OSX 10.9-only. Added GL tests for all implemented objects. KHR_debug is selected first, if that is not available, fall back to EXT_debug_label. If neither is available, the functions are no-op. I hope EXT_debug_label gets replaced by KHR_debug later, thus it is now only "emulated" through KHR_debug enums.
13 years ago
/**
Removed autolink-preventing % noise and unneeded HTML from root namespace.
12 years ago
* @brief Texture label
Support for object labels from KHR_debug and EXT_debug_label. Can't test EXT_debug_label, as that is apparently OSX 10.9-only. Added GL tests for all implemented objects. KHR_debug is selected first, if that is not available, fall back to EXT_debug_label. If neither is available, the functions are no-op. I hope EXT_debug_label gets replaced by KHR_debug later, thus it is now only "emulated" through KHR_debug enums.
13 years ago
*
* The result is *not* cached, repeated queries will result in repeated
Updates for debug output in ES 3.2.
8 years ago
* OpenGL calls. If OpenGL 4.3 / OpenGL ES 3.2 is not supported and
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* neither @gl_extension{KHR,debug} (covered also by
* @gl_extension{ANDROID,extension_pack_es31a}) nor @gl_extension{EXT,debug_label}
Updates for debug output in ES 3.2.
8 years ago
* desktop or ES extension is available, this function returns empty
* string.
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @see @fn_gl_keyword{GetObjectLabel} or
* @fn_gl_extension_keyword{GetObjectLabel,EXT,debug_label} with
Support for object labels from KHR_debug and EXT_debug_label. Can't test EXT_debug_label, as that is apparently OSX 10.9-only. Added GL tests for all implemented objects. KHR_debug is selected first, if that is not available, fall back to EXT_debug_label. If neither is available, the functions are no-op. I hope EXT_debug_label gets replaced by KHR_debug later, thus it is now only "emulated" through KHR_debug enums.
13 years ago
* @def_gl{TEXTURE}
First-class WebGL support, part 2: no debug output.
11 years ago
* @requires_gles Debug output is not available in WebGL.
Support for object labels from KHR_debug and EXT_debug_label. Can't test EXT_debug_label, as that is apparently OSX 10.9-only. Added GL tests for all implemented objects. KHR_debug is selected first, if that is not available, fall back to EXT_debug_label. If neither is available, the functions are no-op. I hope EXT_debug_label gets replaced by KHR_debug later, thus it is now only "emulated" through KHR_debug enums.
13 years ago
*/
GL: port label() / setLabel() away from std::string. All the tests were updated to explicitly check that non-null-terminated strings get handled properly (the GL label APIs have an explicit size, so it *should*, but just in case). Also, because various subclasses override the setter to return correct type for method chaining and the override has to be deinlined to avoid relying on a StringView include, the tests are now explicitly done for each leaf class, instead of the subclass The <string> being removed from the base class for all GL objects may affect downstream projects which relied on it being included. In case of Magnum, the breakages were already fixed in the previous commit. Compile time improvement for the MagnumGL library alone is 0.2 second or 4% (6.1 seconds before, 5.9 after). Not bad, given that there's three more files to compile and strings are still heavily used in other GL debug output APIs and all shader stuff. For a build of just the GL library and all tests, it goes down from 28.9 seconds to 28.1. Most tests also still rely quite heavily on std::stringstream for debug output testing, so the numbers still could go further.
5 years ago
Containers::String label();
Support for object labels from KHR_debug and EXT_debug_label. Can't test EXT_debug_label, as that is apparently OSX 10.9-only. Added GL tests for all implemented objects. KHR_debug is selected first, if that is not available, fall back to EXT_debug_label. If neither is available, the functions are no-op. I hope EXT_debug_label gets replaced by KHR_debug later, thus it is now only "emulated" through KHR_debug enums.
13 years ago
/**
* @brief Set texture label
* @return Reference to self (for method chaining)
*
Updates for debug output in ES 3.2.
8 years ago
* Default is empty string. If OpenGL 4.3 / OpenGL ES 3.2 is not
Doxygen: renamed \extension[2] to \gl_extension[2]. So it's clear that this is not a Vulkan extension link.
8 years ago
* supported and neither @gl_extension{KHR,debug} (covered also by
* @gl_extension{ANDROID,extension_pack_es31a}) nor @gl_extension{EXT,debug_label}
Updates for debug output in ES 3.2.
8 years ago
* desktop or ES extension is available, this function does nothing.
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @see @ref maxLabelLength(), @fn_gl_keyword{ObjectLabel} or
* @fn_gl_extension_keyword{LabelObject,EXT,debug_label} with
Support for object labels from KHR_debug and EXT_debug_label. Can't test EXT_debug_label, as that is apparently OSX 10.9-only. Added GL tests for all implemented objects. KHR_debug is selected first, if that is not available, fall back to EXT_debug_label. If neither is available, the functions are no-op. I hope EXT_debug_label gets replaced by KHR_debug later, thus it is now only "emulated" through KHR_debug enums.
13 years ago
* @def_gl{TEXTURE}
First-class WebGL support, part 2: no debug output.
11 years ago
* @requires_gles Debug output is not available in WebGL.
Support for object labels from KHR_debug and EXT_debug_label. Can't test EXT_debug_label, as that is apparently OSX 10.9-only. Added GL tests for all implemented objects. KHR_debug is selected first, if that is not available, fall back to EXT_debug_label. If neither is available, the functions are no-op. I hope EXT_debug_label gets replaced by KHR_debug later, thus it is now only "emulated" through KHR_debug enums.
13 years ago
*/
GL: port label() / setLabel() away from std::string. All the tests were updated to explicitly check that non-null-terminated strings get handled properly (the GL label APIs have an explicit size, so it *should*, but just in case). Also, because various subclasses override the setter to return correct type for method chaining and the override has to be deinlined to avoid relying on a StringView include, the tests are now explicitly done for each leaf class, instead of the subclass The <string> being removed from the base class for all GL objects may affect downstream projects which relied on it being included. In case of Magnum, the breakages were already fixed in the previous commit. Compile time improvement for the MagnumGL library alone is 0.2 second or 4% (6.1 seconds before, 5.9 after). Not bad, given that there's three more files to compile and strings are still heavily used in other GL debug output APIs and all shader stuff. For a build of just the GL library and all tests, it goes down from 28.9 seconds to 28.1. Most tests also still rely quite heavily on std::stringstream for debug output testing, so the numbers still could go further.
5 years ago
AbstractTexture& setLabel(Containers::StringView label);
First-class WebGL support, part 2: no debug output.
11 years ago
#endif
Support for object labels from KHR_debug and EXT_debug_label. Can't test EXT_debug_label, as that is apparently OSX 10.9-only. Added GL tests for all implemented objects. KHR_debug is selected first, if that is not available, fall back to EXT_debug_label. If neither is available, the functions are no-op. I hope EXT_debug_label gets replaced by KHR_debug later, thus it is now only "emulated" through KHR_debug enums.
13 years ago
Doc++
14 years ago
/**
Follow OpenGL terminology for texture binding units. Until now the textures were bound to layers, which was rather confusing, especially when binding layered textures to layers (gaah). Also the wording might have implied that each texture must be in some layer in order to make it usable in shader. This is no longer the case with (yet unimplemented) bindless texture, so another reason to remove the confusion. All occurences of texture layers were replaced texture binding units to follow OpenGL naming. It was mostly in the docs, except for already-deprecated *Layer enums in shaders, but they will be removed soon anyway.
12 years ago
* @brief Bind texture to given texture unit
Doc++
14 years ago
*
GL: drop support for EXT_direct_state_access.
7 years ago
* If neither @gl_extension{ARB,direct_state_access} (part of OpenGL
* 4.5) nor @gl_extension{ARB,multi_bind} (part of OpenGL 4.4) is
* available, the texture unit is made active before binding the
* texture.
Use functions to bind textures to shader rather than *Texture::bind(). Why did I do this: * It is more clean, shorter and nice looking with method chaining, i.e. instead of: shader.setColor(...) .setOtherParam(5); texture1.bind(MyShader::Texture1Layer); texture2.bind(MyShader::Texture2Layer); We now have this: shader.setColor(...) .setOtherParam(5) .setTexture1(texture1) .setTexture2(texture2); * It is now also clear which texture type is expected, the layer constant did not say anything about type. * Also it is possible to use new features (multi bind, bindless textures etc.) while preserving the same public API. The only potential disadvantage is that the textures don't stay bound like uniform values do, but this become a non-issue with bindless textures. As usual, the old way is now deprecated and will be removed in some future release.
12 years ago
* @note This function is meant to be used only internally from
* @ref AbstractShaderProgram subclasses. See its documentation
* for more information.
Support ARB_multi_bind for textures.
12 years ago
* @see @ref bind(Int, std::initializer_list<AbstractTexture*>),
* @ref unbind(), @ref Shader::maxCombinedTextureImageUnits(),
Documentation updates for the new theme in the root namespace. Also adding search keywords for hopefully all OpenGL APIs.
8 years ago
* @fn_gl_keyword{BindTextureUnit}, @fn_gl_keyword{BindTextures},
* eventually @fn_gl_keyword{ActiveTexture} and
* @fn_gl_keyword{BindTexture}
Doc++
14 years ago
*/
Follow OpenGL terminology for texture binding units. Until now the textures were bound to layers, which was rather confusing, especially when binding layered textures to layers (gaah). Also the wording might have implied that each texture must be in some layer in order to make it usable in shader. This is no longer the case with (yet unimplemented) bindless texture, so another reason to remove the confusion. All occurences of texture layers were replaced texture binding units to follow OpenGL naming. It was mostly in the docs, except for already-deprecated *Layer enums in shaders, but they will be removed soon anyway.
12 years ago
void bind(Int textureUnit);
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
GL: make AbstractTexture destructor protected. Right now there isn't any work done in subclass destructors, but that may change in the future due to driver workarounds being added. So better be prepared.
7 years ago
#if !defined(MAGNUM_BUILD_DEPRECATED) || defined(DOXYGEN_GENERATING_OUTPUT)
protected: /* Destructor was public before */
#endif
/**
* @brief Destructor
*
* Deletes associated OpenGL texture.
