Re: PEAR website

On Friday 06 July 2001 15:48, Martin Jansen wrote: > On Fri, 6 Jul 2001 15:00:34 +0200, Tomas V.V.Cox wrote: > > > >I see three needed task, that can be done with the same script (in > > cron or elsewhere). This is: get packages from cvs, build them with > > the "pear package" and publish them in the website. So users could > > copy the location of the package and launch "pear install url". I > > can help on this if someone can give me access. > > I had the same thoughts a few days ago: My idea was > to checkout the different packages from CVS (/pear > and /php4/pear) and to read each package.xml in the > root directories. > > The information from package.xml can e.g. be stored in a > database. The source files are packed together into a .tar.gz- > and/or a .bz2-file. Build a package with all the packages. Why? I'll explain my idea: Cron job: - checkout modules - go throw all the dirs - do a "pear package package.xml" - parse the package.xml and store useful data in a database - Move the package to a public dir Web interface: - List, search, etc avaible packages - Once you get the desired package, you can download it and install localy or copy the url and install remotely. I think we should distribute pear packages, yeah perhaps a tarball with all the pear packs could be also a good idea. Tomas V.V.Cox