Hey Blake - On Jan 30, 2008 11:52 AM, Blake Sullivan <[EMAIL PROTECTED]> wrote:
> This is more of an issue, however the other problem is that the > ServletContext and Session still don't expose the atomic operations of > ConcurrentMap. Neither do they document what object they do lock on. > Without either of these pieces, it is impossible to perform compound > operations on these Objects atomically safely. Sure, makes sense. If we have requirements for direct access to the ConcurrentHashMap atomic APIs, then I agree we'll need our own ConcurrentHashMap (Servlet API isn't going to provide this for us). If the main concern is avoiding the annoying lock contention that results from use of Hashtable in ServletContext implementations, I would prefer to skip adding a new API and instead focus on getting ServletContext implementations to move over to ConcurrentHashMap. Andy
