On 25/07/2013 18:46, Robert Collins wrote:
On 26 July 2013 10:18, Monty Taylor <[email protected]> wrote:
Hey guys!
Fred made a generalization of the nova testing tool that checks that you
are adding a test with your patch:
We should check it out and see if we can run it across OpenStack.
Mmm, just to say I'm skeptical about such a thing. For several reasons:
- when fixing things with broken tests, you don't add tests (but you
do fix them)
That's taken into consideration by the script. If the change only deals
with tests, it's marked OK. On the same vein, if the change is doc or
translation only, it's marked OK too.
- 'a test' is the wrong metric. 'Enough tests' is the right one : can
this do cyclometric complexity + type inference analysis? [Which is
what good reviewers are doing in their head].
- when you measure something you /get it/ : and if what you measure
isn't what you want, thats a bad thing.
The tool isn't designed to replace reviewers but to help them do their
review job more efficiently by running automatically what some reviewers
are running manually already on their own.
Cheers,
--
Fred - May the Source be with you
_______________________________________________
OpenStack-Infra mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
