Path: utzoo!mnetor!uunet!husc6!hao!oddjob!gargoyle!ihnp4!cbosgd!mandrill!hal!ncoast!allbery
From: allbery@ncoast.UUCP (Brandon Allbery)
Newsgroups: comp.unix.wizards
Subject: Re: getlogin(3C) returns bogus name
Message-ID: <6868@ncoast.UUCP>
Date: 19 Dec 87 19:23:23 GMT
References: <10848@brl-adm.ARPA>
Reply-To: allbery@ncoast.UUCP (Brandon Allbery)
Followup-To: comp.unix.wizards
Organization: Cleveland Public Access UN*X, Cleveland, Oh
Lines: 29

As quoted from <10848@brl-adm.ARPA> by jlunny@NSWC-G.arpa:
+---------------
| i've noticed this problem within the AT&T environment on a Pyramid.
| The problem would surface in some of my background processes.  it is a
| bug in getlogin(), but it was decided it was not important enough to
| be fixed here at this site (like alot of Pyramid's little quirks, but who's
| here to complain).
+---------------

The problem on the Pyramid is probably a bad translation of a BSD /etc/utmp
(I assume you boot using BSD files rather than AT&T files; I'm told one could
use either) to AT&T format.

On the 3B1 (the original problem), the problem is caused by the fact that the
only way to create a new window is to open /dev/window, which metamorphoses
into a /dev/w[0-9] automagically.  Needless to say, this is cleaner for a
program than the way used to get e.g. ptys, but requires rooting through /dev
if the program wants the name of the *real* window device it's using.  (On
3.51 it seems to work correctly; my "w", at least, says I'm logged in on "w1",
and it *doesn't* do any mucking around with window devices.)

A quick fix:  "ln" /dev/window to "/dev/wdev01" and "/dev/wdev02", then set
up a getty on each.  However, this can backfire if someone writes to a user
from the console -- if "write" doesn't track windows the respondent could
end up writing to an otherwise-empty window....
-- 
Brandon S. Allbery		      necntc!ncoast!allbery@harvard.harvard.edu
 {hoptoad,harvard!necntc,cbosgd,sun!mandrill!hal,uunet!hnsurg3}!ncoast!allbery
			Moderator of comp.sources.misc