(Just a status report)
The bugs regarding multicolumns & linebreaks appears to be weeded out.
(linebreaks impossible in upper left cell, bad LaTeX output as reported,
something else I can't remember now... missing rows)
What's left is the problem with the ::Clone() method (which *should* have
been a copy constructor), and the funny behaviour of the cursor in the
Extra options input fields. I think I'll just leave the strange behaviour
with top and bottom lines as it is; it's possible to get a table right with
a little extra work, apart from positioning text in multicolumn cells, it's
wysiwyg..ish.
There's one thing I can't figure out though: If I add a row at the very end
of a table, the previous row retains its bottom line (thus the appended row
is separated from the rest of the table, like the head of the table is). Is
that intentional or a misfeature?
As anyone who spends a few days of roller-coasting on gdb in the table
source code can see, the code is a complete mess. I still don't understand
how it works. I don't see any consensus in how cells are numbered for
example, it varies between different parts of the code, with hacks here and
there to tape things together. ;)
I think it's an excellent example of what happens whith a design that isn't
documented properly (or not documented at all, let alone designed at all).
I don't know if LaTeX2.09 had widths ("\multicolumn{3}{p{20mm}}{.." f.ex.)
on multicolumn cells, but to the best of my knowledge LaTeX2e has not.
LyXTable implements that though -- maybe it's a relic from LaTeX2.09 days,
maybe a misunderstanding somewhere along the way. LaTeX doesn't seem to
barf on it though, so I'll just ignore it.
I'll have a patch ready tomorrow, or the day after that. (knock on wood ;)
Joacim
-
With both feet on the ground, you won't get very far.
-- Loesje
