xmltok.c

一套很值得分析的短信SMS开发源代码。是我今年早些时候从taobao上买来的。但我现在也没看完（先说清楚

  }
  else {
    if (versionPtr)
      *versionPtr = val;
    if (versionEndPtr)
      *versionEndPtr = ptr;
    if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) {
      *badPtr = ptr;
      return 0;
    }
    if (!name) {
      if (isGeneralTextEntity) {
	/* a TextDecl must have an EncodingDecl */
	*badPtr = ptr;
	return 0;
      }
      return 1;
    }
  }
  if (XmlNameMatchesAscii(enc, name, nameEnd, KW_encoding)) {
    int c = toAscii(enc, val, end);
    if (!(ASCII_a <= c && c <= ASCII_z) && !(ASCII_A <= c && c <= ASCII_Z)) {
      *badPtr = val;
      return 0;
    }
    if (encodingName)
      *encodingName = val;
    if (encoding)
      *encoding = encodingFinder(enc, val, ptr - enc->minBytesPerChar);
    if (!parsePseudoAttribute(enc, ptr, end, &name, &nameEnd, &val, &ptr)) {
      *badPtr = ptr;
      return 0;
    }
    if (!name)
      return 1;
  }
  if (!XmlNameMatchesAscii(enc, name, nameEnd, KW_standalone) || isGeneralTextEntity) {
    *badPtr = name;
    return 0;
  }
  if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_yes)) {
    if (standalone)
      *standalone = 1;
  }
  else if (XmlNameMatchesAscii(enc, val, ptr - enc->minBytesPerChar, KW_no)) {
    if (standalone)
      *standalone = 0;
  }
  else {
    *badPtr = val;
    return 0;
  }
  while (isSpace(toAscii(enc, ptr, end)))
    ptr += enc->minBytesPerChar;
  if (ptr != end) {
    *badPtr = ptr;
    return 0;
  }
  return 1;
}

static
int checkCharRefNumber(int result)
{
  switch (result >> 8) {
  case 0xD8: case 0xD9: case 0xDA: case 0xDB:
  case 0xDC: case 0xDD: case 0xDE: case 0xDF:
    return -1;
  case 0:
    if (latin1_encoding.type[result] == BT_NONXML)
      return -1;
    break;
  case 0xFF:
    if (result == 0xFFFE || result == 0xFFFF)
      return -1;
    break;
  }
  return result;
}

int XmlUtf8Encode(int c, char *buf)
{
  enum {
    /* minN is minimum legal resulting value for N byte sequence */
    min2 = 0x80,
    min3 = 0x800,
    min4 = 0x10000
  };

  if (c < 0)
    return 0;
  if (c < min2) {
    buf[0] = (c | UTF8_cval1);
    return 1;
  }
  if (c < min3) {
    buf[0] = ((c >> 6) | UTF8_cval2);
    buf[1] = ((c & 0x3f) | 0x80);
    return 2;
  }
  if (c < min4) {
    buf[0] = ((c >> 12) | UTF8_cval3);
    buf[1] = (((c >> 6) & 0x3f) | 0x80);
    buf[2] = ((c & 0x3f) | 0x80);
    return 3;
  }
  if (c < 0x110000) {
    buf[0] = ((c >> 18) | UTF8_cval4);
    buf[1] = (((c >> 12) & 0x3f) | 0x80);
    buf[2] = (((c >> 6) & 0x3f) | 0x80);
    buf[3] = ((c & 0x3f) | 0x80);
    return 4;
  }
  return 0;
}

int XmlUtf16Encode(int charNum, unsigned short *buf)
{
  if (charNum < 0)
    return 0;
  if (charNum < 0x10000) {
    buf[0] = charNum;
    return 1;
  }
  if (charNum < 0x110000) {
    charNum -= 0x10000;
    buf[0] = (charNum >> 10) + 0xD800;
    buf[1] = (charNum & 0x3FF) + 0xDC00;
    return 2;
  }
  return 0;
}

struct unknown_encoding {
  struct normal_encoding normal;
  int (*convert)(void *userData, const char *p);
  void *userData;
  unsigned short utf16[256];
  char utf8[256][4];
};

int XmlSizeOfUnknownEncoding(void)
{
  return sizeof(struct unknown_encoding);
}

static
int unknown_isName(const ENCODING *enc, const char *p)
{
  int c = ((const struct unknown_encoding *)enc)
	  ->convert(((const struct unknown_encoding *)enc)->userData, p);
  if (c & ~0xFFFF)
    return 0;
  return UCS2_GET_NAMING(namePages, c >> 8, c & 0xFF);
}

