[ILUG] integer overflow exception

[Brian Foster]

  | Date: Tue, 05 Jun 2007 18:38:27 +0100
  | From: Pádraig Brady <P at draigBrady.com>
  | 
  | I was wondering under what circumstances
  | (compilers/CPUs) that the following program
  | generates an integer overflow exception.
  | I think none, but I'm reviewing code
  | that assumes otherwise.
  | I understand MIPS generates an exception
  | for signed arithmetic only.
  | 
  | #include <stdio.h>
  | int main(void) {
  |     unsigned int i;
  |     for(;;i++;) {
  |         if (i==0) {
  |             printf("wrap\n");
  |             fflush(stdout);
  |         }
  |     }
  | }

 ignoring the failure to initialise ‘i‛ (and the
 extra semicolon), that code must — not cannot,
 but must not — ever generate an exception IF the
 compiler in question is C99 (and, AFAICR, C89)
 conformant.

 reason?  it's “unsigned”.  and, by definition,
 unsigned integer maths is computed modulo N,
 for an appropriate N.  (e.g., 2³²-1 for the
 classic 32-bit two-complement integer.)

 OTOH, if it were signed, than the behaviour on
 overflow is not defined, and so an exception
 would be Ok.  but, like you, I cannot think of
 any system where such an exception is the norm,
 and am not certain I know of any system where
 such an exception exists?

cheers!
	-blf-
-- 
▶ ▶  I AM CURRENTLY LOOKING FOR A JOB!  ◀ ◀ | Brian Foster
Experienced (>25 yrs) software engineer:    |        Montpellier, FRANCE
 • Unix, Linux, embedded, design-for-test;  | Stop E$$o (ExxonMobile)!
 • Software/hardware co-design, debugging;  |     http:/www.stopesso.com
 • Kernels, drivers, filesystems, &tc;    Résumé (CV) & contact details:
 • IDL, automated testing, process, &tc.   http://www.blf.utvinternet.ie