Thanks Georg,

Those ids are good news. Now I need to experiment a bit, then things wil get 
clearer I think.

Intermediate Format looks like a good choice for such a hack but I've 
downloaded a FOP snapshot (805900) and the ``IntermediateFormatTestSuite`` 
takes forever (running Ant default target under Windows XP with `Ant-1.7.0`). 
Hope this will be fixed soon!

Yesterday evening I had a look at the FOP code with your other proposal about 
Knuth Sequence in mind. Well, that's too hard for me yet and I'll take the 
"easy" (hum) way for the moment.



-----Message d'origine-----
De : Georg Datterl []
Envoyé : jeudi 20 août 2009 10:11
À :
Objet : AW: Referencing multiple pages for index entries

Hi Laurent,

The index page is indeed the very last page, but since I still have the fo 
file, I can add further indizes and cover pages later. I just don't have them 
at the moment and I don't need them for the correct page numbers of the index, 
so I can look for the last page. You might as well use a table to display your 
entries and find the table by id instead of the page. Actually, I wouldn't need 
to find the page at all, since I search for the entries directly in my xpath 
statement, but stripping the unneeded nodes reduces the time needed to find all 
the page numbers from 2 hours to 2 minutes with even medium sized documents!

prod-id is the area tree name for the id attribute of fo-elements. Each 
page-number-citation element gets a (of course) unique id and I keep a map of 
which elements belong to which entries. Then I get the ids and the actual page 
numbers from the area tree and can strip page-number-citation elements from the 
fo file. Using id/prod-id is robust, since it can't be confused by user content 
and fop requires ids to be unique anyway.

Mit freundlichen Grüßen

Georg Datterl

__________ Information provenant d'ESET NOD32 Antivirus, version de la base des 
signatures de virus 4350 (20090820) __________

Le message a été vérifié par ESET NOD32 Antivirus.

Reply via email to