15:21 <flip214> when swank starts to listen, SETUP-SERVER calls an announce function and passed the TCP port in - but not the address it is listening on.

15:21 <flip214> This makes a difference when ::1 is bound but the client tries to reach 127.1 ...

15:22 <flip214> so a) should the announce function receive the TCP address as well (requiring either new functions or a second return value to LOCAL-PORT)

15:22 <flip214> or b) should we communicate it via a special

15:22 <flip214> or c) should the client "just" test both? what if it tries a.b.c.d, but ::anything would be correct?

15:23 <flip214> I guess a) would be the best solution, but I'd like to hear some feedback.

15:30 <flip214> d) CREATE-SERVER could accept an ANNOUNCE-FN argument (like SETUP-SERVER) that I could use - SETUP-S doesn't take an INTERFACE argument any more

17:47 <luis> a) sounds sensible.

17:47 <luis> Not sure why you it would require new functions or new return values rather than an extra argument to the announce function

17:48 <luis> although, create-server / setup-server will listen on the address that you explicitly asked, won't it?

17:49 <luis> So you should know that a priori.

19:29 <flip214> luis: well, passing an address means knowing whether there _is_ a ::1 or 127.0.0.1.

22:57 <luis> flip214: hmm, the default for *loopback-interface* is "localhost", if the client uses the same approach shouldn't it get the same result? In any case, having the announce receive the address too seems sensible.