Bug #105030 for Time-Piece: 02core_dst.t two failures in 1.30 in irix

Sat Jun 06 21:31:55 2015 JHI [...] cpan.org - Ticket created

Subject:

02core_dst.t two failures in 1.30

With blead's Time-Piece (1.30) two subtests of 02core_dst.t fail in IRIX 6.5.1 (all other tests pass): ./perl -Ilib cpan/Time-Piece/t/02core_dst.t ... ok 56 not ok 57 # Failed test at cpan/Time-Piece/t/02core_dst.t line 128. # '%z' # doesn't match '(?^:-0400|EDT)' ok 58 ok 59 not ok 60 # Failed test at cpan/Time-Piece/t/02core_dst.t line 133. # '%z' # doesn't match '(?^:-0500|EST)' # Looks like you failed 2 tests of 60. If you want to skip the tests in IRIX, $^O is simply 'irix'.

Sun Jun 07 18:08:35 2015 ESAYM [...] cpan.org - Correspondence added

On Sat Jun 06 21:31:55 2015, JHI wrote: Show quoted text

> With blead's Time-Piece (1.30) two subtests of 02core_dst.t fail in > IRIX 6.5.1 (all other tests pass): > > ./perl -Ilib cpan/Time-Piece/t/02core_dst.t > ... > ok 56 > not ok 57 > # Failed test at cpan/Time-Piece/t/02core_dst.t line 128. > # '%z' > # doesn't match '(?^:-0400|EDT)' > ok 58 > ok 59 > not ok 60 > # Failed test at cpan/Time-Piece/t/02core_dst.t line 133. > # '%z' > # doesn't match '(?^:-0500|EST)' > # Looks like you failed 2 tests of 60. > > If you want to skip the tests in IRIX, $^O is simply 'irix'.

Thank you for the report. Some of the tests in 02core_dst.t have shown to be somewhat fragile. I am probably going to have to modify them or simply limit them to Linux hosts. I will push out a fix in the next week or two.

Sun Jun 07 18:08:36 2015 The RT System itself - Status changed from 'new' to 'open'

Sun Jun 07 18:12:03 2015 jhi [...] iki.fi - Correspondence added

Subject:	Re: [rt.cpan.org #105030] 02core_dst.t two failures in 1.30
Date:	Sun, 07 Jun 2015 18:11:50 -0400
To:	bug-Time-Piece [...] rt.cpan.org, JHI [...] cpan.org
From:	Jarkko Hietaniemi <jhi [...] iki.fi>

On Sunday-201506-07 18:08, Samuel Smith via RT wrote: Show quoted text

> Thank you for the report. Some of the tests in 02core_dst.t have shown to be somewhat fragile. I am probably going to have to modify them or simply limit them to Linux hosts.

Thanks. Take a look at the testers results? I would suggest trying to leave in at least FreeBSD (assuming it passes...), to get a little bit more coverage.

Sun Jun 07 20:21:03 2015 jhi [...] iki.fi - Correspondence added

Subject:	Re: [rt.cpan.org #105030] 02core_dst.t two failures in 1.30
Date:	Sun, 07 Jun 2015 20:20:48 -0400
To:	bug-Time-Piece [...] rt.cpan.org
From:	Jarkko Hietaniemi <jhi [...] iki.fi>

On Sunday-201506-07 18:12, jhi@iki.fi via RT wrote: Show quoted text

> Thanks. > > Take a look at the testers results? I would suggest trying to leave in > at least FreeBSD (assuming it passes...), to get a little bit more coverage.

Duh, and of course osx (darwin). But what I mean in general, it's a little sad if testing only for glibc strftime, essentially...

Tue Jun 09 12:37:06 2015 JHI [...] cpan.org - Correspondence added

Show quoted text

> I will push out a fix in the next week or two.

As http://www.cplusplus.com/reference/ctime/strftime/ states, the %z is C99, and Perl doesn't require that. (and even if it did, there's no guarantee that a particular strftime implements %z...)

Tue Jun 16 16:19:07 2015 ESAYM [...] cpan.org - Correspondence added

On Sun Jun 07 18:12:03 2015, jhi@iki.fi wrote: Show quoted text

> On Sunday-201506-07 18:08, Samuel Smith via RT wrote:

> > Thank you for the report. Some of the tests in 02core_dst.t have > > shown to be somewhat fragile. I am probably going to have to modify > > them or simply limit them to Linux hosts.

> > Thanks. > > Take a look at the testers results? I would suggest trying to leave > in > at least FreeBSD (assuming it passes...), to get a little bit more > coverage.

Coverage is a good thing, I agree. Part of the issue is that 02core_dst.t was added for a specific reason that is not really valid anymore. In 1.29, strftime() used a somewhat butchered tm struct to generate a string from and this was causing issues with %z and %Z. I think that has been fixed by simply passing the epoch to strftime in 1.30. So the issues present in 1.29 should not occur again, but more tests usually are a good thing. Except when they prevent one from compiling Perl itself... And I guess that is the issue with dual life modules and strict tests :) I have skipped the tests for the non mainline *nix derivatives for now. I wanted to included Win32, but cpantesters is reporting all failures for Windows 7 that I cannot reproduce on my own Win7 box.

Sun Oct 18 08:48:42 2015 JHI [...] cpan.org - Subject changed from '02core_dst.t two failures in 1.30' to '02core_dst.t two failures in 1.30 in irix'

Wed Oct 21 14:21:11 2015 ESAYM [...] cpan.org - Correspondence added

Ignore extra tests for Irix in 1.31.

Wed Oct 21 14:21:13 2015 ESAYM [...] cpan.org - Status changed from 'open' to 'resolved'

Wed Oct 21 14:21:14 2015 ESAYM [...] cpan.org - Fixed in 1.31 added