R devel - Jan 2007 - is it necessary to always register C routines with R_registerRoutines?

I am writing bindings to the neural network simulator SNNS.
At present I have used only .C interface, now I'm studying .Call interface.

I have adapted the example from page 77 of r-exts.pdf, however, it crashes
R.
I use MingW as recommended by Duncan Murdoch.
Please, tell me what I am missing.
The code is below.

Thank you.

SEXP snns_getVersion(void)
{SEXP version;
 char *v;
  
  PROTECT(version=NEW_CHARACTER(15));
  v=CHARACTER_POINTER(version);
  strcpy(v,krui_getVersion());
  UNPROTECT(1);
  return version;
}
> sessionInfo()
R version 2.4.1 (2006-12-18) 
i386-pc-mingw32 

locale:
LC_COLLATE=Russian_Russia.1251;LC_CTYPE=Russian_Russia.1251;LC_MONETARY=Russian_Russia.1251;LC_NUMERIC=C;LC_TIME=Russian_Russia.1251

attached base packages:
[1] "stats"     "graphics"  "grDevices"
"utils"     "datasets"  "methods"
[7] "base"     
 

-- 
View this message in context:
http://www.nabble.com/is-it-necessary-to-always-register-C-routines-with-R_registerRoutines--tf3063486.html#a8519680
Sent from the R devel mailing list archive at Nabble.com.

sorry, forgot to mention.

krui_getVersion returns char *

SEXP snns_getVersion(void)
{SEXP version;
 char *v;
  
  PROTECT(version=NEW_CHARACTER(15));
  v=CHARACTER_POINTER(version);
  strcpy(v,krui_getVersion());
  UNPROTECT(1);
  return version;
}


-- 
View this message in context:
http://www.nabble.com/is-it-necessary-to-always-register-C-routines-with-R_registerRoutines--tf3063486.html#a8520292
Sent from the R devel mailing list archive at Nabble.com.

No, in answer to the subject line.

On Tue, 23 Jan 2007, Vladimir Eremeev wrote:
> I am writing bindings to the neural network simulator SNNS.
> At present I have used only .C interface, now I'm studying .Call
interface.
>
> I have adapted the example from page 77 of r-exts.pdf, however, it crashes
> R.
Page numbers depend on what paper it is formatted for, but on A4 this is 
nothing like the example on page 77.
> I use MingW as recommended by Duncan Murdoch.
> Please, tell me what I am missing.
> The code is below.
>
> Thank you.
>
> SEXP snns_getVersion(void)
> {SEXP version;
> char *v;
>
>  PROTECT(version=NEW_CHARACTER(15));
>  v=CHARACTER_POINTER(version);
>  strcpy(v,krui_getVersion());
I very much doubt that does what you think it does (but you have not told 
us).  I think the S-compatible Rdefines.h macros are really confusing for 
use in R, as R and S differ in how character vectors are stored.  In 
particular, use of CHARACTER_POINTER is likely to get into trouble as it 
is necessary to use SET_STRING_ELT in R.  (Similarly, CREATE_STRING_VECTOR 
does no such thing: it creates an element of a character vector.)

In this case

 	version = mkString(krui_getVersion());

would seem to be all that is needed.  But if you want to do it by hand,

     PROTECT(version=NEW_CHARACTER(1));
     SET_STRING_ELT(version, 0, mkChar(krui_getVersion()));
     UNPROTECT(1);

is the sort of thing needed.
>  UNPROTECT(1);
>  return version;
> }
>
>> sessionInfo()
> R version 2.4.1 (2006-12-18)
> i386-pc-mingw32
>
> locale:
>
LC_COLLATE=Russian_Russia.1251;LC_CTYPE=Russian_Russia.1251;LC_MONETARY=Russian_Russia.1251;LC_NUMERIC=C;LC_TIME=Russian_Russia.1251
>
> attached base packages:
> [1] "stats"     "graphics"  "grDevices"
"utils"     "datasets"  "methods"
> [7] "base"
>
>
>
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595