The infrastructure/ tests: macOS / 1 (of 1) job is reporting Unexpected results: 1 on many recent PRs. Sample of affected runs (all from different contributors, all against master):
On at least #58587's run, the unexpected test was /infrastructure/testdriver/bidi/emulation/set_touch_override.https.html, failing with webdriver.error.JavascriptErrorException: javascript error (500): A JavaScript exception occured: cannot serialize cyclic structures. The same error string is carried as KNOWN-INTERMITTENT-ERROR on several sibling BiDi testdriver tests in the same run, so it looks like a Safari STP BiDi serialization issue affecting a cluster of testdriver tests. The set_touch_override.https.html.ini metadata has if product != "chrome": ERROR, but wptrunner reports expected OK, so something may also be wrong with how that metadata is being honoured.
This is effectively blocking #58587 (red check on a PR ready for review) and red-marking many other macOS-infra checks. Would appreciate a look.
The
infrastructure/ tests: macOS / 1 (of 1)job is reportingUnexpected results: 1on many recent PRs. Sample of affected runs (all from different contributors, all against master):check-pinned-chromium-revisiontestrunner-loader-typingOn at least #58587's run, the unexpected test was
/infrastructure/testdriver/bidi/emulation/set_touch_override.https.html, failing withwebdriver.error.JavascriptErrorException: javascript error (500): A JavaScript exception occured: cannot serialize cyclic structures.The same error string is carried as KNOWN-INTERMITTENT-ERROR on several sibling BiDi testdriver tests in the same run, so it looks like a Safari STP BiDi serialization issue affecting a cluster of testdriver tests. Theset_touch_override.https.html.inimetadata hasif product != "chrome": ERROR, but wptrunner reportsexpected OK, so something may also be wrong with how that metadata is being honoured.This is effectively blocking #58587 (red check on a PR ready for review) and red-marking many other macOS-infra checks. Would appreciate a look.