/// <summary>
/// Attempts to parse an element tag at the current location. If the tag is parsed,
/// the parser position is advanced to the end of the tag name and true is returned.
/// Otherwise, false is returned and the current parser position does not change.
/// </summary>
/// <param name="parser"></param>
/// <param name="tag"></param>
private bool ParseTag(TextParser parser, out string tag)
{
tag = null;
int pos = 0;
Debug.Assert(parser.Peek() == HtmlRules.TagStart);
char c = parser.Peek(++pos);
if (c == '!' || c == '?')
{
c = parser.Peek(++pos);
}
if (HtmlRules.IsTagCharacter(c))
{
while (HtmlRules.IsTagCharacter(parser.Peek(++pos)))
{
;
}
// Move past '<'
parser.MoveAhead();
// Extract tag name
int length = pos - 1;
tag = parser.Text.Substring(parser.Position, length);
parser.MoveAhead(length);
return(true);
}
// No tag found at this position
return(false);
}
/// <summary>
/// Moves the parser position to the closing tag for the given tag name.
/// If the closing tag is not found, the parser position is set to the end
/// of the text and false is returned.
/// </summary>
/// <param name="parser">Parser object.</param>
/// <param name="tag">Tag name for which the closing tag is being searched.</param>
/// <param name="content">Returns the content before the closing tag</param>
/// <returns></returns>
private bool ParseToClosingTag(TextParser parser, string tag, out string content)
{
string endTag = $"</{tag}";
int start = parser.Position;
// Position assumed to just after open tag
Debug.Assert(parser.Position > 0 && parser.Peek(-1) == HtmlRules.TagEnd);
while (!parser.EndOfText)
{
parser.MoveTo(endTag, true);
// Check that we didn't just match the first part of a longer tag
if (!HtmlRules.IsTagCharacter(parser.Peek(endTag.Length)))
{
content = parser.Extract(start, parser.Position);
parser.MoveAhead(endTag.Length);
parser.MoveTo(HtmlRules.TagEnd);
parser.MoveAhead();
return(true);
}
parser.MoveAhead();
}
content = null;
return(false);
}
/// <summary>
/// Parses the given HTML string into a number of root nodes.
/// </summary>
/// <param name="html">The HTML text to parse.</param>
public IEnumerable <HtmlNode> ParseChildren(string html)
{
TextParser parser = new TextParser(html);
HtmlElementNode rootNode = new HtmlElementNode("[Root]");
HtmlElementNode parentNode = rootNode;
string tag;
bool selfClosing;
// Loop until end of input
while (!parser.EndOfText)
{
if (parser.Peek() == HtmlRules.TagStart)
{
// Test for CDATA segments, which we store but do not parse. This includes comments.
CDataDefinition definition = HtmlRules.CDataDefinitions.FirstOrDefault(dd => parser.MatchesCurrentPosition(dd.StartText, dd.IgnoreCase));
if (definition != null)
{
parentNode.Children.Add(ParseCDataNode(parser, definition));
continue;
}
// Closing tag
if (parser.Peek(1) == HtmlRules.ForwardSlash)
{
parser.MoveAhead(2);
tag = parser.ParseWhile(c => HtmlRules.IsTagCharacter(c));
if (tag.Length > 0)
{
if (parentNode.TagName.Equals(tag, HtmlRules.TagStringComparison))
{
// Should never have matched parent if the top-level node
Debug.Assert(!parentNode.IsTopLevelNode);
parentNode = parentNode.ParentNode;
}
else
{
// Handle mismatched closing tag
int tagPriority = HtmlRules.GetTagPriority(tag);
while (!parentNode.IsTopLevelNode && tagPriority > HtmlRules.GetTagPriority(parentNode.TagName))
{
parentNode = parentNode.ParentNode;
}
if (parentNode.TagName.Equals(tag, HtmlRules.TagStringComparison))
{
Debug.Assert(!parentNode.IsTopLevelNode);
parentNode = parentNode.ParentNode;
}
}
}
parser.MoveTo(HtmlRules.TagEnd);
parser.MoveAhead();
continue;
}
// Open tag
if (ParseTag(parser, out tag))
{
HtmlTagFlag flags = HtmlRules.GetTagFlags(tag);
if (flags.HasFlag(HtmlTagFlag.HtmlHeader))
{
parentNode.Children.Add(ParseHtmlHeader(parser));
}
else if (flags.HasFlag(HtmlTagFlag.XmlHeader))
{
parentNode.Children.Add(ParseXmlHeader(parser));
}
else
{
// Parse attributes
HtmlAttributeCollection attributes = ParseAttributes(parser);
// Parse rest of tag
if (parser.Peek() == HtmlRules.ForwardSlash)
{
parser.MoveAhead();
parser.MovePastWhitespace();
selfClosing = true;
}
else
{
selfClosing = false;
}
parser.MoveTo(HtmlRules.TagEnd);
parser.MoveAhead();
// Add node
HtmlElementNode node = new HtmlElementNode(tag, attributes);
while (!HtmlRules.TagMayContain(parentNode.TagName, tag) && !parentNode.IsTopLevelNode)
{
Debug.Assert(parentNode.ParentNode != null);
parentNode = parentNode.ParentNode;
}
parentNode.Children.Add(node);
if (flags.HasFlag(HtmlTagFlag.CData))
{
// CDATA tags are treated as elements but we store and do not parse the inner content
if (!selfClosing)
{
if (ParseToClosingTag(parser, tag, out string content) && content.Length > 0)
{
node.Children.Add(new HtmlCDataNode(string.Empty, string.Empty, content));
}
}
}
else
{
if (selfClosing && flags.HasFlag(HtmlTagFlag.NoSelfClosing))
{
selfClosing = false;
}
if (!selfClosing && !flags.HasFlag(HtmlTagFlag.NoChildren))
{
parentNode = node; // Node becomes new parent
}
}
}
continue;
}
}
// Text node
int start = parser.Position;
// Text must be at least 1 character (handle '<' that is not part of a tag)
parser.MoveAhead();
parser.MoveTo(HtmlRules.TagStart);
Debug.Assert(parser.Position > start);
parentNode.Children.Add(new HtmlTextNode(parser.Extract(start, parser.Position)));
}
//
return(rootNode.Children);
}