17 Oct
2013
17 Oct
'13
4:48 p.m.
On 08.10.2013 03:15, Øyvind Kolbu wrote:
Ref the thread from september,
<http://lists.sip-router.org/pipermail/sr-users/2013-September/079760.html>
I've fixed SRV lookups to work when use_dns_failover is on and the
domain has no _sip._udp record. An example is cisco.com:
% host -t SRV _sip._udp.cisco.com
Host _sip._udp.cisco.com not found: 3(NXDOMAIN)
% host -t SRV _sip._tcp.cisco.com
_sip._tcp.cisco.com has SRV record 1 0 5060 vcsgw.cisco.com.
% host -t SRV _sips._tcp.cisco.com
_sips._tcp.cisco.com has SRV record 1 0 5061 vcsgw.cisco.com.
Currently attempting to call an uri at cisco.com will fail, unless
manually forcing the protocol.
While fixing the the issue I refactored a lot of code and fixed a plenty of
comments as the lookup code has clearly been copy pasted more than once.
Patch is attached and should probably be reviewed thoroughly as it touches
pretty core usage.
Testet some more with tls and found a bug in the patch. *proto must be
updated after choosing a protocol:
--- a/resolve.c
+++ b/resolve.c
@@ -1581,6 +1581,7 @@ struct hostent* no_naptr_srv_sip_resolvehost(str*
name, unsigned short* port, ch
he=srv_sip_resolvehost(&srv_name, 0, port,
proto, 1, 0);
#endif
if (he!=0) {
+ *proto = srv_proto_list[i].proto;
return he;
}
}
--
Øyvind