(coded-charset-entity-reference-alist): Add setting for
[chise/xemacs-chise.git.1] / src / text-coding.c
index 09afa27..7071934 100644 (file)
@@ -1,8 +1,8 @@
 /* Code conversion functions.
    Copyright (C) 1991, 1995 Free Software Foundation, Inc.
    Copyright (C) 1995 Sun Microsystems, Inc.
-   Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2008, 2011
-     MORIOKA Tomohiko
+   Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2008, 2011,
+     2012, 2013, 2016 MORIOKA Tomohiko
 
 This file is part of XEmacs.
 
@@ -38,6 +38,9 @@ Boston, MA 02111-1307, USA.  */
 #include "mule-ccl.h"
 #include "chartab.h"
 #endif
+#ifdef USE_CONCORD_OBJECT_SYSTEM_TO_COMPOSE
+#include <cos.h>
+#endif
 #include "file-coding.h"
 
 Lisp_Object Qcoding_system_error;
@@ -107,10 +110,11 @@ Lisp_Object Qshort, Qno_ascii_eol, Qno_ascii_cntl, Qseven, Qlock_shift;
 #endif
 #ifdef UTF2000
 Lisp_Object Qutf_8_mcs;
-Lisp_Object Qdisable_composition;
+Lisp_Object Qdisable_composition, Qenable_decomposition;
 Lisp_Object Qccs_priority_list;
 Lisp_Object Quse_entity_reference;
 Lisp_Object Qd, Qx, QX;
+Lisp_Object Vdecomposition_feature_list;
 #endif
 Lisp_Object Qencode, Qdecode;
 
@@ -848,8 +852,10 @@ character set.  Recognized properties are:
                `coding-system-property' will return nil.)
 
 'disable-composition
-     If non-nil, composition/decomposition for combining characters
-     are disabled.
+     If non-nil, composition for combining characters is disabled.
+
+'enable-decomposition
+     If non-nil, decomposition for combining characters is enabled.
 
 'use-entity-reference
      If non-nil, SGML style entity-reference is used for non-system-characters.
@@ -1006,6 +1012,8 @@ if TYPE is 'ccl:
 #ifdef UTF2000
        else if (EQ (key, Qdisable_composition))
          CODING_SYSTEM_DISABLE_COMPOSITION (codesys) = !NILP (value);
+       else if (EQ (key, Qenable_decomposition))
+         CODING_SYSTEM_ENABLE_DECOMPOSITION (codesys) = !NILP (value);
        else if (EQ (key, Quse_entity_reference))
          CODING_SYSTEM_USE_ENTITY_REFERENCE (codesys) = !NILP (value);
 #endif
@@ -1071,6 +1079,8 @@ if TYPE is 'ccl:
              CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 1) = value;
            else if (EQ (key, Qcharset_g2))
              CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 2) = value;
+           else if (EQ (key, Qcharset_g3))
+             CODING_SYSTEM_ISO2022_INITIAL_CHARSET (codesys, 3) = value;
            else
              signal_simple_error ("Unrecognized property", key);
          }
@@ -1490,6 +1500,8 @@ Return the PROP property of CODING-SYSTEM.
 #ifdef UTF2000
   else if (EQ (prop, Qdisable_composition))
     return XCODING_SYSTEM_DISABLE_COMPOSITION (coding_system) ? Qt : Qnil;
+  else if (EQ (prop, Qenable_decomposition))
+    return XCODING_SYSTEM_ENABLE_DECOMPOSITION (coding_system) ? Qt : Qnil;
   else if (EQ (prop, Quse_entity_reference))
     return XCODING_SYSTEM_USE_ENTITY_REFERENCE (coding_system) ? Qt : Qnil;
   else if (EQ (prop, Qccs_priority_list))
@@ -2196,8 +2208,8 @@ do {                                              \
     }                                          \
 } while (0)
 
-INLINE_HEADER void DECODE_ADD_UCS_CHAR(Emchar c, unsigned_char_dynarr* dst);
-INLINE_HEADER void
+static void DECODE_ADD_UCS_CHAR(Emchar c, unsigned_char_dynarr* dst);
+static void
 DECODE_ADD_UCS_CHAR(Emchar c, unsigned_char_dynarr* dst)
 {
   if ( c <= 0x7f )
@@ -2327,7 +2339,11 @@ struct decoding_stream
 
