xmltok_impl.c

很牛的GUI源码wxWidgets-2.8.0.zip 可在多种平台下运行.

      *nextTokPtr = ptr + MINBPC(enc);
      return XML_TOK_NAME_QUESTION;
    default:
      *nextTokPtr = ptr;
      return XML_TOK_INVALID;
    }
  }
  return -tok;
}

static int PTRCALL
PREFIX(attributeValueTok)(const ENCODING *enc, const char *ptr,
                          const char *end, const char **nextTokPtr)
{
  const char *start;
  if (ptr == end)
    return XML_TOK_NONE;
  start = ptr;
  while (ptr != end) {
    switch (BYTE_TYPE(enc, ptr)) {
#define LEAD_CASE(n) \
    case BT_LEAD ## n: ptr += n; break;
    LEAD_CASE(2) LEAD_CASE(3) LEAD_CASE(4)
#undef LEAD_CASE
    case BT_AMP:
      if (ptr == start)
        return PREFIX(scanRef)(enc, ptr + MINBPC(enc), end, nextTokPtr);
      *nextTokPtr = ptr;
      return XML_TOK_DATA_CHARS;
    case BT_LT:
      /* this is for inside entity references */
      *nextTokPtr = ptr;
      return XML_TOK_INVALID;
    case BT_LF:
      if (ptr == start) {
        *nextTokPtr = ptr + MINBPC(enc);
        return XML_TOK_DATA_NEWLINE;
      }
      *nextTokPtr = ptr;
      return XML_TOK_DATA_CHARS;
    case BT_CR:
      if (ptr == start) {
        ptr += MINBPC(enc);
        if (ptr == end)
          return XML_TOK_TRAILING_CR;
        if (BYTE_TYPE(enc, ptr) == BT_LF)
          ptr += MINBPC(enc);
        *nextTokPtr = ptr;
        return XML_TOK_DATA_NEWLINE;
      }
      *nextTokPtr = ptr;
      return XML_TOK_DATA_CHARS;
    case BT_S:
      if (ptr == start) {
        *nextTokPtr = ptr + MINBPC(enc);
        return XML_TOK_ATTRIBUTE_VALUE_S;
      }
      *nextTokPtr = ptr;
      return XML_TOK_DATA_CHARS;
    default:
      ptr += MINBPC(enc);
      break;
    }
  }
  *nextTokPtr = ptr;
  return XML_TOK_DATA_CHARS;
}

static int PTRCALL
PREFIX(entityValueTok)(const ENCODING *enc, const char *ptr,
                       const char *end, const char **nextTokPtr)
{
  const char *start;
  if (ptr == end)
    return XML_TOK_NONE;
  start = ptr;
  while (ptr != end) {
    switch (BYTE_TYPE(enc, ptr)) {
#define LEAD_CASE(n) \
    case BT_LEAD ## n: ptr += n; break;
    LEAD_CASE(2) LEAD_CASE(3) LEAD_CASE(4)
#undef LEAD_CASE
    case BT_AMP:
      if (ptr == start)
        return PREFIX(scanRef)(enc, ptr + MINBPC(enc), end, nextTokPtr);
      *nextTokPtr = ptr;
      return XML_TOK_DATA_CHARS;
    case BT_PERCNT:
      if (ptr == start) {
        int tok =  PREFIX(scanPercent)(enc, ptr + MINBPC(enc),
                                       end, nextTokPtr);
        return (tok == XML_TOK_PERCENT) ? XML_TOK_INVALID : tok;
      }
      *nextTokPtr = ptr;
      return XML_TOK_DATA_CHARS;
    case BT_LF:
      if (ptr == start) {
        *nextTokPtr = ptr + MINBPC(enc);
        return XML_TOK_DATA_NEWLINE;
      }
      *nextTokPtr = ptr;
      return XML_TOK_DATA_CHARS;
    case BT_CR:
      if (ptr == start) {
        ptr += MINBPC(enc);
        if (ptr == end)
          return XML_TOK_TRAILING_CR;
        if (BYTE_TYPE(enc, ptr) == BT_LF)
          ptr += MINBPC(enc);
        *nextTokPtr = ptr;
        return XML_TOK_DATA_NEWLINE;
      }
      *nextTokPtr = ptr;
      return XML_TOK_DATA_CHARS;
    default:
      ptr += MINBPC(enc);
      break;
    }
  }
  *nextTokPtr = ptr;
  return XML_TOK_DATA_CHARS;
}

