xmlparse.cpp

C++ class libraries for network-centric, portable applications, integrated perfectly with the C++ St

  if (eventEndPtr && eventPtr)
    return (int)(eventEndPtr - eventPtr);
  return 0;
}

const char * XMLCALL
XML_GetInputContext(XML_Parser parser, int *offset, int *size)
{
#ifdef XML_CONTEXT_BYTES
  if (eventPtr && buffer) {
    *offset = (int)(eventPtr - buffer);
    *size   = (int)(bufferEnd - buffer);
    return buffer;
  }
#endif /* defined XML_CONTEXT_BYTES */
  return (char *) 0;
}

XML_Size XMLCALL
XML_GetCurrentLineNumber(XML_Parser parser)
{
  if (eventPtr && eventPtr >= positionPtr) {
    XmlUpdatePosition(encoding, positionPtr, eventPtr, &position);
    positionPtr = eventPtr;
  }
  return position.lineNumber + 1;
}

XML_Size XMLCALL
XML_GetCurrentColumnNumber(XML_Parser parser)
{
  if (eventPtr && eventPtr >= positionPtr) {
    XmlUpdatePosition(encoding, positionPtr, eventPtr, &position);
    positionPtr = eventPtr;
  }
  return position.columnNumber;
}

void XMLCALL
XML_FreeContentModel(XML_Parser parser, XML_Content *model)
{
  FREE(model);
}

void * XMLCALL
XML_MemMalloc(XML_Parser parser, size_t size)
{
  return MALLOC(size);
}

void * XMLCALL
XML_MemRealloc(XML_Parser parser, void *ptr, size_t size)
{
  return REALLOC(ptr, size);
}

void XMLCALL
XML_MemFree(XML_Parser parser, void *ptr)
{
  FREE(ptr);
}

void XMLCALL
XML_DefaultCurrent(XML_Parser parser)
{
  if (defaultHandler) {
    if (openInternalEntities)
      reportDefault(parser,
                    internalEncoding,
                    openInternalEntities->internalEventPtr,
                    openInternalEntities->internalEventEndPtr);
    else
      reportDefault(parser, encoding, eventPtr, eventEndPtr);
  }
}

const XML_LChar * XMLCALL
XML_ErrorString(enum XML_Error code)
{
  static const XML_LChar* const message[] = {
    0,
    XML_L("out of memory"),
    XML_L("syntax error"),
    XML_L("no element found"),
    XML_L("not well-formed (invalid token)"),
    XML_L("unclosed token"),
    XML_L("partial character"),
    XML_L("mismatched tag"),
    XML_L("duplicate attribute"),
    XML_L("junk after document element"),
    XML_L("illegal parameter entity reference"),
    XML_L("undefined entity"),
    XML_L("recursive entity reference"),
    XML_L("asynchronous entity"),
    XML_L("reference to invalid character number"),
    XML_L("reference to binary entity"),
    XML_L("reference to external entity in attribute"),
    XML_L("XML or text declaration not at start of entity"),
    XML_L("unknown encoding"),
    XML_L("encoding specified in XML declaration is incorrect"),
    XML_L("unclosed CDATA section"),
    XML_L("error in processing external entity reference"),
    XML_L("document is not standalone"),
    XML_L("unexpected parser state - please send a bug report"),
    XML_L("entity declared in parameter entity"),
    XML_L("requested feature requires XML_DTD support in Expat"),
    XML_L("cannot change setting once parsing has begun"),
    XML_L("unbound prefix"),
    XML_L("must not undeclare prefix"),
    XML_L("incomplete markup in parameter entity"),
    XML_L("XML declaration not well-formed"),
    XML_L("text declaration not well-formed"),
    XML_L("illegal character(s) in public id"),
    XML_L("parser suspended"),
    XML_L("parser not suspended"),
    XML_L("parsing aborted"),
    XML_L("parsing finished"),
    XML_L("cannot suspend in external parameter entity"),
    XML_L("reserved prefix (xml) must not be undeclared or bound to another namespace name"),
    XML_L("reserved prefix (xmlns) must not be declared or undeclared"),
    XML_L("prefix must not be bound to one of the reserved namespace names")
  };
  if (code > 0 && code < sizeof(message)/sizeof(message[0]))
    return message[code];
  return NULL;
}

const XML_LChar * XMLCALL
XML_ExpatVersion(void) {

  /* V1 is used to string-ize the version number. However, it would
     string-ize the actual version macro *names* unless we get them
     substituted before being passed to V1. CPP is defined to expand
     a macro, then rescan for more expansions. Thus, we use V2 to expand
     the version macros, then CPP will expand the resulting V1() macro
     with the correct numerals. */
  /* ### I'm assuming cpp is portable in this respect... */

#define V1(a,b,c) XML_L(#a)XML_L(".")XML_L(#b)XML_L(".")XML_L(#c)
#define V2(a,b,c) XML_L("expat_")V1(a,b,c)

  return V2(XML_MAJOR_VERSION, XML_MINOR_VERSION, XML_MICRO_VERSION);

