const char* TiXmlElement::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
{
p = SkipWhiteSpace( p, encoding );
TiXmlDocument* document = GetDocument();
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, 0, 0, encoding );
return 0;
}
if ( data )
{
data->Stamp( p, encoding );
location = data->Cursor();
}
if ( *p != '<' )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, p, data, encoding );
return 0;
}
p = SkipWhiteSpace( p+1, encoding );
// Read the name.
const char* pErr = p;
p = ReadName( p, &value, encoding );
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_FAILED_TO_READ_ELEMENT_NAME, pErr, data, encoding );
return 0;
}
TIXML_STRING endTag ("</");
endTag += value;
// Check for and read attributes. Also look for an empty
// tag or an end tag.
while ( p && *p )
{
pErr = p;
p = SkipWhiteSpace( p, encoding );
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, pErr, data, encoding );
return 0;
}
if ( *p == '/' )
{
++p;
// Empty tag.
if ( *p != '>' )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_EMPTY, p, data, encoding );
return 0;
}
return (p+1);
}
else if ( *p == '>' )
{
// Done with attributes (if there were any.)
// Read the value -- which can include other
// elements -- read the end tag, and return.
++p;
p = ReadValue( p, data, encoding ); // Note this is an Element method, and will set the error if one happens.
if ( !p || !*p ) {
// We were looking for the end tag, but found nothing.
// Fix for [ 1663758 ] Failure to report error on bad XML
if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data, encoding );
return 0;
}
// We should find the end tag now
// note that:
// </foo > and
// </foo>
// are both valid end tags.
if ( StringEqual( p, endTag.c_str(), false, encoding ) )
{
p += endTag.length();
p = SkipWhiteSpace( p, encoding );
if ( p && *p && *p == '>' ) {
++p;
return p;
}
if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data, encoding );
return 0;
}
else
{
if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data, encoding );
return 0;
}
}
else
{
// Try to read an attribute:
TiXmlAttribute* attrib = new TiXmlAttribute();
if ( !attrib )
{
return 0;
}
attrib->SetDocument( document );
pErr = p;
p = attrib->Parse( p, data, encoding );
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, pErr, data, encoding );
delete attrib;
return 0;
}
// Handle the strange case of double attributes:
#ifdef TIXML_USE_STL
TiXmlAttribute* node = attributeSet.Find( attrib->NameTStr() );
#else
TiXmlAttribute* node = attributeSet.Find( attrib->Name() );
#endif
if ( node )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, pErr, data, encoding );
delete attrib;
return 0;
}
attributeSet.Add( attrib );
}
}
return p;
}
void TiXmlBase::PutString( const TIXML_STRING& str, TIXML_STRING* outString )
{
int i=0;
while( i<(int)str.length() )
{
int c = str[i];
if ( c == '&'
&& i < ( (int)str.length() - 2 )
&& str[i+1] == '#'
&& str[i+2] == 'x' )
{
// Hexadecimal character reference.
// Pass through unchanged.
// © -- copyright symbol, for example.
while ( i<(int)str.length() )
{
outString->append( str.c_str() + i, 1 );
++i;
if ( str[i] == ';' )
break;
}
}
else if ( c == '&' )
{
outString->append( entity[0].str, entity[0].strLength );
++i;
}
else if ( c == '<' )
{
outString->append( entity[1].str, entity[1].strLength );
++i;
}
else if ( c == '>' )
{
outString->append( entity[2].str, entity[2].strLength );
++i;
}
else if ( c == '\"' )
{
outString->append( entity[3].str, entity[3].strLength );
++i;
}
else if ( c == '\'' )
{
outString->append( entity[4].str, entity[4].strLength );
++i;
}
else if ( c < 32 || c > 126 )
{
// Easy pass at non-alpha/numeric/symbol
// 127 is the delete key. Below 32 is symbolic.
TCHAR buf[ 32 ];
wsprintf( buf, TEXT("&#x%04X;"), (unsigned) ( c & 0xffff ) );
outString->append( buf, lstrlen( buf ) );
++i;
}
else
{
TCHAR realc = (TCHAR) c;
outString->append( &realc, 1 );
++i;
}
}
}
void TiXmlBase::EncodeString( const TIXML_STRING& str, TIXML_STRING* outString )
{
int i=0;
while( i<(int)str.length() )
{
unsigned char c = (unsigned char) str[i];
if ( c == '&'
&& i < ( (int)str.length() - 2 )
&& str[i+1] == '#'
&& str[i+2] == 'x' )
{
// Hexadecimal character reference.
