ferencd@0: /*
ferencd@0: www.sourceforge.net/projects/tinyxml
ferencd@0: Original code by Lee Thomason (www.grinninglizard.com)
ferencd@0: 
ferencd@0: This software is provided 'as-is', without any express or implied 
ferencd@0: warranty. In no event will the authors be held liable for any 
ferencd@0: damages arising from the use of this software.
ferencd@0: 
ferencd@0: Permission is granted to anyone to use this software for any 
ferencd@0: purpose, including commercial applications, and to alter it and 
ferencd@0: redistribute it freely, subject to the following restrictions:
ferencd@0: 
ferencd@0: 1. The origin of this software must not be misrepresented; you must 
ferencd@0: not claim that you wrote the original software. If you use this
ferencd@0: software in a product, an acknowledgment in the product documentation
ferencd@0: would be appreciated but is not required.
ferencd@0: 
ferencd@0: 2. Altered source versions must be plainly marked as such, and 
ferencd@0: must not be misrepresented as being the original software.
ferencd@0: 
ferencd@0: 3. This notice may not be removed or altered from any source 
ferencd@0: distribution.
ferencd@0: */
ferencd@0: 
ferencd@0: #include <ctype.h>
ferencd@0: #include <stddef.h>
ferencd@0: 
ferencd@0: #include "tinyxml.h"
ferencd@0: 
ferencd@0: //#define DEBUG_PARSER
ferencd@0: #if defined( DEBUG_PARSER )
ferencd@0: #	if defined( DEBUG ) && defined( _MSC_VER )
ferencd@0: #		include <windows.h>
ferencd@0: #		define TIXML_LOG OutputDebugString
ferencd@0: #	else
ferencd@0: #		define TIXML_LOG printf
ferencd@0: #	endif
ferencd@0: #endif
ferencd@0: 
ferencd@0: // Note tha "PutString" hardcodes the same list. This
ferencd@0: // is less flexible than it appears. Changing the entries
ferencd@0: // or order will break putstring.	
ferencd@0: TiXmlBase::Entity TiXmlBase::entity[ TiXmlBase::NUM_ENTITY ] = 
ferencd@0: {
ferencd@0: 	{ "&amp;",  5, '&' },
ferencd@0: 	{ "&lt;",   4, '<' },
ferencd@0: 	{ "&gt;",   4, '>' },
ferencd@0: 	{ "&quot;", 6, '\"' },
ferencd@0: 	{ "&apos;", 6, '\'' }
ferencd@0: };
ferencd@0: 
ferencd@0: // Bunch of unicode info at:
ferencd@0: //		http://www.unicode.org/faq/utf_bom.html
ferencd@0: // Including the basic of this table, which determines the #bytes in the
ferencd@0: // sequence from the lead byte. 1 placed for invalid sequences --
ferencd@0: // although the result will be junk, pass it through as much as possible.
ferencd@0: // Beware of the non-characters in UTF-8:	
ferencd@0: //				ef bb bf (Microsoft "lead bytes")
ferencd@0: //				ef bf be
ferencd@0: //				ef bf bf 
ferencd@0: 
ferencd@0: const unsigned char TIXML_UTF_LEAD_0 = 0xefU;
ferencd@0: const unsigned char TIXML_UTF_LEAD_1 = 0xbbU;
ferencd@0: const unsigned char TIXML_UTF_LEAD_2 = 0xbfU;
ferencd@0: 
ferencd@0: const int TiXmlBase::utf8ByteTable[256] = 
ferencd@0: {
ferencd@0: 	//	0	1	2	3	4	5	6	7	8	9	a	b	c	d	e	f
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x00
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x10
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x20
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x30
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x40
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x50
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x60
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x70	End of ASCII range
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x80 0x80 to 0xc1 invalid
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0x90 
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0xa0 
ferencd@0: 		1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	// 0xb0 
ferencd@0: 		1,	1,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	// 0xc0 0xc2 to 0xdf 2 byte
ferencd@0: 		2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	2,	// 0xd0
ferencd@0: 		3,	3,	3,	3,	3,	3,	3,	3,	3,	3,	3,	3,	3,	3,	3,	3,	// 0xe0 0xe0 to 0xef 3 byte
ferencd@0: 		4,	4,	4,	4,	4,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1,	1	// 0xf0 0xf0 to 0xf4 4 byte, 0xf5 and higher invalid
ferencd@0: };
ferencd@0: 
ferencd@0: 
ferencd@0: void TiXmlBase::ConvertUTF32ToUTF8( unsigned long input, char* output, int* length )
ferencd@0: {
ferencd@0: 	const unsigned long BYTE_MASK = 0xBF;
ferencd@0: 	const unsigned long BYTE_MARK = 0x80;
ferencd@0: 	const unsigned long FIRST_BYTE_MARK[7] = { 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
ferencd@0: 
ferencd@0: 	if (input < 0x80) 
ferencd@0: 		*length = 1;
ferencd@0: 	else if ( input < 0x800 )
ferencd@0: 		*length = 2;
ferencd@0: 	else if ( input < 0x10000 )
ferencd@0: 		*length = 3;
ferencd@0: 	else if ( input < 0x200000 )
ferencd@0: 		*length = 4;
ferencd@0: 	else
ferencd@0: 		{ *length = 0; return; }	// This code won't covert this correctly anyway.
ferencd@0: 
ferencd@0: 	output += *length;
ferencd@0: 
ferencd@0: 	// Scary scary fall throughs.
ferencd@0: 	switch (*length) 
ferencd@0: 	{
ferencd@0: 		case 4:
ferencd@0: 			--output; 
ferencd@0: 	                *output = static_cast<char>((input | BYTE_MARK) & BYTE_MASK);
ferencd@0:             input >>= 6;
ferencd@0: 		case 3:
ferencd@0: 			--output; 
ferencd@0: 	                *output = static_cast<char>((input | BYTE_MARK) & BYTE_MASK);
ferencd@0:             input >>= 6;
ferencd@0: 		case 2:
ferencd@0: 			--output; 
ferencd@0: 	                *output = static_cast<char>((input | BYTE_MARK) & BYTE_MASK);
ferencd@0:             input >>= 6;
ferencd@0: 		case 1:
ferencd@0: 			--output; 
ferencd@0: 	                *output = static_cast<char>(input | FIRST_BYTE_MARK[*length]);
ferencd@0: 	        default:
ferencd@0: 	                break;
ferencd@0: 	}
ferencd@0: }
ferencd@0: 
ferencd@0: 
ferencd@0: /*static*/ int TiXmlBase::IsAlpha( unsigned char anyByte, TiXmlEncoding /*encoding*/ )
ferencd@0: {
ferencd@0: 	// This will only work for low-ascii, everything else is assumed to be a valid
ferencd@0: 	// letter. I'm not sure this is the best approach, but it is quite tricky trying
ferencd@0: 	// to figure out alhabetical vs. not across encoding. So take a very 
ferencd@0: 	// conservative approach.
ferencd@0: 
ferencd@0: //	if ( encoding == TIXML_ENCODING_UTF8 )
ferencd@0: //	{
ferencd@0: 		if ( anyByte < 127 )
ferencd@0: 			return isalpha( anyByte );
ferencd@0: 		else
ferencd@0: 			return 1;	// What else to do? The unicode set is huge...get the english ones right.
ferencd@0: //	}
ferencd@0: //	else
ferencd@0: //	{
ferencd@0: //		return isalpha( anyByte );
ferencd@0: //	}
ferencd@0: }
ferencd@0: 
ferencd@0: 
ferencd@0: /*static*/ int TiXmlBase::IsAlphaNum( unsigned char anyByte, TiXmlEncoding /*encoding*/ )
ferencd@0: {
ferencd@0: 	// This will only work for low-ascii, everything else is assumed to be a valid
ferencd@0: 	// letter. I'm not sure this is the best approach, but it is quite tricky trying
ferencd@0: 	// to figure out alhabetical vs. not across encoding. So take a very 
ferencd@0: 	// conservative approach.
ferencd@0: 
ferencd@0: //	if ( encoding == TIXML_ENCODING_UTF8 )
ferencd@0: //	{
ferencd@0: 		if ( anyByte < 127 )
ferencd@0: 			return isalnum( anyByte );
ferencd@0: 		else
ferencd@0: 			return 1;	// What else to do? The unicode set is huge...get the english ones right.
ferencd@0: //	}
ferencd@0: //	else
ferencd@0: //	{
ferencd@0: //		return isalnum( anyByte );
ferencd@0: //	}
ferencd@0: }
ferencd@0: 
ferencd@0: 
ferencd@0: class TiXmlParsingData
ferencd@0: {
ferencd@0: 	friend class TiXmlDocument;
ferencd@0:   public:
ferencd@0: 	void Stamp( const char* now, TiXmlEncoding encoding );
ferencd@0: 
ferencd@0: 	const TiXmlCursor& Cursor() const	{ return cursor; }
ferencd@0: 
ferencd@0:   private:
ferencd@0: 	// Only used by the document!
