[Zope-CMF] Re: [dev] tseaver-catalog_events branch

[yuppie]

Hi!


Florent Guillaume wrote:
> Jens Vagelpohl wrote:
>>
>>> 1.) The ContentishSublocations adapter is not registered by default, 
>>> just for unit tests. Why?
>>>
>>> 2.) Instead of using a customized ISublocations adapter and 
>>> _recurseOpaques in handleObjectEvent: Can't we just add a new 
>>> subscriber that dispatches to opaque items like 
>>> dispatchToSublocations dispatches to sublocations?
>>
>> I think Tres would know best what the original aim was.
> 
> I think both ways work. Whichever is best is a matter of taste. Having a 
> second subscriber would mean two recursions more or less in parallel 
> (even though in practice I don't think it would matter); I think I 
> prefer the new ISublocations adapter way.

This is the code on the branch:

http://svn.zope.org/CMF/branches/tseaver-catalog_events/CMFCore/CMFCatalogAware.py?rev=67268&view=markup

There are two reasons why I don't like the new ISublocations adapter:

1.) ISublocations might be used for other purposes as well. Adding 
opaque items to the list returned by sublocations seems not to be in 
sync with the way the adapter is implemented in Zope 3.

2.) It relies on the fact that SimpleItem implements a dummy 
objectValues method.


But the more important issue is that currently no subscriber is used. 
Calling _recurseOpaques from handleObjectEvent *is* nasty and the XXX 
comment should be resolved.

>>> 3.) Does this branch make the unmerged part of the 
>>> efge-1.5-five-compatible branch obsolete?
>>
>> Not sure what's in there.
> 
> I haven't looked at it in a while or compared it to Tres's, but keep in 
> mind it was just an early experiment to "show the way", I'm pretty sure 
> Tres's work obsoletes it.

So either someone has to compare those branches or the 
efge-1.5-five-compatible branch can be removed without looking at it again.


Cheers,

	Yuppie