Hi Eric,

On Sat, Apr 30, 2011 at 07:18:49AM -0600, Eric Blake wrote:
> On 04/30/2011 01:16 AM, Gary V. Vaughan wrote:
> >>     argc is an int, therefore by definition, it is <= INT_MAX.
> >>     
> >> -  assert (0 < argc && argc <= INT_MAX);
> >> +  assert (0 < argc);
> >>    for (i = 1; i < (unsigned) argc; i++)
> >>      {
> >>        const char *arg = ARG((int) i);
> > 
> > Since argc is immediately cast to an unsigned int, I think the test was
> > supposed to be:
> > 
> >   assert (0 < argc && argc <= UINT_MAX);
> 
> Except that INT_MAX  < UINT_MAX, so that's also redundant.  The real
> trick was that gcc was issuing spurious warnings if it didn't have proof
> that argc was positive, so all we really needed was the (0 < argc) term;
> my mistake in the earlier patch was adding a redundant term at the same
> time.

Quite right.  I guess lack of caffeine was affecting my sense of logic.
Sorry for the noise.

Cheers,
-- 
Gary V. Vaughan (gary AT gnu DOT org)

Attachment: pgp369raaqAeJ.pgp
Description: PGP signature

_______________________________________________
M4-patches mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/m4-patches

Reply via email to