X-Git-Url: http://git.chise.org/gitweb/?a=blobdiff_plain;f=src%2Ffile-coding.c;h=3068c896823dac45e334b378d4adf0bf92d709c6;hb=4082f470d7c6075a5c04c1bdf6e9e3f27c9700d5;hp=7a7eee3ad32196326ec51dce11096b75eb16d8ca;hpb=44e716ef11bd794a51f8c5b56c4f3f10a7dbf217;p=chise%2Fxemacs-chise.git-

diff --git a/src/file-coding.c b/src/file-coding.c
index 7a7eee3..3068c89 100644
--- a/src/file-coding.c
+++ b/src/file-coding.c
@@ -25,6 +25,7 @@ Boston, MA 02111-1307, USA.  */
 
 #include <config.h>
 #include "lisp.h"
+
 #include "buffer.h"
 #include "elhash.h"
 #include "insdel.h"
@@ -244,10 +245,46 @@ static Lisp_Object mark_coding_system (Lisp_Object, void (*) (Lisp_Object));
 static void print_coding_system (Lisp_Object, Lisp_Object, int);
 static void finalize_coding_system (void *header, int for_disksave);
 
+#ifdef MULE
+static const struct lrecord_description ccs_description_1[] = {
+  { XD_LISP_OBJECT, offsetof(charset_conversion_spec, from_charset), 2 },
+  { XD_END }
+};
+
+static const struct struct_description ccs_description = {
+  sizeof(charset_conversion_spec),
+  ccs_description_1
+};
+  
+static const struct lrecord_description ccsd_description_1[] = {
+  XD_DYNARR_DESC(charset_conversion_spec_dynarr, &ccs_description),
+  { XD_END }
+};
+
+static const struct struct_description ccsd_description = {
+  sizeof(charset_conversion_spec_dynarr),
+  ccsd_description_1
+};
+#endif
+
+static const struct lrecord_description coding_system_description[] = {
+  { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, name), 2 },
+  { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, mnemonic), 3 },
+  { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, eol_lf), 3 },
+#ifdef MULE
+  { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, iso2022.initial_charset), 4 },
+  { XD_STRUCT_PTR,  offsetof(struct Lisp_Coding_System, iso2022.input_conv),  1, &ccsd_description },
+  { XD_STRUCT_PTR,  offsetof(struct Lisp_Coding_System, iso2022.output_conv), 1, &ccsd_description },
+  { XD_LISP_OBJECT, offsetof(struct Lisp_Coding_System, ccl.decode), 2 },
+#endif
+  { XD_END }
+};
+
 DEFINE_LRECORD_IMPLEMENTATION ("coding-system", coding_system,
 			       mark_coding_system, print_coding_system,
 			       finalize_coding_system,
-			       0, 0, struct Lisp_Coding_System);
+			       0, 0, coding_system_description,
+			       struct Lisp_Coding_System);
 
 static Lisp_Object
 mark_coding_system (Lisp_Object obj, void (*markobj) (Lisp_Object))
@@ -1583,26 +1620,65 @@ determine_real_coding_system (Lstream *stream, Lisp_Object *codesys_in_out,
   if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT ||
       *eol_type_in_out == EOL_AUTODETECT)
     {
+      unsigned char random_buffer[4096];
+      int nread;
+      Lisp_Object coding_system = Qnil;
 
-      while (1)
+      nread = Lstream_read (stream, random_buffer, sizeof (random_buffer));
+      if (nread)
 	{
-	  unsigned char random_buffer[4096];
-	  int nread;
+	  unsigned char *cp = random_buffer;
 
-	  nread = Lstream_read (stream, random_buffer, sizeof (random_buffer));
-	  if (!nread)
-	    break;
-	  if (detect_coding_type (&decst, random_buffer, nread,
-				  XCODING_SYSTEM_TYPE (*codesys_in_out) !=
-				  CODESYS_AUTODETECT))
-	    break;
-	}
+	  while (cp < random_buffer + nread)
+	    {
+	      if ((*cp++ == 'c') && (cp < random_buffer + nread) &&
+		  (*cp++ == 'o') && (cp < random_buffer + nread) &&
+		  (*cp++ == 'd') && (cp < random_buffer + nread) &&
+		  (*cp++ == 'i') && (cp < random_buffer + nread) &&
+		  (*cp++ == 'n') && (cp < random_buffer + nread) &&
+		  (*cp++ == 'g') && (cp < random_buffer + nread) &&
+		  (*cp++ == ':') && (cp < random_buffer + nread))
+		{
+		  unsigned char coding_system_name[4096 - 6];
+		  unsigned char *np = coding_system_name;
 
