On Sat, Feb 26, 2011 at 12:47 AM, Andy Allan <gravityst...@gmail.com> wrote: > Without fully understanding things - do you? Instead of splitting the
Well, I'm sure that it's theoretically possible to write virtually any code in such a way that I don't need the intermediate state. But it might be less readable. And it's harder. And probably the Actions need to be clearer about their postconditions. For example, it should be possible to know for certain where the new way generated by SplitWayAction will figure in the node's parentWays. (It's probably deterministic, but it's not documented.) Without having access to intermediate state, code is probably also going to be more fragile, because you can't do sanity checks. > way and deleting one half[1] are you not instead figuring out which > nodes to delete from the way? If you look at my commit, I actually wrote two completely different versions - because I couldn't figure out for the life of me why the first one (the one that's not commented out) didn't work. > Given that you know before you start > which end of the ways you'll delete from, then it's probably > possible[2] to do it that way. That works for the "delete all nodes between the junction and the centre of the roundabout" approach. It doesn't work for the "delete the way that is between the junction and the centre of the roundabout" approach, because I don't know which way that is without looking through junction.parentWays. (Could be missing something obvious here.) Steve _______________________________________________ Potlatch-dev mailing list Potlatch-dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/potlatch-dev