[PEAK] What's the right way to remove entries from an EntityDM

[Phillip J. Eby]

At 03:40 PM 5/20/2005 -0400, Erik Rose wrote:
>On Mar 21, 2005, at 11:28 AM, Phillip J. Eby wrote:
>>> From what i've seen in it's interface there's no mention on this
>>>subject, nor there is any in QueryDM. I've overritten __delitem__ so
>>>that it works as current __getitem__ does, by the oid.
>>
>>It's entirely up to you.  I personally favor 'dm.remove(ob)', as to me 
>>it's like removing an item from a list or set.  peak.schema will use this 
>>approach when it arrives, as it will treat a database as just a giant set 
>>of objects.  So 'db.add(ob)' and 'db.remove(ob)' will be used to add and 
>>remove objects.
>
>Are you changing your mind, then, about the name of newItem() in EntityDM?

No.  'newItem()' and 'add()' are not synonyms; one creates a new instance 
associated with a DM, the other adds an already-created instance to a database.


>  I used delItem() in the code I'm working on for consistency with 
> newItem(), but if you want to stick an add=newItem in EntityDM to 
> canonicalize it, I'm all for it, as that would, as you intimated, be more 
> Pythonic.

add=newItem won't work, because it doesn't do the same thing.  It's 
probably not a bad idea, however, to create an 'add' method, and perhaps a 
stub 'remove' method.  I'm just a bit reluctant to invest the effort, 
versus working on peak.schema.  However, if you'd like to write up a patch 
(granting all rights, of course), I'll take a look at including it.