Regression testing

[Peter Jones]

On Tue, 2005-03-22 at 09:44 -0600, Jason L Tibbitts III wrote:
> >>>>> "MAH" == Mike A Harris <mharris at www.linux.org.uk> writes:
> 
> MAH> The openssh tests also consume 99.9% of all CPUs even in 8way
> MAH> boxes, and do this for like 16 hours straight.
> 
> Are they broken or just very thorough?  Is there a reasonable subset
> that could be run in a reasonable amount of time that would still make
> sure that, say, the compiler didn't miscompile some bit of code?
> 
> Having a test suite that takes two weeks to run on a box that a mere
> mortal might own seems counterproductive, if the goal is to have mere
> mortals running it.

Well, part of this particular case is a scheduling problem -- openssh's
test suite doesn't use SMP, IIRC, but we don't schedule something else
to build when the box isn't fully utilized.

-- 
        Peter