R help - Feb 2022 - Using .Fortran in R - how can I use file i/o flexibly OR arrays of variable length as arguments?

I have a legacy Fortran code many thousands of lines long, a 
free-standing program that I now want to adapt with minimum re-writing 
to drive an R package, using .Fortran to pass control to a version of 
that legacy code re-written as a subroutine. BUT the output I want to 
return to R is copious, and variable in length, the length being 
computed within the Fortran code (it is a variable-dimension 
simulation). I'm looking for a solution general enough to include in a 
package that would pass CRAN checks.

I can think of two broad approaches

(1) to write this output to files in the Fortran code, to be read back 
later in R.? Of course I know about realpr, etc., but this is pretty 
ugly - apparently there's no binary i/o and no way to direct to a named 
file.

(2) to? dynamically allocate memory in the Fortran code for this output, 
and find a way for the address of this memory to be available in the 
result returned to R by .Fortran. The nearest to a possible solution 
that I can find in "Writing R extensions" involve use of C concepts, 
poorly explained, and I am not sure whether and how these could be used 
in my context.

Advice or information please!

-- 
Peter Green, p.j.green at bristol.ac.uk, peter.green at uts.edu.au
-- 
Emeritus Professor of Statistics & Professorial Research Fellow, University
of Bristol, Bristol BS8 1UG, UK
Phone: +44 117 428 4845; Fax: +44 117 928 7999
-- 
Distinguished Professor of Statistics, University of Technology, Sydney
Room: CB07.05.051; PO Box 123, Broadway NSW 2007, Australia
Phone: +61 2 9514 1742; Fax: +61 2 9514 2260

On Fri, 4 Feb 2022 10:21:11 +0000
Peter Green <mapjg at bristol.ac.uk> wrote:
> the output I want to return to R is copious, and variable in length,
> the length being computed within the Fortran code (it is a
> variable-dimension simulation)
Is there a way to separate the calculation of the output size from the
rest of the subroutine? If yes, the simplest option would be to pass
pre-allocated vectors to the .Fortran() call.

Unfortunately, .Fortran() is not expressive enough for more complicated
use cases. Memory that you intend to return to R must be allocated
using its own allocator/garbage collector system, which (like any other
call into R from user code) may longjmp() away from your functions on
interrupt, error or allocation failure, abandoning any resources the
garbage collector doesn't know about. In turn, this requires the use of
C and the PROTECT() macro. In theory, you could use Fortran 2003 "C
interoperability" to call C entry points, but I expect that to be very
inconvenient (at least due to the lack of CPP macros).

Since you mention CRAN checks, this might be a better question for
<r-package-devel at r-project.org>, not R-help.

-- 
Best regards,
Ivan