Phillip Helbig <[log in to unmask]> wrote:
...
>Another fine point, possibly relevant. At least in the past, if ug_temp
>is floating-point, then one can't use an expression involving it in a
>declaration. IIRC. (I think floating-point operations were disallowed
>to make it easier for cross-compilers or something.) I believe at least
>some aspects of this restriction were lifted recently.
Fortran 77 had this kind of constraint, but it didn't have the SIZE
intrinsic. Fortran 90 (and later) doesn't allow most floating-point
intrinsics functions in a specification expression (or an initialization
expression, for that matter). But a simple addition is permitted (in fact,
all intrinsic operators are). So, in any program which uses the SIZE
intrinsic, the floating point operation should also be available - both being
F90 features.
--
J. Giles
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|