On Sat, 01 Jan 2005 13:00:25 +0100, Nils Philippsen <nphilipp@xxxxxxxxxx> wrote: > As Levente has mentioned, perhaps this could be a bit speedier: > > 5. pickling/unpickling the yum internal data structures instead of > storing/reading XML; depending on how efficient (un)pickling is, this > might be faster though maybe at the cost of that we can't reuse yum > caches between architectures (I think we can live with that if python > doesn't pickle/unpickle portably) It already does that. It only parses XML when you see "MD Read". The rest of the time it loads pickled data. The downside is that loading pickles is very memory-hungry. -- Konstantin Ryabitsev Zlotniks, INC