ferencd@0: 	TiXmlParsingData( const char* start, int _tabsize, int row, int col )
ferencd@0: 	{
ferencd@0: 		assert( start );
ferencd@0: 		stamp = start;
ferencd@0: 		tabsize = _tabsize;
ferencd@0: 		cursor.row = row;
ferencd@0: 		cursor.col = col;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	TiXmlCursor		cursor;
ferencd@0: 	const char*		stamp;
ferencd@0: 	int				tabsize;
ferencd@0: };
ferencd@0: 
ferencd@0: 
ferencd@0: void TiXmlParsingData::Stamp( const char* now, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	assert( now );
ferencd@0: 
ferencd@0: 	// Do nothing if the tabsize is 0.
ferencd@0: 	if ( tabsize < 1 )
ferencd@0: 	{
ferencd@0: 		return;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	// Get the current row, column.
ferencd@0: 	int row = cursor.row;
ferencd@0: 	int col = cursor.col;
ferencd@0: 	const char* p = stamp;
ferencd@0: 	assert( p );
ferencd@0: 
ferencd@0: 	while ( p < now )
ferencd@0: 	{
ferencd@0: 		// Treat p as unsigned, so we have a happy compiler.
ferencd@0: 	        const unsigned char* pU = reinterpret_cast<const unsigned char*>(p);
ferencd@0: 
ferencd@0: 		// Code contributed by Fletcher Dunn: (modified by lee)
ferencd@0: 		switch (*pU) {
ferencd@0: 			case 0:
ferencd@0: 				// We *should* never get here, but in case we do, don't
ferencd@0: 				// advance past the terminating null character, ever
ferencd@0: 				return;
ferencd@0: 
ferencd@0: 			case '\r':
ferencd@0: 				// bump down to the next line
ferencd@0: 				++row;
ferencd@0: 				col = 0;				
ferencd@0: 				// Eat the character
ferencd@0: 				++p;
ferencd@0: 
ferencd@0: 				// Check for \r\n sequence, and treat this as a single character
ferencd@0: 				if (*p == '\n') {
ferencd@0: 					++p;
ferencd@0: 				}
ferencd@0: 				break;
ferencd@0: 
ferencd@0: 			case '\n':
ferencd@0: 				// bump down to the next line
ferencd@0: 				++row;
ferencd@0: 				col = 0;
ferencd@0: 
ferencd@0: 				// Eat the character
ferencd@0: 				++p;
ferencd@0: 
ferencd@0: 				// Check for \n\r sequence, and treat this as a single
ferencd@0: 				// character.  (Yes, this bizarre thing does occur still
ferencd@0: 				// on some arcane platforms...)
ferencd@0: 				if (*p == '\r') {
ferencd@0: 					++p;
ferencd@0: 				}
ferencd@0: 				break;
ferencd@0: 
ferencd@0: 			case '\t':
ferencd@0: 				// Eat the character
ferencd@0: 				++p;
ferencd@0: 
ferencd@0: 				// Skip to next tab stop
ferencd@0: 				col = (col / tabsize + 1) * tabsize;
ferencd@0: 				break;
ferencd@0: 
ferencd@0: 			case TIXML_UTF_LEAD_0:
ferencd@0: 				if ( encoding == TIXML_ENCODING_UTF8 )
ferencd@0: 				{
ferencd@0: 					if ( *(p+1) && *(p+2) )
ferencd@0: 					{
ferencd@0: 						// In these cases, don't advance the column. These are
ferencd@0: 						// 0-width spaces.
ferencd@0: 						if ( *(pU+1)==TIXML_UTF_LEAD_1 && *(pU+2)==TIXML_UTF_LEAD_2 )
ferencd@0: 							p += 3;	
ferencd@0: 						else if ( *(pU+1)==0xbfU && *(pU+2)==0xbeU )
ferencd@0: 							p += 3;	
ferencd@0: 						else if ( *(pU+1)==0xbfU && *(pU+2)==0xbfU )
ferencd@0: 							p += 3;	
ferencd@0: 						else
ferencd@0: 							{ p +=3; ++col; }	// A normal character.
ferencd@0: 					}
ferencd@0: 				}
ferencd@0: 				else
ferencd@0: 				{
ferencd@0: 					++p;
ferencd@0: 					++col;
ferencd@0: 				}
ferencd@0: 				break;
ferencd@0: 
ferencd@0: 			default:
ferencd@0: 				if ( encoding == TIXML_ENCODING_UTF8 )
ferencd@0: 				{
ferencd@0: 					// Eat the 1 to 4 byte utf8 character.
ferencd@0: 					int step = TiXmlBase::utf8ByteTable[*((const unsigned char*)p)];
ferencd@0: 					if ( step == 0 )
ferencd@0: 						step = 1;		// Error case from bad encoding, but handle gracefully.
ferencd@0: 					p += step;
ferencd@0: 
ferencd@0: 					// Just advance one column, of course.
ferencd@0: 					++col;
ferencd@0: 				}
ferencd@0: 				else
ferencd@0: 				{
ferencd@0: 					++p;
ferencd@0: 					++col;
ferencd@0: 				}
ferencd@0: 				break;
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 	cursor.row = row;
ferencd@0: 	cursor.col = col;
ferencd@0: 	assert( cursor.row >= -1 );
ferencd@0: 	assert( cursor.col >= -1 );
ferencd@0: 	stamp = p;
ferencd@0: 	assert( stamp );
ferencd@0: }
ferencd@0: 
ferencd@0: 
ferencd@0: const char* TiXmlBase::SkipWhiteSpace( const char* p, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	if ( !p || !*p )
ferencd@0: 	{
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 	if ( encoding == TIXML_ENCODING_UTF8 )
ferencd@0: 	{
ferencd@0: 		while ( *p )
ferencd@0: 		{
ferencd@0: 			const unsigned char* pU = (const unsigned char*)p;
ferencd@0: 			
ferencd@0: 			// Skip the stupid Microsoft UTF-8 Byte order marks
ferencd@0: 			if (	*(pU+0)==TIXML_UTF_LEAD_0
ferencd@0: 				 && *(pU+1)==TIXML_UTF_LEAD_1 
ferencd@0: 				 && *(pU+2)==TIXML_UTF_LEAD_2 )
ferencd@0: 			{
ferencd@0: 				p += 3;
ferencd@0: 				continue;
ferencd@0: 			}
ferencd@0: 			else if(*(pU+0)==TIXML_UTF_LEAD_0
ferencd@0: 				 && *(pU+1)==0xbfU
ferencd@0: 				 && *(pU+2)==0xbeU )
ferencd@0: 			{
ferencd@0: 				p += 3;
ferencd@0: 				continue;
ferencd@0: 			}
ferencd@0: 			else if(*(pU+0)==TIXML_UTF_LEAD_0
ferencd@0: 				 && *(pU+1)==0xbfU
ferencd@0: 				 && *(pU+2)==0xbfU )
ferencd@0: 			{
ferencd@0: 				p += 3;
ferencd@0: 				continue;
ferencd@0: 			}
ferencd@0: 
ferencd@0: 			if ( IsWhiteSpace( *p ) )		// Still using old rules for white space.
ferencd@0: 				++p;
ferencd@0: 			else
ferencd@0: 				break;
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 	else
ferencd@0: 	{
ferencd@0: 		while ( *p && IsWhiteSpace( *p ) )
ferencd@0: 			++p;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	return p;
ferencd@0: }
ferencd@0: 
ferencd@0: #ifdef TIXML_USE_STL
ferencd@0: /*static*/ bool TiXmlBase::StreamWhiteSpace( std::istream * in, TIXML_STRING * tag )
ferencd@0: {
ferencd@0: 	for( ;; )
ferencd@0: 	{
ferencd@0: 		if ( !in->good() ) return false;
ferencd@0: 
ferencd@0: 		int c = in->peek();
ferencd@0: 		// At this scope, we can't get to a document. So fail silently.
