magnum

Commit Graph

Author	SHA1	Message	Date
Vladimír Vondruš	cc9db2d610	MeshTools: show the more common way of using custom shader attributes.	3 years ago
Vladimír Vondruš	b07d1aae8b	MeshTools: fix concatenate() for more than 4 GB of data. This isn't as easy to auto-test, unfortunately, so hopefully the comment will ensure it doesn't regress again.	3 years ago
Vladimír Vondruš	fdfa6b31e2	Trade: fix MeshData checks for index/vertex data over 4 GB.	3 years ago
Vladimír Vondruš	582776aeef	Trade: additional sanity assertions for {Mesh,Scene}Data bounds checks.	3 years ago
Vladimír Vondruš	d0aca29ba5	Trade: rewrite MeshData bound assertions similarly to SceneData. There it's done in a significantly more robust way, without relying on `begin < end` but rather a negative stride.	3 years ago
Vladimír Vondruš	0679cccdf3	Trade: test handling of MeshData and SceneData over 4 GB. Passes for SceneData but fails for MeshData due to 32-bit types used by accident. The two also have a vastly different calculations in the range checks, should unify that first.	3 years ago
Vladimír Vondruš	51f9a31d3c	Trade: reorder test case definitions to match declaration order.	3 years ago
Vladimír Vondruš	7e79216c34	doc: use a shorthand expression for populating a GL::Buffer.	3 years ago
Vladimír Vondruš	5f0aa9582d	doc: mention the new SpngImporter in the file format tables.	3 years ago
Vladimír Vondruš	d4cc7dc318	Math: MSVC 2015, nobody asked you to fail here.	3 years ago
Vladimír Vondruš	3e59bf23a1	Doc++, to sync with Python bindings.	3 years ago
Vladimír Vondruš	62c395708e	Math: ability to construct a Color3 from RG and B.	3 years ago
Vladimír Vondruš	b38d3eea89	Shaders: add SSBO support to all shaders. For when there's so much to render that it wouldn't fit into an UBO and splitting draw calls and binding buffers under an offset is unwanted overhead.	3 years ago
Vladimír Vondruš	f1df71a000	Shaders: remove various redundant ifdefs.	3 years ago
Vladimír Vondruš	0bc0287ad4	Shaders: no, we don't use any textures in the Line shader. Bad copypaste, bad.	3 years ago
Vladimír Vondruš	521489ac2e	Shaders: don't assert for ARB_uniform_buffer_object presence twice. Done this way only in the Phong shader, everywhere else it's just the MAGNUM_ASSERT_GL_EXTENSION_SUPPORTED() macro. Some WIP code that I forgot to clean up?	3 years ago
Vladimír Vondruš	a2b4c85847	Shaders: don't pass light count that gets unused to format(). Not sure what was this about originally.	3 years ago
Vladimír Vondruš	b338e35f0c	Shaders: fix instanced MeshVisualizer test on ES2 and WebGL 1. Seems like I never got around to test this variant? Huh.	3 years ago
Vladimír Vondruš	8153adbc1c	Shaders: don't format & pass _LOCATION defines on ES < 3.1 and WebGL. Saves quite a few unnecessary allocations and string operations.	3 years ago
Vladimír Vondruš	1847c7201d	Shaders: use explicit texture location and binding on ES3.1+. No reason not to, reduces the amount of unnecessary driver calls quite significantly.	3 years ago
Vladimír Vondruš	7707341e47	Shaders: minor cleanup.	3 years ago
Vladimír Vondruš	218aef5016	Shaders: don't define & format irrelevant skinning values for UBOs. The UBO use case doesn't need any initializers or uniform locations for these, it's just wasted allocations.	3 years ago
Vladimír Vondruš	fdd9d798b3	Shaders: decouple light array size and per-draw light count. So it's possible to have light culling enabled on, say, 64 lights, but with only at most 3 applied each draw, allowing the shader compiler to unroll the loop if it makes sense. This also better prepares for SSBO support where the total light count would be unbounded and thus the value ignored, and thus the value can be 0.	3 years ago
Vladimír Vondruš	a00952cf95	Shaders: light count isn't used to test anything here.	3 years ago
Vladimír Vondruš	6995ad4fa7	Shaders: update outdated docs about passing stuff to constructors. There's a Configuration class now.	3 years ago
Vladimír Vondruš	1a33796784	Shaders: um what was this comment about, even? Probably some leftover from when I tried a bitmask-based approach (which failed to deliver as it was too slow to iterate through in the shader)? Uhh.	3 years ago
Vladimír Vondruš	f2fd6bdccf	Shaders: use zero per-vertex count to disable skinning, not total count. This prepares for SSBO support where the total count is unbounded (and thus the value is ignored, thus it can be 0). Also regroup the doc paragraphs so it's clear what's related to UBO usage and what applies to classic uniforms as well.	3 years ago
