On 25/10/11 17:19 -0400, Greg Blomquist wrote:
On 10/25/2011 03:55 PM, Matt Wagner wrote:
> On Tue, Oct 25, 2011 at 03:38:45PM -0400, Chris Alfonso wrote:
>> Unfortunately, with the tests taking so long to run, with 30 minutes allocated
you will probably only get through one patch. I know folks have looked into speeding up
the specs, but to no avail. I think we will make a lot of progress and let the ack-ers
spend more time developing if we rely on developers to ensure the tests pass on before
they post, or after a rebase and repost.
I actually just ran all the tests (as part of a patch) and found that
the specs didn't take /nearly/ as long as they used to. At the end of
last sprint it took on the neighborhood of 30 min to run the specs.
Now, it took about 13 min.
Chris' estimate of 30min is still valid since it takes about 13min to
run all the cukes, too. So, total test time runs ~30min. But, I'm
really impressed that the specs have gotten so much better recently.
I would propose we try to address this pain point be determining _why_
the tests are so slow. There is always more that can be done to speed
them up, whether it be better use of mocks, changes to configuration,
or something else. IMO, this should be a story in redmine, I added it
here:
https://www.aeolusproject.org/redmine/issues/2654
-j
>
> Yes, amen! I can't say I've never mistakenly sent a broken patch, but I
> think you're spot-on here -- make sure all tests pass before you send
> your patch, rather than leaving it to the tester to discover.
>
> -- Matt
> _______________________________________________
> aeolus-devel mailing list
> aeolus-devel(a)lists.fedorahosted.org
>
https://fedorahosted.org/mailman/listinfo/aeolus-devel
_______________________________________________
aeolus-devel mailing list
aeolus-devel(a)lists.fedorahosted.org
https://fedorahosted.org/mailman/listinfo/aeolus-devel