Skip Menu |

Preferred bug tracker

Please visit the preferred bug tracker to report your issue.

This queue is for tickets about the Test-File CPAN distribution.

Report information
The Basics
Id: 91862
Status: open
Priority: 0/
Queue: Test-File
People
Owner: BAREFOOT [...] cpan.org
Requestors: ether [...] cpan.org
Cc:
AdminCc:
Bug Information
Severity: (no value)
Broken in: 1.36
Fixed in: (no value)

History Show all quoted text
  Sat Jan 04 19:28:44 2014 ether [...] cpan.org - Ticket created
Subject: Test failures

Message body is not shown because it is too large.

  Tue Jun 03 03:06:39 2014 dwheeler [...] cpan.org - Correspondence added
I'm seeing the same thing, making my Travis tests fail. :-( https://s3.amazonaws.com/archive.travis-ci.org/jobs/26616619/log.txt Will work around it for now.
  Tue Jun 03 03:06:40 2014 The RT System itself - Status changed from 'new' to 'open'
  Wed Jun 04 20:16:07 2014 BAREFOOT [...] cpan.org - Correspondence added
Hey, sorry for dropping the ball on this one guys. Show quoted text
>> This may be a duplicate/reoccurence of #89849...
No, that doesn't appear to be the case. The files that it's trying to test just aren't there. Now, I see a potential race condition in the way that the code checks for the presence of the test files. And I can see from Show quoted text
> https://s3.amazonaws.com/archive.travis-ci.org/jobs/26616619/log.txt
that David at least is running his tests in parallel (i.e. using -jX). I'm guessing ether is as well. Now, why this would only show up in 5.19.x ... who can say? Perhaps some op was optimized to the point where Perl can complete an entire test file in one process in the time between the other process creating the directory and creating the test files? That's just a guess. Anyway, could I ask one (or both) of you to try it with the -j switch turned off to see if that fixes the problem? If so, I'll see if I can figure out a way to set up the common files a bit more gracefully. But I'd like to know that my theory is correct before heading down that road. Thx!
  Wed Jun 04 20:16:08 2014 BAREFOOT [...] cpan.org - Taken
  Wed Jun 04 22:17:01 2014 ether [...] cpan.org - Correspondence added
On 2014-06-04 17:16:07, BAREFOOT wrote: Show quoted text
> Hey, sorry for dropping the ball on this one guys.
Are you receiving cpantesters reports? By default, they are sent out in a batch, daily. Your settings can be adjusted here: https://prefs.cpantesters.org/ Show quoted text
> ..that David at least is running his tests in parallel (i.e. using -jX). > I'm guessing ether is as well.
Yes, I have HARNESS_OPTIONS=j9 in my .bashrc. And, when I undefine HARNESS_OPTIONS, the tests pass. Show quoted text
> Now, why this would only show up in > 5.19.x ... who can say?
Are you sure it's only the case in 5.19.x? I just tried with 5.18.2 and saw precisely the same results - test failures with -j9, no failures without. Never forget about Occam's Razor :) Show quoted text
> Anyway, could I ask one (or both) of you to try it with the -j switch > turned off to see if that fixes the problem? If so, I'll see if I can > figure out a way to set up the common files a bit more gracefully.
File::Temp will give you a temporary directory with a unique name, so its presence won't interfere with other tests running simultaneously.
  Tue Jun 10 03:55:46 2014 BAREFOOT [...] cpan.org - Correspondence added
Show quoted text
> Are you receiving cpantesters reports? By default, they are sent out > in a batch, daily. Your settings can be adjusted here: > https://prefs.cpantesters.org/
Yes, ether, I receive CPAN Testers reports. And I know how to change my settings. Thank you. Show quoted text
> Are you sure it's only the case in 5.19.x?
Nope. Just haven't seen it in any other versions yet. Or didn't recognize it as the same issue if I did. Show quoted text
> File::Temp will give you a temporary directory with a unique name, so > its > presence won't interfere with other tests running simultaneously.
Yes, I'm familiar with File::Temp. I presume brian d foy didn't want to create a bunch of different directories that all had the same files in them, although that was done before my time. But that's one way to fix it, true. I'll try to take a look soon and figure out what the best solution is.
  Tue Jun 10 10:24:18 2014 bdfoy [...] cpan.org - Correspondence added
Subject: Re: [rt.cpan.org #91862] Test failures
Date: Tue, 10 Jun 2014 10:23:28 -0400
To: bug-Test-File [...] rt.cpan.org
From: brian d foy <bdfoy [...] cpan.org>
When I started this module, parallel testing wasn't really a thing so I didn't worry about it. It's a bit more complicated to recreate a test directory for every test file, but File::Temp would be the way to do it. I used to ask for a way for a module to mark itself as not parallelizable so the harness could handle these situations. I still think that's a good idea, but I don't think it will ever happen.
  Tue Jun 10 11:29:01 2014 ether [...] cpan.org - Correspondence added
On 2014-06-10 07:24:18, BDFOY wrote: Show quoted text
> I used to ask for a way for a module to mark itself as not > parallelizable so the harness could handle these situations. I still > think that's a good idea, but I don't think it will ever happen.
Test::Harness can handle this, via its "rulesfile".
  Tue Jun 10 11:30:16 2014 ether [...] cpan.org - Correspondence added
On 2014-06-10 08:29:01, ETHER wrote: Show quoted text
> On 2014-06-10 07:24:18, BDFOY wrote: >
> > I used to ask for a way for a module to mark itself as not > > parallelizable so the harness could handle these situations. I still > > think that's a good idea, but I don't think it will ever happen.
> > Test::Harness can handle this, via its "rulesfile".
er TAP::Harness (Test::Harness is the legacy version).
  Tue Jun 10 15:30:25 2014 bdfoy [...] cpan.org - Correspondence added
Subject: Re: [rt.cpan.org #91862] Test failures
Date: Tue, 10 Jun 2014 15:29:36 -0400
To: bug-Test-File [...] rt.cpan.org
From: brian d foy <bdfoy [...] cpan.org>
That's for version 3.31, released a couple of days ago.
  Tue Jun 10 16:05:31 2014 ether [...] cpan.org - Correspondence added
Subject: Re: [rt.cpan.org #91862] Test failures
Date: Tue, 10 Jun 2014 13:05:17 -0700
To: brian d foy via RT <bug-Test-File [...] rt.cpan.org>
From: Karen Etheridge <ether [...] cpan.org>
On Tue, Jun 10, 2014 at 03:30:26PM -0400, brian d foy via RT wrote: Show quoted text
> <URL: https://rt.cpan.org/Ticket/Display.html?id=91862 > > That's for version 3.31, released a couple of days ago.
Interesting! I'd known about it since last year at least, so I didn't realize that it only just made it into a release.
  Wed Sep 10 23:48:11 2014 bdfoy [...] cpan.org - Correspondence added
This should be fixed in 1.39