/// <summary>
/// Parses the stream of html tokens starting
/// from the name of top-level element.
/// Returns XmlElement representing the top-level
/// html element
/// </summary>
private XmlElement ParseHtmlContent()
{
// Create artificial root elelemt to be able to group multiple top-level elements
// We create "html" element which may be a duplicate of real HTML element, which is ok, as HtmlConverter will swallow it painlessly..
XmlElement htmlRootElement = _document.CreateElement
(
"html",
XhtmlNamespace);
OpenStructuringElement(htmlRootElement);
while (_htmlLexicalAnalyzer.NextTokenType != HtmlTokenType.Eof)
{
switch (_htmlLexicalAnalyzer.NextTokenType)
{
case HtmlTokenType.OpeningTagStart:
_htmlLexicalAnalyzer.GetNextTagToken();
if (_htmlLexicalAnalyzer.NextTokenType == HtmlTokenType.Name)
{
string htmlElementName = _htmlLexicalAnalyzer.NextToken.ToLower();
_htmlLexicalAnalyzer.GetNextTagToken();
// Create an element
XmlElement htmlElement = _document.CreateElement
(
htmlElementName,
XhtmlNamespace);
// Parse element attributes
ParseAttributes(htmlElement);
if (_htmlLexicalAnalyzer.NextTokenType == HtmlTokenType.EmptyTagEnd ||
HtmlSchema.IsEmptyElement(htmlElementName))
{
// It is an element without content (because of explicit slash or based on implicit knowledge aboout html)
AddEmptyElement(htmlElement);
}
else if (HtmlSchema.IsInlineElement(htmlElementName))
{
// Elements known as formatting are pushed to some special
// pending stack, which allows them to be transferred
// over block tags - by doing this we convert
// overlapping tags into normal heirarchical element structure.
OpenInlineElement(htmlElement);
}
else if (HtmlSchema.IsBlockElement(htmlElementName) ||
HtmlSchema.IsKnownOpenableElement(htmlElementName))
{
// This includes no-scope elements
OpenStructuringElement(htmlElement);
}
}
break;
case HtmlTokenType.ClosingTagStart:
_htmlLexicalAnalyzer.GetNextTagToken();
if (_htmlLexicalAnalyzer.NextTokenType == HtmlTokenType.Name)
{
string htmlElementName = _htmlLexicalAnalyzer.NextToken.ToLower();
// Skip the name token. Assume that the following token is end of tag,
// but do not check this. If it is not true, we simply ignore one token
// - this is our recovery from bad xml in this case.
_htmlLexicalAnalyzer.GetNextTagToken();
CloseElement(htmlElementName);
}
break;
case HtmlTokenType.Text:
AddTextContent(_htmlLexicalAnalyzer.NextToken);
break;
case HtmlTokenType.Comment:
AddComment(_htmlLexicalAnalyzer.NextToken);
break;
}
_htmlLexicalAnalyzer.GetNextContentToken();
}
// Get rid of the artificial root element
if (htmlRootElement.FirstChild is XmlElement &&
htmlRootElement.FirstChild == htmlRootElement.LastChild &&
htmlRootElement.FirstChild.LocalName.ToLower() == "html")
{
htmlRootElement = (XmlElement)htmlRootElement.FirstChild;
}
return(htmlRootElement);
}