ferencd@0: 		if ( !IsWhiteSpace( c ) || c <= 0 )
ferencd@0: 			return true;
ferencd@0: 
ferencd@0: 		*tag += (char) in->get();
ferencd@0: 	}
ferencd@0: }
ferencd@0: 
ferencd@0: /*static*/ bool TiXmlBase::StreamTo( std::istream * in, int character, TIXML_STRING * tag )
ferencd@0: {
ferencd@0: 	//assert( character > 0 && character < 128 );	// else it won't work in utf-8
ferencd@0: 	while ( in->good() )
ferencd@0: 	{
ferencd@0: 		int c = in->peek();
ferencd@0: 		if ( c == character )
ferencd@0: 			return true;
ferencd@0: 		if ( c <= 0 )		// Silent failure: can't get document at this scope
ferencd@0: 			return false;
ferencd@0: 
ferencd@0: 		in->get();
ferencd@0: 		*tag += (char) c;
ferencd@0: 	}
ferencd@0: 	return false;
ferencd@0: }
ferencd@0: #endif
ferencd@0: 
ferencd@0: // One of TinyXML's more performance demanding functions. Try to keep the memory overhead down. The
ferencd@0: // "assign" optimization removes over 10% of the execution time.
ferencd@0: //
ferencd@0: const char* TiXmlBase::ReadName( const char* p, TIXML_STRING * name, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	// Oddly, not supported on some comilers,
ferencd@0: 	//name->clear();
ferencd@0: 	// So use this:
ferencd@0: 	*name = "";
ferencd@0: 	assert( p );
ferencd@0: 
ferencd@0: 	// Names start with letters or underscores.
ferencd@0: 	// Of course, in unicode, tinyxml has no idea what a letter *is*. The
ferencd@0: 	// algorithm is generous.
ferencd@0: 	//
ferencd@0: 	// After that, they can be letters, underscores, numbers,
ferencd@0: 	// hyphens, or colons. (Colons are valid ony for namespaces,
ferencd@0: 	// but tinyxml can't tell namespaces from names.)
ferencd@0: 	if (    p && *p 
ferencd@0: 		 && ( IsAlpha( (unsigned char) *p, encoding ) || *p == '_' ) )
ferencd@0: 	{
ferencd@0: 		const char* start = p;
ferencd@0: 		while(		p && *p
ferencd@0: 				&&	(		IsAlphaNum( (unsigned char ) *p, encoding ) 
ferencd@0: 						 || *p == '_'
ferencd@0: 						 || *p == '-'
ferencd@0: 						 || *p == '.'
ferencd@0: 						 || *p == ':' ) )
ferencd@0: 		{
ferencd@0: 			//(*name) += *p; // expensive
ferencd@0: 			++p;
ferencd@0: 		}
ferencd@0: 		if ( p-start > 0 ) {
ferencd@0: 			name->assign( start, p-start );
ferencd@0: 		}
ferencd@0: 		return p;
ferencd@0: 	}
ferencd@0: 	return 0;
ferencd@0: }
ferencd@0: 
ferencd@0: const char* TiXmlBase::GetEntity( const char* p, char* value, int* length, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	// Presume an entity, and pull it out.
ferencd@0:     TIXML_STRING ent;
ferencd@0: 	int i;
ferencd@0: 	*length = 0;
ferencd@0: 
ferencd@0: 	if ( *(p+1) && *(p+1) == '#' && *(p+2) )
ferencd@0: 	{
ferencd@0: 		unsigned long ucs = 0;
ferencd@0: 		ptrdiff_t delta = 0;
ferencd@0: 		unsigned mult = 1;
ferencd@0: 
ferencd@0: 		if ( *(p+2) == 'x' )
ferencd@0: 		{
ferencd@0: 			// Hexadecimal.
ferencd@0: 			if ( !*(p+3) ) return 0;
ferencd@0: 
ferencd@0: 			const char* q = p+3;
ferencd@0: 			q = strchr( q, ';' );
ferencd@0: 
ferencd@0: 			if ( !q || !*q ) return 0;
ferencd@0: 
ferencd@0: 			delta = q-p;
ferencd@0: 			--q;
ferencd@0: 
ferencd@0: 			while ( *q != 'x' )
ferencd@0: 			{
ferencd@0: 				if ( *q >= '0' && *q <= '9' )
ferencd@0: 					ucs += mult * (*q - '0');
ferencd@0: 				else if ( *q >= 'a' && *q <= 'f' )
ferencd@0: 					ucs += mult * (*q - 'a' + 10);
ferencd@0: 				else if ( *q >= 'A' && *q <= 'F' )
ferencd@0: 					ucs += mult * (*q - 'A' + 10 );
ferencd@0: 				else 
ferencd@0: 					return 0;
ferencd@0: 				mult *= 16;
ferencd@0: 				--q;
ferencd@0: 			}
ferencd@0: 		}
ferencd@0: 		else
ferencd@0: 		{
ferencd@0: 			// Decimal.
ferencd@0: 			if ( !*(p+2) ) return 0;
ferencd@0: 
ferencd@0: 			const char* q = p+2;
ferencd@0: 			q = strchr( q, ';' );
ferencd@0: 
ferencd@0: 			if ( !q || !*q ) return 0;
ferencd@0: 
ferencd@0: 			delta = q-p;
ferencd@0: 			--q;
ferencd@0: 
ferencd@0: 			while ( *q != '#' )
ferencd@0: 			{
ferencd@0: 				if ( *q >= '0' && *q <= '9' )
ferencd@0: 					ucs += mult * (*q - '0');
ferencd@0: 				else 
ferencd@0: 					return 0;
ferencd@0: 				mult *= 10;
ferencd@0: 				--q;
ferencd@0: 			}
ferencd@0: 		}
ferencd@0: 		if ( encoding == TIXML_ENCODING_UTF8 )
ferencd@0: 		{
ferencd@0: 			// convert the UCS to UTF-8
ferencd@0: 			ConvertUTF32ToUTF8( ucs, value, length );
ferencd@0: 		}
ferencd@0: 		else
ferencd@0: 		{
ferencd@0: 			*value = (char)ucs;
ferencd@0: 			*length = 1;
ferencd@0: 		}
ferencd@0: 		return p + delta + 1;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	// Now try to match it.
ferencd@0: 	for( i=0; i<NUM_ENTITY; ++i )
ferencd@0: 	{
ferencd@0: 		if ( strncmp( entity[i].str, p, entity[i].strLength ) == 0 )
ferencd@0: 		{
ferencd@0: 			assert( strlen( entity[i].str ) == entity[i].strLength );
ferencd@0: 			*value = entity[i].chr;
ferencd@0: 			*length = 1;
ferencd@0: 			return ( p + entity[i].strLength );
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	// So it wasn't an entity, its unrecognized, or something like that.
ferencd@0: 	*value = *p;	// Don't put back the last one, since we return it!
ferencd@0: 	//*length = 1;	// Leave unrecognized entities - this doesn't really work.
ferencd@0: 					// Just writes strange XML.
ferencd@0: 	return p+1;
ferencd@0: }
ferencd@0: 
ferencd@0: 
ferencd@0: bool TiXmlBase::StringEqual( const char* p,
ferencd@0: 							 const char* tag,
ferencd@0: 							 bool ignoreCase,
ferencd@0: 							 TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	assert( p );
ferencd@0: 	assert( tag );
ferencd@0: 	if ( !p || !*p )
ferencd@0: 	{
ferencd@0: 		assert( 0 );
ferencd@0: 		return false;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	const char* q = p;
ferencd@0: 
ferencd@0: 	if ( ignoreCase )
ferencd@0: 	{
ferencd@0: 		while ( *q && *tag && ToLower( *q, encoding ) == ToLower( *tag, encoding ) )
ferencd@0: 		{
ferencd@0: 			++q;
ferencd@0: 			++tag;
ferencd@0: 		}
ferencd@0: 
ferencd@0: 		if ( *tag == 0 )
ferencd@0: 			return true;
ferencd@0: 	}
ferencd@0: 	else
ferencd@0: 	{
ferencd@0: 		while ( *q && *tag && *q == *tag )
ferencd@0: 		{
ferencd@0: 			++q;
ferencd@0: 			++tag;
ferencd@0: 		}
ferencd@0: 
ferencd@0: 		if ( *tag == 0 )		// Have we found the end of the tag, and everything equal?
ferencd@0: 			return true;
ferencd@0: 	}
ferencd@0: 	return false;
ferencd@0: }
ferencd@0: 
ferencd@0: const char* TiXmlBase::ReadText(	const char* p, 
ferencd@0: 									TIXML_STRING * text, 
ferencd@0: 									bool trimWhiteSpace, 
ferencd@0: 									const char* endTag, 
ferencd@0: 									bool caseInsensitive,
ferencd@0: 									TiXmlEncoding encoding )
ferencd@0: {
ferencd@0:     *text = "";
ferencd@0: 	if (    !trimWhiteSpace			// certain tags always keep whitespace
ferencd@0: 		 || !condenseWhiteSpace )	// if true, whitespace is always kept
ferencd@0: 	{
ferencd@0: 		// Keep all the white space.
