/* shindou7_no_ucase.c - Don Yang (uguu.org)

   02/24/07
                                                                           */

#include<assert.h>
#include<stdio.h>
#include<stdlib.h>
#include<string.h>
#ifdef _WIN32
   #include<io.h>
   #include<fcntl.h>
#endif

typedef enum
{
   UTF8        = 1,
   SJIS        = 2,
   EUCJP       = 4,
   UCS2LE      = 8,
   UCS2BE      = 16
} Encoding;

static FILE *infile;
static unsigned char *tmp, *buffer;
static char *const_str = " \0!\0 \1\xff\1\xef\xbc\x81I\xa1\xaa";
static int a,
           state,
           buffer_size, read_size,
           i, i0, last_char_offset,
           c,
           utf8_kana,
           sjis_kana,
           eucjp_kana,
           encoding;

static unsigned char parse_ucs2[] =
{
   2
};
static unsigned char parse_utf8[] =
{
   0xf0, 0xf7,
      0x80, 0xbf,
      0x80, 0xbf,
      0x80, 0xbf,
      4,
   0xe0, 0xef,
      0x80, 0xbf,
      0x80, 0xbf,
      3,
   0xc0, 0xdf,
      0x80, 0xbf,
      2,
   1
};
static unsigned char parse_sjis[] =
{
   0x81, 0x9f,
      0x40, 0xfc,
      2,
   0xe0, 0xef,
      0x40, 0xfc,
      2,
   1
};
static unsigned char parse_eucjp[] =
{
   0xa1, 0xa8,
      0xa1, 0xfe,
      2,
   0xad, 0xad,
      0xa1, 0xfe,
      2,
   0xb0, 0xf4,
      0xa1, 0xfe,
      2,
   0xf9, 0xfc,
      0xa1, 0xfe,
      2,
   0x8e, 0x8e,
      0xa1, 0xfe,
      2,
   0x8f, 0x8f,
      0xa1, 0xfe,
      0xa1, 0xfe,
      3,
   1
};
static unsigned char *parser;

static void Output(void *buffer, int size)
{
   fwrite(buffer, size, 1, stdout);
}

static void ExtendBuffer(void)
{
   if( read_size == buffer_size )
   {
      tmp = buffer;
      buffer = malloc((buffer_size *= 2) + 3);
      assert(buffer != NULL);
      memcpy(buffer, tmp, read_size);
      free(tmp);
   }
   read_size += fread(buffer + read_size, 1, buffer_size - read_size, infile);
}

static void NextChar(void)
{
   unsigned char *p;
   int byte_offset;
   c = 0;
   byte_offset = 0;
   for(p = parser; *p > 4 && c >= 0; p++)
   {
      if( buffer[i + byte_offset] < *p || buffer[i + byte_offset] > p[1] )
      {
         /* Byte not within range */
         if( byte_offset )
         {
            /* Not matching first byte -> character does not fit encoding */
            c = -1;
         }
         else
         {
            /* Try next range */
            for(; *p > 4; p++);
         }
      }
      else
      {
         /* Byte within range */
         byte_offset++;
         p++;
      }
   }
   for(byte_offset = 0; c >= 0 && byte_offset < *p; byte_offset++)
      c = (c << 8) | buffer[i++];
   if( encoding & UCS2LE )
      c = (c >> 8) | ((c & 255) << 8);
}

static void DetectFileEncoding(void)
{
   read_size =
   utf8_kana =
   sjis_kana =
   eucjp_kana = 0;
   for(encoding = UTF8 | SJIS | EUCJP;
       !feof(infile) && (encoding & (encoding - 1));)
   {
      ExtendBuffer();

      if( !(encoding & (UCS2LE|UCS2BE)) )
      {
         if( buffer[0] == 0xff && buffer[1] == 0xfe )
         {
            encoding = UCS2LE;
         }
         else if( buffer[0] == 0xfe && buffer[1] == 0xff )
         {
            encoding = UCS2BE;
         }
         else
         {
            for(i = 0; i < read_size - 1; i += 2)
            {
               if( buffer[i] == 0x0a && !buffer[i + 1] )
               {
                  encoding = UCS2LE;
                  goto finalize;
               }
               if( !buffer[i] && buffer[i + 1] == 0x0a )
               {
                  encoding = UCS2BE;
                  goto finalize;
               }
            }

            if( buffer[0] == 0xef &&
                buffer[1] == 0xbb &&
                buffer[2] == 0xbf )
            {
               encoding = UTF8;
            }
            else
            {
               parser = parse_utf8;
               for(i = 0; i < read_size - 3 && encoding & UTF8;)
               {
                  NextChar();
                  if( c < 0 )
                  {
                     encoding &= ~UTF8;
                  }
                  else if( (c >= 0xe38181 && c <= 0xe381bf) ||
                           (c >= 0xe38280 && c <= 0xe38293) )
                  {
                     utf8_kana++;
                  }
               }

               parser = parse_sjis;
               for(i = 0; i < read_size - 1 && encoding & SJIS;)
               {
                  NextChar();
                  if( c < 0 )
                  {
                     encoding &= ~SJIS;
                  }
                  else if( c >= 0x829f && c <= 0x82f1 )
                  {
                     sjis_kana++;
                  }
               }

               parser = parse_eucjp;
               for(i = 0; i < read_size - 2 && encoding & EUCJP;)
               {
                  NextChar();
                  if( c < 0 )
                  {
                     encoding &= ~EUCJP;
                  }
                  else if( c >= 0xa4a1 && c <= 0xa4f3 )
                  {
                     eucjp_kana++;
                  }
               }
            }
         }
      }
   }

