On Mon, Nov 03, 2003 at 01:13:47PM +0100, Lars Gullik Bj�nnes wrote:
> Andre Poenitz <[EMAIL PROTECTED]> writes:
> 
> | This is just for the record:
> >
> | The attached patch is a long shot at getting rid of inset locking.
> >
> | <stifflip mode="british understatement">
> |     Slightly off the mark, though.
> | </stifflip>
> >
> | It should apply against last Thursday's CVS but is full of conflicts
> | now. I won't bother updating it as it completely breaks anything
> | releated to mouseclicks (positioning and selection) in nested insets.
> >
> | However, it should convey a rough idea of what needs to be done and
> | what can be gained... 
> 
> Oooo big!
> And not only locking code, but reindentation, ws changes,

I humbly follow your lead *cough*

> formatting, other code cleaup, obfuscation etc. :-)
> 
> This certainly needs splitting up.

No, it will simply die. The positioning stuff is completely broken.
No way to fix... 

The main problem here is that in this praticular case the mathed
approach is wrong, so I can't simply copy stuff. In mathed we have
everything relative to the top left of outermost inset.  Translated to
the whole of LyX this would mean 'relative coordinates to upper left of
document', i.e. absolute coordinates. 

Unfortunately we can't do that as updating these is to expensive (as
seen by the recent gain when switching from absolute row.y to relative
row.y.), so in the end we need to stay with the current inset/*.C way to
have relative coordinates within each inset (and adopt that for IU in
mathed/*...). This unfortunately needs some 'coordinate adjustment' from
time to time, so 'simply removing code' does not work too well...

Andre'

Reply via email to