Yes, that's what I did actually:

Client (DEALER, CURVE) --------- (STREAM) Proxy (STREAM) ------------ (CURVE, DEALER) Worker

The proxy works as a ROUTER/ROUTER since it receives the identity of both client and server on each side. I first register the workers and then for the first message of a client, I pair its identity with the one of a worker, so that the same client is always routed to the same worker.

To ease the test, I have just one client and one worker.

But here, I did not recall the curve proxy topic since the problem arise before the handcheck.


Le 06/12/2013 14:01, Pieter Hintjens a écrit :
Please read my previous emails about proxying the CURVE handshake.
What I said still applies. The proxy has to parse the client's CURVE
commands and reply with valid CURVE commands. Using a STREAM socket it
can read the raw frames but then MUST pass them onto the worker, which
will generate the valid replies, which the proxy must then send back
to the client.

On Fri, Dec 6, 2013 at 12:59 PM, Laurent Alebarde <[email protected]> wrote:
Hi Devs,

I have connected a CURVE-DEALER socket with a ZMQ_STREAM one. Since
ZMQ-STREAM is not a mechanism declared in stream_engine, I don't expect it
creates any mechanism object.

But, if I have not make mistakes, when the CURVE-DEALER socket sends its
greeting, which includes "CURVE", then stream_engine creates an unexpected
curve_client_t object. I assume this object is related to the ZMQ-STREAM
socket since the options are the default ones, while I defined an identity
to the CURVE-DEALER socket.

On the other side, a "NULL" greeting is received, I assume onto the
CURVE-DEALER socket from the ZMQ-STREAM socket, which is also not expected I
think.

test_stream.cpp passes but does it pass only because it is a NULL to STREAM
communication ?

If this is true, I think it is not a desirable behaviour.

Are you aware of that ? Do you confirm ? What do you think about it ?

Cheers,


Laurent.


_______________________________________________
zeromq-dev mailing list
[email protected]
http://lists.zeromq.org/mailman/listinfo/zeromq-dev

_______________________________________________
zeromq-dev mailing list
[email protected]
http://lists.zeromq.org/mailman/listinfo/zeromq-dev

_______________________________________________
zeromq-dev mailing list
[email protected]
http://lists.zeromq.org/mailman/listinfo/zeromq-dev

Reply via email to