Craig Dedo writes:
> Indeed, having this "feature" on by default could be considered
> non-standard-conforming, since it changes the interpretation of
> what is otherwise standard-conforming code. If it is on by
> default, you should complain LOUDLY to the vendor.
I agree that it is an annoying default and merits loud complaints.
But it is a subject of substantial debate whether or not it makes a
compiler non-standard-conforming. I suppose the qualifier "could"
makes it accurate in that some people could (and indeed, some
apparently do) consider it to be non-standard-conforming. But other
people consider the first group to be wrong.
I'm not in the mood to re-cycle the same old debate, as it doesn't
seem very constructive. I'll just point out that there is not
unanimity for either side.
The question arises because "\" is not in the Fortran 95 character
set. People disagree about whether or not it has to be characterized
as another character that is "representable by the processor".
Someone will probably post an explanation of why they think that
one side or the other of this debate is the only answer a rational
person could come up with. If so, I'll just pre-announce that I
disagree with whatever anyone posts along such lines. That way,
I won't feel obligated to post another followup saying that. I
take the position that "\" have to be so characterized, but
I acknowledge rational arguments on both sides.
In the f2k draft, it is part of the Fortran character set, making
the question moot.
--
Richard Maine | Good judgment comes from experience;
[log in to unmask] | experience comes from bad judgment.
| -- Mark Twain
|