Bug #101232 for Date-Calc: Test suite started to fail since 2015-01-01

Thu Jan 01 05:25:18 2015 SREZIC [...] cpan.org - Ticket created

Subject:

Test suite started to fail since 2015-01-01

See http://matrix.cpantesters.org/?dist=Date-Calc;reports=1#sl=7,1 for a time-sorted overview of test reports. All reports in the new year 2015 are fail reports.

Thu Jan 01 11:25:28 2015 $_ = 'spro^^*%*^6ut# [...] &$%*c>#!^!#&!pan.org'; y/a-z.@//cd; print - Correspondence added

On Thu Jan 01 05:25:18 2015, SREZIC wrote: Show quoted text

> See http://matrix.cpantesters.org/?dist=Date-Calc;reports=1#sl=7,1 for > a time-sorted overview of test reports. All reports in the new year > 2015 are fail reports.

Makes ‘Happy New Year!’ rather ironic. :-) Oh, and Happy New Year, BTW!

Thu Jan 01 11:25:28 2015 The RT System itself - Status changed from 'new' to 'open'

Fri Jan 02 01:18:14 2015 ANDK [...] cpan.org - Correspondence added

For the record: Date-Calc-XS is also affected

Fri Jan 02 09:38:13 2015 jeppe [...] ioslo.net - Correspondence added

From:

jeppe [...] ioslo.net

It looks like the tests that fail refer to the PP version. The test scripts which fail start with BEGIN { $Date::Calc::XS_DISABLE = $Date::Calc::XS_DISABLE = 1; } .. and they refer to Decode_Date_EU2, Decode_Date_US2, Decode_Date_EU and Decode_Date_US. I've taken some time to rewrite one of the test files using a proper test harness, and easily find that the tests fail since we're now closer to 2064 than we are to 1964.

Subject:

f016.t

#!perl -w BEGIN { eval { require bytes; }; } use Test::Most; BEGIN { $Date::Calc::XS_DISABLE = $Date::Calc::XS_DISABLE = 1; } use Date::Calc qw( Decode_Date_EU Decode_Date_US ); use strict; # ====================================================================== # ($year,$mm,$dd) = Decode_Date_EU($buffer); # ($year,$mm,$dd) = Decode_Date_US($buffer); # ====================================================================== plan tests => 25; # format: [[parameters], [expected output]] my $eu_tests = [ [ ["3.1.64"], [ 1964, 1, 3 ] ], [ ["3 1 64"], [ 1964, 1, 3 ] ], [ ["03.01.64"], [ 1964, 1, 3 ] ], [ ["03/01/64"], [ 1964, 1, 3 ] ], [ [ "3. Ene 1964", 4 ], [ 1964, 1, 3 ] ], [ [ "Geburtstag: 3. Januar '64 in Backnang/Württemberg", 3 ], [ 1964, 1, 3 ] ], [ ["03-Jan-64"], [ 1964, 1, 3 ] ], [ [ "3.Jan1964", 6 ], [ 1964, 1, 3 ] ], [ [ "3Jan64", 0 ], [ 1964, 1, 3 ] ], [ ["030164"], [ 1964, 1, 3 ] ], [ [ "3ja64", ], [ 1964, 1, 3 ] ], [ [ "3164", ], [ 1964, 1, 3 ] ], [ ["28.2.1995"], [ 1995, 2, 28 ] ], [ ["29.2.1995"], [] ], ]; foreach my $testcase (@$eu_tests) { my @arr = Decode_Date_EU( @{ $testcase->[0] } ); is_deeply( \@arr, $testcase->[1], "parse " . join( ',', @{ $testcase->[0] } ) . " ok" ); } my $us_tests = [ [ ["1 3 64"], [ 1964, 1, 3 ] ], [ ["01/03/64"], [ 1964, 1, 3 ] ], [ ["Jan 3 '64"], [ 1964, 1, 3 ] ], [ ["Jan 3 1964"], [ 1964, 1, 3 ] ], [ ["===> January 3rd 1964 (birthday)"], [ 1964, 1, 3 ] ], [ ["Jan31964"], [ 1964, 1, 3 ] ], [ ["Jan364"], [ 1964, 1, 3 ] ], [ ["ja364"], [ 1964, 1, 3 ] ], [ ["1364"], [ 1964, 1, 3 ] ], [ ["2.28.1995"], [ 1996, 2, 28 ] ], [ ["2.29.1995"], [] ] ]; foreach my $testcase (@$us_tests) { my @arr = Decode_Date_US( @{ $testcase->[0] } ); is_deeply( \@arr, $testcase->[1], "parse " . join( ',', @{ $testcase->[0] } ) . " ok" ); } __END__

Wed Jan 07 07:56:09 2015 contyk [...] cpan.org - Cc contyk added

Wed Jan 07 08:59:29 2015 contyk [...] cpan.org - Correspondence added

I think all cases with a two-digit year in the date are affected, both PP and XS as the logic is the same in both. I think just checking if the result is 2064 now is sufficient (and correct). A better option would be computing the correct year in the test itself so this wouldn't break in 2115 again but... well... The attached patch works with Date::Pcalc, too.

