Bug #103457 for Signals-XSIG: t/02-SIGhash.t sometimes fails

Sat Apr 11 05:53:55 2015 SREZIC [...] cpan.org - Ticket created

Subject:

t/02-SIGhash.t sometimes fails

A pass/fail overview may be seen here: http://analysis.cpantesters.org/reports_by_field?distv=Signals-XSIG-0.13;field=fail%3At%2F02-SIGhash.t

Tue Nov 22 12:17:45 2016 MOB [...] cpan.org - Correspondence added

This is mysterious and intermittent. I'll add more diagnostics to the tests and monitor the CPAN testers reports.

Tue Nov 22 12:17:45 2016 The RT System itself - Status changed from 'new' to 'open'

Wed Apr 26 01:26:57 2017 SREZIC [...] cpan.org - Correspondence added

On 2016-11-22 12:17:45, MOB wrote: Show quoted text

> This is mysterious and intermittent. I'll add more diagnostics to the > tests and monitor the CPAN testers reports.

It happened again on one of my smokers: ... # Failed test 'SIGUSR1(1) handler undef after %SIG cleared' # at t/02-SIGhash.t line 24. # Defined handlers: USR1 # Looks like you failed 1 test of 14. t/02-SIGhash.t ............ Dubious, test returned 1 (wstat 256, 0x100) Failed 1/14 subtests ...

Wed Apr 26 01:27:05 2017 SREZIC [...] cpan.org - Broken in 0.14 added

Wed Apr 26 15:51:20 2017 MOB [...] cpan.org - Correspondence added

Well then I'm sorry I haven't uploaded my additional diagnostics for this issue and we missed this opportunity. Watch for 0.15_01 coming soon. -- Marty On Wed Apr 26 01:26:57 2017, SREZIC wrote: Show quoted text

> On 2016-11-22 12:17:45, MOB wrote:

> > This is mysterious and intermittent. I'll add more diagnostics to the > > tests and monitor the CPAN testers reports.

> > It happened again on one of my smokers: > > ... > > # Failed test 'SIGUSR1(1) handler undef after %SIG cleared' > # at t/02-SIGhash.t line 24. > # Defined handlers: USR1 > # Looks like you failed 1 test of 14. > t/02-SIGhash.t ............ > Dubious, test returned 1 (wstat 256, 0x100) > Failed 1/14 subtests > ...

Thu Apr 27 02:34:15 2017 SREZIC [...] cpan.org - Correspondence added

CPAN Testers is slow these days, so I am attaching a fail report with the additional diagnostics. -- Slaven On 2017-04-26 15:51:20, MOB wrote: Show quoted text

> Well then I'm sorry I haven't uploaded my additional diagnostics for > this issue and we missed this opportunity. > Watch for 0.15_01 coming soon. > > -- Marty > > > On Wed Apr 26 01:26:57 2017, SREZIC wrote:

> > On 2016-11-22 12:17:45, MOB wrote:

> > > This is mysterious and intermittent. I'll add more diagnostics to > > > the > > > tests and monitor the CPAN testers reports.

> > > > It happened again on one of my smokers: > > > > ... > > > > # Failed test 'SIGUSR1(1) handler undef after %SIG cleared' > > # at t/02-SIGhash.t line 24. > > # Defined handlers: USR1 > > # Looks like you failed 1 test of 14. > > t/02-SIGhash.t ............ > > Dubious, test returned 1 (wstat 256, 0x100) > > Failed 1/14 subtests > > ...

Subject:

fail.Signals-XSIG-0.1501.amd64-freebsd.10.1-release.1493249418.74530.rpt

Download fail.Signals-XSIG-0.1501.amd64-freebsd.10.1-release.1493249418.74530.rpt
application/octet-stream 43.5k

Message body not shown because it is not plain text.

Thu Apr 27 12:21:56 2017 MOB [...] cpan.org - Correspondence added

Thank you, Slaven. You are really going above and beyond the call of duty here. Signals-XSIG-0.15_02 has been uploaded, which just might make this problem go away. -- Marty On Thu Apr 27 02:34:15 2017, SREZIC wrote: Show quoted text

> CPAN Testers is slow these days, so I am attaching a fail report with > the additional diagnostics. > > -- Slaven > > On 2017-04-26 15:51:20, MOB wrote:

> > Well then I'm sorry I haven't uploaded my additional diagnostics for > > this issue and we missed this opportunity. > > Watch for 0.15_01 coming soon. > > > > -- Marty > > > >

