Re: Managing EOF caching
Re: Managing EOF caching
- Subject: Re: Managing EOF caching
- From: email@hidden
- Date: Sun, 22 Jan 2006 13:44:10 +1300
Hello Art;
Thank you for this documentation reference -- very interesting.
Snapshots are reference-counted ... So as long as a reference to a
snapshot from any editing context exists, the snapshot will remain.
But when the last reference is broken (e.g., when an editing
context is freed), the snapshot will be freed.
I see what you mean from the EODatabase javadoc in that there is
definitely reference counting supported, but are you sure this is
still used by the EOF machinery in WO5? The thing that makes me
wonder is that EC-centric reference-counting doesn't entirely
correlate to the experience I've had with EC behaviours in WO5.2. I
can keep one EC instance for processes which must load up at least
many hundreds of thousands of EO-s over a long period and there are
generally no memory blow-outs which one might otherwise expect.
I would have *guessed* that a cunning arrangement between weak and
strong EC-EO-snapshot references were being used to ensure snapshots
were allowed to be destroyed appropriately using GC. I would really
love to see some concrete WO5-centric tech note about the workings of
"fetch timestamp lag" and "EC-based snapshot maintenace" so one could
be more certain about some of these aspects of EOF.
cheers.
___
Andrew Lindesay
www.lindesay.co.nz
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden