RE: ancestor/subsequent descendant test, problem resta

Hi

Thank you Ken and Michael for your assistance. It's clear that I didn't
really describe the problem very intelligently so here's another go.

Anongst other stuff my input documents contain sections, and sections can
feature as descendants of other sections (either as children or as lower
order descendants). Certain sections are marked with a particular attribute
which I described as "significant".

I think the key that I missed out is how this significance affects later
processing of these documents: in one stylesheet which runs across these
documents any sections with the significant attribute invoke a new
result-document.

The condition which I am trying to detect is where there will be content
"left over". I suppose you could think of it as a place where the tree model
and the linear model are in conflict.

So in my example

  <!-- F1 -->
    <section>
      <include>
        <section>.A.</section>
  <!-- F2 -->
          <section sig="Y">
            <section>.B.</section>
            <include>
              <section>.C.</section>
            </include>
          </section>
  <!-- F3 -->
          <section sig="Y">
            <section>.D.</section>
          </section>
        </include>
      </section>

In this case everything is OK, there are no hanging sections.

And in my modified example

  <!-- F1 -->
    <section>
      <include>
        <section>.A.</section>
  <!-- F2 -->
        <section sig="Y">
          <section>.B.</section>
          <include>
            <section>.C.</section>
          </include>
        </section>
  <!-- F1 -->
        <section>
          <section>.D.</section>
        </section>
      </include>
    </section>

This case is not OK, the .D. section region is left hanging when the F2
result document is completed and thus it goes back into F1.

All the sections have ids (which I've left out of the example) which can be
used to report a problem.

I hope this is clearer.

Cheers
Trevor

-----Original Message-----
From: G. Ken Holman [mailto:gkholman@xxxxxxxxxxxxxxxxxxxx] 
Sent: Saturday, 28 March 2009 5:15 a.m.
To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
Subject: Re:  ancestor/subsequent descendant test


I'm afraid I cannot readily understand what you want, but I can focus 
on your choice of wording to make an attempt at an answer.