[haXe] Xml Namespaces

[Luke Bayes]

Hey Nicolas,

I'm pretty concerned about performance issues with a regexp based XML
parser. Have you found this approach provides acceptable performance? Do you
have a picture of how this affects xml file size boundaries?

Have you considered writing an XML adapter class? Would it make sense to
leave the native parsers in place whenever possible and just wrap the
implementation up behind some consistent interface?

It seems like JavaScript is already pretty slow with XML, and I'm afraid
that a regex parser would make it even slower. I'm also concerned about
going backwards with the SWF xml implementations. I know we can't get
cross-platform e4x support, but the parsing performance would be hard to
give up.

Do other folks have thoughts on this subject? It seems like XML is a pretty
big, integral part of development these days. When discussing XML support
between neko, js and swf, I would push for performance above consistency.
When going from one browser implementation to another, consistency becomes
much more critical.

Until we have a viable cross-environment GUI framework, it seems like a
pretty rare use case that one would take a running SWF application and try
to compile to JS or go the other way.

OK - I have to get to work now!


Luke Bayes
http://www.asserttrue.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.motion-twin.com/pipermail/haxe/attachments/20070703/f79c0634/attachment.htm