utf8.c

广泛使用的邮件服务器！同时

 * Accepts: source sized text
 *	    pointer to return sized text
 *	    EUC parameter table
 *	    canonicalization function
 */

void utf8_text_euc (SIZEDTEXT *text,SIZEDTEXT *ret,void *tab,ucs4cn_t cv,
		    ucs4de_t de)
{
  unsigned long i;
  unsigned char *s;
  unsigned int pass,c,c1,ku,ten;
  struct utf8_eucparam *p1 = (struct utf8_eucparam *) tab;
  struct utf8_eucparam *p2 = p1 + 1;
  struct utf8_eucparam *p3 = p1 + 2;
  unsigned short *t1 = (unsigned short *) p1->tab;
  unsigned short *t2 = (unsigned short *) p2->tab;
  unsigned short *t3 = (unsigned short *) p3->tab;
  for (pass = 0,s = NIL,ret->size = 0; pass <= 1; pass++) {
    for (i = 0; i < text->size;) {
				/* not CS0? */
      if ((c = text->data[i++]) & BIT8) {
				/* yes, must have another high byte */
	if ((i >= text->size) || !((c1 = text->data[i++]) & BIT8))
	  c = UBOGON;		/* out of space or bogon */
	else switch (c) {	/* check 8bit code set */
	case EUC_CS2:		/* CS2 */
	  if (p2->base_ku) {	/* CS2 set up? */
	    if (p2->base_ten)	/* yes, multibyte? */
	      c = ((i < text->size) && ((c = text->data[i++]) & BIT8) &&
		   ((ku = (c1 & BITS7) - p2->base_ku) < p2->max_ku) &&
		   ((ten = (c & BITS7) - p2->base_ten) < p2->max_ten)) ?
		     t2[(ku*p2->max_ten) + ten] : UBOGON;
	    else c = ((c1 >= p2->base_ku) && (c1 < p2->max_ku)) ?
	      c1 + ((unsigned long) p2->tab) : UBOGON;
	  }	  
	  else {		/* CS2 not set up */
	    c = UBOGON;		/* swallow byte, say bogon */
	    if (i < text->size) i++;
	  }
	  break;
	case EUC_CS3:		/* CS3 */
	  if (p3->base_ku) {	/* CS3 set up? */
	    if (p3->base_ten)	/* yes, multibyte? */
	      c = ((i < text->size) && ((c = text->data[i++]) & BIT8) &&
		   ((ku = (c1 & BITS7) - p3->base_ku) < p3->max_ku) &&
		   ((ten = (c & BITS7) - p3->base_ten) < p3->max_ten)) ?
		     t3[(ku*p3->max_ten) + ten] : UBOGON;
	    else c = ((c1 >= p3->base_ku) && (c1 < p3->max_ku)) ?
	      c1 + ((unsigned long) p3->tab) : UBOGON;
	  }	  
	  else {		/* CS3 not set up */
	    c = UBOGON;		/* swallow byte, say bogon */
	    if (i < text->size) i++;
	  }
	  break;

	default:
	  if (((ku = (c & BITS7) - p1->base_ku) >= p1->max_ku) ||
	      ((ten = (c1 & BITS7) - p1->base_ten) >= p1->max_ten)) c = UBOGON;
	  else if (((c = t1[(ku*p1->max_ten) + ten]) == UBOGON) &&
		   /* special hack for JIS X 0212: merge rows less than 10 */
		   ku && (ku < 10) && t3 && p3->base_ten)
	    c = t3[((ku - (p3->base_ku - p1->base_ku))*p3->max_ten) + ten];
	}
      }
				/* convert if second pass */
      if (pass) UTF8_WRITE_BMP (s,c,cv,de)
      else UTF8_COUNT_BMP (ret->size,c,cv,de);
    }
    if (!pass) (s = ret->data = (unsigned char *)
		fs_get (ret->size + 1))[ret->size] =NIL;
  }
}


