On Tue, Sep 02, 2008 at 05:42:17PM -0700, chromatic wrote: > On Tuesday 02 September 2008 16:53:31 Barbie wrote: > > > > http://www.nntp.perl.org/group/perl.cpan.testers/2008/08/msg2029059.html > > > Please note that I don't care *why* this occurred. > > > But you're willing to shoot down cpan-testers just because they're the > > messenger! Thanks. > > I believe that someone interested in reporting test results for CPAN > distributions should use working tools which do not spew useless and > incorrect information into my inbox and on search.cpan.org's distribution > pages.
CPANPLUS is a working tool, however it has made a design decision to try and be helpful to the user, if there is a build tool that is unavailable. That design decision obviously doesn't work for you, but it does for others. cpan-testers are reporting what a user would experience using CPANPLUS, but not having Module::Build, and having problems with installing Module::Build. While you may not be interested in being made aware of those potential failures, others are. > I didn't ask for anyone to set up automatic tests for my distributions. I'm > happy to get useful information out of them -- but when there's an automated > process seemingly spiraling out of control with a configuration that couldn't > possibly tell the difference between a pass and a fail, you'd better believe > that there's something wrong Personally I'm not seeing this spiral out of control, and the configuration doesn't know that *you* don't want to know about this potential issue. If the design decision is wrong, which should hopefully be a non-issue once Perl 5.10 is in use everywhere, then discussing this with the CPANPLUS developers is required. Judging from a further thread, David Golden is looking at writing a patch to reverse this design decision in CPANPLUS. > Like I said, I don't care where the problem is. I don't use CPANPLUS. I'm > happy with people complaining to CPANPLUS if that's where the bug is. I'm > not happy with getting FAIL reports from CPAN Testers clients that can't > possibly produce PASS reports. This is not my bug. No it's not necessarily a bug. It's a design decision you don't agree with. Personally I would be interested to know that this situation may arise, and potentially adjust how I inform the user that they may have to install Module::Build by hand (as that is often the only way to install it). > Notifying me accomplishes precisely nothing positive. But for others it does. Currently we are unable to know an authors personal preference for results like this, and until the work currently on-going to try and allow for that is available, there are likely to be some reports that you personally are not going to find useful. > Putting a little black mark on search.cpan.org > for my distributions misleads my potential users into believing that I don't > write good code. That's a big mistaken leap of understanding. A FAIL report in most cases has nothing to do with not writing good code. In the vast majority of cases it highlights that the environment it is being installed into is different to what the author expects, due to not having that environment available to them when developing the code. > Short of opting out of CPAN Testers entirely (and I'm not looking forward to > updating and uploading new versions of 30+ distributions to do so) and asking > CPAN Testers to fix the brokenness over the period of a couple of years, > exactly what do you recommend I do? If there is a specific tester that is raising these kinds of reports, ask them to add you or the distribution to an exclusion list. If you have a problem with a design decision, highlight that the design decision is incorrect and discuss with the appropriate parties to get it changed. > What are you supposed to do about it? Ask in the appropriate places (as listed in the reports), to understand *why* this is failing. > If that's worthy of sarcastic dismissal ... I can't see how it helps > anything though. Agreed. However, I found your response to the situation equally non-constructive. Ignoring *why* this got reported and deriding the efforts of cpan-testers doesn't help either. Barbie. -- Birmingham Perl Mongers <http://birmingham.pm.org> Memoirs Of A Roadie <http://barbie.missbarbell.co.uk>
