Skip Menu |

This queue is for tickets about the Apache-ParseLog CPAN distribution.

Report information
The Basics
Id: 42838
Status: open
Priority: 0/
Queue: Apache-ParseLog

People
Owner: Nobody in particular
Requestors: bob [...] bobcatos.com
Cc:
AdminCc:

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



Subject: Apache::ParseLog fails test on Perl 5.10.0
Date: Tue, 27 Jan 2009 14:17:06 -0600 (CST)
To: bug-Apache-ParseLog [...] rt.cpan.org
From: Bob McClure <bob [...] bobcatos.com>
Details: Apache::ParseLog v1.02 This is perl, v5.10.0 built for x86_64-linux-thread-multi Linux bobcat.bobcatos.com 2.6.27.9-73.fc9.x86_64 #1 SMP Tue Dec 16 14:54:03 EST 2008 x86_64 x86_64 x86_64 GNU/Linux t/08_custom...... Failed 1/11 subtests Digging into the test I find that test 5 (user) fails. Poking some instrumentation into it, I find that it counts 29 rather than the expected 30 entries in $user{'-'}. Looking at http://cpantesters.perl.org/show/Apache-ParseLog.html, I see that it regularly fails in Perl v5.{9,10,11}.*. On another server with Perl v5.8.8, it built, tested, and installed just fine. If you don't have a Perl v5.10 system to debug on, send me some suggestions and I will be happy to do some debugging for you. Cheers, -- Bob McClure, Jr. Bobcat Open Systems, Inc. bob@bobcatos.com http://www.bobcatos.com "Blessed are those who hunger and thirst for righteousness, for they will be filled." Matthew 5:6 (NIV)
Binary search reveals that it was patch 29279 that marks the changing behaviour. http://public.activestate.com/cgi-bin/perlbrowse?patch_num=29279&show_patch=Show+Patch Or in the git repo it is c74340f9cdee6010339b6bfd0e8b0dc8bc875344 I stepped through the test script in the debugger and compared current perl with an old perl. I think the major bug is in Apache::ParseLog in that it does not check if the match in line 1182 is succeeding. It continues running for the broken logfile line 18 and the empty line 31. Because the capturing variables were not properly reset in older perls the empty logfile line was counted by Apache::ParseLog with $USER eq "-". So the fix in perl reveals a bug in Apache::ParseLog.