* @see @ref BufferTexture::wrap(), @ref CubeMapTexture::wrap(),
* @ref CubeMapTextureArray::wrap(),
* @ref MultisampleTexture::wrap(), @ref RectangleTexture::wrap(),
* @ref Texture::wrap(), @ref TextureArray::wrap(),
* @ref release(), @fn_gl_keyword{DeleteTextures}
*/
/* Not virtual to avoid vtable overhead, however while subclasses are
all the same size (and thus no risk of memory leaks from additional
members) these might perform cleanup steps on destruction (driver
workarouds...), so we have to disallow deletion from base pointer */
~AbstractTexture();
Doxygen: ignore specific namespaces instead of whole directories. Less maintenance burden with `#ifndef DOXYGEN_GENERATING_OUTPUT`, removed the ones which are not needed anymore.
13 years ago
#ifdef DOXYGEN_GENERATING_OUTPUT
private:
#else
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
protected:
Doxygen: ignore specific namespaces instead of whole directories. Less maintenance burden with `#ifndef DOXYGEN_GENERATING_OUTPUT`, removed the ones which are not needed anymore.
13 years ago
#endif
Using new type aliases in whole root Magnum namespace.
13 years ago
template<UnsignedInt textureDimensions> struct DataHelper {};
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
Compressed image support, part 14: queries for compressed block sizes.
11 years ago
#ifndef MAGNUM_TARGET_GLES
static Int compressedBlockDataSize(GLenum target, TextureFormat format);
#endif
Split Texture into TextureArray, MultisampleTexture and RectangleTexture. Each texture has slightly different usage requirements and having everything under one generic class is not worth the additional runtime checks and whatnot. The current way with Texture::Target enum (hopefully not too widely used) is now deprecated and will be removed in some future release. However general Texture1D/2D/3D usage is not changed in any way.
12 years ago
explicit AbstractTexture(GLenum target);
Ability to instantiate OpenGL objects without creating them. Similarly to what's now done with NoInit tags for Containers::Array and all math types such as Vector, there's now NoCreate tag for creating wrappers without actually creating the underlying OpenGL object. The instance is then equivalent to moved-from state. Useful to avoid needless creation/deletion of OpenGL object in case you would overwrite the instance later anyway: Mesh mesh{NoCreate}; std::unique_ptr<Buffer> indices, vertices; std::tie(mesh, indices, vertices) = MeshTools:compile(...);
11 years ago
explicit AbstractTexture(NoCreateT, GLenum target) noexcept: _target{target}, _id{0}, _flags{ObjectFlag::DeleteOnDestruction} {}
Ability to wrap and release existing OpenGL objects. Allows for better interaction with third-party libraries. I should probably write a new documentation page about OpenGL wrapping.
11 years ago
explicit AbstractTexture(GLuint id, GLenum target, ObjectFlags flags) noexcept: _target{target}, _id{id}, _flags{flags} {}
Split Texture into TextureArray, MultisampleTexture and RectangleTexture. Each texture has slightly different usage requirements and having everything under one generic class is not worth the additional runtime checks and whatnot. The current way with Texture::Target enum (hopefully not too widely used) is now deprecated and will be removed in some future release. However general Texture1D/2D/3D usage is not changed in any way.
12 years ago
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
void MAGNUM_GL_LOCAL createIfNotAlready();
Properly create cube map texture before calling ARB_DSA functions. It might happen that the user is calling ARB_DSA-only functions like CubeMapTexture::subImage() the texture was created using glGenTextures() and not using ARB_DSA, for example because the extension was disabled and then not bound or used at all, which makes the texture "not created yet". This is not needed for internal (...ImplementationDSA()) functions because these are always called only if the texture was also created using glCreateTextures(). Basically doing the same that's in AbstractTexture itself but for some reason was omitted here. The internal ...ImplementationDSA() functions might break in case the object is created externally using glGen*(), not bound or used at all and then the class is created using ::wrap(), but that's highly unprobable (why would anyone do that?).
11 years ago
Implemented ARB_shader_image_load_store.
10 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
void bindImageInternal(Int imageUnit, Int level, bool layered, Int layer, ImageAccess access, ImageFormat format);
#endif
Follow OpenGL terminology for texture binding units. Until now the textures were bound to layers, which was rather confusing, especially when binding layered textures to layers (gaah). Also the wording might have implied that each texture must be in some layer in order to make it usable in shader. This is no longer the case with (yet unimplemented) bindless texture, so another reason to remove the confusion. All occurences of texture layers were replaced texture binding units to follow OpenGL naming. It was mostly in the docs, except for already-deprecated *Layer enums in shaders, but they will be removed soon anyway.
12 years ago
/* Unlike bind() this also sets the texture binding unit as active */
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
void MAGNUM_GL_LOCAL bindInternal();
Fixed CubeMapTexture wrapping, simplified code of texture classes. Texture now doesn't have it's own redundant copy of target, but uses the one from AbstractTexture. Subsequently, all DataHelper functions now accept "untyped" GLenum as target for better flexibility. CubeMapTexture is not based on Texture2D anymore, all references to cube map textures were removed from Target enums, so it's now not possible to create cube map texture other way than using CubeMapTexture class, which is how it should be done in first place. The wrapping mode fix is thus now trivial.
14 years ago
Ability to set base and max mip level on textures. Curiously this functionality is not available on ES2, just the max limit is supported by APPLE_texture_max_level extension.
12 years ago
#ifndef MAGNUM_TARGET_GLES2
void setBaseLevel(Int level);
#endif
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !(defined(MAGNUM_TARGET_WEBGL) && defined(MAGNUM_TARGET_GLES2))
Ability to set base and max mip level on textures. Curiously this functionality is not available on ES2, just the max limit is supported by APPLE_texture_max_level extension.
12 years ago
void setMaxLevel(Int level);
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
Split the OpenGL layer out, pt 12: generic sampler state. The Sampler class was split into GL::Sampler (which is now mostly just a placeholder for implementing OpenGL sampler objects), pairs of generic / GL-specific SamplerFormat / GL::SamplerFormat, SamplerMipmap / GL::SamplerMipmap, SamplerWrapping / GL::SamplerWrapping enums and the GL-specific GL::SamplerCompareMode, GL::SamplerCompareFunction, GL::SamplerDepthStencilMode enums. The old Sampler class is marked as deprecated and aliases its enum to the generic enums (or to the GL-specific ones in case the generic versions are not available).
8 years ago
void setMinificationFilter(SamplerFilter filter, SamplerMipmap mipmap);
void setMagnificationFilter(SamplerFilter filter);
Implemented texture LOD parameters.
12 years ago
#ifndef MAGNUM_TARGET_GLES2
void setMinLod(Float lod);
void setMaxLod(Float lod);
#endif
#ifndef MAGNUM_TARGET_GLES
void setLodBias(Float bias);
#endif
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#ifndef MAGNUM_TARGET_WEBGL
Split Texture into TextureArray, MultisampleTexture and RectangleTexture. Each texture has slightly different usage requirements and having everything under one generic class is not worth the additional runtime checks and whatnot. The current way with Texture::Target enum (hopefully not too widely used) is now deprecated and will be removed in some future release. However general Texture1D/2D/3D usage is not changed in any way.
12 years ago
void setBorderColor(const Color4& color);
Implemented EXT_texture_border_clamp AEP extension. Replaced NV_texture_border_clamp enum values with EXT ones, as this is a superset of the NV extension.
11 years ago
#ifndef MAGNUM_TARGET_GLES2
Support for setting border color on integer textures. EXT_texture_integer implementation is now complete. Just GL 3.0 subset, though, as apparently glClearColorI*() is not part of it.
12 years ago
void setBorderColor(const Vector4i& color);
void setBorderColor(const Vector4ui& color);
Integer texture borders are not available in OpenGL ES.
12 years ago
#endif
Implemented EXT_texture_border_clamp AEP extension. Replaced NV_texture_border_clamp enum values with EXT ones, as this is a superset of the NV extension.
11 years ago
#endif
Split Texture into TextureArray, MultisampleTexture and RectangleTexture. Each texture has slightly different usage requirements and having everything under one generic class is not worth the additional runtime checks and whatnot. The current way with Texture::Target enum (hopefully not too widely used) is now deprecated and will be removed in some future release. However general Texture1D/2D/3D usage is not changed in any way.
12 years ago
void setMaxAnisotropy(Float anisotropy);
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#ifndef MAGNUM_TARGET_WEBGL
Renamed various sRGB-related GL APIs for consistency with naming in Math. As usual, the old names are deprecated aliases to the new things.
8 years ago
void setSrgbDecode(bool decode);
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
Implemented ARB_texture_swizzle.
12 years ago
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
Implemented ARB_texture_swizzle.
12 years ago
template<char r, char g, char b, char a> void setSwizzle() {
setSwizzleInternal(Implementation::TextureSwizzle<r>::Value,
Implementation::TextureSwizzle<g>::Value,
Implementation::TextureSwizzle<b>::Value,
Implementation::TextureSwizzle<a>::Value);
}
void setSwizzleInternal(GLint r, GLint g, GLint b, GLint a);
#endif
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !(defined(MAGNUM_TARGET_WEBGL) && defined(MAGNUM_TARGET_GLES2))
Split the OpenGL layer out, pt 12: generic sampler state. The Sampler class was split into GL::Sampler (which is now mostly just a placeholder for implementing OpenGL sampler objects), pairs of generic / GL-specific SamplerFormat / GL::SamplerFormat, SamplerMipmap / GL::SamplerMipmap, SamplerWrapping / GL::SamplerWrapping enums and the GL-specific GL::SamplerCompareMode, GL::SamplerCompareFunction, GL::SamplerDepthStencilMode enums. The old Sampler class is marked as deprecated and aliases its enum to the generic enums (or to the GL-specific ones in case the generic versions are not available).
8 years ago
void setCompareMode(SamplerCompareMode mode);
void setCompareFunction(SamplerCompareFunction function);
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
GL: implement ANGLE_stencil_texturing and WEBGL_stencil_texturing.
2 years ago
#ifndef MAGNUM_TARGET_GLES2
Split the OpenGL layer out, pt 12: generic sampler state. The Sampler class was split into GL::Sampler (which is now mostly just a placeholder for implementing OpenGL sampler objects), pairs of generic / GL-specific SamplerFormat / GL::SamplerFormat, SamplerMipmap / GL::SamplerMipmap, SamplerWrapping / GL::SamplerWrapping enums and the GL-specific GL::SamplerCompareMode, GL::SamplerCompareFunction, GL::SamplerDepthStencilMode enums. The old Sampler class is marked as deprecated and aliases its enum to the generic enums (or to the GL-specific ones in case the generic versions are not available).
