| Subject: | t/05_extended.t failures for some core perl releases |
I just attempted to install MooseX-App 1.16 and encountered the following test suite failure:
[2013-03-04 01:33:23]
[2013-03-04 01:33:23] # Failed test 'Message ok'
[2013-03-04 01:33:23] # at t/05_extended.t line 124.
[2013-03-04 01:33:23] # 'Could be
[2013-03-04 01:33:23] # custom2
[2013-03-04 01:33:23] # custom1 '
[2013-03-04 01:33:23] # doesn't match '(?-xism:Could be
[2013-03-04 01:33:23] # custom1
[2013-03-04 01:33:23] # custom2)'
[2013-03-04 01:33:23] # Looks like you failed 1 test of 3.
[2013-03-04 01:33:23]
[2013-03-04 01:33:23] # Failed test 'Test ambiguous options'
[2013-03-04 01:33:23] # at t/05_extended.t line 127.
[2013-03-04 01:33:23] # Looks like you failed 1 test of 16.
[2013-03-04 01:33:23] t/05_extended.t ........
This occurs for an odd subset of the platforms and compilers I build it for. Here's the status page from my build
system, which installs things for perl 5.10 through 5.16, on 4 distinct binary platforms (32 and 64 bit RHEL 5 and 6):
EFS Deploy State for perl5/MooseX-App/1.16-build001:
Platform/Compiler source depends configure build test install verify archive
==========================================================================
===============================
x86-32.rhel.5/5.10 pass[2m,27s] pass[2m,59s] pass[36s] pass[31s] fail[58s] none none none
x86-32.rhel.5/5.12 pass[36s] pass[58s] pass[32s] pass[22s] pass[39s] none none none
x86-32.rhel.5/5.14 pass[36s] pass[52s] pass[32s] pass[22s] pass[40s] none none none
x86-32.rhel.5/5.16 pass[2m,26s] pass[2m,56s] pass[37s] pass[29s] pass[55s] none none none
x86-32.rhel.6/5.10 pass[44s] pass[1m,11s] pass[29s] pass[19s] fail[42s] none none none
x86-32.rhel.6/5.12 pass[34s] pass[49s] pass[26s] pass[17s] pass[31s] none none none
x86-32.rhel.6/5.14 pass[36s] pass[47s] pass[27s] pass[17s] pass[32s] none none none
x86-32.rhel.6/5.16 pass[34s] pass[49s] pass[27s] pass[16s] pass[31s] none none none
x86-64.rhel.5/5.10 pass[2m,24s] pass[3m,1s] pass[50s] pass[29s] fail[59s] none none none
x86-64.rhel.5/5.12 pass[38s] pass[1m,6s] pass[24s] pass[14s] pass[26s] none none none
x86-64.rhel.5/5.14 pass[41s] pass[50s] pass[30s] pass[21s] fail[40s] none none none
x86-64.rhel.5/5.16 pass[2m,27s] pass[3m,33s] pass[43s] pass[26s] pass[55s] none none none
x86-64.rhel.6/5.10 pass[44s] pass[1m,6s] pass[28s] pass[19s] fail[42s] none none none
x86-64.rhel.6/5.12 pass[34s] pass[46s] pass[22s] pass[11s] pass[22s] none none none
x86-64.rhel.6/5.14 pass[35s] pass[49s] pass[27s] pass[16s] fail[36s] none none none
x86-64.rhel.6/5.16 pass[35s] pass[49s] pass[27s] pass[15s] pass[30s] none none none
The failures are for 5.10 on all platforms, but for 5.14 on the 64bit platforms. FWIW, the actual releases of perl used are
5.10.1, 5.12.4, 5.14.3, and 5.16.2.
Give the nature of the error, I wonder if the root cause is something like the hacking algorithm sorting keys differently by
default on certain platforms. I have not looked into the code at all, so that's shooting from the hip.
In any case, we're skipping this release, and using 1.09, which passed all tests on all platforms. If you would like us to try
out patches that might correct the problem, let me know, and I'll be happy to give them a shot.