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.
9477 Commits
21 Branches
19 Tags
70 MiB
Tree: bc884428f8
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'bc884428f8'
${ noResults }
magnum/src/Magnum/GL/Implementation/TextureState.cpp

560 lines
27 KiB
Raw Normal View History Unescape

De-inline TextureState constructor/destructor. It is heavier than it looks. Also fixed some includes in Implementation subdir.
13 years ago
/*
This file is part of Magnum.
Updated copyright year.
6 years ago
Copyright © 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019,
Updated copyright year.
4 years ago
2020, 2021, 2022 Vladimír Vondruš <mosra@centrum.cz>
De-inline TextureState constructor/destructor. It is heavier than it looks. Also fixed some includes in Implementation subdir.
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.
*/
#include "TextureState.h"
Adapted to Corrade changes. Mostly missing includes now.
7 years ago
#include <tuple>
GL: rework Context to not use any STL containers. We're going to eventually include this class in all Application classes (need that in order to inherit a to-be-created Configuration class) and the <string> and <vector> would be just too much. This change caused magnum-gl-info.wasm (WebGL 2 build) to go down from 247 to 245 kB. Not much, but that's I guess because there's still a lot other vectors of strings elsewhere. There's a lot more places to clean up, will do those in separate commits. This change is the most atomic I could do, and it introduces a breaking change to all APIs that returned a std::vector or a std::string. Fortunately (or as I hope) those weren't used that much, so it shouldn't cause build breakages for that many people. Quite a lot of the optimization ideas is borrowed from the new Vk library -- such as "interning" the driver workaround strings to avoid allocating their copies.
5 years ago
#include <Corrade/Containers/StringView.h>
Missing include. Oh well.
12 years ago
#include <Corrade/Utility/Assert.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/AbstractTexture.h"
#include "Magnum/GL/CubeMapTexture.h"
Implemented EXT_texture_buffer AEP extension.
11 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
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/BufferTexture.h"
Moved global data from texture classes to per-context state.
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
#include "Magnum/GL/Context.h"
#include "Magnum/GL/Extensions.h"
Moved global data from texture classes to per-context state.
12 years ago
Ability to reset internal state tracker.
12 years ago
#include "State.h"
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 { namespace Implementation {
De-inline TextureState constructor/destructor. It is heavier than it looks. Also fixed some includes in Implementation subdir.
13 years ago
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
using namespace Containers::Literals;
GL: put texture/image state trackers into the single allocation also.
5 years ago
TextureState::TextureState(Context& context,
Containers::ArrayView<std::pair<GLenum, GLuint>> bindings,
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
Containers::ArrayView<std::tuple<GLuint, GLint, GLboolean, GLint, GLenum>> imageBindings,
#endif
Containers::StaticArrayView<Implementation::ExtensionCount, const char*> extensions):
maxSize{},
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !(defined(MAGNUM_TARGET_WEBGL) && defined(MAGNUM_TARGET_GLES2))
max3DSize{},
#endif
maxCubeMapSize{},
Hide texture state members which are not used on ES. Shaved eight bytes off total memory footprint, yay!
12 years ago
#ifndef MAGNUM_TARGET_GLES2
maxArrayLayers{},
#endif
Properly initialize variables for texture limits in state tracker. And I was wondering how it is possible that there new NVidia drivers were allowing up to 32 PB textures on my humble laptop graphics :)
12 years ago
#ifndef MAGNUM_TARGET_GLES
maxRectangleSize{}, maxBufferSize{},
#endif
GL: put texture/image state trackers into the single allocation also.
5 years ago
/* This value got queried in State already to allocate the bindings array,
reuse it here */
maxTextureUnits{GLint(bindings.size())},
Max texture LOD bias _is_ available in OpenGL ES 3.
12 years ago
#ifndef MAGNUM_TARGET_GLES2
Hide texture state members which are not used on ES. Shaved eight bytes off total memory footprint, yay!
12 years ago
maxLodBias{0.0f},
#endif
GL: put texture/image state trackers into the single allocation also.
5 years ago
maxMaxAnisotropy(0.0f), currentTextureUnit(0),
Enabled multisample texture limit queries in ES 3.1.
12 years ago
#ifndef MAGNUM_TARGET_GLES2
GL: put texture/image state trackers into the single allocation also.
5 years ago
maxColorSamples(0), maxDepthSamples(0), maxIntegerSamples(0),
Enabled multisample texture limit queries in ES 3.1.
12 years ago
#endif
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
#ifndef MAGNUM_TARGET_GLES
GL: put texture/image state trackers into the single allocation also.
5 years ago
bufferOffsetAlignment(0),
#endif
bindings{bindings}
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
, imageBindings{imageBindings}
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
#endif
Moved global data from texture classes to per-context state.
12 years ago
{
ARB_direct_state_access support for creating textures.
12 years ago
/* Create implementation */
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::ARB::direct_state_access::Index] =
Extensions::ARB::direct_state_access::string();
ARB_direct_state_access support for creating textures.
12 years ago
createImplementation = &AbstractTexture::createImplementationDSA;
} else
#endif
{
createImplementation = &AbstractTexture::createImplementationDefault;
}
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
/* Single bind implementation */
Initial ARB_multi_bind usage in *Texture::bind().