8 years ago
void setDepthStencilMode(SamplerDepthStencilMode mode);
Implementation of ARB_stencil_texturing.
12 years ago
#endif
GL: implement {ARB,OES,EXT}_texture_view. "Funny" how this is the only API where I can't use glCreateTextures(). Like, it would have been so easy to just stop teaching glGen*() and all their quirks and "this ID exists but it's not an object until you bind it somewhere actually" to people altogether, BUT NO! FFS.
3 years ago
#if !(defined(MAGNUM_TARGET_GLES2) && defined(MAGNUM_TARGET_WEBGL))
void viewInternal(AbstractTexture& original, TextureFormat internalFormat, Int levelOffset, Int levelCount, Int layerOffset, Int layerCount);
#endif
Split Texture into TextureArray, MultisampleTexture and RectangleTexture. Each texture has slightly different usage requirements and having everything under one generic class is not worth the additional runtime checks and whatnot. The current way with Texture::Target enum (hopefully not too widely used) is now deprecated and will be removed in some future release. However general Texture1D/2D/3D usage is not changed in any way.
12 years ago
void invalidateImage(Int level);
void generateMipmap();
Ability to retrieve texture images.
13 years ago
#ifndef MAGNUM_TARGET_GLES
Mutable image views can be passed via const&. We're not changing the view parameters, only writing to the memory it points to.
7 years ago
template<UnsignedInt dimensions> void image(GLint level, const BasicMutableImageView<dimensions>& image);
GL: fill appropriate ImageFlags when using Texture::image() and friends. Array texture and cubemap (array) subimage queries get ImageFlag*D::Array, cubemap whole image ImageFlag3D::CubeMap, cubemap array whole image get ImageFlag3D::CubeMap and ImageFlag3D::Array. No flags are checked when uploading images or when downloading images to views -- using an array texture as a cube map and vice versa is a valid case, and others are probably also, so there's no point. The flag restrictions will come into place when ImageFlag*D::YUp / YDown etc. are a thing.
4 years ago
template<UnsignedInt dimensions> void image(GLint level, Image<dimensions>& image, ImageFlags<dimensions> flags);
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
template<UnsignedInt dimensions> void image(GLint level, BufferImage<dimensions>& image, BufferUsage usage);
Mutable image views can be passed via const&. We're not changing the view parameters, only writing to the memory it points to.
7 years ago
template<UnsignedInt dimensions> void compressedImage(GLint level, const BasicMutableCompressedImageView<dimensions>& image);
GL: fill appropriate ImageFlags when using Texture::image() and friends. Array texture and cubemap (array) subimage queries get ImageFlag*D::Array, cubemap whole image ImageFlag3D::CubeMap, cubemap array whole image get ImageFlag3D::CubeMap and ImageFlag3D::Array. No flags are checked when uploading images or when downloading images to views -- using an array texture as a cube map and vice versa is a valid case, and others are probably also, so there's no point. The flag restrictions will come into place when ImageFlag*D::YUp / YDown etc. are a thing.
4 years ago
template<UnsignedInt dimensions> void compressedImage(GLint level, CompressedImage<dimensions>& image, ImageFlags<dimensions> flags);
Compressed image support, part 8: compressed image download. Only full images now, no tests yet.
11 years ago
template<UnsignedInt dimensions> void compressedImage(GLint level, CompressedBufferImage<dimensions>& image, BufferUsage usage);
Mutable image views can be passed via const&. We're not changing the view parameters, only writing to the memory it points to.
7 years ago
template<UnsignedInt dimensions> void subImage(GLint level, const RangeTypeFor<dimensions, Int>& range, const BasicMutableImageView<dimensions>& image);
GL: fill appropriate ImageFlags when using Texture::image() and friends. Array texture and cubemap (array) subimage queries get ImageFlag*D::Array, cubemap whole image ImageFlag3D::CubeMap, cubemap array whole image get ImageFlag3D::CubeMap and ImageFlag3D::Array. No flags are checked when uploading images or when downloading images to views -- using an array texture as a cube map and vice versa is a valid case, and others are probably also, so there's no point. The flag restrictions will come into place when ImageFlag*D::YUp / YDown etc. are a thing.
4 years ago
template<UnsignedInt dimensions> void subImage(GLint level, const RangeTypeFor<dimensions, Int>& range, Image<dimensions>& image, ImageFlags<dimensions> flags);
Support for ARB_get_texture_sub_image. Only thing missing are compressed texture queries, but that needs a lot more work.
11 years ago
template<UnsignedInt dimensions> void subImage(GLint level, const RangeTypeFor<dimensions, Int>& range, BufferImage<dimensions>& image, BufferUsage usage);
Mutable image views can be passed via const&. We're not changing the view parameters, only writing to the memory it points to.
7 years ago
template<UnsignedInt dimensions> void compressedSubImage(GLint level, const RangeTypeFor<dimensions, Int>& range, const BasicMutableCompressedImageView<dimensions>& image);
GL: fill appropriate ImageFlags when using Texture::image() and friends. Array texture and cubemap (array) subimage queries get ImageFlag*D::Array, cubemap whole image ImageFlag3D::CubeMap, cubemap array whole image get ImageFlag3D::CubeMap and ImageFlag3D::Array. No flags are checked when uploading images or when downloading images to views -- using an array texture as a cube map and vice versa is a valid case, and others are probably also, so there's no point. The flag restrictions will come into place when ImageFlag*D::YUp / YDown etc. are a thing.
4 years ago
template<UnsignedInt dimensions> void compressedSubImage(GLint level, const RangeTypeFor<dimensions, Int>& range, CompressedImage<dimensions>& image, ImageFlags<dimensions> flags);
Compressed image support, part 15: compressed subimage download. NVidia reports compressed block data size in bits instead of bytes, thus all the test currently fail. (Another) workaround incoming.
11 years ago
template<UnsignedInt dimensions> void compressedSubImage(GLint level, const RangeTypeFor<dimensions, Int>& range, CompressedBufferImage<dimensions>& image, BufferUsage usage);
Ability to retrieve texture images.
13 years ago
#endif
Move constructor and move assignment for textures.
14 years ago
GLenum _target;
Reworked binding of textures to shaders. According to OpenGL specification the sampler uniform takes only texture layer and not any particular texture. In OpenGL 4.2 it is possible to explicitly specify desired texture layer in the uniform, so I suppose it's desired to have shader with fixed texture layers and on the other hand be able to bind texture to any layer, not only one -- the exact opposite of how it was in Magnum before. The shader should now specify in its public API which texture is on which layer and explicitly set the layer uniforms upon construction. All setUniform() functions taking textures as argument were thus removed, as they confuse things a lot. The textures now have bind() function which takes layer as argument, so it's now sufficient to only bind the texture before drawing the mesh without messing with sampler uniform.
14 years ago
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
private:
GL: move friend declarations down into class privates. So the header is not horrifying right from the top.
6 years ago
friend Implementation::TextureState;
friend AbstractFramebuffer;
friend CubeMapTexture;
Added workaround for NVidia reporting compressed block size in bits. Instead of bytes as per specification of glGetInternalformativ().
10 years ago
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
static Int MAGNUM_GL_LOCAL compressedBlockDataSizeImplementationDefault(GLenum target, TextureFormat format);
static Int MAGNUM_GL_LOCAL compressedBlockDataSizeImplementationBitsWorkaround(GLenum target, TextureFormat format);
Added workaround for NVidia reporting compressed block size in bits. Instead of bytes as per specification of glGetInternalformativ().
10 years ago
#endif
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
static void MAGNUM_GL_LOCAL unbindImplementationDefault(GLint textureUnit);
Added *Texture::unbind(). Will be used in fallback implementation of multi-bind.
12 years ago
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
static void MAGNUM_GL_LOCAL unbindImplementationMulti(GLint textureUnit);
static void MAGNUM_GL_LOCAL unbindImplementationDSA(GLint textureUnit);
Added *Texture::unbind(). Will be used in fallback implementation of multi-bind.
12 years ago
#endif
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
static void MAGNUM_GL_LOCAL bindImplementationFallback(GLint firstTextureUnit, Containers::ArrayView<AbstractTexture* const> textures);
Support ARB_multi_bind for textures.
12 years ago
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
static void MAGNUM_GL_LOCAL bindImplementationMulti(GLint firstTextureUnit, Containers::ArrayView<AbstractTexture* const> textures);
Support ARB_multi_bind for textures.
12 years ago
#endif
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL createImplementationDefault(AbstractTexture& self);
ARB_direct_state_access support for creating textures.
12 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL createImplementationDSA(AbstractTexture& self);
ARB_direct_state_access support for creating textures.
12 years ago
#endif
Compressed image support, part 15: compressed subimage download. NVidia reports compressed block data size in bits instead of bytes, thus all the test currently fail. (Another) workaround incoming.
11 years ago
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
template<UnsignedInt dimensions> std::size_t MAGNUM_GL_LOCAL compressedSubImageSize(TextureFormat format, const Math::Vector<dimensions, Int>& size);
Compressed image support, part 15: compressed subimage download. NVidia reports compressed block data size in bits instead of bytes, thus all the test currently fail. (Another) workaround incoming.
11 years ago
#endif
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL bindImplementationDefault(AbstractTexture& self, GLint textureUnit);
Updated OpenGL ES support in textures.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL bindImplementationMulti(AbstractTexture& self, GLint textureUnit);
static void MAGNUM_GL_LOCAL bindImplementationDSA(AbstractTexture& self, GLint textureUnit);
GL: new "intel-windows-half-baked-dsa-texture-bind" workaround. This is broken as well, in a novel and interesting way. Haha.
7 years ago
#ifdef CORRADE_TARGET_WINDOWS
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL bindImplementationDSAIntelWindows(AbstractTexture& self, GLint textureUnit);
GL: new "intel-windows-half-baked-dsa-texture-bind" workaround. This is broken as well, in a novel and interesting way. Haha.