/* Convert ASCII + double-byte sized text to UTF-8
 * Accepts: source sized text
 *	    pointer to return sized text
 *	    conversion table
 *	    canonicalization function
 */

void utf8_text_dbyte (SIZEDTEXT *text,SIZEDTEXT *ret,void *tab,ucs4cn_t cv,
		      ucs4de_t de)
{
  unsigned long i;
  unsigned char *s;
  unsigned int c,c1,ku,ten;
  struct utf8_eucparam *p1 = (struct utf8_eucparam *) tab;
  unsigned short *t1 = (unsigned short *) p1->tab;
  for (ret->size = i = 0; i < text->size;) {
    if ((c = text->data[i++]) & BIT8) {
				/* special hack for GBK: 0x80 is Euro */
      if ((c == 0x80) && (t1 == (unsigned short *) gb2312tab)) c = UCS2_EURO;
      else c = ((i < text->size) && (c1 = text->data[i++]) &&
		((ku = c - p1->base_ku) < p1->max_ku) &&
		((ten = c1 - p1->base_ten) < p1->max_ten)) ?
	     t1[(ku*p1->max_ten) + ten] : UBOGON;
    }
    UTF8_COUNT_BMP (ret->size,c,cv,de)
  }
  (s = ret->data = (unsigned char *) fs_get (ret->size + 1))[ret->size] = NIL;
  for (i = 0; i < text->size;) {
    if ((c = text->data[i++]) & BIT8) {
				/* special hack for GBK: 0x80 is Euro */
      if ((c == 0x80) && (t1 == (unsigned short *) gb2312tab)) c = UCS2_EURO;
      else c = ((i < text->size) && (c1 = text->data[i++]) &&
		((ku = c - p1->base_ku) < p1->max_ku) &&
		((ten = c1 - p1->base_ten) < p1->max_ten)) ?
	     t1[(ku*p1->max_ten) + ten] : UBOGON;
    }
    UTF8_WRITE_BMP (s,c,cv,de)	/* convert UCS-2 to UTF-8 */
  }
}

/* Convert ASCII + double byte 2 plane sized text to UTF-8
 * Accepts: source sized text
 *	    pointer to return sized text
 *	    conversion table
 *	    canonicalization function
 */

void utf8_text_dbyte2 (SIZEDTEXT *text,SIZEDTEXT *ret,void *tab,ucs4cn_t cv,
		       ucs4de_t de)
{
  unsigned long i;
  unsigned char *s;
  unsigned int c,c1,ku,ten;
  struct utf8_eucparam *p1 = (struct utf8_eucparam *) tab;
  struct utf8_eucparam *p2 = p1 + 1;
  unsigned short *t = (unsigned short *) p1->tab;
  for (ret->size = i = 0; i < text->size;) {
    if ((c = text->data[i++]) & BIT8) {
      if ((i >= text->size) || !(c1 = text->data[i++]))
	c = UBOGON;		/* out of space or bogon */
      else if (c1 & BIT8)	/* high vs. low plane */
	c = ((ku = c - p2->base_ku) < p2->max_ku &&
	     ((ten = c1 - p2->base_ten) < p2->max_ten)) ?
	       t[(ku*(p1->max_ten + p2->max_ten)) + p1->max_ten + ten] :UBOGON;
      else c = ((ku = c - p1->base_ku) < p1->max_ku &&
		((ten = c1 - p1->base_ten) < p1->max_ten)) ?
		  t[(ku*(p1->max_ten + p2->max_ten)) + ten] : UBOGON;
    }
    UTF8_COUNT_BMP (ret->size,c,cv,de)
  }
  (s = ret->data = (unsigned char *) fs_get (ret->size + 1))[ret->size] = NIL;
  for (i = 0; i < text->size;) {
    if ((c = text->data[i++]) & BIT8) {
      if ((i >= text->size) || !(c1 = text->data[i++]))
	c = UBOGON;		/* out of space or bogon */
      else if (c1 & BIT8)	/* high vs. low plane */
	c = ((ku = c - p2->base_ku) < p2->max_ku &&
	     ((ten = c1 - p2->base_ten) < p2->max_ten)) ?
	       t[(ku*(p1->max_ten + p2->max_ten)) + p1->max_ten + ten] :UBOGON;
      else c = ((ku = c - p1->base_ku) < p1->max_ku &&
		((ten = c1 - p1->base_ten) < p1->max_ten)) ?
		  t[(ku*(p1->max_ten + p2->max_ten)) + ten] : UBOGON;
    }
    UTF8_WRITE_BMP (s,c,cv,de)	/* convert UCS-2 to UTF-8 */
  }
}

