On Dec 22, 2006, at 13:29, Simon Pepping wrote:
Hi Simon,
<snip />
Uncertain status. The linked bug is resolved, but the description does
not match the description of the bug:
<li>
Omitting fo:table-column or having fo:table-column without a
column-width
and attempting to create columns implicitly from the first
table row is not implemented, yet (<a href="http://
issues.apache.org/bugzilla/show_bug.cgi?id=35656">Bugzilla #35656</
a>).
</li>
As far as I can see, the testcases that I enabled with the patch
solving the bug indicate that the described issue basically has been
resolved:
table_table-layout_fixed_2.xml shows that FOP correctly handles
columns in case none were specified
table-column_first-row-width does the same for the column-widths
OTOH, these testcases are *very* basic, namely: only test behaviour
in case there is no spanning going on... I'll see if I can add some
checks for creating columns off of cells spanning multiple columns,
although I'm not entirely sure what the behaviour should be:
<fo:table>
<fo:table-body>
<fo:table-cell starts-row="true"
width="20mm" number-columns-spanned="2">
...
Should we create two implicit columns, and distribute the width evenly?
Currently in the code (TableBody.java), two implicit columns are
created, but the cell's width is not yet automatically distributed.
The latter happens only if the cell spans only one column (no
distribution, but simple one-on-one transfer).
Cheers,
Andreas
