On Fri, Nov 07, 2003 at 10:56:31AM +0100, Alfredo Braunstein wrote: > Andre Poenitz wrote: > > > Better. Outer text is fully functional, cursor navigation in text inset > > seems to be ok not, even with selection. > ^^^ > Freudian lapsus? ;-)
;-) > > Tables and math are non-functional, neither are mouse clicks in nested > > insets. > > why mouse clicks? Mostly because that's the area where the current architecture differs most from the IL architecture. With IL we had 'four phase' dispatch in the 'complicated' insets: First a switch to try to make sense of mouse events, then calling the_locking_inset->dispatch, then special handling by the inset and finally pass to the LyXText's dispatch. Nowadays we just have one big switch with special inset handling and the default branch dispatches to the LyXText. The first two phases are basically moved to the global cursor. This is much simpler, but the replacement for the first phase is not yet functional. > > I'd like to commit that nevertheless. This is in a better shape than we > > had after the big undo change and the remaining issues should be fairly > > local (some update missing etc) and fixing this kind of stuff works > > better in CVS where e.g. Alfredo has direct access than in patches send > > over the list. > > I still don't understand well how much is to be done, so I cannot really > advise. (I'll give of course a hand with the fixing, though) We are 80% through, apart from the mouse it's just straightforward stuff left. I'll take care of the mouse stuff myself. The rest may take a few days due to the amount of code involved, but generally there are no ugly corner cases involved, so it should get straightened out quickly. I've already math now back in a working state btw. Andre' -- Those who desire to give up Freedom in order to gain Security, will not have, nor do they deserve, either one. (T. Jefferson or B. Franklin or both...)
