Bug #109843 for App-RecordStream: Modification of non-creatable hash value attempted, subscript "zoo"

Sat Nov 28 04:38:16 2015 SREZIC [...] cpan.org - Ticket created

Subject:

Modification of non-creatable hash value attempted, subscript "zoo"

The assert.t test may fail with the message: # Failed test 'Error contains expression' # at /usr/local/src/CPAN/build/2015112712/App-RecordStream-4.0.17-hu8nYc/tests/RecordStream/Operation/assert.t line 44. # 'Modification of non-creatable hash value attempted, subscript "zoo" at (eval 13) line 1, <GEN2> line 2. # ' # doesn't match '(?-xism:\{\{zoo\}\})' ... A sample fail report: http://www.cpantesters.org/cpan/report/43e0b736-9535-11e5-8bd1-8dc1e0bfc7aa An overview of all fail/pass reports at CPAN Testers: http://matrix.cpantesters.org/?dist=App-RecordStream%204.0.17

Sat Nov 28 12:52:54 2015 tsibley [...] cpan.org - Correspondence added

Show quoted text

> # 'Modification of non-creatable hash value attempted, > # subscript "zoo" at (eval 13) line 1, <GEN2> line 2.'

Thanks for following up, Slaven. I noticed the spate of these yesterday, from your smokers and a few others, and dug into them a bit. I can't reproduce locally, and they smell like an upstream module problem to me. Very little code changed between App-RecordStream 4.0.16 and 4.0.17. The error itself seems to be something that should rarely happen unless XS is involved¹, which this dist has none of (although XS deps are optionally used). I wonder, could you schedule a re-run of App-RecordStream-4.0.16 for me on your smokers? It saw no testers failures originally. If it still passes, then that suggests the issue with 4.0.17 is in the dist itself after all. If it now fails, then it suggests the root problem lies elsewhere. Thanks, Thomas ¹ https://metacpan.org/pod/perlguts#AVs-HVs-and-undefined-values

Sat Nov 28 12:52:54 2015 The RT System itself - Status changed from 'new' to 'open'

Sat Nov 28 13:19:51 2015 tsibley [...] cpan.org - Correspondence added

On Sat Nov 28 09:52:54 2015, TSIBLEY wrote: Show quoted text

> I wonder, could you schedule a re-run of App-RecordStream-4.0.16 for > me on your smokers? It saw no testers failures originally. If it > still passes, then that suggests the issue with 4.0.17 is in the dist > itself after all. If it now fails, then it suggests the root problem > lies elsewhere.

Nevermind! I've tracked down the issue to Cpanel::JSON::XS: https://github.com/rurban/Cpanel-JSON-XS/issues/45 Cheers, Thomas

Sat Nov 28 14:06:12 2015 SREZIC [...] cpan.org - Correspondence added

On 2015-11-28 13:19:51, TSIBLEY wrote: Show quoted text

> On Sat Nov 28 09:52:54 2015, TSIBLEY wrote:

> > I wonder, could you schedule a re-run of App-RecordStream-4.0.16 for > > me on your smokers? It saw no testers failures originally. If it > > still passes, then that suggests the issue with 4.0.17 is in the dist > > itself after all. If it now fails, then it suggests the root problem > > lies elsewhere.

> > Nevermind! I've tracked down the issue to Cpanel::JSON::XS: > > https://github.com/rurban/Cpanel-JSON-XS/issues/45

I can confirm this: with the newest Cpanel::JSON::XS the previously passing perls report now fails for 4.0.17.

Thu Feb 18 12:35:33 2016 tsibley [...] cpan.org - Status changed from 'open' to 'rejected'