magnum

Commit Graph

Author	SHA1	Message	Date
Vladimír Vondruš	25640c4f99	Vk: add an XFAIL for pool allocation failure on NV as well. Not a problem, important is that it succeeds when it should succeed.	2 years ago
Vladimír Vondruš	6f298f4d9f	Vk: allow the allocated buffer memory to be larger in the test. Fixes an assertion in Utility::copy() on NVidia.	2 years ago
Vladimír Vondruš	0c7e5c53e7	Shaders: 1.0/0.0 doesn't produce +inf on NVidia, but a NaN. This caused MeshToolsCompileGLTest to fail in a strange way, and PhongGLTest::renderLowLightAngle() as well. Which looked rare enough that I first suspected some random driver bug, but apparently it was all caused by these using the default infinity range instead of explicitly calling setLightRanges() on the shader. The test is now updated to explicitly verify the default value when a setter isn't called, to catch this problem better in case it reappears in a different form elsewhere.	2 years ago
Vladimír Vondruš	5c047bc128	Shaders: work around a bogus shader compilation warning on NVidia.	2 years ago
Vladimír Vondruš	72382caf72	Shaders: zero-initialize a partially uploaded texture in the test. Otherwise the difference fluctuates and makes the test appear flaky.	2 years ago
Vladimír Vondruš	0a027b4469	MeshTools,Shaders: update test image comparison thresholds for NVidia.	2 years ago
Vladimír Vondruš	13fff0a3eb	Text: prevent minor DistanceFieldGlyphCacheGL differences on ES2. And hint that in TextureTools::DistanceFieldGL docs as well.	2 years ago
Vladimír Vondruš	0137865e9c	TextureTools: NVidia gives the same error on both desktop and ES.	2 years ago
Vladimír Vondruš	60a78956bb	TextureTools: fix test errors on ES2 with EXT_texture_storage. And verify that no GL error happens also in the assertion tests.	2 years ago
Vladimír Vondruš	e3ef801dae	GL: another NVidia-specific XFAIL for BufferTexture. This also doesn't seem too important.	2 years ago
Vladimír Vondruš	0f1ffe9dae	GL: add XFAILs for some NVidia-specific behavior in ShaderProgram tests. Neither of those are really critically important. Failure not being a failure is fine (just don't make the shader broken in the first place), multiple color output bindings should be in the shader source anyway.	2 years ago
Vladimír Vondruš	e9b07ef4ce	GL: clean up instanced data access in Texture GL tests. This was quite a mess, the whole array name copypasted to each and every access. I mean, yeah, originally I thought this would be the* usage pattern, but oh god this resulted in SO MANY copypaste errors. Which ultimately means the two annoying NVidia failures in the TextureGLTest related to pixel storage in compressed 3D images are no longer failing as a result of cleanup.	2 years ago
Vladimír Vondruš	657f7164a8	modules: fix FindSDL2 on Windows if SDL's config is present. The CI builds that began failing in the previous commit should work now again. Sorry for the regression.	2 years ago
Vladimír Vondruš	b567c0a5c9	package/ci: download a SDL version with CMake configs on AppVeyor. A regression was introduced in `2a8e550b57` that affects only Windows builds. The version downloaded on the Windows CI unfortunately doesn't contain CMake configs, so this went unnoticed.	2 years ago
Vladimír Vondruš	fbd2adb9e8	Text: fix warning on a non-assert build.	2 years ago
Vladimír Vondruš	adfd0a80bd	doc: updated credits and changelog.	2 years ago
Andreas Leroux	a9daddfb63	Platform: add GlfwApplication::tickEvent() and setMinimalLoopPeriod(). Co-authored-by: Vladimír Vondruš <mosra@centrum.cz>	2 years ago
Vladimír Vondruš	ba0fbe3eec	Platform: make Sdl2Application::setMinimalLoopPeriod() unit-safe. I wanted to add this for GlfwApplication, only to realize the timer there is a double, in seconds, so accepting integral milliseconds there felt very weird. Let's use the fancy new time types instead. Also updated the docs to (hopefully) clarify that setSwapInterval() has to be called in order for the interaction between the two to work properly. As usual, the old variant taking untyped milliseconds is a deprecated alias to this one.	2 years ago
Vladimír Vondruš	f7cdd61211	Platform: opt-in test for Sdl2Application::tickEvent().	2 years ago
Vladimír Vondruš	734ddd0d75	Platform: test constant redraw in both SDL2 and GLFW applications.	2 years ago
Vladimír Vondruš	f84deb0ef6	Platform: minor cleanup in EmscriptenApplication test code.	2 years ago
Vladimír Vondruš	2a8e550b57	modules: make FindSDL2 delegate to SDL2's own config file, if possible. Now that the Find module doesn't need to get copied to all downstream repositories, I can finally work on it in peace.	2 years ago
Vladimír Vondruš	9208922c2b	GL: Emscripten's Clang 10 actually isn't Clang 10.	2 years ago
Vladimír Vondruš	97cb586d5b	doc: fix mistyped extension name.	2 years ago
Vladimír Vondruš	8b832382a8	package/ci: build on Emscripten with emulated corrade-rc.	2 years ago
