Skip Menu |

This queue is for tickets about the lib-filter CPAN distribution.

Report information
The Basics
Id: 117117
Status: open
Priority: 0/
Queue: lib-filter
People
Owner: Nobody in particular
Requestors: SREZIC [...] cpan.org
Cc:
AdminCc:
Bug Information
Severity: (no value)
Broken in: 0.25
Fixed in: (no value)

History Show all quoted text
  Mon Aug 22 01:22:46 2016 SREZIC [...] cpan.org - Ticket created
Subject: Random test failures
I can just point to http://matrix.cpantesters.org/?dist=lib-filter%200.25 but give no more information. The two fail reports generated on my smokers (to be delivered in a few hours) look different then the present ones on CPAN Testers: there seem to be a busy loop in t/filter.t, eventually failing when exceeding the configured CPU limit (1h). And it's random, I cannot reproduce it with every run.
  Tue Aug 23 10:51:27 2016 PERLANCAR [...] cpan.org - Correspondence added
On Mon, 22 Aug 2016 05:22:46 GMT, SREZIC wrote: Show quoted text
> I can just point to http://matrix.cpantesters.org/?dist=lib- > filter%200.25 but give no more information. > > The two fail reports generated on my smokers (to be delivered in a few > hours) look different then the present ones on CPAN Testers: there > seem to be a busy loop in t/filter.t, eventually failing when > exceeding the configured CPU limit (1h). And it's random, I cannot > reproduce it with every run.
I'm wild guessing that the CPU limit might be caused by a diagnostic routine which tries to find the path of all installed modules using PERLANCAR::Module::List, as some machines might have a lot of modules installed plus there might be slow network disks or weird filesystems? I'm removing that code with just a code that tries to find the path of single modules. This won't eliminate all failures, as I found that some machines lack the "base.pm" core module, which is assumed to be installed by the test. Release 0.26.
  Tue Aug 23 10:51:28 2016 The RT System itself - Status changed from 'new' to 'open'
  Tue Aug 23 22:34:58 2016 PERLANCAR [...] cpan.org - Correspondence added
As for the failure when testing related to core module 'base', the problem is that base is a dual-live module. So the path to the base.pm can be core path or non-core path, depending on which version of base.pm got installed. Example: http://www.cpantesters.org/cpan/report/0d039ed0-694d-11e6-9451-9b92aab8e0c0 So the test needs to be changed to use a non-dual-live module. I'm now picking overload.pm. Released 0.26.