#undef V1
#undef V2
}

XML_Expat_Version XMLCALL
XML_ExpatVersionInfo(void)
{
  XML_Expat_Version version;

  version.major = XML_MAJOR_VERSION;
  version.minor = XML_MINOR_VERSION;
  version.micro = XML_MICRO_VERSION;

  return version;
}

const XML_Feature * XMLCALL
XML_GetFeatureList(void)
{
  static const XML_Feature features[] = {
    {XML_FEATURE_SIZEOF_XML_CHAR,  XML_L("sizeof(XML_Char)"),
     sizeof(XML_Char)},
    {XML_FEATURE_SIZEOF_XML_LCHAR, XML_L("sizeof(XML_LChar)"),
     sizeof(XML_LChar)},
#ifdef XML_UNICODE
    {XML_FEATURE_UNICODE,          XML_L("XML_UNICODE"), 0},
#endif
#ifdef XML_UNICODE_WCHAR_T
    {XML_FEATURE_UNICODE_WCHAR_T,  XML_L("XML_UNICODE_WCHAR_T"), 0},
#endif
#ifdef XML_DTD
    {XML_FEATURE_DTD,              XML_L("XML_DTD"), 0},
#endif
#ifdef XML_CONTEXT_BYTES
    {XML_FEATURE_CONTEXT_BYTES,    XML_L("XML_CONTEXT_BYTES"),
     XML_CONTEXT_BYTES},
#endif
#ifdef XML_MIN_SIZE
    {XML_FEATURE_MIN_SIZE,         XML_L("XML_MIN_SIZE"), 0},
#endif
#ifdef XML_NS
    {XML_FEATURE_NS,               XML_L("XML_NS"), 0},
#endif
    {XML_FEATURE_END,              NULL, 0}
  };

  return features;
}

/* Initially tag->rawName always points into the parse buffer;
   for those TAG instances opened while the current parse buffer was
   processed, and not yet closed, we need to store tag->rawName in a more
   permanent location, since the parse buffer is about to be discarded.
*/
static XML_Bool
storeRawNames(XML_Parser parser)
{
  TAG *tag = tagStack;
  while (tag) {
    int bufSize;
    int nameLen = sizeof(XML_Char) * (tag->name.strLen + 1);
    char *rawNameBuf = tag->buf + nameLen;
    /* Stop if already stored.  Since tagStack is a stack, we can stop
       at the first entry that has already been copied; everything
       below it in the stack is already been accounted for in a
       previous call to this function.
    */
    if (tag->rawName == rawNameBuf)
      break;
    /* For re-use purposes we need to ensure that the
       size of tag->buf is a multiple of sizeof(XML_Char).
    */
    bufSize = nameLen + ROUND_UP(tag->rawNameLength, sizeof(XML_Char));
    if (bufSize > tag->bufEnd - tag->buf) {
      char *temp = (char *)REALLOC(tag->buf, bufSize);
      if (temp == NULL)
        return XML_FALSE;
      /* if tag->name.str points to tag->buf (only when namespace
         processing is off) then we have to update it
      */
      if (tag->name.str == (XML_Char *)tag->buf)
        tag->name.str = (XML_Char *)temp;
      /* if tag->name.localPart is set (when namespace processing is on)
         then update it as well, since it will always point into tag->buf
      */
      if (tag->name.localPart)
        tag->name.localPart = (XML_Char *)temp + (tag->name.localPart -
                                                  (XML_Char *)tag->buf);
      tag->buf = temp;
      tag->bufEnd = temp + bufSize;
      rawNameBuf = temp + nameLen;
    }
    memcpy(rawNameBuf, tag->rawName, tag->rawNameLength);
    tag->rawName = rawNameBuf;
    tag = tag->parent;
  }
  return XML_TRUE;
}

static enum XML_Error PTRCALL
contentProcessor(XML_Parser parser,
                 const char *start,
                 const char *end,
                 const char **endPtr)
{
  enum XML_Error result = doContent(parser, 0, encoding, start, end, 
                                    endPtr, (XML_Bool)!ps_finalBuffer);
  if (result == XML_ERROR_NONE) {
    if (!storeRawNames(parser))
      return XML_ERROR_NO_MEMORY;
  }
  return result;
}

static enum XML_Error PTRCALL
externalEntityInitProcessor(XML_Parser parser,
                            const char *start,
                            const char *end,
                            const char **endPtr)
{
  enum XML_Error result = initializeEncoding(parser);
  if (result != XML_ERROR_NONE)
    return result;
  processor = externalEntityInitProcessor2;
  return externalEntityInitProcessor2(parser, start, end, endPtr);
}

