Just to have another data point: Travis itself doesn't help much as
well. From time to time, I get timeouts when starting Wildfly from
within the integration tests. Just increasing the timeout is not a
proper solution, but there's not much we can do at Travis...
Based on my experience, Travis is the problem "most" of the time.
On 01.12.2015 02:35, mike thompson wrote:
So I think you all probably already know what I’m talking about. We have an
indeterministic suite of integration tests. So, for every PR you have to submit it X times
(2 to 7 is normal) before the PR can be merged and by that time it might already need to
be rebased. Is this on anyones radar to look at?
hawkular-dev mailing list