sb-bsd-sockets:get-host-by-name gives NIL with only loopback active
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
SBCL |
Confirmed
|
Low
|
Unassigned |
Bug Description
I cannot readily start SWANK while being on the train, because the defined TCP
port isn't used.
Investigating gives this:
sb-bsd-
(lo, 127.0.0.1) is active.
As soon as I give eth0 an ipv4 address (no matter which one, even with link
down) it works again.
SBCL 1.0.47.0.debian
Linux 2.6.38-2-amd64 #1 SMP Thu Apr 7 04:28:07 UTC 2011 x86_64 GNU/Linux
Here's a transcript:
(defparameter *l* "127.0.0.1")
(defun resolve (n)
(car (sb-bsd-
(let ((s (make-instance ' sb-bsd-
(sb-bsd-
(print s))
;;; result:
(let ((s (make-instance ' sb-bsd-
(sb-bsd-
(print s))
0: (RESOLVE "127.0.0.1")
1: (SB-BSD-
1: SB-BSD-
1: (SB-BSD-
1: SB-BSD-
#<SB-BSD-
;; when ok:
0: (RESOLVE "127.0.0.1")
1: (SB-BSD-
2: (SB-BSD-
2: SB-BSD-
2: (SB-BSD-
2: SB-BSD-
2: (SB-BSD-
2: SB-BSD-
1: SB-BSD-
1: (SB-BSD-
1: SB-BSD-
0: RESOLVE returned #(127 0 0 1)
#<SB-BSD-
Changed in sbcl: | |
importance: | Undecided → Low |
status: | New → Confirmed |
It's not a DNS problem; 1) the eth0 link isn't necessary; 2) 127.0.0.1 doesn't need /etc/hosts; 3) the name of the machine would return 127.0.1.1 via /etc/hosts.