/*======================================================================*\
|*		Editor mined						*|
|*		CJK character set <-> Unicode mapping tables		*|
\*======================================================================*/

#include "mined.h"
#include "charcode.h"
#include "termprop.h"


/*======================================================================*\
|*			Character properties				*|
\*======================================================================*/

FLAG
no_char (c)
  unsigned long c;
{
  return c == CHAR_UNKNOWN || c == CHAR_INVALID;
}

FLAG
no_unichar (u)
  unsigned long u;
{
  return u == CHAR_UNKNOWN || u == CHAR_INVALID;
}

/**
   Check if character is a control character in current encoding.
   (Should be more generic...)
 */
int
iscontrol (c)
  unsigned long c;
{
  if (mapped_text) {
	unsigned long u = lookup_encodedchar (c);
	return u == '\177' || (! no_unichar (u) && u < ' ');
  } else if (utf8_text) {
	if (unassigned_single_width) {
		if (rxvt_version > 0) {
			/* handle weird mapping of non-Unicode ranges */
			if (c < 0x80000000) {
				c &= 0x1FFFFF;
			}
		}
	}
	return c == '\177' || c < ' ';
  } else if (cjk_text) {
	return c == '\177' || c < ' ';
  } else {
	return c == '\177' || (c & '\177') < ' ';
  }
}

/**
   Return display indication for a control character.
 */
character
controlchar (c)
  character c;
{
  if (c == '\177') {
	return '?';
  } else {
	return c + '@';
  }
}


/**
   Return the isolated form of an ALEF character.
 */
unsigned long
isolated_alef (unichar)
  unsigned long unichar;
{
	if (unichar == 0x0622) {
		/* ALEF WITH MADDA ABOVE */
		return 0xFE81;
	} else if (unichar == 0x0623) {
		/* ALEF WITH HAMZA ABOVE */
		return 0xFE83;
	} else if (unichar == 0x0625) {
		/* ALEF WITH HAMZA BELOW */
		return 0xFE87;
	} else if (unichar == 0x0627) {
		/* ALEF */
		return 0xFE8D;
	} else {
		/* ? -> ALEF SYMBOL */
		return 0x2135;
	}
}


/**
   Return max value in current encoding.
 */
unsigned long
max_char_value ()
{
  if (cjk_text) switch (text_encoding_tag) {
	case 'G': return 0xFFFFFFFF;
	case 'C': return 0x8EFFFFFF;
	case 'J': return 0x8FFFFF;
	default: return 0xFFFF;
  } else if (utf8_text) {
	return 0x7FFFFFFF;
  } else {
	return 0xFF;
  }
}


/**
   Convert CJK character in current text encoding to byte sequence.
 */
int
cjkencode (cjkchar, buf)
  unsigned long cjkchar;
  character * buf;
{
  return cjkencode_char (False, cjkchar, buf);
}

static
int
multi_char (term, c)
  FLAG term;
  character c;
{
  if (term) {
	return (character) c >= 0x80 && (! cjk_term || term_encoding_tag != 'S' || (character) c < 0xA1 || (character) c > 0xDF);
  } else {
	return multichar (c);
  }
}

/**
   Convert CJK character in terminal or text encoding to byte sequence.
 */
int
cjkencode_char (term, cjkchar, buf)
  FLAG term;
  unsigned long cjkchar;
  character * buf;
{
  int len = 0;
  int i;
  char encoding_tag = term ? term_encoding_tag : text_encoding_tag;

  if (cjkchar >= 0x1000000) {
	i = (cjkchar >> 16) & 0xFF;
	if (encoding_tag == 'G' && cjkchar >= 0x80000000
	 && i >= '0' && i <= '9') {
		len = 4;
	} else if (encoding_tag == 'C' && (cjkchar >> 24) == 0x8E) {
		len = 4;
	}
  } else if (cjkchar >= 0x10000) {
	if (encoding_tag == 'J' && (cjkchar >> 16) == 0x8F) {
		len = 3;
	}
  } else if (cjkchar >= 0x8000 && (cjkchar & 0xFF) > 0 &&
		multi_char (term, (character) (cjkchar >> 8))) {
	len = 2;
  } else if (cjkchar < 0x100 && ! multi_char (term, cjkchar)) {
	len = 1;
  }

