On Wed, 2008-01-16 at 23:06 +0000, sebb wrote: > Repeating "mvn test" shows the same Failure. > > Odd. > > The failures look rather similar - why should the first one be ignored? > Perhaps a bug in Maven or Surefire. >
Probably because they are not thrown in the test case itself but in the #tearDown method. It is just a theory. It can well be the code simply fails to propagate the exception correctly. I'll investigate. Writing synchronous test cases for asynchronous I/O is not an easy thing. Test cases are very ugly and flaky because most of the processing logic needs to be run in worker threads. You never know whether the event does not trigger because the I/O reactor still needs time or something got screwy. So, lots of things can potentially go wrong. > If I then run "mvn clean", then "mvn test" the stacktrace appears, but > it is not treated as a failure, even if I run "mvn test" several times > without an intervening clean. > > Very odd. > > Tried running "mvn test" in a fresh source tree: > - first time, no stack trace at all > - second and 3rd time, stack trace and failure recorded. > - mvn clean, then mvn test: stack trace and failure recorded. > > The stacktrace seems to be repeatable now ... anything I can do to > debug it further? > I think I have found the problem spot using my wife's dual-core Mac. I suspect I have never been seeing those problems because my primary Linux system is single-core. Thanks for all you help Oleg --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]