Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
---+
   Reporter:  vanschelven  |Owner:  nobody
   Type:  New feature  |   Status:  new
  Milestone:   |Component:  Core (Other)
Version:  SVN  | Severity:  Normal
 Resolution:   | Keywords:
   Triage Stage:  Accepted |Has patch:  1
Needs documentation:  0|  Needs tests:  0
Patch needs improvement:  1|
---+
Changes (by jaddison):

 * milestone:  1.3 =>


-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-updates@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
---+
   Reporter:  vanschelven  |Owner:  nobody
   Type:  New feature  |   Status:  new
  Milestone:  1.3  |Component:  Core (Other)
Version:  SVN  | Severity:  Normal
 Resolution:   | Keywords:
   Triage Stage:  Accepted |Has patch:  1
Needs documentation:  0|  Needs tests:  0
Patch needs improvement:  1|
---+
Changes (by jaddison):

 * type:   => New feature
 * severity:   => Normal


-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-updates@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
--+-
   Reporter:  vanschelven | Owner:  nobody
 Status:  new | Milestone:  1.3   
  Component:  Core framework  |   Version:  SVN   
 Resolution:  |  Keywords:
   Triage Stage:  Accepted| Has patch:  1 
Needs documentation:  0   |   Needs tests:  0 
Patch needs improvement:  1   |  
--+-
Changes (by russellm):

  * owner:  russellm => nobody


Comment:

 If we were talking about a critical component of Django, I might agree to
 taking a "fix the problem, then fix the fix" approach. In fact, I did just
 that with #9964 only days ago. However, I don't think this problem falls
 into that category. Django has had model inheritance for a while, and this
 is the first time this problem has been reported. It requires a complex
 model inheritance structure and a reliance on automated form ordering
 before it becomes a problem.

 I'm sorry that deferring this will be inconvenient to you, but I'd ask you
 to also consider *my* inconvenience -- the fact that landing a suboptimal
 bugfix *increases* my work as a framework maintainer in the long run,
 because at some point, I'll have to reconcile two disparate
 representations of base classes.

 Two issues of protocol:

  * Backwards compatibility isn't strictly an issue here, because _meta
 isn't stable API. There are certain aspects of _meta that I'd be very
 pained to change (like the basic field management stuff that has been that
 way since the beginning), but I don't think parents falls into that list.

  * Please don't assign bugs to other people without their permission.

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-updates@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
--+-
   Reporter:  vanschelven | Owner:  russellm
 Status:  new | Milestone:  1.3 
  Component:  Core framework  |   Version:  SVN 
 Resolution:  |  Keywords:  
   Triage Stage:  Accepted| Has patch:  1   
Needs documentation:  0   |   Needs tests:  0   
Patch needs improvement:  1   |  
--+-
Changes (by ambv):

  * owner:  nobody => russellm


Comment:

 The reason I introduced `model_bases` instead of rearranging half the
 framework around changes in `parents` was to keep the patch minimal (e.g.
 easier to get right). On top of that, I believe in backwards compatibility
 and I'm not sure if there isn't some 3rd party code using `parents` in
 ways that could have been broken by changing that.

 I am willing to help with any refactoring of Meta necessary. However, this
 is naturally milestone 1.4, e.g. over 9 months away. Could you please
 reassess whether we could integrate the patch as it is? Without that
 change I will need to basically fork Django until another stable release.
 This is not easy to monkey patch.

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-updates@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
--+-
   Reporter:  vanschelven | Owner:  nobody
 Status:  new | Milestone:  1.3   
  Component:  Core framework  |   Version:  SVN   
 Resolution:  |  Keywords:
   Triage Stage:  Accepted| Has patch:  1 
Needs documentation:  0   |   Needs tests:  0 
Patch needs improvement:  1   |  
--+-
Changes (by russellm):

  * needs_better_patch:  0 => 1
  * stage:  Ready for checkin => Accepted


Comment:

 This patch appears to be headed broadly in the right direction, and it
 works, but I'm not convinced that it's quite right yet.

 In particular, the fact that it is introducing a new `model_bases`
 attribute to the Meta class when `parents` already exists concerns me.
 Granted, the two aren't completely identical -- parents doesn't include
 any abstract classes, for example -- but it strikes me that the right
 solution here isn't just to add a new block of almost identical metadata,
 but to adjust the existing metadata, and adjust the existing usage.

 The Meta class is in bad need of some housekeeping -- I'd rather not
 knowingly add to that housekeeping work.

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-updates@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
-+--
   Reporter:  vanschelven| Owner:  nobody
 Status:  new| Milestone:  1.3   
  Component:  Core framework |   Version:  SVN   
 Resolution: |  Keywords:
   Triage Stage:  Ready for checkin  | Has patch:  1 
Needs documentation:  0  |   Needs tests:  0 
Patch needs improvement:  0  |  
-+--
Changes (by julien):

  * stage:  Accepted => Ready for checkin


Comment:

 I've updated the patch so that it applies to trunk. I've also renamed the
 'bases' attribute to 'model_bases' to avoid any confusion.

 I'm RFC'ing this patch as I think it solves an annoying inconsistency.
 I've just got one little reservation: possibly this will be backwards
 incompatible for people who were relying on that inconsistency. Although I
 don't think this is a big deal. I'll leave the core devs to make a call on
 this.

