Skip Menu |

This queue is for tickets about the DateTime-Calendar-FrenchRevolutionary CPAN distribution.

Report information
The Basics
Id: 61230
Status: resolved
Priority: 0/
Queue: DateTime-Calendar-FrenchRevolutionary

People
Owner: Nobody in particular
Requestors: J2N-FORGET [...] orange.fr
Cc:
AdminCc:

Bug Information
Severity: Normal
Broken in: 0.07
Fixed in: 0.09



Subject: Failed tests in 45-lastday.t
In a few CPAN testers reports, substests 6 and 7 fail in 45-lastday.t. In all other CPAN testers reports and on all systems I have access to, they succeed. In version 0.08, a new testfile 46-lastday.t was introduced to use the same dates and check them with the day and month methods instead of using the iso8601 method. These tests succeeds even on systems where 45-lastday.t fail.
The reason is painted here: http://analysis.cpantesters.org/reports_by_field?distv=DateTime-Calendar-FrenchRevolutionary-0.08;field=conf%3Anvsize If nvsize is > 8 then the tests fail. Systems defining uselongdouble on 64bit machines get nvsize=16 or nvsize=12. I don't know how the decision is made between 12 and 16. Anyway, the tests fail on systems with nvsize != 8. Here is the verbose test output: % ./Build test --test_files=t/45_lastday.t --verbose t/45_lastday.t .. 1..7 ok 1 ok 2 ok 3 ok 4 ok 5 not ok 6 : expected 0211-13-05T8:01:01, got 0211-13-05T8:01:00 not ok 7 : expected 0212-13-06T8:01:01, got 0212-13-06T8:01:00 Failed 2/7 subtests HTH, Regards,
Le Ven 24 Déc 2010 07:01:39, ANDK a écrit : Show quoted text
> The reason is painted here: > > http://analysis.cpantesters.org/reports_by_field?distv=DateTime- > Calendar-FrenchRevolutionary-0.08;field=conf%3Anvsize > > If nvsize is > 8 then the tests fail. > Systems defining uselongdouble on > 64bit machines get nvsize=16 or nvsize=12. > I don't know how the decision > is made between 12 and 16. Anyway, the > tests fail on systems with nvsize > != 8. > > Here is the verbose test output: > > % ./Build test --test_files=t/45_lastday.t --verbose > t/45_lastday.t .. > 1..7 > ok 1 > ok 2 > ok 3 > ok 4 > ok 5 > not ok 6 : expected 0211-13-05T8:01:01, got 0211-13-05T8:01:00 > not ok 7 : expected 0212-13-06T8:01:01, got 0212-13-06T8:01:00 > Failed 2/7 subtests > > > HTH, Regards,
Yes, it helps me very much. Not only I know what triggers the problem, but also I know a bit more about the CPAN admins' achievements and the CPAN Testers' contribution to the Perl community. Thank you Andreas, thank you the other admins and testers and a happy new year to you all.
The problem with floating arithmetic is solved in version 0.09.