Closed GoogleCodeExporter closed 8 years ago
jared said (I'm agnostic):
Actually, I think we should remove the create* methods from ForwardingMap, as
Matt
suggests.
That way, the Forwarding* classes forward and do nothing else, which is more
intuitive. Also, we avoid adding three member variables to every ForwardingMap
instance, even though they're redundant in many implementations.
While that's a little inconvenient for people who are currently using the
ForwardingMap.create* methods, it follows the "when in doubt, leave it out"
principle.
Original comment by kevin...@gmail.com
on 12 Oct 2009 at 6:17
Thinking about this, I've realized that none of the similar methods in
ForwardingMultiset (entrySet, elementSet) or ForwardingMultimap (four of them)
are
doing this, so we have a big inconsistency that has to get resolved one way or
the
other.
It also looks like ForwardingMap got these fields in a somewhat accidental
manner.
So it seems like a good idea to back out of this, although how exactly to do it
in
the least disruptive manner is an open question.
Original comment by kevin...@gmail.com
on 16 Oct 2009 at 10:46
unassigning because I'm not working on this.
I still think it should get done for 1.0 if possible.
Original comment by kevin...@gmail.com
on 29 Oct 2009 at 6:43
Original comment by kevin...@gmail.com
on 30 Oct 2009 at 6:45
Original comment by kevin...@gmail.com
on 2 Nov 2009 at 10:16
Original issue reported on code.google.com by
kevin...@gmail.com
on 17 Sep 2009 at 6:42