Mailkit: Converting HtmlBody to pdf using iTextSharp XMLWorker throws "The document has no pages"

Question

I'm trying to convert the HtmlBody of the e-mails I get from a mailserver using Mailkit and looks like iTextSharp doesn't really like the html I'm passing it.

My method works well with a "sample" html code but I get a The document has no pages error message which looks like it's thrown when the html is no html anymore.

public void GenerateHtmlFromBody(UniqueId uid)
{
    var email = imap.Inbox.GetMessage(uid);
    Byte[] bytes;

    using (var ms = new MemoryStream())
    {
        using (var doc = new Document())
        {
            using (var writer = PdfWriter.GetInstance(doc, ms))
            {
                doc.Open();

                //Sample HTML and CSS
                var example_html = @"<p>This <em>is </em><span class=""headline"" style=""text-decoration: underline;"">some</span> <strong>sample <em> text</em></strong><span style=""color: red;"">!!!</span></p>";
                var example_css = @".headline{font-size:200%}";

                using (var srHtml = new StringReader(email.HtmlBody))
                {
                    //Parse the HTML
                    iTextSharp.tool.xml.XMLWorkerHelper.GetInstance().ParseXHtml(writer, doc, srHtml);
                }
                doc.Close();
            }
        }
        bytes = ms.ToArray();
    }
    var testFile = Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.Desktop), "processedMailPdf.pdf");
    System.IO.File.WriteAllBytes(testFile, bytes);
}

I'm accesing to MimeMessage.HtmlBody and debugging, looks like it's, in fact, html.

Here is a link to pastebin for checking the HtmlBody of the MimeMessage because I hit the character limit here.

What am I missing? Thanks.

EDIT: I've tried using the HTMLWorker (which is deprecated) and it's not stable. It worked with one e-mail but not with others. Of course it wasn't a solution, but it finally generated a pdf from Mailkit, which was "something".

Answer 1

Looks like you're facing two issues with HtmlBody:

1. It may be plain text.
2. When [X]HTML, it is not well-formed.

Anytime there's a possibility you're dealing with a string that is not well-formed XML, your best bet is to use a parser like HtmlAgilityPack to clean up the mess. Here's a simple helper method using XPath to cover both issues above, and UPDATED based on comments to remove HtmlCommentNodes that break iText XML Worker:

string FixBrokenMarkup(string broken)
{
    HtmlDocument h = new HtmlDocument()
    {
        OptionAutoCloseOnEnd = true,
        OptionFixNestedTags = true,
        OptionWriteEmptyNodes = true
    };
    h.LoadHtml(broken);

    // UPDATED to remove HtmlCommentNode
    var comments = h.DocumentNode.SelectNodes("//comment()");
    if (comments != null) 
    {
        foreach (var node in comments) { node.Remove(); }
    }

    return h.DocumentNode.SelectNodes("child::*") != null
        //                            ^^^^^^^^^^
        // XPath above: string plain-text or contains markup/tags
        ? h.DocumentNode.WriteTo()
        : string.Format("<span>{0}</span>", broken);
}

And for completeness, code to generate the PDF. Tested and working with the pastebin link you provided above:

var fixedMarkup = FixBrokenMarkup(PASTEBIN);
// swap initialization to verify plain-text works too
// var fixedMarkup = FixBrokenMarkup("some text");

using (var stream = new MemoryStream())
{
    using (var document = new Document())
    {
        PdfWriter writer = PdfWriter.GetInstance(document, stream);
        document.Open();
        using (var stringReader = new StringReader(fixedMarkup))
        {
            XMLWorkerHelper.GetInstance().ParseXHtml(
                writer, document, stringReader
            );
        }
    }
    File.WriteAllBytes(OUTPUT, stream.ToArray());
}