Subject:

Date-Calc-6.3-century.patch

Message body is not shown because it is too large.

Sat Jan 17 21:19:12 2015 gregoa [...] cpan.org - Correspondence added

For the record: Same for Date::Palc.

Sat Jan 17 21:20:19 2015 gregoa [...] cpan.org - Cc GREGOA added

Sun Jan 18 13:49:40 2015 gregoa [...] cpan.org - Correspondence added

This should read Date::Pcalc ...

Sun Jan 18 18:56:29 2015 gregoa [...] cpan.org - Correspondence added

Here's an alternative patch which calculates the year. - Not sure which approach is preferrable ...

Subject:

moving_window.patch

Message body is not shown because it is too large.

Tue Jan 27 10:53:12 2015 dam [...] cpan.org - Correspondence added

Here's yet another patch which tries to keep what I guess is the original idea of the tests -- see if a two-digit year from 50 years ago (64) maps to the correct full year (1964).

Subject:

century.patch

Message body is not shown because it is too large.

Tue Jan 27 10:54:33 2015 dam [...] cpan.org - Cc DAM added

Tue Jan 27 15:48:28 2015 HMBRAND [...] cpan.org - Severity Important added

Wed Feb 04 09:16:11 2015 LGEHLEN [...] cpan.org - Correspondence added

Do I understand correctly that only the test suite is broken, not Date::Calc itself? Thank you for clarifying, Lutz

Wed Feb 04 15:49:58 2015 gregoa [...] debian.org - Correspondence added

Subject:	Re: [rt.cpan.org #101232] Test suite started to fail since 2015-01-01
Date:	Wed, 4 Feb 2015 21:49:31 +0100
To:	Lutz Gehlen via RT <bug-Date-Calc [...] rt.cpan.org>
From:	gregor herrmann <gregoa [...] debian.org>

On Wed, 04 Feb 2015 09:16:11 -0500, Lutz Gehlen via RT wrote: Show quoted text

> Do I understand correctly that only the test suite is broken, not Date::Calc itself? > Thank you for clarifying,

That's my understanding. The tests use the 2-digit-year '64' and check that it expands to 1964; which worked until 2014-12-31, since 2015-01-01 it expands to 2064. And this behaviour is consistent with the module documentation. Cheers, gregor -- .''`. Homepage: http://info.comodo.priv.at/ - OpenPGP key 0xBB3A68018649AA06 : :' : Debian GNU/Linux user, admin, and developer - https://www.debian.org/ `. `' Member of VIBE!AT & SPI, fellow of the Free Software Foundation Europe `- NP: Fleetwood Mac: Little Lies

Download signature.asc
application/pgp-signature 949b

Message body not shown because it is not plain text.

Thu Feb 05 17:03:27 2015 gregoa [...] cpan.org - Correspondence added

Here's the final patch from Damyan we uploaded to Debian right now. (Applies to -XS and Date::PCalc as well with an offset). Cheers, gregor

Subject:

century-tests.patch

Message body is not shown because it is too large.

Mon Mar 02 16:03:37 2015 RSAVAGE [...] cpan.org - Correspondence added

I've emailed the author of this module asking for co-maint. I'm not an XS programmer, but I can fix a few things.

Mon Mar 02 16:03:39 2015 RSAVAGE [...] cpan.org - Fixed in 6.3 added

Wed Mar 04 16:26:55 2015 SREZIC [...] cpan.org - Fixed in 6.3 deleted

Wed Mar 04 19:00:10 2015 RSAVAGE [...] cpan.org - Correspondence added

I'm discussing the patches in private emails with Steffen. Hold your breath.....

Fri Mar 06 17:54:53 2015 STBEY [...] cpan.org - Correspondence added

On Wed Mar 04 19:00:10 2015, RSAVAGE wrote: Show quoted text

> I'm discussing the patches in private emails with Steffen. Hold your > breath.....

I am indeed currently working on a new release. If all goes well it should find its way onto CPAN this weekend. BTW, thanks a lot for all the patches you guys (and gals?) sent in! Unfortunately using Moving_Window() is no option because it makes the tests tautological (ad absurdum), because the Decode_Date_X() functions use it internally. Testing against 2064 is just as bad an idea as it was to test against 1964 in the first place, so I'll adopt a more general approach, in order for this bug not to bite us again next century (when I will no longer be around to fix it! ;-) ). Just in case this module will somehow live that long... ;-)

Fri Mar 06 19:14:56 2015 STBEY [...] cpan.org - Correspondence added

On Fri Mar 06 17:54:53 2015, STBEY wrote: Show quoted text

> On Wed Mar 04 19:00:10 2015, RSAVAGE wrote:

> > I'm discussing the patches in private emails with Steffen. Hold your > > breath.....

