2 www.sourceforge.net/projects/tinyxml
3 Original code (2.0 and earlier )copyright (c) 2000-2002 Lee Thomason (www.grinninglizard.com)
5 This software is provided 'as-is', without any express or implied
6 warranty. In no event will the authors be held liable for any
7 damages arising from the use of this software.
9 Permission is granted to anyone to use this software for any
10 purpose, including commercial applications, and to alter it and
11 redistribute it freely, subject to the following restrictions:
13 1. The origin of this software must not be misrepresented; you must
14 not claim that you wrote the original software. If you use this
15 software in a product, an acknowledgment in the product documentation
16 would be appreciated but is not required.
18 2. Altered source versions must be plainly marked as such, and
19 must not be misrepresented as being the original software.
21 3. This notice may not be removed or altered from any source
30 //#define DEBUG_PARSER
31 #if defined( DEBUG_PARSER )
32 # if defined( DEBUG ) && defined( _MSC_VER )
34 # define TIXML_LOG OutputDebugString
36 # define TIXML_LOG printf
40 // Note tha "PutString" hardcodes the same list. This
41 // is less flexible than it appears. Changing the entries
42 // or order will break putstring.
43 TiXmlBase::Entity TiXmlBase::entity[ NUM_ENTITY ] =
48 { """, 6, '\"' },
52 // Bunch of unicode info at:
53 // http://www.unicode.org/faq/utf_bom.html
54 // Including the basic of this table, which determines the #bytes in the
55 // sequence from the lead byte. 1 placed for invalid sequences --
56 // although the result will be junk, pass it through as much as possible.
57 // Beware of the non-characters in UTF-8:
58 // ef bb bf (Microsoft "lead bytes")
62 const unsigned char TIXML_UTF_LEAD_0 = 0xefU;
63 const unsigned char TIXML_UTF_LEAD_1 = 0xbbU;
64 const unsigned char TIXML_UTF_LEAD_2 = 0xbfU;
66 const int TiXmlBase::utf8ByteTable[256] =
68 // 0 1 2 3 4 5 6 7 8 9 a b c d e f
69 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x00
70 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x10
71 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x20
72 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x30
73 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x40
74 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x50
75 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x60
76 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x70 End of ASCII range
77 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x80 0x80 to 0xc1 invalid
78 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0x90
79 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0xa0
80 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 0xb0
81 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // 0xc0 0xc2 to 0xdf 2 byte
82 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // 0xd0
83 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, // 0xe0 0xe0 to 0xef 3 byte
84 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
88 void TiXmlBase::ConvertUTF32ToUTF8( unsigned long input, char* output, int* length )
90 const unsigned long BYTE_MASK = 0xBF;
91 const unsigned long BYTE_MARK = 0x80;
92 const unsigned long FIRST_BYTE_MARK[7] = { 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
96 else if ( input < 0x800 )
98 else if ( input < 0x10000 )
100 else if ( input < 0x200000 )
103 { *length = 0; return; } // This code won't covert this correctly anyway.
107 // Scary scary fall throughs.
112 *output = (char)((input | BYTE_MARK) & BYTE_MASK);
116 *output = (char)((input | BYTE_MARK) & BYTE_MASK);
120 *output = (char)((input | BYTE_MARK) & BYTE_MASK);
124 *output = (char)(input | FIRST_BYTE_MARK[*length]);
129 /*static*/ int TiXmlBase::IsAlpha( unsigned char anyByte, TiXmlEncoding /*encoding*/ )
131 // This will only work for low-ascii, everything else is assumed to be a valid
132 // letter. I'm not sure this is the best approach, but it is quite tricky trying
133 // to figure out alhabetical vs. not across encoding. So take a very
134 // conservative approach.
136 // if ( encoding == TIXML_ENCODING_UTF8 )
139 return isalpha( anyByte );
141 return 1; // What else to do? The unicode set is huge...get the english ones right.
