Re: 1.4.9: wrong results in signed division

[Eric Blake]

According to Cesar Strauss on 4/21/2007 5:30 PM:
> Hi,
> 
> The eval macro in m4-1.4.9 gives wrong results in signed division:
> 
> $ ./m4
> eval(-100/10)
> 429496719
> (expected: -10)

Indeed, this is a regression.  Thanks for the report.

> This comment near the code seems relevant:
> 
>      /* Minimize undefined C behavior on overflow.  This code assumes
>      that the implementation-defined overflow when casting
>      unsigned to signed is a silent twos-complement
>      wrap-around.  */
> 
> I don't think any overflow handling is needed here, because the
> quotient is always smaller than the dividend, in signed integer
> division.

No, consider INT_MIN/-1.  With 2's complement, it is undefined according
to C (there is no way to represent -INT_MIN in an int).  It turns out that
many 2's complement architectures do wraparound and come up with the
answer INT_MIN, while Intel causes SIGFPE.  So special handling IS needed,
just not what I wrote, nor your reversion.  I'll come up with something
shortly.

-- 
Don't work too hard, make some time for fun as well!

Eric Blake             address@hidden

signature.asc
Description: OpenPGP digital signature