magnum

Commit Graph

Author	SHA1	Message	Date
Vladimír Vondruš	8357442ad7	DebugTools: don't try to use system plugin dirs in CompareImage test. Due to that, all the tests got skipped on the CI and moreover when ABI was broken, tests were failing hard due to ABI mismatches with system installed plugins. No good. Some variants still need to access the system plugin dir (such as checks for compressed data), but those are minimal. Sanitized them nevertheless so a broken system-installed plugin doesn't break the test.	7 years ago
Vladimír Vondruš	9005b3c9aa	Platform: expose also window events via Sdl2Application::anyEvent().	7 years ago
Guillaume Jacquemin	daf1deab3e	Sdl2Application: add glContext() accessor.	7 years ago
Vladimír Vondruš	875ad29dde	DebugTools: the visibility.h header is needed always.	7 years ago
Vladimír Vondruš	94ca8273e1	No need to leak GL headers everywhere.	7 years ago
Vladimír Vondruš	8e8010c40b	Document and print helpful error messages for GL-dependent features.	7 years ago
Vladimír Vondruš	053f406b7d	Platform: make SDL2 and GLFW apps buildable without TARGET_GL again. I'm pretty sure this worked correctly in 2018.04.	7 years ago
Vladimír Vondruš	63e914909f	GL: fix unused variable warning in a test.	7 years ago
Vladimír Vondruš	cb22975d1b	Trade: minor cleanup. These are linked transitively from the Magnum library.	7 years ago
Vladimír Vondruš	8b748ace7d	Text: properly report failures from AbstractFontConverter.	7 years ago
Vladimír Vondruš	a5634f4bf4	Text: fix annoying warnings on GCC 4.8.	7 years ago
Vladimír Vondruš	834c5fe40d	Text: API-independent base for glyph caches. Allows the Font and FontConverter plugins be built without TARGET_GL enabled. That was the last piece missing for making the magnum-plugins repo completely GL-free.	7 years ago
Vladimír Vondruš	215d50a087	Text: allow empty files in AbstractFont. In particular, MagnumFont can be an empty.	7 years ago
Thibault Jochem	28e300da02	zero-initialize utf8 buffer.	7 years ago
Vladimír Vondruš	75cc1725ad	CMake: link both GlfwApplication and Sdl2Application to EGL, if needed. This is already done in the FindMagnum module for both, but not in the source tree -- for SDL2 it was done in the FindSDL2 module (seems a strange place) and for GLFW nowhere. To make things consistent, I'm doing that in the Platform CMakeLists now and removed it from FindSDL2. Hopefully this doesn't break anyone's workflow (static builds of SDL2/GLFW?). In that case we would need to re-add it to the Find modules as well.	7 years ago
Vladimír Vondruš	3d0e237dbf	DebugTools: texelFetch() doesn't work w/o min filter on SwiftShader. WTF. I have no explanation for this.	7 years ago
Vladimír Vondruš	8ba2cac0ee	DebugTools, MeshTools, TextureTools: handle broken gl_VertexID. DebugTools::textureSubImage(), MeshTools::fullScreenTriangle() and TextureTools::distanceField() all plagued by this on SwiftShader.	7 years ago
Vladimír Vondruš	5bf09f0cea	MeshTools: update fullScreenTriangle() docs.	7 years ago
Vladimír Vondruš	cf3ad5c700	MeshTools: test fullScreenTriangle(). Turns out the ES3 version doesn't work on SwiftShader due to broken gl_VertexID.	7 years ago
Vladimír Vondruš	67cf914864	TextureTools: avoid the test looking for plugins we didn't supply.	7 years ago
Vladimír Vondruš	9e3690670a	TextureTools: explicitly force operation sequencing in the test. On GCC 4.8 at least, release() gets called on the image before size(), causing size() to be zero.	7 years ago
Vladimír Vondruš	4ff67d53b1	TextureTools: don't benchmark GPU if we don't have time queries. Makes a loud BOOM on SwiftShader.	7 years ago
Vladimír Vondruš	40e0150e09	DebugTools: add license header to shader sources.	7 years ago
Vladimír Vondruš	06d2f0b598	DebugTools: check for different error string based on impl color read fmt.	7 years ago
Vladimír Vondruš	bbf2e15287	DebugTools: give up and skip R8 screenshot() tests on ES2. It's... not worth the pains.	7 years ago
Vladimír Vondruš	80fc86ccba	DebugTools: add some fuziness to ObjectRenderer test. SwiftShader misplaces two pixels. No biggie.	7 years ago
Vladimír Vondruš	28d89611f3	DebugTools: fix textureSubImage() for float textures on ES3. By accident the returned image had zero size.	7 years ago
Vladimír Vondruš	efcc4e0d1c	DebugTools: harden the test for screenshot().	7 years ago
