From: corona-dev-bounces@xxxxxxxxxxx [mailto:corona-dev-bounces@xxxxxxxxxxx] On Behalf Of O'Flynn, Dennis
Sent: Tuesday, March 13, 2007 1:32
PM
To: Corona
development
Subject: RE: [corona-dev] Ref
count for container.
See my comments within the original email…
-----Original Message-----
From: corona-dev-bounces@xxxxxxxxxxx
[mailto:corona-dev-bounces@xxxxxxxxxxx]
On Behalf Of Marcin Okraszewski
Sent: Tuesday, March 13, 2007 1:05 PM
To: Corona development
Subject: [corona-dev] Ref count for container.
If I'm correct we currently keep container open ref count based on
users. I think we should have it based on client ID (Corona UUID).
Otherwise user wouldn't be able to open two clients for the same
container.
[O'Flynn,
Dennis] +1
[Glenn]
There is really on one container it is shared through the
ContainerManager. The reference count was so that the container would
remain cached in memory until no users were accessing it. The “close”
really just indicated to the ContainerManager that it should decrement
reference count and if count was zero the container could be removed from the cache.
So, I don’t really care if a user has 1 or 10 clients open I only care
that a user still is using the container and it needs to stay in the cache.
Maybe I’m not understanding something about the implementation…
This also mean that open/close event shouldn't indicate user's
availability. It just shows that a client opened container. For user,
there should be some online/offline events. User goes online when he
opens first client. He goes offline, when all clients are closed.
[O'Flynn,
Dennis] +1
[Glenn]
I thought we were only interested in other Users that were active in my
Container – since all of the views are within in the context of a
container. I’m only interested in users that have the same
container open as I do. This approach gives me the ability to NOT be
available to others in a Containers I don’t have open, even though I may
be “online”
This means that also some "keep-alive" message should be
sent, to detect
crashed clients. Otherwise if a client crashes, the container will
never
close.
[O'Flynn,
Dennis] This should be defined as a RESPOSITORY-DESCRIPTOR to specify some web
service as well as the frequency to post a keep-alive message. It would
be nice to limit only sending keep-alive messages to when the client has not
post any other container events.
What do you think? Should I enter a bug for this?
[O'Flynn,
Dennis] +1
Marcin
_______________________________________________
corona-dev mailing list
corona-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/corona-dev
The contents of this e-mail are intended for the named addressee only. It
contains information that may be confidential. Unless you are the named
addressee or an authorized designee, you may not copy or use it, or disclose it
to anyone else. If you received it in error please notify us immediately and
then destroy it.