Question

I’m trying to populate a grid with some data extracted from linkedin, im just trying to get it working for my own learning curve, BUT if I remove the line

MessageBox.Show("asdfasdfasdf")

the list “messages” only has 1 item, if I include the line above it does whats expected and I get 15 messages

Can someone explain?

public void extract_messages_received(object sender, RoutedEventArgs e)
{
    triggered = false;
    System.Windows.Forms.WebBrowser browser = new System.Windows.Forms.WebBrowser();
    browser.Navigate(new Uri(@"http://www.linkedin.com/inbox/messages/received"));
    browser.DocumentCompleted += new WebBrowserDocumentCompletedEventHandler(browser_DocumentCompleted);
}

private void LoadMessages(string url)
{
    txtOutput.Text = @"http://www.linkedin.com" + url.Substring(6, url.Length - 6);
    if (!urls.Contains(url))
    {
        urls.Add(url);
        WebBrowser browser = new WebBrowser();
        browser.Navigate(new Uri(txtOutput.Text);

        loaded_message = false;
        browser.DocumentCompleted += new WebBrowserDocumentCompletedEventHandler(ReadMessages);
    }
}

private void ReadMessages(object sender, WebBrowserDocumentCompletedEventArgs e)
{
    if (loaded_message == false)
    {        
        string url = ((WebBrowser)sender).Url.ToString();
        int loc1 = url.IndexOf("itemID") + 7;
        int loc2 = url.IndexOf("&", loc1);
        IEnumerable<string> name = null;
        IEnumerable<string> odate = null;
        IEnumerable<string> photo = null;
        IEnumerable<string> subject = null;
        IEnumerable<string> headline = null;
        string body = "";
        string id = url.Substring(loc1, loc2 - loc1);
        //System.Windows.MessageBox.Show("READ");
        foreach (HtmlElement element in ((WebBrowser)sender).Document.GetElementsByTagName("div"))
        {
            if (element.GetAttribute("classname").Equals("inbox-item-body"))
            {
                body = element.InnerText;
            }
            if (element.GetAttribute("classname").Equals("inbox-item-header"))
            {
                var doc = new HtmlAgilityPack.HtmlDocument();
                doc.LoadHtml(element.InnerHtml);
                name = from foo in doc.DocumentNode.SelectNodes("//a[@class='fn']") select foo.InnerText;
                odate = from foo in doc.DocumentNode.SelectNodes("//p[@class='date']") select foo.InnerText;
                photo = from foo in doc.DocumentNode.SelectNodes("//img[@class='photo']") select foo.Attributes["src"].Value;
                subject = from foo in doc.DocumentNode.SelectNodes("//h3") select foo.InnerText;
                headline = from foo in doc.DocumentNode.SelectNodes("//span[@class='headline']") select foo.InnerText;
            }
        }

        // ****
        MessageBox.Show("asdfasdfasdf");
        // ****

        messages.Add(new Messages()
        {
            ID = id,
            Subject = subject.First().ToString(),
            Headline = headline.First().ToString(),
            Sender = name.First().ToString(),
            Photo = photo.First().ToString(),
            SendDate = odate.First().ToString(),
            Body = body
        });

           // dataMessages.ItemsSource = messages;
    }
    loaded_message = true;
}

void browser_DocumentCompleted(object sender, WebBrowserDocumentCompletedEventArgs e)
{
    if (!triggered)
    {
        triggered = true;
        System.Windows.Forms.WebBrowser web = sender as System.Windows.Forms.WebBrowser;
        foreach (HtmlElement element in web.Document.GetElementsByTagName("ol"))
        {
            if (element.GetAttribute("classname").Contains("inbox-list "))
            {
                WebBrowser browser = new WebBrowser();
                browser.Navigate("about:blank");
                browser.Document.Write(element.InnerHtml);
                HtmlElementCollection hrefTags = null;
                hrefTags = browser.Document.GetElementsByTagName("a");
                foreach (HtmlElement a in hrefTags)
                {
                    if (a.OuterHtml.Contains("displayMBox"))
                    {
                        LoadMessages(a.GetAttribute("href"));
                    }
                }
            }
        }
    }       
}

Answer 1

This is a timing issue.

When you have the message box in there, loaded_message doesn’t get set to true until after you close the message box, so the other events are processing up until the message box as well, with none of them settings loaded_message to true until you close the first message box.

If you close the messagebox quickly enough, you will probably see some number beteween 1 and 15.

Let’s take a more simplistic example:

    private void Form1_Load(object sender, EventArgs e)
    {

        for (int i = 0; i < 5; i++)
        {
            WebBrowser wb = new WebBrowser();
            wb.DocumentCompleted += wb_DocumentCompleted;
            wb.Navigate("http://www.stackoverflow.com");
        }
    }

    bool shown = false;
    void wb_DocumentCompleted(object sender, WebBrowserDocumentCompletedEventArgs e)
    {
        if (!shown)
        {
            Console.WriteLine(shown);
            MessageBox.Show(shown.ToString());
            shown = true;
        }
    }

Now, if you watch the console, you will see a few false show up before the first message box is shown. When I close the message box, I then see 4 more messageboxes because those were already queued up and waiting to show before shown got set to true. If I comment out the messagebox, then I get only one message box shown and one false in the console.

Now, the question becomes, why did you add and need to check the loaded_message boolean variable.

My guess is that you only want to load each message only once. If that is the case, keep track of each URL in a dictionary and maintain a bool for each URL:

    Dictionary<string, bool> loadedUrls = new Dictionary<string, bool>();
    private void Form1_Load(object sender, EventArgs e)
    {

        for (int i = 0; i < 5; i++)
        {
            WebBrowser wb = new WebBrowser();
            wb.DocumentCompleted += wb_DocumentCompleted;
            string url = "http://stackoverflow.com/" + i;

            loadedUrls.Add(url, false);
            wb.Navigate(url);
        }
    }

    bool shown = false;
    void wb_DocumentCompleted(object sender, WebBrowserDocumentCompletedEventArgs e)
    {

        if (loadedUrls.ContainsKey(e.Url.OriginalString) && loadedUrls[e.Url.OriginalString] == false)
        {
            loadedUrls[e.Url.OriginalString] = true;
            Console.WriteLine(shown);
            shown = true;
        }
    }

I left shown in there to demonstrate that this new approach now works for each pass in the document completed event. Your output window should have a false followed by 4 true.