Home Original page

old SimpleXMLRPCLib classes

Jacob Smullyan smulloni at bracknell.smullyan.org
Mon Nov 4 16:19:59 EST 2002 
In article <mailman.1036434666.12451.python-list at python.org>, Brian Quinlan wrote:
> Jacob wrote:
> So the code doesn't follow the signature proposal described at:
> http://www.ampoliros.com/en/dev/sdk/xmlrpc/sysmethodsig.html
  [... example of system.methodSignature returning type information followed ...]

No, it doesn't, because I didn't know of this page or, apparently,
grok the real purpose of system.methodSignature; I thought it
documented the signature of the underlying (Python) callable, not the
xmlrpc method itself.  I can't find definitive specs for the system
methods -- most web resources cite the currently nonexistent
http://xmlrpc.usefulinc.com/doc/reserved.html -- but I now see
that Eric Kidd's listMethods implementation is probably meant to
behave as you suggest.

My module may be correct were it to implement system.methodHelp so
that it return both the Python method signature (as opposed to the
xmlrpc method signature) and the docstring, while ceasing to
automatically implement system.methodSignature; but should I bother?
Does anyone actually call this method programmatically as part of a
discovery routine, or is it only useful as human-readable
documentation, in which case conformance to the "spec", such as it is,
is arguably not of great importance?

Cheers,

js
More information about the Python-list mailing list