Re: Ticket #15363 -- Normalization of the queryset methods name

2013-03-06 Thread charettes
Unfortunately this would break in subtle in confusing ways in inheritance scenarios. Say you've defined the following Manager subclass prior to Django 1.6 and we're only raising a warning on Manager.get_query_set calls.: class ActiveManager(models.Manager): def get_query_set(self, *args,

Re: Ticket #15363 -- Normalization of the queryset methods name

2013-03-06 Thread Jacob Kaplan-Moss
On Wed, Mar 6, 2013 at 12:21 PM, charettes wrote: > The use of the metaclass make sure you're not shooting yourself in the foot > all and all the metaclass subclassing trouble should only concern a minority > of users who most probably know what they're doing. Ah, OK, that's the missing part I wa

Re: Ticket #15363 -- Normalization of the queryset methods name

2013-03-06 Thread Aymeric Augustin
There's a good, although non-obvious, reason. Without the metaclass, a method overridden in the subclass with the old name would be silently skipped. -- Aymeric. Le 6 mars 2013 à 16:30, Jacob Kaplan-Moss a écrit : > Hm. > > I'm +1 on cleaning up the names but do we *really* have to use

Re: Ticket #15363 -- Normalization of the queryset methods name

2013-03-06 Thread Jacob Kaplan-Moss
Hm. I'm +1 on cleaning up the names but do we *really* have to use a metaclass for this? Seems to me this is "gratuitous use of a metaclass" territory, especially given the shenanigans you have to go through with subclassing the metaclass. I'd be a lot happier with this patch if it just did t

Ticket #15363 -- Normalization of the queryset methods name

2013-03-05 Thread charettes
Since the consensus as shifted toward not making a break-everything 2.0 I'm planning to normalize the queryset methods name early in the 1.6 development cycle to spot any breakages. The affected classes are the following: - *django.contrib.admin.options.BaseModelAdmin* (*queryset* -­>