Re: -pedantic [was Re: Unused dummy argument]

["N.M. Maclaren" <nmm1 at cam dot ac dot uk>]

On Dec 7 2012, Steve Kargl wrote:
> > >
> > >You may not (and I suspect do not want) to use -pedantic
> > >because it changes the ranges of the various integer kinds.
> >
> > Eh?  That's not what the documentation says.  ... I would appreciate
> > clarification.
>
> The comment does not accurately reflect the old discussion.  One
> can go read the mailing list archive and the PR audit trails to
> see the full issue.  From arith.c,
>
>      /* See PRs 13490 and 17912, related to integer ranges.
> 	 The pedantic_min_int exists for range checking when a program
> 	 is compiled with -pedantic, and reflects the belief that
> 	 Standard Fortran requires integers to be symmetrical, i.e.
> 	 every negative integer must have a representable positive
> 	 absolute value, and vice versa.  */

Well, yes, of course.  That has always been the case.

> Is this line of code portable?
>
>   i = iabs(i)

Not if the argument goes outside the symmetric range, and never has
been.  It's not portable in C, either, despite common belief.  But
using constants and constant expressions where that occurs is the
cause of most of the problems, and is precisely why any author of
portable code DOES want a warning!  The compiler does not warn on
that code, in the absence of other issues.

I have used a system where using that on the most negative integer
aborted the program.


Regards,
Nick Maclaren.