On Tue, Jan 19, 2010 at 12:47 PM, David Cantrell <da...@cantrell.org.uk> wrote: > On Mon, Jan 18, 2010 at 09:53:38PM +0100, Andreas J. Koenig wrote: >> On Sat, 16 Jan 2010 22:02:46 +0100, j...@cpan.org (Johan Vromans) said: >> > http://www.nntp.perl.org/group/perl.cpan.testers/2010/01/msg6656308.html >> > I therefore must assume it's a TESTER problem. >> Yes, we're suffering already for a few days but today Barbie fixed the >> testers master database by hiding the broken reports that come from a >> broken and so far unstopped broken test machine. > > But this is one of Chris's reports, not APOCAL's.
Yes -- that pretty clearly should have been discarded and not sent at all. There's something about 'prereq_data' not being recognized. Aha! EekBoek bundles M::B 0.28, which didn't have 'prereq_data'. Chris -- I think CPANPLUS needs to check the exit code of 'Build prereq_data' and fall back on scraping _build/prereqs if it's not OK. Never underestimate the ability of authors/users to find the wrong module, I guess. :-) Johan -- Would you please consider updating the bundled M::B? You might want to see if the instructions in Module::Build::Bundling are sufficient to automate it for you easily. David