On 20 Mar 2011, at 00:13, David Chisnall wrote: > Hi Everyone, > > The connection.m test has been failing for me intermittently ever since I > started running the test suite, but today I finally persuaded it to fail in > the debugger. I don't really know the NSConnection code very well, but this > is the back trace:
I increased the timeout in the test. But ... that test is not a very good one anyway ... all it's doing is running the runloop to give another thread a chance to connect back to the main thread, and seeing if an exception gets raised. There's no check to make sure that the other thread actually succeeded in connecting to the main thread, and there's no mechanism to stop the test earlier if the thread connects quickly, so increasing the timeout means the test takes longer to run. This test is really a good candidate for a rewrite when someone has time. _______________________________________________ Gnustep-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnustep-dev