#ifdef XML_DTD

static int PTRCALL
PREFIX(ignoreSectionTok)(const ENCODING *enc, const char *ptr,
                         const char *end, const char **nextTokPtr)
{
  int level = 0;
  if (MINBPC(enc) > 1) {
    size_t n = end - ptr;
    if (n & (MINBPC(enc) - 1)) {
      n &= ~(MINBPC(enc) - 1);
      end = ptr + n;
    }
  }
  while (ptr != end) {
    switch (BYTE_TYPE(enc, ptr)) {
    INVALID_CASES(ptr, nextTokPtr)
    case BT_LT:
      if ((ptr += MINBPC(enc)) == end)
        return XML_TOK_PARTIAL;
      if (CHAR_MATCHES(enc, ptr, ASCII_EXCL)) {
        if ((ptr += MINBPC(enc)) == end)
          return XML_TOK_PARTIAL;
        if (CHAR_MATCHES(enc, ptr, ASCII_LSQB)) {
          ++level;
          ptr += MINBPC(enc);
        }
      }
      break;
    case BT_RSQB:
      if ((ptr += MINBPC(enc)) == end)
        return XML_TOK_PARTIAL;
      if (CHAR_MATCHES(enc, ptr, ASCII_RSQB)) {
        if ((ptr += MINBPC(enc)) == end)
          return XML_TOK_PARTIAL;
        if (CHAR_MATCHES(enc, ptr, ASCII_GT)) {
          ptr += MINBPC(enc);
          if (level == 0) {
            *nextTokPtr = ptr;
            return XML_TOK_IGNORE_SECT;
          }
          --level;
        }
      }
      break;
    default:
      ptr += MINBPC(enc);
      break;
    }
  }
  return XML_TOK_PARTIAL;
}

#endif /* XML_DTD */

static int PTRCALL
PREFIX(isPublicId)(const ENCODING *enc, const char *ptr, const char *end,
                   const char **badPtr)
{
  ptr += MINBPC(enc);
  end -= MINBPC(enc);
  for (; ptr != end; ptr += MINBPC(enc)) {
    switch (BYTE_TYPE(enc, ptr)) {
    case BT_DIGIT:
    case BT_HEX:
    case BT_MINUS:
    case BT_APOS:
    case BT_LPAR:
    case BT_RPAR:
    case BT_PLUS:
    case BT_COMMA:
    case BT_SOL:
    case BT_EQUALS:
    case BT_QUEST:
    case BT_CR:
    case BT_LF:
    case BT_SEMI:
    case BT_EXCL:
    case BT_AST:
    case BT_PERCNT:
    case BT_NUM:
#ifdef XML_NS
    case BT_COLON:
#endif
      break;
    case BT_S:
      if (CHAR_MATCHES(enc, ptr, ASCII_TAB)) {
        *badPtr = ptr;
        return 0;
      }
      break;
    case BT_NAME:
    case BT_NMSTRT:
      if (!(BYTE_TO_ASCII(enc, ptr) & ~0x7f))
        break;
    default:
      switch (BYTE_TO_ASCII(enc, ptr)) {
      case 0x24: /* $ */
      case 0x40: /* @ */
        break;
      default:
        *badPtr = ptr;
        return 0;
      }
      break;
    }
  }
  return 1;
}

/* This must only be called for a well-formed start-tag or empty
   element tag.  Returns the number of attributes.  Pointers to the
   first attsMax attributes are stored in atts.
*/

static int PTRCALL
PREFIX(getAtts)(const ENCODING *enc, const char *ptr,
                int attsMax, ATTRIBUTE *atts)
{
  enum { other, inName, inValue } state = inName;
  int nAtts = 0;
  int open = 0; /* defined when state == inValue;
                   initialization just to shut up compilers */

