Jean-Marc Lasgouttes wrote: > What bothers me with this patch is that I never really understood what > the semantics of niceInsert are (wrt other inserts). So how can we be > sure that the patch does not break something else?
We can be sure that it does not break anything that was not already broken before Michael did fix the "insert delimiters around selection" bug, because the part of the patch that touches niceInsert() simply reverts that fix. What makes me also sure that this patch is correct is the fact that niceInsert() is used in 1.3 unconditionally, too. I am pretty sure that the following description of niceInsert(arg) is correct: Insert the arg at the current cursor possition. If there is a selection, and if arg has cells, put the selction into the first cell of arg. Otherwise, replace the selection with arg. Georg