7 years ago
#endif
GL: finalize the apple-buffer-texture-unbind-on-buffer-modify workaround. Followup to 24cc971b1f43c61aa157228f6ea9916007ffb3cd, covering the remaining case.
6 years ago
#ifdef CORRADE_TARGET_APPLE
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL bindImplementationAppleBufferTextureWorkaround(AbstractTexture& self, GLint textureUnit);
GL: finalize the apple-buffer-texture-unbind-on-buffer-modify workaround. Followup to 24cc971b1f43c61aa157228f6ea9916007ffb3cd, covering the remaining case.
6 years ago
#endif
Updated OpenGL ES support in textures.
14 years ago
#endif
Using EXT_direct_state_access in textures. Saves unnecessary calls to glActiveTexture() and glBindTexture().
14 years ago
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL parameterImplementationDefault(AbstractTexture& self, GLenum parameter, GLint value);
static void MAGNUM_GL_LOCAL parameterImplementationDefault(AbstractTexture& self, GLenum parameter, GLfloat value);
Implemented ARB_texture_swizzle.
12 years ago
#ifndef MAGNUM_TARGET_GLES2
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL parameterImplementationDefault(AbstractTexture& self, GLenum parameter, const GLint* values);
Implemented ARB_texture_swizzle.
12 years ago
#endif
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL parameterImplementationDefault(AbstractTexture& self, GLenum parameter, const GLfloat* values);
Implemented EXT_texture_border_clamp AEP extension. Replaced NV_texture_border_clamp enum values with EXT ones, as this is a superset of the NV extension.
11 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL parameterIImplementationDefault(AbstractTexture& self, GLenum parameter, const GLuint* values);
static void MAGNUM_GL_LOCAL parameterIImplementationDefault(AbstractTexture& self, GLenum parameter, const GLint* values);
Updates for texture border clamp support in ES 3.2.
9 years ago
#ifdef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL parameterIImplementationEXT(AbstractTexture& self, GLenum parameter, const GLuint* values);
static void MAGNUM_GL_LOCAL parameterIImplementationEXT(AbstractTexture& self, GLenum parameter, const GLint* values);
Updates for texture border clamp support in ES 3.2.
9 years ago
#endif
Updated OpenGL ES support in textures.
14 years ago
#endif
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL parameterImplementationDSA(AbstractTexture& self, GLenum parameter, GLint value);
static void MAGNUM_GL_LOCAL parameterImplementationDSA(AbstractTexture& self, GLenum parameter, GLfloat value);
static void MAGNUM_GL_LOCAL parameterImplementationDSA(AbstractTexture& self, GLenum parameter, const GLint* values);
static void MAGNUM_GL_LOCAL parameterImplementationDSA(AbstractTexture& self, GLenum parameter, const GLfloat* values);
static void MAGNUM_GL_LOCAL parameterIImplementationDSA(AbstractTexture& self, GLenum parameter, const GLuint* values);
static void MAGNUM_GL_LOCAL parameterIImplementationDSA(AbstractTexture& self, GLenum parameter, const GLint* values);
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
#endif
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL setMaxAnisotropyImplementationNoOp(AbstractTexture& self, GLfloat);
static void MAGNUM_GL_LOCAL setMaxAnisotropyImplementationArbOrExt(AbstractTexture& self, GLfloat anisotropy);
Added no-op fallback for unsupported EXT_texture_filter_anisotropic. This extension is ubiquitous, but to make users' life even easier we now provide no-op fallback for both Sampler::maxAnisotropy() and Texture::setMaxAnisotropy().
13 years ago
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL getLevelParameterImplementationDefault(AbstractTexture& self, GLint level, GLenum parameter, GLint* values);
Enabled texture image size queries in ES 3.1.
12 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL getLevelParameterImplementationDSA(AbstractTexture& self, GLint level, GLenum parameter, GLint* values);
Implementation for texture level parameter query.
13 years ago
#endif
Enabled texture image size queries in ES 3.1.
12 years ago
#endif
Implementation for texture level parameter query.
13 years ago
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL mipmapImplementationDefault(AbstractTexture& self);
Updated OpenGL ES support in textures.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL mipmapImplementationDSA(AbstractTexture& self);
Updated OpenGL ES support in textures.
14 years ago
#endif
Using EXT_direct_state_access in textures. Saves unnecessary calls to glActiveTexture() and glBindTexture().
14 years ago
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageImplementationFallback(AbstractTexture& self, GLsizei levels, TextureFormat internalFormat, const Math::Vector<1, GLsizei>& size);
static void MAGNUM_GL_LOCAL storageImplementationDefault(AbstractTexture& self, GLsizei levels, TextureFormat internalFormat, const Math::Vector<1, GLsizei>& size);
static void MAGNUM_GL_LOCAL storageImplementationDSA(AbstractTexture& self, GLsizei levels, TextureFormat internalFormat, const Math::Vector<1, GLsizei>& size);
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
#endif
Hide fallback texture storage implementation from ES3 builds. Not needed there.
11 years ago
#if !defined(MAGNUM_TARGET_GLES) || defined(MAGNUM_TARGET_GLES2)
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageImplementationFallback(AbstractTexture& self, GLsizei levels, TextureFormat internalFormat, const Vector2i& size);
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
#if !(defined(MAGNUM_TARGET_WEBGL) && defined(MAGNUM_TARGET_GLES2))
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageImplementationDefault(AbstractTexture& self, GLsizei levels, TextureFormat internalFormat, const Vector2i& size);
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageImplementationDSA(AbstractTexture& self, GLsizei levels, TextureFormat internalFormat, const Vector2i& size);
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
#endif
Hide fallback texture storage implementation from ES3 builds. Not needed there.
11 years ago
#if !defined(MAGNUM_TARGET_GLES) || (defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL))
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageImplementationFallback(AbstractTexture& self, GLsizei levels, TextureFormat internalFormat, const Vector3i& size);
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
#if !(defined(MAGNUM_TARGET_WEBGL) && defined(MAGNUM_TARGET_GLES2))
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageImplementationDefault(AbstractTexture& self, GLsizei levels, TextureFormat internalFormat, const Vector3i& size);
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageImplementationDSA(AbstractTexture& self, GLsizei levels, TextureFormat internalFormat, const Vector3i& size);
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
#endif
Added MultisampleTexture::setStorage(). With this function multisample texture implementation is pretty much done. Also implemented all the tests.
12 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageMultisampleImplementationFallback(AbstractTexture& self, GLsizei samples, TextureFormat internalFormat, const Vector2i& size, GLboolean fixedsamplelocations);
static void MAGNUM_GL_LOCAL storageMultisampleImplementationFallback(AbstractTexture& self, GLsizei samples, TextureFormat internalFormat, const Vector3i& size, GLboolean fixedsamplelocations);
Enabled 2D multisample textures in ES 3.1.
12 years ago
#endif
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageMultisampleImplementationDefault(AbstractTexture& self, GLsizei samples, TextureFormat internalFormat, const Vector2i& size, GLboolean fixedsamplelocations);
static void MAGNUM_GL_LOCAL storageMultisampleImplementationDefault(AbstractTexture& self, GLsizei samples, TextureFormat internalFormat, const Vector3i& size, GLboolean fixedsamplelocations);
Updates for multisample array texture storage in ES 3.2.
9 years ago
#ifdef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageMultisampleImplementationOES(AbstractTexture& self, GLsizei samples, TextureFormat internalFormat, const Vector3i& size, GLboolean fixedsamplelocations);
Updates for multisample array texture storage in ES 3.2.
9 years ago
#endif
Enabled 2D multisample textures in ES 3.1.
12 years ago
#endif
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL storageMultisampleImplementationDSA(AbstractTexture& self, GLsizei samples, TextureFormat internalFormat, const Vector2i& size, GLboolean fixedsamplelocations);
static void MAGNUM_GL_LOCAL storageMultisampleImplementationDSA(AbstractTexture& self, GLsizei samples, TextureFormat internalFormat, const Vector3i& size, GLboolean fixedsamplelocations);
Added MultisampleTexture::setStorage(). With this function multisample texture implementation is pretty much done. Also implemented all the tests.
12 years ago
#endif
Ability to retrieve texture images.
13 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL getImageImplementationDefault(AbstractTexture& self, GLint level, PixelFormat format, PixelType type, std::size_t dataSize, GLvoid* data);
static void MAGNUM_GL_LOCAL getImageImplementationDSA(AbstractTexture& self, GLint level, PixelFormat format, PixelType type, std::size_t dataSize, GLvoid* data);
static void MAGNUM_GL_LOCAL getImageImplementationRobustness(AbstractTexture& self, GLint level, PixelFormat format, PixelType type, std::size_t dataSize, GLvoid* data);
Compressed image support, part 8: compressed image download. Only full images now, no tests yet.
11 years ago
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL getCompressedImageImplementationDefault(AbstractTexture& self, GLint level, std::size_t dataSize, GLvoid* data);
static void MAGNUM_GL_LOCAL getCompressedImageImplementationDSA(AbstractTexture& self, GLint level, std::size_t dataSize, GLvoid* data);
static void MAGNUM_GL_LOCAL getCompressedImageImplementationRobustness(AbstractTexture& self, GLint level, std::size_t dataSize, GLvoid* data);
Ability to retrieve texture images.
13 years ago
#endif
Updated OpenGL ES support in textures.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL subImageImplementationDefault(AbstractTexture& self, GLint level, const Math::Vector<1, GLint>& offset, const Math::Vector<1, GLsizei>& size, PixelFormat format, PixelType type, const GLvoid* data);
static void MAGNUM_GL_LOCAL subImageImplementationDSA(AbstractTexture& self, GLint level, const Math::Vector<1, GLint>& offset, const Math::Vector<1, GLsizei>& size, PixelFormat format, PixelType type, const GLvoid* data);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL compressedSubImageImplementationDefault(AbstractTexture& self, GLint level, const Math::Vector<1, GLint>& offset, const Math::Vector<1, GLsizei>& size, CompressedPixelFormat format, const GLvoid* data, GLsizei dataSize);
static void MAGNUM_GL_LOCAL compressedSubImageImplementationDSA(AbstractTexture& self, GLint level, const Math::Vector<1, GLint>& offset, const Math::Vector<1, GLsizei>& size, CompressedPixelFormat format, const GLvoid* data, GLsizei dataSize);
Updated OpenGL ES support in textures.
