Jim Fulton  <[EMAIL PROTECTED]> wrote:
> Note that I prefer:
>     c = IMenuItems(content, None)
>     if c is None:
>        return ''
>     ...
> > than testing directly for the provided interface, as it gives an  
> > opportunity for an adapter to do its job. It's probably slower  though, 
> > and maybe not a pattern generally used for marker interfaces.  Opinions ?
> IMO, testing for an interface is sometimes preferable to using an adapter.
> Some people feel very strongly that you should never test for an interface
> -- I don't.
> It seems silly to add an adapter just to avoid using a test
> (for religious reasons iow).
> In particular, providing adapters to handle cases where an object
> doesn't provide some service and can't really be adapted to provide
> a service seems really silly to me.

I agree with that. My intent is more to provide "adaptation point" where
you are really allowing something to extend your framework
(aspect-oriented style).


Florent Guillaume, Nuxeo (Paris, France)   CTO, Director of R&D
+33 1 40 33 71 59   http://nuxeo.com   [EMAIL PROTECTED]
Zope3-users mailing list

Reply via email to