In paleolimbot/anrpackageusingc: Demonstrates Calling 'C' Code

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

anrpackageusingc

The goal of anrpackageusingc is to demonstrate the structure of calling a C function in a package. For a guide on actually writing C code using the R API, see the appropriate chapter of the R Packages book, the appropriate chapter of Avanced R and the r-internals collection of unofficial R API documentation.

Installation

You can install the development version from GitHub with:

# install.packages("remotes")
remotes::install_github("paleolimbot/anrpackageusingc")

This package is designed so that you can fork, clone, and modify the C code.

Example

This package contains one function that calls a C function.

library(anrpackageusingc)
my_function("an input object")

The R function my_function() lives in R/my-function.R:

my_function <- function(input) {
  .Call(anrpackageusingc_c_my_function, input)
}

The C function anrpackageusingc_c_my_function() lives in src/my-function.c. I use the prefix anrpackageusingc_c_...() when I'm declaring a function that will get registered and called from R because I have a tool that generates src/init.c using a regex and because I read something a long time ago that suggested that exported C functions might need unique names (I don't know if this is still true). You can call the functions anything you want but it's helpful if you have a naming convention for all the functions of this type in your package.

SEXP anrpackageusingc_c_my_function(SEXP input_sexp) {
  return R_NilValue;
}

It's registered in src/init.c. The 1 at the end of the registration line is because the function accepts one argument; the NULL, NULL, 0 is to signify the end of the array (kind of like how \0 signifies the end of a null-terminated string).

SEXP anrpackageusingc_c_my_function(SEXP input_sexp);

static const R_CallMethodDef CallEntries[] = {
  {"anrpackageusingc_c_my_function", (DL_FUNC) &anrpackageusingc_c_my_function, 1},
  {NULL, NULL, 0}
};

void R_init_anrpackageusingc(DllInfo *dll) {
  R_registerRoutines(dll, NULL, CallEntries, NULL, NULL);
  R_useDynamicSymbols(dll, FALSE);
}

If you want to add a function, you'll have to copy the definition and a line to CallEntries. The function can live in any .c file of your choosing.

Finally, we need to add the useDynLib() line in the NAMESPACE. I use the automatically populated roxygen line that happens when you call usethis::use_package_doc() and usethis::use_c("some-c-filename") for this.

#' @useDynLib anrpackageusingc, .registration = TRUE

There are some variations on these components but this should get you started. From scratch, I created this package using a bunch of usethis helpers:

create_package("anrpackageusingc")
# modify Title and Description in DESCRIPTION file
use_package_doc()
use_c("my-function")
# write skeleton of anrpackageusingc_c_my_function() that returns R_NilValue
use_c("init.c")
# copy boilerplate from a previous package using C like this one
use_r("my-function")
# write wrapper using .Call()
use_test("my-function")
# write a test that calls my_function()
use_git()
use_github()

Exercises

1. Modify the body of anrpackageusingc_c_my_function() to (1) print something using Rprintf() and (2) return an R object. The easiest R objects to create are strings (Rf_mkString("some text")), integers (Rf_ScalarInteger(1234)), and doubles (Rf_ScalarReal(1234)).

2. Add a new C function called anrpackageusingc_c_my_new_function() that accepts two parameters (SEXP item1, SEXP item2) and returns the first one. Add the function to init.c and create an R wrapper called my_new_function().

3. Add a new C function in a new file that adds 2 to a numeric vector (you can use the example from the "Accessing vector data" section of the appropriate chapter of Avanced R). Make sure you can call it from R!

paleolimbot/anrpackageusingc documentation built on Dec. 22, 2021, 6:37 a.m.