On 20 Jan 2009, at 19:28, Andreas Delmelle wrote:
<snip />
I think that taking it into account for both (offset and width) is
the correct way to implement it: the block-container should be
offset by 77pt from the page's edge (which is the same as the region-
body's edge here).
... and then the inner block should be offset another 77pt from the
block-container's start-edge (= 154pt from the page's edge), unless
you would reset start-indent to 0pt on the block.
The nested block does get the correct width (the cyan background
should be covered by magenta at the end-edge).
So this is wrong. The nested block should be another 77pt less wide
than it currently is.
I'll debug it to check, but seems like a trivial fix at first glance.
I noticed that
BlockContainerLayoutManager.getNextKnuthElementsAbsolute() assumes
zero for both the X- and Y-offset of the content-rectangle. Not sure
if this is intended, or was a slight oversight.
For a quick-fix, naively copying the code for computing those offsets
from getNextKnuthElements() and replacing the zero arguments passed to
updateRelDims() (both occurrences) seems to produce the correct result
for this case.
A definitive fix should obviously try to avoid the code duplication,
perhaps by computing those offsets in getNextKnuthElements() /before/
the conditional branching to getNextKnuthElementsAbsolute().
Cheers
Andreas
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
