On Tue, Jan 08, 2008 at 09:02:06AM -0500, David Golden wrote: > On Jan 8, 2008 4:51 AM, Barbie <[EMAIL PROTECTED]> wrote: > > > I think your intention of send_skipfile above is the same as > > exclude_dists, but it should be decided by the local tester not the > > author. > > That's a a political choice. There are some authors who have > complained that their distribution is old or for some specialized > hardware/software configuration and just shouldn't be tested. It > doesn't really matter to me -- I just want to be able to fix things > once and let everyone benefit.
The author should remove them from CPAN to BACKPAN if they no longer wish them to be used. For other scenarios the author can ignore the reports on cpan-testers. Providing they don't get sent reports I don't see why users should find out the hard way that the install might not work for them. > > I'll amend exclude_dists in YACSmoke and look to adding the skipfile > > setting. > > As a side note about YACSmoke config -- given that you already have > "exclude_dists" as a funky HERE document of lines (which is not > "standard" INI format, I think), They are included with Config::IniFiles :) > I would suggest adding an > "exclude_dists_file" option to let that file be maintained separately > if people wish. Then our approaches will be harmonized. This I think has merit, although for backwards compatibility I think I'll implement both. > On the issue of excluding entirely authors -- that seemed the easier > way of not CC'ing someone rather than going distribution by > distribution. But rather than having to maintain "dists" and > "authors" separately, I realized that the CPAN::Dist->pretty_id has > both: DAGOLDEN/CPAN-Reporter-1.0602.tar.gz -- that makes it easy for > one regex to hit either one. Providing the detection matches correctly, such that a distribution exclusion doesn't match an author, e.g. "MAC" for "Net-MAC" should not be confused with the PAUSE ID "MAC", then that's fine. I've currently implemented it as '^' as the first character always matches the AUTHOR/DIST pattern, otherwise it matches the distribution only. The latter is for backwards compatibility. > What's your timeline on releasing YACSmoke changes? No idea. I think the first implementation will simply be to move the CPANPLUS code to a lower level plugin, with some outstanding enhancements and fixes, then a further implementation will include the CPAN/CPAN-Reporter plugin. If consuming some of Dave Cantrell's script into the plugin works, then I could release it all at once. However, I let Rob do the releases as it's really his baby, I just like feeding it ;) Plus it gives me the opportunity to have someone sanity check what I've done. > I was getting > near to whipping up a quick CPAN::Reporter::Smoker myself for my own > use and I'll hold off if the rest of the framework is there (though it > might be more heavyweight than I was considering). It might seem a bit more heavyweight, but YACSmoke can provide the bulk testing framework, that CPAN-Reporter doesn't currently have access to. What I would like to do though is use a report module that can create all the necessary authomated parts (as per CPAN-Reporter) for the CPANPLUS plugin. It's currently the lacking part of the reporting process of YACSmoke. > Is it in a repo somewhere? Rob has the original code on sourceforge, but sourceforge no longer seems to like me as although I can read the repo I can't write to it :( I've since changed laptops 4 times in the last few months, so it may be worth me trying again. Cheers, Barbie. -- Birmingham Perl Mongers - http://birmingham.pm.org Memoirs Of A Roadie - http://barbie.missbarbell.co.uk