#ifdef JISTOUNICODE		/* Japanese */
/* Convert Shift JIS sized text to UTF-8
 * Accepts: source sized text
 *	    pointer to return sized text
 *	    canonicalization function
 */

void utf8_text_sjis (SIZEDTEXT *text,SIZEDTEXT *ret,ucs4cn_t cv,
		     ucs4de_t de)
{
  unsigned long i;
  unsigned char *s;
  unsigned int c,c1,ku,ten;
  for (ret->size = i = 0; i < text->size;) {
    if ((c = text->data[i++]) & BIT8) {
				/* half-width katakana */
      if ((c >= MIN_KANA_8) && (c < MAX_KANA_8)) c += KANA_8;
      else if (i >= text->size) c = UBOGON;
      else {			/* Shift-JIS */
	c1 = text->data[i++];
	SJISTOJIS (c,c1);
	c = JISTOUNICODE (c,c1,ku,ten);
      }
    }
				/* compromise - do yen sign but not overline */
    else if (c == JISROMAN_YEN) c = UCS2_YEN;
    UTF8_COUNT_BMP (ret->size,c,cv,de)
  }
  (s = ret->data = (unsigned char *) fs_get (ret->size + 1))[ret->size] = NIL;
  for (i = 0; i < text->size;) {
    if ((c = text->data[i++]) & BIT8) {
				/* half-width katakana */
      if ((c >= MIN_KANA_8) && (c < MAX_KANA_8)) c += KANA_8;
      else {			/* Shift-JIS */
	c1 = text->data[i++];
	SJISTOJIS (c,c1);
	c = JISTOUNICODE (c,c1,ku,ten);
      }
    }
				/* compromise - do yen sign but not overline */
    else if (c == JISROMAN_YEN) c = UCS2_YEN;
    UTF8_WRITE_BMP (s,c,cv,de)	/* convert UCS-2 to UTF-8 */
  }
}
#endif

/* Convert ISO-2022 sized text to UTF-8
 * Accepts: source sized text
 *	    pointer to returned sized text
 *	    canonicalization function
 */

void utf8_text_2022 (SIZEDTEXT *text,SIZEDTEXT *ret,ucs4cn_t cv,ucs4de_t de)
{
  unsigned long i;
  unsigned char *s;
  unsigned int pass,state,c,co,gi,gl,gr,g[4],ku,ten;
  for (pass = 0,s = NIL,ret->size = 0; pass <= 1; pass++) {
    gi = 0;			/* quell compiler warnings */
    state = I2S_CHAR;		/* initialize engine */
    g[0]= g[2] = I2CS_ASCII;	/* G0 and G2 are ASCII */
    g[1]= g[3] = I2CS_ISO8859_1;/* G1 and G3 are ISO-8850-1 */
    gl = I2C_G0; gr = I2C_G1;	/* left is G0, right is G1 */
    for (i = 0; i < text->size;) {
      c = text->data[i++];
      switch (state) {		/* dispatch based upon engine state */
      case I2S_ESC:		/* ESC seen */
	switch (c) {		/* process intermediate character */
	case I2C_MULTI:		/* multibyte character? */
	  state = I2S_MUL;	/* mark multibyte flag seen */
	  break;
        case I2C_SS2:		/* single shift GL to G2 */
	case I2C_SS2_ALT:	/* Taiwan SeedNet */
	  gl |= I2C_SG2;
	  break;
        case I2C_SS3:		/* single shift GL to G3 */
	case I2C_SS3_ALT:	/* Taiwan SeedNet */
	  gl |= I2C_SG3;
	  break;
        case I2C_LS2:		/* shift GL to G2 */
	  gl = I2C_G2;
	  break;
        case I2C_LS3:		/* shift GL to G3 */
	  gl = I2C_G3;
	  break;
        case I2C_LS1R:		/* shift GR to G1 */
	  gr = I2C_G1;
	  break;
        case I2C_LS2R:		/* shift GR to G2 */
	  gr = I2C_G2;
	  break;
        case I2C_LS3R:		/* shift GR to G3 */
	  gr = I2C_G3;
	  break;
	case I2C_G0_94: case I2C_G1_94: case I2C_G2_94:	case I2C_G3_94:
	  g[gi = c - I2C_G0_94] = (state == I2S_MUL) ? I2CS_94x94 : I2CS_94;
	  state = I2S_INT;	/* ready for character set */
	  break;
	case I2C_G0_96:	case I2C_G1_96: case I2C_G2_96:	case I2C_G3_96:
	  g[gi = c - I2C_G0_96] = (state == I2S_MUL) ? I2CS_96x96 : I2CS_96;
	  state = I2S_INT;	/* ready for character set */
	  break;
	default:		/* bogon */
	  if (pass) *s++ = I2C_ESC,*s++ = c;
	  else ret->size += 2;
	  state = I2S_CHAR;	/* return to previous state */
	}
	break;

