revset: use `subset &` in bare `p2()`

revset: use `subset &` in bare `p2()`

This takes advantage of the `fullreposet` smartness with a nice
speedup. It's a similar speedup to `p1()` when a merge is in progress
(the non merge case is already lightning fast anyway.)

--- a/mercurial/revset.py	Wed Sep 17 10:59:52 2014 -0700
+++ b/mercurial/revset.py	Wed Sep 17 11:00:03 2014 -0700
@@ -1217,7 +1217,9 @@
         ps = repo[x].parents()
         try:
             p = ps[1].rev()
-            return subset.filter(lambda r: r == p)
+            if p >= 0:
+                return subset & baseset([p])
+            return baseset([])
         except IndexError:
             return baseset([])