Kjetil, Sorry about this, I should have caught it faster
Could you please implement Serguei's recommendation? http://www.nntp.perl.org/group/perl.cpan.testers.discuss/2010/11/msg2262.html It[File::Temp]'s not broken, you're using it wrong. # TEST my ($newfh, $newfilename) = tempfile(); my $content; $ini->WriteConfig($newfilename); $newfh is opened filehandle for $newfilename, and you're trying to open this file again in your code. You should close $newfh first, if you're going to use $newfilename and not $newfh. -----Original Message----- From: Kjetil Kjernsmo [mailto:kje...@kjernsmo.net] Sent: Wednesday, January 25, 2012 10:47 AM To: Susanto, Leo Cc: cpan-testers-discuss@perl.org Subject: Re: RDF::Trine::Store::File test On Wednesday 25. January 2012 11.55.28 Susanto, Leo wrote: > It seems that it is borked on Strawberry too: > http://www.cpantesters.org/cpan/report/4d73c9bd-6c2f-1014-886d-b1ed885448ee > Ah, yes! > Somehow I believe something about File::Util is causing the problem. I'll > include cpan-testers-discuss, maybe one of the wizard can help you. Great, thanks! > C:\CPANTesters\cpan\build\RDF-Trine-Store-File-0.1-_oDzuD>C:\CPANTesters\Per > l5.1 0.1\bin\perl.exe "-MExtUtils::Command::MM" "-e" "test_harness(1, > 'inc', 'blib\li b', 'blib\arch')" t/file-manip.t > t/file-manip.t .. > ok 1 - use RDF::Trine::Store::File; > ok 2 - Store object OK > ok 3 - Store has one statement according to size > ok 4 - Store has one statement according to count > ok 5 - Etag is 32 chars long, only hex > # Sleep one second to ensure new etag > PROCESS TERMINATED DUE TO ERRORS > # Tests were run but no plan was declared and done_testing() was not seen. > > File::Util can't open > ╗.\C\\temp\xLcqJO9pv4½ > because no such file or directory exists. Aha, interesting! This is what I do at this point: https://metacpan.org/source/KJETILK/RDF-Trine-Store-File-0.1/t/file-manip.t#L31 In the first part of the tests, I let RDF::Trine::File::Store do the work. At the end of it, I create an etag, which is basically just right now a hash digest of the file mtime. I test this for sanity, and then I want to ensure it changes and since mtime has a one-second resolution (that's perhaps not portable), I sleep one second before proceeding. Then, I let File::Util work on the file, and thus I load from the same filename. At this point, it seems from the above output that the File::Temp-created file doesn't exist anymore on Windows. Could this be the real problem, and if so, how should I solve it? Kjetil
