Angus Leeming wrote: > Alfredo Braunstein wrote: > >> Andre Poenitz wrote: >> >>> On Sat, Mar 27, 2004 at 01:42:30PM +0100, Alfredo Braunstein wrote: >>>> btw, we should have a right past-the-end position for dociterator >>>> (and all derivatives). >>> >>> We do. >>> >>> It's an empty cursor slice stack. (which is, btw, different from >>> the last possible cursor position, i.e. really one ++ after this). >>> >>> You get this thing by using the default constructor. >> >> I know. But this is not good enough, because it doesn't allow >> pit = end(); > > If that's (pit == end()), then try (!pit)
No. > If it really is (pit = end()), then try (pit = paragraph_iterator()) That's what I meant. >> --pit; > > Why can't you add the appropriate operator? Because it's clearly impossible ;-) There is no information in pit at all, how could --pit magically come up with the last position? Alfredo
