On Tue, 2010-10-12 at 17:01 -0400, Bill Nottingham wrote:
James Laska (jlaska(a)redhat.com) said:
> > No, this is well before the tree gets to that state. It would need to
> > be implemented as a check on git push.
>
> We attempt to validate comps in this test using the comps.rng in git
> (see results logs posted earlier). I'm not tremendously familiar with
> that part of the test, but I believe it should catch that?
You're validating it after the repo has already been made. In this case,
it was broken such that the repo couldn't even be made.
Ah, thanks for the clarification.
James