ferencd@0: 		while (	   p && *p
ferencd@0: 				&& !StringEqual( p, endTag, caseInsensitive, encoding )
ferencd@0: 			  )
ferencd@0: 		{
ferencd@0: 			int len;
ferencd@0: 			char cArr[4] = { 0, 0, 0, 0 };
ferencd@0: 			p = GetChar( p, cArr, &len, encoding );
ferencd@0: 			text->append( cArr, len );
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 	else
ferencd@0: 	{
ferencd@0: 		bool whitespace = false;
ferencd@0: 
ferencd@0: 		// Remove leading white space:
ferencd@0: 		p = SkipWhiteSpace( p, encoding );
ferencd@0: 		while (	   p && *p
ferencd@0: 				&& !StringEqual( p, endTag, caseInsensitive, encoding ) )
ferencd@0: 		{
ferencd@0: 			if ( *p == '\r' || *p == '\n' )
ferencd@0: 			{
ferencd@0: 				whitespace = true;
ferencd@0: 				++p;
ferencd@0: 			}
ferencd@0: 			else if ( IsWhiteSpace( *p ) )
ferencd@0: 			{
ferencd@0: 				whitespace = true;
ferencd@0: 				++p;
ferencd@0: 			}
ferencd@0: 			else
ferencd@0: 			{
ferencd@0: 				// If we've found whitespace, add it before the
ferencd@0: 				// new character. Any whitespace just becomes a space.
ferencd@0: 				if ( whitespace )
ferencd@0: 				{
ferencd@0: 					(*text) += ' ';
ferencd@0: 					whitespace = false;
ferencd@0: 				}
ferencd@0: 				int len;
ferencd@0: 				char cArr[4] = { 0, 0, 0, 0 };
ferencd@0: 				p = GetChar( p, cArr, &len, encoding );
ferencd@0: 				if ( len == 1 )
ferencd@0: 					(*text) += cArr[0];	// more efficient
ferencd@0: 				else
ferencd@0: 					text->append( cArr, len );
ferencd@0: 			}
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 	if ( p && *p )
ferencd@0: 		p += strlen( endTag );
ferencd@0: 	return ( p && *p ) ? p : 0;
ferencd@0: }
ferencd@0: 
ferencd@0: #ifdef TIXML_USE_STL
ferencd@0: 
ferencd@0: void TiXmlDocument::StreamIn( std::istream * in, TIXML_STRING * tag )
ferencd@0: {
ferencd@0: 	// The basic issue with a document is that we don't know what we're
ferencd@0: 	// streaming. Read something presumed to be a tag (and hope), then
ferencd@0: 	// identify it, and call the appropriate stream method on the tag.
ferencd@0: 	//
ferencd@0: 	// This "pre-streaming" will never read the closing ">" so the
ferencd@0: 	// sub-tag can orient itself.
ferencd@0: 
ferencd@0: 	if ( !StreamTo( in, '<', tag ) ) 
ferencd@0: 	{
ferencd@0: 		SetError( TIXML_ERROR_PARSING_EMPTY, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 		return;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	while ( in->good() )
ferencd@0: 	{
ferencd@0: 		int tagIndex = (int) tag->length();
ferencd@0: 		while ( in->good() && in->peek() != '>' )
ferencd@0: 		{
ferencd@0: 			int c = in->get();
ferencd@0: 			if ( c <= 0 )
ferencd@0: 			{
ferencd@0: 				SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 				break;
ferencd@0: 			}
ferencd@0: 			(*tag) += (char) c;
ferencd@0: 		}
ferencd@0: 
ferencd@0: 		if ( in->good() )
ferencd@0: 		{
ferencd@0: 			// We now have something we presume to be a node of 
ferencd@0: 			// some sort. Identify it, and call the node to
ferencd@0: 			// continue streaming.
ferencd@0: 			TiXmlNode* node = Identify( tag->c_str() + tagIndex, TIXML_DEFAULT_ENCODING );
ferencd@0: 
ferencd@0: 			if ( node )
ferencd@0: 			{
ferencd@0: 				node->StreamIn( in, tag );
ferencd@0: 				bool isElement = node->ToElement() != 0;
ferencd@0: 				delete node;
ferencd@0: 				node = 0;
ferencd@0: 
ferencd@0: 				// If this is the root element, we're done. Parsing will be
ferencd@0: 				// done by the >> operator.
ferencd@0: 				if ( isElement )
ferencd@0: 				{
ferencd@0: 					return;
ferencd@0: 				}
ferencd@0: 			}
ferencd@0: 			else
ferencd@0: 			{
ferencd@0: 				SetError( TIXML_ERROR, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 				return;
ferencd@0: 			}
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 	// We should have returned sooner.
ferencd@0: 	SetError( TIXML_ERROR, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: }
ferencd@0: 
ferencd@0: #endif
ferencd@0: 
ferencd@0: const char* TiXmlDocument::Parse( const char* p, TiXmlParsingData* prevData, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	ClearError();
ferencd@0: 
ferencd@0: 	// Parse away, at the document level. Since a document
ferencd@0: 	// contains nothing but other tags, most of what happens
ferencd@0: 	// here is skipping white space.
ferencd@0: 	if ( !p || !*p )
ferencd@0: 	{
ferencd@0: 		SetError( TIXML_ERROR_DOCUMENT_EMPTY, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	// Note that, for a document, this needs to come
ferencd@0: 	// before the while space skip, so that parsing
ferencd@0: 	// starts from the pointer we are given.
ferencd@0: 	location.Clear();
ferencd@0: 	if ( prevData )
ferencd@0: 	{
ferencd@0: 		location.row = prevData->cursor.row;
ferencd@0: 		location.col = prevData->cursor.col;
ferencd@0: 	}
ferencd@0: 	else
ferencd@0: 	{
ferencd@0: 		location.row = 0;
ferencd@0: 		location.col = 0;
ferencd@0: 	}
ferencd@0: 	TiXmlParsingData data( p, TabSize(), location.row, location.col );
ferencd@0: 	location = data.Cursor();
ferencd@0: 
ferencd@0: 	if ( encoding == TIXML_ENCODING_UNKNOWN )
ferencd@0: 	{
ferencd@0: 		// Check for the Microsoft UTF-8 lead bytes.
ferencd@0: 		const unsigned char* pU = (const unsigned char*)p;
ferencd@0: 		if (	*(pU+0) && *(pU+0) == TIXML_UTF_LEAD_0
ferencd@0: 			 && *(pU+1) && *(pU+1) == TIXML_UTF_LEAD_1
ferencd@0: 			 && *(pU+2) && *(pU+2) == TIXML_UTF_LEAD_2 )
ferencd@0: 		{
ferencd@0: 			encoding = TIXML_ENCODING_UTF8;
ferencd@0: 			useMicrosoftBOM = true;
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 
ferencd@0:     p = SkipWhiteSpace( p, encoding );
ferencd@0: 	if ( !p )
ferencd@0: 	{
ferencd@0: 		SetError( TIXML_ERROR_DOCUMENT_EMPTY, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	while ( p && *p )
ferencd@0: 	{
ferencd@0: 		TiXmlNode* node = Identify( p, encoding );
ferencd@0: 		if ( node )
ferencd@0: 		{
ferencd@0: 			p = node->Parse( p, &data, encoding );
ferencd@0: 			LinkEndChild( node );
ferencd@0: 		}
ferencd@0: 		else
ferencd@0: 		{
ferencd@0: 			break;
ferencd@0: 		}
ferencd@0: 
ferencd@0: 		// Did we get encoding info?
ferencd@0: 		if (    encoding == TIXML_ENCODING_UNKNOWN
ferencd@0: 			 && node->ToDeclaration() )
ferencd@0: 		{
ferencd@0: 			TiXmlDeclaration* dec = node->ToDeclaration();
ferencd@0: 			const char* enc = dec->Encoding();
ferencd@0: 			assert( enc );
ferencd@0: 
ferencd@0: 			if ( *enc == 0 )
ferencd@0: 				encoding = TIXML_ENCODING_UTF8;
ferencd@0: 			else if ( StringEqual( enc, "UTF-8", true, TIXML_ENCODING_UNKNOWN ) )
ferencd@0: 				encoding = TIXML_ENCODING_UTF8;
ferencd@0: 			else if ( StringEqual( enc, "UTF8", true, TIXML_ENCODING_UNKNOWN ) )
ferencd@0: 				encoding = TIXML_ENCODING_UTF8;	// incorrect, but be nice
ferencd@0: 			else 
ferencd@0: 				encoding = TIXML_ENCODING_LEGACY;
ferencd@0: 		}
ferencd@0: 
ferencd@0: 		p = SkipWhiteSpace( p, encoding );
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	// Was this empty?
ferencd@0: 	if ( !firstChild ) {
ferencd@0: 		SetError( TIXML_ERROR_DOCUMENT_EMPTY, 0, 0, encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	// All is well.
