On Sun, Jan 29, 2006 at 04:45:51AM +1100, Adam Kennedy wrote: > > I haven't been able to find (although I haven't looked to hard) for a > documented set of result codes. But either DEPFAIL or N/A makes sense.
See CPAN::YACSmoke pod. Just before Christmas I completed the TestersGuide pod, which also explains it in more detail. That version has been released to CPAN, but is available via CVS on SourceForge. > If not, I certainly plan to deal with it that way. N/As cascade. Agreed. > Without knowing how many exist now, I hope to define a more > comprehensive set of HTTP-like codes, but that design element is still > flexible. There are only 4 that are submitted to cpan-testers, PASS, FAIL, NA and UNKNOWN (no tests supplied with the distribution). CPAN::YACSmoke also deals with ABORTED and UNGRADED, which are referenced internally, as well as IGNORED (a subsequent version PASSes) and NONE (where a version has been cleared of it's previous grade by manual intervention. Changing these HTTP like codes, might okay for an internal representation, but would require ALOT of work to change several CPAN modules and ensure all the testers upgraded. There is also the fact that all existing reports are in the system and not going to change. Although I may have misunderstood your intentions. > >Is there a mailing list or wiki or some > >such? (Subversion repository?) > > No infrastructure for now, until it's actually announced, but the IRC > channel is logged ala #perl6, so you can visit when you wish, and catch > up when you care. I use IRC very rarely, as due to work and family taking up most of my time ... when I'm not organising a conference ;) Now that more tools are being created, it seems more appropriate to discuss these elsewhere than on a QA list. I'll investigate setting something up on the Birmingham.pm server, unless there is somewhere else that would be more appropriate. Cheers, Barbie.