add pedantic whitespace mode

3 files changed, 192 insertions, 7 deletions

diff --git a/tinyxml2.cpp b/tinyxml2.cpp
index 4b561b3..901e395 100755
--- a/tinyxml2.cpp
+++ b/tinyxml2.cpp
@@ -715,7 +715,7 @@ bool XMLUtil::ToUnsigned64(const char* str, uint64_t* value) {
 }

 

 

-char* XMLDocument::Identify( char* p, XMLNode** node )

+char* XMLDocument::Identify( char* p, XMLNode** node, bool first )

 {

     TIXMLASSERT( node );

     TIXMLASSERT( p );

@@ -767,9 +767,19 @@ char* XMLDocument::Identify( char* p, XMLNode** node )
         p += dtdHeaderLen;

     }

     else if ( XMLUtil::StringEqual( p, elementHeader, elementHeaderLen ) ) {

-        returnNode =  CreateUnlinkedNode<XMLElement>( _elementPool );

-        returnNode->_parseLineNum = _parseCurLineNum;

-        p += elementHeaderLen;

+

+        // Preserve whitespace pedantically before closing tag, when it's immediately after opening tag

+        if (WhitespaceMode() == PEDANTIC_WHITESPACE && first && p != start && *(p + elementHeaderLen) == '/') {

+            returnNode = CreateUnlinkedNode<XMLText>(_textPool);

+            returnNode->_parseLineNum = startLine;

+            p = start;	// Back it up, all the text counts.

+            _parseCurLineNum = startLine;

+        }

+        else {

+            returnNode = CreateUnlinkedNode<XMLElement>(_elementPool);

+            returnNode->_parseLineNum = _parseCurLineNum;

+            p += elementHeaderLen;

+        }

     }

     else {

         returnNode = CreateUnlinkedNode<XMLText>( _textPool );

@@ -1070,14 +1080,16 @@ char* XMLNode::ParseDeep( char* p, StrPair* parentEndTag, int* curLineNumPtr )
 	if (_document->Error())

 		return 0;

 

+	bool first = true;

 	while( p && *p ) {

         XMLNode* node = 0;

 

-        p = _document->Identify( p, &node );

+        p = _document->Identify( p, &node, first );

         TIXMLASSERT( p );

         if ( node == 0 ) {

             break;

         }

+        first = false;

 

        const int initialLineNum = node->_parseLineNum;

 

diff --git a/tinyxml2.h b/tinyxml2.h
index b0c8b6c..45d6980 100755
--- a/tinyxml2.h
+++ b/tinyxml2.h
@@ -1704,7 +1704,8 @@ private:
 

 enum Whitespace {

     PRESERVE_WHITESPACE,

-    COLLAPSE_WHITESPACE

+    COLLAPSE_WHITESPACE,

+    PEDANTIC_WHITESPACE

 };

 

 

@@ -1915,7 +1916,7 @@ public:
 	void DeepCopy(XMLDocument* target) const;

 

 	// internal

-    char* Identify( char* p, XMLNode** node );

+    char* Identify( char* p, XMLNode** node, bool first );

 

 	// internal

 	void MarkInUse(const XMLNode* const);

diff --git a/xmltest.cpp b/xmltest.cpp
index c3ce079..ae97604 100755
--- a/xmltest.cpp
+++ b/xmltest.cpp
@@ -1869,6 +1869,178 @@ int main( int argc, const char ** argv )
 		XMLTest( "Whitespace  all space", true, 0 == doc.FirstChildElement()->FirstChild() );

 	}

 

+	// ----------- Preserve Whitespace ------------

+	{

+		const char* xml = "<element>This  is  &apos;  \n\n text &apos;</element>";

+		XMLDocument doc(true, PRESERVE_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with whitespace preserved", false, doc.Error());

+		XMLTest("Whitespace preserved", "This  is  '  \n\n text '", doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element> This \nis &apos;  text  &apos;  </element>";

+		XMLDocument doc(true, PRESERVE_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with whitespace preserved", false, doc.Error());

