15 Oct
2025
15 Oct
'25
2:02 p.m.
On Wed Oct 15 14:02:01 2025 +0000, Nikolay Sivov wrote:
There appear to be crashes in Win10 CI tests run. Not sure if it's caused by this MR, but even if not we won't see results because of the crash.
The crash was occurring after `test_h264_output_alignment()` was called but before a trace at the first line of the function could emit output. That makes no sense and I'm fairly sure it's something to do with an earlier test, so I moved the call above some others. Now there's a race condition after waiting for `MESessionTopologyStatus`, where the frame size may have already been updated. Looks like this test is not meaningful, so how about I remove the `MESessionTopologyStatus` check point entirely?