>>>>> On Mon, 24 Dec 2007 17:26:47 +0000, David Cantrell <[EMAIL PROTECTED]>
>>>>> said:
dc> Michael G Schwern wrote:
>> [1] It can be argued that bleadperl testers should probably not email
>> authors
I tend to agree.
dc> I'd argue that they should, as problems found testing against bleadperl
dc> seem to end up being problems in the next stable release.
The argument is not about the usefulness of the reports per se, just
about the practice to send CPAN authors unsolicited email.
dc> Personally
dc> I'd prefer to at least have the opportunity to fix my modules before
dc> ordinary users (who never touch a dev perl) will ever see the problem.
As author you have already the option to fetch the reports from
cpantesters either by visiting the site or by using MARCEL's
App::sync_cpantesters. Perhaps needs more propagation.
dc> Additionally, results sent to the cpan-testers list are largely
dc> un-monitored AFAIK - it's really just a convenient way of feeding nntp
dc> and the various webby tools. If by testing a module we find an
dc> honest-to-goodness bug in bleadperl, it's far more likely to get noticed
dc> and reported to p5p if the *module* author is notified cos he's the one
dc> who is most likely to be paying attention.
This is rather an argument against sending because the author has the
right to be spared from the utter nonsense[1] that happens in
bleadperl and in a separate thread it has already been confirmed that
this is not desireable.
dc> And no, p5p definitely shouldn't be fed cpan-testers failure reports
dc> from dev perls directly! Even ignoring the load on the perl.org mail
dc> server, I can bet you that most people would just procmail them to
dc> /dev/null because most are irrelevant to p5p - they're the same bugs in
dc> modules that are also caught on stable perls.
You hit the nail on the head: we need better monitoring of the test
reports. But this doesn't necessarily mean that we have the right to
send the reports to the authors.
--
andreas
[1] not that it happens every day, but every now and then a patch
breaks lots of modules at once.
