--- Comment #3 from Andreas L. Delmelle <> 2011-03-23 
13:10:19 EDT ---

Without even looking at the code, I can almost feel what is the issue here (see
also my notes about the partial cause for Bugzilla 49801):
ElementListUtils.calcContentLength(), which is used in list-layout to generate
the 'list-parts'.
During element-list buildup, this method calculates a content-length that does
not take into account space-resolution (i.e. all the spaces are still
unresolved SpaceElements).
By the time the algorithm has to compute the most optimal break, those elements
will have been resolved. Some may have disappeared, others add a forced space
of a certain width.

Something similar is bound to pop up for table-cells, sooner or later, as it
also uses calcContentLength().

BTW: it just occurred to me that we may want to have that method return a
MinOptMax, rather than an int. As such, it could more easily take into account
stretchability of spaces, without necessarily having to know more about the
actual breaks or space-resolution...?

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

Reply via email to