// Pass through unchanged.
// © -- copyright symbol, for example.
//
// The -1 is a bug fix from Rob Laveaux. It keeps
// an overflow from happening if there is no ';'.
// There are actually 2 ways to exit this loop -
// while fails (error case) and break (semicolon found).
// However, there is no mechanism (currently) for
// this function to return an error.
while ( i<(int)str.length()-1 )
{
outString->append( str.c_str() + i, 1 );
++i;
if ( str[i] == ';' )
break;
}
}
else if ( c == '&' )
{
outString->append( entity[0].str, entity[0].strLength );
++i;
}
else if ( c == '<' )
{
outString->append( entity[1].str, entity[1].strLength );
++i;
}
else if ( c == '>' )
{
outString->append( entity[2].str, entity[2].strLength );
++i;
}
else if ( c == '\"' )
{
outString->append( entity[3].str, entity[3].strLength );
++i;
}
else if ( c == '\'' )
{
outString->append( entity[4].str, entity[4].strLength );
++i;
}
else if ( c < 32 )
{
// Easy pass at non-alpha/numeric/symbol
// Below 32 is symbolic.
char buf[ 32 ];
#if defined(TIXML_SNPRINTF)
TIXML_SNPRINTF( buf, sizeof(buf), "&#x%02X;", (unsigned) ( c & 0xff ) );
#else
sprintf( buf, "&#x%02X;", (unsigned) ( c & 0xff ) );
#endif
//*ME: warning C4267: convert 'size_t' to 'int'
//*ME: Int-Cast to make compiler happy ...
outString->append( buf, (int)strlen( buf ) );
++i;
}
else
{
//char realc = (char) c;
//outString->append( &realc, 1 );
*outString += (char) c; // somewhat more efficient function call.
++i;
}
}
}
const char* TiXmlElement::Parse( const char* p )
{
p = SkipWhiteSpace( p );
TiXmlDocument* document = GetDocument();
if ( !p || !*p || *p != '<' )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT );
return false;
}
p = SkipWhiteSpace( p+1 );
// Read the name.
p = ReadName( p, &value );
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_FAILED_TO_READ_ELEMENT_NAME );
return false;
}
TIXML_STRING endTag ("</");
endTag += value;
endTag += ">";
// Check for and read attributes. Also look for an empty
// tag or an end tag.
while ( p && *p )
{
p = SkipWhiteSpace( p );
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES );
return 0;
}
if ( *p == '/' )
{
++p;
// Empty tag.
if ( *p != '>' )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_EMPTY );
return 0;
}
return (p+1);
}
else if ( *p == '>' )
{
// Done with attributes (if there were any.)
// Read the value -- which can include other
// elements -- read the end tag, and return.
++p;
p = ReadValue( p ); // Note this is an Element method, and will set the error if one happens.
if ( !p || !*p )
return 0;
// We should find the end tag now
if ( StringEqual( p, endTag.c_str(), false ) )
{
p += endTag.length();
return p;
}
else
{
if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG );
return 0;
}
}
else
{
// Try to read an element:
TiXmlAttribute attrib;
attrib.SetDocument( document );
p = attrib.Parse( p );
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT );
return 0;
}
SetAttribute( attrib.Name(), attrib.Value() );
}
}
return p;
}
const char* TiXmlElement::Parse( const char* p, TiXmlParsingData* data )
{
p = SkipWhiteSpace( p );
TiXmlDocument* document = GetDocument();
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, 0, 0 );
return 0;
}
// TiXmlParsingData data( p, prevData );
if ( data )
{
data->Stamp( p );
location = data->Cursor();
}
if ( *p != '<' )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, p, data );
return 0;
}
p = SkipWhiteSpace( p+1 );
// Read the name.
const char* pErr = p;
p = ReadName( p, &value );
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_FAILED_TO_READ_ELEMENT_NAME, pErr, data );
return 0;
}
TIXML_STRING endTag ("</");
endTag += value;
endTag += ">";
// Check for and read attributes. Also look for an empty
// tag or an end tag.
while ( p && *p )
{
pErr = p;
p = SkipWhiteSpace( p );
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, pErr, data );
return 0;
}
if ( *p == '/' )
{
++p;
// Empty tag.
if ( *p != '>' )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_EMPTY, p, data );
return 0;
}
return (p+1);
}
else if ( *p == '>' )
{
// Done with attributes (if there were any.)
