I agree that "normal" should be the non-NIO sender and NIO sender is for advanced use cases.
thanks, dims On 7/6/07, Asankha C. Perera <[EMAIL PROTECTED]> wrote:
Paul This is what happens in Synapse.. but I do not promote the NIOSender to simple clients as it is designed to send out many requests and not a single message etc. So for a typical client scenario the NIO sender is not suitable from how I see it. asankha Paul Fremantle wrote: > Asankha, Dims, > > I'm wondering what happens if I have the following scenario: > * Anonymous HTTP Req/Resp > * NIO Sender > * Callback in the client. > > How many threads are used? Which pools do they come out of? Are there > any blocking threads? > > Here is what I think should happen: > > The application thread should hand off control to one of the NIO > sender threads. Once the message is sent, no threads should remain > processing anything to do with that request. Once the response comes > back the NIO reciever thread should launch a worker thread to execute > the callback. > > Paul > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
-- Davanum Srinivas :: http://davanum.wordpress.com --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
