[R] Using R as Shared Library

[Thorsten Jolitz]

Dirk Eddelbuettel <edd at debian.org> writes:

> On 12 August 2012 at 07:56, Michael Weylandt wrote:
> | On Aug 12, 2012, at 6:09 AM, Thorsten Jolitz
> | <tjolitz at googlemail.com> wrote:
> | > Thorsten Jolitz <tjolitz at googlemail.com> writes:
> | > Let me reformulate my question (since I managed to make 'native' calls
> | > to C functions in libR by now): 
> | > I still wonder whats 'in there' in libR - only the core C functions, or
> | > all the functions written in R itself too? And what packages are
> | > included?
> | 
> | Just C functions.  
>
> Moreover, many of these are marked 'non visible' and cannot be accessed.  You
> probably want to consider
>
>   a) the standalone R math library, available eg on Debian/Ubuntu as package
>      r-mathlib which gets you a subset of R (distribution functions, random
>      numbers, ...) for use in other programs, or
>
>   b) the embedding API of R
>
> Both of these are documented in the 'Writing R Extension' manual that came
> with R. 
>
> If you are interested in b), you may also want to look at the RInside project
> which makes embedding a lot easier.  The simplest example is just 
>
>     #include <RInside.h>                    // for the embedded R via RInside
>     int main(int argc, char *argv[]) {
>         RInside R(argc, argv);              // create an embedded R instance 
>         R["txt"] = "Hello, world!\n";	    // assign a char* (string) to 'txt'
>         R.parseEvalQ("cat(txt)");           // eval the init string, ignoring any returns
>         exit(0);
>     }
>
> which passes a string to the embedded R interpreter and calls an R function
> to display it. You can send full R objects back and forth thanks to Rcpp, and
> there are over a dozen examples included in the packages, as well as more
> advanced use of embedded R within the context of MPI (for parallel
> computing), Qt (for GUIs and much more) or Wt (for web applications).


Thanks, that looks very interesting indeed. 

-- 
cheers,
Thorsten