- Use `=>ucs' instead of `->ucs'.

[chise/xemacs-chise.git] / src / mule-charset.c

diff --git a/src/mule-charset.c b/src/mule-charset.c
index ee90d44..8453ace 100644 (file)
--- a/src/mule-charset.c
+++ b/src/mule-charset.c
@@ -385,8 +385,8 @@ Lisp_Object Qreverse_direction_charset;
 Lisp_Object Qleading_byte;
 Lisp_Object Qshort_name, Qlong_name;
 #ifdef UTF2000
 Lisp_Object Qleading_byte;
 Lisp_Object Qshort_name, Qlong_name;
 #ifdef UTF2000

-Lisp_Object Qmin_code, Qmax_code;
-Lisp_Object Qmother, Qconversion, Q94x60;
+Lisp_Object Qmin_code, Qmax_code, Qcode_offset;
+Lisp_Object Qmother, Qconversion, Q94x60, Q94x94x60;

 #endif
 
 Lisp_Object Qascii,
 #endif
 
 Lisp_Object Qascii,

@@ -986,9 +986,51 @@ get_unallocated_leading_byte (int dimension)
 Emchar
 decode_builtin_char (Lisp_Object charset, int code_point)
 {
 Emchar
 decode_builtin_char (Lisp_Object charset, int code_point)
 {

+  Lisp_Object mother = XCHARSET_MOTHER (charset);

   int final;
 
   int final;
 

-  if (EQ (charset, Vcharset_chinese_big5))
+  if ( CHARSETP (mother) )
+    {
+      int code = code_point;
+
+      if ( XCHARSET_CONVERSION (charset) == CONVERSION_94x60 )
+       {
+         int row = code_point >> 8;
+         int cell = code_point & 255;    
+
+         if (row < 16 + 32)
+           return -1;
+         else if (row < 16 + 32 + 30)
+           code = (row - (16 + 32)) * 94 + cell - 33;
+         else if (row < 18 + 32 + 30)
+           return -1;
+         else if (row < 18 + 32 + 60)
+           code = (row - (18 + 32)) * 94 + cell - 33;
+       }
+      else if ( XCHARSET_CONVERSION (charset) == CONVERSION_94x94x60 )
+       {
+         int plane = code_point >> 16;
+         int row = (code_point >> 8) & 255;
+         int cell = code_point & 255;    
+
+         if (row < 16 + 32)
+           return -1;
+         else if (row < 16 + 32 + 30)
+           code
+             = (plane - 33) * 94 * 60
+             + (row - (16 + 32)) * 94
+             + cell - 33;
+         else if (row < 18 + 32 + 30)
+           return -1;
+         else if (row < 18 + 32 + 60)
+           code
+             = (plane - 33) * 94 * 60
+             + (row - (18 + 32)) * 94
+             + cell - 33;
+       }
+      return DECODE_CHAR (mother, code + XCHARSET_CODE_OFFSET(charset));
+    }
+  else if (EQ (charset, Vcharset_chinese_big5))

     {
       int c1 = code_point >> 8;
       int c2 = code_point & 0xFF;
     {
       int c1 = code_point >> 8;
       int c2 = code_point & 0xFF;

@@ -1084,115 +1126,130 @@ charset_code_point (Lisp_Object charset, Emchar ch)
   else
     {
       Lisp_Object mother = XCHARSET_MOTHER (charset);
   else
     {
       Lisp_Object mother = XCHARSET_MOTHER (charset);

+      int min = XCHARSET_MIN_CODE (charset);
+      int max = XCHARSET_MAX_CODE (charset);
+      int code;

 
       if ( CHARSETP (mother) )
 
       if ( CHARSETP (mother) )

+       code = charset_code_point (mother, ch);
+      else
+       code = ch;
+      if ( (min <= code) && (code <= max) )

        {
        {

-         int min = XCHARSET_MIN_CODE (charset);
-         int max = XCHARSET_MAX_CODE (charset);
-         int code = charset_code_point (mother, ch);
+         int d = code - XCHARSET_CODE_OFFSET (charset);