145 // return isalpha( anyByte );
150 /*static*/ int TiXmlBase::IsAlphaNum( unsigned char anyByte, TiXmlEncoding /*encoding*/ )
152 // This will only work for low-ascii, everything else is assumed to be a valid
153 // letter. I'm not sure this is the best approach, but it is quite tricky trying
154 // to figure out alhabetical vs. not across encoding. So take a very
155 // conservative approach.
157 // if ( encoding == TIXML_ENCODING_UTF8 )
160 return isalnum( anyByte );
162 return 1; // What else to do? The unicode set is huge...get the english ones right.
166 // return isalnum( anyByte );
171 class TiXmlParsingData
173 friend class TiXmlDocument;
175 void Stamp( const char* now, TiXmlEncoding encoding );
177 const TiXmlCursor& Cursor() { return cursor; }
180 // Only used by the document!
181 TiXmlParsingData( const char* start, int _tabsize, int row, int col )
196 void TiXmlParsingData::Stamp( const char* now, TiXmlEncoding encoding )
200 // Do nothing if the tabsize is 0.
206 // Get the current row, column.
207 int row = cursor.row;
208 int col = cursor.col;
209 const char* p = stamp;
214 // Treat p as unsigned, so we have a happy compiler.
215 const unsigned char* pU = (const unsigned char*)p;
217 // Code contributed by Fletcher Dunn: (modified by lee)
220 // We *should* never get here, but in case we do, don't
221 // advance past the terminating null character, ever
225 // bump down to the next line
231 // Check for \r\n sequence, and treat this as a single character
238 // bump down to the next line
245 // Check for \n\r sequence, and treat this as a single
246 // character. (Yes, this bizarre thing does occur still
247 // on some arcane platforms...)
257 // Skip to next tab stop
258 col = (col / tabsize + 1) * tabsize;
261 case TIXML_UTF_LEAD_0:
262 if ( encoding == TIXML_ENCODING_UTF8 )
264 if ( *(p+1) && *(p+2) )
266 // In these cases, don't advance the column. These are
268 if ( *(pU+1)==TIXML_UTF_LEAD_1 && *(pU+2)==TIXML_UTF_LEAD_2 )
270 else if ( *(pU+1)==0xbfU && *(pU+2)==0xbeU )
272 else if ( *(pU+1)==0xbfU && *(pU+2)==0xbfU )
275 { p +=3; ++col; } // A normal character.
286 if ( encoding == TIXML_ENCODING_UTF8 )
288 // Eat the 1 to 4 byte utf8 character.
289 int step = TiXmlBase::utf8ByteTable[*((const unsigned char*)p)];
291 step = 1; // Error case from bad encoding, but handle gracefully.
294 // Just advance one column, of course.
307 assert( cursor.row >= -1 );
308 assert( cursor.col >= -1 );
314 const char* TiXmlBase::SkipWhiteSpace( const char* p, TiXmlEncoding encoding )
320 if ( encoding == TIXML_ENCODING_UTF8 )
324 const unsigned char* pU = (const unsigned char*)p;
326 // Skip the stupid Microsoft UTF-8 Byte order marks
327 if ( *(pU+0)==TIXML_UTF_LEAD_0
328 && *(pU+1)==TIXML_UTF_LEAD_1
329 && *(pU+2)==TIXML_UTF_LEAD_2 )
334 else if(*(pU+0)==TIXML_UTF_LEAD_0
341 else if(*(pU+0)==TIXML_UTF_LEAD_0
349 if ( IsWhiteSpace( *p ) ) // Still using old rules for white space.
357 while ( *p && IsWhiteSpace( *p ) )
365 /*static*/ bool TiXmlBase::StreamWhiteSpace( std::istream * in, TIXML_STRING * tag )
369 if ( !in->good() ) return false;
372 // At this scope, we can't get to a document. So fail silently.
373 if ( !IsWhiteSpace( c ) || c <= 0 )
376 *tag += (char) in->get();
380 /*static*/ bool TiXmlBase::StreamTo( std::istream * in, int character, TIXML_STRING * tag )
382 //assert( character > 0 && character < 128 ); // else it won't work in utf-8
386 if ( c == character )
388 if ( c <= 0 ) // Silent failure: can't get document at this scope
398 // One of TinyXML's more performance demanding functions. Try to keep the memory overhead down. The
399 // "assign" optimization removes over 10% of the execution time.
