wgsl: Filter atan2 tests based on if const-eval or not

[zoddicus]

Rewrites how test cases are generated for atan2, so that if running in const-eval unbounded results will not be generated, since those will cause compilation errors.

Fixes #3088

---

Requirements for PR author:

• All missing test coverage is tracked with "TODO" or .unimplemented().
• New helpers are /** documented */ and new helper files are found in helper_index.txt.
• Test behaves as expected in a WebGPU implementation. (If not passing, explain above.)

Requirements for reviewer sign-off:

• Tests are properly located in the test tree.
• Test descriptions allow a reader to "read only the test plans and evaluate coverage completeness", and accurately reflect the test code.
• Tests provide complete coverage (including validation control cases). Missing coverage MUST be covered by TODOs.
• Helpers and types promote readability and maintainability.

When landing this PR, be sure to make any necessary issue status updates.

[github-actions]

Previews, as seen when this build job started (f104bc2):
Run tests | View tsdoc

[jiangzhaoming]

Looks good to me, thanks

[jiangzhaoming]

Just wonder, but why I don't get error on constant cases with current implementation?

[zoddicus]

Just wonder, but why I don't get error on constant cases with current implementation?

I think it is only some specific devices that have an issue with y == 0, or near 0 (and other boundary conditions). It is related to if they do FTZ and how they implement atan2. It is only recently with adding Android bots to Chrome that we regularly saw runs that were failing because of this.

That being said, this implies that some bots are being too tolerant of bad inputs, since these should have been failing in general. validation tests should be catching when const-eval is being too tolerant, but that coverage is yet to be implemented.