Paul Goyette <[email protected]> writes: > The timeout happens because something much earlier in the test took > longer than expected to complete, causing the entire test run to > exceed the timeout. The timeout is 4 hours, and a test run usually > completes in about 2 1/2 hours.
Is there an easy way to look for tests that take a lot longer? Perhaps running the tests on a checkout before the suspect change, and after, and loading both the csv formats in a spreadsheet, and looking at the ratio of execuation times? It would be interesting to have a database of historical norms for each test and some way to warn if a value is consistently outside them.
pgpXa_s9LOTy6.pgp
Description: PGP signature
