Fixes #9283 : firstLeadingBit returns signed semantics result instead…

[39ali]

… of unsigned for u32 input 0xFFFFFFFF on Metal backend

Connections
Fixes #9283

Description
fixes a small logic bug between u32 and i32 in firstLeadingBit

Testing
ran all tests

Squash or Rebase?

Checklist

• Run cargo fmt.
• Run taplo format.
• Run cargo clippy --tests. If applicable, add:
  • --target wasm32-unknown-unknown
• Run cargo xtask test to run tests.
• If this contains user-facing changes, add a CHANGELOG.md entry.

[teoxoy]

The snapshots need to be regenerated.
Also, is there a CTS test we could enable for this?

[andyleiserson]

There is a CTS test, webgpu:shader,validation,expression,call,builtin,firstLeadingBit:*. To me it looks like it does cover this case, but it's not failing either before or after this change, which is odd. (There is one failure unrelated to this fix prior to #9262.)

You can run it with:

cargo xtask cts 'webgpu:shader,validation,expression,call,builtin,firstLeadingBit:*'

[39ali]

@andyleiserson i was gonna checkout why but the cts cmd needs some fixing on macos

panicked at xtask/src/util.rs:91:52:git --versiondid not have the expected structure: failed to parse version number 2 ("3 (Apple Git-146)") asu8

[teoxoy]

I think webgpu:shader,validation,expression,call,builtin,firstLeadingBit:* should be passing, this should be tested by the execution test instead: webgpu:shader,execution,expression,call,builtin,firstLeadingBit:*.

[teoxoy]

Change looks good!

[teoxoy]

Could you addd webgpu:shader,execution,expression,call,builtin,firstLeadingBit:* to the list of passing tests? And I will merge this.

[teoxoy]

Thanks!

[andyleiserson]

I was reading the source for the execution test but running the validation test. 🫠