Hi, I am going to try to create a basic service, after that I'll be able to understand better how the sources work. Until then I have insufficient competence to decide which of the solutions you suggested is better.
My hunch is that it is better to alter the XML schema generation mechanism such that it generates the schema I want, instead of turning off validation or using soft validation (because I will lose all the checks that lxml would otherwise do for me) and parsing the string myself at a higher level. I've used XPath and I even made some screencasts about it: http://www.youtube.com/watch?v=7RiW2g-AKbs http://www.youtube.com/watch?v=PgWfF-Ut0zM I think it would be great if we could make such screencasts about rpclib, that would definitely make it more accessible to inexperienced folks such as myself. My current priority is to focus on the creation of these custom classes, but I will happily accept to work on the video tutorials or write more documentation, if someone else can help me by creating those classes (assuming that they're relatively easy to implement for someone familiar with the intimate details of rpclib). Alex _______________________________________________ Soap mailing list [email protected] http://mail.python.org/mailman/listinfo/soap
