On Thursday 10 April 2008, Guy Hulbert wrote: > On Thu, 2008-10-04 at 19:48 +0300, Shlomi Fish wrote: > > And if anyone has a preference for the location of the data file or > > files (or > > __DATA__) - voice his opinion now.[1] > > This approach only works for new modules.
Depends what you mean by "new". If you mean version-0.01-freshly-uploaded-to-the-CPAN then no - it can be applied to existing distribution. If you mean that I would need to convert my distribution to have this meta-data, then the answer is yes - it requires explicit choice. But I see this as a feature, not a limitation, because I want the author (or uploader) to have a conscious choice of what is specified. My original link gave a way for users to assign tags and classifications of their own on a "more-gives-strength" basis. But this has no place in the distribution's tar.gz itself. This should be better be discussed in the rethinking-cpan mailing list not here. My plan is for M::B to read both the parameters given to the constructor and the ones in the YAML file, and integrate them somehow (probably with the constructor params gaining precedence or over-riding the ones in the YAML file). > Did you look at the stuff > Jonathan Rockway advertised here: > > http://git.jrock.us/?p=MetaCPAN.git;a=summary > > In particular, this description is only 3 sentences. > > http://git.jrock.us/?p=MetaCPAN.git;a=blob;f=HOWTO;h=95a43b5958b6066dfe3364 >87c3a0349515411fcf;hb=HEAD > > When I get some time for this, my first step will be to see what this > can do. I don't see anything of relevance in these links, at least not as text that I can read and understand. Are you sure you got them right? And please copy-and-paste the relevant parts to your message. Regards, Shlomi Fish --------------------------------------------------------------------- Shlomi Fish [EMAIL PROTECTED] Homepage: http://www.shlomifish.org/ I'm not an actor - I just play one on T.V.
