Signedness bug in eval

[Peter Clifton]

Please CC me to any response, as I'm not on any M4 mailinglist.

PCB (Layout software) uses M4 to generate its graphical footprints. With
M4 1.4.9 this has broken...

The change in behaviour is maths with eval(..)

For example:

eval(-50/2)
2147483623
eval(-2/2)
2147483647

Old versions of M4 produce the expected results.

Looking at the changelog, could this be caused by:


2007-03-15  Eric Blake  <address@hidden>

        Avoid Tandem/NSK's broken long long (without a symmetric unsigned
        long long, it confuses gnulib).
        * configure.ac (AC_TYPE_LONG_LONG_INT): Declare long long broken
        if unsigned long long doesn't work.
        * THANKS: Update.
        Reported by Matthew Woehlke.


Thanks for looking into this - hopefully if a fix is released, the
change will get picked up by distros quickly - as this breaks our (and
probably other) applications,

Best regards,

-- 
Peter Clifton

Electrical Engineering Division,
Engineering Department,
University of Cambridge,
9, JJ Thomson Avenue,
Cambridge
CB3 0FA

Tel: +44 (0)7729 980173 - (No signal in the lab!)