Bug #113695 for Test-NoLeaks: t/01-noleaks.t fails with recent Test::TCP

Tue Apr 12 02:10:29 2016 SREZIC [...] cpan.org - Ticket created

Subject:

t/01-noleaks.t fails with recent Test::TCP

There's a high probability (though not 100%) that t/01-noleaks.t fails if Test::TCP 2.15 is installed. The test failure looks like this: # Failed test 'non-closed sockets cause FD leaks report' # at t/01-noleaks.t line 73. # Looks like you failed 1 test of 9. t/01-noleaks.t ....... Dubious, test returned 1 (wstat 256, 0x100) Failed 1/9 subtests Statistical analysis of test reports on my smoker machines (negative theta is bad): **************************************************************** Regression 'mod:Test::TCP' **************************************************************** Name Theta StdErr T-stat [0='const'] 1.0000 0.1157 8.64 [1='eq_2.12'] 0.0000 0.2004 0.00 [2='eq_2.13'] 0.0000 0.1217 0.00 [3='eq_2.14'] 0.0000 0.1294 0.00 [4='eq_2.15'] -0.8571 0.1312 -6.53 R^2= 0.829, N= 37, K= 5 ****************************************************************

Wed Apr 13 02:26:13 2016 ivan [...] binary.com - Correspondence added

CC:	Torsten Förtsch <torsten [...] binary.com>
Subject:	Re: [rt.cpan.org #113695] t/01-noleaks.t fails with recent Test::TCP
Date:	Wed, 13 Apr 2016 09:25:59 +0300
To:	bug-Test-NoLeaks [...] rt.cpan.org
From:	Ivan Baidakou <ivan [...] binary.com>

Just checked with Test::TCP v2.16 (released a few hours ago) and all OK. It seems it was the leak in Test::TCP itself :) https://metacpan.org/changes/distribution/Test-TCP On Tue, Apr 12, 2016 at 9:10 AM, Slaven_Rezic via RT < bug-Test-NoLeaks@rt.cpan.org> wrote: Show quoted text

> Tue Apr 12 02:10:29 2016: Request 113695 was acted upon. > Transaction: Ticket created by SREZIC > Queue: Test-NoLeaks > Subject: t/01-noleaks.t fails with recent Test::TCP > Broken in: 0.04 > Severity: (no value) > Owner: Nobody > Requestors: SREZIC@cpan.org > Status: new > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=113695 > > > > There's a high probability (though not 100%) that t/01-noleaks.t fails if > Test::TCP 2.15 is installed. > > The test failure looks like this: > > # Failed test 'non-closed sockets cause FD leaks report' > # at t/01-noleaks.t line 73. > # Looks like you failed 1 test of 9. > t/01-noleaks.t ....... > Dubious, test returned 1 (wstat 256, 0x100) > Failed 1/9 subtests > > > Statistical analysis of test reports on my smoker machines (negative theta > is bad): > > **************************************************************** > Regression 'mod:Test::TCP' > **************************************************************** > Name Theta StdErr T-stat > [0='const'] 1.0000 0.1157 8.64 > [1='eq_2.12'] 0.0000 0.2004 0.00 > [2='eq_2.13'] 0.0000 0.1217 0.00 > [3='eq_2.14'] 0.0000 0.1294 0.00 > [4='eq_2.15'] -0.8571 0.1312 -6.53 > > R^2= 0.829, N= 37, K= 5 > **************************************************************** > >

-- WBR, Ivan Baidakou

Wed Apr 13 02:26:14 2016 The RT System itself - Status changed from 'new' to 'open'

Wed Apr 13 02:36:46 2016 SREZIC [...] cpan.org - Correspondence added

RT-Send-CC:

torsten [...] binary.com, ivan [...] binary.com

Confirmed, with Test::TCP 2.16 it looks better. So maybe only thing left is to make sure Test-NoLeaks is never built with a Test::TCP 2.15 --- e.g. by adding something like $test_tcp_minimum_version = 0; if (eval { require Test::TCP; $Test::TCP::VERSION == 2.15 }) { $test_tcp_minimum_version = '2.16'; } to Makefile.PL On 2016-04-13 02:26:13, ivan@binary.com wrote: Show quoted text

> Just checked with Test::TCP v2.16 (released a few hours ago) and all OK. It > seems it was the leak in Test::TCP itself :) > > https://metacpan.org/changes/distribution/Test-TCP > > On Tue, Apr 12, 2016 at 9:10 AM, Slaven_Rezic via RT < > bug-Test-NoLeaks@rt.cpan.org> wrote: >

> > Tue Apr 12 02:10:29 2016: Request 113695 was acted upon. > > Transaction: Ticket created by SREZIC > > Queue: Test-NoLeaks > > Subject: t/01-noleaks.t fails with recent Test::TCP > > Broken in: 0.04 > > Severity: (no value) > > Owner: Nobody > > Requestors: SREZIC@cpan.org > > Status: new > > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=113695 > > > > > > > There's a high probability (though not 100%) that t/01-noleaks.t fails if > > Test::TCP 2.15 is installed. > > > > The test failure looks like this: > > > > # Failed test 'non-closed sockets cause FD leaks report' > > # at t/01-noleaks.t line 73. > > # Looks like you failed 1 test of 9. > > t/01-noleaks.t ....... > > Dubious, test returned 1 (wstat 256, 0x100) > > Failed 1/9 subtests > > > > > > Statistical analysis of test reports on my smoker machines (negative theta > > is bad): > > > > **************************************************************** > > Regression 'mod:Test::TCP' > > **************************************************************** > > Name Theta StdErr T-stat > > [0='const'] 1.0000 0.1157 8.64 > > [1='eq_2.12'] 0.0000 0.2004 0.00 > > [2='eq_2.13'] 0.0000 0.1217 0.00 > > [3='eq_2.14'] 0.0000 0.1294 0.00 > > [4='eq_2.15'] -0.8571 0.1312 -6.53 > > > > R^2= 0.829, N= 37, K= 5 > > **************************************************************** > > > >

> >