DatabaseErrorTest failing sporadically

[DeepDiver1975]

Observation

2025-03-11T10:04:03.6602480Z 34/35 Test #25: DatabaseErrorTest ................Subprocess killed***Exception:  62.75 sec
...
2025-03-11T10:04:16.7638340Z QFATAL : TestDatabaseError::testDatabaseError(Vfs::Off) 25-03-11 10:03:12:987 [ fatal default ]:  ASSERT: "_end == std::chrono::steady_clock::time_point {}" in file /Users/runner/work/client/client/src/common/chronoelapsedtimer.cpp, line 37
2025-03-11T10:04:16.7638600Z FAIL!  : TestDatabaseError::testDatabaseError(Vfs::Off) Received a fatal error.
2025-03-11T10:04:16.7638850Z    Loc: [/Users/runner/work/client/client/src/common/chronoelapsedtimer.cpp(37)]
2025-03-11T10:04:16.7639010Z Totals: 1 passed, 1 failed, 0 skipped, 0 blacklisted, 12099ms

This assert happens when ChronoElapsedTimer::stop() is called consecutively.
Occurrences:

client/src/libsync/httplogger.cpp

Line 183 in 3a94ba5

ctx->timer.stop();

client/src/libsync/syncengine.cpp

Line 675 in 3a94ba5

_duration.stop();

Hypothesis

Within the HttpLogger the timer is reset and stopped under various conditions. I suspect that under some circumstances timer.stop() caan be called twice or a related ::reset() is missing.

[modSpike]

I am seeing intermittent crashes in some of the sync tests, too, which "magically go away" on re-run locally. I think it very well could be the same thing.

[modSpike]

this was not actually fixed by the linked PR - the tests failed with same issues described here on first run. as before, rerunning the tests "fixes" it

[DeepDiver1975]

https://github.com/owncloud/client/actions/runs/13808828494/job/38625517125?pr=12081

2025-03-12T10:29:23.2460980Z QFATAL : TestDatabaseError::testDatabaseError(Vfs::Off) 25-03-12 10:28:09:958 [ fatal default ]:	ASSERT: "_end == std::chrono::steady_clock::time_point {}" in file /Users/runner/work/client/client/src/common/chronoelapsedtimer.cpp, line 37