utf8.c

广泛使用的邮件服务器！同时

{
  unsigned long i;
  if (!script) return (SCRIPT *) &utf8_scvalid[0];
  else if (*script && (strlen (script) < 128))
    for (i = 0; utf8_scvalid[i].name; i++)
      if (!compare_cstring (script,utf8_scvalid[i].name))
	return (SCRIPT *) &utf8_scvalid[i];
  return NIL;			/* failed */
}


/* Look up charset name or return entire table
 * Accepts: charset name or NIL
 * Returns: charset table entry or NIL if unknown
 */

const CHARSET *utf8_charset (char *charset)
{
  unsigned long i;
  if (!charset) return (CHARSET *) &utf8_csvalid[0];
  else if (*charset && (strlen (charset) < 128))
    for (i = 0; utf8_csvalid[i].name; i++)
      if (!compare_cstring (charset,utf8_csvalid[i].name))
	return (CHARSET *) &utf8_csvalid[i];
  return NIL;			/* failed */
}

/* Validate charset and generate error message if invalid
 * Accepts: bad character set
 * Returns: NIL if good charset, else error message string
 */

#define BADCSS "[BADCHARSET ("
#define BADCSE ")] Unknown charset: "

char *utf8_badcharset (char *charset)
{
  char *msg = NIL;
  if (!utf8_charset (charset)) {
    char *s,*t;
    unsigned long i,j;
				/* calculate size of header, trailer, and bad
				 * charset plus charset names */
    for (i = 0, j = sizeof (BADCSS) + sizeof (BADCSE) + strlen (charset) - 2;
	 utf8_csvalid[i].name; i++)
      j += strlen (utf8_csvalid[i].name) + 1;
				/* not built right */
    if (!i) fatal ("No valid charsets!");
				/* header */
    for (s = msg = (char *) fs_get (j), t = BADCSS; *t; *s++ = *t++);
				/* each charset */
    for (i = 0; utf8_csvalid[i].name; *s++ = ' ', i++)
      for (t = utf8_csvalid[i].name; *t; *s++ = *t++);
				/* back over last space, trailer */
    for (t = BADCSE, --s; *t; *s++ = *t++);
				/* finally bogus charset */
    for (t = charset; *t; *s++ = *t++);
    *s++ = '\0';		/* finally tie off string */
    if (s != (msg + j)) fatal ("charset msg botch");
  }
  return msg;
}

/* Convert charset labelled sized text to UTF-8
 * Accepts: source sized text
 *	    charset
 *	    pointer to returned sized text if non-NIL
 *	    flags
 * Returns: T if successful, NIL if failure
 */

long utf8_text (SIZEDTEXT *text,char *charset,SIZEDTEXT *ret,long flags)
{
  ucs4cn_t cv = (flags & U8T_CASECANON) ? ucs4_titlecase : NIL;
  ucs4de_t de = (flags & U8T_DECOMPOSE) ? ucs4_decompose_recursive : NIL;
  const CHARSET *cs = (charset && *charset) ?
    utf8_charset (charset) : utf8_infercharset (text);
  if (cs) return (text && ret) ? utf8_text_cs (text,cs,ret,cv,de) : LONGT;
  if (ret) {			/* no conversion possible */
    ret->data = text->data;	/* so return source */
    ret->size = text->size;
  }
  return NIL;			/* failure */
}


/* Operations used in converting data */

#define UTF8_COUNT_BMP(count,c,cv,de) {		\
  void *more = NIL;				\
  if (cv) c = (*cv) (c);			\
  if (de) c = (*de) (c,&more);			\
  do count += UTF8_SIZE_BMP(c);			\
  while (more && (c = (*de) (U8G_ERROR,&more)));\
}

#define UTF8_WRITE_BMP(b,c,cv,de) {		\
  void *more = NIL;				\
  if (cv) c = (*cv) (c);			\
  if (de) c = (*de) (c,&more);			\
  do UTF8_PUT_BMP (b,c)				\
  while (more && (c = (*de) (U8G_ERROR,&more)));\
}

