I don't mind removing it. It's there to detect substantial slow-downs in common implementations.
On Wed, May 26, 2010 at 1:08 AM, Robin Anil <[email protected]> wrote: > Oh Found it. the clover coverage instrumentation was slowing down the test > and there was a time check that was failing > > assertTrue(timeMS < 1000L * (long) allowedTimeSec); > > Sean, Is it ok to disable the time check ? for a nightly coverage > instrumentation. > > I think its not ok for keeping running time as a check. Tests run in a lot > of places, it should be correctness check. > Maybe time based tests can be moved to another nightly job, sort of running > a set of benchmarks and draw a graph around the timings > > Robin >
