[comp.unix.aux] Need bi-directional getty for A/UX 1.1

alexis@dasys1.UUCP (Alexis Rosen) (07/14/89)

One of the big problems with SVR2 is that you can't easily use one line for
both dialing in and dialing out. That's exactly what we need to do, though.
So the question is, how can I do that with A/UX 1.1?

I have been told that what we need is a "bi-directional getty". Sounds good
to me. IF it works.

Another alternative is HoneyDanBer UUCP. Does it exist for the Mac? Does it
solve the problem?

One hack which occured to me is that when we want UUCP to dial out, we tell
it to look at the line, and if it's not used, modify the inittab file so
it won't respawn getty on that line, then kill the getty, and then get on
with things. As soon as it's done it restores the inittab to its original
state. Is this crackbrained scheme feasible? What about reasonable? Is there
a better way?

HELP!!!

Thanks,
Alexis Rosen
temporarily at uunet!actnyc!jsb
or try alexis@rascal.ics.utexas.edu
PLEASE DON'T reply to this account. Mail here doesn't work too well.

debra@alice.UUCP (Paul De Bra) (07/16/89)

In article <10219@dasys1.UUCP> jsb@actnyc.UUCP writes:
}
}One of the big problems with SVR2 is that you can't easily use one line for
}both dialing in and dialing out. That's exactly what we need to do, though.
}So the question is, how can I do that with A/UX 1.1?
}
}I have been told that what we need is a "bi-directional getty". Sounds good
}to me. IF it works.
}
}Another alternative is HoneyDanBer UUCP. Does it exist for the Mac? Does it
}solve the problem?

Nope, it doesn't really matter which uucp you have. You need a bi-directional
getty (usually called uugetty, sometimes a combination of getty and ungetty)
together with a uucp that knows how to deal with that uugetty.

}One hack which occured to me is that when we want UUCP to dial out, we tell
}it to look at the line, and if it's not used, modify the inittab file so
}it won't respawn getty on that line, then kill the getty, and then get on
}with things. As soon as it's done it restores the inittab to its original
}state. Is this crackbrained scheme feasible? What about reasonable? Is there
}a better way?

I have used similar tricks with older Unix versions (SCO Xenix 2.2, BSD4.2,
etc.)
The safest way is not to "modify" the inittab file but to keep 2 copies
of it, one with the line enabled, another with the line disabled, and to
copy the one you want to "inittab". Under some circumstances editing the
existing inittab may fail and leave you with an empty inittab, and then
you're in deep sh*t.

Paul.
-- 
------------------------------------------------------
|debra@research.att.com   | uunet!research!debra     |
------------------------------------------------------