From: Tanaka Akira <akr@...>
Date: 2007-12-31T12:38:14+09:00
Subject: Re: [Spam] Re: Wide strings to ruby strings

In article <20071230084654.90944E0483@mail.bc9.jp>,
  Nobuyoshi Nakada <nobu@ruby-lang.org> writes:

> X11, and Solaris (IIRC).

I tested on Solaris.

% uname -a
SunOS solaris 5.11 snv_70b i86pc i386 i86pc
% cat wdump.c 
#include <wchar.h>
#include <stdlib.h>
#include <stdio.h>
#include <locale.h>

int main(int argc, char **argv)
{
  char *s;
  wint_t wc;

  s = setlocale(LC_ALL, "");
  if (s == NULL) { perror("setlocale"); exit(1); }

  printf("sizeof(wchar_t):%d\n", sizeof(wchar_t));

  while ((wc = getwchar()) != WEOF) {
    printf("0x%lx\n", (long)wc);
  }

  if (ferror(stdin)) { perror("getwchar"); exit(1); }

  return 0;
}
% ./ruby -e 'print "\u3042"'|LANG=ja_JP.UTF-8 ./wdump 
sizeof(wchar_t):4
0x3042
% ./ruby -e 'print "\u3042"'|iconv -f UTF-8 -t eucJP|LANG=ja_JP.eucJP ./wdump 
sizeof(wchar_t):4
0x30001222

HIRAGANA LETTER A (U+3042) is represented differently on
ja_JP.UTF-8 and ja_JP.eucJP.

We don't know the internal of wchar_t.
(__STDC_ISO_10646__ is optional.)

But the encoding of mbs is known: nl_langinfo(CODESET).

So I recommend wcs -> mbs -> UTF-8.
-- 
Tanaka Akira