> > I am indeed currently working on a new release. > If all goes well it should find its way onto CPAN this weekend. > > BTW, thanks a lot for all the patches you guys (and gals?) sent in! > > Unfortunately using Moving_Window() is no option because it makes the > tests tautological (ad absurdum), > because the Decode_Date_X() functions use it internally. > > Testing against 2064 is just as bad an idea as it was to test against > 1964 in the first place, > so I'll adopt a more general approach, in order for this bug not to > bite us again next century (when I will no longer be around to fix it! > ;-) ). > > Just in case this module will somehow live that long... ;-)

Please test the following release candidates and let me know if they test alright: Date::Calc 6.4 Sat 07-Mar-2015 01:00:08 CET 2505ac448904945613237ec636695f4e 172879 Date-Calc-6.4.tar.gz 2012 Date-Calc-6.4.README http://guest.engelschall.com/~sb/download/pkg/Date-Calc-6.4.tar.gz Date::Calc::XS 6.4 Sat 07-Mar-2015 01:04:59 CET db42e9ea3b14adcc6d95c2b5913a172c 85842 Date-Calc-XS-6.4.tar.gz 2018 Date-Calc-XS-6.4.README http://guest.engelschall.com/~sb/download/pkg/Date-Calc-XS-6.4.tar.gz

Sat Mar 07 14:38:32 2015 SREZIC [...] cpan.org - Correspondence added

On 2015-03-06 19:14:56, STBEY wrote: Show quoted text

> On Fri Mar 06 17:54:53 2015, STBEY wrote:

> > On Wed Mar 04 19:00:10 2015, RSAVAGE wrote:

> > > I'm discussing the patches in private emails with Steffen. Hold > > > your > > > breath.....

> > > > I am indeed currently working on a new release. > > If all goes well it should find its way onto CPAN this weekend. > > > > BTW, thanks a lot for all the patches you guys (and gals?) sent in! > > > > Unfortunately using Moving_Window() is no option because it makes the > > tests tautological (ad absurdum), > > because the Decode_Date_X() functions use it internally. > > > > Testing against 2064 is just as bad an idea as it was to test against > > 1964 in the first place, > > so I'll adopt a more general approach, in order for this bug not to > > bite us again next century (when I will no longer be around to fix > > it! > > ;-) ). > > > > Just in case this module will somehow live that long... ;-)

> > > Please test the following release candidates and let me know if they > test alright: > Date::Calc 6.4 Sat 07-Mar-2015 01:00:08 CET > 2505ac448904945613237ec636695f4e 172879 Date-Calc-6.4.tar.gz 2012 > Date-Calc-6.4.README > http://guest.engelschall.com/~sb/download/pkg/Date-Calc-6.4.tar.gz > Date::Calc::XS 6.4 Sat 07-Mar-2015 01:04:59 CET > db42e9ea3b14adcc6d95c2b5913a172c 85842 Date-Calc-XS-6.4.tar.gz 2018 > Date-Calc-XS-6.4.README > http://guest.engelschall.com/~sb/download/pkg/Date-Calc-XS-6.4.tar.gz

Tested with perl 5.20.2 on a debian/wheezy machine: Date-Calc-6.4 passes all tests, Date-Calc-XS-6.4 also passes all tests if Date::Calc 6.4 is found in the @INC path.

Sat Mar 07 14:43:11 2015 STBEY [...] cpan.org - Correspondence added

On Sat Mar 07 14:38:32 2015, SREZIC wrote: Show quoted text

> On 2015-03-06 19:14:56, STBEY wrote:

> > On Fri Mar 06 17:54:53 2015, STBEY wrote:

> > > On Wed Mar 04 19:00:10 2015, RSAVAGE wrote:

> > > > I'm discussing the patches in private emails with Steffen. Hold > > > > your > > > > breath.....

> > > > > > I am indeed currently working on a new release. > > > If all goes well it should find its way onto CPAN this weekend. > > > > > > BTW, thanks a lot for all the patches you guys (and gals?) sent in! > > > > > > Unfortunately using Moving_Window() is no option because it makes > > > the > > > tests tautological (ad absurdum), > > > because the Decode_Date_X() functions use it internally. > > > > > > Testing against 2064 is just as bad an idea as it was to test > > > against > > > 1964 in the first place, > > > so I'll adopt a more general approach, in order for this bug not to > > > bite us again next century (when I will no longer be around to fix > > > it! > > > ;-) ). > > > > > > Just in case this module will somehow live that long... ;-)

> > > > > > Please test the following release candidates and let me know if they > > test alright: > > Date::Calc 6.4 Sat 07-Mar-2015 01:00:08 CET > > 2505ac448904945613237ec636695f4e 172879 Date-Calc-6.4.tar.gz 2012 > > Date-Calc-6.4.README > > http://guest.engelschall.com/~sb/download/pkg/Date-Calc-6.4.tar.gz > > Date::Calc::XS 6.4 Sat 07-Mar-2015 01:04:59 CET > > db42e9ea3b14adcc6d95c2b5913a172c 85842 Date-Calc-XS-6.4.tar.gz 2018 > > Date-Calc-XS-6.4.README > > http://guest.engelschall.com/~sb/download/pkg/Date-Calc-XS-6.4.tar.gz

> > Tested with perl 5.20.2 on a debian/wheezy machine: Date-Calc-6.4 > passes all tests, Date-Calc-XS-6.4 also passes all tests if Date::Calc > 6.4 is found in the @INC path.

Thanks a lot!! The latter is indeed a prerequisite.