Ahh I was wondering.
Anyways I still think it would make sense to have control over the address.
For example I start a swank server in a virtual windows machine and need to bind it to the virtual network interface given by qemu (which can't be seen outside the physical host), so there would be no need for a ssh setup.
The way I do it now is: (let ((swank::*loopback-interface* "10.0.2.2")) (create-server)) Which is not nice as *loopback-interface* is not exported.
I would rather prefer to do: (create-server :host "10.0.2.2") Which I think is a clean solution.
-Frank
On Wed, 2013-09-04 at 22:45 +0400, Stas Boukarev wrote:
Frank fau@riseup.net writes:
Hello,
Currently create-server binds to *loopback-interface* which happens to be "127.0.0.1". BTW: *loopback-interface* is not exported, maybe it should?
For remote access binding to the loopback is not really helpful. What about adding a :host keyword to create-server to expose the address we want the server to bind to?
I added a patch for your review. The patch is against the latest slime versions available from quicklisp.
It is intended to be used with an ssh tunnel, because there isn't any kind security in the swank protocol.