Re: optimization/4371

Re: optimization/4371

[Martin Sebor]

The following reply was made to PR optimization/4371; it has been noted by GNATS.

From: Martin Sebor <sebor@roguewave.com>
To: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, jsm28@cam.ac.uk,
 	sebor@roguewave.com
Cc:  
Subject: Re: optimization/4371
Date: Thu, 20 Sep 2001 16:51:32 -0600

 I'm sure you're right about the undefined behavior, but most hardware
 simply wraps the result around (i86 certainly does, as does SPARC or PA).
 That's usually also the expected behavior. None of the compilers that I
 work with (HP aCC, VisualAge, SunPro, MIPSpro, to name just a few) gets
 stuck in an infinite loop like gcc does. So even though the compiler may
 be allowed to do so I think it could and should behave more reasonably.
 A quality implementation could even warn about the portability of the
 code but an infinite loop is most certainly not desirable.
 
 Regards
 Martin
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=4371&database=gcc