12 years ago
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
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
/* Extension name added below */
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
if((context.detectedDriver() & Context::DetectedDriver::IntelWindows) &&
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
!context.isDriverWorkaroundDisabled("intel-windows-half-baked-dsa-texture-bind"_s))
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
{
unbindImplementation = &AbstractTexture::unbindImplementationDefault;
bindImplementation = &AbstractTexture::bindImplementationDSAIntelWindows;
} else
#endif
{
unbindImplementation = &AbstractTexture::unbindImplementationDSA;
bindImplementation = &AbstractTexture::bindImplementationDSA;
}
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
} else if(context.isExtensionSupported<Extensions::ARB::multi_bind>()) {
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
/* Extension name added below */
Added *Texture::unbind(). Will be used in fallback implementation of multi-bind.
12 years ago
unbindImplementation = &AbstractTexture::unbindImplementationMulti;
Initial ARB_multi_bind usage in *Texture::bind().
12 years ago
bindImplementation = &AbstractTexture::bindImplementationMulti;
Added *Texture::unbind(). Will be used in fallback implementation of multi-bind.
12 years ago
Initial ARB_multi_bind usage in *Texture::bind().
12 years ago
} else
#endif
{
Added *Texture::unbind(). Will be used in fallback implementation of multi-bind.
12 years ago
unbindImplementation = &AbstractTexture::unbindImplementationDefault;
GL: one more iteration on the Apple buffer texture workaround. This broke DSA-less contexts with ARB_multi_bind present.
6 years ago
/* This is additionally modified below for the
apple-buffer-texture-unbind-on-buffer-modify workaround */
Initial ARB_multi_bind usage in *Texture::bind().
12 years ago
bindImplementation = &AbstractTexture::bindImplementationDefault;
}
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
/* Multi bind implementation */
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::multi_bind>()) {
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::ARB::multi_bind::Index] =
Extensions::ARB::multi_bind::string();
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
bindMultiImplementation = &AbstractTexture::bindImplementationMulti;
} else
#endif
{
bindMultiImplementation = &AbstractTexture::bindImplementationFallback;
}
Moved global data from texture classes to per-context state.
12 years ago
/* DSA/non-DSA implementation */
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::ARB::direct_state_access::Index] =
Extensions::ARB::direct_state_access::string();
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
parameteriImplementation = &AbstractTexture::parameterImplementationDSA;
parameterfImplementation = &AbstractTexture::parameterImplementationDSA;
parameterivImplementation = &AbstractTexture::parameterImplementationDSA;
parameterfvImplementation = &AbstractTexture::parameterImplementationDSA;
parameterIuivImplementation = &AbstractTexture::parameterIImplementationDSA;
parameterIivImplementation = &AbstractTexture::parameterIImplementationDSA;
getLevelParameterivImplementation = &AbstractTexture::getLevelParameterImplementationDSA;
mipmapImplementation = &AbstractTexture::mipmapImplementationDSA;
subImage1DImplementation = &AbstractTexture::subImageImplementationDSA;
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
subImage2DImplementation = &AbstractTexture::subImage2DImplementationDSA;
subImage3DImplementation = &AbstractTexture::subImage3DImplementationDSA;
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
compressedSubImage1DImplementation = &AbstractTexture::compressedSubImageImplementationDSA;
compressedSubImage2DImplementation = &AbstractTexture::compressedSubImageImplementationDSA;
compressedSubImage3DImplementation = &AbstractTexture::compressedSubImageImplementationDSA;
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
setBufferImplementation = &BufferTexture::setBufferImplementationDSA;
setBufferRangeImplementation = &BufferTexture::setBufferRangeImplementationDSA;
Moved global data from texture classes to per-context state.
12 years ago
} else
#endif
{
parameteriImplementation = &AbstractTexture::parameterImplementationDefault;
parameterfImplementation = &AbstractTexture::parameterImplementationDefault;
Implemented ARB_texture_swizzle.
12 years ago
#ifndef MAGNUM_TARGET_GLES2
parameterivImplementation = &AbstractTexture::parameterImplementationDefault;
#endif
Moved global data from texture classes to per-context state.
12 years ago
parameterfvImplementation = &AbstractTexture::parameterImplementationDefault;
Updates for texture border clamp support in ES 3.2.
9 years ago
#ifndef MAGNUM_TARGET_GLES
Renamed internal function to avoid name clashes.
12 years ago
parameterIuivImplementation = &AbstractTexture::parameterIImplementationDefault;
parameterIivImplementation = &AbstractTexture::parameterIImplementationDefault;
Enabled texture image size queries 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)
Moved global data from texture classes to per-context state.
12 years ago
getLevelParameterivImplementation = &AbstractTexture::getLevelParameterImplementationDefault;
#endif
mipmapImplementation = &AbstractTexture::mipmapImplementationDefault;
#ifndef MAGNUM_TARGET_GLES
subImage1DImplementation = &AbstractTexture::subImageImplementationDefault;
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
compressedSubImage1DImplementation = &AbstractTexture::compressedSubImageImplementationDefault;
Moved global data from texture classes to per-context state.
12 years ago
#endif
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
subImage2DImplementation = &AbstractTexture::subImage2DImplementationDefault;
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
compressedSubImage2DImplementation = &AbstractTexture::compressedSubImageImplementationDefault;
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#if !(defined(MAGNUM_TARGET_WEBGL) && defined(MAGNUM_TARGET_GLES2))
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
subImage3DImplementation = &AbstractTexture::subImage3DImplementationDefault;
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
compressedSubImage3DImplementation = &AbstractTexture::compressedSubImageImplementationDefault;
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#endif
Moved global data from texture classes to per-context state.