+		  while ( (cp < random_buffer + nread)
+			  && ((*cp == ' ') || (*cp == '\t')) )
+		    {
+		      cp++;
+		    }
+		  while ( (cp < random_buffer + nread) &&
+			  (*cp != ' ') && (*cp != '\t') && (*cp != ';') )
+		    {
+		      *np++ = *cp++;
+		    }
+		  *np = 0;
+		  coding_system
+		    = Ffind_coding_system (intern (coding_system_name));
+		  break;
+		}
+	    }
+	  if (EQ(coding_system, Qnil))
+	    do{
+	      if (detect_coding_type (&decst, random_buffer, nread,
+				      XCODING_SYSTEM_TYPE (*codesys_in_out)
+				      != CODESYS_AUTODETECT))
+		break;
+	      nread = Lstream_read (stream,
+				    random_buffer, sizeof (random_buffer));
+	      if (!nread)
+		break;
+	    } while(1);
+	}
       *eol_type_in_out = decst.eol_type;
       if (XCODING_SYSTEM_TYPE (*codesys_in_out) == CODESYS_AUTODETECT)
-	*codesys_in_out = coding_system_from_mask (decst.mask);
+	{
+	  if (EQ(coding_system, Qnil))
+	    *codesys_in_out = coding_system_from_mask (decst.mask);
+	  else
+	    *codesys_in_out = coding_system;
+	}
     }
-
   /* If we absolutely can't determine the EOL type, just assume LF. */
   if (*eol_type_in_out == EOL_AUTODETECT)
     *eol_type_in_out = EOL_LF;
@@ -1725,6 +1801,62 @@ do {								\
 /* C should be a binary character in the range 0 - 255; convert
    to internal format and add to Dynarr DST. */
 
+#ifdef UTF2000
+#define DECODE_ADD_BINARY_CHAR(c, dst) \
+do {						\
+  if (BYTE_ASCII_P (c))				\
+    Dynarr_add (dst, c);			\
+  else						\
+    {						\
+      Dynarr_add (dst, (c >> 6) | 0xc0);	\
+      Dynarr_add (dst, (c & 0x3f) | 0x80);	\
+    }						\
+} while (0)
+
+INLINE void
+DECODE_ADD_UCS_CHAR(Emchar c, unsigned_char_dynarr* dst)
+{
+  if ( c <= 0x7f )
+    {
+      Dynarr_add (dst, c);
+    }
+  else if ( c <= 0x7ff )
+    {
+      Dynarr_add (dst, (c >> 6) | 0xc0);
+      Dynarr_add (dst, (c & 0x3f) | 0x80);
+    }
+  else if ( c <= 0xffff )
+    {
+      Dynarr_add (dst,  (c >> 12) | 0xe0);
+      Dynarr_add (dst, ((c >>  6) & 0x3f) | 0x80);
+      Dynarr_add (dst,  (c        & 0x3f) | 0x80);
+    }
+  else if ( c <= 0x1fffff )
+    {
+      Dynarr_add (dst,  (c >> 18) | 0xf0);
+      Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
+      Dynarr_add (dst, ((c >>  6) & 0x3f) | 0x80);
+      Dynarr_add (dst,  (c        & 0x3f) | 0x80);
+    }
+  else if ( c <= 0x3ffffff )
+    {
+      Dynarr_add (dst,  (c >> 24) | 0xf8);
+      Dynarr_add (dst, ((c >> 18) & 0x3f) | 0x80);
+      Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
+      Dynarr_add (dst, ((c >>  6) & 0x3f) | 0x80);
+      Dynarr_add (dst,  (c        & 0x3f) | 0x80);
+    }
+  else
+    {
+      Dynarr_add (dst,  (c >> 30) | 0xfc);
+      Dynarr_add (dst, ((c >> 24) & 0x3f) | 0x80);
+      Dynarr_add (dst, ((c >> 18) & 0x3f) | 0x80);
+      Dynarr_add (dst, ((c >> 12) & 0x3f) | 0x80);
+      Dynarr_add (dst, ((c >>  6) & 0x3f) | 0x80);
+      Dynarr_add (dst,  (c        & 0x3f) | 0x80);
+    }
+}
+#else
 #define DECODE_ADD_BINARY_CHAR(c, dst)		\
 do {						\
   if (BYTE_ASCII_P (c))				\
@@ -1740,6 +1872,7 @@ do {						\
       Dynarr_add (dst, c);			\
     }						\
 } while (0)
+#endif
 
