Re: c/4359

Re: c/4359

[Reichelt]

The following reply was made to PR c/4359; it has been noted by GNATS.

From: Reichelt <reichelt@igpm.rwth-aachen.de>
To: otto@ii.uib.no, gcc-gnats@gcc.gnu.org, gcc-bugs@gcc.gnu.org
Cc:  
Subject: Re: c/4359
Date: Wed, 19 Sep 2001 16:47:04 +0200

 Hi,
 
 I was able to reproduce the problem on an IRIR 6.5 computer with
 gcc 3.0.2 20010903 (prerelease). It might be target-dependant,
 since everything works fine with gcc 3.0 on i686-pc-linux-gnu.
 
 Here is a smaller testcase which triggers the bug with -O3 *and* -O2:
 
 void e();
 
 void f()
 {
     int i, j;
 
     switch (j) {
     case 2:
     case 3:
     case 5:
     case 6:
       break;
     case 4:
       i = i && j;
       break;
     default:
       e();
     }
 }
 
 (Just compile with "gcc -O2 -c" or "gcc -O3 -c".)
 
 By the way, the bug seems to be a regression from 2.95.3.
 
 Greetings,
 Volker Reichelt
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=4359&database=gcc
 
 

Re: c/4359

[Otto Skrove Bagge]

The following reply was made to PR c/4359; it has been noted by GNATS.

From: Otto Skrove Bagge <otto@ii.uib.no>
To: Reichelt <reichelt@igpm.rwth-aachen.de>
Cc: gcc-gnats@gcc.gnu.org, gcc-bugs@gcc.gnu.org
Subject: Re: c/4359 
Date: Wed, 19 Sep 2001 17:52:00 +0200

 >
 >Hi,
 >
 >I was able to reproduce the problem on an IRIR 6.5 computer with
 >gcc 3.0.2 20010903 (prerelease). It might be target-dependant,
 >since everything works fine with gcc 3.0 on i686-pc-linux-gnu.
 
 That sounds reasonable; I haven't had any trouble with this on GNU/Linux.
 I've also had unconfirmed reports of successful builds on Solaris and
 powerpc-apple-darwin1.0.
 
 I also suspect that it might be outputting wrong code for -O2 -- compiling with
 gcc 3.0 (IRIX) introduces a bug in the program which isn't there when compiled
 with gcc 2.95.2 (IRIX) or with gcc version 3.0.1 20010811 (Debian prerelease)
 for i386-linux.  This problem does not occur when compiling without 
 optimisation.
 
 >Here is a smaller testcase which triggers the bug with -O3 *and* -O2:
 
 Yes, that seems to do it...
 
 -otto