RE: [xslt] str:tokenize behavior

["Vakoc, Mark" <Mark_Vakoc jdedwards com>]

>  Okay I finally addressed this in what I think is the correct way.
> 1/ this is expected to be a node set, so it should allocated as such
> 2/ the problem of freeing the tree is independant and addressed by
>  allocating and freeing it with new APIs

I thought I would give you a quick update on my progress with this.
Thus far I have experienced zero problems with the new method.  I have
approximately half of my twenty or so extensions that return node-sets
converted to the new approach.  So far it has introduced no issues in my
testing.  Of note is that my testing thus far is functional and I have
not performed any memory leak testing.

It is taking some time to convert my functions b/c testing them is not
always easy and the new method cannot be implemented using the single
macro I used before to create a (every changing) RTF object so I have to
manually change more of the code (mainly when adding actual nodes to the
nodeset instead of the 'container').

But basically so far I am very pleased with this approach.  I had to
make some minor changes to get stuff to build on win32 like adding the
RVT functions to libxslt.def.src and such.  Also some conversions from
void* to other types need casts.  I will submit the simple patches when
everything is ironed out.

Great work.  I will let you know the final results when I complete
implementation and testing.

Mark