Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!think.com!spool.mu.edu!uunet!hotmomma!sdb
From: sdb%hotmomma@uunet.uu.net (Scott Ballantyne)
Newsgroups: comp.sys.3b1
Subject: Re: Failure of iswind()
Message-ID: <1991May26.153643.11279@hotmomma.UUCP>
Date: 26 May 91 15:36:43 GMT
Article-I.D.: hotmomma.1991May26.153643.11279
References: <1991May24.203238.7990@sci.ccny.cuny.edu>
	<1991May25.040152.8151@ceilidh.beartrack.com>
Sender: sdb%hotmomma@uunet.uu.net (Scott Ballantyne)
Organization: ScotSoft Research
Lines: 20
In-Reply-To: dnichols@ceilidh.beartrack.com's message of 25 May 91 04: 01:52 GMT

>>>>> On 25 May 91 04:01:52 GMT, dnichols@ceilidh.beartrack.com (DoN Nichols) said:

DoN> 	Yes, it happens when logged in to a tty port.  I presume that
DoN> it could happen via the OBM as well.  Iswind() must be quite
DoN> brain-damaged.  I added code to my first copy of klondike to
DoN> check if /dev/tty = w? to avoid this, but that is a kluge.

>Isn't this some bizarre sort of security hole?!?

DoN> 	At least a denial-of-service one.  Maybe we need to rewrite iswind().


I find the WIOCGETD ioctl works well for determining whether or not
you are actually on a bitmap.  iswind () has never worked for me :-(

sdb

-- 
{sdb%hotmomma@uunet.uu.net | hotmomma!sdb@uunet.uu.net | uunet!hotmomma!sdb}
"A hacker is a machine for turning caffeine into programs"