   unsigned combined_char_count;
   Emchar combined_chars[16];
+#ifdef USE_CONCORD_OBJECT_SYSTEM_TO_COMPOSE
+  COS_object combining_table;
+#else
   Lisp_Object combining_table;
+#endif /* USE_CONCORD_OBJECT_SYSTEM_TO_COMPOSE */
 #endif
   struct detection_state decst;
 };
@@ -2469,7 +2485,11 @@ reset_decoding_stream (struct decoding_stream *str)
   str->bom_flag = 0;
   str->er_counter = 0;
   str->combined_char_count = 0;
+#ifdef USE_CONCORD_OBJECT_SYSTEM_TO_COMPOSE
+  str->combining_table = COS_NIL;
+#else
   str->combining_table = Qnil;
+#endif /* USE_CONCORD_OBJECT_SYSTEM_TO_COMPOSE */
 #endif
   if (CODING_SYSTEM_TYPE (str->codesys) == CODESYS_AUTODETECT
       || CODING_SYSTEM_EOL_TYPE (str->codesys) == EOL_AUTODETECT)
@@ -3414,11 +3434,22 @@ char_encode_as_entity_reference (Emchar ch, char* buf)
        char_type = Qnil;
       if (!NILP (ccs = Ffind_charset (ccs)))
        {
-         int code_point = charset_code_point (ccs, ch, 0);
+         int code_point
+           = charset_code_point (ccs, ch,
+                                 NILP (char_type) ?
+                                 CHAR_ALL : CHAR_ISOLATED_ONLY );
 
          if ( (code_point >= 0)
-              && (NILP (char_type)
-                  || DECODE_CHAR (ccs, code_point, 0) != ch) )
+              && ( NILP (char_type)
+#if 1
+                   || ( charset_code_point (ccs, ch, CHAR_DEFINED_ONLY)
+                        == -1 )
+#endif
+#if 0
+                   || ( DECODE_CHAR (ccs, code_point, 0) != ch )
+#endif
+                   )
+              )
            {
              Lisp_Object ret;
 
@@ -3489,9 +3520,27 @@ COMPOSE_FLUSH_CHARS (struct decoding_stream *str, unsigned_char_dynarr* dst)
   for (i = 0; i < str->combined_char_count; i++)
     decode_add_er_char (str, str->combined_chars[i], dst);
   str->combined_char_count = 0;
+#ifdef USE_CONCORD_OBJECT_SYSTEM_TO_COMPOSE
+  str->combining_table = COS_NIL;
+#else
   str->combining_table = Qnil;
+#endif /* USE_CONCORD_OBJECT_SYSTEM_TO_COMPOSE */
 }
 
+extern CONCORD_DS concord_current_env;
+
+#if 0
+static int
+concord_setup_env_maybe ()
+{
+  if (concord_current_env == NULL)
+    {
+      concord_open_env ("/usr/local/share/chise/1.0/db/");
+    }
+  return 0;
+}
+#endif
+
 void COMPOSE_ADD_CHAR (struct decoding_stream *str, Emchar character,
                       unsigned_char_dynarr* dst);
 void
@@ -3500,6 +3549,72 @@ COMPOSE_ADD_CHAR (struct decoding_stream *str,
 {
   if (CODING_SYSTEM_DISABLE_COMPOSITION (str->codesys))
     decode_add_er_char (str, character, dst);
+#ifdef USE_CONCORD_OBJECT_SYSTEM_TO_COMPOSE
+  else if (!cos_cons_p (str->combining_table))
+    {
+      COS_object ret;
+
+      /* concord_setup_env_maybe (); */
+      open_chise_data_source_maybe ();
+      ret = concord_object_get_attribute (cos_make_char (character),
+                                             COS_COMPOSITION);
+
+      if (!cos_cons_p (ret))
+       decode_add_er_char (str, character, dst);
+      else
+       {
+         str->combined_chars[0] = character;
+         str->combined_char_count = 1;
+         str->combining_table = ret;
+       }
+    }
+  else
+    {
+      COS_object ret
+       = cos_cdr (cos_assoc (cos_make_char (character),
+                             str->combining_table));
+
+      if (cos_char_p (ret))
+       {
+         Emchar char2 = cos_char_id (ret);
+         COS_object ret2;
+
+         /* concord_setup_env_maybe (); */
+         open_chise_data_source_maybe ();
+         ret2 = concord_object_get_attribute (ret, COS_COMPOSITION);
+
+         if (!cos_cons_p (ret2))
+           {
+             decode_add_er_char (str, char2, dst);
+             str->combined_char_count = 0;
+             str->combining_table = COS_NIL;
+           }
+         else
+           {
+             str->combined_chars[0] = char2;
+             str->combined_char_count = 1;
+             str->combining_table = ret2;
+           }
+       }
+      else
+       {
+         /* concord_setup_env_maybe (); */
+         open_chise_data_source_maybe ();
+         ret = concord_object_get_attribute (cos_make_char (character),
+                                                 COS_COMPOSITION);
+
+         COMPOSE_FLUSH_CHARS (str, dst);
+         if (!cos_cons_p (ret))
+           decode_add_er_char (str, character, dst);
+         else
+           {
+             str->combined_chars[0] = character;
+             str->combined_char_count = 1;
+             str->combining_table = ret;
+           }
+       }
+    }
+#else
   else if (!CONSP (str->combining_table))
     {
       Lisp_Object ret
@@ -3555,6 +3670,7 @@ COMPOSE_ADD_CHAR (struct decoding_stream *str,
            }
        }
     }
+#endif /* USE_CONCORD_OBJECT_SYSTEM_TO_COMPOSE */
 }
 #else /* not UTF2000 */
 #define COMPOSE_FLUSH_CHARS(str, dst)
@@ -4658,25 +4774,42 @@ char_encode_utf8 (struct encoding_stream *str, Emchar ch,
 
