will land on green

@fkorotkov Hard to say if it's our problem or yours, but the tests-windows shard here spun around doing nothing for like an hour. Then it did some more stuff and then eventually timed out. At a guess, there's some sort of resource starvation going on in the Windows cluster.

@Hixie me and @gspencergoog saw that but there were issues with reporting logs for such timed out tasks. I fixed it yesterday and at least now there are logs presented every time (they were missing sometimes for timed out tasks).

It seems it was hanging on running slivers_evil_test.dart.

If you are interested I can pretty easy add a way to run windows containers on GCP instead of Azure. The only problem will be that Windows Contianers on GCP will start in 4-5 minutes instead of 1:30-2 on Azure. But we can check if it will be more stable.

@fkorotkov It looks like it spent 30 minutes on one test, then 30 minutes on another test. Hard to tell for sure exactly what was happening, but I've never seen that on any other platform, which is why I suspect some sort of resource starvation (e.g. maybe a disk read took a really long time). I don't have a strong opinion on how Cirrus implements things. :-)

Full logs of failing run: 5708719757197312-test_all.log

@fkorotkov I ran it again and this time it was 60 minutes on a different test (C:\Windows\Temp\flutter sdk\packages\flutter_localizations\test\date_picker_test.dart: DayPicker shows dates for ar_AR). I'll rerun it again.

Log of that failing run: 5694495647596544-test_all.log