Docbook XSL and RELAX NG documents

[Matthew Burgess <matthew@linuxfromscratch.org>]

Hi all!

I'm trying to process an XML document using the Docbook XSL stylesheets 
(1.67.2).  The document itself is a valid Docbook RELAX NG ('gin') 
document - validated using both 'xmllint' and 'jing'.

Processing the document with:

`xsltproc --xinclude --nonet --output index.html 
/usr/share/xml/docbook/xsl-stylesheets-1.67.2/xhtml/docbook.xsl index.xml`

and

`xmllint --xinclude index.xml --output index-big.xml && java 
org.apache.xalan.xslt.Process -in index-big.xml -xsl 
/usr/share/xml/docbook/xsl-stylesheets-1.67.2/xhtml/docbook.xsl -out 
index.html`

both fail with lots of warnings about templates not found:

"Request for title of element with no title: book
No template matches book.
No template matches info in book.
No template matches title in info.
...
(repeat for every occurence of a Docbook tag!)
"

Now, google suggests, from the first error message, that this might be a 
problem with the implementation of exsl:node-set() 
(http://sources.redhat.com/ml/docbook-apps/2004-q3/msg00405.html). 
Having got exactly the same problems with both 'xsltproc' and 'xalan', 
I'm not sure that's what I'm suffering from here.  The comments in 
xhtml/docbook.xsl suggests the use of exsl:node-set() to strip the RELAX 
NG namespace is a hack, albeit a necessary one.

So, has anyone successfully processed a Docbook RELAX NG document using 
the Docbook XSL stylehseets?  If so, which XSLT processor did you use, 
and can you see anything obviously wrong with the XML files under 
http://svn.linuxfromscratch.org/viewcvs.cgi/branches/LFS-RNG/BOOK/ that 
would cause the problems I'm seeing?

Thanks in advance,

Matt.