Re: [Mingw-users] how to convert Microsoft setlocale language string into POSIX language string ?

[Tor L. <tm...@ik...>]

> Is there a way to convert this string into a POSIX compatible string ?
>  "French_France.1252" => "fr_FR.ISO8859-1

A small matter of programming;)

Also, isn't it just so that using ISO639 language codes and ISO3166
country codes is a common convention in many implementations, not a
standard? http://www.opengroup.org/onlinepubs/009695399/functions/setlocale.html
says that "the contents of this string are implementation-defined".

If you really need to know the "POSIX-style" equivalent of the C
locale (which might be different from the Win32 thread locale in
"exotic" or pathological cases), you need to resort to string
comparisons, as far as I know.

If knowing the "POSIX-style" equivalent of the Win32 thread locale is
sufficient, the below code snippet might help. From GLib, so adapt as
necessary. (This code snippet is written by me, and I hereby place it
in the public domain.)

This code snippet doesn't provide the character codeset part, so use
something like sprintf(buf, "CP%d", GetACP()) for that. (GLib
uses/requires UTF-8 everywhere, so there is no reason for this GLib
function to provide the codeset part.)

--tml

gchar *
g_win32_getlocale (void)
{
  LCID lcid;
  LANGID langid;
  gchar *ev;
  gint primary, sub;
  char iso639[10];
  char iso3166[10];
  const gchar *script = NULL;

  /* Let the user override the system settings through environment
   * variables, as on POSIX systems. Note that in GTK+ applications
   * since GTK+ 2.10.7 setting either LC_ALL or LANG also sets the
   * Win32 locale and C library locale through code in gtkmain.c.
   */
  if (((ev = getenv ("LC_ALL")) != NULL && ev[0] != '\0')
      || ((ev = getenv ("LC_MESSAGES")) != NULL && ev[0] != '\0')
      || ((ev = getenv ("LANG")) != NULL && ev[0] != '\0'))
    return g_strdup (ev);

  lcid = GetThreadLocale ();

  if (!GetLocaleInfo (lcid, LOCALE_SISO639LANGNAME, iso639, sizeof (iso639)) ||
      !GetLocaleInfo (lcid, LOCALE_SISO3166CTRYNAME, iso3166, sizeof (iso3166)))
    return g_strdup ("C");

  /* Strip off the sorting rules, keep only the language part.  */
  langid = LANGIDFROMLCID (lcid);

  /* Split into language and territory part.  */
  primary = PRIMARYLANGID (langid);
  sub = SUBLANGID (langid);

  /* Handle special cases */
  switch (primary)
    {
    case LANG_AZERI:
      switch (sub)
        {
        case SUBLANG_AZERI_LATIN:
          script = "@Latn";
          break;
        case SUBLANG_AZERI_CYRILLIC:
          script = "@Cyrl";
          break;
        }
      break;
    case LANG_SERBIAN:          /* LANG_CROATIAN == LANG_SERBIAN */
      switch (sub)
        {
        case SUBLANG_SERBIAN_LATIN:
        case 0x06: /* Serbian (Latin) - Bosnia and Herzegovina */
          script = "@Latn";
          break;
        }
      break;
    case LANG_UZBEK:
      switch (sub)
        {
        case SUBLANG_UZBEK_LATIN:
          script = "@Latn";
          break;
        case SUBLANG_UZBEK_CYRILLIC:
          script = "@Cyrl";
          break;
        }
      break;
    }
  return g_strconcat (iso639, "_", iso3166, script, NULL);
}