Dieter, Thanks, you've added a new room to my brain :) Possibly that point should be made in The Book since the discussion of Pluggable Brains comes at the end of the chapter. Cheers, Chris On Fri, 24 Nov 2000, Dieter Maurer wrote:
Chris Gray writes:
The Zope book discusses calling a ZSQL Method via a URL, saying that a request for a URL like:
http://localhost:8080/zsql_method/parameter/value
"will return a result object", but what I get is the index_html document at the root, although I do get a rendered record with: The ZPublisher works as follows:
It traverses the ZODB guided by the components in your URL. It it arrives at the URL's end, it checks whether the reached object is callable (this is the case for all kinds of methods and documents). If it is, the object is called. If it is not, ZPublisher checks, whether the object has (or can acquire) a method "index_html" (this usually is the case). It then calls this method.
In your case:
ZPublisher will reach the result object (as the Zope book describes). Unless you have wrapped this object into a custom class (aka brain, see "Advanced" tab), this is a "Record" object. A "Record" object is not callable. "index_html" is looked for, found in a folder above and rendered.
That's what you see.
What can you do?
Wrap you results into a custom brain that either defines a "__call__" (that would need to be a Python defined class) or an "index_html" (this may be a non-persistent ZClass).
Dieter