>I think that the best solution for the JXR2 stuff, since I wrote the JXR1
stuff
>and have a decent perspective, is to keep the current parser (it is simple,
>clean, small) and just have it generate SAX events. This way we can have
the
>best of both worlds. A small fast HTML output engine and a more robuts and
open
>XML output so that we can do XSLT (not that the latter will be slower).
>
I think this is the Crux of the issue. There is no is no need to produce
actually XML what is needed is an Open API for handling tree structures as
events. SAX does this and even further XSLT does this as there is no need
for anything other than events to be parse into the engine. Jazilla uses SAX
for it's HTML parsing.
It is even possible to parse in the style sheet as compiled code, and I
think this is defiantly something to be considered and from the reports I've
had of people doing this it works very well.
As far as the generated code goes. I think we're all agreed that checking it
in is a bad thing. As to whether it's a good idea to use it that's a
different issue that need to be decided on a case by case basis. Personally
I think Alexandria's dependency on Castor is a bit excessive for what is
gained but as far as Java parsing goes as long as it's reasonable efficient
and allows us to keep in step with developments in the JDK I don't have a
problem with Antlr (I would say that as I added it ;o))
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