ferencd@0: 	return p;
ferencd@0: }
ferencd@0: 
ferencd@0: void TiXmlDocument::SetError( int err, const char* pError, TiXmlParsingData* data, TiXmlEncoding encoding )
ferencd@0: {	
ferencd@0: 	// The first error in a chain is more accurate - don't set again!
ferencd@0: 	if ( error )
ferencd@0: 		return;
ferencd@0: 
ferencd@0: 	assert( err > 0 && err < TIXML_ERROR_STRING_COUNT );
ferencd@0: 	error   = true;
ferencd@0: 	errorId = err;
ferencd@0: 	errorDesc = errorString[ errorId ];
ferencd@0: 
ferencd@0: 	errorLocation.Clear();
ferencd@0: 	if ( pError && data )
ferencd@0: 	{
ferencd@0: 		data->Stamp( pError, encoding );
ferencd@0: 		errorLocation = data->Cursor();
ferencd@0: 	}
ferencd@0: }
ferencd@0: 
ferencd@0: 
ferencd@0: TiXmlNode* TiXmlNode::Identify( const char* p, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	TiXmlNode* returnNode = 0;
ferencd@0: 
ferencd@0: 	p = SkipWhiteSpace( p, encoding );
ferencd@0: 	if( !p || !*p || *p != '<' )
ferencd@0: 	{
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	p = SkipWhiteSpace( p, encoding );
ferencd@0: 
ferencd@0: 	if ( !p || !*p )
ferencd@0: 	{
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	// What is this thing? 
ferencd@0: 	// - Elements start with a letter or underscore, but xml is reserved.
ferencd@0: 	// - Comments: <!--
ferencd@0: 	// - Decleration: <?xml
ferencd@0: 	// - Everthing else is unknown to tinyxml.
ferencd@0: 	//
ferencd@0: 
ferencd@0: 	const char* xmlHeader = { "<?xml" };
ferencd@0: 	const char* commentHeader = { "<!--" };
ferencd@0: 	const char* dtdHeader = { "<!" };
ferencd@0: 	const char* cdataHeader = { "<![CDATA[" };
ferencd@0: 
ferencd@0: 	if ( StringEqual( p, xmlHeader, true, encoding ) )
ferencd@0: 	{
ferencd@0: 		#ifdef DEBUG_PARSER
ferencd@0: 			TIXML_LOG( "XML parsing Declaration\n" );
ferencd@0: 		#endif
ferencd@0: 		returnNode = new TiXmlDeclaration();
ferencd@0: 	}
ferencd@0: 	else if ( StringEqual( p, commentHeader, false, encoding ) )
ferencd@0: 	{
ferencd@0: 		#ifdef DEBUG_PARSER
ferencd@0: 			TIXML_LOG( "XML parsing Comment\n" );
ferencd@0: 		#endif
ferencd@0: 		returnNode = new TiXmlComment();
ferencd@0: 	}
ferencd@0: 	else if ( StringEqual( p, cdataHeader, false, encoding ) )
ferencd@0: 	{
ferencd@0: 		#ifdef DEBUG_PARSER
ferencd@0: 			TIXML_LOG( "XML parsing CDATA\n" );
ferencd@0: 		#endif
ferencd@0: 		TiXmlText* text = new TiXmlText( "" );
ferencd@0: 		text->SetCDATA( true );
ferencd@0: 		returnNode = text;
ferencd@0: 	}
ferencd@0: 	else if ( StringEqual( p, dtdHeader, false, encoding ) )
ferencd@0: 	{
ferencd@0: 		#ifdef DEBUG_PARSER
ferencd@0: 			TIXML_LOG( "XML parsing Unknown(1)\n" );
ferencd@0: 		#endif
ferencd@0: 		returnNode = new TiXmlUnknown();
ferencd@0: 	}
ferencd@0: 	else if (    IsAlpha( *(p+1), encoding )
ferencd@0: 			  || *(p+1) == '_' )
ferencd@0: 	{
ferencd@0: 		#ifdef DEBUG_PARSER
ferencd@0: 			TIXML_LOG( "XML parsing Element\n" );
ferencd@0: 		#endif
ferencd@0: 		returnNode = new TiXmlElement( "" );
ferencd@0: 	}
ferencd@0: 	else
ferencd@0: 	{
ferencd@0: 		#ifdef DEBUG_PARSER
ferencd@0: 			TIXML_LOG( "XML parsing Unknown(2)\n" );
ferencd@0: 		#endif
ferencd@0: 		returnNode = new TiXmlUnknown();
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	if ( returnNode )
ferencd@0: 	{
ferencd@0: 		// Set the parent, so it can report errors
ferencd@0: 		returnNode->parent = this;
ferencd@0: 	}
ferencd@0: 	return returnNode;
ferencd@0: }
ferencd@0: 
ferencd@0: #ifdef TIXML_USE_STL
ferencd@0: 
ferencd@0: void TiXmlElement::StreamIn (std::istream * in, TIXML_STRING * tag)
ferencd@0: {
ferencd@0: 	// We're called with some amount of pre-parsing. That is, some of "this"
ferencd@0: 	// element is in "tag". Go ahead and stream to the closing ">"
ferencd@0: 	while( in->good() )
ferencd@0: 	{
ferencd@0: 		int c = in->get();
ferencd@0: 		if ( c <= 0 )
ferencd@0: 		{
ferencd@0: 			TiXmlDocument* document = GetDocument();
ferencd@0: 			if ( document )
ferencd@0: 				document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 			return;
ferencd@0: 		}
ferencd@0: 		(*tag) += (char) c ;
ferencd@0: 		
ferencd@0: 		if ( c == '>' )
ferencd@0: 			break;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	if ( tag->length() < 3 ) return;
ferencd@0: 
ferencd@0: 	// Okay...if we are a "/>" tag, then we're done. We've read a complete tag.
ferencd@0: 	// If not, identify and stream.
ferencd@0: 
ferencd@0: 	if (    tag->at( tag->length() - 1 ) == '>' 
ferencd@0: 		 && tag->at( tag->length() - 2 ) == '/' )
ferencd@0: 	{
ferencd@0: 		// All good!
ferencd@0: 		return;
ferencd@0: 	}
ferencd@0: 	else if ( tag->at( tag->length() - 1 ) == '>' )
ferencd@0: 	{
ferencd@0: 		// There is more. Could be:
ferencd@0: 		//		text
ferencd@0: 		//		cdata text (which looks like another node)
ferencd@0: 		//		closing tag
ferencd@0: 		//		another node.
ferencd@0: 		for ( ;; )
ferencd@0: 		{
ferencd@0: 			StreamWhiteSpace( in, tag );
ferencd@0: 
ferencd@0: 			// Do we have text?
ferencd@0: 			if ( in->good() && in->peek() != '<' ) 
ferencd@0: 			{
ferencd@0: 				// Yep, text.
ferencd@0: 				TiXmlText text( "" );
ferencd@0: 				text.StreamIn( in, tag );
ferencd@0: 
ferencd@0: 				// What follows text is a closing tag or another node.
ferencd@0: 				// Go around again and figure it out.
ferencd@0: 				continue;
ferencd@0: 			}
ferencd@0: 
ferencd@0: 			// We now have either a closing tag...or another node.
ferencd@0: 			// We should be at a "<", regardless.
ferencd@0: 			if ( !in->good() ) return;
ferencd@0: 			assert( in->peek() == '<' );
ferencd@0: 			int tagIndex = (int) tag->length();
ferencd@0: 
ferencd@0: 			bool closingTag = false;
ferencd@0: 			bool firstCharFound = false;
ferencd@0: 
ferencd@0: 			for( ;; )
ferencd@0: 			{
ferencd@0: 				if ( !in->good() )
ferencd@0: 					return;
ferencd@0: 
ferencd@0: 				int c = in->peek();
ferencd@0: 				if ( c <= 0 )
ferencd@0: 				{
ferencd@0: 					TiXmlDocument* document = GetDocument();
ferencd@0: 					if ( document )
ferencd@0: 						document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 					return;
ferencd@0: 				}
ferencd@0: 				
ferencd@0: 				if ( c == '>' )
ferencd@0: 					break;
ferencd@0: 
ferencd@0: 				*tag += (char) c;
ferencd@0: 				in->get();
ferencd@0: 
ferencd@0: 				// Early out if we find the CDATA id.
