Description of problem:
In a document where the <indexterm> tag is used, all generated PDF files have a line split after the indexed term. See "Steps to Reproduce" below. This does not occur in generated HTML.


Version-Release number of selected component (if applicable):
Publican 2.1 - on Scientific Linux 6.0 (RHEL 6.0)
Publican 2.3 - on Windows 2000 & XP.
Publican 2.5 on Fedora 14.


How reproducible:
100%


Steps to Reproduce:
1. Create a new book.
2. Enter the following in one of the chapters:

<para>Blah blah blah publican<indexterm><primary>Publican</primary></indexterm> blah blah blah...</para>

3. Build pdf document.

  
Actual results:

The generated xsl/fo is:

<fo:block ...>
    Blah blah blah publican
    <fo:block ...>
        <fo:wrapper id="id3189566">
            <!--Publican-->
        </fo:wrapper>       
    </fo:block>
    blah blah blah...
</fo:block>


And the text in the PDF looks like this:

Blah blah blah publican
blah blah blah...


Expected results:

Blah blah blah publican blah blah blah...



Additional info:
There appears to be an <fo:block> wrapped around the <fo:wrapper>, this causes the linefeed after the indexed term. This is caused in pdf.xsl in which there is a template commented as being copied in from index.xsl from Docbook XSL version 1.72. The template name is "indexterm".

This template has the <fo:block> stuff added in two places which is not in Docbook XSL version 1.75 plus the test on $axf.extensions in Publican is different from that in Docbook which checks both $axf.extensions and $fop1.extensions.

If I comment out the two <fo:block> lines in pdf.xsl, it works as I wish, however, that may introduce other troubles elsewhere.