2008-6-8

00:00 cramm
empty, Alex_Gaynor: I'm trying to make django-pyodbc (http://code.google.com/p/django-pyodbc/) work
00:00 empty
cool.
00:00 cramm
it's based on #5246
00:00 DjangoBot
00:00 cramm
but it's totally broken
00:02 cramm
imvho pyodbc is a better option than adodb because I would like to use it also from unix via unixodbc and freetds
00:03 cramm
i've since dropped support of pymssql and trying to focus in sing pyodbc. It's supposed to support ms sql 2000 and 2005
00:03 empty
yah
00:04 empty
cramm: that limit code for 2000 is interesting. But I suppose in a single select that's the only way.
01:09 Gulopine
woohoo!
01:09 Gulopine just got a patch committed with 0% alteration!
01:10 Gulopine
of course, that change means i gotta go rewrite part of the book :(
01:10 C8E
ehe
02:33 ubernostrum
Oh, how I love thoughtful and informed commentary.
02:34 Alex_Gaynor
Hehe, blog posts are fun
18:16 jbronn
I'm looking at #7367, and I'm wondering is this guy is abusing the ORM by trying to have two OneToOneFields to `Place` from `Restaurant`? Or is this a bug, considering if both O2O relations point to different `Place` models?
18:16 DjangoBot
18:17 jacobkm
jbronn: one sec; I'll look
18:19 jacobkm
jbronn: yeah, if there's two one-to-one relationships, they're not one-to-one :)
18:19 Alex_Gaynor
Looks like an intermediary model issue to me
18:19 jacobkm
OneToOneField is really just for "is-a" relationships, hence the PK munging. I think this is a case of "this word does not mean what you think it means"
18:20 jbronn
Cool, ok if i close it as invalid then?
18:20 jacobkm
Alex_Gaynor: I think it's a subclass of FK(unique=True) not generating a single related descriptor on the related class. If we did that (which is an open ticket elsewhere) then everyone would be pretty happy with FK(unique=True)
18:20 jacobkm
jbronn: yes, please do
18:21 Alex_Gaynor
jacobkm: Sorry wasn't clear, I meant this data could be represented with an intermediary m2m
18:22 jacobkm
Alex_Gaynor: sure, you could do it that way, too
18:24 Alex_Gaynor
The createsuperuser command was added to the django-admin/manage.py page, however I don't think it belongs their, since it isn't a django wide thing, it assumes use of the admin, any opinions?
18:25 jacobkm
Alex_Gaynor: 99% of the time you'll be using django.ontrib.auth, so I'm fine with it being part of that page. It should mention that it's part of auth, though (does it?)
18:25 Alex_Gaynor
I'll check
18:25 Alex_Gaynor
No it doesn't
18:25 Alex_Gaynor
I'll make a quick diff
18:26 jacobkm
cool
18:30 Alex_Gaynor
Here's the diff: http://dpaste.com/55502/ do you want me to open a ticket?
18:32 jezdez
jacobkm: I could have a look at #5522 if you need more management commands :)
18:32 DjangoBot
18:34 jacobkm
jezdez: why did *_messages end up in django/core/management/base? They really don't belong there...
18:37 jezdez
I think I put them there in accordance with the copy_helper function, though I have no hard feelings moving them to their own django/core/managment/commands/*.py
18:38 jacobkm
jezdez: yeah, they should *definitely* be sub-commands. I might not get around to actually aplying the patch right away since it's sort just rearranging deck chairs, but once I've got some time it's pretty much a no-brainer.
18:41 jezdez
cool, great to hear that. I had some i18n work last week with pinax and a djangodash project which reminded me of the usefulness of this reorganisation :) I'll work on a new patch that moves the worker function to the subcommands and would love to actually include a solution for #7180/#5463
18:41 DjangoBot
18:42 jezdez
#5463
18:42 DjangoBot
18:43 jacobkm
jezdez: on #5463 the patch is fine; the default just needs to stay ".html" instead of the big list the patch uses right now.
18:43 DjangoBot
18:43 jacobkm
Wait, isn't 7180 a duplicate of 5463?
18:44 jacobkm
Yeah, it totally is.
18:44 jezdez
yup
18:48 jezdez
ok, I will include the patch on #5463 in the new patch for #5522
18:48 DjangoBot
18:48 DjangoBot
18:49 jacobkm
jezdez: no, please try to keep them seperate; merging tickets together is a major pain.
18:49 jacobkm
feel free to make one patch depend on the other and mark it as such, but please keep things in their place
18:49 jezdez
ok, no prob.
18:50 Gulopine
yeah, don't be like #5361 :)
18:50 DjangoBot
18:50 jezdez
eeep :)
20:11 SimonW
Is there a test application for newforms-admin that exercises most of the features anywhere?
20:11 SimonW
I want to do some work on newforms-admin packages but I need a relatively complete testing environment