Rails core - Apr 2009 - odd behavior with accepts_nested_attributes_for on a belongs_to association

A fairly common UI requirement is for a user to be able to choose an
existing object or create a new one, e.g. select or supply an address
for a shipment. The accepts_nested_attributes_for code goes some way
towards supporting this use case, but falls short in a couple of
areas. The basic problem is that the parameters hash might include
both an ''address_id'' value (from a select box of existing
addresses
from which to choose) and an ''address_attributes'' value (from
the
embedded model form), and the behavior of
accepts_nested_attributes_for in this situation is not well-defined
and is in some cases sensitive to the order in which the hash keys are
processed.

I''ve written up a simple rails app with a unit test suite which
defines the current behavior and exposes at least a few oddities
therein:

http://github.com/dball/nested_test/tree/master

There are probably some other corner cases I failed to consider in
this test suite, but it''s at least a start.

I am working around the problem at least partially in my controllers
by massaging the parameters before sending them to the model, but a
more general solution is probably worthwhile. I figured I''d check here
to see if others agree before digging in and trying to patch
active_record itself.

- donald

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---

Donald,

Yes, please create a patch.

Thanks,
Eloy

On Apr 12, 2009, at 12:01 AM, Donald Ball wrote:
>
> A fairly common UI requirement is for a user to be able to choose an
> existing object or create a new one, e.g. select or supply an address
> for a shipment. The accepts_nested_attributes_for code goes some way
> towards supporting this use case, but falls short in a couple of
> areas. The basic problem is that the parameters hash might include
> both an ''address_id'' value (from a select box of existing
addresses
> from which to choose) and an ''address_attributes'' value
(from the
> embedded model form), and the behavior of
> accepts_nested_attributes_for in this situation is not well-defined
> and is in some cases sensitive to the order in which the hash keys are
> processed.
>
> I''ve written up a simple rails app with a unit test suite which
> defines the current behavior and exposes at least a few oddities
> therein:
>
> http://github.com/dball/nested_test/tree/master
>
> There are probably some other corner cases I failed to consider in
> this test suite, but it''s at least a start.
>
> I am working around the problem at least partially in my controllers
> by massaging the parameters before sending them to the model, but a
> more general solution is probably worthwhile. I figured I''d check
here
> to see if others agree before digging in and trying to patch
> active_record itself.
>
> - donald
>
> >

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core+unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---