This may not be a bug, but it might impact current metabase testers,
so I want to ask here first.
My smoker just got around to testing the new CPAN-Reporter-1.19_03.
It's trying to run this:
.../metabase-profile --output
/home/bq_cdavis/.cpanreporter/~/.cpanreporter/metabase_id.json
despite the existence of ~/.cpanreporter/metabase_id.json
And the reason for that is because this is false:
File::Spec->file_name_is_absolute('~/.cpanreporter/metabase_id.json');
And I'm using that path because that's what the last set of
documentation that I read on the Testers Wiki seemed / did suggest
when we made the change to metabase last year. And it has worked until
now. Of course, it's a simple workaround to update my
~/.cpanreporter/config.ini, but what is the impact for existing
metabase users? Do we:
1) Fix my config.ini to not include a ~ and warn against that in the docs
2) Add an additional check for ~ to CPAN::Reporter
3) File a bug against File::Spec (I'm surprised this doesn't work, anyone else?)
Cheers,
Chad
On Tue, Jun 28, 2011 at 21:16, David Golden <xda...@gmail.com> wrote:
> I just uploaded a new dev release of CPAN::Reporter. Thank you to
> Christian and Breno for contributing, and to Breno in particular for
> patiently, but persistently prodding me into action.
>
> I'd appreciate if people could try it out. Even if you don't use
> CPAN::Reporter, I'd appreciate if you could read through the
> documentation and see if what I've said about Metabase makes sense.
>
> -- David
>
> 1.19_03 2011-06-28 14:44:19 America/New_York
>
> [NEW FEATURES]
> - During configuration, if 'Metabase' is the requested transport and
> a Metabase profile file is not found, configuration will offer to
> run metabase-profile to create it.
>
> - For Metabase transport (only), the 'id_file' may have a relative
> file path, treated as relative to the .cpanreporter directory.
>
> [BUG FIXES]
> - updated CPAN::Reporter to go with Metabase by default, instead of
> the deprecated SMTP transport system. This provides more up-to-date
> information for new testers on how to setup their systems and
> addresses RT#64316 (and maybe also RT#61735) [Breno G. de Oliveira]
>
> - Replaced use of Tee.pm with Capture::Tiny as the two don't play
> well together on some systems. [Christian Walde]
>
> [DOCUMENTATION]
> - Docs have been substantially re-written to reflect the new HTTP-based
> transport instead of email-based transport. [Breno G. de Oliveira and
> David Golden]
>
