merge_pr_49235

[wptrunner] Run --enable-sanitizer tests with their own executors (#49235)

Running non-crashtests with the crashtest executor likely doesn't
provide high-fidelity coverage because the test completes on the load
event by default. This means that code paths normally exercised after
the load event aren't actually run under sanitization. Also, some
tests require testdriver, which the crashtest executor doesn't implement
yet, to progress.

Now, simply run each test with its corresponding type's executor and
suppress functional failures (i.e., not timeout or crash) at the end.
This also simplifies the result coercion, which didn't interact
correctly with #47903.

Tested locally:

./wpt run -y chrome infrastructure --metadata infrastructure/metadata --enable-sanitizer