X-Git-Url: http://git.chise.org/gitweb/?a=blobdiff_plain;ds=sidebyside;f=src%2Ftext-coding.c;h=8015b7cae2d67c0521aa008207d02d096f7a1708;hb=99acb31ad5095183b2c9c7c45093edf5cd378890;hp=b28c2d73260341713cc872b6ae93ad802c0559d7;hpb=6f39425d9242fa7af570309aef32cae510bfad44;p=chise%2Fxemacs-chise.git.1

diff --git a/src/text-coding.c b/src/text-coding.c
index b28c2d7..8015b7c 100644
--- a/src/text-coding.c
+++ b/src/text-coding.c
@@ -1,7 +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 MORIOKA Tomohiko
+   Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2008, 2011
+     MORIOKA Tomohiko
 
 This file is part of XEmacs.
 
@@ -110,6 +111,7 @@ Lisp_Object Qdisable_composition;
 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;
 
@@ -2320,6 +2322,7 @@ struct decoding_stream
   unsigned char counter;
 #endif
 #ifdef UTF2000
+  char bom_flag;
   unsigned char er_counter;
   unsigned char er_buf[ER_BUF_SIZE];
 
@@ -2464,6 +2467,7 @@ reset_decoding_stream (struct decoding_stream *str)
   str->counter = 0;
 #endif /* MULE */
 #ifdef UTF2000
+  str->bom_flag = 0;
   str->er_counter = 0;
   str->combined_char_count = 0;
   str->combining_table = Qnil;
@@ -3411,11 +3415,22 @@ 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;
 
@@ -3474,7 +3489,7 @@ char_encode_as_entity_reference (Emchar ch, char* buf)
 /************************************************************************/
 /*                          character composition                       */
 /************************************************************************/
-extern Lisp_Object Qcomposition;
+extern Lisp_Object Qcomposition, Qrep_decomposition;
 
 INLINE_HEADER void
 COMPOSE_FLUSH_CHARS (struct decoding_stream *str, unsigned_char_dynarr* dst);
@@ -4504,6 +4519,7 @@ decode_coding_utf8 (Lstream *decoding, const Extbyte *src,
   eol_type_t eol_type	= str->eol_type;
   unsigned char counter	= str->counter;
 #ifdef UTF2000
+  int bom_flag = str->bom_flag;
   Lisp_Object ccs
     = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (DECODING_STREAM_DATA
 					     (decoding)->codesys, 0);
@@ -4519,11 +4535,20 @@ decode_coding_utf8 (Lstream *decoding, const Extbyte *src,
 	      COMPOSE_FLUSH_CHARS (str, dst);
 	      decode_flush_er_chars (str, dst);
 	      DECODE_HANDLE_EOL_TYPE (eol_type, c, flags, dst);
+
+	      if ( bom_flag == 0 )
+		bom_flag = -1;
+
 	      DECODE_ADD_UCS_CHAR (c, dst);
 	    }
 	  else if ( c < 0xC0 )
-	    /* decode_add_er_char (str, c, dst); */
-	    COMPOSE_ADD_CHAR (str, c, dst);
+	    {	      
+	      if ( bom_flag == 0 )
+		bom_flag = -1;
+
+	      /* decode_add_er_char (str, c, dst); */
+	      COMPOSE_ADD_CHAR (str, c, dst);
+	    }
 	  else
 	    {
 	      /* decode_flush_er_chars (str, dst); */
@@ -4561,6 +4586,17 @@ decode_coding_utf8 (Lstream *decoding, const Extbyte *src,
 	    {
 	      Emchar char_id;
 
+	      if ( bom_flag == 0 )
+		{
+		  if ( cpos == 0xFEFF )
+		    {
+		      bom_flag = 1;
+		      goto decoded;
+		    }
+		  else
+		    bom_flag = -1;
+		}
+
 	      if (!NILP (ccs))
 		{
 		  char_id = decode_defined_char (ccs, cpos, 0);
@@ -4571,6 +4607,7 @@ decode_coding_utf8 (Lstream *decoding, const Extbyte *src,
 	      else
 		char_id = cpos;
 	      COMPOSE_ADD_CHAR (str, char_id, dst);
+	    decoded:
 	      cpos = 0;
 	      counter = 0;
 	    }
@@ -4603,6 +4640,9 @@ decode_coding_utf8 (Lstream *decoding, const Extbyte *src,
   str->flags	= flags;
   str->cpos	= cpos;
   str->counter	= counter;
+#ifdef UTF2000
+  str->bom_flag = bom_flag;
+#endif
 }
 
 void
@@ -4630,10 +4670,43 @@ char_encode_utf8 (struct encoding_stream *str, Emchar ch,
 
       if ( (code_point < 0) || (code_point > 0xEFFFF) )
 	{
-	  Lisp_Object map
-	    = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, 1);
-	  Lisp_Object ret;
+	  Lisp_Object rest = Vdecomposition_feature_list;
+	  Lisp_Object decomp_f;
+	  Lisp_Object seq = Qnil;
+	  Lisp_Object map, ret;
+	  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);
+	    }
+
+	  if ( CONSP (seq) )
+	    {
+	      Lisp_Object base = Fcar (seq);
+
+	      seq = Fcdr (seq);
+	      if ( CHARP (base) && CONSP (seq) )
+		{
+		  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;
+		    }
+		}
+	    }
+
+	  map = CODING_SYSTEM_ISO2022_INITIAL_CHARSET (str->codesys, 1);
 	  if ( !NILP (map)
 	       && INTP (ret = Fchar_feature (make_char (ch),
 					     map, Qnil,
@@ -6507,6 +6580,14 @@ 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