Weird. I have a feeling there is some reason for this hiding back
around 2000. Which is wrong now. But I think originally the idea
was to stop filling people's screens full of the stupid irrelevant
inet6 stuff [1]
But the diff is really really crazy. It is like you are avoiding
removing some historical logic. I have a feeling a couple of people
need to dig into this and find why it did what it did, and decide it
is BS.
[1] "inet6 is too good and original, and using arp would be terrible, so
we must invent a new thing that uses multicast, even if it has side
effects that if you enable an interface for inet6, even without an
address, you get a very long line that wraps, and then the person trying
to review their inet4 configuration on a machine with 3 interfaces
cannot read it on a 80x24 screen because the details they are looking
scroll off the top". Really I think this was our logic, which was caused
by a ridiculous design choice.
Jan Klemkow <j.klem...@wemelug.de> wrote:
> Hi,
>
> ifconfig doesn't print ipv6 addresses if its used with media option.
>
> # ifconfig -A
> vio0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
> ...
> inet 10.0.1.65 netmask 0xffffff00 broadcast 10.0.1.255
> inet6 fe80::5054:ff:fe6a:b6fd%vio0 prefixlen 64 scopeid 0x1
> inet6 fc00:1::1 prefixlen 64
> inet 192.168.0.1 netmask 0xffffff00 broadcast 192.168.0.255
>
> # ifconfig -A media
> vio0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
> ...
> supported media:
> media autoselect
> inet 10.0.1.65 netmask 0xffffff00 broadcast 10.0.1.255
> inet 192.168.0.1 netmask 0xffffff00 broadcast 192.168.0.255
>
> As the diff below shows, afp is NULL by default, but set to inet if
> there is an additional program parameter. At the end, no specific
> address family is assumed if afp is NULL. Thus, the diff below
> introduces a new variable to remember if a specific address family was
> set by the user or not for printing all interface addresses.
>
> The regression test of ifconfig(8) is passing with the diff below.
n>
> ok?
>
> bye,
> Jan
>
> Index: ifconfig.c
> ===================================================================
> RCS file: /cvs/src/sbin/ifconfig/ifconfig.c,v
> retrieving revision 1.461
> diff -u -p -r1.461 ifconfig.c
> --- ifconfig.c 18 Jan 2023 21:57:10 -0000 1.461
> +++ ifconfig.c 23 Jan 2023 12:11:30 -0000
> @@ -746,6 +746,7 @@ const struct afswtch {
> };
>
> const struct afswtch *afp; /*the address family being set or asked about*/
> +const struct afswtch *pafp; /*the address family being used for printing*/
>
> char joinname[IEEE80211_NWID_LEN];
> size_t joinlen;
> @@ -840,7 +841,7 @@ main(int argc, char *argv[])
> if (argc > 0) {
> for (afp = rafp = afs; rafp->af_name; rafp++)
> if (strcmp(rafp->af_name, *argv) == 0) {
> - afp = rafp;
> + pafp = afp = rafp;
> argc--;
> argv++;
> break;
> @@ -1216,7 +1217,7 @@ printif(char *name, int ifaliases)
> (ifa->ifa_addr->sa_family == AF_INET &&
> ifaliases == 0 && noinet == 0))
> continue;
> - if ((p = afp) != NULL) {
> + if ((p = pafp) != NULL) {
> if (ifa->ifa_addr->sa_family == p->af_af)
> p->af_status(1);
> } else {
> @@ -3514,7 +3515,7 @@ status(int link, struct sockaddr_dl *sdl
>
> proto_status:
> if (link == 0) {
> - if ((p = afp) != NULL) {
> + if ((p = pafp) != NULL) {
> p->af_status(1);
> } else for (p = afs; p->af_name; p++) {
> ifr.ifr_addr.sa_family = p->af_af;
>