Vladimír Vondruš	282f95d723	CMake: install and automatically use Find modules of dependencies. In short this means the users no longer need to care about FindSDL2 etc., it'll just get done automatically. They don't need to drag along FindMagnum / FindCorrade anymore either, but having them gives a much more reasonable error message if the library cannot be located at all, so I still recommend having them.	2 years ago
Vladimír Vondruš	f7079cd174	modules: skip all target setup in FindMagnum if the component isn't found. Before, if e.g. find_package(Magnum REQUIRED Vk) was called and the installed Magnum didn't have Magnum::Vk enabled at all, it would still try to call find_package(Vulkan) and do all other setup which could then fail as well, resulting in way too many noise printed. Now the decision about the _FOUND variable is moved all the way to the top, and the actual target setup is only done afterwards.	2 years ago
Vladimír Vondruš	a24a62b9c8	CMake: mark variables that shouldn't need to be toggled as advanced.	2 years ago
Vladimír Vondruš	edb08411ea	Updated toolchains submodule.	2 years ago
Vladimír Vondruš	ebe9ecaa9b	modules: update FindCorrade.cmake, FindNodeJs is no longer needed. Corrade now installs it itself and uses it when appropriate, removing the need for downstream projects to carry it along.	2 years ago
Vladimír Vondruš	eef921c238	It's October already?!	2 years ago
Vladimír Vondruš	60631459d9	Math: cross-link dot product and cross product docs.	2 years ago
Vladimír Vondruš	eb790d7cc2	GL: clean up remaining functions for extension-dependent functionality. All others are either already converted to call the underlying GL APIs directly, or they cannot do that.	2 years ago
Vladimír Vondruš	9d99e00bd1	MeshTools: make generateIndices() less restrictive. Sometimes it's desirable to just make a mesh indexed unconditionally, which wasn't really possible with this thing, and having to special-case that at the call site is very annoying. Now it makes a mesh trivially indexed if it has any other primitive than the ones listed, and if it's already indexed, it's just passed through.	2 years ago
Vladimír Vondruš	7e2ba1fd81	MeshTools: properly test the output primitive from generateIndices().	2 years ago
Vladimír Vondruš	6733550e97	MeshTools: added generateTrivialIndices(). Because writing a for loop for this is annoying, and using std::iota() isn't any better either.	2 years ago
Vladimír Vondruš	15d3759a23	MeshTools: doc++	2 years ago
Vladimír Vondruš	134c625f48	MeshTools: fix the generateQuadIndices() test to be more stable.	2 years ago
Vladimír Vondruš	45528e85ef	doc: update copyright updating docs. Very meta.	2 years ago
Vladimír Vondruš	363717467d	package/ci: reduce parallelism on the Vulkan build, OOMs too often.	2 years ago
Vladimír Vondruš	8da46ef9dc	GL: advertise fake MAGNUM_compressed_texture_astc_{ldr,hdr}. Because everything is a trash fire. Ugh.	2 years ago
Vladimír Vondruš	6236ebaa68	GL: geez, Emscripten, can't you just fix those warnings already.	2 years ago
Vladimír Vondruš	a95fdb5598	GL: implement EXT_clip_control and WEBGL_polygon_mode on Emscripten. Following the spirit of extension-based functionality, the entrypoints are available always but do something (i.e., call the actual WebGL API) only if the extension is advertised. Which it is only on Emscripten 3.1.66+ because older versions don't have the corresponding entrypoints, so there it's marked as disabled. Additionally, EXT_polygon_offset_clamp is now also working on 3.1.66+, but there's no wrapper for it yet.	2 years ago
Vladimír Vondruš	e967cf551e	external: entrypoints for WebGL exts implemented in Emscripten 3.1.66. This is also the first time they just stopped bothering and introduced a name that is neither in the official gl.xml nor in the google-private ANGLE-specific gl.xml. So yeah, gotta make my own gl.xml. I wonder how other people do this. Do they just type out the headers themselves? Or what's the process for this fucking trash fire these days??	2 years ago
Vladimír Vondruš	d2141d0a8e	external: some extensions got new GL entrypoints since last time.	2 years ago
Vladimír Vondruš	aaab54772b	Add isCompressedPixelFormat{Normalized,FloatingPoint,Srgb}() helpers. To have feature parity with similar helpers for non-compressed PixelFormat.	2 years ago
Vladimír Vondruš	9b4fd3f45b	Doc++	2 years ago
Vladimír Vondruš	f589f03214	Move a source-local array closer to the function that uses it.	2 years ago
Vladimír Vondruš	6a2c053302	doc: show C++14 and C++20 constexpr as [constexpr] in the docs. Ideally it'd say the standard version as well, but that requires adding a feature to m.css. Not that important right now.	2 years ago
Vladimír Vondruš	a056d7de92	TextureTools: don't include Vector2 in DistanceFieldGL.h. This makes the header pretty much weightless so it doesn't hurt if it's included directly in Text::DistanceFieldGlyphCache etc.	2 years ago

1 2 3 4 5 ...

11092 Commits (25640c4f9979fa308c74dba93c8aadee889609ff) All Branches Search

11092 Commits (25640c4f9979fa308c74dba93c8aadee889609ff)

All Branches