On Mon, Mar 14, 2005 at 03:32:28PM +0100, Juergen Spitzmueller wrote:
> Martin Vermeer wrote:
> > I think I know where the problem is, not how to fix it :-(
>
> Me too. It seems that MOUSE_MOTION moves inside nested insets despite of
>
> // only accept motions to places not deeper nested than the real anchor
> if (bvcur.anchor_.hasPart(cur)) {
>
> No idea why.
>
> JürgenIt looks to me that MOUSE_MOTION functions right in insettabular... the problem is in BufferView_pimpl, I think, based on the backtrace. Actually I believe that in workAreaDispatch the whole building of a temporary cursor, dispatching to an inset near it etc. up to and including the call to update, should simply be disabled for LFUN_MOUSE_MOTION... just like what happens a few lines lower. I think this temporary cursor is our villain: it is within the deeper inset but its anchor is not. Educated guess. Does this make sense? - Martin
pgpnM1iUiQVdK.pgp
Description: PGP signature