14 years ago
#endif
Using EXT_direct_state_access in textures. Saves unnecessary calls to glActiveTexture() and glBindTexture().
14 years ago
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL imageImplementationDefault(AbstractTexture& self, GLenum target, GLint level, TextureFormat internalFormat, const Vector2i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage&);
SVGA3D: work around broken glTex[ture][Sub]Image*D() for arrays. SVGA3D has broken handling of glTex[ture][Sub]Image*D() for 1D arrays, 2D arrays, 3D textures and cube map textures where it uploads just the first slice in the last dimension. This is only with copies from host memory, not with buffer images. Seems to be fixed in Mesa 13, but I have no such system to verify that on. Relevant commit in the Mesa sources: https://github.com/mesa3d/mesa/commit/2aa9ff0cda1f6ad97c83d5583fab7a84efabe19e This is one of the uglier workarounds -- I had to reintroduce multiple code paths for glTexImage() which were removed when implementing ARB_DSA and the workaround consists of basically a bunch of functions that are slicing the image and calling the original implementations with each slice.
9 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL imageImplementationSvga3DSliceBySlice(AbstractTexture& self, GLenum target, GLint level, TextureFormat internalFormat, const Vector2i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage&);
SVGA3D: work around broken glTex[ture][Sub]Image*D() for arrays. SVGA3D has broken handling of glTex[ture][Sub]Image*D() for 1D arrays, 2D arrays, 3D textures and cube map textures where it uploads just the first slice in the last dimension. This is only with copies from host memory, not with buffer images. Seems to be fixed in Mesa 13, but I have no such system to verify that on. Relevant commit in the Mesa sources: https://github.com/mesa3d/mesa/commit/2aa9ff0cda1f6ad97c83d5583fab7a84efabe19e This is one of the uglier workarounds -- I had to reintroduce multiple code paths for glTexImage() which were removed when implementing ARB_DSA and the workaround consists of basically a bunch of functions that are slicing the image and calling the original implementations with each slice.
9 years ago
#endif
/* Had to put explicit "2D" in the name so it's not overloaded and
Clang is able to pass it as template parameter to
subImageImplementationSvga3DSliceBySlice(). GCC had no problem with
the original. */
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL subImage2DImplementationDefault(AbstractTexture& self, GLint level, const Vector2i& offset, const Vector2i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage&);
SVGA3D: work around broken glTex[ture][Sub]Image*D() for arrays. SVGA3D has broken handling of glTex[ture][Sub]Image*D() for 1D arrays, 2D arrays, 3D textures and cube map textures where it uploads just the first slice in the last dimension. This is only with copies from host memory, not with buffer images. Seems to be fixed in Mesa 13, but I have no such system to verify that on. Relevant commit in the Mesa sources: https://github.com/mesa3d/mesa/commit/2aa9ff0cda1f6ad97c83d5583fab7a84efabe19e This is one of the uglier workarounds -- I had to reintroduce multiple code paths for glTexImage() which were removed when implementing ARB_DSA and the workaround consists of basically a bunch of functions that are slicing the image and calling the original implementations with each slice.
9 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
template<void(*original)(AbstractTexture&, GLint, const Vector2i&, const Vector2i&, PixelFormat, PixelType, const GLvoid*, const PixelStorage&)> static void MAGNUM_GL_LOCAL subImageImplementationSvga3DSliceBySlice(AbstractTexture& self, GLint level, const Vector2i& offset, const Vector2i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage& storage);
SVGA3D: work around broken glTex[ture][Sub]Image*D() for arrays. SVGA3D has broken handling of glTex[ture][Sub]Image*D() for 1D arrays, 2D arrays, 3D textures and cube map textures where it uploads just the first slice in the last dimension. This is only with copies from host memory, not with buffer images. Seems to be fixed in Mesa 13, but I have no such system to verify that on. Relevant commit in the Mesa sources: https://github.com/mesa3d/mesa/commit/2aa9ff0cda1f6ad97c83d5583fab7a84efabe19e This is one of the uglier workarounds -- I had to reintroduce multiple code paths for glTexImage() which were removed when implementing ARB_DSA and the workaround consists of basically a bunch of functions that are slicing the image and calling the original implementations with each slice.
9 years ago
#endif
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL compressedSubImageImplementationDefault(AbstractTexture& self, GLint level, const Vector2i& offset, const Vector2i& size, CompressedPixelFormat format, const GLvoid* data, GLsizei dataSize);
Forgot to hide DSA functions when targetting ES.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL subImage2DImplementationDSA(AbstractTexture& self, GLint level, const Vector2i& offset, const Vector2i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage&);
static void MAGNUM_GL_LOCAL compressedSubImageImplementationDSA(AbstractTexture& self, GLint level, const Vector2i& offset, const Vector2i& size, CompressedPixelFormat format, const GLvoid* data, GLsizei dataSize);
Forgot to hide DSA functions when targetting ES.
14 years ago
#endif
Using EXT_direct_state_access in textures. Saves unnecessary calls to glActiveTexture() and glBindTexture().
14 years ago
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !(defined(MAGNUM_TARGET_WEBGL) && defined(MAGNUM_TARGET_GLES2))
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL imageImplementationDefault(AbstractTexture& self, GLint level, TextureFormat internalFormat, const Vector3i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage&);
SVGA3D: work around broken glTex[ture][Sub]Image*D() for arrays. SVGA3D has broken handling of glTex[ture][Sub]Image*D() for 1D arrays, 2D arrays, 3D textures and cube map textures where it uploads just the first slice in the last dimension. This is only with copies from host memory, not with buffer images. Seems to be fixed in Mesa 13, but I have no such system to verify that on. Relevant commit in the Mesa sources: https://github.com/mesa3d/mesa/commit/2aa9ff0cda1f6ad97c83d5583fab7a84efabe19e This is one of the uglier workarounds -- I had to reintroduce multiple code paths for glTexImage() which were removed when implementing ARB_DSA and the workaround consists of basically a bunch of functions that are slicing the image and calling the original implementations with each slice.
9 years ago
#ifndef MAGNUM_TARGET_WEBGL
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL imageImplementationSvga3DSliceBySlice(AbstractTexture& self, GLint level, TextureFormat internalFormat, const Vector3i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage&);
SVGA3D: work around broken glTex[ture][Sub]Image*D() for arrays. SVGA3D has broken handling of glTex[ture][Sub]Image*D() for 1D arrays, 2D arrays, 3D textures and cube map textures where it uploads just the first slice in the last dimension. This is only with copies from host memory, not with buffer images. Seems to be fixed in Mesa 13, but I have no such system to verify that on. Relevant commit in the Mesa sources: https://github.com/mesa3d/mesa/commit/2aa9ff0cda1f6ad97c83d5583fab7a84efabe19e This is one of the uglier workarounds -- I had to reintroduce multiple code paths for glTexImage() which were removed when implementing ARB_DSA and the workaround consists of basically a bunch of functions that are slicing the image and calling the original implementations with each slice.
9 years ago
#endif
/* Had to put explicit "3D" in the name so it's not overloaded and
Clang is able to pass it as template parameter to
subImageImplementationSvga3DSliceBySlice(). GCC had no problem with
the original. */
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL subImage3DImplementationDefault(AbstractTexture& self, GLint level, const Vector3i& offset, const Vector3i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage&);
SVGA3D: work around broken glTex[ture][Sub]Image*D() for arrays. SVGA3D has broken handling of glTex[ture][Sub]Image*D() for 1D arrays, 2D arrays, 3D textures and cube map textures where it uploads just the first slice in the last dimension. This is only with copies from host memory, not with buffer images. Seems to be fixed in Mesa 13, but I have no such system to verify that on. Relevant commit in the Mesa sources: https://github.com/mesa3d/mesa/commit/2aa9ff0cda1f6ad97c83d5583fab7a84efabe19e This is one of the uglier workarounds -- I had to reintroduce multiple code paths for glTexImage() which were removed when implementing ARB_DSA and the workaround consists of basically a bunch of functions that are slicing the image and calling the original implementations with each slice.
9 years ago
#ifndef MAGNUM_TARGET_WEBGL
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
template<void(*original)(AbstractTexture&, GLint, const Vector3i&, const Vector3i&, PixelFormat, PixelType, const GLvoid*, const PixelStorage&)> static void MAGNUM_GL_LOCAL subImageImplementationSvga3DSliceBySlice(AbstractTexture& self, GLint level, const Vector3i& offset, const Vector3i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage& storage);
SVGA3D: work around broken glTex[ture][Sub]Image*D() for arrays. SVGA3D has broken handling of glTex[ture][Sub]Image*D() for 1D arrays, 2D arrays, 3D textures and cube map textures where it uploads just the first slice in the last dimension. This is only with copies from host memory, not with buffer images. Seems to be fixed in Mesa 13, but I have no such system to verify that on. Relevant commit in the Mesa sources: https://github.com/mesa3d/mesa/commit/2aa9ff0cda1f6ad97c83d5583fab7a84efabe19e This is one of the uglier workarounds -- I had to reintroduce multiple code paths for glTexImage() which were removed when implementing ARB_DSA and the workaround consists of basically a bunch of functions that are slicing the image and calling the original implementations with each slice.
9 years ago
#endif
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL compressedSubImageImplementationDefault(AbstractTexture& self, GLint level, const Vector3i& offset, const Vector3i& size, CompressedPixelFormat format, const GLvoid* data, GLsizei dataSize);
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
Forgot to hide DSA functions when targetting ES.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL subImage3DImplementationDSA(AbstractTexture& self, GLint level, const Vector3i& offset, const Vector3i& size, PixelFormat format, PixelType type, const GLvoid* data, const PixelStorage&);
SVGA3D: work around broken glTex[ture][Sub]Image*D() for arrays. SVGA3D has broken handling of glTex[ture][Sub]Image*D() for 1D arrays, 2D arrays, 3D textures and cube map textures where it uploads just the first slice in the last dimension. This is only with copies from host memory, not with buffer images. Seems to be fixed in Mesa 13, but I have no such system to verify that on. Relevant commit in the Mesa sources: https://github.com/mesa3d/mesa/commit/2aa9ff0cda1f6ad97c83d5583fab7a84efabe19e This is one of the uglier workarounds -- I had to reintroduce multiple code paths for glTexImage() which were removed when implementing ARB_DSA and the workaround consists of basically a bunch of functions that are slicing the image and calling the original implementations with each slice.
