This WPT status shows that the first sub-test is passing on Chrome M142. This is wrong: this could only happen if experimental web-platform features are enabled.
In reality, both sub-tests there have always been failing in Chrome, and Chrome got a fix recently only as an experimental web-platform feature! I confirmed that local WPT runs on Chrome Stable match this failing expectation today, so does the wpt.live outcome.
Surprisingly, this problem does not occur in all tests. Here is a WPT showing a correct result---the Chrome Stable result here would be green if experimental features were enabled but it is red as expected.
Clearly the dashboard has a problem somewhere! Is it enabling experimental web-platform features for some WPTs? Or mixing up Chrome Canary result with Stable?