At Wed, 18 Aug 2004 19:59:20 +0900, Mauricio FernáÏdez wrote: > > I think the former is better. The idea is making RPA's packages > (including associated metadata, etc) good enough to allow them to be > used as the basis for other packages. In other words, taking Debian as > an example, it would be possible to generate a debianized source tree > from a RPA port. Now, direct .rps/.rpa => .deb conversion IS possible, but > it is better to respect Debian/FreeBSD/Gentoo/etc 's normal procedures. > > rpa-base is essentially repackager-friendly (but it will soon become > better), and a number of parameters (installation paths, handling > of rdoc/ri documentation) can be adapted to comply with the desired > filesystem/etc standards. This sounds very good :) I hope you'll reach this goal, this should result in less work for Debian/FreeBSD/Gentoo/etc package maintainers, and hopefully more ruby-related and more up-to-date packages on those systems. > > Or that those systems live next to each other ? Or is it > > This is the current situation. rpa-base tries to play nice with other > port/pkg managers and will not overwrite any file unless you indicate > it's OK. Ok, but for me personally, this is still not a good solution, because i want to keep 'native' package files and rpa-installed package files separated.. but... > > mainly meant to use those tools as a regular user with the purpose of > > having your own versions of the ruby libraries you want without > > needing superuser access ? > > You can also use rpa-base that way: just set the $prefix to some directory > you own when installing rpa-base for the first time. On my system, I > have 3 rpa-base installed: one in $HOME/usr (where I have the latest > stable snapshot of Ruby too), one in $HOME/ruby1.9 (1.9 CVS) and the > system-wide one in /usr/local. this seems perfect for now, this way i can install things with rpa-base without messing up my system. Thanks for the explanation and for the work and time you put in rpa. Ruben