9 years ago
#ifndef MAGNUM_TARGET_WEBGL
#endif
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL compressedSubImageImplementationDSA(AbstractTexture& self, GLint level, const Vector3i& offset, const Vector3i& size, CompressedPixelFormat format, const GLvoid* data, GLsizei dataSize);
Forgot to hide DSA functions when targetting ES.
14 years ago
#endif
Using EXT_direct_state_access in textures. Saves unnecessary calls to glActiveTexture() and glBindTexture().
14 years ago
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL invalidateImageImplementationNoOp(AbstractTexture& self, GLint level);
Extension-aware Buffer and *Texture data invalidation. If ARB_invalidate_subdata is not available, these functions do nothing, instead of crashing on null pointer dereference. It results in more convenient usage, enabling users to call them whenever they want, improving performance on implementations which supports that.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL invalidateImageImplementationARB(AbstractTexture& self, GLint level);
Extension-aware Buffer and *Texture data invalidation. If ARB_invalidate_subdata is not available, these functions do nothing, instead of crashing on null pointer dereference. It results in more convenient usage, enabling users to call them whenever they want, improving performance on implementations which supports that.
14 years ago
#endif
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL invalidateSubImageImplementationNoOp(AbstractTexture& self, GLint level, const Vector3i& offset, const Vector3i& size);
Extension-aware Buffer and *Texture data invalidation. If ARB_invalidate_subdata is not available, these functions do nothing, instead of crashing on null pointer dereference. It results in more convenient usage, enabling users to call them whenever they want, improving performance on implementations which supports that.
14 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
static void MAGNUM_GL_LOCAL invalidateSubImageImplementationARB(AbstractTexture& self, GLint level, const Vector3i& offset, const Vector3i& size);
Extension-aware Buffer and *Texture data invalidation. If ARB_invalidate_subdata is not available, these functions do nothing, instead of crashing on null pointer dereference. It results in more convenient usage, enabling users to call them whenever they want, improving performance on implementations which supports that.
14 years ago
#endif
Refactoring of internal OpenGL object names.
14 years ago
GLuint _id;
Ability to wrap and release existing OpenGL objects. Allows for better interaction with third-party libraries. I should probably write a new documentation page about OpenGL wrapping.
11 years ago
ObjectFlags _flags;
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
};
Fixed wrong automagic texture target for 3D textures. While GL_TEXTURE_2D == GL_TEXTURE_1D + 1, GL_TEXTURE_3D is something absolutely different and cannot be computed like this. Now done via AbstractTexture::DataHelper, which is cleaner and better readable.
15 years ago
#ifndef DOXYGEN_GENERATING_OUTPUT
Port to OpenGL ES 2. Desktop OpenGL and OpenGL ES 2 support can be switched using CMake TARGET_GLES option. All functionality not supported in ES is marked in documentation. If targetting OpenGL ES, GLES2/gl2.h is included instead of GLEW. Mesh class now uses VAOs only in desktop OpenGL, in ES the buffers are bound on each draw call.
14 years ago
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
template<> struct MAGNUM_GL_EXPORT AbstractTexture::DataHelper<1> {
Compressed image support, part 14: queries for compressed block sizes.
11 years ago
static Math::Vector<1, GLint> compressedBlockSize(GLenum target, TextureFormat format);
Don't spread cubemap texture level parameter query workarounds elsewhere. Pre-DSA code paths need to specify for which face we are querying the level parameters, which meant that all other calls had to specify the (implicit) target too. I'm also preparing to put a cubemap-specific workaround in the level parameter query and that really shouldn't be present in the generic implementation for all texture types. The other place where a specific target is needed is in setImage() implementations, but these are rather big chunks of code and I don't feel like copying these verbatim to cubemap implementation just to isolate the workaround in one place.
10 years ago
static Math::Vector<1, GLint> imageSize(AbstractTexture& texture, GLint level);
Ability to retrieve texture image size in given mip level.
13 years ago
Deprecate the weird Magnum::Array class. Its only use was for specifying N-dimensional SamplerWrapping because, compared to a Math::Vector, it had an implicit constructor from a single value (whereas the Vector has it explicit). I solved that by simply adding a few single-value overloads where it mattered. There, done. No need for this weird thing and confusion with Containers::Array anymore. All places that used it now use Math::VectorN<SamplerWrapping>, but the class is still included for backwards compatibility purposes, together with providing implicit conversion from and to a Vector.
5 years ago
static void setWrapping(AbstractTexture& texture, const Math::Vector<1, SamplerWrapping>& wrapping);
Greatly simplified Texture::setWrapping(). Temporarily removed the check for not setting repeat wrap mode on rectangle textures.
14 years ago
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setStorage(AbstractTexture& texture, GLsizei levels, TextureFormat internalFormat, const Math::Vector<1, GLsizei>& size);
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
Compressed image support, part 2: renamed ImageReference to ImageView. With pixel pack/unpack support it will be possible to create views onto sub-images, renamed the class to reflect that. The old Magnum/ImageReference.h and ImageReference types are now aliases to ImageView.h and ImageView types, are marked as deprecated and will be removed in future release.
11 years ago
static void setImage(AbstractTexture& texture, GLint level, TextureFormat internalFormat, const ImageView1D& image);
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setImage(AbstractTexture& texture, GLint level, TextureFormat internalFormat, BufferImage1D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedImage(AbstractTexture& texture, GLint level, const CompressedImageView1D& image);
static void setCompressedImage(AbstractTexture& texture, GLint level, CompressedBufferImage1D& image);
Added classes Image and BufferedImage. These classes are meant to be used in the same texture updating functions as Trade::ImageData due to static polymorphism. In addition to Trade::ImageData, which is read-only, these classes support updating the dimensions and data. Image2D and ImageBuffer2D can update the data also from framebuffer.
14 years ago
Compressed image support, part 2: renamed ImageReference to ImageView. With pixel pack/unpack support it will be possible to create views onto sub-images, renamed the class to reflect that. The old Magnum/ImageReference.h and ImageReference types are now aliases to ImageView.h and ImageView types, are marked as deprecated and will be removed in future release.
11 years ago
static void setSubImage(AbstractTexture& texture, GLint level, const Math::Vector<1, GLint>& offset, const ImageView1D& image);
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setSubImage(AbstractTexture& texture, GLint level, const Math::Vector<1, GLint>& offset, BufferImage1D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedSubImage(AbstractTexture& texture, GLint level, const Math::Vector<1, GLint>& offset, const CompressedImageView1D& image);
static void setCompressedSubImage(AbstractTexture& texture, GLint level, const Math::Vector<1, GLint>& offset, CompressedBufferImage1D& image);
Implemented ARB_invalidate_subdata GL 4.3 extension. Framebuffer invalidation is also available in OpenGL ES 3.0 and ES 2.0 using EXT_discard_framebuffer extension.
14 years ago
Get rid of pointers in internal texture implementation.
12 years ago
static void invalidateSubImage(AbstractTexture& texture, GLint level, const Math::Vector<1, GLint>& offset, const Math::Vector<1, GLint>& size);
Fixed wrong automagic texture target for 3D textures. While GL_TEXTURE_2D == GL_TEXTURE_1D + 1, GL_TEXTURE_3D is something absolutely different and cannot be computed like this. Now done via AbstractTexture::DataHelper, which is cleaner and better readable.
15 years ago
};
Port to OpenGL ES 2. Desktop OpenGL and OpenGL ES 2 support can be switched using CMake TARGET_GLES option. All functionality not supported in ES is marked in documentation. If targetting OpenGL ES, GLES2/gl2.h is included instead of GLEW. Mesh class now uses VAOs only in desktop OpenGL, in ES the buffers are bound on each draw call.
14 years ago
#endif
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
template<> struct MAGNUM_GL_EXPORT AbstractTexture::DataHelper<2> {
Compressed image support, part 14: queries for compressed block sizes.
11 years ago
#ifndef MAGNUM_TARGET_GLES
static Vector2i compressedBlockSize(GLenum target, TextureFormat format);
#endif
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
Don't spread cubemap texture level parameter query workarounds elsewhere. Pre-DSA code paths need to specify for which face we are querying the level parameters, which meant that all other calls had to specify the (implicit) target too. I'm also preparing to put a cubemap-specific workaround in the level parameter query and that really shouldn't be present in the generic implementation for all texture types. The other place where a specific target is needed is in setImage() implementations, but these are rather big chunks of code and I don't feel like copying these verbatim to cubemap implementation just to isolate the workaround in one place.
10 years ago
static Vector2i imageSize(AbstractTexture& texture, GLint level);
Ability to retrieve texture image size in given mip level.
13 years ago
#endif
Deprecate the weird Magnum::Array class. Its only use was for specifying N-dimensional SamplerWrapping because, compared to a Math::Vector, it had an implicit constructor from a single value (whereas the Vector has it explicit). I solved that by simply adding a few single-value overloads where it mattered. There, done. No need for this weird thing and confusion with Containers::Array anymore. All places that used it now use Math::VectorN<SamplerWrapping>, but the class is still included for backwards compatibility purposes, together with providing implicit conversion from and to a Vector.
5 years ago
static void setWrapping(AbstractTexture& texture, const Math::Vector2<SamplerWrapping>& wrapping);
Greatly simplified Texture::setWrapping(). Temporarily removed the check for not setting repeat wrap mode on rectangle textures.
14 years ago
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setStorage(AbstractTexture& texture, GLsizei levels, TextureFormat internalFormat, const Vector2i& size);
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setStorageMultisample(AbstractTexture& texture, GLsizei samples, TextureFormat internalFormat, const Vector2i& size, GLboolean fixedSampleLocations);
Enabled 2D multisample textures in ES 3.1.
12 years ago
#endif
Added MultisampleTexture::setStorage(). With this function multisample texture implementation is pretty much done. Also implemented all the tests.
12 years ago
Compressed image support, part 2: renamed ImageReference to ImageView. With pixel pack/unpack support it will be possible to create views onto sub-images, renamed the class to reflect that. The old Magnum/ImageReference.h and ImageReference types are now aliases to ImageView.h and ImageView types, are marked as deprecated and will be removed in future release.
