Re: c/10176: Broken nested Do loop: simple example. gcc 3.2 redhat (fwd)

Re: c/10176: Broken nested Do loop: simple example. gcc 3.2 redhat (fwd)

[Wolfgang Bangerth]

The following reply was made to PR c/10176; it has been noted by GNATS.

From: Wolfgang Bangerth <bangerth@ticam.utexas.edu>
To: gcc-gnats@gcc.gnu.org
Cc:  
Subject: Re: c/10176: Broken nested Do loop: simple example. gcc 3.2 redhat
 (fwd)
Date: Fri, 21 Mar 2003 09:03:47 -0600 (CST)

 ---------- Forwarded message ----------
 Date: 20 Mar 2003 18:49:53 -0800
 From: Ernst Berg <ernstberg@bigvalley.net>
 To: bangerth@dealii.org
 Subject: Re: c/10176: Broken nested Do loop: simple example. gcc 3.2 redhat
 
 It would need to look for a power of 3.. I was looking at the parity
 language of it.
  I just ran it again to get an example and it worked.  I don't get it. 
 
 I am sorry.. I'm not sure why it wouldn't work before even after I
 powered off and rebooted.. 
 
 False alarm.
 
  This is a related function to the Collatz problem 3x+1   if( odd )
                                                    -----
                                                      2    if ( even )
 
 
 
 On Thu, 2003-03-20 at 17:51, bangerth@dealii.org wrote:
 > Synopsis: Broken nested Do loop: simple example. gcc 3.2 redhat
 > 
 > State-Changed-From-To: open->feedback
 > State-Changed-By: bangerth
 > State-Changed-When: Fri Mar 21 01:51:28 2003
 > State-Changed-Why:
 >     Your program 4x.c is essentially equivalent to this one:
 >     --------------------------
 >     int main(void)
 >     {
 >       unsigned int count = 1;
 >       do {
 >         if( count % 3 ) { count *= 4; count += 2; }
 >         else            { count /= 3; }
 >         
 >         if( count == 1 ) break;
 >         if( count == 3 ) break;
 >       } while(1);
 >       return(0);
 >     }
 >     ----------------------
 >     What happens is that you never jump out of the inner loop.
 >     Unfortunately, I don't see right away why the loop should
 >     terminate at all. gcc seems to have the same problem, as
 >     does icc. Can you give a simple argument why it should?
 >     
 >     Thanks
 >       Wolfgang
 > 
 > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=10176
 
 
 
 ----------------------------------------------------------------------------------------------
 50% faster, 400% cheaper, one phone line for voice and data, v.92, national coverage, all for less  than $8.50 a month http://www.BigValley.net