🚀 go-pugleaf

In <1ep9qjINNsni@agate.berkeley.edu> actize@garnet.berkeley.edu () writes:

>I program in C, and most of our mathematical libraries are in Fortran 77.
>Our online documentation describes a few things, such as 
>1.	Fortran calls by address
>2.	The order of array storage is different.

>Does anybody have experience with this, and advice ?

>David
>actize@garnet.berkeley.edu

If you're programming on a Sun, the SunOS fortran manual has a whole chapter
on this. IBM's RS-6000 has a similar document, though it isn't as detailed.

A few more differences (assuming you're on a Unix system):

All fortran routines are converted to lowercase and have an underscore appended
to the end of the function name.

Strings are passed in an odd manner--there is an extra argument at the *end*
of the argument list for each string passed. This is a long (*not* a long *)
and it holds the length of the string in bytes.

Subroutines return int *'s. Functions return the corresponding type in C (passed
by reference, of course).

You must link in libF77.a (or libV77.a if you need VMS compatibility).

int * <--> INTEGER or INTEGER*4
long * <--> INTEGER*8
char * <--> CHARACTER or INTEGER*1 or CHARACTER*n or LOGICAL
double * <--> DOUBLE PRECISION or REAL*8
float * <--> REAL or REAL*4

Complex numbers correspond to a structure of two of the appropriate floating
point type.

If you're programming a VMS system, all this is BS.

Good luck.

Mike Gross
Physics Board
Univ of California                 GO SLUGS!!!!!
Santa Cruz, CA 95064