Priority: P1: Critical
Affects Version/s: 1.1
Fix Version/s: None
See this test run and search for tst_QDateTime.
Its block of test output starts normally, but search for "Executing instruction 332" and you'll see the block ends with no Totals line, no "Finished testing" line - and 18 tests are unmentioned.
The test whose output I want to see should assert, but there's no sign of the test, much less my expected QFATAL (I'm making sure my regression test does actually fail before I fix it).
Nor is there a crash-log indicating that the test crashed.
Nor does Coin say anything at all about tst_QDateTime ending prematurely.
Same happens for tst_QAlgorithms (instruction 342), tst_QLine (370), tst_QRect (398), tst_QImage (488) and I have to guess there are more. On the way to that, notice that "Executing instruction 334" interrupts the output for tst_QDate.
Something is severely fishy here.
Are the tests crashing in some way that the agent fails to recognise ?
Is the agent discarding data ?
(Does it buffer the data on its way to logging ? If so, is the buffer finite ?)
Is something killing the tests without logging that it's done so ?