Pages printed twice in index of PDF.

[Martin Miehe]

I made some further investigations:

Zitat von Bruce Dubbs <bruce.dubbs at gmail.com>:

> Martin Miehe wrote:
>> In the indices of current stable book (and in 6.3 as well) are several
>> page numbers printed twice after one entry. E.g. at
>> Programs -> acinstall
>> Programs -> aclocal
>> Programs -> addpart
>> Programs -> autoconf
>>
>> and so on.
>>
>> Has this been done purposely or is it a side effect and should be fixed?
>
> Moving to -dev.  Please answer there.
>
> This is only an artifact of pdf, not the html.

Indeed. It is done by fop. There are some XML-PDF-converters which  
handle page numbers better, but they are not free. In the HTML version  
are no page numbers at all.

> Looking at  automake.xml, there is
>
> <varlistentry id="aclocal"> ...
> <indexterm zone="ch-system-automake aclocal">
>
> and
>
> <varlistentry id="aclocal-version"> ...
> <indexterm zone="ch-system-automake aclocal-version">
>
> but aclocal-version is not in the index (pdf or html).  I think that  
> >  this is an
> xsl problem, but I don't know how to fix it.

aclocal-version is in the index with version replaced by the version number.

The reason why the page numbers are printed twice is that the  
attribute zone contains two ids. The ID of the entry aclocal itself  
and the ID of the chapter where aclocal is treated.

The fast&easy way to get rid of duplicate page numbers is to throw out  
one of the IDs.

There is a more sophisticated way: By applying the parameter  
make.index.markup at the appropriate places in a first pass of fop a  
PDF is generated that contains the index as XML code. This code can be  
extracted and merged with the lfs-full.xml. Of course <index/> has to  
be removed. In a second pass of fop the final PDF will be generated.
More informations are available at:
http://www.xml.com/pub/a/2004/07/14/dbndx.html?page=2
http://docbook.sourceforge.net/release/xsl/current/doc/fo/make.index.markup.html
http://www.cranesoftwrights.com/resources/bbi/index.htm

I am in no way deep enough involved in LFS to decide what shall be  
done at this point. What is the best?

Regards,  Martin Miehe.