401 const char* TiXmlBase::ReadName( const char* p, TIXML_STRING * name, TiXmlEncoding encoding )
403 // Oddly, not supported on some comilers,
409 // Names start with letters or underscores.
410 // Of course, in unicode, tinyxml has no idea what a letter *is*. The
411 // algorithm is generous.
413 // After that, they can be letters, underscores, numbers,
414 // hyphens, or colons. (Colons are valid ony for namespaces,
415 // but tinyxml can't tell namespaces from names.)
417 && ( IsAlpha( (unsigned char) *p, encoding ) || *p == '_' ) )
419 const char* start = p;
421 && ( IsAlphaNum( (unsigned char ) *p, encoding )
427 //(*name) += *p; // expensive
431 name->assign( start, p-start );
438 const char* TiXmlBase::GetEntity( const char* p, char* value, int* length, TiXmlEncoding encoding )
440 // Presume an entity, and pull it out.
445 if ( *(p+1) && *(p+1) == '#' && *(p+2) )
447 unsigned long ucs = 0;
454 if ( !*(p+3) ) return 0;
457 q = strchr( q, ';' );
459 if ( !q || !*q ) return 0;
466 if ( *q >= '0' && *q <= '9' )
467 ucs += mult * (*q - '0');
468 else if ( *q >= 'a' && *q <= 'f' )
469 ucs += mult * (*q - 'a' + 10);
470 else if ( *q >= 'A' && *q <= 'F' )
471 ucs += mult * (*q - 'A' + 10 );
481 if ( !*(p+2) ) return 0;
484 q = strchr( q, ';' );
486 if ( !q || !*q ) return 0;
493 if ( *q >= '0' && *q <= '9' )
494 ucs += mult * (*q - '0');
501 if ( encoding == TIXML_ENCODING_UTF8 )
503 // convert the UCS to UTF-8
504 ConvertUTF32ToUTF8( ucs, value, length );
511 return p + delta + 1;
514 // Now try to match it.
515 for( i=0; i<NUM_ENTITY; ++i )
517 if ( strncmp( entity[i].str, p, entity[i].strLength ) == 0 )
519 assert( strlen( entity[i].str ) == entity[i].strLength );
520 *value = entity[i].chr;
522 return ( p + entity[i].strLength );
526 // So it wasn't an entity, its unrecognized, or something like that.
527 *value = *p; // Don't put back the last one, since we return it!
528 //*length = 1; // Leave unrecognized entities - this doesn't really work.
529 // Just writes strange XML.
534 bool TiXmlBase::StringEqual( const char* p,
537 TiXmlEncoding encoding )
551 while ( *q && *tag && ToLower( *q, encoding ) == ToLower( *tag, encoding ) )
562 while ( *q && *tag && *q == *tag )
568 if ( *tag == 0 ) // Have we found the end of the tag, and everything equal?
574 const char* TiXmlBase::ReadText( const char* p,
578 bool caseInsensitive,
579 TiXmlEncoding encoding )
582 if ( !trimWhiteSpace // certain tags always keep whitespace
583 || !condenseWhiteSpace ) // if true, whitespace is always kept
585 // Keep all the white space.
587 && !StringEqual( p, endTag, caseInsensitive, encoding )
591 char cArr[4] = { 0, 0, 0, 0 };
592 p = GetChar( p, cArr, &len, encoding );
593 text->append( cArr, len );
598 bool whitespace = false;
600 // Remove leading white space:
601 p = SkipWhiteSpace( p, encoding );
603 && !StringEqual( p, endTag, caseInsensitive, encoding ) )
605 if ( *p == '\r' || *p == '\n' )
610 else if ( IsWhiteSpace( *p ) )
617 // If we've found whitespace, add it before the
618 // new character. Any whitespace just becomes a space.
