public void ParseHtml(object sender, RoutedEventArgs e)
{
Tokens.Items.Clear();
ReconstructedHTML.Clear();
ParseErrors.Items.Clear();
using (TextReader stringReader = new StringReader(HtmlInput.Text))
{
using (HtmlTokenizer tokenizer = new Tokenizer(
Fragment.IsChecked.HasValue && Fragment.IsChecked.Value
? new HtmlTextTokenizer(stringReader, FragmentContainer.Text)
: new HtmlTextTokenizer(stringReader),
Tokens)) {
using (HtmlReader reader = new HtmlReader(tokenizer)) {
reader.ParseError += new EventHandler<ParseErrorEventArgs>(reader_ParseError);
while (reader.Read()) {
switch (reader.NodeType) {
case XmlNodeType.Text:
case XmlNodeType.Whitespace:
case XmlNodeType.SignificantWhitespace:
ReconstructedHTML.AppendText(reader.Value);
break;
case XmlNodeType.Comment:
ReconstructedHTML.AppendText("<!--");
ReconstructedHTML.AppendText(reader.Value);
ReconstructedHTML.AppendText("-->");
break;
case XmlNodeType.Element:
ReconstructedHTML.AppendText("<");
ReconstructedHTML.AppendText(reader.Name);
if (reader.MoveToFirstAttribute()) {
do {
ReconstructedHTML.AppendText(" ");
ReconstructedHTML.AppendText(reader.Name);
ReconstructedHTML.AppendText("=");
ReconstructedHTML.AppendText(reader.QuoteChar.ToString());
ReconstructedHTML.AppendText(reader.Value);
ReconstructedHTML.AppendText(reader.QuoteChar.ToString());
} while (reader.MoveToNextAttribute());
}
ReconstructedHTML.AppendText(">");
break;
case XmlNodeType.EndElement:
ReconstructedHTML.AppendText("</");
ReconstructedHTML.AppendText(reader.Name);
ReconstructedHTML.AppendText(">");
break;
default:
ReconstructedHTML.AppendText("###");
ReconstructedHTML.AppendText(Enum.GetName(typeof(XmlNodeType), reader.NodeType));
ReconstructedHTML.AppendText("###");
break;
}
}
}
}
}
}
private void DoTest(string input, string fragmentContainer, string expectedOutput, string[] parseErrors)
{
expectedOutput.Trim();
input = input.Replace("\n", Environment.NewLine);
expectedOutput = expectedOutput.Replace("\n", Environment.NewLine);
Trace.WriteLine("Input:");
Trace.WriteLine(input);
Trace.WriteLine("");
if (parseErrors == null || parseErrors.Length == 0) {
Trace.WriteLine("No expected error.");
} else {
Trace.WriteLine("Expected errors:");
foreach (string parseError in parseErrors) {
Trace.WriteLine(parseError);
}
}
Trace.WriteLine("");
Trace.WriteLine("Expected:");
Trace.WriteLine(expectedOutput);
StringBuilder actualOutput = new StringBuilder(expectedOutput.Length);
try {
HtmlReader htmlReader;
if (fragmentContainer == null) {
htmlReader = new HtmlReader(new StringReader(input));
} else {
htmlReader = new HtmlReader(new StringReader(input), fragmentContainer);
}
using (htmlReader) {
htmlReader.ParseError += new EventHandler<ParseErrorEventArgs>(htmlReader_ParseError);
using (XmlReader reader = new LintXmlReader(htmlReader)) {
while (reader.Read()) {
if (reader.NodeType == XmlNodeType.EndElement) {
continue;
}
actualOutput.Append("| ");
if (reader.Depth > 0) {
actualOutput.Append(' ', reader.Depth * 2);
}
switch (reader.NodeType) {
case XmlNodeType.DocumentType:
actualOutput.Append("<!DOCTYPE");
//if (reader.Name.Length > 0) {
actualOutput.Append(' ').Append(reader.Name);
//}
//string publicId = reader.GetAttribute("PUBLIC");
//string systemId = reader.GetAttribute("SYSTEM");
//if (publicId != null) {
// actualOutput.Append("PUBLIC ")
// .Append('"').Append(publicId.Replace("\n", Environment.NewLine)).Append('"');
//} else if (systemId != null) {
// actualOutput.Append("SYSTEM");
//}
//if (systemId != null) {
// actualOutput.Append(' ').Append('"').Append(systemId.Replace("\n", Environment.NewLine)).Append('"');
//}
actualOutput.Append('>');
break;
case XmlNodeType.Element:
actualOutput.Append("<").Append(reader.Name).Append('>');
if (reader.MoveToFirstAttribute()) {
do {
actualOutput.AppendLine();
actualOutput.Append("| ");
actualOutput.Append(' ', reader.Depth * 2);
actualOutput.Append(reader.Name).Append("=\"").Append(reader.Value.Replace("\n", Environment.NewLine).Replace("\"", """)).Append('"');
} while (reader.MoveToNextAttribute());
reader.MoveToElement();
}
break;
case XmlNodeType.EndElement:
throw new InvalidOperationException("EndElement should have been handled above !?");
case XmlNodeType.Comment:
actualOutput.Append("<!-- ").Append(reader.Value.Replace("\n", Environment.NewLine)).Append(" -->");
break;
case XmlNodeType.Whitespace:
case XmlNodeType.Text:
actualOutput.Append('"').Append(reader.Value.Replace("\n", Environment.NewLine).Replace("\"", """)).Append('"');
break;
default:
Assert.Fail("Unexpected token type: {0}", reader.NodeType);
break;
}
actualOutput.AppendLine();
}
}
}
Assert.AreEqual(expectedOutput, actualOutput.ToString().Trim());
Assert.AreEqual(parseErrors == null ? 0 : parseErrors.Length,
this.parseErrors);
} catch (NotImplementedException nie) {
// Amnesty for those that confess
Trace.WriteLine("");
Trace.WriteLine(nie);
#if NUNIT
Assert.Ignore("Not Implemented");
#else
Assert.Inconclusive("Not Implemented");
#endif
} finally {
if (this.parseErrors == 0) {
Trace.WriteLine("");
Trace.WriteLine("No actual error.");
}
Trace.WriteLine("");
Trace.WriteLine("Actual:");
Trace.WriteLine(actualOutput);
}
}