  for (i = len - 1; i >= 0; i --) {
	buf [i] = cjkchar & 0xFF;
	cjkchar = cjkchar >> 8;
	if (buf [i] == '\0') {
		len = 0;
	}
  }
  buf [len] = '\0';

  return len;
}

/**
   convert Unicode character to UTF-8
 */
int
utfencode (unichar, buf)
  unsigned long unichar;
  character * buf;
{
  int len;

  if (unichar < 0x80) {
	len = 1;
	* buf ++ = unichar;
  } else if (unichar < 0x800) {
	len = 2;
	* buf ++ = 0xC0 | (unichar >> 6);
	* buf ++ = 0x80 | (unichar & 0x3F);
  } else if (unichar < 0x10000) {
	len = 3;
	* buf ++ = 0xE0 | (unichar >> 12);
	* buf ++ = 0x80 | ((unichar >> 6) & 0x3F);
	* buf ++ = 0x80 | (unichar & 0x3F);
  } else if (unichar < 0x200000) {
	len = 4;
	* buf ++ = 0xF0 | (unichar >> 18);
	* buf ++ = 0x80 | ((unichar >> 12) & 0x3F);
	* buf ++ = 0x80 | ((unichar >> 6) & 0x3F);
	* buf ++ = 0x80 | (unichar & 0x3F);
  } else if (unichar < 0x4000000) {
	len = 5;
	* buf ++ = 0xF8 | (unichar >> 24);
	* buf ++ = 0x80 | ((unichar >> 18) & 0x3F);
	* buf ++ = 0x80 | ((unichar >> 12) & 0x3F);
	* buf ++ = 0x80 | ((unichar >> 6) & 0x3F);
	* buf ++ = 0x80 | (unichar & 0x3F);
  } else if (unichar < 0x80000000) {
	len = 6;
	* buf ++ = 0xFC | (unichar >> 30);
	* buf ++ = 0x80 | ((unichar >> 24) & 0x3F);
	* buf ++ = 0x80 | ((unichar >> 18) & 0x3F);
	* buf ++ = 0x80 | ((unichar >> 12) & 0x3F);
	* buf ++ = 0x80 | ((unichar >> 6) & 0x3F);
	* buf ++ = 0x80 | (unichar & 0x3F);
  } else {
	len = 0;
  }
  * buf = '\0';
  return len;
}

/**
   Convert character to byte sequence.
 */
char *
encode_char (c)
  unsigned long c;
{
  static char buf [7];
  if (utf8_text) {
	(void) utfencode (c, buf);
  } else if (cjk_text) {
	(void) cjkencode (c, buf);
  } else {
	buf [0] = c;
	buf [1] = '\0';
  }
  return buf;
}


/**
   Check if a CJK character is encoded in the defined code range of 
   the currently active text encoding.
 */
FLAG
valid_cjk (cjkchar, cjkbytes)
  unsigned long cjkchar;
  character * cjkbytes;
{
  return valid_cjkchar (False, cjkchar, cjkbytes);
}

/**
   Check if a CJK character is encoded in the defined code range of 
   the terminal or text encoding.
 */
FLAG
valid_cjkchar (term, cjkchar, cjkbytes)
  FLAG term;
  unsigned long cjkchar;
  character * cjkbytes;
{
  character cjkbuf [5];
  char encoding_tag = term ? term_encoding_tag : text_encoding_tag;

  if (cjkchar < 0x80) {
	return True;
  }

  if (! cjkbytes) {
	cjkbytes = cjkbuf;
	(void) cjkencode_char (term, cjkchar, cjkbytes);
  }