625 char cArr[4] = { 0, 0, 0, 0 };
626 p = GetChar( p, cArr, &len, encoding );
628 (*text) += cArr[0]; // more efficient
630 text->append( cArr, len );
635 p += strlen( endTag );
641 void TiXmlDocument::StreamIn( std::istream * in, TIXML_STRING * tag )
643 // The basic issue with a document is that we don't know what we're
644 // streaming. Read something presumed to be a tag (and hope), then
645 // identify it, and call the appropriate stream method on the tag.
647 // This "pre-streaming" will never read the closing ">" so the
648 // sub-tag can orient itself.
650 if ( !StreamTo( in, '<', tag ) )
652 SetError( TIXML_ERROR_PARSING_EMPTY, 0, 0, TIXML_ENCODING_UNKNOWN );
658 int tagIndex = (int) tag->length();
659 while ( in->good() && in->peek() != '>' )
664 SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
672 // We now have something we presume to be a node of
673 // some sort. Identify it, and call the node to
674 // continue streaming.
675 TiXmlNode* node = Identify( tag->c_str() + tagIndex, TIXML_DEFAULT_ENCODING );
679 node->StreamIn( in, tag );
680 bool isElement = node->ToElement() != 0;
684 // If this is the root element, we're done. Parsing will be
685 // done by the >> operator.
693 SetError( TIXML_ERROR, 0, 0, TIXML_ENCODING_UNKNOWN );
698 // We should have returned sooner.
699 SetError( TIXML_ERROR, 0, 0, TIXML_ENCODING_UNKNOWN );
704 const char* TiXmlDocument::Parse( const char* p, TiXmlParsingData* prevData, TiXmlEncoding encoding )
708 // Parse away, at the document level. Since a document
709 // contains nothing but other tags, most of what happens
710 // here is skipping white space.
713 SetError( TIXML_ERROR_DOCUMENT_EMPTY, 0, 0, TIXML_ENCODING_UNKNOWN );
717 // Note that, for a document, this needs to come
718 // before the while space skip, so that parsing
719 // starts from the pointer we are given.
723 location.row = prevData->cursor.row;
724 location.col = prevData->cursor.col;
731 TiXmlParsingData data( p, TabSize(), location.row, location.col );
732 location = data.Cursor();
734 if ( encoding == TIXML_ENCODING_UNKNOWN )
736 // Check for the Microsoft UTF-8 lead bytes.
737 const unsigned char* pU = (const unsigned char*)p;
738 if ( *(pU+0) && *(pU+0) == TIXML_UTF_LEAD_0
739 && *(pU+1) && *(pU+1) == TIXML_UTF_LEAD_1
740 && *(pU+2) && *(pU+2) == TIXML_UTF_LEAD_2 )
742 encoding = TIXML_ENCODING_UTF8;
743 useMicrosoftBOM = true;
747 p = SkipWhiteSpace( p, encoding );
750 SetError( TIXML_ERROR_DOCUMENT_EMPTY, 0, 0, TIXML_ENCODING_UNKNOWN );
756 TiXmlNode* node = Identify( p, encoding );
759 p = node->Parse( p, &data, encoding );
760 LinkEndChild( node );
767 // Did we get encoding info?
768 if ( encoding == TIXML_ENCODING_UNKNOWN
769 && node->ToDeclaration() )
771 TiXmlDeclaration* dec = node->ToDeclaration();
772 const char* enc = dec->Encoding();
776 encoding = TIXML_ENCODING_UTF8;
777 else if ( StringEqual( enc, "UTF-8", true, TIXML_ENCODING_UNKNOWN ) )
778 encoding = TIXML_ENCODING_UTF8;
779 else if ( StringEqual( enc, "UTF8", true, TIXML_ENCODING_UNKNOWN ) )
780 encoding = TIXML_ENCODING_UTF8; // incorrect, but be nice
782 encoding = TIXML_ENCODING_LEGACY;
785 p = SkipWhiteSpace( p, encoding );
790 SetError( TIXML_ERROR_DOCUMENT_EMPTY, 0, 0, encoding );
798 void TiXmlDocument::SetError( int err, const char* pError, TiXmlParsingData* data, TiXmlEncoding encoding )
800 // The first error in a chain is more accurate - don't set again!
