Michelle,
I won't be able to attend the JDO meeting tomorrow, because of two
other, one-time meetings that came up this week.
Michael B.,
you probably also need to attend the EJB3 phone conference, which says
"urgent" and just got scheduled for tomorrow 8 a.m. (it doesn't say how
long it will take).
A quick status update on my tasks:
Michelle Caisse wrote:
Double-check locking in the PMF (Martin)
I've been discussing with Marina and Michael B. a couple of code changes
on the PersistenceManagerFactoryImpl this week. We're closing in on a
final solution, which is bigger than we had anticipated originally.
While we started with just a direct fix of the broken synchronization
code (double-checked locking), we found that the code for the JTA<->PM
bookkeeping can be significantly improved: by using a thread-local PM
lookup, we don't need a 'pmCache' table! (so, no synchronization at all)
Michael, Craig, and Marina, find the details in the email thread:
Re: code review PersistenceManagerFactoryImpl
I'm currently working on a few more code simplifications, which should
be done very soon. Marina will help testing in the appserver; and if
not too much work, Michael B. has offered to test my changes in JDO
(also need some EJBImplHelper changes backported?).
We definitely need a thorough, final code review with Craig before
committing any changes.
[May 13 2005] AI Martin Zaun will investigate JSR 294 (Java 5) to see
impact on enhancer.
Still open (no progress since last week). I hope to get to finish and
test my JDK5 classfile changes next week.
Martin
