[Rd] Calling FORTRAN function from R issue?

[Dominick Samperi]

Hello,

I am trying to call the BLAS Level1 function zdotc from R via
a .C call like this:

#include "R.h"
#include "R_ext/BLAS.h"

void testzdotc() {
    Rcomplex zx[3], zy[3], ret_val;

    zx[0].r = 1.0; zx[0].i = 0.0;
    zx[1].r = 2.0; zx[0].i = 0.0;
    zx[2].r = 3.0; zx[0].i = 0.0;

    zy[0].r = 1.0; zy[0].i = 0.0;
    zy[1].r = 2.0; zy[0].i = 0.0;
    zy[2].r = 3.0; zy[0].i = 0.0;

    int n=3, incx=1, incy=1;
    F77_CALL(zdotc)(&ret_val, &n, zx, &incx, zy, &incy);
    Rprintf("ret_val = %f, %f\n", ret_val.r, ret_val.i);
}

This does not work. When I run '.C('testzdotc')' there is
typically a delay for a second or so, then I get: 0.0, 0.0
instead of the correct ans: 14.0, 0.0.

Section 5.2 of the R manual (on Extending R) says that only
FORTRAN subroutines can be called (not functions), probably
because of the non-standard way the compilers map FORTRAN
function names to symbols in the DLL.

This is consistent with the interface prototype for the BLAS
routine zdotc contained in <R>/include/R_ext/BLAS.h, namely,

BLAS_extern Rcomplex
    F77_NAME(zdotc)(Rcomplex * ret_val, int *n,
		    Rcomplex *zx, int *incx, Rcomplex *zy, int *incy);

But this seems to BOTH return a result, and pass the result
as the first argument?

On the other hand, this is not consistent with the standard
FORTRAN definition for zdotc that is contained in
<R>/src/extra/blas/cmplxblas.f, where the first argument is
n, not ret_val. Consequently, it is not clear where the wrapper
is defined that is called via the prototype.

My search is complicated by the fact that the libraries
libR.so, libRblas.so, libRlapack.so are stripped.

When I install the standard (FORTRAN-based) BLAS on my
system (Fedora 16), I find that zdotc is defined in the traditional
way (without adjustment to the first argument). This is probably
irrelevant because R does not use it in my configuration.

I found some documentation on Intel FORTRAN that seems to
suggest that the first argument on the C side is always the
same as (a pointer to) the FORTRAN function return value, but
this is not so if I use the standard definition of zdotc.f with
gfortran.

Any ideas?

Thanks,
Dominick