On Fri, Oct 12, 2012 at 1:53 PM, Kevin Winchester <
[email protected]> wrote:

> Hi,
>
> I have an application under development that uses SSHD TCP/IP Port
> Forwarding for the client and server sides of a communication interface.  I
> have need of two features that I do not believe are possible currently in
> the SSHD client:
>
> 1. At times, I would like to forward the same port on different local
> addresses.  The current TcpipForwardSupport class tracks the forwards by
> port number, ignoring the local address, making this impossible.  As of my
> contribution in SSHD-192, I can replace TcpipForwardSupport with my own
> factory-injected class.  I plan to do this, but I thought I would offer to
> add this functionality to the standard TcpipForwardSupport if that is
> something that would be received favorably.
>


Well, you can always create a patch and attach your patch if you already
have it.  Worst case, it won't be applied ;-)
Though, I'm not sure to understand what you mean by "forwarding the same
port on different local addresses" ...

>
> 2. At other times in my application, I do not really care what local port
> is used for forwarding.  The InetSocketAddress class documentation states
> that one can specify 0 as the port number to request this functionality
> when binding.  I am not sure if this will work through SSHD/MINA.  And if
> it does work, the trouble would be finding out what local port was bound,
> since I would need to connect to it.  Is there any way to do this
> currently?  What about if a new method were added to the ClientSession
> interface, something like:
>
> int startLocalPortForwarding(**String localHostname, SshdSocketAddress
> remote);
>
> that would return the port number bound.  Looking at the code, I guess it
> might require a Future of some sort since the binding is likely not done
> immediately, but that conveys the general idea.
>

There's a really easy workaround though.  You can open a server socket,
check the bound port, close the socket and use the port when calling sshd.
 But yeah, I suppose we should enhance sshd to support that use case.


>
> As with my previous questions, I am happy to contribute in any way I can
> to the development of these features.  I would only request an indication
> that these features would be acceptable, as well as some guidance as to the
> preferred approach for achieving the requested functionality.
>
> Thanks,
> Kevin
>



-- 
------------------------
Guillaume Nodet
------------------------
Blog: http://gnodet.blogspot.com/
------------------------
FuseSource, Integration everywhere
http://fusesource.com

Reply via email to