Bug #107061 for Parse-RecDescent: Test fails for perl >= 5.18.0

Sun Sep 13 07:46:07 2015 SREZIC [...] cpan.org - Ticket created

Subject:

Test fails for perl >= 5.18.0

See http://fast-matrix.cpantesters.org/?dist=Parse-RecDescent%201.967011 for an overview. With perl 5.20.3 the test log looks like this: # Failed test 'parsers match' # at t/reproducible.t line 420. # Failed test 'parsers match' # at t/reproducible.t line 420. # Failed test 'parsers match' # at t/reproducible.t line 420. # Failed test 'parsers match' # at t/reproducible.t line 420. # Failed test 'parsers match' # at t/reproducible.t line 420. # Failed test 'parsers match' # at t/reproducible.t line 420. # Looks like you failed 6 tests of 41. t/reproducible.t .......... Dubious, test returned 6 (wstat 1536, 0x600) Failed 6/41 subtests

Sun Sep 13 10:24:53 2015 jtbraun [...] CPAN.org - Correspondence added

Thanks. I'll build a new perl and investigate. Presumably the changes to the way hashing works in recent Perls cause the output to change on each invocation, and the original patch submission pre-dated these changes.

Sun Sep 13 10:24:53 2015 The RT System itself - Status changed from 'new' to 'open'

Sun Sep 13 11:05:48 2015 jtbraun [...] CPAN.org - Correspondence added

Fix committed in: https://github.com/jtbraun/Parse-RecDescent/commit/06dc4d8c779d69706665b07646fac5ebb5615c74 Will appear as v1.967012 on CPAN shortly. Root cause: a typo. I camel-cased $Data::Dumper::SortKeys mistakenly (it's actually Sortkeys), and it only showed up as a test failure with 5.18, because 5.18 randomizes hashes per-hash, not per-process. Initial testing of v1.967011 before release was done with 5.14. *sigh*

Sun Sep 13 11:05:50 2015 jtbraun [...] CPAN.org - Status changed from 'open' to 'resolved'