 #define DECODE_OUTPUT_PARTIAL_CHAR(ch)	\
 do {					\
@@ -1799,6 +1932,9 @@ struct decoding_stream
   /* Additional information (the state of the running CCL program)
      used by the CCL decoder. */
   struct ccl_program ccl;
+
+  /* counter for UTF-8 or UCS-4 */
+  unsigned char counter;
 #endif
   struct detection_state decst;
 };
@@ -1933,6 +2069,7 @@ reset_decoding_stream (struct decoding_stream *str)
     {
       setup_ccl_program (&str->ccl, CODING_SYSTEM_CCL_DECODE (str->codesys));
     }
+  str->counter = 0;
 #endif /* MULE */
   str->flags = str->ch = 0;
 }
@@ -2116,6 +2253,7 @@ mule_decode (Lstream *decoding, CONST unsigned char *src,
       decode_coding_utf8 (decoding, src, dst, n);
       break;
     case CODESYS_CCL:
+      str->ccl.last_block = str->flags & CODING_STATE_END;
       ccl_driver (&str->ccl, src, dst, n, 0, CCL_MODE_DECODING);
       break;
     case CODESYS_ISO2022:
@@ -2394,7 +2532,11 @@ reset_encoding_stream (struct encoding_stream *str)
 	str->iso2022.register_right = 1;
 	str->iso2022.current_charset = Qnil;
 	str->iso2022.current_half = 0;
+#ifdef UTF2000
+	str->iso2022.current_char_boundary = 0;
+#else
 	str->iso2022.current_char_boundary = 1;
+#endif
 	break;
       }
     case CODESYS_CCL:
@@ -2528,6 +2670,7 @@ mule_encode (Lstream *encoding, CONST unsigned char *src,
       encode_coding_utf8 (encoding, src, dst, n);
       break;
     case CODESYS_CCL:
+      str->ccl.last_block = str->flags & CODING_STATE_END;
       ccl_driver (&str->ccl, src, dst, n, 0, CCL_MODE_ENCODING);
       break;
     case CODESYS_ISO2022:
@@ -2687,10 +2830,16 @@ decode_coding_sjis (Lstream *decoding, CONST unsigned char *src,
 	    {
 	      unsigned char e1, e2;
 
-	      Dynarr_add (dst, LEADING_BYTE_JAPANESE_JISX0208);
 	      DECODE_SJIS (ch, c, e1, e2);
+#ifdef UTF2000
+	      DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_japanese_jisx0208,
+					    e1 & 0x7F,
+					    e2 & 0x7F), dst);
+#else
+	      Dynarr_add (dst, LEADING_BYTE_JAPANESE_JISX0208);
 	      Dynarr_add (dst, e1);
 	      Dynarr_add (dst, e2);
+#endif
 	    }
 	  else
 	    {
@@ -2706,8 +2855,13 @@ decode_coding_sjis (Lstream *decoding, CONST unsigned char *src,
 	    ch = c;
 	  else if (BYTE_SJIS_KATAKANA_P (c))
 	    {
+#ifdef UTF2000
+	      DECODE_ADD_UCS_CHAR(MAKE_CHAR(Vcharset_katakana_jisx0201,
+					    c & 0x7F, 0), dst);
+#else
 	      Dynarr_add (dst, LEADING_BYTE_KATAKANA_JISX0201);
 	      Dynarr_add (dst, c);
+#endif
 	    }
 	  else
 	    DECODE_ADD_BINARY_CHAR (c, dst);
@@ -2732,10 +2886,82 @@ encode_coding_sjis (Lstream *encoding, CONST unsigned char *src,
   unsigned int flags  = str->flags;
   unsigned int ch     = str->ch;
   eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
+#ifdef UTF2000
+  unsigned char char_boundary = str->iso2022.current_char_boundary;
+#endif
 
   while (n--)
     {
       c = *src++;
+#ifdef UTF2000
+      switch (char_boundary)
+	{
+	case 0:
+	  if ( c >= 0xfc )
+	    {
+	      ch = c & 0x01;
+	      char_boundary = 5;
+	    }
+	  else if ( c >= 0xf8 )
+	    {
+	      ch = c & 0x03;
+	      char_boundary = 4;
+	    }
+	  else if ( c >= 0xf0 )
+	    {
+	      ch = c & 0x07;
+	      char_boundary = 3;
+	    }
+	  else if ( c >= 0xe0 )
+	    {
+	      ch = c & 0x0f;
+	      char_boundary = 2;
+	    }
+	  else if ( c >= 0xc0 )
+	    {
+	      ch = c & 0x1f;
+	      char_boundary = 1;
+	    }
+	  else
+	    {
+	      ch = 0;
+	      if (c == '\n')
+		{
+		  if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
+		    Dynarr_add (dst, '\r');
+		  if (eol_type != EOL_CR)
+		    Dynarr_add (dst, c);
+		}
+	      else
+		Dynarr_add (dst, c);
+	      char_boundary = 0;
+	    }
+	  break;
+	case 1:
+	  ch = ( ch << 6 ) | ( c & 0x3f );
+	  {
+	    Lisp_Object charset;
+	    unsigned int c1, c2, s1, s2;
+	    
+	    BREAKUP_CHAR (ch, charset, c1, c2);
+	    if (EQ(charset, Vcharset_katakana_jisx0201))
+	      {
+		Dynarr_add (dst, c1 | 0x80);
+	      }
+	    else if (EQ(charset, Vcharset_japanese_jisx0208))
+	      {
+		ENCODE_SJIS (c1 | 0x80, c2 | 0x80, s1, s2);
+		Dynarr_add (dst, s1);
+		Dynarr_add (dst, s2);
+	      }
+	  }
+	  char_boundary = 0;
+	  break;
+	default:
+	  ch = ( ch << 6 ) | ( c & 0x3f );
+	  char_boundary--;
+	}
+#else
       if (c == '\n')
 	{
 	  if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
@@ -2772,10 +2998,14 @@ encode_coding_sjis (Lstream *encoding, CONST unsigned char *src,
 	      ch = 0;
 	    }
 	}
+#endif
     }
 
