[Rd] Accessing C++ code from R

[Simon Urbanek]

On Apr 4, 2007, at 2:46 PM, Jos Elkink wrote:

> Thank you - that indeed fixes the first problem (about cout). It  
> still leaves the second error, though:
>
> main.cpp:52: error: brace-enclosed initializer used to initialize  
> 'R_NativePrimitiveArgType*'
>
> Any ideas about that?
>

Sure, do what the compiler tells you to - use a reference instead:

static R_NativePrimitiveArgType runSimulation_t[] =
   {INTSXP, INTSXP, INTSXP, REALSXP, REALSXP, REALSXP, INTSXP,  
INTSXP, INTSXP,
    INTSXP, INTSXP, INTSXP, REALSXP, REALSXP, INTSXP, INTSXP, INTSXP,
    INTSXP, INTSXP, INTSXP, INTSXP, INTSXP, INTSXP, INTSXP, REALSXP,
    REALSXP, REALSXP};

R_CMethodDef cMethods[] = {
   {"runSimulation", (DL_FUNC) &runSimulation, 27, runSimulation_t},
   {NULL, NULL, 0}
};

BTW have a look at registration.c in R sources for an elegant way to  
make this less error-prone when registering multiple functions (and  
looking at your argument list ... you should seriously consider  
using .Call :P)

Cheers,
Simon



>
> On 4/4/07, Simon Urbanek <simon.urbanek at r-project.org> wrote:
>> Jos,
>>
>> you omitted the crucial part but my guess would be that all you have
>> to do is to put #include <iostream> *before* R headers includes (R
>> defines "length" which throws off iostream).
>> Also note that using cout is not a good idea, because it is not
>> captured by R, you should be using Rprintf instead.
>>
>> Cheers,
>> Simon
>>
>> On Apr 4, 2007, at 1:44 PM, Jos Elkink wrote:
>>
>>> Hi,
>>>
>>> I am trying to use existing C++ code from R. I have no problems
>>> compiling C code and using it in R, but with C++ I'm running into
>>> problems.
>>>
>>> Here's the compiler output:
>>>
>>> Macintosh-10:~/Desktop/dissertation/Model - CPP version/R labguest$
>>> g++ -I/Library/Frameworks/R.framework/Resources/include
>>> -I/Library/Frameworks/R.framework/Resources/include/i386 *.cpp
>>> In file included from /usr/include/c++/4.0.0/bits/locale_facets.h:
>>> 1533,
>>>                  from /usr/include/c++/4.0.0/bits/basic_ios.h:44,
>>>                  from /usr/include/c++/4.0.0/ios:50,
>>>                  from /usr/include/c++/4.0.0/ostream:44,
>>>                  from /usr/include/c++/4.0.0/iostream:44,
>>>                  from main.cpp:5:
>>> /usr/include/c++/4.0.0/bits/codecvt.h:217:45: error: macro "length"
>>> passed 4 arguments, but takes just 1
>>> /usr/include/c++/4.0.0/bits/codecvt.h:216: error: expected ';'
>>> before 'const'
>>> /usr/include/c++/4.0.0/bits/codecvt.h:220: error: expected `;'
>>> before 'int'
>>> main.cpp:51: error: brace-enclosed initializer used to initialize
>>> 'R_NativePrimitiveArgType*'
>>>
>>>
>>> Line 51 refers to the standard part to register the function in R:
>>>
>>> R_CMethodDef cMethods[] = {
>>>   {"runSimulation", (void* (*) ()) &runSimulation, 27, {INTSXP,
>>> INTSXP, INTSXP, REALSXP, REALSXP, REALSXP, INTSXP, INTSXP, INTSXP,
>>> INTSXP, INTSXP, INTSXP, REALSXP, REALSXP, INTSXP, INTSXP, INTSXP,
>>> INTSXP, INTSXP, INTSXP, INTSXP, INTSXP, INTSXP, INTSXP, REALSXP,
>>> REALSXP, REALSXP}},
>>>   {NULL, NULL, 0}
>>> };
>>>
>>>
>>> I had to add the explicit cast of &runSimulation to get rid of one
>>> compiler error, but do not know how to deal with this one.
>>> Encapsulating this part of the code in
>>>
>>> extern "C" {
>>> }
>>>
>>> has no effect.
>>>
>>> runSimulation is just a regular function, with a lot of int* and
>>> double* parameters, and it is not part of a class. It does inside
>>> create a class, though.
>>>
>>> Outline:
>>>
>>> void runSimulation(int *p_1 ... etc) {
>>>
>>>    try {
>>>      CModel Model();
>>>      Model.Run();
>>>    } catch (CException& e) {
>>>      cout << "exception" << endl;
>>>    }
>>> }
>>>
>>> Encapsulating this function in extern "C" also has no effect.
>>>
>>> Anyone any ideas what I should do? The system I am using is an
>>> Intel-based Mac.
>>>
>>> Thanks in advance!
>>>
>>> Jos
>>>
>>> ______________________________________________
>>> R-devel at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>>
>>>
>>
>>
>
>
> -- 
> Jos Elkink
> PhD Candidate, Trinity College Dublin
> HMDC Fellow, Harvard University
> http://jaeweb.cantr.net
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
>