Vladimír Vondruš	3f0b298d9b	DebugTools: disable bufferData() tests on drivers w/o map_buffer_range. Was crashing, haha.	7 years ago
Vladimír Vondruš	211949d3ce	DebugTools: improve the reverse format mapping in screenshot(). Previously it was going through all following ifs even after it found the matching value. Using a lambda to do early returns.	7 years ago
Vladimír Vondruš	796bdb6731	GL: new pseudo-extension MAGNUM_shader_vertex_id. Will be used to distinguish if the driver supports gl_VertexID. This is not the case on SwiftShader ES3 contexts (and could be elsewhere, probably).	7 years ago
Vladimír Vondruš	74183eb8b9	GL: improve extension docs.	7 years ago
Vladimír Vondruš	3d08c641d8	GL: yet another workaround for SwiftShader, now with XFB buffer bindings.	7 years ago
Vladimír Vondruš	579d7b489f	GL: use proper literals in a test.	7 years ago
Vladimír Vondruš	a52d6944ce	GL: enable TimerQuery test everywhere except on WebGL. What the heck is wrong with me, why is this class always put in a wrong place.	7 years ago
Vladimír Vondruš	d8a2e70617	GL: XFAIL on SwiftShader with integer data type mismatches. Accepted everywhere else (or so is my experience).	7 years ago
Vladimír Vondruš	aa033e2e97	GL: XFAIL in PixelStorageGLTest if the driver wants to read RGBA. The data are arranged for RGB, no idea why this wants RGBA. Probably because ES?	7 years ago
Vladimír Vondruš	5912467f06	GL: follow ES restrictions in multi-attachment framebuffer tests.	7 years ago
Vladimír Vondruš	fdac0a7150	GL: expect less precision on ES2 tests that use RGBA4 renderbuffers.	7 years ago
Vladimír Vondruš	2b7a9b7eaf	GL: fix AbstractShaderProgram compute test on ES 3.1 contexts.	7 years ago
Vladimír Vondruš	24210a00f9	GL: disable ANGLE_instanced_arrays and OES_texture_3D on ES2 SwiftShader. Turns out the extensions are advertised but the ANGLE-suffixed and OES-suffixed entrypoints are missing, leading to crashes.	7 years ago
Vladimír Vondruš	c9937f73a0	GL: disable indexed Mesh tests on drivers w/o indexed extensions.	7 years ago
Vladimír Vondruš	f5e9ba7e8a	GL: wait, there's not just three ES instanced extensions, there's FIVE. Like, why can't they just depend on each other or something?	7 years ago
Vladimír Vondruš	44a77ed918	GL: fixed broken links to extensions. These were originally working around some inconsistencies in the upstream naming, but that's fixed since.	7 years ago
Vladimír Vondruš	1ae982020d	GL: detect ANGLE using the renderer string on non-WebGL platforms. No need to use such hacks there.	7 years ago
Vladimír Vondruš	b76d1226e0	GL: detect SwiftShader.	7 years ago
Vladimír Vondruš	6a982f7841	Adapted to Corrade changes.	7 years ago
Vladimír Vondruš	127c8b624b	GL: engrish.	7 years ago
Vladimír Vondruš	d6e0186dbf	GL: actually fix the VAO / element array buffer binding state interaction. Fixing the new (and now failing test) from the previous commit. In setIndexBuffer(), I was resetting the state tracker to a VAO state that was about to be set in the very next step, and then, when doing that next step the state tracker "optimized away" the state change because it thought it was already done (even though it wasn't). Reordering the two operations fixes it. In comparison to how this was meant to be done in the original `169031fb7b`, the new way should do the same but additionally avoid a bunch of redundant state calls. Let's hope no more bugs related to this appear.	7 years ago
Vladimír Vondruš	ee16c25125	GL: further harden VAO and index buffer binding state interaction tests. This one crashes. Turns out the `169031fb7b` contained a random temporary test state instead of the real solution (and so the comment didn't even match the code, it should have been resetting that to 0). That also made some tests fail with DSA disabled, but none of the tests were actual Mesh tests, just accidentally hitting the problematic code path. I took the opportunity to look at this more closely and investigate why this failed -- turns out, in setIndexBuffer(), I was resetting the state tracker to a VAO state that was about to be set in the very next step, and then, when doing that next step the state tracker "optimized away" the state change because it thought it was already done (even though it wasn't). The new test in MeshGLTest covers this particular case.	7 years ago

... 11 12 13 14 15 ...

3159 Commits (d64a67ab6358dc472b83b9b082e259c4fa3aaed5)