Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!zaphod.mps.ohio-state.edu!usc!elroy.jpl.nasa.gov!sdd.hp.com!apollo!apollo.hp.com!weber_w
From: weber_w@apollo.HP.COM (Walt Weber)
Newsgroups: comp.sys.apollo
Subject: Re: getty parity problem
Keywords: getty parity
Message-ID: <4f487956.20b6d@apollo.HP.COM>
Date: 18 Jan 91 18:33:00 GMT
References: <1991Jan17.214034.1951@quintro.uucp>
Sender: root@apollo.HP.COM
Reply-To: weber_w@apollo.hp.com
Organization: Hewlett-Packard Chelmsford Response Center
Lines: 45


In article <1991Jan17.214034.1951@quintro.uucp>, Bryan Province writes:
|> 
|> A while back I posted an article about a problem with getty.  Several people
|> said they were experiencing the same problem.  I worked some with the
|> HP/Apollo support line but I couldn't convince them that they had a problem.
|> The problem is this.
|> 
|> When the login process first starts up it uses 7 bits, even parity, and 1
|> stop bit even though I have it setup to use 8,n,1.  After you type in your
|> username at what looks like a garbled login prompt the line gets reset back
|> to 8,n,1 and the Password: prompt looks fine and so does everything else
|> after that.  After you logout the login prompt goes back to 7,e,1 and the
|> cycle starts all over again.  I have seen the problem only when using the
|> Procomm terminal emmulator on a PC.  I have noticed that other people have
|> seen the problem with other PC terminal emmulators.

Bryan -

The problem is that getty (bsd4.3 and bsd4.3-tahoe versions) use an INTERNAL
routine which ALWAYS generates 7bit even parity when accepting input. The first
input (login name) is accepted by getty using it's internal parity routine. 
Getty only sets the line to the gettytab setting just before it is going to
exec() /bin/login. Now, you could make an argument that the behavior is wrong,
and I'm going to dodge that issue. We CAN say, however, that the behavior is
the way that BSD is designed to operate, and Apollo appears to conform to that
behavior model.

I say "appears" because this isn't an official response from HP, Apollo Systems
Division of HP, the HP response center organization; it's "from Walt". (I'd
have gone on more, but my "insert_disclaimer" macro just blew up :-).

I know this doesn't do much more than define the issues more clearly. I hope
it's been helpful. I HAVE seen users config their PC lines for 7 bit/ignore
parity during the login sequence, and then switch over to 8/n/1 for the login
session; it generally involves some kind of scripting language on the PC end.

|> Bryan Province -Glenayre Corp., Quincy, IL- quintro!bep@lll-winken.llnl.gov

...walt...

Walt Weber                           Hewlett Packard Response Center
508-256-6600x8315                    Chelmsford, MA, USA
   "The power of accurate observation is commonly called cynicism
    by those who have not got it" -George Bernard Shaw