Paul Eggert <[EMAIL PROTECTED]> wrote: > Jim Meyering <[EMAIL PROTECTED]> writes: >> this new version of the macro always fails for signed >> accumulators. Why? > > I'll be a little long-winded here, not for your benefit (you know this > stuff!) but for others who are reading along.
Thanks for the explanation and fix. > 2005-07-05 Paul Eggert <[EMAIL PROTECTED]> > > * src/system.h (DECIMAL_DIGIT_ACCUMULATE): Generate a hard error > (not just a warning) if GCC is used and the types don't match. > > --- system.h.~1.130.~ 2005-07-04 23:14:06.000000000 -0700 > +++ system.h 2005-07-05 15:12:42.000000000 -0700 > @@ -811,13 +811,18 @@ ptr_align (void const *ptr, size_t align > then don't update Accum and return false to indicate it would > overflow. Otherwise, set Accum to that new value and return true. > Verify at compile-time that Type is Accum's type, and that Type is > - unsigned. Accum must be an object, so that we can take its address. > - Accum and Digit_val may be evaluated multiple times. */ > + unsigned. Accum must be an object, so that we can take its > + address. Accum and Digit_val may =be evaluated multiple times. I've removed that stray `='. _______________________________________________ Bug-coreutils mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-coreutils
