[Bug 54790] New: quartz:systemclock - test_advise() sometimes gets an unexpected even in Wine (GitLab CI)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=54790

Bug ID: 54790
           Summary: quartz:systemclock - test_advise() sometimes gets an
                    unexpected even in Wine (GitLab CI)
           Product: Wine
           Version: unspecified
          Hardware: x86-64
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: quartz
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

quartz:systemclock - test_advise() sometimes gets an unexpected even in Wine:

systemclock.c:242: Test failed: Event should not be signaled.

This mostly happens on the GitLab CI but it did happen once in the nightly
WineTest runs:

* 2022-11-15 on test-linux-32 for MR!1370
* 2022-12-14 on fg-deb64-t32
* 2022-12-19 on test-linux-32 for MR!1846
* 2023-03-28 on test-linux-32 for MR!2483
* 2023-04-05 on test-linux-32 for MR!2588

The GitLab CI and fg-* test configurations are the most likely to be busy with
other tasks while the tests are running so it is likely a timing issue caused
by having too much load on the system running the test.

It should be noted that this never happened on Windows though, particularly not
on the fgtb-w10pro64 VM which has the same host as the fg-* test
configurations. So this seems to be a  Wine issue.

[Bug 54790] quartz:systemclock - test_advise() sometimes gets an unexpected even in Wine (GitLab CI)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=54790

François Gouget  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |source, testcase

[Bug 54790] quartz:systemclock - test_advise() sometimes gets an unexpected even in Wine (GitLab CI)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=54790

Zeb Figura  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #1 from Zeb Figura  ---
Yeah, so we can increase the tolerance again, although I kind of have to wonder
at what point we say this is the fault of the VM for not providing consistent
enough timing. Not that I guess there's anything we can do about it if it
doesn't.

I doubt there's anything fundamentally specific to Wine about it. At best I can
only imagine that the Linux kernel is less consistent with timing than Windows.

[Bug 54790] quartz:systemclock - test_advise() sometimes gets an unexpected even in Wine (GitLab CI)

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=54790

--- Comment #2 from Zeb Figura  ---
I guess the thing is, if we aren't going to allow tests like this to run in
real-time, unconstrained by other processes fighting for the same CPU, then we
should probably delete them altogether (or just mark them interactive-only).
That's a bit unfortunate, but it's also true that CI can only go so far.