Re: RFC: The Future of Solving Dependency Problems in APT (SAT, CUDF)
On Do, 2010-12-23 at 16:14 +0100, Pietro Abate wrote:
> It is important to notice, that picosat will be able to tell you a
> solution, but probably it won't be a good one. These solvers should be
> able to tell you the best solution accordingly to given objective
> function.
picosat allows us to specify a lot of things, like keeping installed
packages if possible, or preferring candidate versions over
non-candidate ones. In my tests, it created wonderful solutions.
> A small whish item since we are here: can we also expose the function
> that computes the installation order. It would be nice to somehow
> formalize this algorithm and let the "user" access the list in some
> other way then throught a debugging option...
Different part of APT.
--
Julian Andres Klode - Debian Developer, Ubuntu Member
See http://wiki.debian.org/JulianAndresKlode and http://jak-linux.org/.
Reply to: