Bug #129181 for Data-TableAutoSum: Test failures (hash randomization?)

Sun Apr 14 09:58:38 2019 SREZIC [...] cpan.org - Ticket created

Subject:

Test failures (hash randomization?)

The test suite fails on almost all of my smoker systems. The test log output is quite long because of warnings and usually truncated on CPAN Testers (see http://www.cpantesters.org/cpan/report/8926edfe-52c6-11e9-ab3e-5b43707c0951 for a sample report). The actual failure seems to be: ... t/result.t ............ ok Name "main::STORED_TALE" used only once: possible typo at t/store_read.t line 32. # Failed test 'Read stored table has the same as_string representantion as the original (ARRAY(0x5637bc4f9018)xARRAY(0x5637bc5486f8) table)' # at t/store_read.t line 37. # got: '\t Col 0\t Col 1\t Col 2\t Col 3\t Col 4\t Col 5\t Col 6\t Col 7\t Col 8\t Col 9\t Col 10\t Col 11\t Col 12\t Col 13\t Col 14\t Col 15\t Col 16\t Col 17\t Col 18\t Col 19\t Col 20\t Col 21\t Col 22\t Col 23\t Col 24\t Col 25\t Col 26\t Col 27\t Col 28\t Col 29\t Col 30\t Col 31\t Col 32\t Col 33\t Col 34\t Col 35\t Col 36\t Col 37\t Col 38\t Col 39\t Col 40\t Col 41\t Col 42\t Col 43\t Col 44\t Col 45\t Col 46\t Col 47\t Col 48\t Col 49\t Sum\n ... (more rows following) ... For perls < 5.18 I don't see any fail reports, so maybe the cause here is hash randomization introduced in newer perls?

Sun Apr 28 05:16:14 2019 bigj [...] kamelfreund.de - Correspondence added

Thanks for the report, it had nothing to do with hashing, but I fixed: - these many warnings (have correctly increased with perl versions), I developped it with Perl 5.6 :-o - storing the file with immediate flushing - main problem was that there were small floating errors when calculating random numbers truncating on 3 digits. I now round them in the tests to integers to avoid such stuff. Could it be that Perl < 5.18 test suite run perl with 32 bit or anything else. Anyway, the module worked, but tests failed :-o Also, somehow I doubt anyone will still use it :-o - but who knows. I'd use pandas instead for sure nowadays :-) Version 0.09 with this fix is uploading and Version 0.10 including a link to the github repo in https://github.com/hanfried/Data-TableAutoSum is uploading. When tests are running over next days, I'll mark it as resolved. On Sun Apr 14 09:58:38 2019, SREZIC wrote: Show quoted text

> The test suite fails on almost all of my smoker systems. The test log > output is quite long because of warnings and usually truncated on CPAN > Testers (see http://www.cpantesters.org/cpan/report/8926edfe-52c6- > 11e9-ab3e-5b43707c0951 for a sample report). > > The actual failure seems to be: > > ... > t/result.t ............ ok > Name "main::STORED_TALE" used only once: possible typo at > t/store_read.t line 32. > > # Failed test 'Read stored table has the same as_string > representantion as the original > (ARRAY(0x5637bc4f9018)xARRAY(0x5637bc5486f8) table)' > # at t/store_read.t line 37. > # got: '\t Col 0\t Col 1\t Col 2\t Col 3\t Col 4\t Col > 5\t Col 6\t Col 7\t Col 8\t Col 9\t Col 10\t Col 11\t > Col 12\t Col 13\t Col 14\t Col 15\t Col > 16\t Col 17\t Col 18\t Col 19\t Col 20\t > Col 21\t Col 22\t Col 23\t Col 24\t Col > 25\t Col 26\t Col 27\t Col 28\t Col 29\t > Col 30\t Col 31\t Col 32\t Col 33\t Col > 34\t Col 35\t Col 36\t Col 37\t Col 38\t > Col 39\t Col 40\t Col 41\t Col 42\t Col > 43\t Col 44\t Col 45\t Col 46\t Col 47\t > Col 48\t Col 49\t Sum\n > ... (more rows following) ... > > For perls < 5.18 I don't see any fail reports, so maybe the cause here > is hash randomization introduced in newer perls?

Sun Apr 28 05:16:15 2019 The RT System itself - Status changed from 'new' to 'open'

Sun Apr 28 05:16:15 2019 bigj [...] kamelfreund.de - Status changed from 'open' to 'patched'

Mon Apr 29 05:31:34 2019 bigj [...] kamelfreund.de - Correspondence added

All tests are now green, so I'll mark it as resolved.

Mon Apr 29 05:31:34 2019 bigj [...] kamelfreund.de - Status changed from 'patched' to 'resolved'

Mon Apr 29 05:31:34 2019 bigj [...] kamelfreund.de - Fixed in 0.09 added