/// compose a wide char from a sequence of getchars
///  this is a slow lane routine, called if no specialised code
///  isavailable.
static int GETW(int sno) {
  StreamDesc *st = GLOBAL_Stream + sno;
  int ch = GETC();

  if (ch == -1)
    return post_process_weof(st);

  switch (st->encoding) {
  case ENC_OCTET:
  // no error detection, all characters are ok.
  case ENC_ISO_LATIN1:
    return post_process_read_wchar(ch, 1, st);
  // 7 bits code, anything above is bad news
  case ENC_ISO_ASCII:
    if (ch & 0x80) {
      /* error */
    }
    return post_process_read_wchar(ch, 1, st);
  // default OS encoding, depends on locale.
  case ENC_ISO_ANSI: {
    char buf[8];
    int out;
    wchar_t wch;
    mbstate_t mbstate;

    memset((void *)&(mbstate), 0, sizeof(mbstate_t));
    buf[0] = ch;
    int n = 1;
    while ((out = mbrtowc(&wch, buf, 1, &(mbstate))) != 1) {
      int ch = buf[0] = GETC();
      n++;
      if (ch == -1)
        return post_process_weof(st);
    }
    return post_process_read_wchar(wch, n, st);
  }
  // UTF-8 works o 8 bits.
  case ENC_ISO_UTF8: {
    int wch;
    unsigned char buf[8];

    if (ch < 0x80) {
      return post_process_read_wchar(ch, 1, st);
    }
    // if ((ch - 0xc2) > (0xf4-0xc2)) return UTF8PROC_ERROR_INVALIDUTF8;
    if (ch < 0xe0) { // 2-byte sequence
                     // Must have valid continuation character
      int c1 = buf[0] = GETC();
      if (c1 == -1)
        return post_process_weof(st);
      // if (!utf_cont(*str)) return UTF8PROC_ERROR_INVALIDUTF8;
      wch = ((ch & 0x1f) << 6) | (c1 & 0x3f);
      return post_process_read_wchar(wch, 2, st);
    }
    if (ch < 0xf0) { // 3-byte sequence
      // if ((str + 1 >= end) || !utf_cont(*str) || !utf_cont(str[1]))
      //   return UTF8PROC_ERROR_INVALIDUTF8;
      // Check for surrogate chars
      // if (ch == 0xed && *str > 0x9f)
      //    return UTF8PROC_ERROR_INVALIDUTF8;
      int c1 = GETC();
      if (c1 == -1)
        return post_process_weof(st);
      int c2 = GETC();
      if (c2 == -1)
        return post_process_weof(st);
      wch = ((ch & 0xf) << 12) | ((c1 & 0x3f) << 6) | (c2 & 0x3f);
      return post_process_read_wchar(wch, 3, st);
    } else {
      int c1 = GETC();
      if (c1 == -1)
        return post_process_weof(st);
      int c2 = GETC();
      if (c2 == -1)
        return post_process_weof(st);
      int c3 = GETC();
      if (c3 == -1)
        return post_process_weof(st);
      wch = ((ch & 7) << 18) | ((c1 & 0x3f) << 12) | ((c2 & 0x3f) << 6) |
            (c3 & 0x3f);
      return post_process_read_wchar(wch, 4, st);
    }
  }
  case ENC_UTF16_LE: // check http://unicode.org/faq/utf_bom.html#utf16-3
                     // little-endian: start with big shot
    {
      int wch;
      int c1 = GETC();
      if (c1 == -1)
        return post_process_weof(st);
      wch = (c1 << 8) + ch;
      if (wch >= 0xd800 && wch < 0xdc00) {
        int c2 = GETC();
        if (c2 == -1)
          return post_process_weof(st);
        int c3 = GETC();
        if (c3 == -1)
          return post_process_weof(st);
        wch = wch + (((c3 << 8) + c2) << wch) + SURROGATE_OFFSET;
        return post_process_read_wchar(wch, 4, st);
      }
      return post_process_read_wchar(wch, 2, st);
    }

  case ENC_UTF16_BE: // check http://unicode.org/faq/utf_bom.html#utf16-3
                     // little-endian: start with big shot
    {
      int wch;
      int c1 = GETC();
      if (c1 == -1)
        return post_process_weof(st);
      wch = (c1) + (ch << 8);
      if (wch >= 0xd800 && wch < 0xdc00) {
        int c3 = GETC();
        if (c3 == -1)
          return post_process_weof(st);
        int c2 = GETC();
        if (c2 == -1)
          return post_process_weof(st);
        wch = (((c3 << 8) + c2) << 10) + wch + SURROGATE_OFFSET;
        return post_process_read_wchar(wch, 4, st);
      }
      return post_process_read_wchar(wch, 2, st);
    }

  case ENC_UCS2_BE: // check http://unicode.org/faq/utf_bom.html#utf16-3
                    // little-endian: start with big shot
    {
      int wch;
      int c1 = GETC();
      if (c1 == -1)
        return post_process_weof(st);
      wch = (c1) + (ch << 8);
      return post_process_read_wchar(wch, 2, st);
    }

  case ENC_UCS2_LE: // check http://unicode.org/faq/utf_bom.html#utf16-3
                    // little-endian: start with big shot
    {
      int wch;
      int c1 = GETC();
      if (c1 == -1)
        return post_process_weof(st);
      wch = (c1 << 8) + ch;

      return post_process_read_wchar(wch, 2, st);
    }

  case ENC_ISO_UTF32_BE: // check http://unicode.org/faq/utf_bom.html#utf16-3
    // little-endian: start with big shot
    {
      int wch = ch;
      {
        int c1 = GETC();
        if (c1 == -1)
          return post_process_weof(st);
        wch = wch + c1;
      }
      {
        int c1 = GETC();
        if (c1 == -1)
          return post_process_weof(st);
        wch = (wch << 8) + c1;
      }
      {
        int c1 = GETC();
        if (c1 == -1)
          return post_process_weof(st);
        wch = (wch << 8) + c1;
      }
      return post_process_read_wchar(wch, 4, st);
    }
  case ENC_ISO_UTF32_LE: // check http://unicode.org/faq/utf_bom.html#utf16-3
    // little-endian: start with big shot
    {
      int wch = ch;
      {
        int c1 = GETC();
        if (c1 == -1)
          return post_process_weof(st);
        wch += c1 << 8;
      }
      {
        int c1 = GETC();
        if (c1 == -1)
          return post_process_weof(st);
        wch += c1 << 16;
      }
      {
        int c1 = GETC();
        if (c1 == -1)
          return post_process_weof(st);
        wch += c1 << 24;
      }
      return post_process_read_wchar(wch, 4, st);
    }
  default:
    Yap_Error(SYSTEM_ERROR_OPERATING_SYSTEM, MkIntTerm(st->encoding),
              "Bad Encoding\n");
    return -1;
  }
}