Skip Menu |

This queue is for tickets about the CPAN-YACSmoke CPAN distribution.

Report information
The Basics
Id: 18188
Status: open
Priority: 0/
Queue: CPAN-YACSmoke

People
Owner: Nobody in particular
Requestors: jdhedden [...] cpan.org
Cc:
AdminCc:

Bug Information
Severity: Important
Broken in: (no value)
Fixed in: (no value)



Subject: Partial install of failed module
The following test report shows a failed test for the threads module: http://www.nntp.perl.org/group/perl.cpan.testers/298525 Because the test failed, the module should not have been installed. However, the tester's system immediately went on to test threads::shared: http://www.nntp.perl.org/group/perl.cpan.testers/298526 This test report show that threads.pm from the previous (failed) test WAS installed, but the matching binary portions of the module were not.
Subject: Cascading test failures
There is a problems with CPANPLUS whereby failures in one modules are causing failures of other modules. In 4 separate testing report: http://www.nntp.perl.org/group/perl.cpan.testers/298526 http://www.nntp.perl.org/group/perl.cpan.testers/300134 http://www.nntp.perl.org/group/perl.cpan.testers/300601 http://www.nntp.perl.org/group/perl.cpan.testers/300627 the tester's system first tests the threads module which fails. Then the system tests the threads::shared module. This fails as well, but don't because of a problem with the threads::shared module. In each case the failure reads something like: t/av_refs........threads object version 1.05 does not match bootstrap parameter 1.12 at /home/cwill1is/perl/lib/5.8.6/i686-linux-thread-multi-64int/DynaLoader.pm line 253. Compilation failed in require at t/av_refs.t line 37. This shows that CPANPLUS is causing Perl to try to load the threads module using threads.pm from the failed module, and the threads module binaries from the core Perl. This mismatch then results in the above vaulure which results in an erroneous failure report against threads::shared.
On Wed Mar 15 16:40:38 2006, JDHEDDEN wrote: Show quoted text
> The following test report shows a failed test for the threads module: > http://www.nntp.perl.org/group/perl.cpan.testers/298525 > Because the test failed, the module should not have been installed. > > However, the tester's system immediately went on to test threads::shared: > http://www.nntp.perl.org/group/perl.cpan.testers/298526 > This test report show that threads.pm from the previous (failed) test > WAS installed, but the matching binary portions of the module were not.
The failed module is not installed, as so much as it's path has been included in @INC for testing purposes. After a completed 'run' of commands, this @INC addition should be cleared. CPANPLUS offers an API call for this, which doesn't look like it's being used by CPAN::YACSmoke (the module that actually generated the test reports) a call to flush the relevant caches is needed, along these lines: $cb->flush( $_ ) for @caches relevant types are probably these (from perldoc CPANPLUS::Backend: ) ========================= =item * C<modules> Information about modules such as prerequisites and whether installation succeeded, failed, or was not attempted. =item * C<lib> This resets PERL5LIB, which is changed to ensure that while installing modules they are in our @INC. =item * C<load> This resets the cache of modules we've attempted to load, but failed. This enables you to load them again after a failed load, if they somehow have become available. ========================= I'll move this ticket to the cpan::yacsmoke queue, in the hope the author can address this issue. Thanks for reporting,
On Mon Mar 20 09:21:45 2006, JDHEDDEN wrote: Show quoted text
> There is a problems with CPANPLUS whereby failures in one modules are > causing failures of other modules. > > In 4 separate testing report: > http://www.nntp.perl.org/group/perl.cpan.testers/298526 > http://www.nntp.perl.org/group/perl.cpan.testers/300134 > http://www.nntp.perl.org/group/perl.cpan.testers/300601 > http://www.nntp.perl.org/group/perl.cpan.testers/300627 > the tester's system first tests the threads module which fails. Then > the system tests the threads::shared module. This fails as well, but > don't because of a problem with the threads::shared module. In each > case the failure reads something like:
This seems to be the same issue as you reported in: http://rt.cpan.org/Ticket/Display.html?id=18188 i will attempt to merge this ticket with the other one