On Mon, Nov 12, 2007 at 03:41:57PM +0100, Nagy Gabor wrote:
commit 96f8faa6664714943201d86393099dbf7464abc2 Author: Chantry Xavier <shiningxc@gmail.com> Date: Sun Nov 11 10:52:51 2007 -0600
Add two requiredby pactests
One currently should succeed (006), and 005 fails.
requiredby005.py is originally from Nagy Gabor <ngaba@petra.hos.u-szeged.hu>. Well, I'm not sure that pacman should deal with broken localdbs. After we decided that _all_ satisfiers _must_ list the dependency "owner" in their requiredby, this is a broken db. We have a cool testdb stuff (created by you) to help user fix these (an automatized db fix wouldn't be so hard to implement).
I only submitted a pactest because Aaron asked for it, for testing the buggy list implentation : http://www.archlinux.org/pipermail/pacman-dev/2007-November/009925.html So I didn't want it to be submitted. I didn't say it clearly though. Dan submitted it because he found it was useful, and I couldn't find any good arguments against it. Now you bring me one :) I didn't even pay attention to the fact that the db was broken in this pactest. I really dislike this pactest anyway, it isn't interesting at all, it was only meant to show a very specific bug in the list implementation. Also, the title is misleading. I originally called it "broken list". What I meant exactly was : "broken list implementation". So I think this pactest should be removed, and eventually replaced later by other better / more interesting ones. I'm not sure pactest is the best way to test the list implementation, it's too general. If pacman really needs an automated test for its list implementation, then it should probably be a specific test.