On Mon, Dec 03, 2001 at 03:11:42PM +0000, John Levon wrote: > > No, that would require the cursor being a representation of the _path_ from > > the tree root to the current position. Sort of a stack: push inset when > > entering an inset, pop when leaving one. The "actual position" is the top > > of the stack. > > Tree, single path down the tree, they are both no cleaner than the current > system IMHO
Please? The "tree of insets" is already there. That's what LyX is showing all the time: A view on some tree of insets. What is not there is the "single path down the tree", instead each LyXText keeps a "cursor" of its own. We have instead some kind of "distributed cursor": cursor positions in every text inset, that are even there if the cursor is not in the text - just an indication that the implementation does not map the concept very well. And we have a whole lot of fancy "parent retrieval" methods... Andre' PS: > > Is the "full text" an inset? > > no. Why? > Yes, that's ugly, but not really complex ... Nested 'if' and '->' coming in bunches of three or four are complex enough for me... Andre' -- André Pönitz .............................................. [EMAIL PROTECTED]
