xsltproc and DocBook

[Darel Finkbeiner]

This may be the wrong group, so let me know.

My "problem" is this: I am writing my commentary in DocBook 5 and
using the program xsltproc and the docbook5 XSL stylesheets to produce
XHTML output. Since it is a commentary, it has both English and
polytonic Greek with combining diacritics in it. My console and VIM
are both perfectly configured to allow me to edit such documents in a
very natural and easy way, and one in which I can actually read the
Greek that I've typed in.

After processing with xsltproc, all of my beautiful UTF-8 encoded
Greek is being transformed into butt-ugly entity references.

Now, I suppose, "technically speaking", this isn't an issue when
viewing the html document in a browser.... maybe. But I like to be
able to view and "debug" the resulting file in a text editor as I want
to ... additionally, how am I to be sure that the "correct" UTF-8
code points are being used for crucial combining marks ( and by
"correct", I mean the exact code points that I have chosen to use,
since there are alternatives in the unicode standard )? I
specifically chose XHTML output because it is natively UTF-8, so why
convert them to entities in the first place?

My question is, how do I turn off this "feature"? Or can I? Or
should I use a different XSLT processor?

 

[Joseph Kesselman]

Did you specify UTF-8 as your output encoding in the xslutput directive?

If you did, and you're still getting everything converted to character
references... you may want to try another XSLT processor and see if its
serializer does a better job of taking advantage of UTF-8.

 

[Darel Finkbeiner]

Did you specify UTF-8 as your output encoding in the xslutput directive?

If you did, and you're still getting everything converted to character
references... you may want to try another XSLT processor and see if its
serializer does a better job of taking advantage of UTF-8.

It looks like the output method in the XSL stylesheet sets the
encoding correctly.....

<xslutput method="xml" encoding="UTF-8" indent="no" doctype-
public="-//W3C//DTD XHTML 1.0 Transitional//EN" doctype-system="http://
www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"/>

Any suggestions on a good XSLT processor?

 

[Alain Ketterlin]

My question is, how do I turn off this "feature"? Or can I? Or
should I use a different XSLT processor?

You may try to not even mention XHTML in <xslutput> (make a new
"driver" xslt stylesheet, with only <xslutput> and <xsl:include> of
the other xsl). xsltproc should not use any entity then.

-- Alain.

 

[Darel Finkbeiner]

You may try to not even mention XHTML in <xslutput> (make a new
"driver" xslt stylesheet, with only <xslutput> and <xsl:include> of
the other xsl). xsltproc should not use any entity then.

-- Alain.

Amazing... you were absolutely correct. I changed the output to:

<xslutput method="xml" encoding="UTF-8"/>

And suddenly it worked perfectly. Thanks for the tip, Alain!

 

[Joseph Kesselman]

Amazing... you were absolutely correct. I changed the output to:
<xslutput method="xml" encoding="UTF-8"/>
And suddenly it worked perfectly. Thanks for the tip, Alain!

Note that method="xhtml" is actually not defined in the XSLT 1.0
standard... but since XHTML is an XML language, outputting as XML should
work.