[Zope3-Users] Re: What's the rationale of @@ ? Was: index.html

[Jim Fulton]

Jeff Rush wrote:
> Frank Burkhardt wrote:
> 
>>
>> On Tue, Dec 13, 2005 at 02:02:05PM +0100, Andreas Elvers wrote:
>>
>>> what the rationale of @@ anyway ? I've searched for a definition for
>>> this in the Developer Handbook and elsewhere but it's hard to search
>>> for @@ and view.
>>
>>
>> @@ marks a given path component explicitely as a view of the previous 
>> object.
>>
>> Example:
>>
>> http://zopeserver/contents.html
>> Could mean either "The object named 'contents.html' in the root 
>> folder" or
>> (if there is no such object) "The 'contents.html' view (List of contained
>> objects) of the root folder" - the object is preferred.
>>
>> http://zopeserver/@@contents.html
>>
>> Will always display the list of contained objects of the root folder - 
>> no matter
>> if there's a object 'contents.html' or not.
> 
> 
> Hmm, but the '@@' is optional, as I can leave it off and still get the 
> contents.html view if there is no object named contents.html in that 
> folder.
> 
> This would seem to confuse developers, as if there IS an accidental 
> collision of an item name with a view that is unknown to the developer, 
> it will return the view out of nowhere.

No, it will return the view.

> Shouldn't the '@@' should be *required* on views to avoid this ambiguity?

This is a matter of policy. It is straightforward to change. Many people hate
@@ and the ++view++name syntax.  They will often arrange their item and view
names so that they don't overlap.

It might be nice to have easily-configured alternate policies.  For example:

- require @@, as you suggest

- log a warning, or possible raise an error in development mode, if, when you
   get an item, you also find a view with the same name.

Jim

-- 
Jim Fulton           mailto:jim at zope.com       Python Powered!
CTO                  (540) 361-1714            http://www.python.org
Zope Corporation     http://www.zope.com       http://www.zope.org