Re: cvs head has lots of test failures

[Phil Muldoon <pmuldoon at redhat dot com>]

I checked out each day from August 5th.

The test-suite works on 08/05, 08/06, 08/07 with cvs co corresponding to 
those dates.

It fails on 08/08 (ie cvs co -D "2 days ago"). So it looks like 
something changed on that date that caused some issues. So people who 
checked in on that day, should double check their patches and look for 
gotchas

Thanks

Phil

Mark Wielaard wrote:
> On Thu, 2006-08-10 at 09:06 -0400, Elena Zannoni wrote:
>   
> > Mark Wielaard writes:
> >  > 
> >  > We really need an autobuilder to catch such things early.
> >
> > We really need people to run make check before checking stuff in.
> > Please folks be considerate to your fellow coders...
> >     
>
> Unfortunately make check is not always very reliable. We are working on
> some very different platforms. It can be that you are not seeing
> failures on your platform, but others do see them on theirs. And it
> seems make check is often randomly broken :{ Making it hard to see if
> some failure was already there or that it was introduced by your new
> code. I would love to have a zero make check fails checkin policy.
>
> Although the fact that some tests take more than half an hour to run
> make that a little hard in practise.
> http://sources.redhat.com/ml/frysk/2006-q3/msg00160.html
>
> Having an autobuilder on some canonical platform with public test
> results would make comparing results between days/checkins easier at
> least.
>
> Cheers,
>
> Mark
>
>