@foolip Looks like the css-grid tests benefited from reftest-wait, even if we don't do a delayed screenshot. Can these be submitted like this?

@LukeZielinski looks like the test are timing out now because the reftest-wait class is never removed. That makes them pass the stability check, but it's still breaking the tests. Oops :)

@LukeZielinski looks like the test are timing out now because the reftest-wait class is never removed. That makes them pass the stability check, but it's still breaking the tests. Oops :)

Doh! didn't click enough buttons. Abandoning.

@LukeZielinski looks like the test are timing out now because the reftest-wait class is never removed. That makes them pass the stability check, but it's still breaking the tests. Oops :)

Doh! didn't click enough buttons. Abandoning.

..Actually, I realize I just needed to takeScreenshot call - so it looks like this passes if we do takeScrenshot (without the delay)?

So the latest on this is that these tests do seem to pass by simply loading ahem stylesheet, without any explicit waiting or screenshotting. @foolip WDYT?

Well that's pretty funny, wasn't that the original change carved out from #17205?

When we were debugging this a Chrome bug seemed somewhat plausible, so maybe Chrome actually changed and that's why it's now passing.

@LukeZielinski is this a case you were ever able to repro locally by opening the test in a browser manually? If so could you use https://www.chromium.org/developers/bisect-builds-py to verify that the change is due to a Chromium change? If not, then I say let's just merge this since it passes :)

@LukeZielinski is this a case you were ever able to repro locally by opening the test in a browser manually? If so could you use https://www.chromium.org/developers/bisect-builds-py to verify that the change is due to a Chromium change? If not, then I say let's just merge this since it passes :)

This is one of the test that behaves differently locally than on CI unfortunately, so I'm inclined to just submit.