I don't know about the guarantees that implementing j.u.Map implies, but
for persistent maps, the iteration sequence is definitely stable.
(zipmap (map kfn (keys m))
(map vfn (vals m)))
is a common idiom.
If there is no such guarantee for j.u.Map, maybe we should document it for
clojure
Maybe I'm just paranoid, or maybe I'm missing something, but there are some
facts about maps that I believe to be true in general, that I'd like to
count on, and that as far as I can tell are not explicitly documented as
being true. What it comes down to is whether the same (identical, not just