static
int unknown_isNmstrt(const ENCODING *enc, const char *p)
{
  int c = ((const struct unknown_encoding *)enc)
	  ->convert(((const struct unknown_encoding *)enc)->userData, p);
  if (c & ~0xFFFF)
    return 0;
  return UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xFF);
}

static
int unknown_isInvalid(const ENCODING *enc, const char *p)
{
  int c = ((const struct unknown_encoding *)enc)
	   ->convert(((const struct unknown_encoding *)enc)->userData, p);
  return (c & ~0xFFFF) || checkCharRefNumber(c) < 0;
}

static
void unknown_toUtf8(const ENCODING *enc,
		    const char **fromP, const char *fromLim,
		    char **toP, const char *toLim)
{
  char buf[XML_UTF8_ENCODE_MAX];
  for (;;) {
    const char *utf8;
    int n;
    if (*fromP == fromLim)
      break;
    utf8 = ((const struct unknown_encoding *)enc)->utf8[(unsigned char)**fromP];
    n = *utf8++;
    if (n == 0) {
      int c = ((const struct unknown_encoding *)enc)
	      ->convert(((const struct unknown_encoding *)enc)->userData, *fromP);
      n = XmlUtf8Encode(c, buf);
      if (n > toLim - *toP)
	break;
      utf8 = buf;
      *fromP += ((const struct normal_encoding *)enc)->type[(unsigned char)**fromP]
	         - (BT_LEAD2 - 2);
    }
    else {
      if (n > toLim - *toP)
	break;
      (*fromP)++;
    }
    do {
      *(*toP)++ = *utf8++;
    } while (--n != 0);
  }
}

static
void unknown_toUtf16(const ENCODING *enc,
		     const char **fromP, const char *fromLim,
		     unsigned short **toP, const unsigned short *toLim)
{
  while (*fromP != fromLim && *toP != toLim) {
    unsigned short c
      = ((const struct unknown_encoding *)enc)->utf16[(unsigned char)**fromP];
    if (c == 0) {
      c = (unsigned short)((const struct unknown_encoding *)enc)
	   ->convert(((const struct unknown_encoding *)enc)->userData, *fromP);
      *fromP += ((const struct normal_encoding *)enc)->type[(unsigned char)**fromP]
	         - (BT_LEAD2 - 2);
    }
    else
      (*fromP)++;
    *(*toP)++ = c;
  }
}

ENCODING *
XmlInitUnknownEncoding(void *mem,
		       int *table,
		       int (*convert)(void *userData, const char *p),
		       void *userData)
{
  int i;
  struct unknown_encoding *e = mem;
  for (i = 0; i < (int)sizeof(struct normal_encoding); i++)
    ((char *)mem)[i] = ((char *)&latin1_encoding)[i];
  for (i = 0; i < 128; i++)
    if (latin1_encoding.type[i] != BT_OTHER
        && latin1_encoding.type[i] != BT_NONXML
	&& table[i] != i)
      return 0;
  for (i = 0; i < 256; i++) {
    int c = table[i];
    if (c == -1) {
      e->normal.type[i] = BT_MALFORM;
      /* This shouldn't really get used. */
      e->utf16[i] = 0xFFFF;
      e->utf8[i][0] = 1;
      e->utf8[i][1] = 0;
    }
    else if (c < 0) {
      if (c < -4)
	return 0;
      e->normal.type[i] = BT_LEAD2 - (c + 2);
      e->utf8[i][0] = 0;
      e->utf16[i] = 0;
    }
    else if (c < 0x80) {
      if (latin1_encoding.type[c] != BT_OTHER
	  && latin1_encoding.type[c] != BT_NONXML
	  && c != i)
	return 0;
      e->normal.type[i] = latin1_encoding.type[c];
      e->utf8[i][0] = 1;
      e->utf8[i][1] = (char)c;
      e->utf16[i] = c == 0 ? 0xFFFF : c;
    }
    else if (checkCharRefNumber(c) < 0) {
      e->normal.type[i] = BT_NONXML;
      /* This shouldn't really get used. */
      e->utf16[i] = 0xFFFF;
      e->utf8[i][0] = 1;
      e->utf8[i][1] = 0;
    }
    else {
      if (c > 0xFFFF)
	return 0;
      if (UCS2_GET_NAMING(nmstrtPages, c >> 8, c & 0xff))
	e->normal.type[i] = BT_NMSTRT;
      else if (UCS2_GET_NAMING(namePages, c >> 8, c & 0xff))
	e->normal.type[i] = BT_NAME;
      else
	e->normal.type[i] = BT_OTHER;
      e->utf8[i][0] = (char)XmlUtf8Encode(c, e->utf8[i] + 1);
      e->utf16[i] = c;
    }
  }
  e->userData = userData;
  e->convert = convert;
  if (convert) {
    e->normal.isName2 = unknown_isName;
    e->normal.isName3 = unknown_isName;
    e->normal.isName4 = unknown_isName;
    e->normal.isNmstrt2 = unknown_isNmstrt;
    e->normal.isNmstrt3 = unknown_isNmstrt;
    e->normal.isNmstrt4 = unknown_isNmstrt;
    e->normal.isInvalid2 = unknown_isInvalid;
    e->normal.isInvalid3 = unknown_isInvalid;
    e->normal.isInvalid4 = unknown_isInvalid;
  }
  e->normal.enc.utf8Convert = unknown_toUtf8;
  e->normal.enc.utf16Convert = unknown_toUtf16;
  return &(e->normal.enc);
}