/*
         GB 18030 >> GBK >> GB 2312-80
                        GB18030	*?	81-FE	40-7E, 80-FE
                        		81-FE	30-39	81-FE	30-39
         Big5-HKSCS >> Big5
                        Big5	!	A1-FE	40-7E, A1-FE
                       (Big5+	!	89-FE	40-7E, A1-FE)
                        Big5+	!	88-FE	40-7E, A1-FE
                        HKSCS-2004!	87-FE	40-7E, A1-FE
         CNS
         EUC-TW -> CNS 11643-1992
                        EUC-TW	-	A1-FE	A1-FE
                        		8E	A1-A7	A1-FE	A1-FE
*/
  switch (encoding_tag) {
    case 'G':	if (cjkchar > 0xFFFF) {
			return cjkbytes [0] >= 0x81 && cjkbytes [0] <= 0xFE
				&& cjkbytes [1] >= '0' && cjkbytes [1] <= '9'
				&& cjkbytes [2] >= 0x81 && cjkbytes [2] <= 0xFE
				&& cjkbytes [3] >= '0' && cjkbytes [3] <= '9';
		} else {
			return cjkbytes [0] >= 0x81 && cjkbytes [0] <= 0xFE
				&& cjkbytes [1] >= 0x40 && cjkbytes [1] <= 0xFE
				&& cjkbytes [1] != 0x7F;
		}
    case 'B':	return cjkbytes [0] >= 0x87 && cjkbytes [0] <= 0xFE
				&& ((cjkbytes [1] >= 0x40 && cjkbytes [1] <= 0x7E)
				    ||
				    (cjkbytes [1] >= 0xA1 && cjkbytes [1] <= 0xFE)
				   )
				&& cjkbytes [2] == 0;
    case 'C':	return (cjkbytes [0] >= 0xA1 && cjkbytes [0] <= 0xFE
			&& cjkbytes [1] >= 0xA1 && cjkbytes [1] <= 0xFE
			&& cjkbytes [2] == 0)
			||
			(cjkbytes [0] == 0x8E
			&& cjkbytes [1] >= 0xA1 && cjkbytes [1] <= 0xAF
			&& cjkbytes [2] >= 0xA1 && cjkbytes [2] <= 0xFE
			&& cjkbytes [3] >= 0xA1 && cjkbytes [3] <= 0xFE);
/*
         EUC-JP -> JIS X 0208 + JIX X 0212
                        EUC-JP	!*	A1-FE	A1-FE
                        		8F	A1-FE	A1-FE
                        		8E	A1-DF
         Shift-JIS 0213 >> Shift-JIS
                        Shift-JIS	A1-DF
                        		81-9F, E0-EF	40-7E, 80-FC
*/
    case 'J':	return (cjkbytes [0] >= 0xA1 && cjkbytes [0] <= 0xFE
			&& cjkbytes [1] >= 0xA1 && cjkbytes [1] <= 0xFE
			&& cjkbytes [2] == 0
			)
			||
			(cjkbytes [0] == 0x8E
			&& cjkbytes [1] >= 0xA1 && cjkbytes [1] <= 0xDF
			&& cjkbytes [2] == 0)
			||
			(cjkbytes [0] == 0x8F
			&& cjkbytes [1] >= 0xA1 && cjkbytes [1] <= 0xFE
			&& cjkbytes [2] >= 0xA1 && cjkbytes [2] <= 0xFE
			&& cjkbytes [3] == 0);
    case 'S':	return (cjkchar >= 0xA1 && cjkchar <= 0xDF)
			|| (((cjkbytes [0] >= 0x81 && cjkbytes [0] <= 0x9F)
				||
			     (cjkbytes [0] >= 0xE0)
			    )
			    && cjkbytes [1] >= 0x40 && cjkbytes [1] <= 0xFC
			    && cjkbytes [1] != 0x7F
			    && cjkbytes [2] == 0);
/*
         UHC >> KSC
                        EUC-KR	!	A1-FE	A1-FE
                        UHC	*	81-FE	41-5A, 61-7A, 81-FE
         Johab
                        Johab		84-DE, E0-F9	31-7E, 81-FE
*/
    case 'K':	return cjkbytes [0] >= 0x81 && cjkbytes [0] <= 0xFE
			&& ((cjkbytes [1] >= 0x41 && cjkbytes [1] <= 0x5A)
			    ||
			    (cjkbytes [1] >= 0x61 && cjkbytes [1] <= 0x7A)
			    ||
			    (cjkbytes [1] >= 0x81 && cjkbytes [1] <= 0xFE)
			   )
			&& cjkbytes [2] == 0;
    case 'H':	return ((cjkbytes [0] >= 0x84 && cjkbytes [0] <= 0xDE)
			 ||
			(cjkbytes [0] >= 0xE0 && cjkbytes [0] <= 0xF9)
			)
			&&
			((cjkbytes [1] >= 0x31 && cjkbytes [1] <= 0x7E)
			 ||
			 (cjkbytes [1] >= 0x81 && cjkbytes [1] <= 0xFE)
			)
			&& cjkbytes [2] == 0;
    default:	return False;
  }
}


