janos erdelyi

this part of the series on how i implemented trackback in C# revolves around the trackback receiver.

this is based on the specifications on trackback

i will not be presenting one giant block of code like i did with Part 1 since this is a larger piece. i will try to retain a sense of order however.

alrighty, enough preface, let’s jump in.

first off, the trackback page is expecting a POST with a required field named “url“. please note that all of my form and querystring field names will be case-sensitive.

one thing that was initially confusing to me was that in addition to the trackback POST information, the page also needs a piece of querystring info to determine the article in question which is being trackbacked to.

for the sake of this article, the querystring info will be filtered down to a System.Int32 value which is the unique id of the article – aptly named “articleId“.

i will skip the incredibly boring code revolving around converting the querstring System.String info into System.Int32. suffice it to say that if a proper integer cannot be pulled from the querystring, an error is thrown since this can be considered an improper page request.

assuming we have gotten this far, i’d also recommend checking that the article by that id does in fact exist and is active.

i made a handy quick little method to handle all of my messages, since according to the specification, messages must follow a specific format.

private void serverResponse(
	int errorCode,
	string errorMessage
) {
	Response.ClearContent();
	Response.ContentType = "text/xml";
	Response.ContentEncoding = System.Text.Encoding.UTF8;
	
	Response.Write("<?xml version=\"1.0\" encoding=\"utf-8\"?>");
	Response.Write("<response>");
	Response.Write("<error>" + errorCode.ToString() + "</error>");
	if (errorCode != 0) {
		Response.Write("<message>" + errorMessage + "</message>");
	}
	Response.Write("</response>");
	
	Response.End();
	return;
}

you really should take a second to familiarize yourself with the response specification if you have not already. it is very simple and very direct.

so, that error thrown as described above was really calling serverResponse instead.

moving on… once we have determined that a real and active article is being requested, the bare minimum required POST information is the url.

if (Request.Form["url"] == null) {
	serverResponse(1, "The parameter 'url' is required.");
}

i created a plain little private class in which to stuff the trackback information. here it is to save you some typing if you like:

private class TB
{
	public TB () {
		
	}
	
	public string Title {
		get { return title; }
		set { title = value; }
	}
	
	public string Exerpt {
		get { return exerpt; }
		set { exerpt = value; }
	}
	
	public string Url {
		get { return url; }
		set { url = value; }
	}
	
	public string BlogName {
		get { return blogName; }
		set { blogName = value; }
	}
	
	public string IpAddress {
		get { return ipAddress; }
		set { ipAddress = value; }
	}
	
	public int ArticleId {
		get { return articleId; }
		set { articleId = value; }
	}
	
	private string title;//optional
	private string exerpt;//optional
	private string url;//required
	private string blogName;//optional
	private string ipAddress;//not part of the spec. using this for internals
	private int articleId;//also, not part of spec. using for internal needs
}

you may note that i am also storing the ip address of the POSTer. this is for other uses, such as whitelisting or blacklisting. i know that an ip address can be a bit broad, but if you‘re having problems with someone, ip-based blacklisting can be handy. also, i just like to look to see who is tracking back (i’ve been thinking of hooking this up to a geo-ip system).

here we go, stuffing the info in.

tb.Url = Request.Form["url"];
tb.Title = Request.Form["title"];
tb.BlogName = Request.Form["blog_name"];
tb.Exerpt = Request.Form["exerpt"];
tb.IpAddress = Request.UserHostAddress;
tb.ArticleId = articleId;

as i mentioned about case-sensitivity – the specification used those form names in lower-case, so i am as well. i see no need to be friendly about it and check different cases. follow the specification!

next i also check the calling url to see if it at least contains a link to my page in it. this is a minor, but fairly effective manner of suppressing trackback spam.

//check to see that the calling reference actually has our site linked
//if not, it could easily be trackback spam
if (
	checkReference(
	tb.Url,
	blog,
	articleId
) == false
) {
	serverResponse(1, "No reference for this page '" + tb.Url + "' was found on your site. Please verify that the 'url' you have provided is correct.");
}

of course you’ll want the source to the reference checker. here it goes:

private bool checkReference(
	string url,
	com.janoserdelyi.nagry.DataObjects.Blog blog,
	int articleId
) {
	/*i should be looking for two main formats:
	<my url>/ArticleDisplay.aspx?articleId=<articleId>
	<my url>/permalink/article/<articleId>.aspx
	
	let the screen scraping begin!
	*/
	string html = GetRemoteHTTPGet(url, blog);
	
	if (html.Length == 0) {
		return false;
	}
	
	string myUrl1 = "http://" + blog.BlogFqdn + "/ArticleDisplay.aspx?articleId=" + articleId.ToString();
	string myUrl2 = "http://" + blog.BlogFqdn + "/permalink/article/" + articleId.ToString() + ".aspx";
	
	if (html.IndexOf(myUrl2) != -1) {
		return true;
	}
	if (html.IndexOf(myUrl1) != -1) {
		return true;
	}
	
	return false;
}


private string GetRemoteHTTPGet(
	string url,
	com.janoserdelyi.nagry.DataObjects.Blog blog
) {
	// get the HTML from the URL of the url that called this script
	System.Net.WebRequest webRequest = System.Net.WebRequest.Create(url);
	
	if (webRequest is System.Net.HttpWebRequest) {
		((System.Net.HttpWebRequest)webRequest).UserAgent = blog.BlogFqdn + " Trackback Checker";
		((System.Net.HttpWebRequest)webRequest).Referer = "http://" + blog.BlogFqdn + "/";
		((System.Net.HttpWebRequest)webRequest).Timeout = 6000;
	}
	
	System.Net.WebResponse webResponse = webRequest.GetResponse();
	// Get the response stream.
	System.IO.Stream responseStream = webResponse.GetResponseStream();
	
	// Use a StreamReader to read the entire response
	System.IO.StreamReader reader = new System.IO.StreamReader(responseStream, System.Text.Encoding.ASCII);
	string readerToEnd = reader.ReadToEnd();
	reader.Close();
	return readerToEnd;
}

once all of that is done, it’s up to you to insert the info into your database, whatever that may be – xml, postgresql, mysql, mssql, text files, whatever you use.

and then declare success!

serverResponse(0, "Success");

i hope that was helpful.

i’d like to note the main source i used while mucking my way through this (since i knew nothing about trackback prior to writing my own.) An ASP.NET TrackBack handler in C# my implementation is very much largely based on his, but altered to not only fit my code-base, but also to more closely fit my personal coding style preferences. his resource is excellent, however. i really did not find nearly as much online as i was expecting when i first set out, so i hope that adding my version to the heap will help someone out.

feel free to use the code as you see fit. all i ask is that some credit with my name and a link to this blog be present.