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-updates@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
-+--
  Reporter:  vanschelven | Owner:  nobody
Status:  new | Milestone:  1.3   
 Component:  Core framework  |   Version:  SVN   
Resolution:  |  Keywords:
 Stage:  Accepted| Has_patch:  1 
Needs_docs:  0   |   Needs_tests:  0 
Needs_better_patch:  0   |  
-+--
Changes (by ambv):

  * version:  1.2 => SVN
  * milestone:  => 1.3

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-updates@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
-+--
  Reporter:  vanschelven | Owner:  nobody
Status:  new | Milestone:
 Component:  Core framework  |   Version:  1.2   
Resolution:  |  Keywords:
 Stage:  Accepted| Has_patch:  1 
Needs_docs:  0   |   Needs_tests:  0 
Needs_better_patch:  0   |  
-+--
Changes (by anonymous):

  * has_patch:  0 => 1
  * component:  Uncategorized => Core framework

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-upda...@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
+---
  Reporter:  vanschelven| Owner:  nobody
Status:  new| Milestone:
 Component:  Uncategorized  |   Version:  1.2   
Resolution: |  Keywords:
 Stage:  Accepted   | Has_patch:  0 
Needs_docs:  0  |   Needs_tests:  0 
Needs_better_patch:  0  |  
+---
Changes (by russellm):

  * stage:  Unreviewed => Accepted

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-upda...@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
+---
  Reporter:  vanschelven| Owner:  nobody
Status:  new| Milestone:
 Component:  Uncategorized  |   Version:  1.2   
Resolution: |  Keywords:
 Stage:  Unreviewed | Has_patch:  0 
Needs_docs:  0  |   Needs_tests:  0 
Needs_better_patch:  0  |  
+---
Comment (by vanschelven):

 Replying to [comment:3 lrekucki]:
 > One thing that I don't understand is why does field order matter to you
 ? Maybe your problem can be solved easier on a different level.

 The field order is mainly important wherever I use ModelForms.
 True, I can post-hoc fix everything there, but that doesn't seem very DRY.

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-upda...@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
+---
  Reporter:  vanschelven| Owner:  nobody
Status:  new| Milestone:
 Component:  Uncategorized  |   Version:  1.2   
Resolution: |  Keywords:
 Stage:  Unreviewed | Has_patch:  0 
Needs_docs:  0  |   Needs_tests:  0 
Needs_better_patch:  0  |  
+---
Comment (by lrekucki):

 One thing that I don't understand is why does field order matter to you ?
 Maybe your problem can be solved easier on a different level.

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-upda...@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
+---
  Reporter:  vanschelven| Owner:  nobody
Status:  new| Milestone:
 Component:  Uncategorized  |   Version:  1.2   
Resolution: |  Keywords:
 Stage:  Unreviewed | Has_patch:  0 
Needs_docs:  0  |   Needs_tests:  0 
Needs_better_patch:  0  |  
+---
Changes (by lrekucki):

 * cc: lreku...@gmail.com (added)

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-upda...@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

Re: [Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
+---
  Reporter:  vanschelven| Owner:  nobody
Status:  new| Milestone:
 Component:  Uncategorized  |   Version:  1.2   
Resolution: |  Keywords:
 Stage:  Unreviewed | Has_patch:  0 
Needs_docs:  0  |   Needs_tests:  0 
Needs_better_patch:  0  |  
+---
Changes (by vanschelven):

  * needs_better_patch:  => 0
  * needs_tests:  => 0
  * needs_docs:  => 0

Comment:

 The attached file "failed_attempt" shows an initial stab at the problem.
 Running the tests quickly shows that this indeed does not solve it (other
 tests regress); but this is where my understanding of the ORM ends for
 now.

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-upda...@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.

[Django] #14705: Model Field Order not influenced by MRO of superclasses

[Django]

#14705: Model Field Order not influenced by MRO of superclasses
---+
 Reporter:  vanschelven|   Owner:  nobody
   Status:  new|   Milestone:
Component:  Uncategorized  | Version:  1.2   
 Keywords: |   Stage:  Unreviewed
Has_patch:  0  |  
---+
 Looking at the two definitions below,

 {{{
 class OneTwo(AbstractModelOne, AbstractModelTwo):
 pass

 class TwoOne(AbstractModelTwo, AbstractModelOne):
 pass
 }}}

 one (I?) would expect the order of the fields in the second definition to
 be the opposite of the order in the first one. (Since the base classes are
 referenced in the opposite order). This is, however, not the case.

 Django only uses creation_counter to determine the order of the fields.
 Therefor the actual field order is determined by the order in which the
 containing abstract classes are evaluated (i.e. their python code is
 read). If these classes come from multiple sources (apps) the field order
 may end up to be pretty much random (though deterministic).

 The general use case is heavy use of abstract models to mix in information
 from various sources into actual models. In real apps the two above models
 are not very likely, but being able to have control over the field order
 (other than placing the abstract models in a certain order) is important.
 I make heavy use abstract models to be able to make my apps reusable. The
 above situation is an important roadblock to efficient use.

 A patch with a failing test is attached.
 Haven't yet been able to fix this in a non-intrusive way.
 I would be very much obliged if someone can come up with a solution.

-- 
Ticket URL: 
Django 
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To post to this group, send email to django-upda...@googlegroups.com.
To unsubscribe from this group, send email to 
django-updates+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/django-updates?hl=en.