Simon Pepping wrote:
I just created the branch Temp_Interleaved_Page_Line_Breaking and
committed code to it. The branch contains a file BRANCH, which states
the purpose of the branch as follows:
This is branch Temp_Interleaved_Page_Line_Breaking. It holds the
development of interleaved page and line breaking. That is, line
breaking is postponed until the page breaker requires the lines of the
paragraph. This allows the following features:
In a word: WOW!
And I thought you had stopped following the FOP project! How wrong was I!
The page breaker can implement both the total fit strategy for a page
sequence and the best fit strategy for one page (or a range of
pages). The latter strategy allows one to have different widths for
I am curious to know how this works. So if I want to allow changing IPD
in a page-sequence, I specify a property on the page-sequence to
indicate best-fit instead of total-fit?
A single total fit strategy for page and line breaking. Such a
strategy allows one to integrate the placement of side floats in the
page breaking calculations. The setup of such a strategy has not yet
been fully understood.
I have been working on this from March to May, and resumed it
recently. The code I submitted passes 222 out of 341 layout-standard
unit tests. Normal paragraphs are formatted properly. The errors are
in mainly in footnotes, list items, tables, and markers.
The branch was copied from revision 519671, when I started this
work. Next I will work on merging the changes of trunk since then into
Plently still to do then. Keep up the good work!