We Keep Coding

Getting sender IP address from Exchange Server 2003 using WebDAV

I recently wrote an application for our company to process newsletter signup requests via signup emails sent to an inbox on our Exchange 2003 servers using WebDAV. This all works fine.
Now we've realized that for auditing purposes, we need to capture the ip address of where the signup request originated. My question is, is there a way to request the original ip address of the originator of the email with my WebDAV request?
I've browsed through the urn:schemas:mailheader: and the urn:schemas:httpmail: documentation and didn't see a field you can request with this data besides maybe urn:schemas:mailheader:path. But when I make a request to our exchange server with the path in the request, the status for that property comes back 404 not found.
It looks like http://schemas.microsoft.com/cdo/smtpenvelope has a clientipaddress property that would have this information, but that is only applicable to messages still in transit.
Has anyone had to do this before and figured out a way to snag the ip address of the user who originated the email? It probably isn't helpful to the question, but the format of my WebDAV request is below:
string webdav =
#"
<?xml version=""1.0""?>
<D:searchrequest xmlns:D = ""DAV:"">
<D:sql>
SELECT
""DAV:displayname"",
""urn:schemas:httpmail:fromemail"",
""urn:schemas:mailheader:subject"",
""urn:schemas:httpmail:textdescription"",
""urn:schemas:mailheader:date""
FROM
SCOPE('shallow traversal of ""{0}""')
WHERE
""DAV:isfolder"" = false AND
""urn:schemas:httpmail:read"" = false
</D:sql>
</D:searchrequest>
";

This comes back to SMTP more than it comes to Exchange/WebDAV. It really depends on which email service the end user is using. SMTP can pass an email around multiple times before it ends up at the destination. Normally, each hop adds a Received: from header, with some additional information like an IP address.
But, some services, like Google, don't count the user sending the email has a hop, and the originating IP address is a Google SMTP server. So you'll never know the end user IP address from the email. Then, other services may count the end user's public IP address as the first hop. And some other services may add a special header like X-Sender-IP or X-Originating-IP to the message.
So, there isn't a guaranteed way to obtain that information. Part of it has to do with the distributed nature of SMTP, the prevalence of webmail and some privacy concerns. If this information is critical to your auditing, you may want to setup a simple webform which could send an email to this inbox, and then you could add additional information like an IP address to the body of the email.

Detect auto reply emails programmatically

I have scenario where I would have to track the delivery of the emails I send programmatically and flag those recipients who have set either 'Out of Office" OR have the message delivery failed due to over-sized inbox OR if their email ID doesn't exist. Such instances usually send out automated replies. How can I track them? Does .NET (System.Net.Mail) offer any APIs to do it?

There's no set of checks for auto-responses that produce perfect behavior (detect all auto-replies with no false positives), but the following checks have worked well so far:
header Auto-Submitted with value other than no (see RFC
3834)
headers X-Autoreply or X-Autorespond with any value
header Precedence with value auto_reply
I don't yet have any advice for detecting message delivery failure notifications.

sending an email to hotmail inbox

I am trying to send an email using c# MailMessage to a hotmail account but the emails are always going to junk.
How can i send emails directly to hotmail inbox in c#?
I don't have my own smtp server, therefore i have tried using my university smtp and other smtps like google, yahoo... but all emails were sent to junk.
Any solutions?

This is not really a C# question.
Mails are sent to spam depending on their content and their headers so you should check a few things :
You need to have a subject
you shouldn't have spammy words (sex, viagra, love, watches)
You need to have a return-to and a from address headers that match
You need to be consistent with your encoding, if you go for UTF8 send UTF8 text
You shouldn't insert images
Your links inside the mail, if any, should have their text set to the address they are going to
Using these strategies should help your email not being classified as spam.
You can always check the headers of one of the mail that went into spam. It is often described what rules were applied and where the mail failed.

In hotmail, open the Junk folder, click on the message. Hotmail will display options in the body of the message - click on "Wait, it's safe!" Hotmail will move the message to your inbox, and mark the FROM email address as safe.
There is (probably) nothing wrong with your code - it is hotmail identifying the sender and/or subject as being junk.

