Niclas Hedhman wrote:
On Tue, May 5, 2009 at 9:54 AM, Rickard Öberg <[email protected]> wrote:
Fair enough. Do you have any suggestions on how to work around the mentioned
issues? Would it be ok if apply() invalidated all the state and did a
refresh(), i.e. all entities are reloaded, but the references are still
valid.
Not sure, need to measure that. Could be a lot depending on how
effective the ES cache is.
What I can say is definitely Ok for me, is that only the Entities that
were modified are invalidated. The modification sets are small. Not
sure though if that helps.
Well, that could be ok. I mean, the modified ones are the only ones we
KNOW will have changed versions and such. After the logs have been sent
we would then have to clear all the modification events and retain all
get-events. Reloading the modified events would then append new
get-events to the existing log, which would be ok.
I think it could work.
/Rickard
_______________________________________________
qi4j-dev mailing list
[email protected]
http://lists.ops4j.org/mailman/listinfo/qi4j-dev
