Thanks, Malcolm.  I can write it more in the style of your suggestion,
    double precision :: max_allowed
    parameter ( max_allowed = sqrt(sqrt(huge(max_allowed))) )
and hope it is less confusing.

In the meantime, your other suggestion for working around the gfortran
problem,
    Doubtless you could work around it by putting a PRIVATE SQRT statement
    in mod1 (since you probably did not intend to export the intrinsic
    SQRT from there anyway)
seems to solve my problem.

Thanks again.

Dave Smith


________________________________________
From: Fortran 90 List [[log in to unmask]] on behalf of Malcolm Cohen [[log in to unmask]]
Sent: Sunday, November 09, 2014 4:38 PM
To: [log in to unmask]
Subject: Re: ambiguous reference

>There was an interp that resulted in
>
>  integer, parameter :: N = bit_size(n)
>
>being illegal.  How is this one different?

That's a confusing question, since I indicated that I did think it was
(probably) outside the standard...

The interp was Fortran 95, so pretty ancient history, and mostly asking
about things like

INTEGER :: P(complicated_expression_for_lower_bound_1:   &
               & complicated_expression_for_upper_bound_1,   &
               & complicated_expression_for_lower_bound_2:   &
               & complicated_expression_for_upper_bound_2) = &
               & RESHAPE( (/ 11, 21, 12, 22 /), SHAPE(P) )

and the answer was that that was not allowed (plus a whole load of others).

The relevant text in the standard is

  "If a constant expression includes a specification inquiry that depends on
a type parameter or an array bound of an entity specified in the same
specification-part , the type parameter or array bound shall be specified in
a prior specification of the specification-part . The prior specification
may be to the left of the specification inquiry in the same statement, but
shall not be within the same entity-decl ."

Now, in the case of "double precision :: max_allowed...", the question is
whether the specification of the type and kind are considered to occur at
the type-spec, or whether they occur at the (first) occurrence of
"max_allowed".  If the former, there is nothing wrong.   If the latter, it
is beyond the standard.  From the text of the interp answer, one would think
the latter, but the text in the standard itself is not 100% clear.

(And I think that few if any compilers will reject or even warn about this
particular case.)

I rest my case with "at best, confusingly bad style".

Cheers,
--
.............................Malcolm Cohen, NAG Oxford/Tokyo.