/*======================================================================*\
	Conversion tables mapping various CJK encodings to Unicode
\*======================================================================*/

#include "charmaps.h"

struct charmap_table_entry {
	struct encoding_table_entry * table;
	unsigned int * table_len;
	char * charmap;
	char * tag2;
	char tag1;
};

static struct charmap_table_entry charmaps_table [] =
{
#ifdef __TURBOC__
	{pc437_table, & pc437_table_len, "CP437", "PC", 'p'},
	{pc850_table, & pc850_table_len, "CP850", "PL", 'P'},
#else
#include "charmaps.t"
#endif
};


/*======================================================================*\
|*			Configuration string matching			*|
\*======================================================================*/

/**
   matchprefix determines whether its first parameter contains its 
   second parameter matching approximately as an initial prefix.
   The match ignores separating '-', '_', and space characters, 
   and does not match case.
   The algorithm assumes that letters are ASCII as this is used for 
   configuration strings only.
 */
static
int
matchprefix (s, m)
  char * s;
  char * m;
{
  char cs, cm;
  do {
	while (* m == '-' || * m == '_' || * m == ' ') {
		m ++;
	}
	while (* s == '-' || * s == '_' || * s == ' ') {
		s ++;
	}
	if (! * m) {
		return True;
	}
	if (! * s) {
		return False;
	}
	cs = * s;
	if (cs >= 'a' && cs <= 'z') {
		cs = cs - 'a' + 'A';
	}
	cm = * m;
	if (cm >= 'a' && cm <= 'z') {
		cm = cm - 'a' + 'A';
	}
	if (cm != cs) {
		return False;
	}
	s ++;
	m ++;
  } while (True);
}

/**
   matchpart determines whether its first parameter contains its 
   second parameter matching approximately as an initial prefix or 
   as a prefix of any part after a '/' or '>' separator.
   The match ignores separating '-', '_', and space characters, 
   and does not match case.
   The algorithm assumes that letters are ASCII as this is used for 
   configuration strings only.
 */
static
int
matchpart (s, m)
  char * s;
  char * m;
{
  char * p;
  if (matchprefix (s, m)) {
	return True;
  } else {
	p = strpbrk (s, ">/");
	if (p) {
		p ++;
		return matchpart (p, m);
	} else {
		return False;
	}
  }
}


/*======================================================================*\
|*			Mapping tables and functions			*|
\*======================================================================*/

/**
   Terminal character mapping table and its length
 */
static struct encoding_table_entry * terminal_table = (struct encoding_table_entry *) 0;
static unsigned int terminal_table_len = 0;

/**
   Current CJK/Unicode mapping table and its length
 */
static struct encoding_table_entry * text_table = (struct encoding_table_entry *) 0;
static unsigned int text_table_len = 0;

#ifndef use_cjk_tables
/* define dummy tables referred in code */
#ifdef __TURBOC__
#define EMPTY 1
#else
#define EMPTY 0
#endif
static struct encoding_table_entry gb_table [EMPTY];
static struct encoding_table_entry ejis_table [EMPTY];
static struct encoding_table_entry sjis_table [EMPTY];
#endif


