Skip Menu |

This queue is for tickets about the TAP-Formatter-JUnit CPAN distribution.

Report information
The Basics
Id: 69777
Status: resolved
Priority: 0/
Queue: TAP-Formatter-JUnit

People
Owner: GTERMARS [...] cpan.org
Requestors: MARKF [...] cpan.org
Cc:
AdminCc:

Bug Information
Severity: Important
Broken in: (no value)
Fixed in: 0.10



Subject: Timing test occasionally fails
We smoke your module every night as part of our build system. Every so often it does this: Building TAP-Formatter-JUnit GTERMARS/TAP-Formatter-JUnit-0.08.tar.gz ./Build -- OK Running Build test t/formatter.t ......... ok t/passing-todos.t ..... ok t/pod-coverage.t ...... skipped: Test::Pod::Coverage 1.00 required for testing POD t/pod.t ............... skipped: Test::Pod 1.00 required for testing POD t/tap2junit-filter.t .. ok t/tap2junit-name.t .... ok t/tap2junit.t ......... ok # foo # bar # foobar # Failed test '... test timing: slept 2 seconds' # at t/timer.t line 118. # got: '1' # expected: '2' # Looks like you failed 1 test of 14. t/timer.t ............. Dubious, test returned 1 (wstat 256, 0x100) Failed 1/14 subtests I see from CPAN testers we're not the only ones that have this problem.
From: jeppe [...] ioslo.net
I can confirm this also happens in 0.09. Very rarely, but it happens. Creating new 'Build' script for 'TAP-Formatter-JUnit' version '0.09' Building TAP-Formatter-JUnit t/formatter.t ......... ok t/passing-todos.t ..... ok t/pod-coverage.t ...... ok t/pod.t ............... ok t/tap2junit-filter.t .. ok t/tap2junit-name.t .... ok t/tap2junit.t ......... ok # foo # bar # foobar # Failed test '... test timing: (teardown)' # at t/timer.t line 152. # got: '3' # expected: '2' # Looks like you failed 1 test of 21. t/timer.t ............. Dubious, test returned 1 (wstat 256, 0x100) Failed 1/21 subtests
Tests that were timing sensitive have been moved to "xt/", as they're primarily for my own benefit as opposed to seeing if the modules function correctly on other platforms. While not the most elegant way to handle it, this should correct the situation and allow builds/tests to pass successfully on a wider range of systems (in the past, virtual machines routinely caused problems with these timing tests).