11 years ago
static void setImage(AbstractTexture& texture, GLint level, TextureFormat internalFormat, const ImageView2D& image) {
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
setImage(texture, texture._target, level, internalFormat, image);
}
Compressed image support, part 2: renamed ImageReference to ImageView. With pixel pack/unpack support it will be possible to create views onto sub-images, renamed the class to reflect that. The old Magnum/ImageReference.h and ImageReference types are now aliases to ImageView.h and ImageView types, are marked as deprecated and will be removed in future release.
11 years ago
static void setImage(AbstractTexture& texture, GLenum target, GLint level, TextureFormat internalFormat, const ImageView2D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedImage(AbstractTexture& texture, GLint level, const CompressedImageView2D& image) {
setCompressedImage(texture, texture._target, level, image);
}
static void setCompressedImage(AbstractTexture& texture, GLenum target, GLint level, const CompressedImageView2D& image);
Pass image as reference, not pointer to *Texture::set*Image(). Avoids some nullptr-related errors and leads to simpler implementation, as we need to only handle ImageReference and BufferImage types. All other Image classes are implicitly convertible to ImageReference and implicit conversion works only with references, not with pointers. The usage might be clumsy at this point, as Trade::*Importer classes return pointer to Trade::ImageData which then needs to be dereferenced. Hopefully introducing C++14's std::optional (or equivalent) might help solve that issue. Also removed "convenience" overloads for setting e.g. 2D subdata of 3D texture. The convenience overload doesn't cover all subdata cases (only XY plane, not YZ or XZ ones) and overly complicates the implementation. This can be done in the future in Image classes themselves.
13 years ago
#ifndef MAGNUM_TARGET_GLES2
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setImage(AbstractTexture& texture, GLint level, TextureFormat internalFormat, BufferImage2D& image) {
setImage(texture, texture._target, level, internalFormat, image);
}
Get rid of pointers in internal texture implementation.
12 years ago
static void setImage(AbstractTexture& texture, GLenum target, GLint level, TextureFormat internalFormat, BufferImage2D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedImage(AbstractTexture& texture, GLint level, CompressedBufferImage2D& image) {
setCompressedImage(texture, texture._target, level, image);
}
static void setCompressedImage(AbstractTexture& texture, GLenum target, GLint level, CompressedBufferImage2D& image);
Pass image as reference, not pointer to *Texture::set*Image(). Avoids some nullptr-related errors and leads to simpler implementation, as we need to only handle ImageReference and BufferImage types. All other Image classes are implicitly convertible to ImageReference and implicit conversion works only with references, not with pointers. The usage might be clumsy at this point, as Trade::*Importer classes return pointer to Trade::ImageData which then needs to be dereferenced. Hopefully introducing C++14's std::optional (or equivalent) might help solve that issue. Also removed "convenience" overloads for setting e.g. 2D subdata of 3D texture. The convenience overload doesn't cover all subdata cases (only XY plane, not YZ or XZ ones) and overly complicates the implementation. This can be done in the future in Image classes themselves.
13 years ago
#endif
Ability to set texture subdata from image with one less dimensions. It is now possible to call Texture2D::setSubData() with 1D image and Texture3D::setSubData() with 2D image.
14 years ago
Compressed image support, part 2: renamed ImageReference to ImageView. With pixel pack/unpack support it will be possible to create views onto sub-images, renamed the class to reflect that. The old Magnum/ImageReference.h and ImageReference types are now aliases to ImageView.h and ImageView types, are marked as deprecated and will be removed in future release.
11 years ago
static void setSubImage(AbstractTexture& texture, GLint level, const Vector2i& offset, const ImageView2D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedSubImage(AbstractTexture& texture, GLint level, const Vector2i& offset, const CompressedImageView2D& image);
Pass image as reference, not pointer to *Texture::set*Image(). Avoids some nullptr-related errors and leads to simpler implementation, as we need to only handle ImageReference and BufferImage types. All other Image classes are implicitly convertible to ImageReference and implicit conversion works only with references, not with pointers. The usage might be clumsy at this point, as Trade::*Importer classes return pointer to Trade::ImageData which then needs to be dereferenced. Hopefully introducing C++14's std::optional (or equivalent) might help solve that issue. Also removed "convenience" overloads for setting e.g. 2D subdata of 3D texture. The convenience overload doesn't cover all subdata cases (only XY plane, not YZ or XZ ones) and overly complicates the implementation. This can be done in the future in Image classes themselves.
13 years ago
#ifndef MAGNUM_TARGET_GLES2
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setSubImage(AbstractTexture& texture, GLint level, const Vector2i& offset, BufferImage2D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedSubImage(AbstractTexture& texture, GLint level, const Vector2i& offset, CompressedBufferImage2D& image);
Pass image as reference, not pointer to *Texture::set*Image(). Avoids some nullptr-related errors and leads to simpler implementation, as we need to only handle ImageReference and BufferImage types. All other Image classes are implicitly convertible to ImageReference and implicit conversion works only with references, not with pointers. The usage might be clumsy at this point, as Trade::*Importer classes return pointer to Trade::ImageData which then needs to be dereferenced. Hopefully introducing C++14's std::optional (or equivalent) might help solve that issue. Also removed "convenience" overloads for setting e.g. 2D subdata of 3D texture. The convenience overload doesn't cover all subdata cases (only XY plane, not YZ or XZ ones) and overly complicates the implementation. This can be done in the future in Image classes themselves.
13 years ago
#endif
Implemented ARB_invalidate_subdata GL 4.3 extension. Framebuffer invalidation is also available in OpenGL ES 3.0 and ES 2.0 using EXT_discard_framebuffer extension.
14 years ago
Get rid of pointers in internal texture implementation.
12 years ago
static void invalidateSubImage(AbstractTexture& texture, GLint level, const Vector2i& offset, const Vector2i& size);
Fixed wrong automagic texture target for 3D textures. While GL_TEXTURE_2D == GL_TEXTURE_1D + 1, GL_TEXTURE_3D is something absolutely different and cannot be computed like this. Now done via AbstractTexture::DataHelper, which is cleaner and better readable.
15 years ago
};
Split the OpenGL layer out, pt 1: moving files to a new location. Minimal updates (just the include guards) so Git is hopefully able to detect the rename and track the history properly. Everything except Magnum::GL doesn't compile now.
8 years ago
template<> struct MAGNUM_GL_EXPORT AbstractTexture::DataHelper<3> {
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !(defined(MAGNUM_TARGET_WEBGL) && defined(MAGNUM_TARGET_GLES2))
Compressed image support, part 14: queries for compressed block sizes.
11 years ago
#ifndef MAGNUM_TARGET_GLES
static Vector3i compressedBlockSize(GLenum target, TextureFormat format);
#endif
Enabled texture image size queries in ES 3.1.
12 years ago
#ifndef MAGNUM_TARGET_GLES2
Don't spread cubemap texture level parameter query workarounds elsewhere. Pre-DSA code paths need to specify for which face we are querying the level parameters, which meant that all other calls had to specify the (implicit) target too. I'm also preparing to put a cubemap-specific workaround in the level parameter query and that really shouldn't be present in the generic implementation for all texture types. The other place where a specific target is needed is in setImage() implementations, but these are rather big chunks of code and I don't feel like copying these verbatim to cubemap implementation just to isolate the workaround in one place.
10 years ago
static Vector3i imageSize(AbstractTexture& texture, GLint level);
Ability to retrieve texture image size in given mip level.
13 years ago
#endif
Deprecate the weird Magnum::Array class. Its only use was for specifying N-dimensional SamplerWrapping because, compared to a Math::Vector, it had an implicit constructor from a single value (whereas the Vector has it explicit). I solved that by simply adding a few single-value overloads where it mattered. There, done. No need for this weird thing and confusion with Containers::Array anymore. All places that used it now use Math::VectorN<SamplerWrapping>, but the class is still included for backwards compatibility purposes, together with providing implicit conversion from and to a Vector.
5 years ago
static void setWrapping(AbstractTexture& texture, const Math::Vector3<SamplerWrapping>& wrapping);
Greatly simplified Texture::setWrapping(). Temporarily removed the check for not setting repeat wrap mode on rectangle textures.
14 years ago
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setStorage(AbstractTexture& texture, GLsizei levels, TextureFormat internalFormat, const Vector3i& size);
Implemented ARB_texture_storage. Equivalent functionality is in ES3 and also EXT_texture_storage ES2 extension.
14 years ago
Implemented OES_texture_storage_multisample_2d_array AEP extension.
11 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setStorageMultisample(AbstractTexture& texture, GLsizei samples, TextureFormat internalFormat, const Vector3i& size, GLboolean fixedSampleLocations);
Enabled 2D multisample textures in ES 3.1.
12 years ago
#endif
Added MultisampleTexture::setStorage(). With this function multisample texture implementation is pretty much done. Also implemented all the tests.
12 years ago
Compressed image support, part 2: renamed ImageReference to ImageView. With pixel pack/unpack support it will be possible to create views onto sub-images, renamed the class to reflect that. The old Magnum/ImageReference.h and ImageReference types are now aliases to ImageView.h and ImageView types, are marked as deprecated and will be removed in future release.
11 years ago
static void setImage(AbstractTexture& texture, GLint level, TextureFormat internalFormat, const ImageView3D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedImage(AbstractTexture& texture, GLint level, const CompressedImageView3D& image);
Pass image as reference, not pointer to *Texture::set*Image(). Avoids some nullptr-related errors and leads to simpler implementation, as we need to only handle ImageReference and BufferImage types. All other Image classes are implicitly convertible to ImageReference and implicit conversion works only with references, not with pointers. The usage might be clumsy at this point, as Trade::*Importer classes return pointer to Trade::ImageData which then needs to be dereferenced. Hopefully introducing C++14's std::optional (or equivalent) might help solve that issue. Also removed "convenience" overloads for setting e.g. 2D subdata of 3D texture. The convenience overload doesn't cover all subdata cases (only XY plane, not YZ or XZ ones) and overly complicates the implementation. This can be done in the future in Image classes themselves.
