Stuart Henderson wrote: > > > > Personally my take on this is that as long as it's just done as -a > > then it's small and simple to implement (pass a string from args to > > auth_userokay), and there's no other way to provide access to this which > > is an important, though lesser-known, part of bsd_auth. We already trust > > auth_userokay with network-supplied strings for this (e.g. as part of > > the username from ssh) so this doesn't seem to add any real exposure > > risk. > > > > Here's an updated version of Renaud's diff against -current after the change > to auth_userchallenge.
I suppose it's time to bend to user demand. ok. But all these auth changes are burning up your reserve line budget. :) > .Bl -tag -width tenletters > ++.It Fl a Ar style > ++The > ++.Fl a > ++(authentication style) option causes And only one set of plus here please. :)