> >    http://cr.opensolaris.org/~meem/dhcp-tweaks/
 > 
 > ifconfig.c:
 > 
 >   On the ifconfig changes, have we completely ruled out "foo0:0" as an
 >   alias for "foo0"?  It doesn't seem we have, but the "get_lun()"
 >   removal seems to assume so.  (Most likely not a real issue, but I
 >   didn't work through it completely.)

As we discussed in person, only the undocumented "configinfo" subcommand
is using get_lun() -- but even so, the distinction actually seems right
because ":0" should refer to the address and not the interface.  So, I
think all of this is fine.

 >   4143,4159,4161: what is this?  Why BUFSIZ?  Why does the buffer need
 >   to exist at all?  Wouldn't perror(str) work fine?

Sure seems like it would.  I've updated it.

 > The rest looks fine, though I somewhat question the value of making
 > ifparse understand the "ifconfig foo0 ether" (ether-without-an-
 > argument) format.  That form doesn't set any parameters -- it's a
 > query mechanism -- so it seems pretty unlikely to be found in any
 > configuration file ifparse will have to handle.  I think it's sort of
 > like making it understand "-a".
 > 
 > But I don't care much either way, and being in sync is probably better
 > than not.

Right, I put it there just to keep things in-sync.

-- 
meem

Reply via email to