13 years ago
#ifndef MAGNUM_TARGET_GLES2
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setImage(AbstractTexture& texture, GLint level, TextureFormat internalFormat, BufferImage3D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedImage(AbstractTexture& texture, GLint level, CompressedBufferImage3D& image);
Pass image as reference, not pointer to *Texture::set*Image(). Avoids some nullptr-related errors and leads to simpler implementation, as we need to only handle ImageReference and BufferImage types. All other Image classes are implicitly convertible to ImageReference and implicit conversion works only with references, not with pointers. The usage might be clumsy at this point, as Trade::*Importer classes return pointer to Trade::ImageData which then needs to be dereferenced. Hopefully introducing C++14's std::optional (or equivalent) might help solve that issue. Also removed "convenience" overloads for setting e.g. 2D subdata of 3D texture. The convenience overload doesn't cover all subdata cases (only XY plane, not YZ or XZ ones) and overly complicates the implementation. This can be done in the future in Image classes themselves.
13 years ago
#endif
Ability to set texture subdata from image with one less dimensions. It is now possible to call Texture2D::setSubData() with 1D image and Texture3D::setSubData() with 2D image.
14 years ago
Compressed image support, part 2: renamed ImageReference to ImageView. With pixel pack/unpack support it will be possible to create views onto sub-images, renamed the class to reflect that. The old Magnum/ImageReference.h and ImageReference types are now aliases to ImageView.h and ImageView types, are marked as deprecated and will be removed in future release.
11 years ago
static void setSubImage(AbstractTexture& texture, GLint level, const Vector3i& offset, const ImageView3D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedSubImage(AbstractTexture& texture, GLint level, const Vector3i& offset, const CompressedImageView3D& image);
Pass image as reference, not pointer to *Texture::set*Image(). Avoids some nullptr-related errors and leads to simpler implementation, as we need to only handle ImageReference and BufferImage types. All other Image classes are implicitly convertible to ImageReference and implicit conversion works only with references, not with pointers. The usage might be clumsy at this point, as Trade::*Importer classes return pointer to Trade::ImageData which then needs to be dereferenced. Hopefully introducing C++14's std::optional (or equivalent) might help solve that issue. Also removed "convenience" overloads for setting e.g. 2D subdata of 3D texture. The convenience overload doesn't cover all subdata cases (only XY plane, not YZ or XZ ones) and overly complicates the implementation. This can be done in the future in Image classes themselves.
13 years ago
#ifndef MAGNUM_TARGET_GLES2
Full ARB_direct_state_access support in textures. ARB_DSA is now preferred in single-bind cases, as it is easier to use than passing pointers to ARB_multi_bind. ARB_multi_bind was preferred for single-bind previously simply because EXT_DSA was not in core. Because there is a lot to say about feature selection for each function, I took this as a opportunity to remove redundant documentation blocks, just refer to Texture documentation from everywhere and add extension requirements and deprecation where needed, so it's clear for each class what needs what. ARB_DSA also took the opportunity to finally remove all target enum values from function calls and because of that the CubeMapTexture has to handle a bunch of special cases. In order: - CubeMapTexture::imageSize() now doesn't take face parameter and returns one value for all faces. I'm thus now also assuming that the user is sane and called either setStorage() or setImage() with the same size for all faces. In non-ARB_DSA path I'm thus querying only size of +X face and returning it as size for all faces. The old imageSize(Coordinate, Int) overload is still present, but ignores the first parameter and calls imageSize(Int). It is marked as deprecated and will be removed in some future release. - CubeMapTexture::image() now needs to call glTextureSubImage() in ARB_DSA path to make it possible to extract single face. Other code paths (EXT_DSA, Robustness and "default") remain the same. - CubeMapTexture::setSubImage() calls glTextureSubImage3D() in ARB_DSA, because it is not possible to specify face index in glTextureSubImage2D(). Other code paths (EXT_DSA and "default") remain the same. Implementation of these special cases is extracted into CubeMapTexture class to avoid pollution of AbstractTexture with incompatible nonsense.
12 years ago
static void setSubImage(AbstractTexture& texture, GLint level, const Vector3i& offset, BufferImage3D& image);
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
static void setCompressedSubImage(AbstractTexture& texture, GLint level, const Vector3i& offset, CompressedBufferImage3D& image);
Pass image as reference, not pointer to *Texture::set*Image(). Avoids some nullptr-related errors and leads to simpler implementation, as we need to only handle ImageReference and BufferImage types. All other Image classes are implicitly convertible to ImageReference and implicit conversion works only with references, not with pointers. The usage might be clumsy at this point, as Trade::*Importer classes return pointer to Trade::ImageData which then needs to be dereferenced. Hopefully introducing C++14's std::optional (or equivalent) might help solve that issue. Also removed "convenience" overloads for setting e.g. 2D subdata of 3D texture. The convenience overload doesn't cover all subdata cases (only XY plane, not YZ or XZ ones) and overly complicates the implementation. This can be done in the future in Image classes themselves.
13 years ago
#endif
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
Implemented ARB_invalidate_subdata GL 4.3 extension. Framebuffer invalidation is also available in OpenGL ES 3.0 and ES 2.0 using EXT_discard_framebuffer extension.
14 years ago
Get rid of pointers in internal texture implementation.
12 years ago
static void invalidateSubImage(AbstractTexture& texture, GLint level, const Vector3i& offset, const Vector3i& size);
Fixed wrong automagic texture target for 3D textures. While GL_TEXTURE_2D == GL_TEXTURE_1D + 1, GL_TEXTURE_3D is something absolutely different and cannot be computed like this. Now done via AbstractTexture::DataHelper, which is cleaner and better readable.
15 years ago
};
#endif
Ability to wrap and release existing OpenGL objects. Allows for better interaction with third-party libraries. I should probably write a new documentation page about OpenGL wrapping.
11 years ago
inline AbstractTexture::AbstractTexture(AbstractTexture&& other) noexcept: _target{other._target}, _id{other._id}, _flags{other._flags} {
Simplified and tested Texture move constructor/assignment. Move constructor and move assignment is now noexcept and it just swaps the data of the two instances instead of calling GL API, thus it can be inline. Also removed unneded limitations in BufferTexture.
13 years ago
other._id = 0;
}
inline AbstractTexture& AbstractTexture::operator=(AbstractTexture&& other) noexcept {
Use Utility::move() and swap() instead of STL almost everywhere. Partially needed to avoid build breakages because Corrade itself switched as well, partially because a cleanup is always good. Done except for (STL-heavy) code that's deprecated or SceneGraph-related APIs that are still quite full of STL as well.
3 years ago
using Utility::swap;
Properly use the "Std Swap Two Step". I hate this. Why it isn't intuitive like proposed in here? http://ericniebler.com/2014/10/21/customization-point-design-in-c11-and-beyond/
11 years ago
swap(_target, other._target);
swap(_id, other._id);
Ability to wrap and release existing OpenGL objects. Allows for better interaction with third-party libraries. I should probably write a new documentation page about OpenGL wrapping.
11 years ago
swap(_flags, other._flags);
Simplified and tested Texture move constructor/assignment. Move constructor and move assignment is now noexcept and it just swaps the data of the two instances instead of calling GL API, thus it can be inline. Also removed unneded limitations in BufferTexture.
13 years ago
return *this;
}
Ability to wrap and release existing OpenGL objects. Allows for better interaction with third-party libraries. I should probably write a new documentation page about OpenGL wrapping.
11 years ago
inline GLuint AbstractTexture::release() {
const GLuint id = _id;
_id = 0;
return id;
}
GL: clang-cl is this what you want?
6 years ago
#ifdef CORRADE_TARGET_CLANG_CL
/* Otherwise Clang-CL complains these functions are not defined */
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
extern template void AbstractTexture::subImageImplementationSvga3DSliceBySlice<&AbstractTexture::subImage2DImplementationDefault>(AbstractTexture&, GLint, const Vector2i&, const Vector2i&, PixelFormat, PixelType, const GLvoid*, const PixelStorage&);
extern template void AbstractTexture::subImageImplementationSvga3DSliceBySlice<&AbstractTexture::subImage2DImplementationDSA>(AbstractTexture&, GLint, const Vector2i&, const Vector2i&, PixelFormat, PixelType, const GLvoid*, const PixelStorage&);
GL: clang-cl is this what you want?
6 years ago
#endif
#ifndef MAGNUM_TARGET_WEBGL
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
extern template void AbstractTexture::subImageImplementationSvga3DSliceBySlice<&AbstractTexture::subImage3DImplementationDefault>(AbstractTexture&, GLint, const Vector3i&, const Vector3i&, PixelFormat, PixelType, const GLvoid*, const PixelStorage&);
GL: clang-cl is this what you want?
6 years ago
#ifndef MAGNUM_TARGET_GLES
GL: clean up remaining member function pointers in the state tracker. And use static functions with an explicit "self" pointer instead. Those have half the size (8 vs 16 bytes on 64bit x86), which in turn reduces the state tracker memory use by about 750 bytes. On desktop GL with an Intel GPU & Mesa this reduces the state tracker allocation size by almot 10%, from 8.3 kB to 7.6 kB. Not bad. Apart from small memory savings, this also removes the need to include the full class definiton from the State headers on MSVC (because on that compiler the member function pointer size is different based on whether the type definition is known or not, IMAGINE THAT BEING A FEATURE AND NOT A BUG), leading to less header dependencies and better incremental compile times there. This was already done in some cases (and the Vk library used this from the beginning), and as I'm about to add some more extension-dependent functionality it felt like a good time to finish that change, finally. In some cases the *Implementation() could even be dropped in favor of pointing to the GL API directly (such as is already done for various glUniform*() calls), that'd be another step -- this is good enough for now.
3 years ago
extern template void AbstractTexture::subImageImplementationSvga3DSliceBySlice<&AbstractTexture::subImage3DImplementationDSA>(AbstractTexture&, GLint, const Vector3i&, const Vector3i&, PixelFormat, PixelType, const GLvoid*, const PixelStorage&);
GL: clang-cl is this what you want?
6 years ago
#endif
#endif
#endif
GL: removed deprecated aliases to GL functionality in the root namespace. Deprecated for 2018.04, it's been almost a year since. Whoever is using Magnum regularly updated already, and who not can always upgrade gradually (2018.02, 2018.04, 2018.10, 2019.01 etc.).
7 years ago
}}
Splitting non-templated things from Texture to AbstractTexture.
15 years ago
#endif