It's been a while since my last package (it was before git) and i'm trying to update my knowlege about the packaging process.
The problem i'm facing is that there's a test in this Perl distribution that is behaving differently inside pdebuild with cowbuilder. As far as i can see from the tests i performed outside pdebuild, the problem seems to be that there's an extra pipe for fd/0 which makes an odd number of pipes, thus failing the test. I made some changes to the test with quilt and pushed them into the pkg-perl repo, but they're only to help me debug the issue for now. Those changes make the test show some diags, e.g. $ pdebuild --pbuilder cowbuilder ... make[1]: Entering directory `/tmp/buildd/libmojolicious-plugin-cgi-perl-0.07' PERL_DL_NONLAZY=1 /usr/bin/perl "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t t/00-basic.t ... ok t/basic.t ...... ok t/before.t ..... ok t/errlog.t ..... ok t/multipart.t .. ok # 11686 # waiting for 11686 Use of uninitialized value in pattern match (m//) at t/post.t line 35. # pipes: $VAR1 = [ # [ # '/proc/11685/fd/0', # 'pipe:[90940]' # ], # [ # '/proc/11685/fd/1', # 'pipe:[91868]' # ], # [ # '/proc/11685/fd/2', # 'pipe:[91869]' # ], # [ # '/proc/11685/fd/6', # 'pipe:[91868]' # ], # [ # '/proc/11685/fd/7', # 'pipe:[91869]' # ] # ]; # Failed test 'no leaky pipes' # at t/post.t line 41. # Looks like you failed 1 test of 5. t/post.t ....... Dubious, test returned 1 (wstat 256, 0x100) Failed 1/5 subtests t/redirect.t ... ok Does anybody have a clue about where does the fd/0 come from and how could i improve the test? I'll ask the author if it's safe to ignore fd/0 in this case. Any other thoughts? Thanks a lot, Alex
signature.asc
Description: Digital signature
