Jim,
Comments below
Dave Booz
----- Original Message -----
From: "Jim Marino" <[EMAIL PROTECTED]>
To: <[email protected]>
Sent: Wednesday, September 27, 2006 10:11 PM
Subject: Re: interceptors and async invocation
Hi Greg,
Comments below.
On Sep 27, 2006, at 1:32 PM, Greg Dritschler wrote:
In the current implementation the thread switch for an asynchronous
method
invocation occurs after interceptors are given control. Isn't this
backwards?
Won't interceptors need to run on the dispatch thread to
establish transactional and security context, provide monitoring,
etc.?
I don't think so as context will need to be propagated to the thread
the dispatch is made on anyway. Also, the target may be a composite
reference in which case it could be in another VM. Keeping the thread
"switching" in the target invoker simplifies that case IMO.
I would have thought that an interceptor on the dispatch thread is the
means by which the context was set. Seems like you'd need them on
both sides of the dispatch in all cases. For a remote dispatch, the target
would need interceptors to propogate context from the message
to the thread. Maybe you are thinking about some form of optimization
for the local case?
Jim
Greg Dritschler
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
