Hash: SHA1

Chris Rossi wrote:
> On Fri, Jun 5, 2009 at 11:33 AM, Chris McDonough <chr...@plope.com> wrote:
>> - If a "factory" is specified on a route, it will need to point at a
>>   function that had the same call/response convention as a traversal
>>   root factory.  This will break code.  "Context factories" accept
>>   key/value pairs assumed to be items that matched in the URL match.
>>   These would cease working, and would need to be rewritten as root
>>   factories, which accept a WSGI environment.
> My gut feeling is this should be a show-stopper.  Not having easy access to
> the match dict when doing routes just goes too far in terms of breaking
> expectations about how routes should work.

I don't think we'd have to lose the match dict:  one obvious thing would
be to put the dict into the environ before calling the root/context
factory, which would then pick it out and do whatever was appropriate.

> That said, I think the goal of unifying the two methods is very compelling,
> so we should spend some more time thinking about how to accomplish this
> without breaking fundamental expectations with regards to url dispatch. The
> idea of performing a graph traversal starting at a context found by a routes
> match is particularly compelling and reason enough to pursue a unification.
> I'd just suggest thinking about it a little more before starting to move
> code around.  I'll try and devote some brain cycles of my own to the
> problem.

How many folks have real routes-based applications which would need
fixing?  I'm assuming that we would fix anything in our repository;  it
should even be possible to do so in a BBB-compatible way, e.g.::

   def context_factory(environ=None, **kw):
       match_dict = kw.copy()
       if environ is not None: # called as "root factory"

- --
Tres Seaver          +1 540-429-0999          tsea...@palladion.com
Palladion Software   "Excellence by Design"    http://palladion.com
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

Repoze-dev mailing list

Reply via email to