Andre Poenitz wrote:
On Fri, Mar 21, 2008 at 02:24:06AM -0400, rgheck wrote:
OK, here's my attempt once again to fix bug 2178. I've tested this a bit,
but I'm not an extensive user of tables, so I'm sure there are bugs here.
But the general idea seems to work.
Note that this is intended to be a "minimal patch", involving as few
changes as possible to the existing code. It may well be that there are
better ways to do this, longer term, but I am NOT about to start messing
with the InsetTabular code in any serious way. The list of names at the
start of InsetTabular.cpp is enough to scare me off for good. ;-)
Comments welcome.
I would have prefered if the InsetTableCell and the CellData were
combined in a single class (assumng they were always 1:1) and the
whole stuff located in InsetTabular.cpp instead of a separate .h/.cpp
file pair as this is pretty much an implementation detail of table
insets.
But feel free to commit what you have. If it annoys me too much I could
always "fix" it myself ;-)
I think doing what you suggest is indeed best. Merging InsetTableCell
and CellData will mean that the former doesn't need a pointer to the
latter, among other things. But I thought it better to do a "minimal
patch" first so we can make sure it works---fixing 2178 has been an
exercise in frustration---and then fix it up a bit later. But I can
certainly move all the InsetTableCell into InsetTabular.cpp easily enough.
Richard