/* If this enumeration is changed, getEncodingIndex and encodings
must also be changed. */
enum {
  UNKNOWN_ENC = -1,
  ISO_8859_1_ENC = 0,
  US_ASCII_ENC,
  UTF_8_ENC,
  UTF_16_ENC,
  UTF_16BE_ENC,
  UTF_16LE_ENC,
  /* must match encodingNames up to here */
  NO_ENC
};

static const char KW_ISO_8859_1[] = {
  ASCII_I, ASCII_S, ASCII_O, ASCII_MINUS, ASCII_8, ASCII_8, ASCII_5, ASCII_9, ASCII_MINUS, ASCII_1, '\0'
};
static const char KW_US_ASCII[] = {
  ASCII_U, ASCII_S, ASCII_MINUS, ASCII_A, ASCII_S, ASCII_C, ASCII_I, ASCII_I, '\0'
};
static const char KW_UTF_8[] =	{
  ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_8, '\0'
};
static const char KW_UTF_16[] =	{
  ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, '\0'
};
static const char KW_UTF_16BE[] = {
  ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_B, ASCII_E, '\0'
};
static const char KW_UTF_16LE[] = {
  ASCII_U, ASCII_T, ASCII_F, ASCII_MINUS, ASCII_1, ASCII_6, ASCII_L, ASCII_E, '\0'
};

static
int getEncodingIndex(const char *name)
{
  static const char *encodingNames[] = {
    KW_ISO_8859_1,
    KW_US_ASCII,
    KW_UTF_8,
    KW_UTF_16,
    KW_UTF_16BE,
    KW_UTF_16LE,
  };
  int i;
  if (name == 0)
    return NO_ENC;
  for (i = 0; i < (int)(sizeof(encodingNames)/sizeof(encodingNames[0])); i++)
    if (streqci(name, encodingNames[i]))
      return i;
  return UNKNOWN_ENC;
}

/* For binary compatibility, we store the index of the encoding specified
at initialization in the isUtf16 member. */

#define INIT_ENC_INDEX(enc) ((int)(enc)->initEnc.isUtf16)
#define SET_INIT_ENC_INDEX(enc, i) ((enc)->initEnc.isUtf16 = (char)i)

/* This is what detects the encoding.
encodingTable maps from encoding indices to encodings;
INIT_ENC_INDEX(enc) is the index of the external (protocol) specified encoding;
state is XML_CONTENT_STATE if we're parsing an external text entity,
and XML_PROLOG_STATE otherwise.
*/


static
int initScan(const ENCODING **encodingTable,
	     const INIT_ENCODING *enc,
	     int state,
	     const char *ptr,
	     const char *end,
	     const char **nextTokPtr)
{
  const ENCODING **encPtr;