804 assert( err > 0 && err < TIXML_ERROR_STRING_COUNT );
807 errorDesc = errorString[ errorId ];
809 errorLocation.Clear();
810 if ( pError && data )
812 data->Stamp( pError, encoding );
813 errorLocation = data->Cursor();
818 TiXmlNode* TiXmlNode::Identify( const char* p, TiXmlEncoding encoding )
820 TiXmlNode* returnNode = 0;
822 p = SkipWhiteSpace( p, encoding );
823 if( !p || !*p || *p != '<' )
828 p = SkipWhiteSpace( p, encoding );
835 // What is this thing?
836 // - Elements start with a letter or underscore, but xml is reserved.
838 // - Decleration: <?xml
839 // - Everthing else is unknown to tinyxml.
842 const char* xmlHeader = { "<?xml" };
843 const char* commentHeader = { "<!--" };
844 const char* dtdHeader = { "<!" };
845 const char* cdataHeader = { "<![CDATA[" };
847 if ( StringEqual( p, xmlHeader, true, encoding ) )
850 TIXML_LOG( "XML parsing Declaration\n" );
852 returnNode = new TiXmlDeclaration();
854 else if ( StringEqual( p, commentHeader, false, encoding ) )
857 TIXML_LOG( "XML parsing Comment\n" );
859 returnNode = new TiXmlComment();
861 else if ( StringEqual( p, cdataHeader, false, encoding ) )
864 TIXML_LOG( "XML parsing CDATA\n" );
866 TiXmlText* text = new TiXmlText( "" );
867 text->SetCDATA( true );
870 else if ( StringEqual( p, dtdHeader, false, encoding ) )
873 TIXML_LOG( "XML parsing Unknown(1)\n" );
875 returnNode = new TiXmlUnknown();
877 else if ( IsAlpha( *(p+1), encoding )
881 TIXML_LOG( "XML parsing Element\n" );
883 returnNode = new TiXmlElement( "" );
888 TIXML_LOG( "XML parsing Unknown(2)\n" );
890 returnNode = new TiXmlUnknown();
895 // Set the parent, so it can report errors
896 returnNode->parent = this;
903 void TiXmlElement::StreamIn (std::istream * in, TIXML_STRING * tag)
905 // We're called with some amount of pre-parsing. That is, some of "this"
906 // element is in "tag". Go ahead and stream to the closing ">"
912 TiXmlDocument* document = GetDocument();
914 document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
923 if ( tag->length() < 3 ) return;
925 // Okay...if we are a "/>" tag, then we're done. We've read a complete tag.
926 // If not, identify and stream.
928 if ( tag->at( tag->length() - 1 ) == '>'
929 && tag->at( tag->length() - 2 ) == '/' )
934 else if ( tag->at( tag->length() - 1 ) == '>' )
936 // There is more. Could be:
938 // cdata text (which looks like another node)
943 StreamWhiteSpace( in, tag );
946 if ( in->good() && in->peek() != '<' )
949 TiXmlText text( "" );
950 text.StreamIn( in, tag );
952 // What follows text is a closing tag or another node.
953 // Go around again and figure it out.
957 // We now have either a closing tag...or another node.
958 // We should be at a "<", regardless.
959 if ( !in->good() ) return;
960 assert( in->peek() == '<' );
961 int tagIndex = (int) tag->length();
963 bool closingTag = false;
964 bool firstCharFound = false;
974 TiXmlDocument* document = GetDocument();
976 document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
986 // Early out if we find the CDATA id.
987 if ( c == '[' && tag->size() >= 9 )
989 size_t len = tag->size();
990 const char* start = tag->c_str() + len - 9;
991 if ( strcmp( start, "<![CDATA[" ) == 0 ) {
992 assert( !closingTag );
997 if ( !firstCharFound && c != '<' && !IsWhiteSpace( c ) )
999 firstCharFound = true;
1004 // If it was a closing tag, then read in the closing '>' to clean up the input stream.
