Personal CPAN server

dragonchild has asked for the wisdom of the Perl Monks concerning the following question:

What steps do I need to take in order to create a personal CPAN server? The idea here is that I'm organizing my application into a number of CPAN-like distributions and I would like to be able to use CPAN to install them onto my production server(s). But, I cannot upload them to PAUSE because they contain company algorithms. I don't want to mirror CPAN - I want to have my own CPAN server that only contains my modules that is the first server in my CPAN search path.

I've already checked SuperSearch, google, www.cpan.org, and the Camel book, but there isn't much info on the internals of how CPAN works or what CPAN expects as an API.

------
We are the carpenters and bricklayers of the Information Age.

Please remember that I'm crufty and crochety. All opinions are purely mine and all code is untested, unless otherwise specified.

Comment on Personal CPAN server

Replies are listed 'Best First'.
Re: Personal CPAN server by gellyfish (Monsignor) on Jan 19, 2004 at 15:26 UTC
In the first instance I would take a look at CPAN::Site which provides a method of doing roughly what you want /J\	[reply]
Re: Personal CPAN server by derby (Abbot) on Jan 19, 2004 at 15:40 UTC
merlyn had a column on that. Maybe we need a meryln tag for perlmonks? -derby Update: Uggh. I missed the "I don't want to mirror" part of the OP. In my company we work this the other way around. We have several production environments and then we also have a test and development enviroment. When we build perl, we include our "company specific" INC path. The normal flow is to download CPAN modules to dev, test and production. Company specific modules are developed on dev. Then we push through those modules to test and push through to production via rdist . So it's a bit more complicated to set up but if you have more than one production machine - it's much easier to rdist from a test environment than it is to go to each production machine and do CPAN updates. And since our production machines are similar, we really only do a CPAN update on one and rdist those updates out to the rest (but we do need to do a CPAN update for each different architecutre - linux, solaris, etc).	[reply]
Re: Personal CPAN server by Roger (Parson) on Jan 20, 2004 at 01:58 UTC
I thought the CPAN server is just an FTP server, isn't it? I would just poke around inside the CPAN mirror and do some reverse engineering :-D CPAN.html 1k Nov 19 2001 Symbolic link -> authors/id/J/JO/JONO/cpan.html ENDINGS 1k May 1 2001 Symbolic link -> .cpan/ENDINGS MIRRORED.BY 123k Dec 16 11:42 MIRRORING.FROM 3k Jan 11 18:23 README 1k May 26 2002 README.html 1k May 1 2001 Symbolic link -> index.html RECENT 1k May 1 2001 Symbolic link -> indices/RECENT-print RECENT.html 175k Jan 11 03:34 ROADMAP 1k May 26 2002 ROADMAP.html 1k May 26 2002 SITES 21k Dec 16 11:42 SITES.html 42k Dec 16 11:42 authors/ 7k Feb 17 2003 Directory clpa/ 1k Aug 30 2002 Directory disclaimer.html 2k Aug 30 2002 doc/ 1k Sep 8 20:34 Directory index.html 3k Jan 10 22:01 indices/ 1k Jan 11 18:26 Directory jhi/ 1k Feb 18 2003 Directory ls-lR.gz 1k May 1 2001 Symbolic link -> indices/ls-lR.gz misc/ 1k Dec 27 2002 Directory modules/ 1k Oct 1 11:27 Directory <<< looks interesting ports/ 1k Jan 3 20:35 Directory scripts/ 1k May 9 2002 Directory src/ 1k Nov 17 15:45 Directory The CPAN.pm module provides a standard interface to CPAN. There is a Defaultsite variable which can be overrided with your own FTP/CPAN server... I haven't tried this but it would certainly be an interesting little weekend project. `use CPAN; $CPAN::Defaultsite = "ftp://mirror.aarnet.edu.au/pub/CPAN";` [download]	[reply] [d/l]
Re: Personal CPAN server by neilwatson (Priest) on Feb 24, 2006 at 20:51 UTC
Dragonchild, What was your solution? I'm in a similar position. I'd like to know what you did. Neil Watson watson-wilson.ca	[reply]