magnum

Commit Graph

Author	SHA1	Message	Date
Vladimír Vondruš	9ee7e899aa	Ignore warnings of glaring uselessness on clang-cl. I hope those don't fire on regular Windows clang as well, this seems to be specific to the way how MSVC-compatible template instantiation works.	8 months ago
Vladimír Vondruš	507e8845d3	Import GL shader resources on static builds unconditionally. It's faster that way because it doesn't involve a linear lookup, if the resource is already imported it's a constant-time check and it becomes a no-op.	8 months ago
Vladimír Vondruš	f2da1e12b5	Math: accidental stray semicolon. Copypasted 14 times. Sigh.	8 months ago
Vladimír Vondruš	bc5cb3e363	Test: fix Clang missing prototypes warnings harder. I thought I fixed all in `6b58165856` (which should have gone into `1c1ded5869` already), but apparently there were four more instances I missed. Sigh.	8 months ago
Vladimír Vondruš	8f6e35f4ad	package/homebrew: C++14 is nothing to boast with in 2025.	8 months ago
Vladimír Vondruš	06550adb04	package/homebrew: use current master for the stable version. 2020.06 is just too old, and waiting for the time I make next release just to stay "correct" is increasingly counterproductive.	8 months ago
Vladimír Vondruš	b85f0c8dfe	doc: update credits and changelog.	9 months ago
Vladimír Vondruš	f297d593bf	Math: make all trivial mutable getters constexpr under C++14. Technically speaking there are many more opportunities to mark things as C++14 constexpr, such as various operators, but doing so would make them impossible to optimize with various (SIMD) intrisnics unless basically providing two separate implementations for them, one scalar, slow and constexpr and the other fast but not compile time, and dispatching with std::is_constant_evaluated() or some such from C++20. In other words, such functions cannot and will not be marked as constexpr in C++17 or older anyway, because the preference is and still should be runtime perf rather than doing excessive work at compile time, where it's hard to debug or reasonably test etc etc., and the more you rebuild the more you'll hate the cost of extra compile time caused by that. Co-authored-by: Stanislaw Halik <sthalik@misaki.pl>	9 months ago
Vladimír Vondruš	c232a57325	Math: doc++	9 months ago
Vladimír Vondruš	c41a13ecc8	Math: make test using declarations consistent as well.	9 months ago
Vladimír Vondruš	ddde7f77db	Math: rename test cases for consistency with other test files.	9 months ago
Vladimír Vondruš	0fc40bea05	package/ci: use GitHub for downloading OpenAL releases. So I don't needlessly strain a small website, plus the certificate is currently expired.	9 months ago
Vladimír Vondruš	20f66ff0b4	Platform: delete EmscriptenApplication WebGL context only if there is.	9 months ago
Vladimír Vondruš	968b9a59dc	Platform: add a TODO. Of course it's all extremely messy because there's no better way.	9 months ago
Vladimír Vondruš	ac9780f65b	Platform: add test for creating contextless EmscriptenApplication.	9 months ago
Vladimír Vondruš	43c552aa01	Shaders: properly test uniform struct trivial copyability. Somehow this wasn't done for any of the types and then the lack of tests got copied to the Ui library as well. Sigh.	9 months ago
Vladimír Vondruš	67b8559de3	Text: can do this string concatenation at compile time.	9 months ago
Vladimír Vondruš	b04e726be7	package/ci: lcov!!! don't test my patience!	9 months ago
Vladimír Vondruš	b2307ef9f5	package/ci: skip OpenAL tests on the Xcode 15 image, they hang, FFS.	9 months ago
Vladimír Vondruš	b30ec0a5f7	modules: link to static MoltenVK dependencies in FindVulkan.	9 months ago
Vladimír Vondruš	bd1ff4ad94	package/ci: test with static MoltenVK in the macos-static build. Fails to link due to missing dependencies. Fix in the next commit.	9 months ago
Vladimír Vondruš	6b58165856	Test: suppress missing prototypes warnings on Clang. Should have gone into `1c1ded5869` already.	9 months ago
Vladimír Vondruš	f66d1cc1d6	Shaders: more cases of accidental lowp. Should have gone into `b93de590bc` already. No idea what am I doing there.	10 months ago
Vladimír Vondruš	50a3632027	GL: expand notes for the "nv-broken-buffer-dsa" workaround.	10 months ago
Vladimír Vondruš	440653dbb8	package/ci: update to a newer Emscripten image that isn't 404ing. Consistently with what's done in Corrade.	10 months ago
Vladimír Vondruš	d45da3a5a6	GL: this function isn't here.	10 months ago
Vladimír Vondruš	ad9cb8c96e	Shaders: increase test thresholds for GLES3 on NVidia. Not sure why that thing has a bigger difference if texture arrays are used, but only on ES3 and not on desktop.	10 months ago
Vladimír Vondruš	dc77bf980b	GL: doc++	10 months ago
Vladimír Vondruš	2d36cab303	GL: put branch contents on a new line. So line code coverage reports which branches weren't taken by tests. Other libraries need similar treatment but there's too many cases to fix so I'm doing it just here for now.	10 months ago