1005 // If it was not, the streaming will be done by the tag.
1014 TiXmlDocument* document = GetDocument();
1016 document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
1022 // We are done, once we've found our closing tag.
1027 // If not a closing tag, id it, and stream.
1028 const char* tagloc = tag->c_str() + tagIndex;
1029 TiXmlNode* node = Identify( tagloc, TIXML_DEFAULT_ENCODING );
1032 node->StreamIn( in, tag );
1036 // No return: go around from the beginning: text, closing tag, or node.
1043 const char* TiXmlElement::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
1045 p = SkipWhiteSpace( p, encoding );
1046 TiXmlDocument* document = GetDocument();
1050 if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, 0, 0, encoding );
1056 data->Stamp( p, encoding );
1057 location = data->Cursor();
1062 if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, p, data, encoding );
1066 p = SkipWhiteSpace( p+1, encoding );
1069 const char* pErr = p;
1071 p = ReadName( p, &value, encoding );
1074 if ( document ) document->SetError( TIXML_ERROR_FAILED_TO_READ_ELEMENT_NAME, pErr, data, encoding );
1078 TIXML_STRING endTag ("</");
1081 // Check for and read attributes. Also look for an empty
1082 // tag or an end tag.
1086 p = SkipWhiteSpace( p, encoding );
1089 if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, pErr, data, encoding );
1098 if ( document ) document->SetError( TIXML_ERROR_PARSING_EMPTY, p, data, encoding );
1103 else if ( *p == '>' )
1105 // Done with attributes (if there were any.)
1106 // Read the value -- which can include other
1107 // elements -- read the end tag, and return.
1109 p = ReadValue( p, data, encoding ); // Note this is an Element method, and will set the error if one happens.
1111 // We were looking for the end tag, but found nothing.
1112 // Fix for [ 1663758 ] Failure to report error on bad XML
1113 if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data, encoding );
1117 // We should find the end tag now
1121 // are both valid end tags.
1122 if ( StringEqual( p, endTag.c_str(), false, encoding ) )
1124 p += endTag.length();
1125 p = SkipWhiteSpace( p, encoding );
1126 if ( p && *p && *p == '>' ) {
1130 if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data, encoding );
1135 if ( document ) document->SetError( TIXML_ERROR_READING_END_TAG, p, data, encoding );
1141 // Try to read an attribute:
1142 TiXmlAttribute* attrib = new TiXmlAttribute();
1148 attrib->SetDocument( document );
1150 p = attrib->Parse( p, data, encoding );
1154 if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, pErr, data, encoding );
1159 // Handle the strange case of double attributes:
1160 #ifdef TIXML_USE_STL
1161 TiXmlAttribute* node = attributeSet.Find( attrib->NameTStr() );
1163 TiXmlAttribute* node = attributeSet.Find( attrib->Name() );
1167 if ( document ) document->SetError( TIXML_ERROR_PARSING_ELEMENT, pErr, data, encoding );
1172 attributeSet.Add( attrib );
1179 const char* TiXmlElement::ReadValue( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
1181 TiXmlDocument* document = GetDocument();
1183 // Read in text and elements in any order.
1184 const char* pWithWhiteSpace = p;
1185 p = SkipWhiteSpace( p, encoding );
1191 // Take what we have, make a text element.
1192 TiXmlText* textNode = new TiXmlText( "" );
1199 if ( TiXmlBase::IsWhiteSpaceCondensed() )
1201 p = textNode->Parse( p, data, encoding );
1205 // Special case: we want to keep the white space
1206 // so that leading spaces aren't removed.
1207 p = textNode->Parse( pWithWhiteSpace, data, encoding );
1210 if ( !textNode->Blank() )
1211 LinkEndChild( textNode );
1218 // Have we hit a new element or an end tag? This could also be
1219 // a TiXmlText in the "CDATA" style.