// Read the value -- which can include other
// elements -- read the end tag, and return.
++p;
p = ReadValue( p, data ); // Note this is an Element method, and will set the error if one happens.
if ( !p || !*p )
return 0;
// We should find the end tag now
if ( StringEqual( p, endTag.c_str(), false ) )
{
p += endTag.length();
return p;
}
else
{
if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data );
return 0;
}
}
else
{
// Try to read an attribute:
TiXmlAttribute* attrib = new TiXmlAttribute();
if ( !attrib )
{
if ( document ) document->SetError( TIXML_ERROR_OUT_OF_MEMORY, pErr, data );
return 0;
}
attrib->SetDocument( document );
const char* pErr = p;
p = attrib->Parse( p, data );
if ( !p || !*p )
{
if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, pErr, data );
delete attrib;
return 0;
}
// Handle the strange case of double attributes:
TiXmlAttribute* node = attributeSet.Find( attrib->Name() );
if ( node )
{
node->SetValue( attrib->Value() );
delete attrib;
return 0;
}
attributeSet.Add( attrib );
}
}
return p;
}
void TiXmlBase::PutString( const TIXML_STRING& str, TIXML_STRING* outString )
{
int i=0;
while( i<(int)str.length() )
{
unsigned char c = (unsigned char) str[i];
if ( c == '&'
&& i < ( (int)str.length() - 2 )
&& str[i+1] == '#'
&& str[i+2] == 'x' )
{
// Hexadecimal character reference.
// Pass through unchanged.
// © -- copyright symbol, for example.
//
// The -1 is a bug fix from Rob Laveaux. It keeps
// an overflow from happening if there is no ';'.
// There are actually 2 ways to exit this loop -
// while fails (error case) and break (semicolon found).
// However, there is no mechanism (currently) for
// this function to return an error.
while ( i<(int)str.length()-1 )
{
outString->append( str.c_str() + i, 1 );
++i;
if ( str[i] == ';' )
break;
}
}
else if ( c == '&' )
{
outString->append( entity[0].str, entity[0].strLength );
++i;
}
else if ( c == '<' )
{
outString->append( entity[1].str, entity[1].strLength );
++i;
}
else if ( c == '>' )
{
outString->append( entity[2].str, entity[2].strLength );
++i;
}
else if ( c == '\"' )
{
outString->append( entity[3].str, entity[3].strLength );
++i;
}
else if ( c == '\'' )
{
outString->append( entity[4].str, entity[4].strLength );
++i;
}
else if ( c == 0x0d && ( (int)str.length() - 1 )
&& str[i+1] == 0x0a)
{
// insert \n
*outString += '\n';
++i;
++i;
}
else if (isspace (c))
{
// pass through unchagend
*outString += (char) c; // somewhat more efficient function call.
++i;
}
else if (iscntrl (c)) // ( c < 32)
{
std::cerr << "WARNING: control character 0x"
<< std::hex << std::setw(2) << std::setfill('0') << (unsigned)c
<< " in text input at character "
<< std::dec << std::setw(0) << i
<< std::endl;
++i;
}
else
{
// just pass these through, since we've
// declared an encoding that presumably allows
// them
*outString += (char) c; // somewhat more efficient function call.