 
 

-         if ( (min <= code) && (code <= max) )
+         if ( XCHARSET_CONVERSION (charset) == CONVERSION_94x60 )

            {
            {

-             if ( XCHARSET_CONVERSION (charset) == CONVERSION_94x60 )
-               {
-                 int m = code - min;
-                 int row = m  / 94;
-                 int cell = m % 94 + 33;
-
-                 if (row < 30)
-                   row += 16 + 32;
-                 else
-                   row += 18 + 32;
-                 return (row << 8) | cell;
-               }
+             int row  = d / 94;
+             int cell = d % 94 + 33;
+
+             if (row < 30)
+               row += 16 + 32;

              else
              else

-               return code - XCHARSET_CODE_OFFSET (charset);
+               row += 18 + 32;
+             return (row << 8) | cell;

            }
            }

-       }
-    }
-  return range_charset_code_point (charset, ch);
-}
-
-int
-range_charset_code_point (Lisp_Object charset, Emchar ch)
-{
-  int d;
+         else if ( XCHARSET_CONVERSION (charset) == CONVERSION_94x94x60 )
+           {
+             int plane =  d / (94 * 60) + 33;
+             int row   = (d % (94 * 60)) / 94;
+             int cell  =  d %  94 + 33;

 
 

-  if ((XCHARSET_MIN_CODE (charset) <= ch)
-      && (ch <= XCHARSET_MAX_CODE (charset)))
-    {
-      d = ch - XCHARSET_CODE_OFFSET (charset);
-
-      if (XCHARSET_CHARS (charset) == 256)
-       return d;
-      else if (XCHARSET_DIMENSION (charset) == 1)
-       return d + XCHARSET_BYTE_OFFSET (charset);
-      else if (XCHARSET_DIMENSION (charset) == 2)
-       return
-         ((d / XCHARSET_CHARS (charset)
-           + XCHARSET_BYTE_OFFSET (charset)) << 8)
-         | (d % XCHARSET_CHARS (charset) + XCHARSET_BYTE_OFFSET (charset));
-      else if (XCHARSET_DIMENSION (charset) == 3)
-       return
-         ((d / (XCHARSET_CHARS (charset) * XCHARSET_CHARS (charset))
-           + XCHARSET_BYTE_OFFSET (charset)) << 16)
-         | ((d / XCHARSET_CHARS (charset)
-             % XCHARSET_CHARS (charset)
-             + XCHARSET_BYTE_OFFSET (charset)) << 8)
-         | (d % XCHARSET_CHARS (charset) + XCHARSET_BYTE_OFFSET (charset));
-      else /* if (XCHARSET_DIMENSION (charset) == 4) */
-       return
-         ((d / (XCHARSET_CHARS (charset)
-                * XCHARSET_CHARS (charset) * XCHARSET_CHARS (charset))
-           + XCHARSET_BYTE_OFFSET (charset)) << 24)
-         | ((d / (XCHARSET_CHARS (charset) * XCHARSET_CHARS (charset))
-             % XCHARSET_CHARS (charset)
-             + XCHARSET_BYTE_OFFSET (charset)) << 16)
-         | ((d / XCHARSET_CHARS (charset) % XCHARSET_CHARS (charset)
-             + XCHARSET_BYTE_OFFSET (charset)) << 8)
-         | (d % XCHARSET_CHARS (charset) + XCHARSET_BYTE_OFFSET (charset));
-    }
-  else if (XCHARSET_CODE_OFFSET (charset) == 0)
-    {
-      if (XCHARSET_DIMENSION (charset) == 1)
-       {
-         if (XCHARSET_CHARS (charset) == 94)
+             if (row < 30)
+               row += 16 + 32;
+             else
+               row += 18 + 32;
+             return (plane << 16) | (row << 8) | cell;
+           }
+         else if (XCHARSET_CHARS (charset) == 94)

