BlogEngine.NET version 1.3 has a syntax highlighting extension included but is in beta, so I looked around for another syntax highlighter, since this blog heavly uses code snippets. After looking for a while I found this syntax highlighter extension. This extension uses Wilco Bauwer's syntax highlighter library which is impressive. But the main problem with this extension is, it does not handle HTML tags and special HTML characters like &nbsp &lt and &gt very well. Tags and special characters are left as garbage after extension tries to highlight the code. You see something like this

private voidTest&nbsp{

<p>&nbsp;private&nbsp;int&nbsp;i=0;

<p>}

After inspecting SyntaxHighlightingExtension.cs file I saw that the extension matches the source code with a regular epression and feeds the Wilco's highlighter with the raw html (body). This was kind of incomlete implementation causing the side effect I mentioned above. We need to clean html tags and special characters from the raw html(body). So I changed Highlight method of the extension. The resulting method is something like this.

C#-Code: Highlight(HighlightOptions options) method

001 private string Highlight(HighlightOptions options)
002 {
003  string parsed;
004  uint id = NextCodeID();
005  string name = options.Language; 
006
007
008  HighlighterBase highlighter = GetHighlighter(name);
009  if (highlighter != null)
010  {
011   name = highlighter.FullName;
012   highlighter.Parser = htmlParser;
013   
014   string body = Regex.Replace(options.Code,@"<(?![!/]?[>\s])[^>]*>",String.Empty,RegexOptions.CultureInvariant| RegexOptions.IgnoreCase | RegexOptions.Singleline);  
015   body = HttpUtility.HtmlDecode(body);
016   parsed = highlighter.Parse(body);
017   highlighter.ForceReset();
018  }
019  else
020  {
021   name += " (not highlighted)";
022   parsed = options.Code;
023  } 
024
025
026  if (options.DisplayLineNumbers)
027  {
028   string[] lines = parsed.Split(new char[] { '\n' });
029   StringBuilder outputBuffer = new StringBuilder(); 
030
031
032   for (int i = 0; i < lines.Length; i++)
033   {
034    outputBuffer.AppendFormat(linenumberingTemplate, i+1, lines[i]);
035   } 
036
037
038   return string.Format(OutputTemplate, id, name, options.Title, outputBuffer);
039  } 
040
041
042  return string.Format(OutputTemplate, id, name, options.Title, parsed);
043 }

The change I made resides starting from line 12 and ending in line 21. I simply stripped out the html tags with a regular expression and then used HttpUtility.Decode method to decode special HTML characters and feed the parser with normalized body text and voila the extension started performing well.

By the way I want to remind you that if you are using the default editor (TinyMCE) you should copy and paste your source code to a plain text editor like Notepad++ and the copy from Notepad++ and paste to TinyMCE. I think TinyMCE should consider to add something like Paste as PlainText functionality to their editor like FCKEditor.