Re: [bug-gawk] Overflow to Infinity

[arnold]

Eli Zaretskii <address@hidden> wrote:

> Which Gawk results are nonsensical?  Maybe I missed some, but I looked
> at most of them, and they do make sense, more than the C results,
> actually.

nan >  2 -> true
nan >= 2 -> true

-nan >  2 -> true
-nan >= 2 -> true

nan >  inf -> true
nan >= inf -> true

-nan >  inf -> true
-nan >= inf -> true

Possibly others.

> Anyway, if there are nonsensical results, we should try fixing them so
> that they do make sense.  But mimicking C is not necessarily the way
> to do it.

Do you wish to review the code and submit a suggested patch?

The comparison operators are processed in interpret.h, and they
call cmp_scalars() which is in eval.c.

Thanks,

Arnold