Message-Id: <3.0.6.32.19990202154617.0090ada0@pop.netaddress.com>
X-Sender: pderbysh AT pop DOT netaddress DOT com
X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.6 (32)
Date: Tue, 02 Feb 1999 15:46:17 -0500
To: djgpp AT delorie DOT com
From: Paul Derbyshire <pderbysh AT usa DOT net>
Subject: Re: Alignment fault signals
In-Reply-To: <199902021824.NAA00657@envy.delorie.com>
References: <3 DOT 0 DOT 6 DOT 32 DOT 19990202120936 DOT 00906210 AT pop DOT netaddress DOT com>
 <3 DOT 0 DOT 6 DOT 32 DOT 19990202120936 DOT 00906210 AT pop DOT netaddress DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Reply-To: djgpp AT delorie DOT com

At 01:24 PM 2/2/99 -0500, you wrote:

>> On architectures where the CPU enforces strict alignment of code and
>> data, is there a fatal signal raised specific to those systems when
>> an alignment fault occurs?
>
>Yes.  SIGBUS is common, but it's not defined by POSIX because it's
>implementation dependent.

Why not have an ANSI or at least POSIX signal macro for this, which is
perhaps never raised except explicitly by raise() on some architectures,
but is there to be used on other architectures with strict alignment
requirements?

Do you know of any other common macros for an alignment fault signal?
-- 
   .*.  "Clouds are not spheres, mountains are not cones, coastlines are not
-()  <  circles, and bark is not smooth, nor does lightning travel in a
   `*'  straight line."    -------------------------------------------------
        -- B. Mandelbrot  |http://surf.to/pgd.net
_____________________ ____|________     Paul Derbyshire     pderbysh AT usa DOT net
Programmer & Humanist|ICQ: 10423848|