12 years ago
Updates for buffer texture support in ES 3.2. Turns out Buffer::TargetHint::Texture was not enabled for AEP. Fixed that now.
9 years ago
#ifndef MAGNUM_TARGET_GLES
Moved global data from texture classes to per-context state.
12 years ago
setBufferImplementation = &BufferTexture::setBufferImplementationDefault;
setBufferRangeImplementation = &BufferTexture::setBufferRangeImplementationDefault;
#endif
GL: new "intel-windows-broken-dsa-for-cubemaps" workaround. Fixes everything related to cube maps when ARB_DSA is used.
7 years ago
}
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
/* DSA/non-DSA implementation for cubemaps, because Intel (and AMD) Windows
drivers have to be broken in a special way */
GL: new "intel-windows-broken-dsa-for-cubemaps" workaround. Fixes everything related to cube maps when ARB_DSA is used.
7 years ago
#ifndef MAGNUM_TARGET_GLES
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
GL: new "intel-windows-broken-dsa-for-cubemaps" workaround. Fixes everything related to cube maps when ARB_DSA is used.
7 years ago
#ifdef CORRADE_TARGET_WINDOWS
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
if((context.detectedDriver() & Context::DetectedDriver::IntelWindows) && !context.isDriverWorkaroundDisabled("intel-windows-broken-dsa-for-cubemaps"_s)) {
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
getCubeLevelParameterivImplementation = &CubeMapTexture::getLevelParameterImplementationDefault;
cubeSubImageImplementation = &CubeMapTexture::subImageImplementationDefault;
cubeCompressedSubImageImplementation = &CubeMapTexture::compressedSubImageImplementationDefault;
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
} else if((context.detectedDriver() & Context::DetectedDriver::Amd) && !context.isDriverWorkaroundDisabled("amd-windows-cubemap-image3d-slice-by-slice"_s)) {
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
/* This one is not broken, but the others are */
getCubeLevelParameterivImplementation = &CubeMapTexture::getLevelParameterImplementationDSA;
cubeSubImageImplementation = &CubeMapTexture::subImageImplementationDefault;
cubeCompressedSubImageImplementation = &CubeMapTexture::compressedSubImageImplementationDefault;
} else
GL: new "intel-windows-broken-dsa-for-cubemaps" workaround. Fixes everything related to cube maps when ARB_DSA is used.
7 years ago
#endif
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
{
/* Extension name added above */
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: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
getCubeLevelParameterivImplementation = &CubeMapTexture::getLevelParameterImplementationDSA;
cubeSubImageImplementation = &CubeMapTexture::subImageImplementationDSA;
cubeCompressedSubImageImplementation = &CubeMapTexture::compressedSubImageImplementationDSA;
}
GL: new "intel-windows-broken-dsa-for-cubemaps" workaround. Fixes everything related to cube maps when ARB_DSA is used.
7 years ago
} else
#endif
{
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
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
getCubeLevelParameterivImplementation = &CubeMapTexture::getLevelParameterImplementationDefault;
#endif
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
cubeSubImageImplementation = &CubeMapTexture::subImageImplementationDefault;
Compressed image support, part 7: compressed (sub)image upload. No tests yet.
11 years ago
cubeCompressedSubImageImplementation = &CubeMapTexture::compressedSubImageImplementationDefault;
Moved global data from texture classes to per-context state.
12 years ago
}
Updates for texture border clamp support in ES 3.2.
9 years ago
#if defined(MAGNUM_TARGET_GLES) && !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
Updates for buffer texture support in ES 3.2. Turns out Buffer::TargetHint::Texture was not enabled for AEP. Fixed that now.
9 years ago
/* Integer parameter implementation for ES3 */
Updates for texture border clamp support in ES 3.2.
9 years ago
if(context.isVersionSupported(Version::GLES320)) {
parameterIuivImplementation = &AbstractTexture::parameterIImplementationDefault;
parameterIivImplementation = &AbstractTexture::parameterIImplementationDefault;
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
} else if(context.isExtensionSupported<Extensions::EXT::texture_border_clamp>()) {
Updates for texture border clamp support in ES 3.2.
9 years ago
parameterIuivImplementation = &AbstractTexture::parameterIImplementationEXT;
parameterIivImplementation = &AbstractTexture::parameterIImplementationEXT;
} else {
parameterIuivImplementation = nullptr;
parameterIivImplementation = nullptr;
}
Updates for buffer texture support in ES 3.2. Turns out Buffer::TargetHint::Texture was not enabled for AEP. Fixed that now.
9 years ago
/* Buffer texture implementation for ES3 */
if(context.isVersionSupported(Version::GLES320)) {
setBufferImplementation = &BufferTexture::setBufferImplementationDefault;
setBufferRangeImplementation = &BufferTexture::setBufferRangeImplementationDefault;
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
} else if(context.isExtensionSupported<Extensions::EXT::texture_buffer>()) {
Updates for buffer texture support in ES 3.2. Turns out Buffer::TargetHint::Texture was not enabled for AEP. Fixed that now.
9 years ago
setBufferImplementation = &BufferTexture::setBufferImplementationEXT;
setBufferRangeImplementation = &BufferTexture::setBufferRangeImplementationEXT;
} else {
setBufferImplementation = nullptr;
setBufferRangeImplementation = nullptr;
}
Updates for texture border clamp support in ES 3.2.
