Add AdapterLimitsGPUTest that sets all limits to the adapter's

[greggman]

Note: I'm not sure how I refactored device_pool.ts is good but, the old code created one initial device with no descriptor the first time in acquire. It then created another of the type actually wanted by the user. That no longer works becauase the adapter is passed in and so if the first requestDevice succeeds, then the adapter has been used and can't be used for the second one.

Also, assuming this change is approved, I can (and should?) probably refactor LimitsTest to use this vs how it's doing it now. That would end up using cached devices.

Issue: #3363

---

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 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.

[kainino0x]

LGTM. This makes more sense than the thing that tried to get a device twice.

[ben-clayton]

This change seems to be breaking a bunch of things with the node based test runner.

Test that use

  .beforeAllSubcases(t => {
    t.selectDeviceOrSkipTestCase({ requiredFeatures: ['shader-f16'] });
  })

Are now causing exceptions on devices that don't support shader-f16. This results in tests failing instead of being skipped.

I also get the impression that once a device is lost, all subsequent tests fail to reacquire a new device.

[greggman]

sorry about that. I'll fix assp