I was rereading JCIP last night and it seems that we should implement
CacheImpl.getMembers() by making a shallow unmodifiable copy of the
members field list under the same synchronization lock that we use to do
compound members operations. What we are doing now seems unsafe. If I am
missing the point I would like to be shown why :)
In preparing for rewriting the tutorial for 2.0.0, I've updated the
GUI as well. Could you please get this from HEAD and let me know
what you think. Feedback would be very useful.
1) Check out CVS HEAD
2) ./build.sh dist
3) Unzip the JBossCache-core-2.0.0 distro zip file in dist/
4) In the JBossCache-core-2.0.0 directory, run ./build.sh run.demo
5) Do the above AGAIN in another shell window.
You should now have 2 GUI instances, both with embedded beanshell
panes (thanks to Vladimir for his initial work on this).
Now to write up the documentation for the tutorial, taking users
through the cache API. :-) One of the other things I have done is
update and comment the source code for the GUI and View - since I
expect users to look at this source code as a "live" example.
Please let me know your thoughts.
Lead, JBoss Cache
JBoss, a division of Red Hat
Telephone: +44 7786 702 706