I _thought_ that read-only and timeout were added to implement entities
that were 'read-mostly' - things that might be updated outside of the
EJB container occasionially (like a product catalog, say), while keeping
the caching advantage of commit option A. If I'm remembering this
correctly, this would mean that the timeout would extend accross
transactions - maybe overriding commit option B to some extent?
However, if I am remembering correctly, then the timeout should be more
like 300 seconds than 300 milliseconds.
-danch
Dain Sundstrom wrote:
Does anyone remember who originally wrote the time-out code or know the
original goal?
I am working on adding read-only to relationships, and have some questions
on how time-out is supposed to work.
Once a read-only field is loaded in a transaction, is it supposed to be
valid for the length of the transaction, or only for the amount of time in
time-out (300 ms by default)?
If we are in commit option A, should the time-out extend across
transactions?
If we have a locking-strategy enabled (select-for-update is currently the
only strategy), should the time-out be ignored within a transaction (i.e.,
the row is locked so why reload)?
Any help would be greatly appreciated. I'm just guessing right now.
-dain
___
Jboss-development mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-development
___
Jboss-development mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-development
