Jeremias Maerki wrote:
While I was investigating this earlier I modified AbstractBreaker to set
the "force" parameter to "false" on the findBreakingPoints() call. I
found out now that this had the interesting side-effect that footnotes7.xml
produces the same footnotes on both page 1 and page 2. It seems that the
footnote handling relies on the force parameter to be set to "true".
Don't know if this means anything. I simply thought I'd share this. I
reverted that change on my working copy.
(sorry for not answering before)
That's quite a strange behaviour ... :-/
I'm going to investigate its causes.
By the way, I also would like to work a little bit more on footnotes.
The recommendation (18.104.22.168.) states that "There may be limits on how
much space conditionally generated areas can borrow from the
region-referencearea. It is left to the user agent to decide these
At the moment, FOP has an implicit limit, as each page will contain at
least a line from the normal flow, but I would like the limit to be
settable in the document itself: as there are documents with many long
footnotes (for example, a commented version of Dante's divine comedy) I
think it should be the user to decide whether he likes more having at
least some (and how many, precisely) content lines in each page, or to
create "footnote pages" so that the next content lines will be placed in a
page where their cited footnote can be placed too.