- The Typifier was updated to expect `uint`
- Both `glsl` and `spv` frontends where updated to cast the result to `sint`.
- Both `glsl` and `spv` backends where updated to cast the result to `uint`.
- Remove cast in `msl` backend.
- samplerCubeShadow requires vec3 gradients
- Bias must be written after offset
- sampler1D hack on GLES requires vec2 gradients
- textureSize doesn't accept a lod argument for MS samplers
- sampler1DArray hack on GLES requires a y component on imageLoad
Fixes#1642.
Since 1d textures cannot have mipmaps, MSL requires that the `level` argument to
texel accesses and dimension queries be a constexpr 0. For our purposes, just
omit the level argument altogether.
Minimize allocation in the namer. Heap-allocate string only when needed to
provide an owned key for a hash table, or to hold synthesized text. Try to reuse
allocations.
Many unnecessary trailing `_` characters are removed from test output. These
were all superfluous; separators are still inserted where necessary.
Replace uses of `call_unique` with uses of `call` and `call_or`, which becomes
public. It's not clear when `call_unique` is correct to use, and avoiding a few
numeric suffixes here and there isn't worth it.
* Implement lexing for all WGSL number literal types
* Move number literal test cases
* Adjust tests to match WGSL spec on number literals
Suffixes are not type names and currently only a plain `u` is supported
for uints. More specifically, `i` and `f` suffixes or suffixes with
widths in bits like `u32` are not supported at the moment.
* Add more tests for invalid number literal suffixes
* Replace code too new for Rust 1.43
* Implement parsing for hexadecimal integers
* Switch to enum number types, and Bytes for width
* Check for negative and leading zeros in int literals
* Implement parsing of hex floats with hexf-parse
* Update error message tests
* Update snapshot test output files
* Clean up lexer state machine code
* Clean up unexpected token error code
* Move number literal parsing to own submodule
* [hlsl-out] flatten the entry point inputs
Previously, the logic was re-ordering the inputs according to the binding.
This breaks if one of the inputs is a struct. With this change, the struct
fields are also flattened into the fake entry point struct. We also
construct the original arguments at the beginning of the function.
* hlsl-out: completely separate the flattened IO structs from the original IR structs
Previously, we had heuristics to detect if a particular struct needs the fields
to be re-ordered. We'd re-order interface structs without layout, and the detection
was very fragile and easily wrong.
The new logic is spawning separate struct types if we need any re-ordering to happen.
It's solid, there are no heuristics.
Some SPIR-V texture access instructions take coordinates as integers, others as
floats. The types of coordinates in Naga expressions generally match those in
SPIR-V, but Naga indices for arrayed textures are always integers, whereas
SPIR-V combines coordinates and array indices into a single vector, so indices
need to be cast to match the coordinate component type.
This commit makes `write_texture_coordinates` properly cast array indices to
match the coordinates' component type before combining them all into a single
result vector.
Fixes#1186.
* Resurrect texture_storage_* tests
* Test parsing of `var<storage,write>`
* Default storage textures to READ
* Restore default features
* Fix glsl/hlsl/msl/spv front and back ends
* Add missing test outputs
* All-around fixes for the storage access
Co-authored-by: Dzmitry Malyshau <kvarkus@gmail.com>