finalize:
   if( !(encoding & (UCS2LE|UCS2BE)) )
   {
      if( !(encoding & UTF8) ) utf8_kana = -1;
      if( !(encoding & SJIS) ) sjis_kana = -1;
      if( !(encoding & EUCJP) ) eucjp_kana = -1;

      encoding =
         (utf8_kana < sjis_kana)
            ?  (sjis_kana < eucjp_kana) ? EUCJP : SJIS
            :  (utf8_kana < eucjp_kana) ? EUCJP :
                  encoding & UTF8 ? UTF8 : EUCJP;
   }
}

static int IsExclaimationPunct(void)
{
   return
      c == 0x21 ||
      (encoding & (UCS2LE|UCS2BE) && c == 0xff01) ||
      (encoding & UTF8 && c == 0xefbc81) ||
      (encoding & SJIS && c == 0x8149) ||
      (encoding & EUCJP && c == 0xa1aa);
}

static int IsEOSPunct(void)
{
   return
      c == 0x2e || c == 0x3f ||
      (encoding & (UCS2LE|UCS2BE) &&
       (c == 0xff1f || c == 0xff0e || c == 0x3002 || c == 0xff61)) ||
      (encoding & UTF8 &&
       (c == 0xefbc9f || c == 0xefbc8e || c == 0xe38082 || c == 0xefbda1)) ||
      (encoding & SJIS &&
       (c == 0x8148 || c == 0x8144 || c == 0x8142)) ||
      (encoding & EUCJP &&
       (c == 0xa1a9 || c == 0xa1a5 || c == 0xa1a3)) ||
      IsExclaimationPunct();
}

static int IsPunct(void)
{
   return
      c == 0x2c || c == 0x0a ||
      (encoding & (UCS2LE|UCS2BE) &&
       (c == 0xff0c || c == 0x3001 || c == 0xff64)) ||
      (encoding & UTF8 &&
       (c == 0xefbc8c || c == 0xe38081 || c == 0xefbda4)) ||
      (encoding & SJIS &&
       (c == 0x8143 || c == 0x8141)) ||
      (encoding & EUCJP &&
       (c == 0xa1a4 || c == 0xa1a2)) ||
      IsEOSPunct();
}

static void WriteExclamationMark(void)
{
   Output(const_str +
            (c > 255
               ?  encoding & UCS2BE ? 6 :
                  encoding & UCS2LE ? 5 :
                  encoding & UTF8 ? 8 :
                  encoding & SJIS ? 10 : 12
               :  encoding & UCS2BE ? 1 : 2),
          c > 255
            ?  encoding & UTF8 ? 3 : 2
            :  encoding & (UCS2LE|UCS2BE) ? 2 : 1);
}

static void Filter(void)
{
   for(i = i0 = state = 0; !feof(infile) || i < read_size;)
   {
      last_char_offset = i;
      NextChar();
      if( c < 0 )
         i++;
      if( i < read_size - 3 )
      {
         memmove(buffer, buffer + i0, read_size - i0);
         read_size -= i0;
         i -= i0;
         last_char_offset -= i0;
         i0 = 0;

         ExtendBuffer();
      }

      if( state )
      {
         if( IsPunct() )
         {
            if( IsEOSPunct() )
            {
               for(i = i0; i < last_char_offset; i0 = i)
               {
                  NextChar();
                  Output(buffer + i0, i - i0);
                  Output(const_str + (encoding & UCS2BE ? 3 : 0),
                         encoding & (UCS2LE|UCS2BE) ? 2 : 1);
               }

               NextChar();
            }

            if( i0 < last_char_offset )
               Output(buffer + i0, last_char_offset - i0);
            if( IsExclaimationPunct() )
            {
               WriteExclamationMark();
               WriteExclamationMark();
            }
            else if( IsPunct() && c != 0x0a )
            {
               WriteExclamationMark();
            }
            else
            {
               Output(buffer + last_char_offset, i - last_char_offset);
            }
            state = 0;
         }
      }
      else if( c < 0 || IsPunct() )
      {
         Output(buffer + last_char_offset, i - last_char_offset);
      }
      else
      {
         state = 1;
         i0 = last_char_offset;
      }
   }
}

int main(int argc, char **argv)
{
   #ifdef _WIN32
      setmode(fileno(stdin), O_BINARY);
      setmode(fileno(stdout), O_BINARY);
   #endif

   buffer_size = 0x1000;
   buffer = malloc(buffer_size);
   assert(buffer != NULL);

   if( argc > 1 )
   {
      for(a = 1; a < argc; a++)
      {
         if( (infile = fopen(argv[a], "rb")) )
         {
            DetectFileEncoding();
            printf("%s: %s\n", argv[a],
                   encoding != UCS2LE
                     ? encoding != UCS2BE
                        ? encoding != UTF8
                           ? encoding != SJIS
                              ? "EUC-JP"
                              : "Shift_JIS"
                           : "UTF-8"
                        : "UCS-2BE"
                     : "UCS-2LE");

            fclose(infile);
         }
         else
         {
            printf("%s: can not open file\n", argv[a]);
         }
      }
   }
   else
   {
      infile = stdin;
      DetectFileEncoding();
      parser =
         encoding != UTF8
            ?  encoding != SJIS
               ?  encoding != EUCJP
                  ?  parse_ucs2
                  :  parse_eucjp
               :  parse_sjis
            :  parse_utf8;
      Filter();
   }

   free(buffer);
   return 0;
}