slirp and erlang nodes

[Per Hedeland]

Mark NG <markn@REDACTED> wrote:
>I have a small client/server program written in Erlang (so that there is no
>need for sockets).  But I can't get the server and the client program to
>recognise each other when they're run over slirp.  ie: the server is on the
>slirp host, and the client is on the other end of the modem.

Well, slirp is nasty stuff:-) - if I remember right, the "slirp client"
doesn't have its own IP address, but uses the "slirp server"'s address,
right? I'm not sure whether you can get the Erlang distribution to work
at all between the hosts in that scenario - your best bet would probably
be to somehow fool the nodes into thinking that they're running on the
same host.

The distribution mechanism requires that no node is known by more than
one name, and because of this the host part of any node name must be the
"canonical name" of the host - and of course looking up the host part
must yield an IP address that can be used to connect to the node
(together with the port number received from epmd on that node's host).
If you can make all this be true in your slirp setup, it should work - I
think...:-)

--Per Hedeland
per@REDACTED