From: Martin Sevior (msevior@physics.unimelb.edu.au)
Date: Tue Sep 17 2002 - 11:41:21 EDT
On 17 Sep 2002, j.m.maurer wrote:
> > make format Table compile and not crash.
>
> it compiled here... /me is puzzled :)
I suspect it was gcc 2.96 weirdness
> >
> > uwog, changing just a cell property does not rebuild a table which is
> > needed after a cell property change. I put in a couple methods I wrote
> > to do this.
> >
>
> ah... that explains a lot... I'm not very experienced with anything that
> has to do with the piecetable... but I'm learning ...
In this case it is the formating engine. Changing a strux means that the
fp*Container described by the strux must be destroyed and rebuilt with the
new properties the strux has obtained. (Look in the fl*Layout classes at
what happens in the *ChangeStrux methods) I could have put in a couple of
calls in the fl_CellLayout to destroy and rebuild the entire table
fp_TableContainer upon a changeStrux in the cell (which we have to do
anyway) but I decided that for performance reasons it would be better to
do it externally. There are many cases where a set of strux changes to
cells are made sequentially (like inserting or deleting rows,columns etc)
in which case it would be very inefficient to destroy the
fp_TableContainer on each cell change. On the other hand the entire table
DOES need to be rebuilt so I put in the calls to do that.
Cheers
Martin
This archive was generated by hypermail 2.1.4 : Tue Sep 17 2002 - 11:46:31 EDT