static enum XML_Error PTRCALL
externalEntityInitProcessor2(XML_Parser parser,
                             const char *start,
                             const char *end,
                             const char **endPtr)
{
  const char *next = start; /* XmlContentTok doesn't always set the last arg */
  int tok = XmlContentTok(encoding, start, end, &next);
  switch (tok) {
  case XML_TOK_BOM:
    /* If we are at the end of the buffer, this would cause the next stage,
       i.e. externalEntityInitProcessor3, to pass control directly to
       doContent (by detecting XML_TOK_NONE) without processing any xml text
       declaration - causing the error XML_ERROR_MISPLACED_XML_PI in doContent.
    */
    if (next == end && !ps_finalBuffer) {
      *endPtr = next;
      return XML_ERROR_NONE;
    }
    start = next;
    break;
  case XML_TOK_PARTIAL:
    if (!ps_finalBuffer) {
      *endPtr = start;
      return XML_ERROR_NONE;
    }
    eventPtr = start;
    return XML_ERROR_UNCLOSED_TOKEN;
  case XML_TOK_PARTIAL_CHAR:
    if (!ps_finalBuffer) {
      *endPtr = start;
      return XML_ERROR_NONE;
    }
    eventPtr = start;
    return XML_ERROR_PARTIAL_CHAR;
  }
  processor = externalEntityInitProcessor3;
  return externalEntityInitProcessor3(parser, start, end, endPtr);
}

static enum XML_Error PTRCALL
externalEntityInitProcessor3(XML_Parser parser,
                             const char *start,
                             const char *end,
                             const char **endPtr)
{
  int tok;
  const char *next = start; /* XmlContentTok doesn't always set the last arg */
  eventPtr = start;
  tok = XmlContentTok(encoding, start, end, &next);
  eventEndPtr = next;

  switch (tok) {
  case XML_TOK_XML_DECL:
    {
      enum XML_Error result;
      result = processXmlDecl(parser, 1, start, next);
      if (result != XML_ERROR_NONE)
        return result;
      switch (ps_parsing) {
      case XML_SUSPENDED: 
        *endPtr = next;
        return XML_ERROR_NONE;
      case XML_FINISHED:
        return XML_ERROR_ABORTED;
      default:
        start = next;
      }
    }
    break;
  case XML_TOK_PARTIAL:
    if (!ps_finalBuffer) {
      *endPtr = start;
      return XML_ERROR_NONE;
    }
    return XML_ERROR_UNCLOSED_TOKEN;
  case XML_TOK_PARTIAL_CHAR:
    if (!ps_finalBuffer) {
      *endPtr = start;
      return XML_ERROR_NONE;
    }
    return XML_ERROR_PARTIAL_CHAR;
  }
  processor = externalEntityContentProcessor;
  tagLevel = 1;
  return externalEntityContentProcessor(parser, start, end, endPtr);
}

static enum XML_Error PTRCALL
externalEntityContentProcessor(XML_Parser parser,
                               const char *start,
                               const char *end,
                               const char **endPtr)
{
  enum XML_Error result = doContent(parser, 1, encoding, start, end, 
                                    endPtr, (XML_Bool)!ps_finalBuffer);
  if (result == XML_ERROR_NONE) {
    if (!storeRawNames(parser))
      return XML_ERROR_NO_MEMORY;
  }
  return result;
}

static enum XML_Error
doContent(XML_Parser parser,
          int startTagLevel,
          const ENCODING *enc,
          const char *s,
          const char *end,
          const char **nextPtr,
          XML_Bool haveMore)
{
  /* save one level of indirection */
  DTD * const dtd = _dtd;  

  const char **eventPP;
  const char **eventEndPP;
  if (enc == encoding) {
    eventPP = &eventPtr;
    eventEndPP = &eventEndPtr;
  }
  else {
    eventPP = &(openInternalEntities->internalEventPtr);
    eventEndPP = &(openInternalEntities->internalEventEndPtr);
  }
  *eventPP = s;

  for (;;) {
    const char *next = s; /* XmlContentTok doesn't always set the last arg */
    int tok = XmlContentTok(enc, s, end, &next);
    *eventEndPP = next;
    switch (tok) {
    case XML_TOK_TRAILING_CR:
      if (haveMore) {
        *nextPtr = s;
        return XML_ERROR_NONE;
      }
      *eventEndPP = end;
      if (characterDataHandler) {
        XML_Char c = 0xA;
        characterDataHandler(handlerArg, &c, 1);
      }
      else if (defaultHandler)
        reportDefault(parser, enc, s, end);
      /* We are at the end of the final buffer, should we check for 
         XML_SUSPENDED, XML_FINISHED? 
      */
      if (startTagLevel == 0)
        return XML_ERROR_NO_ELEMENTS;
      if (tagLevel != startTagLevel)
        return XML_ERROR_ASYNC_ENTITY;
      *nextPtr = end;
      return XML_ERROR_NONE;
    case XML_TOK_NONE:
      if (haveMore) {
        *nextPtr = s;
        return XML_ERROR_NONE;
      }
      if (startTagLevel > 0) {
        if (tagLevel != startTagLevel)
          return XML_ERROR_ASYNC_ENTITY;
        *nextPtr = s;
        return XML_ERROR_NONE;
      }
      return XML_ERROR_NO_ELEMENTS;
    case XML_TOK_INVALID:
      *eventPP = next;
      return XML_ERROR_INVALID_TOKEN;
    case XML_TOK_PARTIAL:
      if (haveMore)