1220 if ( StringEqual( p, "</", false, encoding ) )
1226 TiXmlNode* node = Identify( p, encoding );
1229 p = node->Parse( p, data, encoding );
1230 LinkEndChild( node );
1238 pWithWhiteSpace = p;
1239 p = SkipWhiteSpace( p, encoding );
1244 if ( document ) document->SetError( TIXML_ERROR_READING_ELEMENT_VALUE, 0, 0, encoding );
1250 #ifdef TIXML_USE_STL
1251 void TiXmlUnknown::StreamIn( std::istream * in, TIXML_STRING * tag )
1253 while ( in->good() )
1258 TiXmlDocument* document = GetDocument();
1260 document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
1275 const char* TiXmlUnknown::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
1277 TiXmlDocument* document = GetDocument();
1278 p = SkipWhiteSpace( p, encoding );
1282 data->Stamp( p, encoding );
1283 location = data->Cursor();
1285 if ( !p || !*p || *p != '<' )
1287 if ( document ) document->SetError( TIXML_ERROR_PARSING_UNKNOWN, p, data, encoding );
1293 while ( p && *p && *p != '>' )
1301 if ( document ) document->SetError( TIXML_ERROR_PARSING_UNKNOWN, 0, 0, encoding );
1308 #ifdef TIXML_USE_STL
1309 void TiXmlComment::StreamIn( std::istream * in, TIXML_STRING * tag )
1311 while ( in->good() )
1316 TiXmlDocument* document = GetDocument();
1318 document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
1325 && tag->at( tag->length() - 2 ) == '-'
1326 && tag->at( tag->length() - 3 ) == '-' )
1336 const char* TiXmlComment::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
1338 TiXmlDocument* document = GetDocument();
1341 p = SkipWhiteSpace( p, encoding );
1345 data->Stamp( p, encoding );
1346 location = data->Cursor();
1348 const char* startTag = "<!--";
1349 const char* endTag = "-->";
1351 if ( !StringEqual( p, startTag, false, encoding ) )
1353 document->SetError( TIXML_ERROR_PARSING_COMMENT, p, data, encoding );
1356 p += strlen( startTag );
1358 // [ 1475201 ] TinyXML parses entities in comments
1359 // Oops - ReadText doesn't work, because we don't want to parse the entities.
1360 // p = ReadText( p, &value, false, endTag, false, encoding );
1362 // from the XML spec:
1364 [Definition: Comments may appear anywhere in a document outside other markup; in addition,
1365 they may appear within the document type declaration at places allowed by the grammar.
1366 They are not part of the document's character data; an XML processor MAY, but need not,
1367 make it possible for an application to retrieve the text of comments. For compatibility,
1368 the string "--" (double-hyphen) MUST NOT occur within comments.] Parameter entity
1369 references MUST NOT be recognized within comments.
1371 An example of a comment:
1373 <!-- declarations for <head> & <body> -->
1377 // Keep all the white space.
1378 while ( p && *p && !StringEqual( p, endTag, false, encoding ) )
1380 value.append( p, 1 );
1384 p += strlen( endTag );
1390 const char* TiXmlAttribute::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
1392 p = SkipWhiteSpace( p, encoding );
1393 if ( !p || !*p ) return 0;
1397 data->Stamp( p, encoding );
1398 location = data->Cursor();
1400 // Read the name, the '=' and the value.
1401 const char* pErr = p;
1402 p = ReadName( p, &name, encoding );
1405 if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, pErr, data, encoding );
1408 p = SkipWhiteSpace( p, encoding );
1409 if ( !p || !*p || *p != '=' )
1411 if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, p, data, encoding );
1416 p = SkipWhiteSpace( p, encoding );
1419 if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, p, data, encoding );
1424 const char SINGLE_QUOTE = '\'';
1425 const char DOUBLE_QUOTE = '\"';
1427 if ( *p == SINGLE_QUOTE )
1430 end = "\'"; // single quote in string
1431 p = ReadText( p, &value, false, end, false, encoding );
1433 else if ( *p == DOUBLE_QUOTE )
1436 end = "\""; // double quote in string
1437 p = ReadText( p, &value, false, end, false, encoding );
1441 // All attribute values should be in single or double quotes.
1442 // But this is such a common error that the parser will try
1443 // its best, even without them.