9 years ago
#endif
Moved global data from texture classes to per-context state.
12 years ago
/* Data invalidation implementation */
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::invalidate_subdata>()) {
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::ARB::invalidate_subdata::Index] =
Extensions::ARB::invalidate_subdata::string();
Moved global data from texture classes to per-context state.
12 years ago
invalidateImageImplementation = &AbstractTexture::invalidateImageImplementationARB;
invalidateSubImageImplementation = &AbstractTexture::invalidateSubImageImplementationARB;
} else
#endif
{
invalidateImageImplementation = &AbstractTexture::invalidateImageImplementationNoOp;
invalidateSubImageImplementation = &AbstractTexture::invalidateSubImageImplementationNoOp;
}
#ifndef MAGNUM_TARGET_GLES
Add workaround for inconsistent cubemap compressed image size on NVidia. The affected test case now passes again.
10 years ago
/* Compressed cubemap image size query implementation (extensions added
above) */
if((context.detectedDriver() & Context::DetectedDriver::NVidia) &&
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
!context.isDriverWorkaroundDisabled("nv-cubemap-inconsistent-compressed-image-size"_s)) {
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>())
Add workaround for inconsistent cubemap compressed image size on NVidia. The affected test case now passes again.
10 years ago
getCubeLevelCompressedImageSizeImplementation = &CubeMapTexture::getLevelCompressedImageSizeImplementationDSANonImmutableWorkaround;
else getCubeLevelCompressedImageSizeImplementation = &CubeMapTexture::getLevelCompressedImageSizeImplementationDefaultImmutableWorkaround;
GL: new "intel-windows-broken-dsa-for-cubemaps" workaround. Fixes everything related to cube maps when ARB_DSA is used.
7 years ago
} else if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()
#ifdef CORRADE_TARGET_WINDOWS
&& (!(context.detectedDriver() & Context::DetectedDriver::IntelWindows) ||
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
context.isDriverWorkaroundDisabled("intel-windows-broken-dsa-for-cubemaps"_s))
GL: new "intel-windows-broken-dsa-for-cubemaps" workaround. Fixes everything related to cube maps when ARB_DSA is used.
7 years ago
#endif
) {
getCubeLevelCompressedImageSizeImplementation = &CubeMapTexture::getLevelCompressedImageSizeImplementationDSA;
Add workaround for inconsistent cubemap compressed image size on NVidia. The affected test case now passes again.
10 years ago
} else {
GL: new "intel-windows-broken-dsa-for-cubemaps" workaround. Fixes everything related to cube maps when ARB_DSA is used.
7 years ago
getCubeLevelCompressedImageSizeImplementation = &CubeMapTexture::getLevelCompressedImageSizeImplementationDefault;
Add workaround for inconsistent cubemap compressed image size on NVidia. The affected test case now passes again.
10 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
/* Image retrieval implementation */
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
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
/* Extension name added above */
getImageImplementation = &AbstractTexture::getImageImplementationDSA;
Compressed image support, part 8: compressed image download. Only full images now, no tests yet.
11 years ago
getCompressedImageImplementation = &AbstractTexture::getCompressedImageImplementationDSA;
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
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
} else if(context.isExtensionSupported<Extensions::ARB::robustness>()) {
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::ARB::robustness::Index] =
Extensions::ARB::robustness::string();
Moved global data from texture classes to per-context state.
12 years ago
getImageImplementation = &AbstractTexture::getImageImplementationRobustness;
Compressed image support, part 8: compressed image download. Only full images now, no tests yet.
11 years ago
getCompressedImageImplementation = &AbstractTexture::getCompressedImageImplementationRobustness;
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
Compressed image support, part 8: compressed image download. Only full images now, no tests yet.
11 years ago
} else {
getImageImplementation = &AbstractTexture::getImageImplementationDefault;
getCompressedImageImplementation = &AbstractTexture::getCompressedImageImplementationDefault;
}
Clarified and fixed cube map texture image retrieval methods. Because ARB_DSA doesn't have any way to extract image of single cube map coordinate, we have to use ARB_get_texture_sub_image instead, thus for cube maps the whole thing is different. That was implemented, but wasn't mentioned in the docs and wasn't properly accounted for in implementation switcher (I was under assumption that ARB_DSA is equivalent to ARB_get_texture_sub_image, which is not).
11 years ago
/* Image retrieval implementation for cube map */
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::get_texture_sub_image>()) {
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::ARB::get_texture_sub_image::Index] =
Extensions::ARB::get_texture_sub_image::string();
Clarified and fixed cube map texture image retrieval methods. Because ARB_DSA doesn't have any way to extract image of single cube map coordinate, we have to use ARB_get_texture_sub_image instead, thus for cube maps the whole thing is different. That was implemented, but wasn't mentioned in the docs and wasn't properly accounted for in implementation switcher (I was under assumption that ARB_DSA is equivalent to ARB_get_texture_sub_image, which is not).
11 years ago
getCubeImageImplementation = &CubeMapTexture::getImageImplementationDSA;
Compressed image support, part 8: compressed image download. Only full images now, no tests yet.
