* Kent Fredric <[EMAIL PROTECTED]> wrote: Hi folks,
> Imo, the cyclic dep problem could be solved as thus, > > A depends B > B depends C||A > > Where C is a minimalist subset of A required for building B, > which is only depended on if A is not present. > A is also a replacement for C. Beware of conflicts ! AFAIK portage cannot handle them by itself, instead aborts and leaves the job to the user. > So the flow would go like such. > > Emerge A: > * depends on b > * A is missing, so depend on C > *emerges C* > *emerges B* > *removes C* <-- otherwise A & C containing the same files = headache > *emerges A* Who exactly controls this flow ? portage or the user ? > Computers exist to make our life easier, not the other way around :) ACK. An "ideal" system would be able to do it all alone, the user just tells what he wants. Of course ideal != realistic, but IMHO its good to go into the ideal direction as far as possible. BTW: some people expressed (within this thread), that Gentoo was too easy to use and should be made more complicated. Funny, isn't it ? Ah, just in case that 99% of the people didn't actually read my postings, I'll just say it again: The circular dependencies between Xserver and drivers do not come from upstream. They're artificial. If we just want an simple-to-use package which gets the Xserver *and* drivers based on useflags (which IMHO is an good idea), it's quite trivial to do this by an virtual/meta package. I've did it and it works good. You can get it from my overlay via CVS: CVSROOT=:pserver:[EMAIL PROTECTED]:/repositories/cvs/oss-qm password: annymous module: oss-qm-overlay cu -- --------------------------------------------------------------------- Enrico Weigelt == metux IT service - http://www.metux.de/ --------------------------------------------------------------------- Please visit the OpenSource QM Taskforce: http://wiki.metux.de/public/OpenSource_QM_Taskforce Patches / Fixes for a lot dozens of packages in dozens of versions: http://patches.metux.de/ --------------------------------------------------------------------- -- [EMAIL PROTECTED] mailing list
