possible regression in http tunnel in 3.2.0.CR1?

[Jeff Mesnil-2]

Hi,

I'm currently testing HornetQ 2.1.0.CR1 for the release and we have a
regression on our Servlet examples.
I am wondering if this is not a regression of Netty 3.2.0.CR1.

We get a warning on the client-side, when we close the Netty channel:

      ChannelFuture closeFuture = channel.close();

      if (!closeFuture.awaitUninterruptibly(10000))
      {
         NettyConnection.log.warn("Timed out waiting for channel to close");
      }

This occurs only on our Servlet transport which uses a
HttpTunnelingClientSocketChannel:

      // if we are a servlet wrap the socketChannelFactory
      if (useServlet)
      {
         ClientSocketChannelFactory proxyChannelFactory = channelFactory;
         channelFactory = new
HttpTunnelingClientSocketChannelFactory(proxyChannelFactory);
      }

I'm still figuring out why the channel is not closed properly but I was
wondering if it rings any bell for you...

Do not hesitate to tell me if you need more info.

cheers,
jeff



-----
-- 
Jeff Mesnil
jmesnil at gmail.com
http://jmesnil.net/weblog/

-- 
View this message in context: http://netty-forums-and-mailing-lists.685743.n2.nabble.com/possible-regression-in-http-tunnel-in-3-2-0-CR1-tp5098982p5098982.html
Sent from the Netty User Group mailing list archive at Nabble.com.