[Rcpp-devel] String encoding (UTF-8 conversion)

[Jeroen Ooms]

On Thu, Dec 11, 2014 at 2:16 PM, Dirk Eddelbuettel <edd at debian.org> wrote:
> Maybe the same way?  ;-)

OK I am not completely back at the C api, and it still doesn't work
when called via Rcpp:

// [[Rcpp::export]]
SEXP echo(SEXP src){
  const char* utf8str = Rf_translateCharUTF8(Rf_asChar(src));
  SEXP out = PROTECT(Rf_allocVector(STRSXP, 1));
  SET_STRING_ELT(out, 0, Rf_mkCharCE(utf8str, CE_UTF8));
  UNPROTECT(1);
  return out;
}

> test
[1] "東京"
> echo(test)
[1] "æ ±äº¬"
> echo(enc2utf8 (test))
[1] "æ ±äº¬"

Also:

> echo("Zürich")
[1] "Zürich"
> echo(enc2utf8("Zürich"))
[1] "Zürich"

The same function works perfectly fine when invoked via .Call(). Does
Rcpp somehow override the CE or attempt to recode strings before
giving them back to R?