Checking Arguments for Function and Subroutine Calls

[Tom Browder]

In my struggle to port a large code base to gfortran, I am finding
some serious bugs that are not warned about (or at least I have not
found one that will do it).  The ones of most concern are situations
where the wrong argument types are used by the caller or the wrong
number of arguments are used by the caller.

Now the old code is f77, and I know gfortran is f90, but, without an
explicit interface, will gfortran ever be able to check such
situations?

Note that the "struggle" was due to non-standard code, Portland
extensions, and chunks (such as list-directed output) that the
standard leaves up to the implementation.  Gfortran has been great,
and it compares well with Sun's compiler so far.

Thanks.

-Tom

Tom Browder
Niceville, Florida
USA