#define UTF8_COUNT(count,c,cv,de) {		\
  void *more = NIL;				\
  if (cv) c = (*cv) (c);			\
  if (de) c = (*de) (c,&more);			\
  do count += utf8_size (c);			\
  while (more && (c = (*de) (U8G_ERROR,&more)));\
}

#define UTF8_WRITE(b,c,cv,de) {			\
  void *more = NIL;				\
  if (cv) c = (*cv) (c);			\
  if (de) c = (*de) (c,&more);			\
  do b = utf8_put (b,c);			\
  while (more && (c = (*de) (U8G_ERROR,&more)));\
}

/* Convert sized text to UTF-8 given CHARSET block
 * Accepts: source sized text
 *	    CHARSET block
 *	    pointer to returned sized text 
 *	    canonicalization function
 *	    decomposition function
 * Returns: T if successful, NIL if failure
 */

long utf8_text_cs (SIZEDTEXT *text,const CHARSET *cs,SIZEDTEXT *ret,
		   ucs4cn_t cv,ucs4de_t de)
{
  ret->data = text->data;	/* default to source */
  ret->size = text->size;
  switch (cs->type) {		/* convert if type known */
  case CT_ASCII:		/* 7-bit ASCII no table */
  case CT_UTF8:			/* variable UTF-8 encoded Unicode no table */
    if (cv || de) utf8_text_utf8 (text,ret,cv,de);
    break;
  case CT_1BYTE0:		/* 1 byte no table */
    utf8_text_1byte0 (text,ret,cv,de);
    break;
  case CT_1BYTE:		/* 1 byte ASCII + table 0x80-0xff */
    utf8_text_1byte (text,ret,cs->tab,cv,de);
    break;
  case CT_1BYTE8:		/* 1 byte table 0x00 - 0xff */
    utf8_text_1byte8 (text,ret,cs->tab,cv,de);
    break;
  case CT_EUC:			/* 2 byte ASCII + utf8_eucparam base/CS2/CS3 */
    utf8_text_euc (text,ret,cs->tab,cv,de);
    break;
  case CT_DBYTE:		/* 2 byte ASCII + utf8_eucparam */
    utf8_text_dbyte (text,ret,cs->tab,cv,de);
    break;
  case CT_DBYTE2:		/* 2 byte ASCII + utf8_eucparam plane1/2 */
    utf8_text_dbyte2 (text,ret,cs->tab,cv,de);
    break;
  case CT_UTF7:			/* variable UTF-7 encoded Unicode no table */
    utf8_text_utf7 (text,ret,cv,de);
    break;
  case CT_UCS2:			/* 2 byte 16-bit Unicode no table */
    utf8_text_ucs2 (text,ret,cv,de);
    break;
  case CT_UCS4:			/* 4 byte 32-bit Unicode no table */
    utf8_text_ucs4 (text,ret,cv,de);
    break;
  case CT_UTF16:		/* variable UTF-16 encoded Unicode no table */
    utf8_text_utf16 (text,ret,cv,de);
    break;
  case CT_2022:			/* variable ISO-2022 encoded no table*/
    utf8_text_2022 (text,ret,cv,de);
    break;
  case CT_SJIS:			/* 2 byte Shift-JIS encoded JIS no table */
    utf8_text_sjis (text,ret,cv,de);
    break;
  default:			/* unknown character set type */
    return NIL;
  }
  return LONGT;			/* return success */
}

/* Reverse mapping routines
 *
 * These routines only support character sets, not all possible charsets.  In
 * particular, they do not support any Unicode encodings or ISO 2022.
 *
 * As a special dispensation, utf8_cstext() and utf8_cstocstext() support
 * support ISO-2022-JP if EUC-JP can be reverse mapped; and utf8_rmaptext()
 * will generated ISO-2022-JP using an EUC-JP rmap if flagged to do so.
 *
 * No attempt is made to map "equivalent" Unicode characters or Unicode
 * characters that have the same glyph; nor is there any attempt to handle
 * combining characters or otherwise do any stringprep.  Maybe later.
 */


