Attention is currently required from: fixeria, laforge, osmith.

pespin has posted comments on this change. ( https://gerrit.osmocom.org/c/libosmo-sccp/+/35242?usp=email )

Change subject: vty: Retrieve IP addr set from sk when dumping xUA server ......................................................................

Patch Set 2:

(3 comments)

File src/osmo_ss7_vty.c:

https://gerrit.osmocom.org/c/libosmo-sccp/+/35242/comment/fe80e270_327033cf PS2, Line 613: char buf[OSMO_SOCK_MAX_ADDRS * INET6_ADDRSTRLEN + OSMO_SOCK_MAX_ADDRS + 2 + 6 + 1];

maybe put it into a define? see also https://gerrit.osmocom. […]

Done

https://gerrit.osmocom.org/c/libosmo-sccp/+/35242/comment/2ac1d177_4ddbc071 PS2, Line 629: <error>

May we let the VTY user know a bit more than:

I really expect this to never happen, because it's only going over here if the fd is set. If unset, the ips are printed from config. So this is only some hint to look at something really weird happening internally. It shouldn't happen.

May we let the VTY user know a bit more than:

I can print some UTF flower emojis too if you want. But probably not worth the effort given that we have APIs already printing stuff in a "standard" or usual way.

https://gerrit.osmocom.org/c/libosmo-sccp/+/35242/comment/f538093a_877e8769 PS2, Line 632: num_hostbuf = ARRAY_SIZE(hostbuf);

From the user perspective, it would be more practical to see those which fit into the buffer, and so […]

Because in libosmocore I'm already using a strbuf and it was easy. In here I'd need to do complex snprintf sutff or change to a strbuf.

This is all really hypotetical cases which I don't expect to happen in general, I think our time is better spent here on other things. The socket won't have more than 32 addresses because to start with internally we limit to it during osmo_sockaddr_multiaddr_init2(), and iirc the kernel sctp stack also limits it to 32.

If someone ever really gets into that, I bet we can then and only then spend time fixing this kind of stuff.