Message194747
| Author |
pitrou |
| Recipients |
eli.bendersky, flox, jcea, pitrou, python-dev, scoder |
| Date |
2013年08月09日.15:53:22 |
| SpamBayes Score |
-1.0 |
| Marked as misclassified |
Yes |
| Message-id |
<94532443.53068424.1376063596900.JavaMail.root@zimbra10-e2.priv.proxad.net> |
| In-reply-to |
<1376062235.29.0.365491025686.issue17741@psf.upfronthosting.co.za> |
| Content |
> > Unless I'm reading it wrong, when _setevents() is called, the
> > internal
> > hooks are rewired to populate the events list, rather than call the
> > corresponding TreeBuilder methods. So, yes, there's a TreeBuilder
> > somewhere, but it stands unused.
>
> Yes, you *are* reading it wrong. For example, the "start" callback
> calls self._start_list, which in turn calls self.target.start(),
> thus calling into the TreeBuilder. That's the thing that constructs
> the elements that the IncrementalParser collects in its events list.
Ah, indeed, my bad. That said, using a custom TreeBuilder-alike
would necessitate changes in the C implementation of XMLParser.
(which, I suppose, is why the _setevents hack exists in the first
place) |
|