Vladimír Vondruš	5d330c28b4	GL: don't try to test texture views w/o texture storage. I wonder how this even passed the test before, given my PKGBUILD also has a variant that disables ARB_texture_storage?	3 years ago
Vladimír Vondruš	b73f8eca0b	GL: document what's the reasonable value for Framebuffer::clearDepth().	3 years ago
Vladimír Vondruš	862c48517f	Math: add Vector2::{r,g}() and Vector3::rg(). For consistency with X/Y accessors. No reason to not have them.	3 years ago
Vladimír Vondruš	73c881f5ae	Math: mention that cross product can be used to calculate triangle area.	3 years ago
Vladimír Vondruš	bd0230865a	Shaders: mark the line shader as experimental. I still expect the API to change slightly to accomodate for line stipple or textured lines (such as possibly having different smoothness alongside and across the segment), plus there are some TODOs that might change how it behaves in certain corner cases.	3 years ago
Vladimír Vondruš	e9e05ecf6c	Shaders: use same index order as MeshTools::compileLines() has. To avoid pointless differences and impossible-to-reproduce bugs.	3 years ago
Vladimír Vondruš	54f7824f2d	doc: update changelog.	3 years ago
Vladimír Vondruš	cfa8604118	Shaders: add a "screenshot" for LineGL. OMG PRETTY!!!	3 years ago
Vladimír Vondruš	f3cc5bd1c1	Shaders,MeshTools: adapt image comparison thresholds to SwiftShader.	3 years ago
Vladimír Vondruš	e615637c21	Shaders,MeshTools: adapt image compare thresholds for NVidia & llvmpipe. Well, most of them, there's two outliers on NVidia that need further investigation.	3 years ago
Vladimír Vondruš	e3c7410ac7	MeshTools: initial utility for compiling a mesh for the line shader. Currently just the bare minimum, more features such as handling multiple contiguous strips and loops inside a single mesh or an overlapping layout will come later.	3 years ago
Vladimír Vondruš	a26d77afb4	MeshTools: remove excessive indentation in a test.	3 years ago
Vladimír Vondruš	327d0cb749	MeshTools: clarify why actually the API isn't on ES2.	3 years ago
Vladimír Vondruš	c8d8fd7e19	Shaders: add debug output for LineVertexAnnotations as well. Need to inspect a generated mesh.	3 years ago
Vladimír Vondruš	c209c27d93	Slight clean up in ImageFlags*D debug printer tests. Not sure why it was printing that random number also for flags that had actual values.	3 years ago
Vladimír Vondruš	dff1a1ce60	Math: add also castInto(T[], T[]) overloads. For generic code, which would otherwise have to invent some SFINAE "use castInto() if the types are different and Utility::copy() otherwise" nastiness in every such case, and that's just annoying.	3 years ago
Vladimír Vondruš	43acb7a5c4	Math: castInto() between [Unsigned]Int and [Unsigned]Long wasn't tested. But it was covered, I suppose due to some other code calling it already.	3 years ago
Vladimír Vondruš	506841f4a9	Shaders: initial implementation of a line shader. The shader requires the input data to be laid out in a rather specific way, and there will be a dedicated MeshTools utility for it in the following commits. For independence though, the shader tests use a custom helper. The initial implementation has certain corner cases which will be eventually resolved. For now they are pinned down with repro cases in the test. But apart from that, it's pretty much usable in practice. Remaining join styles (round and miter-clip) as well as stipple support will eventually follow as well.	3 years ago
Vladimír Vondruš	681c98605b	MeshTools: have a separate assert message for input and output. Otherwise it may be hard to guess which of them is failing. Ideally the assert would also contain the stride vs type size, maybe I do that next time I spend that much time investigating why it asserts.	3 years ago
Vladimír Vondruš	bf347109c9	MeshTools: make generateIndices() work with indexed MeshData. This way one no longer needs to duplicate() the original first, leading to an enlarged size of vertex data for no reason.	3 years ago
Vladimír Vondruš	971c0d90a7	MeshTools: rename test variable for clarity.	3 years ago
Vladimír Vondruš	d88ec88187	MeshTools: generate*Indices() overloads taking a 2D index buffer. So these can be used directly in the generateIndices() implementation that takes a MeshData.	3 years ago
Vladimír Vondruš	cfb5b015f7	MeshTools: test generateIndices() assertions through the top-level API. It delegates to the Into() implementation but that shouldn't mean the assert isn't checked there.	3 years ago

1 2 3 4 5 ...

10314 Commits (cc9db2d610c89829e92935aa91570c2b7e2be66e) All Branches Search

10314 Commits (cc9db2d610c89829e92935aa91570c2b7e2be66e)

All Branches