1445 while ( p && *p // existence
1446 && !IsWhiteSpace( *p ) // whitespace
1447 && *p != '/' && *p != '>' ) // tag end
1449 if ( *p == SINGLE_QUOTE || *p == DOUBLE_QUOTE ) {
1450 // [ 1451649 ] Attribute values with trailing quotes not handled correctly
1451 // We did not have an opening quote but seem to have a
1452 // closing one. Give up and throw an error.
1453 if ( document ) document->SetError( TIXML_ERROR_READING_ATTRIBUTES, p, data, encoding );
1463 #ifdef TIXML_USE_STL
1464 void TiXmlText::StreamIn( std::istream * in, TIXML_STRING * tag )
1466 while ( in->good() )
1469 if ( !cdata && (c == '<' ) )
1475 TiXmlDocument* document = GetDocument();
1477 document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
1482 in->get(); // "commits" the peek made above
1484 if ( cdata && c == '>' && tag->size() >= 3 ) {
1485 size_t len = tag->size();
1486 if ( (*tag)[len-2] == ']' && (*tag)[len-3] == ']' ) {
1487 // terminator of cdata.
1495 const char* TiXmlText::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding encoding )
1498 TiXmlDocument* document = GetDocument();
1502 data->Stamp( p, encoding );
1503 location = data->Cursor();
1506 const char* const startTag = "<![CDATA[";
1507 const char* const endTag = "]]>";
1509 if ( cdata || StringEqual( p, startTag, false, encoding ) )
1513 if ( !StringEqual( p, startTag, false, encoding ) )
1515 document->SetError( TIXML_ERROR_PARSING_CDATA, p, data, encoding );
1518 p += strlen( startTag );
1520 // Keep all the white space, ignore the encoding, etc.
1522 && !StringEqual( p, endTag, false, encoding )
1530 p = ReadText( p, &dummy, false, endTag, false, encoding );
1535 bool ignoreWhite = true;
1537 const char* end = "<";
1538 p = ReadText( p, &value, ignoreWhite, end, false, encoding );
1540 return p-1; // don't truncate the '<'
1545 #ifdef TIXML_USE_STL
1546 void TiXmlDeclaration::StreamIn( std::istream * in, TIXML_STRING * tag )
1548 while ( in->good() )
1553 TiXmlDocument* document = GetDocument();
1555 document->SetError( TIXML_ERROR_EMBEDDED_NULL, 0, 0, TIXML_ENCODING_UNKNOWN );
1569 const char* TiXmlDeclaration::Parse( const char* p, TiXmlParsingData* data, TiXmlEncoding _encoding )
1571 p = SkipWhiteSpace( p, _encoding );
1572 // Find the beginning, find the end, and look for
1573 // the stuff in-between.
1574 TiXmlDocument* document = GetDocument();
1575 if ( !p || !*p || !StringEqual( p, "<?xml", true, _encoding ) )
1577 if ( document ) document->SetError( TIXML_ERROR_PARSING_DECLARATION, 0, 0, _encoding );
1582 data->Stamp( p, _encoding );
1583 location = data->Cursor();
1599 p = SkipWhiteSpace( p, _encoding );
1600 if ( StringEqual( p, "version", true, _encoding ) )
1602 TiXmlAttribute attrib;
1603 p = attrib.Parse( p, data, _encoding );
1604 version = attrib.Value();
1606 else if ( StringEqual( p, "encoding", true, _encoding ) )
1608 TiXmlAttribute attrib;
1609 p = attrib.Parse( p, data, _encoding );
1610 encoding = attrib.Value();
1612 else if ( StringEqual( p, "standalone", true, _encoding ) )
1614 TiXmlAttribute attrib;
1615 p = attrib.Parse( p, data, _encoding );
1616 standalone = attrib.Value();
1620 // Read over whatever it is.
1621 while( p && *p && *p != '>' && !IsWhiteSpace( *p ) )
1628 bool TiXmlText::Blank() const
1630 for ( unsigned i=0; i<value.length(); i++ )
1631 if ( !IsWhiteSpace( value[i] ) )