   str->flags = flags;
   str->ch    = ch;
+#ifdef UTF2000
+  str->iso2022.current_char_boundary = char_boundary;
+#endif
 }
 
 DEFUN ("decode-shift-jis-char", Fdecode_shift_jis_char, 1, 1, 0, /*
@@ -3007,6 +3237,7 @@ static void
 encode_coding_big5 (Lstream *encoding, CONST unsigned char *src,
 		    unsigned_char_dynarr *dst, unsigned int n)
 {
+#ifndef UTF2000
   unsigned char c;
   struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
   unsigned int flags  = str->flags;
@@ -3061,6 +3292,7 @@ encode_coding_big5 (Lstream *encoding, CONST unsigned char *src,
 
   str->flags = flags;
   str->ch    = ch;
+#endif
 }
 
 
@@ -3081,7 +3313,7 @@ Return the corresponding character.
   if (BYTE_BIG5_TWO_BYTE_1_P (b1) &&
       BYTE_BIG5_TWO_BYTE_2_P (b2))
     {
-      int leading_byte;
+      Charset_ID leading_byte;
       Lisp_Object charset;
       DECODE_BIG5 (b1, b2, leading_byte, c1, c2);
       charset = CHARSET_BY_LEADING_BYTE (leading_byte);
@@ -3199,6 +3431,9 @@ Return the UCS code (a positive integer) corresponding to CHARACTER.
   return Fget_char_table (character, mule_to_ucs_table);
 }
 
+#ifdef UTF2000
+#define decode_ucs4 DECODE_ADD_UCS_CHAR
+#else
 /* Decode a UCS-4 character into a buffer.  If the lookup fails, use
    <GETA MARK> (U+3013) of JIS X 0208, which means correct character
    is not found, instead.
@@ -3227,6 +3462,7 @@ decode_ucs4 (unsigned long ch, unsigned_char_dynarr *dst)
       Dynarr_add (dst, 46 + 128);
     }
 }
+#endif
 
 static unsigned long
 mule_char_to_ucs4 (Lisp_Object charset,
@@ -3304,37 +3540,40 @@ decode_coding_ucs4 (Lstream *decoding, CONST unsigned char *src,
   struct decoding_stream *str = DECODING_STREAM_DATA (decoding);
   unsigned int flags = str->flags;
   unsigned int ch    = str->ch;
+  unsigned char counter = str->counter;
 
   while (n--)
     {
       unsigned char c = *src++;
-      switch (flags)
+      switch (counter)
 	{
 	case 0:
 	  ch = c;
-	  flags = 3;
+	  counter = 3;
 	  break;
 	case 1:
 	  decode_ucs4 ( ( ch << 8 ) | c, dst);
 	  ch = 0;
-	  flags = 0;
+	  counter = 0;
 	  break;
 	default:
 	  ch = ( ch << 8 ) | c;
-	  flags--;
+	  counter--;
 	}
     }
-  if (flags & CODING_STATE_END)
+  if (counter & CODING_STATE_END)
     DECODE_OUTPUT_PARTIAL_CHAR (ch);
 
   str->flags = flags;
   str->ch    = ch;
+  str->counter = counter;
 }
 
 static void
 encode_coding_ucs4 (Lstream *encoding, CONST unsigned char *src,
 		    unsigned_char_dynarr *dst, unsigned int n)
 {
+#ifndef UTF2000
   struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
   unsigned int flags = str->flags;
   unsigned int ch = str->ch;
@@ -3463,6 +3702,7 @@ encode_coding_ucs4 (Lstream *encoding, CONST unsigned char *src,
   str->iso2022.current_charset = charset;
 
