Teach qt-testrunner.py how to make sense of tst_selftests' Catch output

A change I'm working on failed its tst_selftests run (due to cmptest producing output that didn't match expectation) and Catch reported the failure. Catch is used by tst_selftests and it doesn't know anything about the XML format Coin asks our usual testlib-based tests to run. So qt-testrunner.py mistakenly warns about the failure, suspecting a crash, although tst_selftests did not crash. (Note that some of the selftests it runs do crash – they're meant to – so don't be mislead by anything related to that in output.) The failure of tst_selftests was on macOS, but I doubt the issue with parsing Catch's output is limited to that platform.

Please teach qt-testrunner.py to recognise failures in tst_selftests, despite their being reported differently to those of other tests.