X-Authentication-Warning: delorie.com: mailnull set sender to djgpp-workers-bounces using -f
From: sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id: <10203031643.AA28229@clio.rice.edu>
Subject: Re: libm signed/unsigned warnings
To: djgpp-workers AT delorie DOT com
Date: Sun, 3 Mar 2002 10:43:23 -0600 (CST)
In-Reply-To: <200203031419.g23EJBH06559@mother.ludd.luth.se> from "Martin Str|mberg" at Mar 03, 2002 03:19:10 PM
X-Mailer: ELM [version 2.5 PL2]
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> How about this:
> Index: djgpp//src/libm/math/e_pow.c

> -       __uint32_t lx,ly;
> +       __uint32_t lx,ly,ju;

>                 if(k>20) {
> -                   j = ly>>(52-k);
> -                   if((j<<(52-k))==ly) yisint = 2-(j&1);
> +                   ju = ly>>(52-k);
> +                   if((ju<<(52-k))==ly) yisint = 2-(ju&1);
>                 } else if(ly==0) {

Once again, I like this much better.  It uses the right type
declaration, and may even fix subtile bugs.  An extra temp 
declaration of the right type may even make the compiler's job
easier.  

That is, assuming that someone didn't purposly use a signed in
this code someplace to get that different behavior ...