John W. Shipman wrote: > For me, one of the big payoffs of validating XML files as they > are input is that it eliminates a lot of error-checking logic > from my application. I never have to worry, 'what if this > attribute is missing or isn't a valid number?' I can just > grab the attribute and coerce it with int().
That's why I added parse-time XML-Schema validation support to lxml 2.0, especially for lxml.objectify. It's the easiest way of making sure the object structure the parser returns is as expected. > May I suggest that if there is some validating parser added > to the library, stick with a basic set of content types that > cover the vast majority of needs. Careful applications > coders will not mind the additional overhead of validating > the rest. I think that's how additions to stdlib work (or at least should work) in general. Batteries included does not mean you also get a free power socket. Stefan _______________________________________________ XML-SIG maillist - XML-SIG@python.org http://mail.python.org/mailman/listinfo/xml-sig
