Jason Carreira wrote:
Oh shit Jason, you're so right. I really need to document our cache stuff better. I _started_Is there a document describing the semantics of the CacheProvider Interface? For instance:
writing something once and didn't get far....
What does is the lock() method do (is it supposed to be local or cluster-wide), and what are the ramifications if it's ignored?
cluster-wide .... if its not implemented, ReadWriteCache won't be threadsafe. Note that
NonstrictReadWrite and Readonly don't use lock()
This stuff is all handled cleverly by the CacheConcurrencyStrategy, if I'm understandingWhat happens if there is a version conflict on update? Will the cache be flushed? Will it have already been updated?
your question correctly.
How are CacheProvider updates related to transactional boundaries?The CacheConcurrencyStrategy is supposed to preserve transaction isolation, without
requiring that the cache provider be transactional.
Don't have time to say much more right now .... just got to Aarhus.
------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ hibernate-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/hibernate-devel