From: armelasselin@hotmail.com   
      
   "Chris F Clark"  a icrit   
   > So, I wonder, how many of you have experienced real compiler bugs,   
   > that you could verify was the compilers fault and not just "mysterious   
   > behavior" that could be fixed by changing the optimizer/debugger   
   > level?   
      
   Hum, it happened once to me for a C/C++ compiler for MIPS (used to be   
   delivered by M... for their PocketPC), was so clever that when   
   executing:   
      
   myfunc (astruc &a)   
      
   for (idx=0; idx<10; ++idx)   
   {   
       mystruc.x = idx;   
       myfunc (mystruc)   
   }   
      
   the compiler found intelligent to keep mystruc.x in a register and never   
   saved it into mystruc.   
   so an aliasing bug, but one incredibly easy to avoid by a good compiler, and   
   rather hard to workaround...   
      
   Armel   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)