Validate sampling operations on unaligned block-compressed textures.

[castano]

Fixes Issue: #4707

---

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.)
• Test have be tested with compatibility mode validation enabled and behave as expected. (If not passing, explain above.)

Requirements for reviewer sign-off:

• Tests are properly located.
• Test descriptions are accurate and complete.
• Tests provide complete coverage (including validation control cases). Missing coverage MUST be covered by TODOs.
• Tests avoid over-parameterization (see case count report).

[github-actions]

Results for build job (at 5f1ad5c):

+webgpu:shader,execution,expression,call,builtin,textureSampleLevel:sampled_2d_coords,compressed_unaligned:* - 600 cases, 1800 subcases (~3/case)
-TOTAL: 287982 cases, 2362149 subcases
+TOTAL: 288582 cases, 2363949 subcases

[castano]

I finally got some time to look into this. The CTS already had all the machinery needed to test these edge cases, so this is a single new test in textureSampleLevel.spec.ts.

One concern that I have, though, is that checkCallResults uses a softwareTexture that come from createTextureWithRandomDataAndGetTexels and is initialized by reading the compressed texture back using textureLoad to sample every texel. The JS implements filtering and address modes, but not block decoding.

A decode bug that affected textureLoad and textureSampleLevel identically on partial blocks would not be caught. Implementing a software decoder for every block compression format is out of scope for this PR.

[castano]

Tested on Chrome Canary 157.0.8089.0 in both core and compatibility mode, on an Apple M4 Pro.

[Kangz]

LGTM!