   /* Verbum caro factum est! */
+#endif
 }
 
 
@@ -3513,37 +3753,38 @@ decode_coding_utf8 (Lstream *decoding, CONST unsigned char *src,
   unsigned int flags  = str->flags;
   unsigned int ch     = str->ch;
   eol_type_t eol_type = str->eol_type;
+  unsigned char counter = str->counter;
 
   while (n--)
     {
       unsigned char c = *src++;
-      switch (flags)
+      switch (counter)
 	{
 	case 0:
 	  if ( c >= 0xfc )
 	    {
 	      ch = c & 0x01;
-	      flags = 5;
+	      counter = 5;
 	    }
 	  else if ( c >= 0xf8 )
 	    {
 	      ch = c & 0x03;
-	      flags = 4;
+	      counter = 4;
 	    }
 	  else if ( c >= 0xf0 )
 	    {
 	      ch = c & 0x07;
-	      flags = 3;
+	      counter = 3;
 	    }
 	  else if ( c >= 0xe0 )
 	    {
 	      ch = c & 0x0f;
-	      flags = 2;
+	      counter = 2;
 	    }
 	  else if ( c >= 0xc0 )
 	    {
 	      ch = c & 0x1f;
-	      flags = 1;
+	      counter = 1;
 	    }
 	  else
 	    {
@@ -3555,11 +3796,11 @@ decode_coding_utf8 (Lstream *decoding, CONST unsigned char *src,
 	  ch = ( ch << 6 ) | ( c & 0x3f );
 	  decode_ucs4 (ch, dst);
 	  ch = 0;
-	  flags = 0;
+	  counter = 0;
 	  break;
 	default:
 	  ch = ( ch << 6 ) | ( c & 0x3f );
-	  flags--;
+	  counter--;
 	}
     label_continue_loop:;
     }
@@ -3569,8 +3810,10 @@ decode_coding_utf8 (Lstream *decoding, CONST unsigned char *src,
 
   str->flags = flags;
   str->ch    = ch;
+  str->counter = counter;
 }
 
+#ifndef UTF2000
 static void
 encode_utf8 (Lisp_Object charset,
 	     unsigned char h, unsigned char l, unsigned_char_dynarr *dst)
@@ -3616,6 +3859,7 @@ encode_utf8 (Lisp_Object charset,
       Dynarr_add (dst,  (code        & 0x3f) | 0x80);
     }
 }
+#endif
 
 static void
 encode_coding_utf8 (Lstream *encoding, CONST unsigned char *src,
@@ -3626,6 +3870,63 @@ encode_coding_utf8 (Lstream *encoding, CONST unsigned char *src,
   unsigned int ch     = str->ch;
   eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
   unsigned char char_boundary = str->iso2022.current_char_boundary;
+#ifdef UTF2000
+
+  while (n--)
+    {
+      unsigned char c = *src++;	  
+      switch (char_boundary)
+	{
+	case 0:
+	  if ( c >= 0xfc )
+	    {
+	      Dynarr_add (dst, c);
+	      char_boundary = 5;
+	    }
+	  else if ( c >= 0xf8 )
+	    {
+	      Dynarr_add (dst, c);
+	      char_boundary = 4;
+	    }
+	  else if ( c >= 0xf0 )
+	    {
+	      Dynarr_add (dst, c);
+	      char_boundary = 3;
+	    }
+	  else if ( c >= 0xe0 )
+	    {
+	      Dynarr_add (dst, c);
+	      char_boundary = 2;
+	    }
+	  else if ( c >= 0xc0 )
+	    {
+	      Dynarr_add (dst, c);
+	      char_boundary = 1;
+	    }
+	  else
+	    {
+	      if (c == '\n')
+		{
+		  if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
+		    Dynarr_add (dst, '\r');
+		  if (eol_type != EOL_CR)
+		    Dynarr_add (dst, c);
+		}
+	      else
+		Dynarr_add (dst, c);
+	      char_boundary = 0;
+	    }
+	  break;
+	case 1:
+	  Dynarr_add (dst, c);
+	  char_boundary = 0;
+	  break;
+	default:
+	  Dynarr_add (dst, c);
+	  char_boundary--;
+	}
+    }
+#else /* not UTF2000 */
   Lisp_Object charset = str->iso2022.current_charset;
 
 #ifdef ENABLE_COMPOSITE_CHARS
@@ -3749,10 +4050,13 @@ encode_coding_utf8 (Lstream *encoding, CONST unsigned char *src,
     }
 #endif
 