/**
   Are mapped text and terminal encodings different?
 */
FLAG
remap_chars ()
{
  return text_table != terminal_table;
}


/**
   List of 2nd characters of 2 Unicode character mappings (mostly accents) 
   for certain 2-character CJK mappings (JIS or HKSCS);
   must be consistent with range and order of according #defines in charcode.h
 */
static unsigned int uni2_accents [] = 
	{0x309A, 0x0300, 0x0301, 0x02E5, 0x02E9, 0x0304, 0x030C};

/**
   Current encoding indications
 */
char text_encoding_tag = '-';
char * text_encoding_flag = "??";	/* for display in flags menu area */
char term_encoding_tag = '-';
static char * current_text_encoding = "";
static char * term_encoding = "";

/**
   Return charmap name of current text encoding.
 */
char *
get_text_encoding ()
{
  if (utf8_text) {
	if (utf16_file) {
		if (utf16_little_endian) {
			return "UTF-16LE";
		} else {
			return "UTF-16BE";
		}
	} else {
		return "UTF-8";
	}
  } else if (! cjk_text && ! mapped_text) {
	return "ISO 8859-1";
  } else {
	return current_text_encoding;
  }
}

/**
   Return charmap name of terminal encoding.
 */
char *
get_term_encoding ()
{
  if (* term_encoding == '\0' && utf8_screen) {
	return "UTF-8";
  } else if (* term_encoding == '\0' && ! cjk_term && ! mapped_term) {
	return "ISO 8859-1";
  } else {
	return term_encoding;
  }
}


static FLAG combined_text;

/**
   Return True if active encoding has combining characters.
 */
FLAG
encoding_has_combining ()
{
  return utf8_text
	|| (cjk_text && 
		(text_encoding_tag == 'G'
		 || text_encoding_tag == 'J'
		 || text_encoding_tag == 'S'))
	|| (mapped_text && combined_text);
}

/**
   Determine if active encoding has combining characters.
 */
static
FLAG
mapping_has_combining (term)
  FLAG term;
{
  unsigned long i;
  for (i = 0; i < 0x100; i ++) {
	unsigned long unichar;
	if (term) {
		unichar = lookup_mappedtermchar (i);
	} else {
		unichar = lookup_encodedchar (i);
	}
	if (term ? term_iscombining (unichar) : iscombining_unichar (unichar)) {
		return True;
	}
  }
  return False;
}

#ifdef split_map_entries
/*
   Decode CJK character value from split table entry.
 */
static
unsigned long
decode_cjk (entrypoi, map_table)
  struct encoding_table_entry * entrypoi;
  struct encoding_table_entry * map_table;
{
  if (map_table == gb_table) {
	if ((unsigned int) entrypoi->cjk_ext == 0xFF) {
		return entrypoi->cjk_base;
	} else {
		return ((entrypoi->cjk_base & 0x00FF) << 24)
			| (entrypoi->cjk_base & 0xFF00)
			| 0x00300030
			| ((((unsigned int) entrypoi->cjk_ext) & 0xF0) << 12)
			| (((unsigned int) entrypoi->cjk_ext) & 0x0F);
	}
  } else {
	if ((unsigned int) entrypoi->cjk_ext >= 0x90) {
		return 0x8E000000 | (((unsigned int) entrypoi->cjk_ext) << 16) | entrypoi->cjk_base;
	} else {
		return (((unsigned int) entrypoi->cjk_ext) << 16) | entrypoi->cjk_base;
	}
  }
}
#endif

static
void
setup_mapping (term, map_table, map_table_len, tag1, tag2)
  FLAG term;
  struct encoding_table_entry * map_table;
  unsigned int map_table_len;
  char tag1;
  char * tag2;
{
  FLAG multi_byte = False;
  unsigned int j;

  if (term) {
	terminal_table = map_table;
	terminal_table_len = map_table_len;
	term_encoding_tag = tag1;
  } else {
	text_table = map_table;
	text_table_len = map_table_len;
	text_encoding_tag = tag1;
	text_encoding_flag = tag2;
  }