/* Convert UTF-8 sized text to charset
 * Accepts: source sized text
 *	    destination charset
 *	    pointer to returned sized text
 *	    substitute character if not in cs, else NIL to return failure
 * Returns: T if successful, NIL if failure
 */


long utf8_cstext (SIZEDTEXT *text,char *charset,SIZEDTEXT *ret,
		  unsigned long errch)
{
  short iso2022jp = !compare_cstring (charset,"ISO-2022-JP");
  unsigned short *rmap = utf8_rmap (iso2022jp ? "EUC-JP" : charset);
  return rmap ? utf8_rmaptext (text,rmap,ret,errch,iso2022jp) : NIL;
}

/* Convert charset labelled sized text to another charset
 * Accepts: source sized text
 *	    source charset
 *	    pointer to returned sized text
 *	    destination charset
 *	    substitute character if not in dest cs, else NIL to return failure
 * Returns: T if successful, NIL if failure
 *
 * This routine has the same restricts as utf8_cstext().
 */

long utf8_cstocstext (SIZEDTEXT *src,char *sc,SIZEDTEXT *dst,char *dc,
		      unsigned long errch)
{
  SIZEDTEXT utf8;
  const CHARSET *scs,*dcs;
  unsigned short *rmap;
  long ret = NIL;
  long iso2022jp;
				/* lookup charsets and reverse map */
  if ((dc && (dcs = utf8_charset (dc))) &&
      (rmap = (iso2022jp = ((dcs->type == CT_2022) &&
			    !compare_cstring (dcs->name,"ISO-2022-JP"))) ?
       utf8_rmap ("EUC-JP") : utf8_rmap_cs (dcs)) &&
      (scs = (sc && *sc) ? utf8_charset (sc) : utf8_infercharset (src))) {
				/* init temporary buffer */
    memset (&utf8,NIL,sizeof (SIZEDTEXT));
				/* source cs equivalent to dest cs? */
    if ((scs->type == dcs->type) && (scs->tab == dcs->tab)) {
      dst->data = src->data;	/* yes, just copy pointers */
      dst->size = src->size;
      ret = LONGT;
    }
				/* otherwise do the conversion */
    else ret = (utf8_text_cs (src,scs,&utf8,NIL,NIL) &&
		utf8_rmaptext (&utf8,rmap,dst,errch,iso2022jp));
				/* flush temporary buffer */
    if (utf8.data && (utf8.data != src->data) && (utf8.data != dst->data))
      fs_give ((void **) &utf8.data);
  }
  return ret;
}

/* Cached rmap */

static const CHARSET *currmapcs = NIL;
static unsigned short *currmap = NIL;


/* Cache and return map for UTF-8 -> character set
 * Accepts: character set name
 * Returns: cached map if character set found, else NIL
 */

unsigned short *utf8_rmap (char *charset)
{
  return (currmapcs && !compare_cstring (charset,currmapcs->name)) ? currmap :
    utf8_rmap_cs (utf8_charset (charset));
}


/* Cache and return map for UTF-8 -> character set given CHARSET block
 * Accepts: CHARSET block
 * Returns: cached map if character set found, else NIL
 */

unsigned short *utf8_rmap_cs (const CHARSET *cs)
{
  unsigned short *ret = NIL;
  if (!cs);			/* have charset? */
  else if (cs == currmapcs) ret = currmap;
  else if (ret = utf8_rmap_gen (cs,currmap)) {
    currmapcs = cs;
    currmap = ret;
  }
  return ret;
}

/* Return map for UTF-8 -> character set given CHARSET block
 * Accepts: CHARSET block
 *	    old map to recycle
 * Returns: map if character set found, else NIL
 */