11 years ago
getCompressedCubeImageImplementation = &CubeMapTexture::getCompressedImageImplementationDSA;
Clarified and fixed cube map texture image retrieval methods. Because ARB_DSA doesn't have any way to extract image of single cube map coordinate, we have to use ARB_get_texture_sub_image instead, thus for cube maps the whole thing is different. That was implemented, but wasn't mentioned in the docs and wasn't properly accounted for in implementation switcher (I was under assumption that ARB_DSA is equivalent to ARB_get_texture_sub_image, which is not).
11 years ago
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
} else if(context.isExtensionSupported<Extensions::ARB::robustness>()) {
Clarified and fixed cube map texture image retrieval methods. Because ARB_DSA doesn't have any way to extract image of single cube map coordinate, we have to use ARB_get_texture_sub_image instead, thus for cube maps the whole thing is different. That was implemented, but wasn't mentioned in the docs and wasn't properly accounted for in implementation switcher (I was under assumption that ARB_DSA is equivalent to ARB_get_texture_sub_image, which is not).
11 years ago
/* Extension name added above */
getCubeImageImplementation = &CubeMapTexture::getImageImplementationRobustness;
Compressed image support, part 8: compressed image download. Only full images now, no tests yet.
11 years ago
getCompressedCubeImageImplementation = &CubeMapTexture::getCompressedImageImplementationRobustness;
Clarified and fixed cube map texture image retrieval methods. Because ARB_DSA doesn't have any way to extract image of single cube map coordinate, we have to use ARB_get_texture_sub_image instead, thus for cube maps the whole thing is different. That was implemented, but wasn't mentioned in the docs and wasn't properly accounted for in implementation switcher (I was under assumption that ARB_DSA is equivalent to ARB_get_texture_sub_image, which is not).
11 years ago
Compressed image support, part 8: compressed image download. Only full images now, no tests yet.
11 years ago
} else {
getCubeImageImplementation = &CubeMapTexture::getImageImplementationDefault;
getCompressedCubeImageImplementation = &CubeMapTexture::getCompressedImageImplementationDefault;
}
Workaround broken full cube map compressed image query on NVidia. The query returns only data of the first cube map face and not the others, worked around that by querying each face separately.
10 years ago
GL: rename & reorganize cubemap internals and test cases for clarity. The 3D subimage setters/queries will get enabled on all platforms now, and naming them "full" just doesn't make sense.
5 years ago
/* 3D compressed cubemap image query implementation (extensions added
Workaround broken full cube map compressed image query on NVidia. The query returns only data of the first cube map face and not the others, worked around that by querying each face separately.
10 years ago
above) */
if((context.detectedDriver() & Context::DetectedDriver::NVidia) &&
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
context.isExtensionSupported<Extensions::ARB::direct_state_access>() &&
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
!context.isDriverWorkaroundDisabled("nv-cubemap-broken-full-compressed-image-query"_s))
GL: rename & reorganize cubemap internals and test cases for clarity. The 3D subimage setters/queries will get enabled on all platforms now, and naming them "full" just doesn't make sense.
5 years ago
getCompressedCubeImage3DImplementation = &CubeMapTexture::getCompressedImageImplementationDSASingleSliceWorkaround;
Workaround broken full cube map compressed image query on NVidia. The query returns only data of the first cube map face and not the others, worked around that by querying each face separately.
10 years ago
else
GL: rename & reorganize cubemap internals and test cases for clarity. The 3D subimage setters/queries will get enabled on all platforms now, and naming them "full" just doesn't make sense.
5 years ago
getCompressedCubeImage3DImplementation = &CubeMapTexture::getCompressedImageImplementationDSA;
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
GL: don't check for Windows-only workarounds everywhere. Apparently nobody tried to run AMDGPU PRO drivers on Linux until now. Sorry.
5 years ago
#ifdef CORRADE_TARGET_WINDOWS
/** @todo those *might* be happening with the proprietary AMD driver on
linux as well, test */
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
if((context.detectedDriver() & Context::DetectedDriver::Amd) &&
context.isExtensionSupported<Extensions::ARB::direct_state_access>() &&
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
!context.isDriverWorkaroundDisabled("amd-windows-cubemap-image3d-slice-by-slice"_s))
GL: rename & reorganize cubemap internals and test cases for clarity. The 3D subimage setters/queries will get enabled on all platforms now, and naming them "full" just doesn't make sense.
5 years ago
getCubeImage3DImplementation = &CubeMapTexture::getImageImplementationDSAAmdSliceBySlice;
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
else if((context.detectedDriver() & Context::DetectedDriver::IntelWindows) &&
context.isExtensionSupported<Extensions::ARB::direct_state_access>() &&
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
!context.isDriverWorkaroundDisabled("intel-windows-broken-dsa-for-cubemaps"_s))
GL: rename & reorganize cubemap internals and test cases for clarity. The 3D subimage setters/queries will get enabled on all platforms now, and naming them "full" just doesn't make sense.
5 years ago
getCubeImage3DImplementation = &CubeMapTexture::getImageImplementationSliceBySlice;
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
else
GL: don't check for Windows-only workarounds everywhere. Apparently nobody tried to run AMDGPU PRO drivers on Linux until now. Sorry.
5 years ago
#endif
GL: enable cube map 3D image up/download on all platforms. There were slice-by-slice workarounds for certain drivers already, so why not reuse them to make this functionality available everywhere.
5 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
getCubeImage3DImplementation = &CubeMapTexture::getImageImplementationDSA;
} else {
GL: rename & reorganize cubemap internals and test cases for clarity. The 3D subimage setters/queries will get enabled on all platforms now, and naming them "full" just doesn't make sense.