//char realc = (char) c;
//outString->append( &realc, 1 );
++i;
}
}
}
void TiXmlBase::EncodeString( const TIXML_STRING& str, TIXML_STRING* outString )
{
int i=0;
while( i<(int)str.length() )
{
unsigned char c = (unsigned char) str[i];
if ( c == '&'
&& i < ( (int)str.length() - 2 )
&& str[i+1] == '#'
&& str[i+2] == 'x' )
{
while ( i<(int)str.length()-1 )
{
outString->append( str.c_str() + i, 1 );
++i;
if ( str[i] == ';' )
break;
}
}
else if ( c == '&' )
{
outString->append( entity[0].str, entity[0].strLength );
++i;
}
else if ( c == '<' )
{
outString->append( entity[1].str, entity[1].strLength );
++i;
}
else if ( c == '>' )
{
outString->append( entity[2].str, entity[2].strLength );
++i;
}
else if ( c == '\"' )
{
outString->append( entity[3].str, entity[3].strLength );
++i;
}
else if ( c == '\'' )
{
outString->append( entity[4].str, entity[4].strLength );
++i;
}
else if ( c < 32 )
{
char buf[ 32 ];
#if defined(TIXML_SNPRINTF)
TIXML_SNPRINTF( buf, sizeof(buf), "&#x%02X;", (unsigned) ( c & 0xff ) );
#else
sprintf( buf, "&#x%02X;", (unsigned) ( c & 0xff ) );
#endif
outString->append( buf, (int)strlen( buf ) );
++i;
}
else
{
*outString += (char) c;
++i;
}
}
}
void TiXmlBase::PutString( const TIXML_STRING& str, TIXML_STRING* outString )
{
int i=0;
while( i<(int)str.length() )
{
unsigned char c = (unsigned char) str[i];
if ( c == '&'
&& i < ( (int)str.length() - 2 )
&& str[i+1] == '#'
&& str[i+2] == 'x' )
{
// Hexadecimal character reference.
// Pass through unchanged.
// © -- copyright symbol, for example.
while ( i<(int)str.length() )
{
outString->append( str.c_str() + i, 1 );
++i;
if ( str[i] == ';' )
break;
}
}
else if ( c == '&' )
{
outString->append( entity[0].str, entity[0].strLength );
++i;
}
else if ( c == '<' )
{
outString->append( entity[1].str, entity[1].strLength );
++i;
}
else if ( c == '>' )
{
outString->append( entity[2].str, entity[2].strLength );
++i;
}
else if ( c == '\"' )
{
outString->append( entity[3].str, entity[3].strLength );
++i;
}
else if ( c == '\'' )
{
outString->append( entity[4].str, entity[4].strLength );
++i;
}
else if ( c < 32 )
{
// Easy pass at non-alpha/numeric/symbol
// Below 32 is symbolic.
char buf[ 32 ];
sprintf( buf, "&#x%02X;", (unsigned) ( c & 0xff ) );
outString->append( buf, strlen( buf ) );
++i;
}
else
{
// Assume everthing else is unicode. c should never actually
// be out of the range of 0-255. Else something has gone strange.
//char realc = (char) c;
//outString->append( &realc, 1 );
*outString += (char) c; // somewhat more efficient function call.
++i;
}
}
}
void TiXmlBase::PutString( const TIXML_STRING& str, TIXML_STRING* outString )
{
TIXML_STRING::size_type i=0;
while( i < static_cast<TIXML_STRING::size_type>(str.length()) )
{
int c = str[i];
if ( c == '&'
&& i < ( static_cast<TIXML_STRING::size_type>(str.length() - 2 ))
&& str[i+1] == '#'
&& str[i+2] == 'x' )
{
// Hexadecimal character reference.
// Pass through unchanged.
// © -- copyright symbol, for example.
while ( i<static_cast<TIXML_STRING::size_type>(str.length()) )
{
outString->append( str.c_str() + i, 1 );
++i;
if ( str[i] == ';' )
break;
}
}
else if ( c == '&' )
{
outString->append( entity[0].str, entity[0].strLength );
++i;
}
else if ( c == '<' )
{
outString->append( entity[1].str, entity[1].strLength );
++i;
}
else if ( c == '>' )
{
outString->append( entity[2].str, entity[2].strLength );
++i;
}
else if ( c == '\"' )
{
outString->append( entity[3].str, entity[3].strLength );
++i;
}
else if ( c == '\'' )
{
outString->append( entity[4].str, entity[4].strLength );
++i;
}
else if ( c < 32 || c > 126 )
{
// Easy pass at non-alpha/numeric/symbol
// 127 is the delete key. Below 32 is symbolic.
char buf[ 32 ];
sprintf( buf, "&#x%02X;", static_cast<unsigned> ( c & 0xff ) );
outString->append( buf, strlen( buf ) );
++i;
}
else
{
char realc = static_cast<char>( c);
outString->append( &realc, 1 );
++i;
}
}
}
// <-- Strange class for a bug fix. Search for STL_STRING_BUG
TiXmlBase::StringToBuffer::StringToBuffer( const TIXML_STRING& str )
: buffer(new char[str.length() + 1]) {
if ( buffer ) {
strcpy( buffer, str.c_str() );
}
}
