[R] Calling C code from R...wrapping C structures

[Duncan Murdoch]

On 5/13/2008 7:51 AM, Nathan Harmston wrote:
> Hi everyone,
> 
> I am currently trying to call some C code from R, specifically calling a
> function which populates a C struct.
> 
> typedef struct{
>    // contents
> } Model;
> 
> void test(Model *m){
>    // fill the struct with crap
> }
> 
> I compile the C code into a shared library, which loads into R properly. My
> simple test functions work (i.e adding numbers etc)
> setModel <- function(){
>    model<-vector("list", 6)
>    name(model) <- c( SET THE NAMES OF THE MODEL HERE )
>    model
> }
> t <- setModel()
> testcode <- function(setModel){
> 
> 
> dyn.load("Simulation.so")
> 
> foo <- .C("test", Model=setModel)
> foo
> }
> 
> testcode(t)
> 
> However I get segfaults whenever I try to access any of the variables
> contained in Model. So my question is: am I doing something wrong here? Is
> list the right data type to wrap a structure in? I cant see any
> documentation about wrapping structs in R.
> 
> Also when I pass strings to C from R they dont seem to be passed at all? Is
> this a problem with R or a problem with me and if so how do I fix it?

It is a problem in your C code.  The .C interface doesn't know about the 
C declaration of the Model structure; it can only pass things it knows 
about.  So your C function test() should have its header set to accept 
what gets passed to it.  In your case, that's a list, so your C function 
receives an array of pointers to SEXP structures; not the easiest things 
to work with.  Most people using .C() pass in simple vectors (numeric, 
integer, etc.), and use the .Call() interface if they really need R 
internals.

Duncan Murdoch


> 
> Many Thanks in advance,
> 
> Nathan
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.