ferencd@0: 				if ( c == '[' && tag->size() >= 9 )
ferencd@0: 				{
ferencd@0: 					size_t len = tag->size();
ferencd@0: 					const char* start = tag->c_str() + len - 9;
ferencd@0: 					if ( strcmp( start, "<![CDATA[" ) == 0 ) {
ferencd@0: 						assert( !closingTag );
ferencd@0: 						break;
ferencd@0: 					}
ferencd@0: 				}
ferencd@0: 
ferencd@0: 				if ( !firstCharFound && c != '<' && !IsWhiteSpace( c ) )
ferencd@0: 				{
ferencd@0: 					firstCharFound = true;
ferencd@0: 					if ( c == '/' )
ferencd@0: 						closingTag = true;
ferencd@0: 				}
ferencd@0: 			}
ferencd@0: 			// If it was a closing tag, then read in the closing '>' to clean up the input stream.
ferencd@0: 			// If it was not, the streaming will be done by the tag.
ferencd@0: 			if ( closingTag )
ferencd@0: 			{
ferencd@0: 				if ( !in->good() )
ferencd@0: 					return;
ferencd@0: 
ferencd@0: 				int c = in->get();
ferencd@0: 				if ( c <= 0 )
ferencd@0: 				{
ferencd@0: 					TiXmlDocument* document = GetDocument();
ferencd@0: 					if ( document )
ferencd@0: 						document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 					return;
ferencd@0: 				}
ferencd@0: 				assert( c == '>' );
ferencd@0: 				*tag += (char) c;
ferencd@0: 
ferencd@0: 				// We are done, once we've found our closing tag.
ferencd@0: 				return;
ferencd@0: 			}
ferencd@0: 			else
ferencd@0: 			{
ferencd@0: 				// If not a closing tag, id it, and stream.
ferencd@0: 				const char* tagloc = tag->c_str() + tagIndex;
ferencd@0: 				TiXmlNode* node = Identify( tagloc, TIXML_DEFAULT_ENCODING );
ferencd@0: 				if ( !node )
ferencd@0: 					return;
ferencd@0: 				node->StreamIn( in, tag );
ferencd@0: 				delete node;
ferencd@0: 				node = 0;
ferencd@0: 
ferencd@0: 				// No return: go around from the beginning: text, closing tag, or node.
ferencd@0: 			}
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: }
ferencd@0: #endif
ferencd@0: 
ferencd@0: const char* TiXmlElement::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	p = SkipWhiteSpace( p, encoding );
ferencd@0: 	TiXmlDocument* document = GetDocument();
ferencd@0: 
ferencd@0: 	if ( !p || !*p )
ferencd@0: 	{
ferencd@0: 		if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, 0, 0, encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	if ( data )
ferencd@0: 	{
ferencd@0: 		data->Stamp( p, encoding );
ferencd@0: 		location = data->Cursor();
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	if ( *p != '<' )
ferencd@0: 	{
ferencd@0: 		if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, p, data, encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	p = SkipWhiteSpace( p+1, encoding );
ferencd@0: 
ferencd@0: 	// Read the name.
ferencd@0: 	const char* pErr = p;
ferencd@0: 
ferencd@0:     p = ReadName( p, &value, encoding );
ferencd@0: 	if ( !p || !*p )
ferencd@0: 	{
ferencd@0: 		if ( document )	document->SetError( TIXML_ERROR_FAILED_TO_READ_ELEMENT_NAME, pErr, data, encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 
ferencd@0:     TIXML_STRING endTag ("</");
ferencd@0: 	endTag += value;
ferencd@0: 
ferencd@0: 	// Check for and read attributes. Also look for an empty
ferencd@0: 	// tag or an end tag.
ferencd@0: 	while ( p && *p )
ferencd@0: 	{
ferencd@0: 		pErr = p;
ferencd@0: 		p = SkipWhiteSpace( p, encoding );
ferencd@0: 		if ( !p || !*p )
ferencd@0: 		{
ferencd@0: 			if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, pErr, data, encoding );
ferencd@0: 			return 0;
ferencd@0: 		}
ferencd@0: 		if ( *p == '/' )
ferencd@0: 		{
ferencd@0: 			++p;
ferencd@0: 			// Empty tag.
ferencd@0: 			if ( *p  != '>' )
ferencd@0: 			{
ferencd@0: 				if ( document ) document->SetError( TIXML_ERROR_PARSING_EMPTY, p, data, encoding );		
ferencd@0: 				return 0;
ferencd@0: 			}
ferencd@0: 			return (p+1);
ferencd@0: 		}
ferencd@0: 		else if ( *p == '>' )
ferencd@0: 		{
ferencd@0: 			// Done with attributes (if there were any.)
ferencd@0: 			// Read the value -- which can include other
ferencd@0: 			// elements -- read the end tag, and return.
ferencd@0: 			++p;
ferencd@0: 			p = ReadValue( p, data, encoding );		// Note this is an Element method, and will set the error if one happens.
ferencd@0: 			if ( !p || !*p ) {
ferencd@0: 				// We were looking for the end tag, but found nothing.
ferencd@0: 				// Fix for [ 1663758 ] Failure to report error on bad XML
ferencd@0: 				if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data, encoding );
ferencd@0: 				return 0;
ferencd@0: 			}
ferencd@0: 
ferencd@0: 			// We should find the end tag now
ferencd@0: 			// note that:
ferencd@0: 			// </foo > and
ferencd@0: 			// </foo> 
ferencd@0: 			// are both valid end tags.
ferencd@0: 			if ( StringEqual( p, endTag.c_str(), false, encoding ) )
ferencd@0: 			{
ferencd@0: 				p += endTag.length();
ferencd@0: 				p = SkipWhiteSpace( p, encoding );
ferencd@0: 				if ( p && *p && *p == '>' ) {
ferencd@0: 					++p;
ferencd@0: 					return p;
ferencd@0: 				}
ferencd@0: 				if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data, encoding );
ferencd@0: 				return 0;
ferencd@0: 			}
ferencd@0: 			else
ferencd@0: 			{
ferencd@0: 				if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data, encoding );
ferencd@0: 				return 0;
ferencd@0: 			}
ferencd@0: 		}
ferencd@0: 		else
ferencd@0: 		{
ferencd@0: 			// Try to read an attribute:
ferencd@0: 			TiXmlAttribute* attrib = new TiXmlAttribute();
ferencd@0: 			if ( !attrib )
ferencd@0: 			{
ferencd@0: 				return 0;
ferencd@0: 			}
ferencd@0: 
ferencd@0: 			attrib->SetDocument( document );
ferencd@0: 			pErr = p;
ferencd@0: 			p = attrib->Parse( p, data, encoding );
ferencd@0: 
ferencd@0: 			if ( !p || !*p )
ferencd@0: 			{
ferencd@0: 				if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, pErr, data, encoding );
ferencd@0: 				delete attrib;
ferencd@0: 				return 0;
ferencd@0: 			}
ferencd@0: 
ferencd@0: 			// Handle the strange case of double attributes:
ferencd@0: 			#ifdef TIXML_USE_STL
ferencd@0: 			TiXmlAttribute* node = attributeSet.Find( attrib->NameTStr() );
ferencd@0: 			#else
ferencd@0: 			TiXmlAttribute* node = attributeSet.Find( attrib->Name() );
ferencd@0: 			#endif
ferencd@0: 			if ( node )
ferencd@0: 			{
ferencd@0: 				if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, pErr, data, encoding );
ferencd@0: 				delete attrib;
ferencd@0: 				return 0;
ferencd@0: 			}
ferencd@0: 
ferencd@0: 			attributeSet.Add( attrib );
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 	return p;
ferencd@0: }
ferencd@0: 
ferencd@0: 
ferencd@0: const char* TiXmlElement::ReadValue( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	TiXmlDocument* document = GetDocument();
ferencd@0: 
ferencd@0: 	// Read in text and elements in any order.
ferencd@0: 	const char* pWithWhiteSpace = p;
ferencd@0: 	p = SkipWhiteSpace( p, encoding );
ferencd@0: 
ferencd@0: 	while ( p && *p )
ferencd@0: 	{
ferencd@0: 		if ( *p != '<' )
ferencd@0: 		{
ferencd@0: 			// Take what we have, make a text element.
ferencd@0: 			TiXmlText* textNode = new TiXmlText( "" );
ferencd@0: 
ferencd@0: 			if ( !textNode )
ferencd@0: 			{
ferencd@0: 			    return 0;
ferencd@0: 			}
ferencd@0: 
ferencd@0: 			if ( TiXmlBase::IsWhiteSpaceCondensed() )
ferencd@0: 			{
ferencd@0: 				p = textNode->Parse( p, data, encoding );
ferencd@0: 			}
ferencd@0: 			else
ferencd@0: 			{
ferencd@0: 				// Special case: we want to keep the white space
ferencd@0: 				// so that leading spaces aren't removed.
