Hey it's me again.

So I'm following the instructions in the Kamailio 5.1.x wiki module page for websocket configuration and it specifically mention to set this block:

onreply_route[WS_REPLY] {

xlog("L_INFO", "[CSeq $cs] Is in WS_REPLY.");

xlog("L_INFO", "[CSeq $cs] Sending $rs $rr to $sel(via[2].host) via $sel(via[2].transport) len: $ml");

xlog("L_INFO", "[CSeq $cs] \n$mb \n");

if(nat_uac_test("64")) {

# Do NAT traversal stuff for replies to a WebSocket connection

# - even if it is not behind a NAT!

# This won't be needed in the future if Kamailio and the

# WebSocket client support Outbound and Path.

add_contact_alias();

}

When I did this and started up Kamailio, it gave me an error:

Jan 3 22:19:43 sjomainkama55 kamailio: ERROR: <core> [core/cfg.y:3309]: yyparse(): cfg. parser: failed to find command nat_uac_test (params 1)

I moved that if statement to a route block and I was able to start it up.

The wiki page for nathelper module specify this function can be run in onreply_route as well:

Tries to guess if client's request originated behind a nat. The parameter determines what heuristics is used.

Meaning of the flags is as follows:

1 - The “Contact” header field is searched for occurrence of RFC1918 or RFC6598 addresses.
2 - the "received" test is used: address in the “Via” header is compared against source IP address of signaling
4 - The Top Most “Via” is searched for occurrence of RFC1918 or RFC6598 addresses
8 - The SDP is searched for occurrence of RFC1918 or RFC6598 addresses
16 - Test if the source port is different from the port in the “Via” header
32 - Test if the source IP address of signaling is a RFC1918 or RFC6598 address
64 - Test if the source connection of signaling is a WebSocket
128 - Test if the “Contact” header URI port differs from the source port of the request (Warning: this is might be legal or even intended combination in non NATted scenarios)

All flags can be bitwise combined, the test returns true if any of the tests identified a NAT.

This function can be used from REQUEST_ROUTE, ONREPLY_ROUTE, FAILURE_ROUTE, BRANCH_ROUTE.

is this some known bug in 5.1.x?

Thanks

Andy Chen