Hi,

I have both "XPATH 2.0" and "XSLT 2.0" and I've looked for this information, but I'm coming up dry.

I am given the task of working with XML files that must be able to round-trip to Adobe's Structured Framemaker.

For the most part this is not a problem, but the XML generated from Framemaker has a <!DOCTYPE Section [...]>. Inside this section are a <!NOTATION cgm SYSTEM ""> and several <!ENTITY Figure SYSTEM "FigureID.cgm" NDATA cgm> entries.

I'm not sure why FM is using this method for including graphics files and other external files, but they are.

My problem is, I bring in one of these files, do some restructuring work but I'm not able to write it out with the <!DOCTYPE Section...> intact.

The way FM does Cross References is equally as bad, they use <?FM MARKER [Cross-Ref]....?> constructs.

So, writing my modified XML back out so it can be imported into FM is not working because all of these constructs are not being written in the file.

Is these something I need to do so I can work with these constructs?

Thanks,

- m

------------------------
GnuPG Key fingerprint = 1AD4 726D E359 A31D 05BF ACE5 CA93 7AD5 D8E3 A876

Michael Hare

__________________
------------------------
Keep Moving Forward

GnuPG Key fingerprint = 1AD4 726D E359 A31D 05BF ACE5 CA93 7AD5 D8E3 A876

Michael Hare

The <?FM ...?> is called a processing-instruction, and can be written out in XSLT using <xsl:processing-instruction>

e.g.:

<xsl:processing-instruction name="FM">MARKER [Cross-Ref]...</xsl:processing-instruction>

http://www.w3schools.com/xsl/el_proc...nstruction.asp

Entity declarations are not part of the standard XSLT instruction set, but some processors provide extra functionality to output this information.

e.g. Saxon: http://www.saxonica.com/documentatio...s/doctype.html

/- Sam Judson : Wrox Technical Editor -/

Thanks Sam, that helped a lot.

Now, is there a way in Saxon to read all the DocType information from the source document?

unparsed-entity-public-id() and unparsed-entity-uri() do not seem to able to read unknown entities.

------------------------
GnuPG Key fingerprint = 1AD4 726D E359 A31D 05BF ACE5 CA93 7AD5 D8E3 A876

Michael Hare

Unparsed entities aren't widely used and aren't well supported by XSLT. You may find the saxon:doctype extension helpful. Alternatively some people try to generate the DTD of an output document "by hand" using disable-output-escaping.

A better way might be to generate the information needed in the form of regular elements and attributes, and then post-process the output through a SAX filter that converts these elements and attributes into the required DTD constructs.

Processing instructions like <?FM MARKER [Cross-Ref]....?> are fully supported by XSLT and shouldn't cause any difficulty.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference