Re: [xslt] Isn't it time to make libxslt multi-threaded?

[Stefan Behnel <stefan_ml behnel de>]

Liam R E Quin, 26.07.2012 14:29:
> libxslt is a f*cking crap optimizer.

Well, it's pretty fast overall, but it's certainly true that it doesn't do
many optimisations. The obvious reason is that these things are really hard
to implement in a low-level language like C. They are hard enough to do in
Java, after all, which still tends to be a more user friendly language than
C (at least for something as high-level as an XSLT processor).

I've been considering to implement an XSLT processor in Python as part of
lxml for a while (i.e. on top of libxml2), simply to make the
implementation of the "obvious" optimisations easier. But so far, I feared
the amount of work that would be necessary in order to get something useful
out of it. And I didn't find anyone who'd pay me for it either ...

Did I mention that this thread is getting slightly off-topic?

Stefan