--- Comment #2 from Andreas L. Delmelle <>  2009-03-22 
02:05:00 PST ---

Debugging this, I see that the last page-master triggers a restart of the
algorithm, but the restarting point is assumed to be the last span change. The
other span changes already added the areas for the earlier blocks to a page,
but that page is discarded. Since the layout for the earlier parts was already
considered 'finished', they are not revisited and disappear.

The issue presents itself always in case of span changes on the last page, when
a last page-master is present. Only the last span will be retained in the

A possible fix may be to treat span changes slightly differently than they are
now, if there is a last page-master. Instead of triggering addAreas()
unconditionally for the whole list, we might be able to limit that to
'everything up to the last break preceding the span change'. 

For the moment, that's still a thought, though... Limiting the list for which
to add the areas seems easy enough, but that's only part of the story. It seems
possible, but I still have to investigate whether it is also doable.

Configure bugmail:
------- You are receiving this mail because: -------
You are the assignee for the bug.

Reply via email to