I am trying to manipulate a string containing HTML-Code and then save the content to a htm-file. Afterwards the htm file is imported to a Word-File. Goal is to append a document formatted in HTML to a Word document. This process is part of a much larger programm and i cannot modify the given parameters.
To easily modify the HTML-Code I thought using XDocument would be a great idea.
So I tried this:
AppendContent(string content, Document doc)
{
string filePath = ...; //somewhere in /AppData/Local
var xDoc = XDocument.Parse(content);
// code left out because irrelevant
// Finding all "img" elements, in order to
// extract the embedded picture and save it as external file
FileHelper.SaveToFile(filePath, xDoc.ToString());
//... After this, the file is appended to the word file (the one in doc)
}
First attempt worked actually, with a small test html. Using any of the big documents I'm trying to append to the word document, cause an exception to be thrown:
XDocument.Parse cannot parse entities like "nbsp" or "uuml" (german ü). I already found out that XML only supports a hand full of predefined entities, so i would have to manually add the definition to the html file. This is not an option, because this operation is supposed to work with ANY Html file.
I found following fix:
var decodedContent = WebUtility.HtmlDecode(content);
var xDoc = XDocument.Parse(decodedContent);
This converts all entities to the representing character. So "uuml" is converted to "ü", etc. This worked until i hit a document that contained the "amp" entity, which is then converted to "&"... and such the XDocument.Parse is complaining again.
I'm looking for a way to convert HTML to unicode-representation ("\0x1234") or a HTML-decode, that does not decode XML-predefined entities.
