Re: [xsl] Merging 2 XML's in to 1 output XML (Performance Issue)

[Elliotte Harold <elharo@xxxxxxxxxxxxxxx>]

Michael Kay wrote:

> I think that Elliotte is referring here to the Saxon 6.5.x code base: I'm
> not aware of any issues in Saxon 8.x where there are acknowledged bugs that
> I have said I won't fix (unless there's a good reason, like I'm expecting
> the spec to change).

That's correct. I'm not paying a huge amount of attention to and 
certainly not depending on XSLT 2 until the spec is actually finished, 
and maybe not even then. There's a lot to like about XSLT 1, and it 
seems to solve my needs and the needs of most users pretty well.

XSLT 2 isn't stable yet; and it isn't well understood in the community. 
There's some nice stuff in XSLT 2, but it's hidden under a pile of 
schema-typing and other extraneous features that strike me as not so 
wise. I'm not yet sure if the goodness in XSLT 2 outweighs the badness 
or not. In any case, I see no reason to move forward just yet.

> Saxon 6.5.x includes a version of the AElfred parser, but I don't regard
> that as part of the Saxon code base and I've declined to fix bugs there
> because it's not my code and you have the option of using a different XML
> parser: it was only included for convenience in the days before there was a
> parser in the JVM.

I wasn't referring to AElfred.

> As far as I'm aware the full list of non-conformances in Saxon 6.5.4 is
> documented at:
>
> http://saxon.sourceforge.net/saxon6.5.4/limitations.html
>
> The reason these weren't fixed when I did the last maintenance release was
> that they affected very few users, but fixing them would involve changing
> substantial amounts of code which would almost certainly introduce more bugs
> than they solved (some of them are also in areas where it's far from clear
> what the correct behaviour is). There's no such thing as zero bugs,
> unfortunately.

The specific bug that convinced me to leave Saxon isn't listed on that 
page. If I recall correctly, it involved TrAX failing to output complete 
document fragments when the output result tree fragment did not have a 
single root.

--
o;?Elliotte Rusty Harold  elharo@xxxxxxxxxxxxxxx
XML in a Nutshell 3rd Edition Just Published!
http://www.cafeconleche.org/books/xian3/
http://www.amazon.com/exec/obidos/ISBN=0596007647/cafeaulaitA/ref=nosim