Re: POE Upgrade error

Good morning.

I could use some help with this problem.

A third person reported the problem in IRC today, and six CPAN testers =
also reported it.  I haven't been able to reproduce the problem or make =
much progress on it.  I sent Patrick a public SSH key off-list, but I =
haven't heard back.  My e-mail may have been caught by a filter =
somewhere.

Here are the six failing CPAN reports.  Can anyone spot a pattern in =
them?

http://cpantesters.org/cpan/report/1f792eb6-ad79-11e1-8b69-15a10df65b4f
http://cpantesters.org/cpan/report/bf31699c-ad6e-11e1-9373-e8c10df65b4f
http://cpantesters.org/cpan/report/e865b668-ad6b-11e1-a563-f7010ef65b4f
http://cpantesters.org/cpan/report/53a384fe-ad68-11e1-9889-91d70df65b4f
http://cpantesters.org/cpan/report/1e015dcc-a303-11e1-89fa-d0960df65b4f
http://cpantesters.org/cpan/report/a783e948-a173-11e1-9d6f-f6dbfa7543f5

It's not a timing problem.  Tests are failing within a few seconds, not =
after a lengthy timeout.  Ignore the elapsed times on the right... =
dubious results don't have them.  System speeds also seem to vary, if =
the previous test's elapsed time is any indication.

[22:07:53] t/30_loops/select/wheel_readwrite.t ................ ok      =
945 ms
[22:07:54] t/30_loops/select/wheel_run.t ......................=20
Dubious, test returned 1 (wstat 256, 0x100)

[03:34:13] t/30_loops/io_poll/wheel_readwrite.t ............... ok     =
1982 ms
[03:34:16] t/30_loops/io_poll/wheel_run.t .....................=20
Dubious, test returned 1 (wstat 256, 0x100)

[04:01:41] t/30_loops/io_poll/wheel_readwrite.t ............... ok      =
964 ms
[04:01:42] t/30_loops/io_poll/wheel_run.t .....................=20
Dubious, test returned 1 (wstat 256, 0x100)

[04:23:11] t/30_loops/select/wheel_readwrite.t ................ ok     =
1081 ms
[04:23:13] t/30_loops/select/wheel_run.t ......................=20
Dubious, test returned 1 (wstat 256, 0x100)

[05:37:43] t/30_loops/select/wheel_readwrite.t ................ ok      =
847 ms
[05:37:45] t/30_loops/select/wheel_run.t ......................=20
Dubious, test returned 1 (wstat 256, 0x100)

Operating systems aren't necessarily an issue.  Neither are they all =
using 64bit integers.  Intel processors are a common thread, but they're =
also passing tests in droves.

osname=3Dlinux, osvers=3D2.6.18-1.2798.fc6, =
archname=3Di686-linux-thread-multi-64int-ld
osname=3Dlinux, osvers=3D2.6.17-1.2142_fc4, =
archname=3Di686-linux-thread-multi
osname=3Dlinux, osvers=3D2.6.18-1.2798.fc6, =
archname=3Di686-linux-thread-multi-64int-ld
osname=3Dlinux, osvers=3D2.6.18-1.2798.fc6, =
archname=3Di686-linux-thread-multi-64int-ld
osname=3Dlinux, osvers=3D2.6.18-1.2798.fc6, =
archname=3Di686-linux-thread-multi-64int-ld
osname=3Dopenbsd, osvers=3D5.1, archname=3DOpenBSD.i386-openbsd-64int

It's affecting a wide variety of Perl versions, all of which are also =
passing tests in droves:

Summary of my perl5 (revision 5 version 14 subversion 2) configuration
Summary of my perl5 (revision 5 version 15 subversion 2) configuration
Summary of my perl5 (revision 5 version 12 subversion 4) configuration
Summary of my perl5 (revision 5 version 14 subversion 2) configuration
Summary of my perl5 (revision 5 version 10 subversion 1) configuration
Summary of my perl5 (revision 5 version 14 subversion 1) configuration

Thanks!

--=20
Rocco Caputo <[email protected]>

On Jun 26, 2012, at 12:14, Tony Wildish wrote:

> Hi,
>=20
>  I don't know if it helps any, but I saw this same error myself a few=20=

> days ago while installing on a system that had some baggage on it.=20
> Clearing up (rm -rf) and installing from scratch made the problem go =
away.
>=20
>  The 'baggage' in question comes from the fact that I don't have root=20=

> access to the system I was installing on, so I install in my own home=20=

> directory. The system has been updated a few times since, and I was=20
> updating in case I'd fallen behind. So it could have been some=20
> incompatibility that wasn't managed properly, or something like that.
>=20
>  I'm afraid I can't give much more detail than that, I didn't really=20=

> pay much attention to it since it went away.
>=20
>  Cheers,
>  Tony.
>=20
> On 6/25/12 7:43 PM, Patrick Amigo wrote:
>> On Jun 25, 2012, at 12:02, Patrick Amigo wrote:
>>=20
>>> Hello,
>>>=20
>>>=20
>>>=20
>>> We are trying to  upgrade POE version 1.289 to current version POE =
1.354.
>>> However, we get the following error on every attempt.
>>>=20
>>>=20
>>>=20
>>> t/30_loops/io_poll/wheel_run.t ..................... 1/103
>>>=20
>>> #   Failed test 'stdio/redirection'
>>>=20
>>> #   at /usr/lib/perl5/site_perl/5.8.8/POE/Test/Loops/wheel_run.pm =
line
>> 597.
>>>=20
>>> #                   'CHILD:'
>>>=20
>>> #     doesn't match '(?-xism:CHILD:PARENT:CHILD:)'