+		XMLTest("Whitespace preserved", " This \nis '  text  '  ", doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>  \n This is &apos; text &apos;  \n</element>";

+		XMLDocument doc(true, PRESERVE_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with whitespace preserved", false, doc.Error());

+		XMLTest("Whitespace preserved", "  \n This is ' text '  \n", doc.FirstChildElement()->GetText());

+	}

+

+	// Following cases are for text that is all whitespace which are not preserved intentionally

+	{

+		const char* xml = "<element> </element>";

+		XMLDocument doc(true, PRESERVE_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with whitespace preserved", false, doc.Error());

+		XMLTest("Whitespace preserved", true, 0 == doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>   </element>";

+		XMLDocument doc(true, PRESERVE_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with whitespace preserved", false, doc.Error());

+		XMLTest("Whitespace preserved", true, 0 == doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>\n\n</element>";

+		XMLDocument doc(true, PRESERVE_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with whitespace preserved", false, doc.Error());

+		XMLTest("Whitespace preserved", true, 0 == doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>  \n</element>";

+		XMLDocument doc(true, PRESERVE_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with whitespace preserved", false, doc.Error());

+		XMLTest("Whitespace preserved", true, 0 == doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element> \n \n </element>";

+		XMLDocument doc(true, PRESERVE_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with whitespace preserved", false, doc.Error());

+		XMLTest("Whitespace preserved", true, 0 == doc.FirstChildElement()->GetText());

+	}

+

+	// ----------- Pedantic Whitespace ------------

+	{

+		const char* xml = "<element>This  is  &apos;  \n\n text &apos;</element>";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", "This  is  '  \n\n text '", doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element> This \nis &apos;  text  &apos;  </element>";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", " This \nis '  text  '  ", doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>  \n This is &apos; text &apos;  \n</element>";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", "  \n This is ' text '  \n", doc.FirstChildElement()->GetText());

+	}

+

+	// Following cases are for text that is all whitespace which is preserved with pedantic mode

+	{

+		const char* xml = "<element> </element>";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", " ", doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>   </element>";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", "   ", doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>\n\n</element>\n";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", "\n\n", doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>  \n</element> \n ";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", "  \n", doc.FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element> \n  \n </element>  ";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", " \n  \n ", doc.FirstChildElement()->GetText());

+	}

+

+	// Following cases are for checking nested elements are still parsed with pedantic whitespace

+	{

+		const char* xml = "<element>\n\t<a> This is nested text </a>\n</element>  ";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse nested elements with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", " This is nested text ", doc.RootElement()->FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>  <b> </b>  </element>\n";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse nested elements with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", " ", doc.RootElement()->FirstChildElement()->GetText());

+	}

+

+	{

+		const char* xml = "<element>  <c attribute=\"test\"/>  </element>\n ";

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.Parse(xml);

+		XMLTest("Parse nested elements with pedantic whitespace", false, doc.Error());

+		XMLTest("Pedantic whitespace", true, 0 == doc.RootElement()->FirstChildElement()->GetText());

+	}

+

+	// Check sample xml can be parsed with pedantic mode

+	{

+		XMLDocument doc(true, PEDANTIC_WHITESPACE);

+		doc.LoadFile("resources/dream.xml");

+		XMLTest("Load dream.xml with pedantic whitespace mode", false, doc.Error());

+

+		XMLTest("Dream", "xml version=\"1.0\"",

+			doc.FirstChild()->ToDeclaration()->Value());

+		XMLTest("Dream", true, doc.FirstChild()->NextSibling()->ToUnknown() != 0);

+		XMLTest("Dream", "DOCTYPE PLAY SYSTEM \"play.dtd\"",

+			doc.FirstChild()->NextSibling()->ToUnknown()->Value());

+		XMLTest("Dream", "And Robin shall restore amends.",

+			doc.LastChild()->LastChild()->LastChild()->LastChild()->LastChildElement()->GetText());

+	}

+

 	{

 		// An assert should not fire.

 		const char* xml = "<element/>";