  /* check if it is a multi-byte mapping table */
  for (j = 0; j < map_table_len; j ++) {
	unsigned long cjki;
#ifdef split_map_entries
		cjki = decode_cjk (& map_table [j], map_table);
#else
		cjki = map_table [j].cjk;
#endif
	if (cjki > 0xFF) {
		multi_byte = True;
		break;
	}
  }

  if (term) {
	if (multi_byte) {
		cjk_term = True;
		mapped_term = False;
		/* combining_screen is auto-detected */
	} else {
		mapped_term = True;
		cjk_term = False;
		/* combining_screen is auto-detected */
	}
  } else {
	if (multi_byte) {
		cjk_text = True;
		mapped_text = False;
		/* combined_text is dynamically enquired with 
		   encoding_has_combining () */
	} else {
		mapped_text = True;
		cjk_text = False;
		combined_text = mapping_has_combining (term);
	}
  }
}

/**
   Set either text or terminal character mapping table.
   Return True on success, False if tag unknown.
 */
static
FLAG
set_char_encoding (term, charmap, tag)
  FLAG term;
  char * charmap;
  char tag;
{
  int i;

  if (charmap && ! term
	&& (streq (":16", charmap) || matchpart ("UTF-16BE", charmap))) {
	utf8_text = True;
	utf16_file = True;
	utf16_little_endian = False;
	cjk_text = False;
	mapped_text = False;
	current_text_encoding = "UTF-16BE";
	text_encoding_flag = "16";
	return True;
  } else if (charmap && ! term
	&& (streq (":61", charmap) || matchpart ("UTF-16LE", charmap))) {
	utf8_text = True;
	utf16_file = True;
	utf16_little_endian = True;
	cjk_text = False;
	mapped_text = False;
	current_text_encoding = "UTF-16LE";
	text_encoding_flag = "61";
	return True;
  } else if (charmap && ! term && streq (":??", charmap)) {
	text_table_len = 0;
	text_encoding_tag = ' ';
	text_encoding_flag = "??";
	utf8_text = False;
	cjk_text = True;
	mapped_text = False;
	current_text_encoding = "[CJK]";
	return True;
  } else if (charmap ? strisprefix ("UTF-8", charmap) : tag == 'U') {
	if (term) {
		utf8_screen = True;
		utf8_input = True;
		term_encoding = "UTF-8";
	} else {
		utf8_text = True;
		utf16_file = False;
		cjk_text = False;
		mapped_text = False;
		current_text_encoding = "UTF-8";
		text_encoding_flag = "U8";
	}
	return True;
  } else if (charmap ? matchpart ("ISO 8859-1", charmap) : tag == 'L') {
	if (term) {
		utf8_screen = False;
		utf8_input = False;
		cjk_term = False;
		term_encoding = "ISO 8859-1";
	} else {
		utf8_text = False;
		cjk_text = False;
		mapped_text = False;
		current_text_encoding = "ISO 8859-1";
		text_encoding_flag = "L1";
	}
	return True;
  } else {
    for (i = 0; i < arrlen (charmaps_table); i ++) {
	if (charmap ? (charmap [0] == ':' ?
			  streq (& charmap [1], charmaps_table [i].tag2)
			: matchpart (charmaps_table [i].charmap, charmap)
		      )
		    : charmaps_table [i].tag1 == tag) {
		if (term) {
			utf8_screen = False;
			utf8_input = False;
			term_encoding = charmaps_table [i].charmap;
		} else {
			utf8_text = False;
			current_text_encoding = charmaps_table [i].charmap;
		}
		setup_mapping (term, 
				charmaps_table [i].table, 
				* charmaps_table [i].table_len, 
				charmaps_table [i].tag1, 
				charmaps_table [i].tag2);
		return True;
	}
    }
  }
  return False;
}

