Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!necntc!ci-dandelion!ulowell!cg-atla!hildreth
From: hildreth@cg-atla.UUCP (Lon Hildreth X7065)
Newsgroups: comp.sources.bugs
Subject: Re: MUSH bug fix
Message-ID: <962@cg-atla.UUCP>
Date: Tue, 20-Oct-87 13:30:29 EDT
Article-I.D.: cg-atla.962
Posted: Tue Oct 20 13:30:29 1987
Date-Received: Thu, 22-Oct-87 02:36:20 EDT
References: <131@dmnhack.UUCP>
Reply-To: hildreth@cg-atla.UUCP (Lon Hildreth X7065)
Organization: Compugraphic Corp. Wilmington, MA
Lines: 26
Summary: recursion in Addstr()

In article <131@dmnhack.UUCP> phb@dmnhack.UUCP writes:
>I got a core dump from an infinite recursion in the function "Addstr".
>A quick scan of it indicates that this might be the problem:
>
>line 151 of rite.c:
>	if (newline == '\n' && !*++s)
>
>Should probably be:
>	++s;
>	if (newline == '\n' && !*s)

No.  Don't be too anxious with this -- clearly, the intent is to not
increment 's' UNLESS 'newline == '\n'.  The real fix is in copy_msg()
where it checks the msg_rect.r_width/r_height to see if they are big
enough to handle *any* text.  Add the code in msgs.c:

	if (msg_rect.r_width < 2 * l_width(curfont) ||
	    msg_rect.r_height < 2 * l_height(curfont))
		return;

at the same place where msg_rect size is already being tested.  Sorry,
I can't supply diffs.
-- 
Lon Hildreth		...!{decvax,ima,ism780c,ulowell}!cg-atla!hildreth
Compugraphic Corp	"You can't beat fun at the old ball park."
Wilmington, MA					- Harry Caray