On Wed 05 Mar 2008 at 12:11AM, Danek Duvall wrote:
> On Tue, Mar 04, 2008 at 11:42:59PM -0800, Dan Price wrote:
>
> > > However, I think before we can institute this, the testsuite needs to
> > > support known failures. Otherwise it's going to be a royal pain in the
> > > butt to run the suite and see dozens of errors you're not fixing and you
> > > can't really do anything about.
> >
> > I'm not clear whether I can do this without a substantial rewrite of
> > unittest. It may be possible, but I've avoided looking at the
> > implementation of unittest itself, in the event that I need to reauthor
> > something equivalent to it at some point. I'm obviously reluctant to do
> > major hacking on something like unittest...
> >
> > Can you live with some inconvenience in the short term?
>
> Short term, yes. But people are going to be scared by lots of noise when
> they run the test suite, and if a) they're required to before putting back
> and b) we're supposed to be adding more known failure cases, then the noise
> is going to get out of control pretty quickly and make this venture a
> failure. :( So I think short term needs to be pretty damned short.
Ok. I've spoken to Steve Purcell (the original author of unittest)
and he was kind enough to give me some very sage advice about how to do
this in a sane way.
I'll take a whack at it and see how it goes.
-dp
--
Daniel Price - Solaris Kernel Engineering - [EMAIL PROTECTED] - blogs.sun.com/dp
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
