Path: utzoo!utgpu!cunews!bcars8!bnrgate!bwdls58!mlord
From: mlord@bwdls58.bnr.ca (Mark Lord)
Newsgroups: comp.os.msdos.apps
Subject: Re: patch and GNU diff woes
Message-ID: <3774@bwdls58.UUCP>
Date: 12 Jul 90 19:25:44 GMT
References: <27000@netnews.upenn.edu>
Sender: mlord@bwdls58.UUCP
Reply-To: mlord@bwdls58.bnr.ca (Mark Lord)
Organization: Bell-Northern Research, Ottawa, Canada
Lines: 35

In article <27000@netnews.upenn.edu> weisen@eniac.seas.upenn.edu (Neil Weisenfeld) writes:
>Sorry that I haven't had time to look at the source code and ferret these
>out myself, but here's the problem I had:

Well.. I just generated an executable for my HPUX machine, and the compiler
spotted an uninitialized local variable in one routine.. perhaps fixing it
might cure some of these other ailments people are noticing.

In file INP.C, there is a procedure called rev_in_string() at the bottom.

This procedure declares a local variable "s", which is then used without
first being given a value.  This instance of "s" should be replaced with
the parameter "string" instead on that one line.

Ie.  the existing line looks like:

	if (strnEQ(string,revision,patlen) && isspace(s[patlen]))

and should be changed to look like

	if (strnEQ(string,revision,patlen) && isspace(string[patlen]))

This is such an obvious bug, I wonder whether it was part of the REAL
distribution of 'patch', or just something introduced here by a mis-applied
patch or something.  I've sent email to the author (lwall) asking about it.

I got some other complaints about illegally mixing pointers and ints in
expressions from the mypatch.dif file.. so beware of this when applying those
patches.  Ideally, turn on ALL error detection in TURBOC and then FIX every
problem reported, even the "simple" warnings.  
-- 
 ___Mark S. Lord__________________________________________
| ..uunet!bnrgate!bmerh614!mlord | Climb Free Or Die (NH) |
| ..uunet!bnrgate!mlord%bmerh614 | Personal views only.   |
|________________________________|________________________|