5 years ago
getCubeImage3DImplementation = &CubeMapTexture::getImageImplementationSliceBySlice;
GL: don't check for Windows-only workarounds everywhere. Apparently nobody tried to run AMDGPU PRO drivers on Linux until now. Sorry.
5 years ago
}
Moved global data from texture classes to per-context state.
12 years ago
#endif
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
/* Texture storage implementation for desktop and ES */
#ifndef MAGNUM_TARGET_WEBGL
Moved global data from texture classes to per-context state.
12 years ago
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::texture_storage>())
Actually make use of EXT_texture_storage on ES2. Currently it was ignored when deciding about *Texture::setStorage() implementation.
12 years ago
#elif defined(MAGNUM_TARGET_GLES2)
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::EXT::texture_storage>())
Moved global data from texture classes to per-context state.
12 years ago
#endif
{
#ifndef MAGNUM_TARGET_GLES
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::ARB::texture_storage::Index] =
Extensions::ARB::texture_storage::string();
Actually make use of EXT_texture_storage on ES2. Currently it was ignored when deciding about *Texture::setStorage() implementation.
12 years ago
#elif defined(MAGNUM_TARGET_GLES2)
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::EXT::texture_storage::Index] =
Extensions::EXT::texture_storage::string();
Moved global data from texture classes to per-context state.
12 years ago
#endif
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
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
storage1DImplementation = &AbstractTexture::storageImplementationDSA;
storage2DImplementation = &AbstractTexture::storageImplementationDSA;
storage3DImplementation = &AbstractTexture::storageImplementationDSA;
Moved global data from texture classes to per-context state.
12 years ago
} else
#endif
{
#ifndef MAGNUM_TARGET_GLES
storage1DImplementation = &AbstractTexture::storageImplementationDefault;
#endif
storage2DImplementation = &AbstractTexture::storageImplementationDefault;
storage3DImplementation = &AbstractTexture::storageImplementationDefault;
}
}
Don't use MAGNUM_TARGET_GLES3 where the condition will pass also for ES4.
11 years ago
#if !defined(MAGNUM_TARGET_GLES) || defined(MAGNUM_TARGET_GLES2)
Actually make use of EXT_texture_storage on ES2. Currently it was ignored when deciding about *Texture::setStorage() implementation.
12 years ago
else {
Moved global data from texture classes to per-context state.
12 years ago
#ifndef MAGNUM_TARGET_GLES
storage1DImplementation = &AbstractTexture::storageImplementationFallback;
#endif
storage2DImplementation = &AbstractTexture::storageImplementationFallback;
storage3DImplementation = &AbstractTexture::storageImplementationFallback;
}
#endif
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
/* Texture storage implementation for WebGL 1.0 */
#elif defined(MAGNUM_TARGET_GLES2)
storage2DImplementation = &AbstractTexture::storageImplementationFallback;
/* Texture storage implementation for WebGL 2.0 */
#else
storage2DImplementation = &AbstractTexture::storageImplementationDefault;
storage3DImplementation = &AbstractTexture::storageImplementationDefault;
#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
/* Storage implementation for multisample textures. The fallback doesn't
have DSA alternative, so it must be handled specially. */
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::texture_storage_multisample>()) {
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::ARB::texture_storage_multisample::Index] =
Extensions::ARB::texture_storage_multisample::string();
Added MultisampleTexture::setStorage(). With this function multisample texture implementation is pretty much done. Also implemented all the tests.
12 years ago
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
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
storage2DMultisampleImplementation = &AbstractTexture::storageMultisampleImplementationDSA;
storage3DMultisampleImplementation = &AbstractTexture::storageMultisampleImplementationDSA;
Added MultisampleTexture::setStorage(). With this function multisample texture implementation is pretty much done. Also implemented all the tests.
12 years ago
} else {
storage2DMultisampleImplementation = &AbstractTexture::storageMultisampleImplementationDefault;
storage3DMultisampleImplementation = &AbstractTexture::storageMultisampleImplementationDefault;
}
} else {
storage2DMultisampleImplementation = &AbstractTexture::storageMultisampleImplementationFallback;
storage3DMultisampleImplementation = &AbstractTexture::storageMultisampleImplementationFallback;
}
First-class WebGL support, part 4: reduced texture functionality.
11 years ago
#elif !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
Enabled 2D multisample textures in ES 3.1.
12 years ago
storage2DMultisampleImplementation = &AbstractTexture::storageMultisampleImplementationDefault;
Updates for multisample array texture storage in ES 3.2.
9 years ago
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
if(context.isVersionSupported(Version::GLES320)) {
Updates for multisample array texture storage in ES 3.2.
9 years ago
storage3DMultisampleImplementation = &AbstractTexture::storageMultisampleImplementationDefault;
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
} else if(context.isExtensionSupported<Extensions::OES::texture_storage_multisample_2d_array>()) {
extensions[Extensions::OES::texture_storage_multisample_2d_array::Index] =
Extensions::OES::texture_storage_multisample_2d_array::string();
Updates for multisample array texture storage in ES 3.2.
9 years ago
storage3DMultisampleImplementation = &AbstractTexture::storageMultisampleImplementationOES;
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
} else {
Updates for multisample array texture storage in ES 3.2.
9 years ago
storage3DMultisampleImplementation = nullptr;
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
}
Added MultisampleTexture::setStorage(). With this function multisample texture implementation is pretty much done. Also implemented all the tests.
