Re: [Fwd: Re: Symbols in function return specifiers]

[Steve Kargl <sgk at troutmask dot apl dot washington dot edu>]

On Sun, Nov 02, 2008 at 07:18:43PM +0100, Tobias Burnus wrote:
> 
> Daniel Kraft wrote:
> > I just got this reply to my clf posting about the "used before typed"
> > thing (PR 32095, IIRC).
> > Robin believes my first example is valid, if I understand him
> > correctly, but this is currently rejected by gfortran (because I
> > implemented it the way I interpreted it, that IMPLICIT types x and
> > that any explicit declaration afterwards can only enforce the same
> > typing).  I believe I did implement the behaviour exactly as Richard
> > Maine answered, which Robin claims is "erroneous".
> >
> > What do you think, is Robin right?
> 
> In general, I have more trust in Richard Maine than in Robin; Richard
> was editor of the Fortran 2003 standard and was involved in revising the
> Fortran standard, which included answering (formal) interpretation
> requests. (I don't know the background of Robin.)
> 
> Richard cites from the standard and while one could argue that certain
> parts are wrongly interpreted or that a relevant part of the standard is
> missed, Robin does not address this but simply claims "x is an integer"
> without proving this with the standard.
> 
> I think Richard's reading of the standard is correct and Robin lacks the
> arguments to convince me otherwise (though this does not completely rule
> out that he is right).
> 

For what its worth, I totally agree with Tobias.  I'd
trust Richard's opinion over Robin's.  With respect to
clf replies I would follow advice from RIchard, Steve
Lionel (Intel), Robert Corbett (Sun), Mike Metcalf,
Dick Hendrickson, Walt Spector, James van Buskirk,
and few others (names escape at the moment).  Robin
isn't among the few others.

-- 
Steve