davis@MASIG2.OCEAN.FSU.EDU (Alan Davis) (12/11/90)
I am coding up some routines in C that I need to make accessable to the rest of
our researchers that only program in the F------ language. There are a couple
of things that I would like to do and TFMs say I should be able to do, but
unfortunately do not seem to work.
First, in Sec. 3.2.1 of the FORTRAN 77 Programmer's guide it states in a
Note that mkf2c has an option for letting function names contain an
underscore, but not such option appears to exist and mkf2c complains about it
when you try to insert one.
Second, I am passing some floating point arguments that I want to prevent
from being promoted to double. mkf2c has an -f flag to do just that, but TFM
says that you should use function prototypes in the C routine. When I do this
mkf2c also complains about a syntax error. If I write the C routine in the
old style and use mkf2c with the -f flag the type conversion does appear to
be prohibited, but this causes the C routine to be written without a prototype
which I would like to avoid. My way out was to code a small fortran wrapper
routine without the prototype which in turn calls the real C routine which I
wrote with prototyping resulting in two function calls from Fortran.
Does anybody have any experience with either of these two bugs. I am running
a 320 VGX with 3.3.1.
ps. Does anyone know the release date of 3.3.2, as I need it to fix a problem
in the gl library on the VGX?
--
Alan Davis |
MS B-174, Love 008A |
Mesoscale Air-Sea Interaction Group | INTERNET davis@masig1.ocean.fsu.edu
Florida State University | SPAN scri1::davis
Tallahassee, FL 32306-3041 |
(904) 644-3798 |
_______________________________________________________________________________blbates@AERO4.LARC.NASA.GOV ("Brent L. Bates AAD/TAB MS361 x42854") (12/13/90)
I was told that 3.3.2 should be shipping "any day now".
Why do you need to use mkf2c? I always though that created more pain
than was necessary. Other machines I have used don't need anything like
mkf2c. If you make the C routines correctly, like using pointers for the
varables passed to the C routines and other similar precausions, it
shouldn't be that difficult. Now I haven't tried any FORTRAN to C on our
4D's, yet, and when I tried it on our 3130, long ago, it never worked to
my statisfaction, but then again there weren't a lot of things that worked
reasonably on the 3130's any way.
--
Brent L. Bates
NASA-Langley Research Center
M.S. 361
Hampton, Virginia 23665-5225
(804) 864-2854
E-mail: blbates@aero4.larc.nasa.gov or blbates@aero8.larc.nasa.gov