            {
            {

-             if (((d = ch - (MIN_CHAR_94
-                             + (XCHARSET_FINAL (charset) - '0') * 94)) >= 0)
-                 && (d < 94))
+             if (XCHARSET_DIMENSION (charset) == 1)

                return d + 33;
                return d + 33;

+             else if (XCHARSET_DIMENSION (charset) == 2)
+               return ((d / 94 + 33) << 8) | (d % 94 + 33);
+             else if (XCHARSET_DIMENSION (charset) == 3)
+               return
+                 (   (d / (94 * 94) + 33) << 16)
+                 |  ((d / 94 % 94   + 33) <<  8)
+                 |   (d % 94        + 33);
+             else /* if (XCHARSET_DIMENSION (charset) == 4) */
+               return
+                 (  (d / (94 * 94 * 94) + 33) << 24)
+                 | ((d / (94 * 94) % 94 + 33) << 16)
+                 | ((d / 94 % 94        + 33) <<  8)
+                 |  (d % 94             + 33);

            }
          else if (XCHARSET_CHARS (charset) == 96)
            {
            }
          else if (XCHARSET_CHARS (charset) == 96)
            {

-             if (((d = ch - (MIN_CHAR_96
-                             + (XCHARSET_FINAL (charset) - '0') * 96)) >= 0)
-                 && (d < 96))
+             if (XCHARSET_DIMENSION (charset) == 1)

                return d + 32;
                return d + 32;

+             else if (XCHARSET_DIMENSION (charset) == 2)
+               return ((d / 96 + 32) << 8) | (d % 96 + 32);
+             else if (XCHARSET_DIMENSION (charset) == 3)
+               return
+                 (   (d / (96 * 96) + 32) << 16)
+                 |  ((d / 96 % 96   + 32) <<  8)
+                 |   (d % 96        + 32);
+             else /* if (XCHARSET_DIMENSION (charset) == 4) */
+               return
+                 (  (d / (96 * 96 * 96) + 32) << 24)
+                 | ((d / (96 * 96) % 96 + 32) << 16)
+                 | ((d / 96 % 96        + 32) <<  8)
+                 |  (d % 96             + 32);

            }
          else
            }
          else

-           return -1;
+           return code - XCHARSET_CODE_OFFSET (charset);

        }
        }

-      else if (XCHARSET_DIMENSION (charset) == 2)
+      else if ( (XCHARSET_CODE_OFFSET (charset) == 0) ||
+               (XCHARSET_CODE_OFFSET (charset)
+                == XCHARSET_MIN_CODE (charset)) )

        {
        {

-         if (XCHARSET_CHARS (charset) == 94)
+         int d;
+
+         if (XCHARSET_DIMENSION (charset) == 1)

            {
            {

-             if (((d = ch - (MIN_CHAR_94x94
-                             + (XCHARSET_FINAL (charset) - '0') * 94 * 94))
-                  >= 0)
-                 && (d < 94 * 94))
-               return (((d / 94) + 33) << 8) | (d % 94 + 33);
+             if (XCHARSET_CHARS (charset) == 94)
+               {
+                 if (((d = ch - (MIN_CHAR_94
+                                 + (XCHARSET_FINAL (charset) - '0') * 94))
+                      >= 0)
+                     && (d < 94))
+                   return d + 33;
+               }
+             else if (XCHARSET_CHARS (charset) == 96)
+               {
+                 if (((d = ch - (MIN_CHAR_96
+                                 + (XCHARSET_FINAL (charset) - '0') * 96))
+                      >= 0)
+                     && (d < 96))
+                   return d + 32;
+               }
+             else
+               return -1;

            }
            }

-         else if (XCHARSET_CHARS (charset) == 96)
+         else if (XCHARSET_DIMENSION (charset) == 2)