unsigned short *utf8_rmap_gen (const CHARSET *cs,unsigned short *oldmap)
{
  unsigned short u,*tab,*rmap;
  unsigned int i,m,ku,ten;
  struct utf8_eucparam *param,*p2;
  switch (cs->type) {		/* is a character set? */
  case CT_ASCII:		/* 7-bit ASCII no table */
  case CT_1BYTE0:		/* 1 byte no table */
  case CT_1BYTE:		/* 1 byte ASCII + table 0x80-0xff */
  case CT_1BYTE8:		/* 1 byte table 0x00 - 0xff */
  case CT_EUC:			/* 2 byte ASCII + utf8_eucparam base/CS2/CS3 */
  case CT_DBYTE:		/* 2 byte ASCII + utf8_eucparam */
  case CT_DBYTE2:		/* 2 byte ASCII + utf8_eucparam plane1/2 */
  case CT_SJIS:			/* 2 byte Shift-JIS */
    rmap = oldmap ? oldmap :	/* recycle old map if supplied else make new */
      (unsigned short *) fs_get (65536 * sizeof (unsigned short));
				/* initialize table for ASCII */
    for (i = 0; i < 128; i++) rmap[i] = (unsigned short) i;
				/* populate remainder of table with NOCHAR */
#define NOCHARBYTE (NOCHAR & 0xff)
#if NOCHAR - ((NOCHARBYTE << 8) | NOCHARBYTE)
    while (i < 65536) rmap[i++] = NOCHAR;
#else
    memset (rmap + 128,NOCHARBYTE,(65536 - 128) * sizeof (unsigned short));
#endif
    break;
  default:			/* unsupported charset type */
    rmap = NIL;			/* no map possible */
  }
  if (rmap) {			/* have a map? */
    switch (cs->type) {		/* additional reverse map actions */
    case CT_1BYTE0:		/* 1 byte no table */
      for (i = 128; i < 256; i++) rmap[i] = (unsigned short) i;
      break;
    case CT_1BYTE:		/* 1 byte ASCII + table 0x80-0xff */
      for (tab = (unsigned short *) cs->tab,i = 128; i < 256; i++)
	if (tab[i & BITS7] != UBOGON) rmap[tab[i & BITS7]] = (unsigned short)i;
      break;
    case CT_1BYTE8:		/* 1 byte table 0x00 - 0xff */
      for (tab = (unsigned short *) cs->tab,i = 0; i < 256; i++)
	if (tab[i] != UBOGON) rmap[tab[i]] = (unsigned short) i;
      break;
    case CT_EUC:		/* 2 byte ASCII + utf8_eucparam base/CS2/CS3 */
      for (param = (struct utf8_eucparam *) cs->tab,
	     tab = (unsigned short *) param->tab, ku = 0;
	   ku < param->max_ku; ku++)
	for (ten = 0; ten < param->max_ten; ten++)
	  if ((u = tab[(ku * param->max_ten) + ten]) != UBOGON)
	    rmap[u] = ((ku + param->base_ku) << 8) +
	      (ten + param->base_ten) + 0x8080;
      break;

    case CT_DBYTE:		/* 2 byte ASCII + utf8_eucparam */
      for (param = (struct utf8_eucparam *) cs->tab,
	     tab = (unsigned short *) param->tab, ku = 0;
	   ku < param->max_ku; ku++)
	for (ten = 0; ten < param->max_ten; ten++)
	  if ((u = tab[(ku * param->max_ten) + ten]) != UBOGON)
	    rmap[u] = ((ku + param->base_ku) << 8) + (ten + param->base_ten);
      break;
    case CT_DBYTE2:		/* 2 byte ASCII + utf8_eucparam plane1/2 */
      param = (struct utf8_eucparam *) cs->tab;
      p2 = param + 1;		/* plane 2 parameters */
				/* only ten parameters should differ */
      if ((param->base_ku != p2->base_ku) || (param->max_ku != p2->max_ku))
	fatal ("ku definition error for CT_DBYTE2 charset");
				/* total codepoints in each ku */
      m = param->max_ten + p2->max_ten;
      tab = (unsigned short *) param->tab;
      for (ku = 0; ku < param->max_ku; ku++) {
	for (ten = 0; ten < param->max_ten; ten++)
	  if ((u = tab[(ku * m) + ten]) != UBOGON)
	    rmap[u] = ((ku + param->base_ku) << 8) + (ten + param->base_ten);
	for (ten = 0; ten < p2->max_ten; ten++)
	  if ((u = tab[(ku * m) + param->max_ten + ten]) != UBOGON)
	    rmap[u] = ((ku + param->base_ku) << 8) + (ten + p2->base_ten);
      }
      break;
    case CT_SJIS:		/* 2 byte Shift-JIS */
      for (ku = 0; ku < MAX_JIS0208_KU; ku++)