Fri Apr 28 01:35:59 2017 SREZIC [...] cpan.org - Correspondence added

Unfortunately not, see the attached file. On 2017-04-27 12:21:56, MOB wrote: Show quoted text

> Thank you, Slaven. You are really going above and beyond the call of > duty here. > > Signals-XSIG-0.15_02 has been uploaded, which just might make this > problem go away. > > -- Marty > > > > On Thu Apr 27 02:34:15 2017, SREZIC wrote:

> > CPAN Testers is slow these days, so I am attaching a fail report with > > the additional diagnostics. > > > > -- Slaven > > > > On 2017-04-26 15:51:20, MOB wrote:

> > > Well then I'm sorry I haven't uploaded my additional diagnostics > > > for > > > this issue and we missed this opportunity. > > > Watch for 0.15_01 coming soon. > > > > > > -- Marty > > > > > >

Subject:

fail.Signals-XSIG-0.1502.x86_64-linux.3.16.0-4-amd64.1493341057.30464.rpt

Download fail.Signals-XSIG-0.1502.x86_64-linux.3.16.0-4-amd64.1493341057.30464.rpt
application/octet-stream 68.7k

Message body not shown because it is not plain text.

Sat Apr 29 14:56:08 2017 SREZIC [...] cpan.org - Correspondence added

A fail report 0.15_03 is attached (current ratio on my smokers: 65 pass vs. 7 fail). On 2017-04-28 01:35:59, SREZIC wrote: Show quoted text

> Unfortunately not, see the attached file. > > On 2017-04-27 12:21:56, MOB wrote:

> > Thank you, Slaven. You are really going above and beyond the call of > > duty here. > > > > Signals-XSIG-0.15_02 has been uploaded, which just might make this > > problem go away. > > > > -- Marty > > > > > > > > On Thu Apr 27 02:34:15 2017, SREZIC wrote:

> > > CPAN Testers is slow these days, so I am attaching a fail report with > > > the additional diagnostics. > > > > > > -- Slaven > > > > > > On 2017-04-26 15:51:20, MOB wrote:

> > > > Well then I'm sorry I haven't uploaded my additional diagnostics > > > > for > > > > this issue and we missed this opportunity. > > > > Watch for 0.15_01 coming soon. > > > > > > > > -- Marty > > > > > > > >

> >

Subject:

fail.Signals-XSIG-0.1503.x86_64-linux-thread-multi.4.8.0-2-amd64.1493457656.16807.rpt

Download fail.Signals-XSIG-0.1503.x86_64-linux-thread-multi.4.8.0-2-amd64.1493457656.16807.rpt
application/octet-stream 58.6k

Message body not shown because it is not plain text.

Sat Apr 29 17:02:47 2017 MOB [...] cpan.org - Correspondence added

OK, I now have a FreeBSD virtual machine that exhibits this problem. Like you Slaven, I see it about 10% of the time. There's a line $XSIG{$_}[0] = undef for keys %_XSIG in the Signals::XSIG::TieSig::CLEAR function. I find that if the very first key returns by 'keys' is 'USR1' or 'USR2', then the test will fail. If the first key is some other signal name, it succeeds (though I didn't try them all). If I change the test to use '__WARN__' or '__DIE__' as the test signals ($s1 and $s2 in t/02-SIGhash.t), and arrange for '__WARN__' and '__DIE__' to be the first signals in the for loop list, then the test also succeeds. And if that works, I have no idea why. On Sat Apr 29 14:56:08 2017, SREZIC wrote: Show quoted text

> A fail report 0.15_03 is attached (current ratio on my smokers: 65 > pass vs. 7 fail). > > On 2017-04-28 01:35:59, SREZIC wrote:

> > Unfortunately not, see the attached file. > > > > On 2017-04-27 12:21:56, MOB wrote:

> > > Thank you, Slaven. You are really going above and beyond the call > > > of > > > duty here. > > > > > > Signals-XSIG-0.15_02 has been uploaded, which just might make this > > > problem go away. > > > > > > -- Marty > > > > > > > > > > > > On Thu Apr 27 02:34:15 2017, SREZIC wrote:

> > > > CPAN Testers is slow these days, so I am attaching a fail report > > > > with > > > > the additional diagnostics. > > > > > > > > -- Slaven > > > > > > > > On 2017-04-26 15:51:20, MOB wrote:

> > > > > Well then I'm sorry I haven't uploaded my additional > > > > > diagnostics > > > > > for > > > > > this issue and we missed this opportunity. > > > > > Watch for 0.15_01 coming soon. > > > > > > > > > > -- Marty > > > > > > > > > >

> > > >