On Thu, May 30, 2013 at 01:29:42PM +0300, Victor Miasnikov wrote: > > Hi! > > >On Tue, May 28, 2013 at 03:53:05PM +0100, David Cantrell wrote: > > > >>The code in my module that has the problem starts at line 186 here: > >> > >> https://github.com/DrHyde/perl-modules-CPAN-ParseDistribution/blob/master/lib/CPAN/ParseDistribution.pm#L186 > > > >I've factored the Unix-specific code into a separate module now, so if > >anyone gets a chance to look at it, it should be easier to fix. > > > Ok, > for All who not have time open 1-10 URLs, see _related_ content: > > http://www.cpantesters.org/cpan/report/6c50e2a4-6c0f-1014-988b-f0c2162c787f
> Output from 'C:\Perl16\site\bin\dmake.exe test': > > C:\Perl16\bin\perl.exe "-MExtUtils::Command::MM" "-e" "test_harness(0, > 'blib\lib', 'blib\arch')" t/*.t > 'tar' is not recognized as an internal or external command, > operable program or batch file. > no AUTHOR_TESTING, skipping a bunch of tests This is not relevant. That's the tests detecting that you don't have tar installed and thus skipping all the tests that require it. > Attempt to free unreferenced scalar: SV 0x3a39a4c, Perl interpreter: > 0x2f23974 at > C:\Perl16\cpan\build\Parallel-ForkManager-1.03-NooagN/blib/lib/Parallel/ForkManager.pm > line 511. > > ( VVM: see later "CORE::exit($x || 0);" ) Here's the details of what's happening in the tests: https://gist.github.com/wchristian/dcace83dc70011ae47ca So I think that these are just warnings and can (for now) be ignored. > t/fileindexer.t ... > Dubious, test returned 15 (wstat 3840, 0xf00) > Failed 9/38 subtests The message "failed 9/38 subtests" is a bit misleading. It actually appears to mean that something is crashing horribly at about line 147 of this file - there are nine tests remaining to be run at that point: https://metacpan.org/source/DCANTRELL/CPAN-ParseDistribution-1.5/t/fileindexer.t This makes sense, as trying to parse the Acme-BadExample distribution is where it tests the magic for having the child process time out. > VVM>> ActiveState 5.16.0 ? Why not 5.16.3? Why not S.B.Perl? > > D.C.> You mean why did my original email point at a test report from perl > 5.16.0? > D.C.>Because, errm, that's where the test report came from. > D.C.> I don't control what build of perl people use. > + > D.C.> . . .could just attempt to debug solely by using the CPAN-testers > but . . . > > > IMHO, need simply answer to "perl people"/tester ( i.e. to Christian Walde) > : > -- please send result of testing with Strawberry Perl 5.18 and/or 5.16.3 Here <http://tinyurl.com/l6am97l> are several other identical test failures on various versions of perl on Windows. -- David Cantrell | Official London Perl Mongers Bad Influence Hail Caesar! Those about to vi ^[ you!