[SuppressMessage("Microsoft.Maintainability", "CA1502")] // Most of the complexity is due to simple switch statements. Not sure if it would truly be less complex if it was split into different methods...
private void Parse()
{
if (m_ParseState == HtmlParseState.BetweenAttributes)
{
m_AttributeLineNumber = m_LineNumber;
m_AttributeLinePosition = m_LinePosition;
}
// keep track of the previous character to find empty element nodes.
char? previousChar = null;
bool setPreviousChar = false; // only set it after the first loop
char currentChar = '\0'; // only setting this here to shut up the compiler.
while (true)
{
// All node types end at the '>' character with the exception of the Text node.
// So, for the Text parse state, peek at the next character rather than read it.
// If it is the '<' character, end the node.
if (m_ParseState == HtmlParseState.Text && (char)((ushort)m_Reader.Peek()) == '<')
{
EndNode(previousChar);
break; // break out of the while loop - done parsing the Text node
}
if (setPreviousChar) previousChar = currentChar;
currentChar = NextChar();
setPreviousChar = true;
switch (m_ParseState)
{
case HtmlParseState.AttributeAfterEquals:
if (currentChar == '>')
{
AddAttribute();
EndNode(previousChar);
break;
}
else if (!Char.IsWhiteSpace(currentChar))
{
if (currentChar == '\'' || currentChar == '"')
{
m_QuoteChar = currentChar;
m_ParseState = HtmlParseState.QuotedAttributeValue;
// continue parsing
}
else // non-quoted attribute value
{
AppendToAttributeValue(currentChar);
m_ParseState = HtmlParseState.AttributeValue;
// continue parsing
}
}
continue;
case HtmlParseState.AttributeBeforeEquals:
if (!Char.IsWhiteSpace(currentChar))
{
if (currentChar == '>')
{
AddAttribute();
EndNode(previousChar);
break;
}
else if (currentChar == '=')
{
m_ParseState = HtmlParseState.AttributeAfterEquals;
// continue parsing
}
else // start of a new attribute name
{
AddAttribute();
StartNewAttribute();
AppendToAttributeName(currentChar);
m_ParseState = HtmlParseState.AttributeName;
// continue parsing
}
}
continue;
case HtmlParseState.AttributeName:
if (Char.IsWhiteSpace(currentChar))
{
m_ParseState = HtmlParseState.AttributeBeforeEquals;
// continue parsing
}
else if (currentChar == '=')
{
m_ParseState = HtmlParseState.AttributeAfterEquals;
// continue parsing
}
else if (currentChar == '>')
{
AddAttribute();
EndNode(previousChar);
break;
}
else
{
AppendToAttributeName(currentChar);
// continue parsing
}
continue;
case HtmlParseState.AttributeValue:
if (Char.IsWhiteSpace(currentChar))
{
AddAttribute();
m_ParseState = HtmlParseState.BetweenAttributes;
// continue parsing
}
else if (currentChar == '>')
{
AddAttribute();
EndNode(previousChar);
break;
}
else
{
AppendToAttributeValue(currentChar);
// continue parsing
}
continue;
case HtmlParseState.BetweenAttributes:
if (!Char.IsWhiteSpace(currentChar))
{
if (currentChar == '>')
{
EndNode(previousChar);
break;
}
else
{
StartNewAttribute();
AppendToAttributeName(currentChar);
m_ParseState = HtmlParseState.AttributeName;
// continue parsing
}
}
continue;
case HtmlParseState.Comment:
if (currentChar == '-')
{
char secondChar = NextChar();
if (secondChar == '-')
{
char thirdChar = NextChar();
int count = 0; // keep track of the number of '-' signs to append to the node value
while (thirdChar == '-')
{
thirdChar = NextChar();
count++;
}
// check if the end of the comment has been reached
if (thirdChar == '>')
{
while (count-- > 0)
{
AppendToValue('-');
}
EndNode(previousChar);
break;
}
else
{
// If there were extra dashes before the final -->, append them to the
// value.
count += 2;
while (count-- > 0)
{
AppendToValue('-');
}
AppendToValue(thirdChar);
// continue parsing
}
}
else
{
AppendToValue(currentChar);
AppendToValue(secondChar);
// continue parsing
}
}
else
{
AppendToValue(currentChar);
// continue parsing
}
continue;
case HtmlParseState.Identifier:
if (currentChar == '>')
{
EndNode(previousChar);
break;
}
else if (Char.IsWhiteSpace(currentChar))
{
m_ParseState = HtmlParseState.BetweenAttributes;
break;
}
else
{
m_Name.Append(currentChar);
// continue parsing
}
continue;
case HtmlParseState.QuotedAttributeValue:
if (currentChar == m_QuoteChar && currentChar != '\0')
{
AddAttribute();
m_ParseState = HtmlParseState.BetweenAttributes;
// continue parsing
}
else
{
AppendToAttributeValue(currentChar);
// continue parsing
}
continue;
case HtmlParseState.Tag:
case HtmlParseState.EndTag:
if (currentChar == '>')
{
EndNode(previousChar);
break;
}
else if (Char.IsWhiteSpace(currentChar))
{
m_ParseState = HtmlParseState.BetweenAttributes;
break;
}
else
{
m_Name.Append(currentChar);
// continue parsing
}
continue;
case HtmlParseState.Text:
// This node type ending was checked above as a special case.
AppendToValue(currentChar);
// continue parsing
continue;
case HtmlParseState.None:
if (currentChar != '<')
{
m_ParseState = HtmlParseState.Text;
StartNode(HtmlNodeType.Text);
AppendToValue(currentChar);
break;
}
else
{
m_ParseState = HtmlParseState.BeginTag;
// continue parsing
}
continue;
case HtmlParseState.BeginTag:
if (currentChar == '!')
{
char secondChar = NextChar();
if (secondChar == '-')
{
char thirdChar = NextChar();
if (thirdChar == '-')
{
m_ParseState = HtmlParseState.Comment;
StartNode(HtmlNodeType.Comment);
break;
}
else if (thirdChar == '>')
{
StartNode(HtmlNodeType.Identifier);
AppendToValue(secondChar);
EndNode(previousChar);
break;
}
}
else if (secondChar == '>')
{
StartNode(HtmlNodeType.Identifier);
EndNode(previousChar);
break;
}
else
{
StartNode(HtmlNodeType.Identifier);
m_Name.Append(secondChar);
m_ParseState = HtmlParseState.Identifier;
// continue parsing
}
}
else if (Char.IsLetter(currentChar))
{
StartNode(HtmlNodeType.Element);
m_ParseState = HtmlParseState.Tag;
m_Name.Append(currentChar);
// continue parsing
}
else if (currentChar == '/')
{
StartNode(HtmlNodeType.EndElement);
m_ParseState = HtmlParseState.EndTag;
// continue parsing
}
else
{
// if a number or symbol appears after a '<', IE treats it
// as text instead of an element.
m_ParseState = HtmlParseState.Text;
StartNode(HtmlNodeType.Text);
AppendToValue('<');
AppendToValue(currentChar);
break;
}
continue;
}
break;
}
}
/// <summary>
/// Finishes parsing the current node.
/// </summary>
private void EndNode(char? previousChar)
{
// if previousChar is "/" and this is an element node, set IsEmptyElement.
if (previousChar == '/' && NodeType == HtmlNodeType.Element)
{
IsEmptyElement = true;
}
else
{
IsEmptyElement = false;
}
m_ParseState = HtmlParseState.None;
}