12 years ago
#endif
Moved global data from texture classes to per-context state.
12 years ago
/* Anisotropic filter implementation */
Support for ARB_texture_filter_anisotropic.
8 years ago
#ifndef MAGNUM_TARGET_GLES
GL: whoops. Broken since 0ceb54ed7df24ec1a79727397d3afe5cfd94142e, but the two code paths actually differ only by an enum name so it didn't cause any crashes. (I wonder why I need two different code paths at all.)
5 years ago
if(context.isExtensionSupported<Extensions::ARB::texture_filter_anisotropic>()) {
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::ARB::texture_filter_anisotropic::Index] =
Extensions::ARB::texture_filter_anisotropic::string();
Support for ARB_texture_filter_anisotropic.
8 years ago
GL: we don't need two code paths that do effectively the same. The enum value is exactly the same for both.
5 years ago
setMaxAnisotropyImplementation = &AbstractTexture::setMaxAnisotropyImplementationArbOrExt;
Support for ARB_texture_filter_anisotropic.
8 years ago
} else
#endif
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::EXT::texture_filter_anisotropic>()) {
GL: avoid populating and sorting (!!) a vector of strings for exts. Ew. At first I tried to just port a growable Array of StringViews (which would already save quite a lot), but then I realized I have a clear upper bound on the extensions and so can use a "counting sort" without having to deduplicate anything after. After the previous (rather minimal) reduction by the Context cleanup, this reduces the size of magnum-gl-info.wasm from 245 to 237 kB. Quite significant, I'd say!
5 years ago
extensions[Extensions::EXT::texture_filter_anisotropic::Index] =
Extensions::EXT::texture_filter_anisotropic::string();
Moved global data from texture classes to per-context state.
12 years ago
GL: we don't need two code paths that do effectively the same. The enum value is exactly the same for both.
5 years ago
setMaxAnisotropyImplementation = &AbstractTexture::setMaxAnisotropyImplementationArbOrExt;
Moved global data from texture classes to per-context state.
12 years ago
} else setMaxAnisotropyImplementation = &AbstractTexture::setMaxAnisotropyImplementationNoOp;
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
/* NVidia workaround for compressed block data size implementation */
if((context.detectedDriver() & Context::DetectedDriver::NVidia) &&
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
!context.isDriverWorkaroundDisabled("nv-compressed-block-size-in-bits"_s))
Added workaround for NVidia reporting compressed block size in bits. Instead of bytes as per specification of glGetInternalformativ().
10 years ago
compressedBlockDataSizeImplementation = &AbstractTexture::compressedBlockDataSizeImplementationBitsWorkaround;
else
compressedBlockDataSizeImplementation = &AbstractTexture::compressedBlockDataSizeImplementationDefault;
#endif
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
/* SVGA3D workaround for array / 3D / cube map texture upload. Overrides
the DSA / non-DSA function pointers set above. */
if((context.detectedDriver() & Context::DetectedDriver::Svga3D) &&
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
!context.isDriverWorkaroundDisabled("svga3d-texture-upload-slice-by-slice"_s)) {
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
image2DImplementation = &AbstractTexture::imageImplementationSvga3DSliceBySlice;
#endif
image3DImplementation = &AbstractTexture::imageImplementationSvga3DSliceBySlice;
#ifndef MAGNUM_TARGET_GLES
Split the OpenGL layer out, pt 16: {Extensions::GL => GL::Extensions}. Was Magnum::GL::Extensions::GL before and the redundancy was completely unnecessary. Potential future extensions coming from GLX, EGL or whatnot will most probably be in the Platform namespace in a completely separate file, so this is not a problem. All code internal to the GL library is affected, not much the outside, as that is handled by the compatibility alias.
8 years ago
if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
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
subImage2DImplementation = &AbstractTexture::subImageImplementationSvga3DSliceBySlice<&AbstractTexture::subImage2DImplementationDSA>;
#endif
subImage3DImplementation = &AbstractTexture::subImageImplementationSvga3DSliceBySlice<&AbstractTexture::subImage3DImplementationDSA>;
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
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
} else
#endif
{
#ifndef MAGNUM_TARGET_GLES
subImage2DImplementation = &AbstractTexture::subImageImplementationSvga3DSliceBySlice<&AbstractTexture::subImage2DImplementationDefault>;
#endif
subImage3DImplementation = &AbstractTexture::subImageImplementationSvga3DSliceBySlice<&AbstractTexture::subImage3DImplementationDefault>;
}
} else
#endif
{
/* These need to be set up in any case */
image2DImplementation = &AbstractTexture::imageImplementationDefault;
#if !(defined(MAGNUM_TARGET_WEBGL) && defined(MAGNUM_TARGET_GLES2))
image3DImplementation = &AbstractTexture::imageImplementationDefault;
#endif
}
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
#ifndef MAGNUM_TARGET_GLES
/* SVGA3D and Intel workaround for cube map texture upload. Overrides the
DSA / non-DSA function pointers set above. */
if((context.detectedDriver() & Context::DetectedDriver::Svga3D) &&
GL: enable cube map 3D image up/download on all platforms. There were slice-by-slice workarounds for certain drivers already, so why not reuse them to make this functionality available everywhere.