      case I2S_MUL:		/* ESC $ */
	switch (c) {		/* process multibyte intermediate character */
	case I2C_G0_94: case I2C_G1_94: case I2C_G2_94:	case I2C_G3_94:
	  g[gi = c - I2C_G0_94] = I2CS_94x94;
	  state = I2S_INT;	/* ready for character set */
	  break;
	case I2C_G0_96:	case I2C_G1_96: case I2C_G2_96:	case I2C_G3_96:
	  g[gi = c - I2C_G0_96] = I2CS_96x96;
	  state = I2S_INT;	/* ready for character set */
	  break;
	default:		/* probably omitted I2CS_94x94 */
	  g[gi = I2C_G0] = I2CS_94x94 | c;
	  state = I2S_CHAR;	/* return to character state */
	}
	break;
      case I2S_INT:
	state = I2S_CHAR;	/* return to character state */
	g[gi] |= c;		/* set character set */
	break;

      case I2S_CHAR:		/* character data */
	switch (c) {
	case I2C_ESC:		/* ESC character */
	  state = I2S_ESC;	/* see if ISO-2022 prefix */
	  break;
	case I2C_SI:		/* shift GL to G0 */
	  gl = I2C_G0;
	  break;
	case I2C_SO:		/* shift GL to G1 */
	  gl = I2C_G1;
	  break;
        case I2C_SS2_ALT:	/* single shift GL to G2 */
	case I2C_SS2_ALT_7:
	  gl |= I2C_SG2;
	  break;
        case I2C_SS3_ALT:	/* single shift GL to G3 */
	case I2C_SS3_ALT_7:
	  gl |= I2C_SG3;
	  break;

	default:		/* ordinary character */
	  co = c;		/* note original character */
	  if (gl & (3 << 2)) {	/* single shifted? */
	    gi = g[gl >> 2];	/* get shifted character set */
	    gl &= 0x3;		/* cancel shift */
	  }
				/* select left or right half */
	  else gi = (c & BIT8) ? g[gr] : g[gl];
	  c &= BITS7;		/* make 7-bit */
	  switch (gi) {		/* interpret in character set */
	  case I2CS_ASCII:	/* ASCII */
	    break;		/* easy! */
	  case I2CS_BRITISH:	/* British ASCII */
				/* Pound sterling sign */
	    if (c == BRITISH_POUNDSTERLING) c = UCS2_POUNDSTERLING;
	    break;
	  case I2CS_JIS_ROMAN:	/* JIS Roman */
	  case I2CS_JIS_BUGROM:	/* old bugs */
	    switch (c) {	/* two exceptions to ASCII */
	    case JISROMAN_YEN:	/* Yen sign */
	      c = UCS2_YEN;
	      break;
				/* overline */
	    case JISROMAN_OVERLINE:
	      c = UCS2_OVERLINE;
	      break;
	    }
	    break;
	  case I2CS_JIS_KANA:	/* JIS hankaku katakana */
	    if ((c >= MIN_KANA_7) && (c < MAX_KANA_7)) c += KANA_7;
	    break;

	  case I2CS_ISO8859_1:	/* Latin-1 (West European) */
	    c |= BIT8;		/* just turn on high bit */
	    break;
	  case I2CS_ISO8859_2:	/* Latin-2 (Czech, Slovak) */
	    c = iso8859_2tab[c];
	    break;
	  case I2CS_ISO8859_3:	/* Latin-3 (Dutch, Turkish) */
	    c = iso8859_3tab[c];
	    break;
	  case I2CS_ISO8859_4:	/* Latin-4 (Scandinavian) */
	    c = iso8859_4tab[c];
	    break;
	  case I2CS_ISO8859_5:	/* Cyrillic */
	    c = iso8859_5tab[c];
	    break;
	  case I2CS_ISO8859_6:	/* Arabic */
	    c = iso8859_6tab[c];
	    break;
	  case I2CS_ISO8859_7:	/* Greek */
	    c = iso8859_7tab[c];
	    break;
	  case I2CS_ISO8859_8:	/* Hebrew */
	    c = iso8859_8tab[c];
	    break;
	  case I2CS_ISO8859_9:	/* Latin-5 (Finnish, Portuguese) */
	    c = iso8859_9tab[c];
	    break;
	  case I2CS_TIS620:	/* Thai */
	    c = tis620tab[c];
	    break;
	  case I2CS_ISO8859_10:	/* Latin-6 (Northern Europe) */
	    c = iso8859_10ta