            {
            {

-             if (((d = ch - (MIN_CHAR_96x96
-                             + (XCHARSET_FINAL (charset) - '0') * 96 * 96))
-                  >= 0)
-                 && (d < 96 * 96))
-               return (((d / 96) + 32) << 8) | (d % 96 + 32);
+             if (XCHARSET_CHARS (charset) == 94)
+               {
+                 if (((d = ch - (MIN_CHAR_94x94
+                                 +
+                                 (XCHARSET_FINAL (charset) - '0') * 94 * 94))
+                      >= 0)
+                     && (d < 94 * 94))
+                   return (((d / 94) + 33) << 8) | (d % 94 + 33);
+               }
+             else if (XCHARSET_CHARS (charset) == 96)
+               {
+                 if (((d = ch - (MIN_CHAR_96x96
+                                 +
+                                 (XCHARSET_FINAL (charset) - '0') * 96 * 96))
+                      >= 0)
+                     && (d < 96 * 96))
+                   return (((d / 96) + 32) << 8) | (d % 96 + 32);
+               }
+             else
+               return -1;

            }
            }

-         else
-           return -1;

        }
     }
   return -1;
        }
     }
   return -1;

@@ -1436,6 +1493,7 @@ character set.  Recognized properties are:
                this character set.
 'dimension     Number of octets used to index a character in this charset.
                Either 1 or 2.  Defaults to 1.
                this character set.
 'dimension     Number of octets used to index a character in this charset.
                Either 1 or 2.  Defaults to 1.

+               If UTF-2000 feature is enabled, 3 or 4 are also available.

 'columns       Number of columns used to display a character in this charset.
                Only used in TTY mode. (Under X, the actual width of a
                character can be derived from the font used to display the
 'columns       Number of columns used to display a character in this charset.
                Only used in TTY mode. (Under X, the actual width of a
                character can be derived from the font used to display the

@@ -1444,6 +1502,7 @@ character set.  Recognized properties are:
 'chars         Number of characters in each dimension (94 or 96).
                Defaults to 94.  Note that if the dimension is 2, the
                character set thus described is 94x94 or 96x96.
 'chars         Number of characters in each dimension (94 or 96).
                Defaults to 94.  Note that if the dimension is 2, the
                character set thus described is 94x94 or 96x96.

+               If UTF-2000 feature is enabled, 128 or 256 are also available.

 'final         Final byte of ISO 2022 escape sequence.  Must be
                supplied.  Each combination of (DIMENSION, CHARS) defines a
                separate namespace for final bytes.  Note that ISO
 'final         Final byte of ISO 2022 escape sequence.  Must be
                supplied.  Each combination of (DIMENSION, CHARS) defines a
                separate namespace for final bytes.  Note that ISO

@@ -1468,6 +1527,13 @@ character set.  Recognized properties are:
                is passed the octets of the character, with the high
                bit cleared and set depending upon whether the value
                of the 'graphic property is 0 or 1.
                is passed the octets of the character, with the high
                bit cleared and set depending upon whether the value
                of the 'graphic property is 0 or 1.

+'mother                [UTF-2000 only] Base coded-charset.
+'code-min      [UTF-2000 only] Minimum code-point of a base coded-charset.
+'code-max      [UTF-2000 only] Maximum code-point of a base coded-charset.
+'code-offset   [UTF-2000 only] Offset for a code-point of a base
+               coded-charset.
+'conversion    [UTF-2000 only] Conversion for a code-point of a base
+               coded-charset (94x60 or 94x94x60).