+#endif /* not UTF2000 */
   str->flags = flags;
   str->ch    = ch;
   str->iso2022.current_char_boundary = char_boundary;
+#ifndef UTF2000
   str->iso2022.current_charset = charset;
+#endif
 
   /* Verbum caro factum est! */
 }
@@ -4652,7 +4956,9 @@ decode_coding_iso2022 (Lstream *decoding, CONST unsigned char *src,
       else
 	{			/* Graphic characters */
 	  Lisp_Object charset;
-	  int lb;
+#ifndef UTF2000
+	  Charset_ID lb;
+#endif
 	  int reg;
 
 	  DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
@@ -4695,6 +5001,22 @@ decode_coding_iso2022 (Lstream *decoding, CONST unsigned char *src,
 		    charset = new_charset;
 		}
 
+#ifdef UTF2000
+	      if (XCHARSET_DIMENSION (charset) == 1)
+		{
+		  DECODE_OUTPUT_PARTIAL_CHAR (ch);
+		  DECODE_ADD_UCS_CHAR
+		    (MAKE_CHAR (charset, c & 0x7F, 0), dst);
+		}
+	      else if (ch)
+		{
+		  DECODE_ADD_UCS_CHAR
+		    (MAKE_CHAR (charset, ch & 0x7F, c & 0x7F), dst);
+		  ch = 0;
+		}
+	      else
+		ch = c;
+#else
 	      lb = XCHARSET_LEADING_BYTE (charset);
 	      switch (XCHARSET_REP_BYTES (charset))
 		{
@@ -4743,6 +5065,7 @@ decode_coding_iso2022 (Lstream *decoding, CONST unsigned char *src,
 		  else
 		    ch = c;
 		}
+#endif
 	    }
 
 	  if (!ch)
@@ -4860,12 +5183,15 @@ encode_coding_iso2022 (Lstream *encoding, CONST unsigned char *src,
   unsigned char char_boundary;
   struct encoding_stream *str = ENCODING_STREAM_DATA (encoding);
   unsigned int flags          = str->flags;
-  unsigned int ch             = str->ch;
+  Emchar ch                   = str->ch;
   Lisp_Coding_System *codesys = str->codesys;
   eol_type_t eol_type         = CODING_SYSTEM_EOL_TYPE (str->codesys);
   int i;
   Lisp_Object charset;
   int half;
+#ifdef UTF2000
+  unsigned int byte1, byte2;
+#endif
 