       if ( (code_point < 0) || (code_point > 0xEFFFF) )
        {
-         Lisp_Object seq = Fchar_feature (make_char (ch),
-                                          Qrep_decomposition, Qnil,
-                                          Qnil, Qnil);
          Lisp_Object map, ret;
 
-         if ( CONSP (seq) )
+         if ( CODING_SYSTEM_ENABLE_DECOMPOSITION (str->codesys) )
            {
-             Lisp_Object base = Fcar (seq);
+             Lisp_Object rest = Vdecomposition_feature_list;
+             Lisp_Object decomp_f;
+             Lisp_Object seq = Qnil;
+             struct gcpro gcpro1;
+
+             while ( CONSP (rest) )
+               {
+                 decomp_f = XCAR (rest);
+                 GCPRO1 (rest);
+                 seq = Fchar_feature (make_char (ch), decomp_f, Qnil,
+                                      Qnil, Qnil);
+                 UNGCPRO;
+                 if ( !NILP (seq) )
+                   break;
+                 rest = XCDR (rest);
+               }
 
-             seq = Fcdr (seq);
-             if ( CHARP (base) && CONSP (seq) )
+             if ( CONSP (seq) )
                {
-                 Lisp_Object comb = Fcar (seq);
+                 Lisp_Object base = Fcar (seq);
 
-                 if ( CHARP (comb) )
+                 seq = Fcdr (seq);
+                 if ( CHARP (base) && CONSP (seq) )
                    {
-                     char_encode_utf8 (str, XCHAR (base), dst, flags);
-                     char_encode_utf8 (str, XCHAR (comb), dst, flags);
-                     return;
+                     Lisp_Object comb = Fcar (seq);
+
+                     if ( CHARP (comb) )
+                       {
+                         char_encode_utf8 (str, XCHAR (base), dst, flags);
+                         char_encode_utf8 (str, XCHAR (comb), dst, flags);
+                         return;
+                       }
                    }
                }
            }
@@ -4694,6 +4827,13 @@ char_encode_utf8 (struct encoding_stream *str, Emchar ch,
                                                  map, Qnil,
                                                  Qnil, Qnil)) )
            code_point = XINT (ret);
+         else if ( !NILP (map =
+                          CODING_SYSTEM_ISO2022_INITIAL_CHARSET
+                          (str->codesys, 3))
+                   && INTP (ret = Fchar_feature (make_char (ch),
+                                                 map, Qnil,
+                                                 Qnil, Qnil)) )
+           code_point = XINT (ret);
          else if (CODING_SYSTEM_USE_ENTITY_REFERENCE (str->codesys))
            {
              char buf[18];
@@ -6426,6 +6566,7 @@ syms_of_file_coding (void)
 #ifdef UTF2000
   defsymbol (&Qutf_8_mcs, "utf-8-mcs");
   defsymbol (&Qdisable_composition, "disable-composition");
+  defsymbol (&Qenable_decomposition, "enable-decomposition");
   defsymbol (&Qccs_priority_list, "ccs-priority-list");
   defsymbol (&Quse_entity_reference, "use-entity-reference");
   defsymbol (&Qd, "d");
@@ -6555,6 +6696,14 @@ and behaviors of various editing commands.
 Setting this to nil does not do anything.
 */ );
   enable_multibyte_characters = 1;
+
+#ifdef UTF2000
+  DEFVAR_LISP ("decomposition-feature-list",
+              &Vdecomposition_feature_list /*
+List of `=decomposition@FOO' feature to encode characters as IVS.
+*/ );
+  Vdecomposition_feature_list = Qnil;
+#endif
 }
 
 void
@@ -6608,6 +6757,7 @@ complex_vars_of_file_coding (void)
   DEFINE_CODESYS_PROP (CODESYS_PROP_CCL,     Qdecode);
 #ifdef UTF2000
   DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK,  Qdisable_composition);
+  DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK,  Qenable_decomposition);
   DEFINE_CODESYS_PROP (CODESYS_PROP_ALL_OK,  Quse_entity_reference);
 #endif
 #endif /* MULE */