  for (ptr += MINBPC(enc);; ptr += MINBPC(enc)) {
    switch (BYTE_TYPE(enc, ptr)) {
#define START_NAME \
      if (state == other) { \
        if (nAtts < attsMax) { \
          atts[nAtts].name = ptr; \
          atts[nAtts].normalized = 1; \
        } \
        state = inName; \
      }
#define LEAD_CASE(n) \
    case BT_LEAD ## n: START_NAME ptr += (n - MINBPC(enc)); break;
    LEAD_CASE(2) LEAD_CASE(3) LEAD_CASE(4)
#undef LEAD_CASE
    case BT_NONASCII:
    case BT_NMSTRT:
    case BT_HEX:
      START_NAME
      break;
#undef START_NAME
    case BT_QUOT:
      if (state != inValue) {
        if (nAtts < attsMax)
          atts[nAtts].valuePtr = ptr + MINBPC(enc);
        state = inValue;
        open = BT_QUOT;
      }
      else if (open == BT_QUOT) {
        state = other;
        if (nAtts < attsMax)
          atts[nAtts].valueEnd = ptr;
        nAtts++;
      }
      break;
    case BT_APOS:
      if (state != inValue) {
        if (nAtts < attsMax)
          atts[nAtts].valuePtr = ptr + MINBPC(enc);
        state = inValue;
        open = BT_APOS;
      }
      else if (open == BT_APOS) {
        state = other;
        if (nAtts < attsMax)
          atts[nAtts].valueEnd = ptr;
        nAtts++;
      }
      break;
    case BT_AMP:
      if (nAtts < attsMax)
        atts[nAtts].normalized = 0;
      break;
    case BT_S:
      if (state == inName)
        state = other;
      else if (state == inValue
               && nAtts < attsMax
               && atts[nAtts].normalized
               && (ptr == atts[nAtts].valuePtr
                   || BYTE_TO_ASCII(enc, ptr) != ASCII_SPACE
                   || BYTE_TO_ASCII(enc, ptr + MINBPC(enc)) == ASCII_SPACE
                   || BYTE_TYPE(enc, ptr + MINBPC(enc)) == open))
        atts[nAtts].normalized = 0;
      break;
    case BT_CR: case BT_LF:
      /* This case ensures that the first attribute name is counted
         Apart from that we could just change state on the quote. */
      if (state == inName)
        state = other;
      else if (state == inValue && nAtts < attsMax)
        atts[nAtts].normalized = 0;
      break;
    case BT_GT:
    case BT_SOL:
      if (state != inValue)
        return nAtts;
      break;
    default:
      break;
    }
  }
  /* not reached */
}

static int PTRFASTCALL
PREFIX(charRefNumber)(const ENCODING *enc, const char *ptr)
{
  int result = 0;
  /* skip &# */
  ptr += 2*MINBPC(enc);
  if (CHAR_MATCHES(enc, ptr, ASCII_x)) {
    for (ptr += MINBPC(enc);
         !CHAR_MATCHES(enc, ptr, ASCII_SEMI);
         ptr += MINBPC(enc)) {
      int c = BYTE_TO_ASCII(enc, ptr);
      switch (c) {
      case ASCII_0: case ASCII_1: case ASCII_2: case ASCII_3: case ASCII_4:
      case ASCII_5: case ASCII_6: case ASCII_7: case ASCII_8: case ASCII_9:
        result <<= 4;
        result |= (c - ASCII_0);
        break;
      case ASCII_A: case ASCII_B: case ASCII_C:
      case ASCII_D: case ASCII_E: case ASCII_F:
        result <<= 4;
        result += 10 + (c - ASCII_A);
        break;
      case ASCII_a: case ASCII_b: case ASCII_c:
      case ASCII_d: case ASCII_e: case ASCII_f:
        result <<= 4;
        result += 10 + (c - ASCII_a);
        break;
      }
      if (result >= 0x110000)
        return -1;
    }
  }
  else {
    for (; !CHAR_MATCHES(enc, ptr, ASCII_SEMI); ptr += MINBPC(enc)) {
      int c = BYTE_TO_ASCII(enc, ptr);
      result *= 10;
      result += (c - ASCII_0);
      if (result >= 0x110000)
        return -1;
    }
  }
  return checkCharRefNumber(result);
}

static int PTRCALL
PREFIX(predefinedEntityName)(const ENCODING *enc, const char *ptr,
                             const char *end)
{
  switch ((end - ptr)/MINBPC(enc)) {
  case 2:
    if (CHAR_MATCHES(enc, ptr + MINBPC(enc), ASCII_t)) {
      switch (BYTE_TO_ASCII(enc, ptr)) {
      case ASCII_l:
        return ASCII_LT;
      case ASCII_g:
        return ASCII_GT;
      }
    }
    break;
  case 3:
    if (CHAR_MATCHES(enc, ptr, ASCII_a)) {
      ptr += MINBPC(enc);
      if (CHAR_MATCHES(enc, ptr, ASCII_m)) {
        ptr += MINBPC(enc);
        if (CHAR_MATCHES(enc, ptr, ASCII_p))
          return ASCII_AMP;
      }
    }
    break;
  case 4:
    switch (BYTE_TO_ASCII(enc, ptr)) {
    case ASCII_q:
      ptr += MINBPC(enc);
      if (CHAR_MATCHES(enc, ptr, ASCII_u)) {
        ptr += MINBPC(enc);
        if (CHAR_MATCHES(enc, ptr, ASCII_o)) {
          ptr += MINBPC(enc);
          if (CHAR_MATCHES(enc, ptr, ASCII_t))
            return ASCII_QUOT;
        }
      }
      break;
    case ASCII_a:
      ptr += MINBPC(enc);
      if (CHAR_MATCHES(enc, ptr, ASCII_p)) {
        ptr += MINBPC(enc);
        if (CHAR_MATCHES(enc, ptr, ASCII_o)) {
          ptr += MINBPC(enc);
          if (CHAR_MATCHES(enc, ptr, ASCII_s))
            return ASCII_APOS;
        }
      }
      break;
    }
  }
  return 0;
}

