The https://testresults.qt.io/coin/api/results/qt/qtdeclarative/a6526c5c966b87a394cec71a79386dd4b95a117d/MacOSMacOS_11_00arm64MacOSMacOS_11_00arm64Clangqtci-macos-11.0-arm64TestOnly/986900e09b34d8efb74398ad9d7bf5bb1b85b9d7/test_1658235870/log.txt.gz

workitem failed after re-running failed testcase. With first run there was two functions failing:

FAIL!  : tst_HoverHandler::deviceCursor(synth, tablet wins)
FAIL!  : tst_HoverHandler::deviceCursor(synth, mouse wins)

No with re-runs only

FAIL!  : tst_HoverHandler::deviceCursor(synth, tablet wins)

kept failing.
Even that the test failed, tests functions was marked as flaky in influxdb.
Now when COIN parses the testresult XML files it checks for testcase name and test function name. In this specific case the test function names were the same. So even if the deviceCursor() kept failing, deviceCursor was also among the passed test functions so COIN considers it as flaky. Which is wrong.