Although it may well be out of your control, junk email filters generally work on a weighting system, so there are things that you can do to make your email look less like spam.
To start with, check that:
your subject doesn't contain all caps or sensitive words (such as "FREE!")
your body has content
the 'from' address for the MailMessage exists.
If those are all fine, have a look through this list, this article on live.com and the Policies, Practices and Guidelines for Hotmail.

What is the content of your mail, is it text that is likely to be regarded as junk mail by spam filters? Have you tried sending to other accounts such as Yahoo or Gmail?

You can't do it from C#.
It is up to the recipient to route the messages to junk or inbox or wherever it decides to put the message. If the sender were allowed to decide where the message went on the client's side, imagine how much bigger of a problem spam would be.
One solution you can employ is to have your target mail account 'whitelist' the sender, but that may not be an option in all cases.

Outlook Add-In - Custom Headers for Email Tracking

We have written an Outlook add-in in C#, that appends a custom header to outgoing messages.
This add-in has to use a library called Redemption to bypass Outlook's security to modify the headers, and this is all working great.
Our problem lies when sending outgoing mail through an exchange server. We use the additional header as such:
Add a References header with an email address that includes an ID for tracking with our system.
This is a standard email header that all mail clients should pass on when replying to messages. So replying to a message will automatically keep the new message tracked.
All of this works just fine if you send an email from an IMAP account setup in Outlook such as GMail.
Problem is, if you send mail via an Exchange account, the Exchange server overwrites the References header and uses it's own proprietary headers: Thread-Index and Thread-Topic. Email standards suggest to use References and In-Reply-To headers. See this link on this issue.
Does anyone know a way around this? Some algorithm to gain us the following:
Add a header (of any name or kind) to emails that includes a 10-digit ID and 3 letter prefix
Replying to this email from all (or most) mail clients preserves the custom header

I think the following algorithm is going to solve our issue:
Our Outlook add-in will set the References and an arbitrary X- header
Our mail filter will look for References, if found use it (if outgoing mail server was Exchange, it will not be present)
If our mail filter finds the arbitrary X- header and a Thread-Index, it will store the data found in the X- header.
Later if the email is replied-to several times, the mail filter will use the Thread-Index to look up the past info to keep the email tracked.
For those wanting to know internals of the Thread-Index header, it is a Base64 encoded string. The first 22 bytes are the original unique portion and each reply adds an additional 5 bytes on to it. We only use the first 22 bytes to identify the email.

Force mail client to use text rather then HTML through mailto:

I'm using SubVersion and TRAC on a C# project I am working on, and I have my TRAC system setup with a email address that can be used to create tickets. In my program I've added a simple "FeedBack" button in my program which sends an email to this address. To open the email I'm just "starting" a mailto link as shown below.
System.Reflection.Assembly assem = System.Reflection.Assembly.GetExecutingAssembly();
string ver = assem.GetName().Version.ToString();
System.Diagnostics.Process.Start("mailto:foo#bar.com?subject=<Provide a title for your feedback here>&body=< Describe the problem you are having or enhancement you would like to suggest here. Please be as descriptive as you can, and if possible list out the actions that will replicate the problem >%0D%0A%0D%0A%0D%0AVersion: "+ver);
The problem I'm running into is if the user is using Outlook and their copy of Outlook is setup to HTML the ticket that gets created ends up having a bunch of HTML code that I have to clean up. Is there some way to notify whatever mail client is handling it to send the email as text rather then HTML?

There's nothing you can do (besides education) on the client - there's nothing in mailto to control a client side program. And, frankly, with the proliferation of web-based email - I think mailto is showing it's age.
Outlook should send a mime/multipart message, with both plain text and HTML parts. I'd guess you could extend or patch Trac to only grab the text/plain portion.
Otherwise, just create a form in your app to capture the email info. Again, if someone is using Hotmail or GMail - mailto is not likely to work anyway (or will open up their unconfigured Outlook Express, where they will dutifully type up an email and press Send. Only it won't go anywhere, because no SMTP servers are configured - so it will languish in the Outbox for years. Not that they will notice though...).