On Sat, 22 Dec 2001 01:38:44 +0100 "Nicolas Chauvat" <[EMAIL PROTECTED]> wrote:
> Anyhow, my pet project for tonight was to write such a graph checker. I (and others i know of) have thought about this problem, it is an intersting problem, i dont think anybody has come up with a good solution yet. Most cycles arent bad, some package depend on themselves (their virtual package). > What I call a dependency cycle is something like (A depends on B and B > depends on A) which basically implies that A and B *have*to* be > installed and removed together. It may not be a mistake for some closely > related packages, but chances are that it will be with cycles of lenght > longer than 2. > These types of cycles arent a problem, as you say they can be installed together. A bad cycle which would make a package uninstallable would be if A Pre-Depends on B and B Depends or Pr-Depends on A (and visa versa) > What I call a conflict dependency problem is something like (A conflicts > with B and B depends on A). I saw at least one case where this was on > purpose to facilitate upgrade, but I would say that it is an actual > problem in most cases. > Some of these will be false alarms as you have to take into account virtual packages (Provides: field). e.g. gaim **> gaim-gnome --> gaim-common --> gaim gaim-gnome provides the virtual package gaim which satisfies gaim-common Best of luck Glenn