Vladimír Vondruš	28f5d5a07c	GL: assume object is created if it's already bound somewhere. The assertions triggered by the tests added in previous commit no longer fire.	10 months ago
Vladimír Vondruš	744b38dada	GL: ensure the texture is actually created when using ARB_multi_bind. With ARB_DSA disabled, the test added in previous commit would assert already during bind(), and not later in label(). This was due to the state tracker being updated too early here.	10 months ago
Vladimír Vondruš	86972cb07f	GL: repro cases for wrap() without ObjectFlag::Created causing asserts. This happened to me from time to time and I always treated it as user error. But when thinking more about it, calling wrap() without actually knowing if the object was already created or not, is a completely valid use case, especially in cases of making non-owning references to existing GL objects that might only get created at some point after. These tests now cause assertions in all cases on GLES, on desktop some only if ARB_direct_state_access is disabled. In all cases the assertion is triggered by querying object label (which is a no-op if no debug extension is present, but it calls createIfNotAlready() always), because that's the simplest, the real-world cases are usually some texture bind() calls. Fix in the next commits.	10 months ago
Vladimír Vondruš	cfb18081bb	GL: these helpers are never used on WebGL, compile them away.	10 months ago
Vladimír Vondruš	ce466be335	GL: add getters for ObjectFlags to all GL object wrappers. Mainly just for diagnostic purposes / testing, not really useful otherwise.	10 months ago
Vladimír Vondruš	6b11916fb8	GL: add debug output operators for ObjectFlags.	10 months ago
Vladimír Vondruš	0abacbb84d	GL: minor.	10 months ago
Vladimír Vondruš	372edc13bc	package/ci: exclude GL and Vk sources from code coverage on Windows. I'm giving up there. Both Clang-cl and MinGW GCC report a far bigger set of lines than their Linux / Mac counterparts (block ends, lines containing just `else`, etc etc.), which wouldn't be a problem on its own, if those lines would be covered. But on Windows there isn't a single CI job that runs through GL/Vk codepaths -- as Mesa isn't easy to get there and SwiftShader is utter crap -- which means there are extra lines in GL/Vk codepaths that are reported as uncovered from Windows CI jobs, while the same lines on Linux and Mac arere excluded and thus there isn't anything that could make them green as well.	10 months ago
Vladimír Vondruš	33c37e8e84	No need to use reinterpret_cast for char* <-> void* casting.	10 months ago
Vladimír Vondruš	82254ff189	Doc++	10 months ago
Vladimír Vondruš	cda1da17b0	Order ImageD raw data accessors before all format-dependent accessors. For consistency with Trade::MeshData, MaterialData and SceneData, where the raw data access is also separated from the formatted access.	10 months ago
Vladimír Vondruš	dd7b0754a0	Test: test that compressed blocks can be 2D/3D even for 1D/2D images.	10 months ago
Vladimír Vondruš	9ce6f147f7	Audio,GL,Vk: no point in including the extension lists in code coverage.	10 months ago
Vladimír Vondruš	05069848a4	GL: new "nv-cubemap-broken-dsa-compressed-subimage-upload" workaround. Yet another broken behavior with compressed textures on NVidia unearthed by the recent changes with compressed block properties being set almost always.	10 months ago
Vladimír Vondruš	ab05558ee2	GL: expand Texture3D compressed subimage upload to investigate a NV bug. Nope, unfixable. I don't even know how many days I have lost on trying to figure out bugs with compressed textures on NVidia. Ugh.	10 months ago
Vladimír Vondruš	239984d4f9	GL: group tests with 3D cubemap images together and name them properly. They are all using desktop-only APIs and most of the compressed variants likely suffer from the same NVidia bug, so it's useful to have them next to each other.	10 months ago
Vladimír Vondruš	06e0118184	GL: of course I pushed a total stupidity right to master. Once again the case of not properly commiting working copy changes before a push. This should have gone to `096f1a5269` already.	10 months ago
Vladimír Vondruš	5c88bc1d58	GL: actually test Buffer::setTargetHint().	10 months ago
Vladimír Vondruš	3d55698cea	GL: exclude debug enum printer values from coverage.	10 months ago
Vladimír Vondruš	c804cca59f	GL: set compressed block sizes only if actually needed. And make resetState() aware of ARB_compressed_texture_pixel_storage so when it's not supported, it doesn't cause the very next compressed image upload or download to fire a GL error due to an unknown state being set.	10 months ago
Vladimír Vondruš	1851f345a7	GL: minor cleanup. Those redundant prefixes were likely from back when the GL wrapper was still a part of the root Magnum library.	10 months ago

1 2 3 4 5 ...

11646 Commits (3b26301bc3fa5818a7b7a14c11cfa378d41f2f7e) All Branches Search

11646 Commits (3b26301bc3fa5818a7b7a14c11cfa378d41f2f7e)

All Branches