Stuart Marks wrote:
Hi all,
Please review an updated webrev for this bug:
http://cr.openjdk.java.net/~smarks/reviews/6896297/webrev.1/
Using ConcurrentHashMap is much nicer in many ways, and it seems to
resolve the JCK failures at least as well as the previous fix did.
This does nothing to remedy the issue of snapshot consistency, which
seems to be covered by 4945726, which we'll postpone (again) to a
future release.
This is nicer and apologies for raising the concern about serialization
compatibility - I incorrectly thought that these tables were declared as
HashMaps.
At L242 you might want to include a comment to remind a future
maintainer why it's a zero length array rather than using size().
-Alan.