> it's really about turning XMLParser's "push" API for events (where the events are pushed into the target object by the parser calling the appropriate methods), into an iterparse style pull API where the events can be retrieved via calls to read_events().

Sorry, but this is ignoring the fact that the target is an *inherent* part of this feature. In fact, the push API of XMLParser forms an integral part of the design. The events that are being collected are a mixture of what the XMLParser produces and what the target produces. You can't think this feature without these two parts.