[usocket-devel] "principle of least surprise"
Chun Tian (binghe)
binghe.lisp at gmail.com
Wed Feb 2 06:44:00 UTC 2011
Hi, James
Your patch was accepted. Actually I consider it as a bug, because we does have HOST-TO-HOSTNAME translations in SOCKET-CONNECT function on CCL, but we forget to do the same thing in SOCKET-LISTEN. This is also partly because I personally do UDP much than TCP so that I call SOCKET-LISTEN rarely.
Any way, thanks very much! I've committed your change as revision 547 in USOCKET trunk.
Regards,
Chun Tian (binghe)
在 2011-2-1,22:26, James Ashley 写道:
> This stupid-simple patch lets you pass a wider variety of address specifications
> into socket-listen (Clozure's make-socket seems to require either an int or a
> string address).
>
> Well, I think it does, and I don't think it has much performance implications.
> But I'm a CL noob, and I won't pretend to know all that much about socket
> programming, so I could be totally wrong about this.
>
> Thanks,
> James
>
> --- backend/openmcl.lisp~ 2011-01-13 18:57:39.000000000 -0600
> +++ backend/openmcl.lisp 2011-01-30 10:11:39.000000000 -0600
> @@ -118,6 +118,7 @@
> (backlog 5)
> (element-type 'character))
> (let* ((reuseaddress (if reuse-address-supplied-p reuse-address
> reuseaddress))
> + (real-host (host-to-hostname host))
> (sock (with-mapped-conditions ()
> (apply #'openmcl-socket:make-socket
> (append (list :connect :passive
> @@ -126,7 +127,7 @@
> :backlog backlog
> :format (to-format element-type))
> (when (ip/= host *wildcard-host*)
> - (list :local-host host)))))))
> + (list :local-host real-host)))))))
> (make-stream-server-socket sock :element-type element-type)))
>
> (defmethod socket-accept ((usocket stream-server-usocket) &key element-type)
>
> _______________________________________________
> usocket-devel mailing list
> usocket-devel at common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/usocket-devel
More information about the usocket-devel
mailing list