  if (ptr == end)
    return XML_TOK_NONE;
  encPtr = enc->encPtr;
  if (ptr + 1 == end) {
    /* only a single byte available for auto-detection */
#ifndef XML_DTD /* FIXME */
    /* a well-formed document entity must have more than one byte */
    if (state != XML_CONTENT_STATE)
      return XML_TOK_PARTIAL;
#endif
    /* so we're parsing an external text entity... */
    /* if UTF-16 was externally specified, then we need at least 2 bytes */
    switch (INIT_ENC_INDEX(enc)) {
    case UTF_16_ENC:
    case UTF_16LE_ENC:
    case UTF_16BE_ENC:
      return XML_TOK_PARTIAL;
    }
    switch ((unsigned char)*ptr) {
    case 0xFE:
    case 0xFF:
    case 0xEF: /* possibly first byte of UTF-8 BOM */
      if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
	  && state == XML_CONTENT_STATE)
	break;
      /* fall through */
    case 0x00:
    case 0x3C:
      return XML_TOK_PARTIAL;
    }
  }
  else {
    switch (((unsigned char)ptr[0] << 8) | (unsigned char)ptr[1]) {
    case 0xFEFF:
      if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
	  && state == XML_CONTENT_STATE)
	break;
      *nextTokPtr = ptr + 2;
      *encPtr = encodingTable[UTF_16BE_ENC];
      return XML_TOK_BOM;
    /* 00 3C is handled in the default case */
    case 0x3C00:
      if ((INIT_ENC_INDEX(enc) == UTF_16BE_ENC
	   || INIT_ENC_INDEX(enc) == UTF_16_ENC)
	  && state == XML_CONTENT_STATE)
	break;
      *encPtr = encodingTable[UTF_16LE_ENC];
      return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
    case 0xFFFE:
      if (INIT_ENC_INDEX(enc) == ISO_8859_1_ENC
	  && state == XML_CONTENT_STATE)
	break;
      *nextTokPtr = ptr + 2;
      *encPtr = encodingTable[UTF_16LE_ENC];
      return XML_TOK_BOM;
    case 0xEFBB:
      /* Maybe a UTF-8 BOM (EF BB BF) */
      /* If there's an explicitly specified (external) encoding
         of ISO-8859-1 or some flavour of UTF-16
         and this is an external text entity,
	 don't look for the BOM,
         because it might be a legal data. */
      if (state == XML_CONTENT_STATE) {
	int e = INIT_ENC_INDEX(enc);
	if (e == ISO_8859_1_ENC || e == UTF_16BE_ENC || e == UTF_16LE_ENC || e == UTF_16_ENC)
	  break;
      }
      if (ptr + 2 == end)
	return XML_TOK_PARTIAL;
      if ((unsigned char)ptr[2] == 0xBF) {
	*nextTokPtr = ptr + 3;
	*encPtr = encodingTable[UTF_8_ENC];
	return XML_TOK_BOM;
      }
      break;
    default:
      if (ptr[0] == '\0') {
	/* 0 isn't a legal data character. Furthermore a document entity can only
	   start with ASCII characters.  So the only way this can fail to be big-endian
	   UTF-16 if it it's an external parsed general entity that's labelled as
	   UTF-16LE. */
	if (state == XML_CONTENT_STATE && INIT_ENC_INDEX(enc) == UTF_16LE_ENC)
	  break;
	*encPtr = encodingTable[UTF_16BE_ENC];
	return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
      }
      else if (ptr[1] == '\0') {
	/* We could recover here in the case:
	    - parsing an external entity
	    - second byte is 0
	    - no externally specified encoding
	    - no encoding declaration
	   by assuming UTF-16LE.  But we don't, because this would mean when
	   presented just with a single byte, we couldn't reliably determine
	   whether we needed further bytes. */
	if (state == XML_CONTENT_STATE)
	  break;
	*encPtr = encodingTable[UTF_16LE_ENC];
	return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
      }
      break;
    }
  }
  *encPtr = encodingTable[INIT_ENC_INDEX(enc)];
  return XmlTok(*encPtr, state, ptr, end, nextTokPtr);
}


#define NS(x) x
#define ns(x) x
#include "xmltok_ns.c"
#undef NS
#undef ns

#ifdef XML_NS

#define NS(x) x ## NS
#define ns(x) x ## _ns

#include "xmltok_ns.c"

#undef NS
#undef ns

ENCODING *
XmlInitUnknownEncodingNS(void *mem,
		         int *table,
		         int (*convert)(void *userData, const char *p),
		         void *userData)
{
  ENCODING *enc = XmlInitUnknownEncoding(mem, table, convert, userData);
  if (enc)
    ((struct normal_encoding *)enc)->type[ASCII_COLON] = BT_COLON;
  return enc;
}

#endif /* XML_NS */