 #ifdef ENABLE_COMPOSITE_CHARS
   /* flags for handling composite chars.  We do a little switcharoo
@@ -4882,6 +5208,216 @@ encode_coding_iso2022 (Lstream *encoding, CONST unsigned char *src,
 #ifdef ENABLE_COMPOSITE_CHARS
  back_to_square_n:
 #endif
+#ifdef UTF2000
+  while (n--)
+    {
+      c = *src++;
+
+      switch (char_boundary)
+	{
+	case 0:
+	  if ( c >= 0xfc )
+	    {
+	      ch = c & 0x01;
+	      char_boundary = 5;
+	    }
+	  else if ( c >= 0xf8 )
+	    {
+	      ch = c & 0x03;
+	      char_boundary = 4;
+	    }
+	  else if ( c >= 0xf0 )
+	    {
+	      ch = c & 0x07;
+	      char_boundary = 3;
+	    }
+	  else if ( c >= 0xe0 )
+	    {
+	      ch = c & 0x0f;
+	      char_boundary = 2;
+	    }
+	  else if ( c >= 0xc0 )
+	    {
+	      ch = c & 0x1f;
+	      char_boundary = 1;
+	    }
+	  else
+	    {
+	      ch = 0;
+
+	      restore_left_to_right_direction (codesys, dst, &flags, 0);
+	      
+	      /* Make sure G0 contains ASCII */
+	      if ((c > ' ' && c < ISO_CODE_DEL) ||
+		  !CODING_SYSTEM_ISO2022_NO_ASCII_CNTL (codesys))
+		{
+		  ensure_normal_shift (str, dst);
+		  iso2022_designate (Vcharset_ascii, 0, str, dst);
+		}
+	      
+	      /* If necessary, restore everything to the default state
+		 at end-of-line */
+	      if (c == '\n' &&
+		  !(CODING_SYSTEM_ISO2022_NO_ASCII_EOL (codesys)))
+		{
+		  restore_left_to_right_direction (codesys, dst, &flags, 0);
+
+		  ensure_normal_shift (str, dst);
+
+		  for (i = 0; i < 4; i++)
+		    {
+		      Lisp_Object initial_charset =
+			CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i);
+		      iso2022_designate (initial_charset, i, str, dst);
+		    }
+		}
+	      if (c == '\n')
+		{
+		  if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
+		    Dynarr_add (dst, '\r');
+		  if (eol_type != EOL_CR)
+		    Dynarr_add (dst, c);
+		}
+	      else
+		{
+		  if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
+		      && fit_to_be_escape_quoted (c))
+		    Dynarr_add (dst, ISO_CODE_ESC);
+		  Dynarr_add (dst, c);
+		}
+	      char_boundary = 0;
+	    }
+	  break;
+	case 1:
+	  ch = ( ch << 6 ) | ( c & 0x3f );
+	  
+	  char_boundary = 0;
+	  if ( (0x80 <= ch) && (ch <= 0x9f) )
+	    {
+	      charmask = (half == 0 ? 0x00 : 0x80);
+	  
+	      if (CODING_SYSTEM_ISO2022_ESCAPE_QUOTED (codesys)
+		  && fit_to_be_escape_quoted (ch))
+		Dynarr_add (dst, ISO_CODE_ESC);
+	      /* you asked for it ... */
+	      Dynarr_add (dst, ch);
+	    }
+	  else
+	    {
+	      int reg;
+
+	      BREAKUP_CHAR (ch, charset, byte1, byte2);
+	      ensure_correct_direction (XCHARSET_DIRECTION (charset),
+					codesys, dst, &flags, 0);
+
+	      /* Now determine which register to use. */
+	      reg = -1;
+	      for (i = 0; i < 4; i++)
+		{
+		  if (EQ (charset, str->iso2022.charset[i]) ||
+		      EQ (charset,
+			  CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, i)))
+		    {
+		      reg = i;
+		      break;
+		    }
+		}
+	      
+	      if (reg == -1)
+		{
+		  if (XCHARSET_GRAPHIC (charset) != 0)
+		    {
+		      if (!NILP (str->iso2022.charset[1]) &&
+			  (!CODING_SYSTEM_ISO2022_SEVEN (codesys) ||
+			   CODING_SYSTEM_ISO2022_LOCK_SHIFT (codesys)))
+			reg = 1;
+		      else if (!NILP (str->iso2022.charset[2]))
+			reg = 2;
+		      else if (!NILP (str->iso2022.charset[3]))
+			reg = 3;
+		      else
+			reg = 0;
+		    }
+		  else
+		    reg = 0;
+		}
+	      
+	      iso2022_designate (charset, reg, str, dst);
+	      
+	      /* Now invoke that register. */
+	      switch (reg)
+		{
+		case 0:
+		  ensure_normal_shift (str, dst);
+		  half = 0;
+		  break;
+		  
+		case 1:
+		  if (CODING_SYSTEM_ISO2022_SEVEN (codesys))
+		    {
+		      ensure_shift_out (str, dst);
+		      half = 0;
+		    }
+		  else
+		    half = 1;
+		  break;
+		  
+		case 2:
+		  if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
+		    {
+		      Dynarr_add (dst, ISO_CODE_ESC);
+		      Dynarr_add (dst, 'N');
+		      half = 0;
+		    }
+		  else
+		    {
+		      Dynarr_add (dst, ISO_CODE_SS2);
+		      half = 1;
+		    }
+		  break;
+		  
+		case 3:
+		  if (CODING_SYSTEM_ISO2022_SEVEN (str->codesys))
+		    {
+		      Dynarr_add (dst, ISO_CODE_ESC);
+		      Dynarr_add (dst, 'O');
+		      half = 0;
+		    }
+		  else
+		    {
+		      Dynarr_add (dst, ISO_CODE_SS3);
+		      half = 1;
+		    }
+		  break;
+		  
+		default:
+		  abort ();
+		}
+	      
+	      charmask = (half == 0 ? 0x00 : 0x80);
+	      
+	      switch (XCHARSET_DIMENSION (charset))
+		{
+		case 1:
+		  Dynarr_add (dst, byte1 | charmask);
+		  break;
+		case 2:
+		  Dynarr_add (dst, byte1 | charmask);
+		  Dynarr_add (dst, byte2 | charmask);
+		  break;
+		default:
+		  abort ();
+		}
+	    }
+	  ch =0;
+	  break;
+	default:
+	  ch = ( ch << 6 ) | ( c & 0x3f );
+	  char_boundary--;
+	}
+    }
+#else /* not UTF2000 */
+
   while (n--)
     {
       c = *src++;
@@ -5119,6 +5655,7 @@ encode_coding_iso2022 (Lstream *encoding, CONST unsigned char *src,
 	    }
 	}
     }
+#endif /* not UTF2000 */
 
 #ifdef ENABLE_COMPOSITE_CHARS
   if (in_composite)
@@ -5132,7 +5669,11 @@ encode_coding_iso2022 (Lstream *encoding, CONST unsigned char *src,
     }
 #endif /* ENABLE_COMPOSITE_CHARS */
 