5 years ago
context.isExtensionSupported<Extensions::ARB::direct_state_access>() &&
!context.isDriverWorkaroundDisabled("svga3d-texture-upload-slice-by-slice"_s)
) {
cubeSubImage3DImplementation = &CubeMapTexture::subImageImplementationDSASliceBySlice;
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
} else if((context.detectedDriver() & Context::DetectedDriver::IntelWindows) &&
GL: enable cube map 3D image up/download on all platforms. There were slice-by-slice workarounds for certain drivers already, so why not reuse them to make this functionality available everywhere.
5 years ago
context.isExtensionSupported<Extensions::ARB::direct_state_access>() &&
!context.isDriverWorkaroundDisabled("intel-windows-broken-dsa-for-cubemaps"_s)
) {
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
cubeSubImage3DImplementation = &CubeMapTexture::subImageImplementationSliceBySlice;
GL: don't check for Windows-only workarounds everywhere. Apparently nobody tried to run AMDGPU PRO drivers on Linux until now. Sorry.
5 years ago
}
#ifdef CORRADE_TARGET_WINDOWS
/** @todo those *might* be happening with the proprietary AMD driver on
linux as well, test */
else if((context.detectedDriver() & Context::DetectedDriver::Amd) &&
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
!context.isDriverWorkaroundDisabled("amd-windows-cubemap-image3d-slice-by-slice"_s)) {
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
/* DSA version is broken (non-zero Z offset not allowed), need to
emulate using classic APIs */
cubeSubImage3DImplementation = &CubeMapTexture::subImageImplementationSliceBySlice;
GL: don't check for Windows-only workarounds everywhere. Apparently nobody tried to run AMDGPU PRO drivers on Linux until now. Sorry.
5 years ago
}
#endif
else if(context.isExtensionSupported<Extensions::ARB::direct_state_access>()) {
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
cubeSubImage3DImplementation = &CubeMapTexture::subImageImplementationDSA;
} else
GL: enable cube map 3D image up/download on all platforms. There were slice-by-slice workarounds for certain drivers already, so why not reuse them to make this functionality available everywhere.
5 years ago
#endif
GL: pile on more workarounds for CubeMapTexture DSA code paths. Ugh why this has to be SO TERRIBLE. I'm starting to have extremely low expectations for similar parts of Vulkan drivers by the same vendors.
7 years ago
{
cubeSubImage3DImplementation = &CubeMapTexture::subImageImplementationSliceBySlice;
}
GL: fix compilation on macOS + ANGLE.
6 years ago
#if defined(CORRADE_TARGET_APPLE) && !defined(MAGNUM_TARGET_GLES)
GL: use StringView literals for all workaround names. These are in most cases the only strings that are used, and I don't think having to call std::strlen() for each of them is a good idea if we don't need to.
5 years ago
if(!context.isDriverWorkaroundDisabled("apple-buffer-texture-unbind-on-buffer-modify"_s)) {
GL: finalize the apple-buffer-texture-unbind-on-buffer-modify workaround. Followup to 24cc971b1f43c61aa157228f6ea9916007ffb3cd, covering the remaining case.
6 years ago
CORRADE_INTERNAL_ASSERT(std::size_t(maxTextureUnits) <= decltype(bufferTextureBound)::Size);
/* Assume ARB_multi_bind is not supported, otherwise we'd need to
implement the workaround also for bindMultiImplementation */
CORRADE_INTERNAL_ASSERT(!context.isExtensionSupported<Extensions::ARB::multi_bind>());
bindImplementation = &AbstractTexture::bindImplementationAppleBufferTextureWorkaround;
GL: one more iteration on the Apple buffer texture workaround. This broke DSA-less contexts with ARB_multi_bind present.
6 years ago
bindInternalImplementation = &AbstractTexture::bindImplementationAppleBufferTextureWorkaround;
} else
GL: finalize the apple-buffer-texture-unbind-on-buffer-modify workaround. Followup to 24cc971b1f43c61aa157228f6ea9916007ffb3cd, covering the remaining case.
6 years ago
#endif
GL: one more iteration on the Apple buffer texture workaround. This broke DSA-less contexts with ARB_multi_bind present.
6 years ago
{
/* bindImplementation already set above */
bindInternalImplementation = &AbstractTexture::bindImplementationDefault;
}
Moved global data from texture classes to per-context state.
12 years ago
}
De-inline TextureState constructor/destructor. It is heavier than it looks. Also fixed some includes in Implementation subdir.
13 years ago
Ability to reset internal state tracker.
12 years ago
void TextureState::reset() {
GL: don't use std::fill_n(). This used to somehow get dragged along with <string> on all platforms we tested on, but once we get rid of <string> we'd have to include <algorithm> instead. Not acceptable, also it's much shorter and less error-prone this way.
4 years ago
for(std::pair<GLenum, GLuint>& i: bindings)
i = {{}, State::DisengagedBinding};
Implemented ARB_shader_image_load_store.
10 years ago
#if !defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
GL: don't use std::fill_n(). This used to somehow get dragged along with <string> on all platforms we tested on, but once we get rid of <string> we'd have to include <algorithm> instead. Not acceptable, also it's much shorter and less error-prone this way.
4 years ago
for(std::tuple<GLuint, GLint, GLboolean, GLint, GLenum>& i: imageBindings)
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
i = std::make_tuple(State::DisengagedBinding, 0, false, 0, 0);
Implemented ARB_shader_image_load_store.
10 years ago
#endif
Ability to reset internal state tracker.
12 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
}}}