Thanks very much, Aditya for your advice. I really appreciate it. I tried your suggestions.
On Mon, 21 Mar 2011, R. Ermers wrote:
Thanks, Aditya, The first problem is adequately solved by adding []. I'll prepare an example for the second problem.
A related problem is referring to sections and other blocks.
<section id="sec:cows"><title>Cows</title> <para>cows cows cows. As I told in <xref linkend="sec:cows"/></para></section>
The title is mapped as follows (lpath): \xmlsetsetup{#1}{section/title}{xml:section/title}
\startxmlsetups xml:section:title \section{\xmlflush{#1}} \stopxmlsetups
The section is not mapped: \startxmlsetups xml:section \xmlflush{#1} \stopxmlsetups
Adding [\xmlatt{#1}{id}] to xml:section:title does not work of course. Using [\xmlatt{#1}{id}] with xml:section puts the id before the \section command. I think one of the other commands has to be used in such a way that the id is flushed forward after the \section command.
Map section to include both id and title. (Untested)
\startxmlsetups xml:section \section[\xmlatt{#1}{id}]{\xmlfirst{#1}{/title}} % or slightly better % \section[\xmlattdef{#1}{id}{+}]{\xmlfirst{#1}{/title}} The section titles and numbers are doubled now, most likely due to other mappings that. The subsection and subsubsection titles need to be mapped as such, a generic mapping for all section heads is not feasible (I think).
% Then flush para separately % \xmlall{#1}{/para}
This does not work, because the sections contain all kinds of elements, not just <para>. <section><blockquote/><section><para>text</para></section><para>text</para></section> I tried \xmlall{#1} with other elements, but the result is a mess. I need an lpath symbol that refers to all underlying elements. I tried {+}, {*}, {//}, etc.
% or % \xmlconcat {#1} {/para} {\par} % or better, add a contents tag around the contents % <section> <title> ... </title> % <contents> ... </contents>
% </section> % and then use % \xmlfirst{#1}{/contents}
I had thought about this, but it would ruin the docbook structure of the document, and my editor programme would no longer accept it. The programme already complains about the change of