Bug#901082: libpgobject-type-json-perl: FTBFS with Perl 5.28: t/02-serialization.t failure
On Sun, Jun 10, 2018 at 12:36:09AM +0300, Niko Tyni wrote: > This bisects to > > https://github.com/makamaka/JSON-PP/commit/87bd6a49bacc3a2634cbb1dd0ce9cc75675bb524 > and I've filed > https://github.com/makamaka/JSON-PP/issues/39 > about it. > > Not sure yet if it's a bug or if others need to adapt. Upstream(s) are somewhat quiet, so just noting that this can be worked around with a build dependency on libjson-xs-perl if there's no better fix. For runtime, maybe a Recommends entry would be appropriate? -- Niko Tyni nt...@debian.org
Bug#901082: libpgobject-type-json-perl: FTBFS with Perl 5.28: t/02-serialization.t failure
On Sat, Jun 09, 2018 at 10:14:22AM +0300, Niko Tyni wrote: > On Fri, Jun 08, 2018 at 11:14:35PM +0200, gregor herrmann wrote: > > On Fri, 08 Jun 2018 21:38:12 +0300, Niko Tyni wrote: > > > > > Source: libpgobject-type-json-perl > > > Version: 2.01-1 > > > Severity: important > > > User: debian-p...@lists.debian.org > > > Usertags: perl-5.28-transition > > > > > > This package fails to build with Perl 5.28 (currently in experimental.) > > > > > > > > > http://perl.debian.net/rebuild-logs/perl-5.28-throwaway/libpgobject-type-json-perl_2.01-1/libpgobject-type-json-perl_2.01-1_amd64-2018-06-08T17:55:53Z.build > > > > > > # Failed test 'Literal serializes correctly' > > > # at t/02-serialization.t line 49. > > > # got: '123' > > > # expected: '"123"' > > > # Looks like you failed 1 test of 23. > > > > > > I don't see an upstream bug or a failing CPAN test report > > > but it fails consistently for me. > > > > I can confirm the failure with 5.28, and it still passes the tests > > with 5.26. > > > > No idea which JSON thing in which part adds/removes the quotation > > marks. All I can think of is the different version of JSON::PP but > > then the problem should show up elsewhere as well. > > It's indeed to do with JSON::PP and not specific to Perl 5.28. > > Reduced to > > perl -MJSON::PP -le '$p=123; "". $p; print > JSON::PP->new->allow_nonref->encode($p)' > > which gives the string "123" on older versions of JSON::PP and the number 123 > on newer ones (at least 2.97001-1). This bisects to https://github.com/makamaka/JSON-PP/commit/87bd6a49bacc3a2634cbb1dd0ce9cc75675bb524 and I've filed https://github.com/makamaka/JSON-PP/issues/39 about it. Not sure yet if it's a bug or if others need to adapt. -- Niko Tyni nt...@debian.org
Bug#901082: libpgobject-type-json-perl: FTBFS with Perl 5.28: t/02-serialization.t failure
Control: retitle -1 libpgobject-type-json-perl: FTBFS with newer versions of JSON::PP: t/02-serialization.t failure On Fri, Jun 08, 2018 at 11:14:35PM +0200, gregor herrmann wrote: > On Fri, 08 Jun 2018 21:38:12 +0300, Niko Tyni wrote: > > > Source: libpgobject-type-json-perl > > Version: 2.01-1 > > Severity: important > > User: debian-p...@lists.debian.org > > Usertags: perl-5.28-transition > > > > This package fails to build with Perl 5.28 (currently in experimental.) > > > > > > http://perl.debian.net/rebuild-logs/perl-5.28-throwaway/libpgobject-type-json-perl_2.01-1/libpgobject-type-json-perl_2.01-1_amd64-2018-06-08T17:55:53Z.build > > > > # Failed test 'Literal serializes correctly' > > # at t/02-serialization.t line 49. > > # got: '123' > > # expected: '"123"' > > # Looks like you failed 1 test of 23. > > > > I don't see an upstream bug or a failing CPAN test report > > but it fails consistently for me. > > I can confirm the failure with 5.28, and it still passes the tests > with 5.26. > > No idea which JSON thing in which part adds/removes the quotation > marks. All I can think of is the different version of JSON::PP but > then the problem should show up elsewhere as well. It's indeed to do with JSON::PP and not specific to Perl 5.28. Reduced to perl -MJSON::PP -le '$p=123; "". $p; print JSON::PP->new->allow_nonref->encode($p)' which gives the string "123" on older versions of JSON::PP and the number 123 on newer ones (at least 2.97001-1). I don't know why CPAN testers don't see this. Maybe there's JSON::XS installed underneath that masks it? -- Niko Tyni nt...@debian.org
Bug#901082: libpgobject-type-json-perl: FTBFS with Perl 5.28: t/02-serialization.t failure
On Fri, 08 Jun 2018 21:38:12 +0300, Niko Tyni wrote: > Source: libpgobject-type-json-perl > Version: 2.01-1 > Severity: important > User: debian-p...@lists.debian.org > Usertags: perl-5.28-transition > > This package fails to build with Perl 5.28 (currently in experimental.) > > > http://perl.debian.net/rebuild-logs/perl-5.28-throwaway/libpgobject-type-json-perl_2.01-1/libpgobject-type-json-perl_2.01-1_amd64-2018-06-08T17:55:53Z.build > > # Failed test 'Literal serializes correctly' > # at t/02-serialization.t line 49. > # got: '123' > # expected: '"123"' > # Looks like you failed 1 test of 23. > > I don't see an upstream bug or a failing CPAN test report > but it fails consistently for me. I can confirm the failure with 5.28, and it still passes the tests with 5.26. No idea which JSON thing in which part adds/removes the quotation marks. All I can think of is the different version of JSON::PP but then the problem should show up elsewhere as well. Cheers, gregor -- .''`. https://info.comodo.priv.at -- Debian Developer https://www.debian.org : :' : OpenPGP fingerprint D1E1 316E 93A7 60A8 104D 85FA BB3A 6801 8649 AA06 `. `' Member VIBE!AT & SPI Inc. -- Supporter Free Software Foundation Europe `- NP: Kante: Ituri signature.asc Description: Digital Signature
Bug#901082: libpgobject-type-json-perl: FTBFS with Perl 5.28: t/02-serialization.t failure
Source: libpgobject-type-json-perl Version: 2.01-1 Severity: important User: debian-p...@lists.debian.org Usertags: perl-5.28-transition This package fails to build with Perl 5.28 (currently in experimental.) http://perl.debian.net/rebuild-logs/perl-5.28-throwaway/libpgobject-type-json-perl_2.01-1/libpgobject-type-json-perl_2.01-1_amd64-2018-06-08T17:55:53Z.build # Failed test 'Literal serializes correctly' # at t/02-serialization.t line 49. # got: '123' # expected: '"123"' # Looks like you failed 1 test of 23. I don't see an upstream bug or a failing CPAN test report but it fails consistently for me. -- Niko Tyni nt...@debian.org