static int PTRCALL
PREFIX(sameName)(const ENCODING *enc, const char *ptr1, const char *ptr2)
{
  for (;;) {
    switch (BYTE_TYPE(enc, ptr1)) {
#define LEAD_CASE(n) \
    case BT_LEAD ## n: \
      if (*ptr1++ != *ptr2++) \
        return 0;
    LEAD_CASE(4) LEAD_CASE(3) LEAD_CASE(2)
#undef LEAD_CASE
      /* fall through */
      if (*ptr1++ != *ptr2++)
        return 0;
      break;
    case BT_NONASCII:
    case BT_NMSTRT:
#ifdef XML_NS
    case BT_COLON:
#endif
    case BT_HEX:
    case BT_DIGIT:
    case BT_NAME:
    case BT_MINUS:
      if (*ptr2++ != *ptr1++)
        return 0;
      if (MINBPC(enc) > 1) {
        if (*ptr2++ != *ptr1++)
          return 0;
        if (MINBPC(enc) > 2) {
          if (*ptr2++ != *ptr1++)
            return 0;
          if (MINBPC(enc) > 3) {
            if (*ptr2++ != *ptr1++)
              return 0;
          }
        }
      }
      break;
    default:
      if (MINBPC(enc) == 1 && *ptr1 == *ptr2)
        return 1;
      switch (BYTE_TYPE(enc, ptr2)) {
      case BT_LEAD2:
      case BT_LEAD3:
      case BT_LEAD4:
      case BT_NONASCII:
      case BT_NMSTRT:
#ifdef XML_NS
      case BT_COLON:
#endif
      case BT_HEX:
      case BT_DIGIT:
      case BT_NAME:
      case BT_MINUS:
        return 0;
      default:
        return 1;
      }
    }
  }
  /* not reached */
}

static int PTRCALL
PREFIX(nameMatchesAscii)(const ENCODING *enc, const char *ptr1,
                         const char *end1, const char *ptr2)
{
  for (; *ptr2; ptr1 += MINBPC(enc), ptr2++) {
    if (ptr1 == end1)
      return 0;
    if (!CHAR_MATCHES(enc, ptr1, *ptr2))
      return 0;
  }
  return ptr1 == end1;
}

static int PTRFASTCALL
PREFIX(nameLength)(const ENCODING *enc, const char *ptr)
{
  const char *start = ptr;
  for (;;) {
    switch (BYTE_TYPE(enc, ptr)) {
#define LEAD_CASE(n) \
    case BT_LEAD ## n: ptr += n; break;
    LEAD_CASE(2) LEAD_CASE(3) LEAD_CASE(4)
#undef LEAD_CASE
    case BT_NONASCII:
    case BT_NMSTRT:
#ifdef XML_NS
    case BT_COLON:
#endif
    case BT_HEX:
    case BT_DIGIT:
    case BT_NAME:
    case BT_MINUS:
      ptr += MINBPC(enc);
      break;
    default:
      return ptr - start;
    }
  }
}

static const char * PTRFASTCALL
PREFIX(skipS)(const ENCODING *enc, const char *ptr)
{
  for (;;) {
    switch (BYTE_TYPE(enc, ptr)) {
    case BT_LF:
    case BT_CR:
    case BT_S:
      ptr += MINBPC(enc);
      break;
    default:
      return ptr;
    }
  }
}

static void PTRCALL
PREFIX(updatePosition)(const ENCODING *enc,
                       const char *ptr,
                       const char *end,
                       POSITION *pos)
{
  while (ptr != end) {
    switch (BYTE_TYPE(enc, ptr)) {
#define LEAD_CASE(n) \
    case BT_LEAD ## n: \
      ptr += n; \
      break;
    LEAD_CASE(2) LEAD_CASE(3) LEAD_CASE(4)
#undef LEAD_CASE
    case BT_LF:
      pos->columnNumber = (unsigned)-1;
      pos->lineNumber++;
      ptr += MINBPC(enc);
      break;
    case BT_CR:
      pos->lineNumber++;
      ptr += MINBPC(enc);
      if (ptr != end && BYTE_TYPE(enc, ptr) == BT_LF)
        ptr += MINBPC(enc);
      pos->columnNumber = (unsigned)-1;
      break;
    default:
      ptr += MINBPC(enc);
      break;
    }
    pos->columnNumber++;
  }
}

#undef DO_LEAD_CASE
#undef MULTIBYTE_CASES
#undef INVALID_CASES
#undef CHECK_NAME_CASE
#undef CHECK_NAME_CASES
#undef CHECK_NMSTRT_CASE
#undef CHECK_NMSTRT_CASES