/**
   Set terminal character code mapping table according to encoding tag.
   Return True on success, False if tag unknown.
 */
FLAG
set_term_encoding (charmap, tag)
  char * charmap;
  char tag;
{
  return set_char_encoding (True, charmap, tag);
}

#define dont_debug_set_text_encoding

/**
   Set character mapping table and text encoding variables according 
   to encoding tag.
   Return True on success, False if tag unknown.
 */
FLAG
set_text_encoding (charmap, tag, debug_tag)
  char * charmap;
  char tag;
  char * debug_tag;
{
#ifdef debug_set_text_encoding
  printf ("set_text_encoding [%s] %s %c\n", debug_tag, charmap, tag);
#endif
  return set_char_encoding (False, charmap, tag);
}

/*
   Look up a Unicode value in a character set mapping table.
   @return CJK value, or CHAR_INVALID if not found
 */
static
unsigned long
unmap_char (unichar, map_table, map_table_len)
  unsigned long unichar;
  struct encoding_table_entry * map_table;
  unsigned int map_table_len;
{
#ifdef split_map_entries
	unsigned char unichar_high = unichar >> 16;
	unsigned short unichar_low = unichar & 0xFFFF;
#endif
	unsigned int i = 0;

	struct encoding_table_entry * map_table_poi = map_table;
	while (i ++ < map_table_len) {
#ifdef split_map_entries
		if (
		    unichar_low == map_table_poi->unicode_low
		 && unichar_high == map_table_poi->unicode_high
		   ) {
			return decode_cjk (map_table_poi, map_table);
		}
#else
		if (
		    unichar == map_table_poi->unicode
		   ) {
			return map_table_poi->cjk;
		}
#endif
		map_table_poi ++;
	}
	return CHAR_INVALID;
}

/*
   Map a character in a character set mapping table.
   @return Unicode value, or CHAR_INVALID if not found
 */
static
unsigned long
map_char (cjk, map_table, map_table_len)
  unsigned long cjk;
  struct encoding_table_entry * map_table;
  unsigned int map_table_len;
{
	int low = 0;
	int high = map_table_len - 1;
	int i;

	unsigned long cjki;

	while (low <= high) {
		i = (low + high) / 2;
#ifdef split_map_entries
		cjki = decode_cjk (& map_table [i], map_table);
#else
		cjki = map_table [i].cjk;
#endif
		if (cjki == cjk) {
#ifdef split_map_entries
			if (map_table [i].unicode_high & 0x80) {
				return 0x80000000 | (uni2_accents [map_table [i].unicode_high & 0x7F] << 16) | (map_table [i].unicode_low);
			} else {
				return (((unsigned long) map_table [i].unicode_high) << 16) | (map_table [i].unicode_low);
			}
#else
			if (map_table [i].unicode & 0x800000) {
				return 0x80000000 | (uni2_accents [(map_table [i].unicode >> 16) & 0x7F] << 16) | (map_table [i].unicode & 0xFFFF);
			} else {
				return map_table [i].unicode;
			}
#endif
		} else if (cjki >= cjk) {
			high = i - 1;
		} else {
			low = i + 1;
		}
	}
	return CHAR_INVALID;
}


/*======================================================================*\
|*		Conversion functions					*|
\*======================================================================*/

/**
   GB18030 algorithmic mapping part
 */
static
unsigned long
gb_to_unicode (gb)
  unsigned long gb;
{
	unsigned int byte2 = (gb >> 16) & 0xFF;
	unsigned int byte3 = (gb >> 8) & 0xFF;
	unsigned int byte4 = gb & 0xFF;
	if (byte2 < '0' || byte2 > '9' || byte3 < 0x81 || byte4 < '0' || byte4 > '9') {
		return CHAR_INVALID;
	}
	return (((((gb >> 24) & 0xFF) - 0x90) * 10
		+ (byte2 - 0x30)) * 126L
		+ (byte3 - 0x81)) * 10L
		+ (byte4 - 0x30)
		+ 0x10000;
}

static
unsigned long
unicode_to_gb (uc)
  unsigned long uc;
{
	unsigned int a, b, c, d;
	uc -= 0x10000;
	d = 0x30 + uc % 10;
	uc /= 10;
	c = 0x81 + uc % 126;
	uc /= 126;
	b = 0x30 + uc % 10;
	uc /= 10;
	a = 0x90 + uc;
	return (a << 24) | (b << 16) | (c << 8) | d;
}


