Fixes#1356.
Output for the test case in that issue is now:
error: the value indexed by a `[]` subscripting expression must not be a pointer
┌─ wgsl:5:14
│
5 │ let a = *pv[3]; // Problematic line
│ ^^ expression is a pointer
Could not parse WGSL
Both a `var` binding and a `let` binding of a pointer to a variable create
entries in the `lookup_ident` tables whose expression is a `LocalVariable`.
However, the `let` should be a pointer, whereas the `var` should be a reference.
This patch changes the tables to hold `TypedExpression` values, thus preserving
the distinction.
The old filenames were probably based on the shader stage, not the shader
function name, which the PR changed to do in response to a review suggestion.
`tests/out/glsl/operators.main.Fragment.glsl` is no longer generated because the
function in question changed to a Compute entry point.
`tests/out/wgsl/multiple_entry_points-glsl.wgsl` is no longer generated because
the input test was removed in 057dc310.
Make the parser code more closely follow the spec's grammar around
`unary_expression`, `postfix_expression`, and `singular_expression`.
Change the handling of postfix expressions (indexing, member/component access,
and swizzling) to apply the indirection at the appropriate time, resulting in
code improvements on all output formats. For example, where we used to generate
the following MSL:
metal::float4 _e13 = bar.matrix[3];
float b = _e13.x;
we now generate, simply:
float b = bar.matrix[3].x;
Propagate WGSL reference types correctly, so that parenthesizing expressions no
longer causes the Load Rule to be applied.
Together with #1332 (already landed), this is a replacement for #1312, and
unblocks #1352.
Fixes#1351.
Regression from ce676cf130.
We need to output (*d).mx rather than *d.mx, at least according to Tint.
wgsl-in seems to handle *d.mx just fine, which is likely a separate bug.
* [spv-in] New two pass parser based
* [spv-in] Allow expressions defined in dominant block in different scopes
* Make the patch non breaking
* [spv-in] Allow scope transfers in phi instructions
* [spv-in] Remove unused stuff
* [spv-in] Handle switch merges as breaks
* Remove no longer needed stuff
* Revert some changes to prepare to merge
* Remove dead code
* Don't spill into local if in scope
* [spv-in] Documentation, comments, some renaming for clarity.
* Address comments
Co-authored-by: Jim Blandy <jimb@red-bean.com>
* Fix GLSL output for non-fallthrough switch cases
Partially reverts 02c74b5002 and
fixes#1309.
* Fix indentation of control-flow WGSL code
* Clean up glsl-out switch case fallthrough handling
Only insert a break if needed
and if a case is fallthrough, insert a comment indicating this.
* Implement fuzzing for the GLSL parser
* Remove arbitrary dependency from naga
Derive `Arbitrary` for proxy objects in `fuzz/fuzz_targets/glsl_parser.rs`
instead.
Fixes#1305.
Ensure that two `back::spv::LocalType::Image` values are sure to be equal (and
hash equal) whenever they would generate the same `OpTypeImage` instruction, so
that the usual duplicate removal via `Writer::lookup_type` works. Accomplish
this by changing the contents of `LocalType::Image` to more closely match the
operands of `OpTypeImage` instructions.
Previously, `LocalType::Image` included a `ImageClass::Storage::access` value,
which did not affect the `OpTypeImage` instruction generated. If two
`LocalType::Image` values differered only in their `access`, then they would get
separate entries in `Writer::lookup_type`, two identical `OpTypeImage`
instructions would be generated, and SPIR-V validation would fail.