ferencd@0: 				p = textNode->Parse( pWithWhiteSpace, data, encoding );
ferencd@0: 			}
ferencd@0: 
ferencd@0: 			if ( !textNode->Blank() )
ferencd@0: 				LinkEndChild( textNode );
ferencd@0: 			else
ferencd@0: 				delete textNode;
ferencd@0: 		} 
ferencd@0: 		else 
ferencd@0: 		{
ferencd@0: 			// We hit a '<'
ferencd@0: 			// Have we hit a new element or an end tag? This could also be
ferencd@0: 			// a TiXmlText in the "CDATA" style.
ferencd@0: 			if ( StringEqual( p, "</", false, encoding ) )
ferencd@0: 			{
ferencd@0: 				return p;
ferencd@0: 			}
ferencd@0: 			else
ferencd@0: 			{
ferencd@0: 				TiXmlNode* node = Identify( p, encoding );
ferencd@0: 				if ( node )
ferencd@0: 				{
ferencd@0: 					p = node->Parse( p, data, encoding );
ferencd@0: 					LinkEndChild( node );
ferencd@0: 				}				
ferencd@0: 				else
ferencd@0: 				{
ferencd@0: 					return 0;
ferencd@0: 				}
ferencd@0: 			}
ferencd@0: 		}
ferencd@0: 		pWithWhiteSpace = p;
ferencd@0: 		p = SkipWhiteSpace( p, encoding );
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	if ( !p )
ferencd@0: 	{
ferencd@0: 		if ( document ) document->SetError( TIXML_ERROR_READING_ELEMENT_VALUE, 0, 0, encoding );
ferencd@0: 	}	
ferencd@0: 	return p;
ferencd@0: }
ferencd@0: 
ferencd@0: 
ferencd@0: #ifdef TIXML_USE_STL
ferencd@0: void TiXmlUnknown::StreamIn( std::istream * in, TIXML_STRING * tag )
ferencd@0: {
ferencd@0: 	while ( in->good() )
ferencd@0: 	{
ferencd@0: 		int c = in->get();	
ferencd@0: 		if ( c <= 0 )
ferencd@0: 		{
ferencd@0: 			TiXmlDocument* document = GetDocument();
ferencd@0: 			if ( document )
ferencd@0: 				document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 			return;
ferencd@0: 		}
ferencd@0: 		(*tag) += (char) c;
ferencd@0: 
ferencd@0: 		if ( c == '>' )
ferencd@0: 		{
ferencd@0: 			// All is well.
ferencd@0: 			return;		
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: }
ferencd@0: #endif
ferencd@0: 
ferencd@0: 
ferencd@0: const char* TiXmlUnknown::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	TiXmlDocument* document = GetDocument();
ferencd@0: 	p = SkipWhiteSpace( p, encoding );
ferencd@0: 
ferencd@0: 	if ( data )
ferencd@0: 	{
ferencd@0: 		data->Stamp( p, encoding );
ferencd@0: 		location = data->Cursor();
ferencd@0: 	}
ferencd@0: 	if ( !p || !*p || *p != '<' )
ferencd@0: 	{
ferencd@0: 		if ( document ) document->SetError( TIXML_ERROR_PARSING_UNKNOWN, p, data, encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 	++p;
ferencd@0:     value = "";
ferencd@0: 
ferencd@0: 	while ( p && *p && *p != '>' )
ferencd@0: 	{
ferencd@0: 		value += *p;
ferencd@0: 		++p;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	if ( !p )
ferencd@0: 	{
ferencd@0: 		if ( document )	
ferencd@0: 			document->SetError( TIXML_ERROR_PARSING_UNKNOWN, 0, 0, encoding );
ferencd@0: 	}
ferencd@0: 	if ( p && *p == '>' )
ferencd@0: 		return p+1;
ferencd@0: 	return p;
ferencd@0: }
ferencd@0: 
ferencd@0: #ifdef TIXML_USE_STL
ferencd@0: void TiXmlComment::StreamIn( std::istream * in, TIXML_STRING * tag )
ferencd@0: {
ferencd@0: 	while ( in->good() )
ferencd@0: 	{
ferencd@0: 		int c = in->get();	
ferencd@0: 		if ( c <= 0 )
ferencd@0: 		{
ferencd@0: 			TiXmlDocument* document = GetDocument();
ferencd@0: 			if ( document )
ferencd@0: 				document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 			return;
ferencd@0: 		}
ferencd@0: 
ferencd@0: 		(*tag) += (char) c;
ferencd@0: 
ferencd@0: 		if ( c == '>' 
ferencd@0: 			 && tag->at( tag->length() - 2 ) == '-'
ferencd@0: 			 && tag->at( tag->length() - 3 ) == '-' )
ferencd@0: 		{
ferencd@0: 			// All is well.
ferencd@0: 			return;		
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: }
ferencd@0: #endif
ferencd@0: 
ferencd@0: 
ferencd@0: const char* TiXmlComment::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	TiXmlDocument* document = GetDocument();
ferencd@0: 	value = "";
ferencd@0: 
ferencd@0: 	p = SkipWhiteSpace( p, encoding );
ferencd@0: 
ferencd@0: 	if ( data )
ferencd@0: 	{
ferencd@0: 		data->Stamp( p, encoding );
ferencd@0: 		location = data->Cursor();
ferencd@0: 	}
ferencd@0: 	const char* startTag = "<!--";
ferencd@0: 	const char* endTag   = "-->";
ferencd@0: 
ferencd@0: 	if ( !StringEqual( p, startTag, false, encoding ) )
ferencd@0: 	{
ferencd@0: 		if ( document )
ferencd@0: 			document->SetError( TIXML_ERROR_PARSING_COMMENT, p, data, encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 	p += strlen( startTag );
ferencd@0: 
ferencd@0: 	// [ 1475201 ] TinyXML parses entities in comments
ferencd@0: 	// Oops - ReadText doesn't work, because we don't want to parse the entities.
ferencd@0: 	// p = ReadText( p, &value, false, endTag, false, encoding );
ferencd@0: 	//
ferencd@0: 	// from the XML spec:
ferencd@0: 	/*
ferencd@0: 	 [Definition: Comments may appear anywhere in a document outside other markup; in addition, 
ferencd@0: 	              they may appear within the document type declaration at places allowed by the grammar. 
ferencd@0: 				  They are not part of the document's character data; an XML processor MAY, but need not, 
ferencd@0: 				  make it possible for an application to retrieve the text of comments. For compatibility, 
ferencd@0: 				  the string "--" (double-hyphen) MUST NOT occur within comments.] Parameter entity 
ferencd@0: 				  references MUST NOT be recognized within comments.
ferencd@0: 
ferencd@0: 				  An example of a comment:
ferencd@0: 
ferencd@0: 				  <!-- declarations for <head> & <body> -->
ferencd@0: 	*/
ferencd@0: 
ferencd@0:     value = "";
ferencd@0: 	// Keep all the white space.
ferencd@0: 	while (	p && *p && !StringEqual( p, endTag, false, encoding ) )
ferencd@0: 	{
ferencd@0: 		value.append( p, 1 );
ferencd@0: 		++p;
ferencd@0: 	}
ferencd@0: 	if ( p && *p ) 
ferencd@0: 		p += strlen( endTag );
ferencd@0: 
ferencd@0: 	return p;
ferencd@0: }
ferencd@0: 
ferencd@0: 
ferencd@0: const char* TiXmlAttribute::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	p = SkipWhiteSpace( p, encoding );
ferencd@0: 	if ( !p || !*p ) return 0;
ferencd@0: 
ferencd@0: 	if ( data )
ferencd@0: 	{
ferencd@0: 		data->Stamp( p, encoding );
ferencd@0: 		location = data->Cursor();
ferencd@0: 	}
ferencd@0: 	// Read the name, the '=' and the value.
ferencd@0: 	const char* pErr = p;
ferencd@0: 	p = ReadName( p, &name, encoding );
ferencd@0: 	if ( !p || !*p )
ferencd@0: 	{
ferencd@0: 		if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, pErr, data, encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 	p = SkipWhiteSpace( p, encoding );
ferencd@0: 	if ( !p || !*p || *p != '=' )
ferencd@0: 	{
ferencd@0: 		if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, p, data, encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	++p;	// skip '='
ferencd@0: 	p = SkipWhiteSpace( p, encoding );
ferencd@0: 	if ( !p || !*p )
ferencd@0: 	{
ferencd@0: 		if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, p, data, encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 	
ferencd@0: 	const char* end;
ferencd@0: 	const char SINGLE_QUOTE = '\'';
ferencd@0: 	const char DOUBLE_QUOTE = '\"';
ferencd@0: 
ferencd@0: 	if ( *p == SINGLE_QUOTE )
ferencd@0: 	{
ferencd@0: 		++p;
ferencd@0: 		end = "\'";		// single quote in string
ferencd@0: 		p = ReadText( p, &value, false, end, false, encoding );
ferencd@0: 	}
ferencd@0: 	else if ( *p == DOUBLE_QUOTE )
ferencd@0: 	{
ferencd@0: 		++p;
ferencd@0: 		end = "\"";		// double quote in string
ferencd@0: 		p = ReadText( p, &value, false, end, false, encoding );
ferencd@0: 	}
ferencd@0: 	else
ferencd@0: 	{
ferencd@0: 		// All attribute values should be in single or double quotes.
