Efficient parsing of XML Posts
Hi, I've been delving into twisted.web, looking at the handling of HTTP requests by the HTTPChannel and Request classes. Something I'd like to do is produce a variant for XML processing which starts parsing raw HTTP message body data into SAX events as soon as it comes in. The idea is to avoid having to cache the incoming raw data in memory (or a temp file) and wait for it all to have arrived before parsing it - instead, we would be handling SAX events generated out of the incoming stream of XML data immediately. This would basically be an efficiency tweak for turning around XML messages more quickly and with less of a memory footprint. Has anything like this been tried already? Does anyone know of any good reasons *not* to do it? I'm particularly thinking of web service calls where a custom parser could parse and validate SOAP RPC messages directly as they came down the wire, without ever constructing an intermediate raw data object or XML DOM. Regards, Dominic
participants (1)
-
Dominic Fox