 */
        (name, doc_string, props))
 {
 */
        (name, doc_string, props))
 {

@@ -1477,8 +1543,8 @@ character set.  Recognized properties are:
   Lisp_Object charset;
   Lisp_Object ccl_program = Qnil;
   Lisp_Object short_name = Qnil, long_name = Qnil;
   Lisp_Object charset;
   Lisp_Object ccl_program = Qnil;
   Lisp_Object short_name = Qnil, long_name = Qnil;

-  int min_code = 0, max_code = 0;

   Lisp_Object mother = Qnil;
   Lisp_Object mother = Qnil;

+  int min_code = 0, max_code = 0, code_offset = 0;

   int byte_offset = -1;
   int conversion = 0;
 
   int byte_offset = -1;
   int conversion = 0;
 

@@ -1578,35 +1644,37 @@ character set.  Recognized properties are:
          }
 
 #ifdef UTF2000
          }
 
 #ifdef UTF2000

+       else if (EQ (keyword, Qmother))
+         {
+           mother = Fget_charset (value);
+         }
+

        else if (EQ (keyword, Qmin_code))
          {
            CHECK_INT (value);
        else if (EQ (keyword, Qmin_code))
          {
            CHECK_INT (value);

-           min_code = XINT (value);
-           if (min_code < 0)
-             {
-               min_code = (~(-1 - min_code)) & 0x7FFFFFFF;
-             }
+           min_code = XUINT (value);

          }
 
        else if (EQ (keyword, Qmax_code))
          {
            CHECK_INT (value);
          }
 
        else if (EQ (keyword, Qmax_code))
          {
            CHECK_INT (value);

-           max_code = XINT (value);
-           if (max_code < 0)
-             {
-               max_code = (~(-1 - max_code)) & 0x7FFFFFFF;
-             }
+           max_code = XUINT (value);

          }
 
          }
 

-       else if (EQ (keyword, Qmother))
+       else if (EQ (keyword, Qcode_offset))

          {
          {

-           mother = Fget_charset (value);
+           CHECK_INT (value);
+           code_offset = XUINT (value);

          }
 
        else if (EQ (keyword, Qconversion))
          {
            if (EQ (value, Q94x60))
              conversion = CONVERSION_94x60;
          }
 
        else if (EQ (keyword, Qconversion))
          {
            if (EQ (value, Q94x60))
              conversion = CONVERSION_94x60;

+           else if (EQ (value, Q94x94x60))
+             conversion = CONVERSION_94x94x60;
+           else
+             signal_simple_error ("Unrecognized conversion", value);

          }
 
 #endif
          }
 
 #endif

@@ -1670,7 +1738,7 @@ character set.  Recognized properties are:
   charset = make_charset (id, name, chars, dimension, columns, graphic,
                          final, direction, short_name, long_name,
                          doc_string, registry,
   charset = make_charset (id, name, chars, dimension, columns, graphic,
                          final, direction, short_name, long_name,
                          doc_string, registry,

-                         Qnil, min_code, max_code, 0, byte_offset,
+                         Qnil, min_code, max_code, code_offset, byte_offset,

                          mother, conversion);
   if (!NILP (ccl_program))
     XCHARSET_CCL_PROGRAM (charset) = ccl_program;
                          mother, conversion);
   if (!NILP (ccl_program))
     XCHARSET_CCL_PROGRAM (charset) = ccl_program;

@@ -2392,11 +2460,13 @@ syms_of_mule_charset (void)
   defsymbol (&Qshort_name, "short-name");
   defsymbol (&Qlong_name, "long-name");
 #ifdef UTF2000
   defsymbol (&Qshort_name, "short-name");
   defsymbol (&Qlong_name, "long-name");
 #ifdef UTF2000

+  defsymbol (&Qmother, "mother");

   defsymbol (&Qmin_code, "min-code");
   defsymbol (&Qmax_code, "max-code");
   defsymbol (&Qmin_code, "min-code");
   defsymbol (&Qmax_code, "max-code");

-  defsymbol (&Qmother, "mother");
+  defsymbol (&Qcode_offset, "code-offset");

   defsymbol (&Qconversion, "conversion");
   defsymbol (&Q94x60, "94x60");
   defsymbol (&Qconversion, "conversion");
   defsymbol (&Q94x60, "94x60");

+  defsymbol (&Q94x94x60, "94x94x60");

 #endif
 
   defsymbol (&Ql2r, "l2r");
 #endif
 
   defsymbol (&Ql2r, "l2r");