| CC: | slaven [...] rezic.de |
| Subject: | Fwd: FAIL Gitalist-0.000002 amd64-freebsd 7.0-release |
| Date: | Thu, 10 Dec 2009 22:53:40 +0000 |
| To: | bug-Gitalist [...] rt.cpan.org |
| From: | Tomas Doran <bobtfish [...] bobtfish.net> |
As always, you are an _absolute legend_.
Thank you very much for the through and comprehensive investigation of
the issue!
Forwarding straight into rt.cpan.org :)
Many thanks
t0m
Begin forwarded message:
Show quoted text
> From: Slaven Rezic <slaven@rezic.de>
> Date: 10 December 2009 21:31:39 GMT
> To: Tomas Doran <bobtfish@bobtfish.net>
> Cc: srezic@cpan.org
> Subject: Re: FAIL Gitalist-0.000002 amd64-freebsd 7.0-release
> Reply-To: slaven@rezic.de
>
> Tomas Doran <bobtfish@bobtfish.net> writes:
>
>> On 8 Dec 2009, at 23:02, srezic@cpan.org wrote:
>>
>
> The first snapshot test in 03legacy_uri.t hangs. When turning
> IPCRUNDEBUG on, this command was listed last:
>
> IPC::Run 0004 [#7(16528) git]: execing /usr/local/bin/git --git-
> dir /mnt/i386/var/tmp/CPAN-build/Gitalist-0.000003-VHEIfI/t/lib/
> repositories/repo1 archive --format=tar --prefix=repo1/
> 145dc3ef5d307be84cb9b325d70bd08aeed0eceb
>
> and then just wait checks are following, without an end.
>
> It seems that Gitalist is doing something similar in
> Gitalist::Git::Util::run_cmd_fh like this one-liner:
>
> /usr/perl5.10.1/bin/perl -MSymbol=geniosym -MIPC::Run=run -e 'for
> ($in,$out,$err) { $_ = geniosym } warn run([qw(/usr/local/bin/git --
> git-dir /mnt/i386/var/tmp/CPAN-build/Gitalist-0.000003-VHEIfI/t/lib/
> repositories/repo1 archive --format=tar --prefix=repo1/
> 145dc3ef5d307be84cb9b325d70bd08aeed0eceb)], ">pipe", $out);'
>
> This one-liner hangs on my FreeBSD system, but not on an Linux system.
> Then I just replaced --format=tar with --format=zip on the FreeBSD
> system --- no hanging anymore. Then I changed the sample git
> repository
> with a larger one, producing a 5MB tar file on the Linux system --- it
> was also hanging there! So probably it's just a question of the
> system's
> buffer size if this command hangs or not. If the buffer size is large
> enough to hold git's output without the need to read it, then
> everything
> is OK.
>
> Maybe you have to use IPC::Run::start instead of IPC::Run::run?
>
> Regards,
> Slaven
>
> --
> Slaven Rezic - slaven <at> rezic <dot> de
>
> tksm - Perl/Tk program for searching and replacing in multiple
> files
> http://ptktools.sourceforge.net/#tksm
>>> Additional comments from tester:
>>>
>>> I have to kill the 03legacy_uri test, because it usually hangs on my
>>> system.
>>
>>
>> Hi..
>>
>> Thanks for the report!
>>
>> Can you tell me which test it hangs at, and/or provide output from
>> perl running that test file, especially with the GITALIST_DEBUG
>> environment variable set?