/*
   mapped_char () converts a Unicode value into an encoded character,
   using the table given as parameter.
 */
static
unsigned long
mapped_char (unichar, map_table, map_table_len)
  unsigned long unichar;
  struct encoding_table_entry * map_table;
  unsigned int map_table_len;
{
	unsigned long cjkchar;

	if (map_table == gb_table && unichar >= 0x10000) {
		return unicode_to_gb (unichar);
	}

	cjkchar = unmap_char (unichar, map_table, map_table_len);
	if (cjkchar != CHAR_INVALID) {
		return cjkchar;
	}

	if (unichar < 0x20) {
		/* transparently return control range (for commands) */
		return unichar;
	} else if (unichar < 0x80) {
		/* transparently map ASCII range unless remapped */
		cjkchar = unichar;
		unichar = map_char (cjkchar, map_table, map_table_len);
		if (! no_unichar (unichar) && unichar != cjkchar) {
			return CHAR_INVALID;
		} else {
			return cjkchar;
		}
	} else {
		/* notify "not found" */
		return CHAR_INVALID;
	}
}

/*
   mappedtermchar () converts a Unicode value into an encoded character,
   using the terminal encoding (terminal_table).
 */
unsigned long
mappedtermchar (unichar)
  unsigned long unichar;
{
	return mapped_char (unichar, terminal_table, terminal_table_len);
}

/*
   encodedchar () converts a Unicode value into an encoded character,
   using the current text encoding (text_table).
 */
unsigned long
encodedchar (unichar)
  unsigned long unichar;
{
  if (cjk_text || mapped_text) {
	return mapped_char (unichar, text_table, text_table_len);
  } else if (utf8_text || unichar < 0x100) {
	return unichar;
  } else {
	return CHAR_INVALID;
  }
}

/*
   encodedchar2 () converts two Unicode values into one JIS character,
   using the current text encoding (text_table).
 */
unsigned long
encodedchar2 (uc1, uc2)
  unsigned long uc1;
  unsigned long uc2;
{
  int i;
  for (i = 0; i < arrlen (uni2_accents); i ++) {
	if (uni2_accents [i] == uc2) {
		unsigned long unichar = uc1 | ((0x80 + i) << uni2tag_shift);
		return mapped_char (unichar, text_table, text_table_len);
	}
  }
  return CHAR_INVALID;
}


/*
   lookup_mapped_char () converts an encoded character to Unicode,
   using the table given as parameter.
 */
static
unsigned long
lookup_mapped_char (cjk, map_table, map_table_len)
  unsigned long cjk;
  struct encoding_table_entry * map_table;
  unsigned int map_table_len;
{
	unsigned long unichar;

	if (map_table == gb_table && cjk >= 0x90000000) {
		return gb_to_unicode (cjk);
	}

	unichar = map_char (cjk, map_table, map_table_len);
	if (! no_unichar (unichar)) {
		return unichar;
	} else if (cjk < 0x80) {
		/* transparently map ASCII range */
		return cjk;
	} else {
		/* notify "not found" */
		return CHAR_INVALID;
	}
}

/*
   lookup_mappedtermchar () converts an encoded character to Unicode,
   using the terminal encoding (terminal_table).
 */
unsigned long
lookup_mappedtermchar (cjk)
  unsigned long cjk;
{
	return lookup_mapped_char (cjk, terminal_table, terminal_table_len);
}

/*
   lookup_encodedchar () converts an encoded character to Unicode,
   using the current text encoding (text_table).
 */
unsigned long
lookup_encodedchar (cjk)
  unsigned long cjk;
{
	return lookup_mapped_char (cjk, text_table, text_table_len);
}


/*======================================================================*\
|*				End					*|
\*======================================================================*/