* brian d foy via RT (bug-CPAN-Site@rt.cpan.org) [081023 12:47]:
Show quoted text> Thu Oct 23 08:47:49 2008: Request 40301 was acted upon.
> Transaction: Ticket created by BDFOY
> Queue: CPAN-Site
> Subject: CPAN::Site::Index shouldn't read raw tar.gz stream
> Ticket <URL:
http://rt.cpan.org/Ticket/Display.html?id=40301 >
>
> CPAN::Site::Index goes through a tar file looking for strings with
> `package` statements in the raw stream. This fails in many cases, such
> as DBI, where there is example code.
Yes, but don't forget: it only does this for your own packages, not
for the common CPAN stuff. So: people can work-around complications
in their own packages themselves. If they would like to use CPAN::Site
for (a part of) the common CPAN as well, then I would advise cpan-mini.
Show quoted text> I've been working on this sort of problem for my MyCPAN::Indexer
> stuff. There is a lot of work there to unpack a distribution, find
> the modules, and pull out the package statements. Feel free
> to steal any of the code. :)
Thanks for pointing it to me. And probably inside Andreas' CPAN code
is the reference implementation for this.
If I make the module better, then the indexer will run noticeably slower
and make it far more complex. For the purpose of CPAN::Site which I have
in mind, the problem is not that big to spend energy on (I can spend my
time better on other modules). Unless you have a purpose of CPAN::Site
which I did not think of, which may shift the balance...
--
Regards,
MarkOv
------------------------------------------------------------------------
Mark Overmeer MSc MARKOV Solutions
Mark@Overmeer.net solutions@overmeer.net
http://Mark.Overmeer.net http://solutions.overmeer.net
