[XML-SIG] PyDOM performance

[Fred L. Drake, Jr.]

Mike Olson writes:
 > I was quite suprised by this.  I don't know enough about the pydom
 > internals to explain why it is slower.  I just always assumed it

  PyDOM pays a *huge* penalty in two places: the proxies used to avoid 
circular references cause a lot of object creation/destruction when
using the document, though I'm not sure it affects construction time
so much.  It also used instances for the internal data format, where
perhaps only lists, tuples and dictionaries are really needed (at the
expense of making the code more obscure).
  I'd love to see the proxies disappear, and just require explicit
calls to a .destroy() method, but that means another massive code
change.


  -Fred

--
Fred L. Drake, Jr.	     <fdrake@acm.org>
Corporation for National Research Initiatives