2008-6-2

13:35 RaceCondition
I've posted an invalid diff file on django trac and it was integrated in the trunk.. should I simply post the correct diff now?
13:36 RaceCondition
the bad diff file was not generated in the django svn repo root dir
13:36 RaceCondition
so it has revision 0 as both source and target revisions
13:42 empty
RaceCondition: was the patch already applied and ticket closed?
14:12 RaceCondition
empty: svn diff -r 7567 shows that its included
14:13 RaceCondition
but the file that it adds to the repository does not exist after svn up
16:18 RaceCondition
I'm having some hard time translating some strings in django -- they're really impossible to translate into Estonian
16:57 RaceCondition
empty: sorry, I mistakenly thought my diff was applied to trunk. actually it hasn't been :)
16:57 empty
yeah just upload a new patch.
16:57 empty
RaceCondition: ^^
16:58 RaceCondition
yep
16:59 RaceCondition
empty: if I don't have admin privileges in Trac, I cannot replace my previously attached diff file which is invalid?
17:00 empty
RaceCondition: we don't replace. Just add another one.
17:00 empty
They like the history.
17:00 RaceCondition
OK, fine
17:00 RaceCondition
I wonder what's taking it so long for the diff to be integrated
17:04 RaceCondition
how can I keep myself logged in to Trac so I wouldn't have to enter my name/email address when commenting a ticket?
17:05 RaceCondition
oh, got it :) I'm a dumbass
17:39 robhudson
wondering if adding a middleware class argument to specify dependency would solve the middleware ordering problem? E.g.: If auth middleware added a `depend = "django.contrib.sessions.middleware.SessionMiddleware"` and you had them in the wrong order in settings.py, it would automatically correct the order for that dependency tree only (so you still have some control). Care should be taken to not have a cycle in the dependency tree (should be a DAG).
17:43 jacobkm
robhudson: not a bad idea, though I'd more prefer it to just warn you about the order than to actually try any sort of automatic correction... but it's kinda compelling
17:43 jacobkm
robhudson: you can do that manually in you're middleware's __init__ by inspecting settings, fwiw.
17:46 robhudson
what I like about a depends attribute is it puts the problem of dependency on the author of the middleware, who should know best about it. Agree that automatically re-ordering might be a bit much, but that would make it "just work" (unless you have cyclic dependencies), which is kinda nice.
17:47 robhudson
I could play around with the idea and add a patch to #730 if interested
17:47 DjangoBot
17:48 robhudson
I could also look at making ./manage.py validate say something about middleware ordering if dependencies were specified.
17:53 jacobkm
robhudson: sure, sounds like a good plan
17:55 robhudson
cool. seems like validate is a good 1st place to start anyway. I'll see what I can come up with. Sounds fun... I like playing with graphs.
17:55 jacobkm
there's a "tree" structure in django.utils now, I think
17:55 robhudson
is there any graph stuff already in Django? any datastructures I can borrow?
17:56 robhudson
ok, I'll check it
17:56 jacobkm
I thought malcolm added one for qsrf - take a look
18:39 lsoto
hi!
18:41 lsoto
can someone give me a hand with my Field.get_db_prep_* confusions?
18:41 lsoto