+#ifdef UTF2000
+  if ( (char_boundary == 0) && flags & CODING_STATE_END)
+#else
   if (char_boundary && flags & CODING_STATE_END)
+#endif
     {
       restore_left_to_right_direction (codesys, dst, &flags, 0);
       ensure_normal_shift (str, dst);
@@ -5195,10 +5736,68 @@ encode_coding_no_conversion (Lstream *encoding, CONST unsigned char *src,
   unsigned int flags  = str->flags;
   unsigned int ch     = str->ch;
   eol_type_t eol_type = CODING_SYSTEM_EOL_TYPE (str->codesys);
+#ifdef UTF2000
+  unsigned char char_boundary = str->iso2022.current_char_boundary;
+#endif
 
   while (n--)
     {
-      c = *src++;
+      c = *src++;	  
+#ifdef UTF2000
+      switch (char_boundary)
+	{
+	case 0:
+	  if ( c >= 0xfc )
+	    {
+	      ch = c & 0x01;
+	      char_boundary = 5;
+	    }
+	  else if ( c >= 0xf8 )
+	    {
+	      ch = c & 0x03;
+	      char_boundary = 4;
+	    }
+	  else if ( c >= 0xf0 )
+	    {
+	      ch = c & 0x07;
+	      char_boundary = 3;
+	    }
+	  else if ( c >= 0xe0 )
+	    {
+	      ch = c & 0x0f;
+	      char_boundary = 2;
+	    }
+	  else if ( c >= 0xc0 )
+	    {
+	      ch = c & 0x1f;
+	      char_boundary = 1;
+	    }
+	  else
+	    {
+	      ch = 0;
+
+	      if (c == '\n')
+		{
+		  if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
+		    Dynarr_add (dst, '\r');
+		  if (eol_type != EOL_CR)
+		    Dynarr_add (dst, c);
+		}
+	      else
+		Dynarr_add (dst, c);
+	      char_boundary = 0;
+	    }
+	  break;
+	case 1:
+	  ch = ( ch << 6 ) | ( c & 0x3f );
+	  Dynarr_add (dst, ch & 0xff);
+	  char_boundary = 0;
+	  break;
+	default:
+	  ch = ( ch << 6 ) | ( c & 0x3f );
+	  char_boundary--;
+	}
+#else /* not UTF2000 */
       if (c == '\n')
 	{
 	  if (eol_type != EOL_LF && eol_type != EOL_AUTODETECT)
@@ -5234,10 +5833,14 @@ encode_coding_no_conversion (Lstream *encoding, CONST unsigned char *src,
 	     untranslatable character, so ignore it */
 	  ch = 0;
 	}
+#endif /* not UTF2000 */
     }
 
   str->flags = flags;
   str->ch    = ch;
+#ifdef UTF2000
+  str->iso2022.current_char_boundary = char_boundary;
+#endif
 }
 
 
@@ -5291,12 +5894,17 @@ convert_to_external_format (CONST Bufbyte *ptr,
 
       for (; ptr < end;)
         {
+#ifdef UTF2000
+          Bufbyte c =
+	    (*ptr < 0xc0) ? *ptr :
+	    ((*ptr & 0x1f) << 6) | (*(ptr+1) & 0x3f);
+#else
           Bufbyte c =
             (BYTE_ASCII_P (*ptr))		   ? *ptr :
             (*ptr == LEADING_BYTE_CONTROL_1)	   ? (*(ptr+1) - 0x20) :
             (*ptr == LEADING_BYTE_LATIN_ISO8859_1) ? (*(ptr+1)) :
             '~';
-
+#endif
           Dynarr_add (conversion_out_dynarr, (Extbyte) c);
           INC_CHARPTR (ptr);
         }
@@ -5655,12 +6263,24 @@ complex_vars_of_file_coding (void)
      list4 (Qeol_type, Qlf,
 	    Qmnemonic, build_string ("Binary")));
 
+#ifdef UTF2000
+  Fmake_coding_system
+    (Qutf8, Qutf8,
+     build_string ("Coding-system of ISO/IEC 10646 UTF-8."),
+     list2 (Qmnemonic, build_string ("UTF8")));
+#endif
+
   Fdefine_coding_system_alias (Qno_conversion, Qraw_text);
 
   /* Need this for bootstrapping */
   coding_category_system[CODING_CATEGORY_NO_CONVERSION] =
     Fget_coding_system (Qraw_text);
 
+#ifdef UTF2000
+  coding_category_system[CODING_CATEGORY_UTF8]
+   = Fget_coding_system (Qutf8);
+#endif
+
 #ifdef MULE
   {
     unsigned int i;