Making it easier to work with shaders https://git.yummers.dev/slang.git/atom?h=master 2025-02-04T21:58:54+00:00 2025-02-04T21:58:54+00:00 cheneym2 acheney@nvidia.com 2025-02-04T21:58:54+00:00 urn:sha1:78a6389aa4b01cfe4f8b4ec34d48b2cbfc092865 * Fix warning about push constants, shaderRecordEXT These resources should not warn about missing vk::binding. * format code --------- Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com> 2025-01-29T19:59:42+00:00 cheneym2 acheney@nvidia.com 2025-01-29T19:59:42+00:00 urn:sha1:2ba6458eba8bd2d4f4d2ffdd452ae089e5b50907 * Fix combined sampler documentation and warning * Update comment, show detailed '-fvk-t-shift' message in warning instead of generic '-fvk-xxx-shift' * format code --------- Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com> 2023-09-08T22:57:00+00:00 Yong He yonghe@outlook.com 2023-09-08T22:57:00+00:00 urn:sha1:26a0b3e04689fee1ec9ec071eacd72faf1efe4eb * Fix attribute highlighting + language server crash. * Fix wave intrinsic. * Fix. * Fix. --------- Co-authored-by: Yong He <yhe@nvidia.com> 2021-02-12T19:31:56+00:00 jsmall-nvidia jsmall@nvidia.com 2021-02-12T19:31:56+00:00 urn:sha1:369279e91dde1b056d8d0e3bb83e7ba3f96321af * #include an absolute path didn't work - because paths were taken to always be relative. * WIP: First pass in supporting output of line error information. * Add support for lexing to better be able to indicate SourceLocation information. * Fix lexer usage in DiagnosticSink in C++ extractor. * Update diagnostics tests to have line location info. * Fixed test expected output that now have source location information in them. * Better handling of tab. * Fix test expected results for tabbing change. * DiagnosticLexer -> DiagnosticSink::SourceLocationLexer Added line continuation tests. * Fix typo. * Added String::appendRepeatedChar * Change to rerun tests. Co-authored-by: Tim Foley <tfoleyNV@users.noreply.github.com> 2019-03-05T22:24:44+00:00 jsmall-nvidia jsmall@nvidia.com 2019-03-05T22:24:44+00:00 urn:sha1:dcd9e574782b87d6280f1db8ee9ba6dbb7c96c8b * Add diagnostic for vk::binding failure. * Add test for vk::binding failure. * Add the expected output for glsl-layout-define.hlsl * * Copy over initialize expr if available when validating unchecked * Fix unloop - because now it always has one parameter (when before it could have none) * Split vk::binding and layout tests with invalid parameters * Removed the diagnostic for 2 ints expected * Added vk::binding that doesn't specify set in vk-bindings.slang * * Fix typo * Improve comments. 2018-11-30T23:03:31+00:00 Tim Foley tfoleyNV@users.noreply.github.com 2018-11-30T23:03:31+00:00 urn:sha1:7f2d2c9a248a9413e236826a4c6473a6fc104714 * Don't look at VK bindings when compiling for D3D and vice versa The compiler had been looking at all the modifiers on a declaration when piecing together binding information, whether or not those modifiers should apply on the chosen target API. This was working in practice because the "layout resource kinds" used by each API target were disjoint, for the most part. This change ensures that we don't even look at modifiers that don't apply on the chosen target, and furthermore adds a new warning that applies if the user is compiling a shader with explicit `register` bindings for Vulkan, if there are no corresponding `[[vk::binding(...)]]` attributes (under the assumption that if they want to be explicit in one case, they probably want to be explicit in all cases). * Allow explicit space/set bindings on parameter blocks The syntax for the D3D case is to specify a `space` in a `register` modifier, without any other register class: ```hlsl ParameterBlock<X> myBlock : regsiter(space999); ``` In the Vulkan case, the user must apply the `[[vk::binding(...)]]` attribute and is expected to use a `binding` of zero: ```hlsl [[vk::binding(0,999)]] ParameterBlock<X> myBlock; ``` This change includes a reflection test for the new capability (where we also confirm that it produces the expected output when compared with fxc), and a test for the diagnostic messages when the user messes up bindings for Vulkan. The implementation itself is fairly straightforward, since the compiler already treats registe spaces/sets as a resource that parameters can consume directly. Note: the test case for explicit parameter block space/set bindings includes some commented out code that lead to a compiler crash. I would like to fix the underlying issue, but it seemed sensible to keep the bug fix out of a change like this that is adding functionality.