Added node_declaration node type for <?xml nodes, added corresponding parse_declaration and format_no_declaration flags and parsing/saving/DOM functionality

git-svn-id: http://pugixml.googlecode.com/svn/trunk@104 99668b35-9821-0410-8761-19e4c4f06640

src/pugixml.cpp

@@ -859,7 +859,7 @@ namespace
 
 						if (!is_chartype(*s, ct_start_symbol)) // bad PI
 							return false;
-						else if (OPTSET(parse_pi))
+						else if (OPTSET(parse_pi) || OPTSET(parse_declaration))
 						{
 							mark = s;
 							SCANWHILE(is_chartype(*s, ct_symbol)); // Read PI target
@@ -880,8 +880,16 @@ namespace
 								if ((mark[0] == 'x' || mark[0] == 'X') && (mark[1] == 'm' || mark[1] == 'M')
 									&& (mark[2] == 'l' || mark[2] == 'L') && mark[3] == 0)
 								{
+									if (OPTSET(parse_declaration))
+									{
+										PUSHNODE(node_declaration);
+
+										cursor->name = mark;
+
+										POPNODE();
 									}
-								else
+								}
+								else if (OPTSET(parse_pi))
 								{
 									PUSHNODE(node_pi); // Append a new node on the tree.
 
@@ -894,15 +902,35 @@ namespace
 							else if ((mark[0] == 'x' || mark[0] == 'X') && (mark[1] == 'm' || mark[1] == 'M')
 								&& (mark[2] == 'l' || mark[2] == 'L') && mark[3] == 0)
 							{
+								if (OPTSET(parse_declaration))
+								{
+									PUSHNODE(node_declaration);
+
+									cursor->name = mark;
+
+									// scan for tag end
+									mark = s;
+
 									SCANFOR(*s == '?' && *(s+1) == '>'); // Look for '?>'.
 									CHECK_ERROR();
-								s += 2;
+
+									// replace ending ? with / to terminate properly
+									*s = '/';
+
+									// parse attributes
+									s = mark;
+
+									goto LOC_ATTRIBUTES;
+								}
 							}
 							else
+							{
+								if (OPTSET(parse_pi))
 								{
 									PUSHNODE(node_pi); // Append a new node on the tree.
 
 									cursor->name = mark;
+								}
 
 								if (is_chartype(ch, ct_space))
 								{
@@ -921,11 +949,14 @@ namespace
 
 								++s; // Step over >
 
+								if (OPTSET(parse_pi))
+								{
 									cursor->value = mark;
 
 									POPNODE();
 								}
 							}
+						}
 						else // not parsing PI
 						{
 							SCANFOR(*s == '?' && *(s+1) == '>'); // Look for '?>'.
@@ -1087,6 +1118,7 @@ namespace
 						}
 						else if (is_chartype(ch, ct_space))
 						{
+						LOC_ATTRIBUTES:
 						    while (*s)
 						    {
 								SKIPWS(); // Eat any whitespace.
@@ -1537,6 +1569,28 @@ namespace
 			if ((flags & format_raw) == 0) writer.write('\n');
 			break;
 		
+		case node_declaration:
+		{
+			writer.write("<?");
+			writer.write(node.name());
+
+			for (xml_attribute a = node.first_attribute(); a; a = a.next_attribute())
+			{
+				writer.write(' ');
+				writer.write(a.name());
+				writer.write('=');
+				writer.write('"');
+
+				text_output_escaped(writer, a.value(), opt1_to_type<1>());
+
+				writer.write('"');
+			}
+
+			writer.write("?>");
+			if ((flags & format_raw) == 0) writer.write('\n');
+			break;
+		}
+
 		default:
 			;
 		}
@@ -2063,6 +2117,7 @@ namespace pugi
 		switch (type())
 		{
 		case node_pi:
+		case node_declaration:
 		case node_element:
 		{
 			bool insitu = _root->name_insitu;
@@ -2100,7 +2155,7 @@ namespace pugi
 
 	xml_attribute xml_node::append_attribute(const char* name)
 	{
-		if (type() != node_element) return xml_attribute();
+		if (type() != node_element && type() != node_declaration) return xml_attribute();
 		
 		xml_attribute a(_root->append_attribute(get_allocator()));
 		a.set_name(name);
@@ -2110,7 +2165,7 @@ namespace pugi
 
 	xml_attribute xml_node::insert_attribute_before(const char* name, const xml_attribute& attr)
 	{
-		if (type() != node_element || attr.empty()) return xml_attribute();
+		if ((type() != node_element && type() != node_declaration) || attr.empty()) return xml_attribute();
 		
 		// check that attribute belongs to *this
 		xml_attribute_struct* cur = attr._attr;
@@ -2136,7 +2191,7 @@ namespace pugi
 
 	xml_attribute xml_node::insert_attribute_after(const char* name, const xml_attribute& attr)
 	{
-		if (type() != node_element || attr.empty()) return xml_attribute();
+		if ((type() != node_element && type() != node_declaration) || attr.empty()) return xml_attribute();
 		
 		// check that attribute belongs to *this
 		xml_attribute_struct* cur = attr._attr;
@@ -2738,8 +2793,11 @@ namespace pugi
 			buffered_writer.write(utf8_bom, 3);
 		}
 
+		if (!(flags & format_no_declaration))
+		{
 			buffered_writer.write("<?xml version=\"1.0\"?>");
 			if (!(flags & format_raw)) buffered_writer.write("\n");
+		}
 
 		node_output(buffered_writer, *this, indent, flags, 0);
 	}

src/pugixml.hpp

@@ -42,7 +42,8 @@ namespace pugi
 		node_pcdata,		///< E.g. '>...<'
 		node_cdata,			///< E.g. '<![CDATA[...]]>'
 		node_comment,		///< E.g. '<!--...-->'
-		node_pi				///< E.g. '<?...?>'
+		node_pi,			///< E.g. '<?...?>'
+		node_declaration	///< E.g. '<?xml ...?>'
 	};
 
 	// Parsing options
@@ -154,6 +155,18 @@ namespace pugi
  	 */
  	const unsigned int parse_wconv_attribute	= 0x0080;
 	
+	/**
+	 * This flag determines if XML document declaration (this node has the form of <?xml ... ?> in XML)
+	 * are to be put in DOM tree. If this flag is off, it is not put in the tree, but is still parsed
+	 * and checked for correctness.
+	 *
+	 * The corresponding node in DOM tree will have type node_declaration, name "xml" and attributes,
+	 * if any.
+	 *
+	 * This flag is off by default.
+	 */
+	const unsigned int parse_declaration		= 0x0100;
+
 	/**
      * This is the default set of flags. It includes parsing CDATA sections (comments/PIs are not
      * parsed), performing character and entity reference expansion, replacing whitespace characters
@@ -187,6 +200,15 @@ namespace pugi
 	 */
 	const unsigned int format_raw		= 0x04;
 	
+	/**
+	 * If this flag is on, no default XML declaration is written to output file.
+	 * This means that there will be no XML declaration in output stream unless there was one in XML document
+	 * (i.e. if it was parsed with parse_declaration flag).
+	 *
+	 * This flag is off by default.
+	 */
+	const unsigned int format_no_declaration = 0x08;
+
 	/**
 	 * This is the default set of formatting flags. It includes indenting nodes depending on their
 	 * depth in DOM tree.