ferencd@0: 		// But this is such a common error that the parser will try
ferencd@0: 		// its best, even without them.
ferencd@0: 		value = "";
ferencd@0: 		while (    p && *p											// existence
ferencd@0: 				&& !IsWhiteSpace( *p )								// whitespace
ferencd@0: 				&& *p != '/' && *p != '>' )							// tag end
ferencd@0: 		{
ferencd@0: 			if ( *p == SINGLE_QUOTE || *p == DOUBLE_QUOTE ) {
ferencd@0: 				// [ 1451649 ] Attribute values with trailing quotes not handled correctly
ferencd@0: 				// We did not have an opening quote but seem to have a 
ferencd@0: 				// closing one. Give up and throw an error.
ferencd@0: 				if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, p, data, encoding );
ferencd@0: 				return 0;
ferencd@0: 			}
ferencd@0: 			value += *p;
ferencd@0: 			++p;
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 	return p;
ferencd@0: }
ferencd@0: 
ferencd@0: #ifdef TIXML_USE_STL
ferencd@0: void TiXmlText::StreamIn( std::istream * in, TIXML_STRING * tag )
ferencd@0: {
ferencd@0: 	while ( in->good() )
ferencd@0: 	{
ferencd@0: 		int c = in->peek();	
ferencd@0: 		if ( !cdata && (c == '<' ) ) 
ferencd@0: 		{
ferencd@0: 			return;
ferencd@0: 		}
ferencd@0: 		if ( c <= 0 )
ferencd@0: 		{
ferencd@0: 			TiXmlDocument* document = GetDocument();
ferencd@0: 			if ( document )
ferencd@0: 				document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 			return;
ferencd@0: 		}
ferencd@0: 
ferencd@0: 		(*tag) += (char) c;
ferencd@0: 		in->get();	// "commits" the peek made above
ferencd@0: 
ferencd@0: 		if ( cdata && c == '>' && tag->size() >= 3 ) {
ferencd@0: 			size_t len = tag->size();
ferencd@0: 			if ( (*tag)[len-2] == ']' && (*tag)[len-3] == ']' ) {
ferencd@0: 				// terminator of cdata.
ferencd@0: 				return;
ferencd@0: 			}
ferencd@0: 		}    
ferencd@0: 	}
ferencd@0: }
ferencd@0: #endif
ferencd@0: 
ferencd@0: const char* TiXmlText::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
ferencd@0: {
ferencd@0: 	value = "";
ferencd@0: 	TiXmlDocument* document = GetDocument();
ferencd@0: 
ferencd@0: 	if ( data )
ferencd@0: 	{
ferencd@0: 		data->Stamp( p, encoding );
ferencd@0: 		location = data->Cursor();
ferencd@0: 	}
ferencd@0: 
ferencd@0: 	const char* const startTag = "<![CDATA[";
ferencd@0: 	const char* const endTag   = "]]>";
ferencd@0: 
ferencd@0: 	if ( cdata || StringEqual( p, startTag, false, encoding ) )
ferencd@0: 	{
ferencd@0: 		cdata = true;
ferencd@0: 
ferencd@0: 		if ( !StringEqual( p, startTag, false, encoding ) )
ferencd@0: 		{
ferencd@0: 			if ( document )
ferencd@0: 				document->SetError( TIXML_ERROR_PARSING_CDATA, p, data, encoding );
ferencd@0: 			return 0;
ferencd@0: 		}
ferencd@0: 		p += strlen( startTag );
ferencd@0: 
ferencd@0: 		// Keep all the white space, ignore the encoding, etc.
ferencd@0: 		while (	   p && *p
ferencd@0: 				&& !StringEqual( p, endTag, false, encoding )
ferencd@0: 			  )
ferencd@0: 		{
ferencd@0: 			value += *p;
ferencd@0: 			++p;
ferencd@0: 		}
ferencd@0: 
ferencd@0: 		TIXML_STRING dummy; 
ferencd@0: 		p = ReadText( p, &dummy, false, endTag, false, encoding );
ferencd@0: 		return p;
ferencd@0: 	}
ferencd@0: 	else
ferencd@0: 	{
ferencd@0: 		bool ignoreWhite = true;
ferencd@0: 
ferencd@0: 		const char* end = "<";
ferencd@0: 		p = ReadText( p, &value, ignoreWhite, end, false, encoding );
ferencd@0: 		if ( p && *p )
ferencd@0: 			return p-1;	// don't truncate the '<'
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: }
ferencd@0: 
ferencd@0: #ifdef TIXML_USE_STL
ferencd@0: void TiXmlDeclaration::StreamIn( std::istream * in, TIXML_STRING * tag )
ferencd@0: {
ferencd@0: 	while ( in->good() )
ferencd@0: 	{
ferencd@0: 		int c = in->get();
ferencd@0: 		if ( c <= 0 )
ferencd@0: 		{
ferencd@0: 			TiXmlDocument* document = GetDocument();
ferencd@0: 			if ( document )
ferencd@0: 				document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
ferencd@0: 			return;
ferencd@0: 		}
ferencd@0: 		(*tag) += (char) c;
ferencd@0: 
ferencd@0: 		if ( c == '>' )
ferencd@0: 		{
ferencd@0: 			// All is well.
ferencd@0: 			return;
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: }
ferencd@0: #endif
ferencd@0: 
ferencd@0: const char* TiXmlDeclaration::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding _encoding )
ferencd@0: {
ferencd@0: 	p = SkipWhiteSpace( p, _encoding );
ferencd@0: 	// Find the beginning, find the end, and look for
ferencd@0: 	// the stuff in-between.
ferencd@0: 	TiXmlDocument* document = GetDocument();
ferencd@0: 	if ( !p || !*p || !StringEqual( p, "<?xml", true, _encoding ) )
ferencd@0: 	{
ferencd@0: 		if ( document ) document->SetError( TIXML_ERROR_PARSING_DECLARATION, 0, 0, _encoding );
ferencd@0: 		return 0;
ferencd@0: 	}
ferencd@0: 	if ( data )
ferencd@0: 	{
ferencd@0: 		data->Stamp( p, _encoding );
ferencd@0: 		location = data->Cursor();
ferencd@0: 	}
ferencd@0: 	p += 5;
ferencd@0: 
ferencd@0: 	version = "";
ferencd@0: 	encoding = "";
ferencd@0: 	standalone = "";
ferencd@0: 
ferencd@0: 	while ( p && *p )
ferencd@0: 	{
ferencd@0: 		if ( *p == '>' )
ferencd@0: 		{
ferencd@0: 			++p;
ferencd@0: 			return p;
ferencd@0: 		}
ferencd@0: 
ferencd@0: 		p = SkipWhiteSpace( p, _encoding );
ferencd@0: 		if ( StringEqual( p, "version", true, _encoding ) )
ferencd@0: 		{
ferencd@0: 			TiXmlAttribute attrib;
ferencd@0: 			p = attrib.Parse( p, data, _encoding );		
ferencd@0: 			version = attrib.Value();
ferencd@0: 		}
ferencd@0: 		else if ( StringEqual( p, "encoding", true, _encoding ) )
ferencd@0: 		{
ferencd@0: 			TiXmlAttribute attrib;
ferencd@0: 			p = attrib.Parse( p, data, _encoding );		
ferencd@0: 			encoding = attrib.Value();
ferencd@0: 		}
ferencd@0: 		else if ( StringEqual( p, "standalone", true, _encoding ) )
ferencd@0: 		{
ferencd@0: 			TiXmlAttribute attrib;
ferencd@0: 			p = attrib.Parse( p, data, _encoding );		
ferencd@0: 			standalone = attrib.Value();
ferencd@0: 		}
ferencd@0: 		else
ferencd@0: 		{
ferencd@0: 			// Read over whatever it is.
ferencd@0: 			while( p && *p && *p != '>' && !IsWhiteSpace( *p ) )
ferencd@0: 				++p;
ferencd@0: 		}
ferencd@0: 	}
ferencd@0: 	return 0;
ferencd@0: }
ferencd@0: 
ferencd@0: bool TiXmlText::Blank() const
ferencd@0: {
ferencd@0: 	for ( unsigned i=0; i<value.length(); i++ )
ferencd@0: 		if ( !IsWhiteSpace( value[i] ) )
ferencd@0: 			return false;
ferencd@0: 	return true;
ferencd@0: }
ferencd@0: