From: Reuben T. <rr...@sc...> - 2004-02-04 18:36:37
|
> > So that's part of our re-org of the existing libraries. > > Yeah. Just means that stdlib will break existing code, slowing uptake. Well, we'll have to minimise that. There's no big problem in duplicating to start with. > I suppose. Given that those functions are likely to be used in groups, it > would be nice to have something really short, but I suppose fn is too likely > to get shadowed by people's locals. Sadly. A lot of the time I'm annoyed by the new naming scheme and having to type table. and string. all the time. > > It's a real pity that next() is linear or map over vectors wouldn't have > > to be different from map over tables. > > It would anyway. Map over vectors needs to miss out the n if present. n is evil. > Also, next() doesn't honour numerical order. Doesn't matter unless the mapped function is assuming something about the underlying list. It shouldn't be. > And why do you say next is linear? Cos it is. Discussed on the list; look at the code if you're not convinced. It's necessary because the table can be rehashed while you're traversing it. I think Squirrel avoids this. -- http://www.mupsych.org/~rrt/ | robber, n. a candid man of affairs (Bierce) |