xTuple Connect Guide

Yes, xTuple needs to be running for the Synchronization to work. This is necessary because xTuple launches the synchronizer and communicates login credentials including database, user and password, so the it knows where to synchronize the data. We had originally thought about running the synchronizer as an independent service on the machine, but we subsequently decided we didn't want create potential security vulnerabilities by storing database credentials on the local machine which would be necessary to run it as a free standing service.

It does not.

When mail is imported into the database a trigger checks to see if a special reference is in the subject or body that has "xtref" and a document code in it. This reference is automatically appended to the subject of any Email that is launched from the xTuple through your Email client, and it is different for each document. Sales orders, for example, might be something like [xtref:S-012345]. This launching of your Email client happens when you click on the Email Address of a Contact or when you click "New" or "Reply" in the Email tab of documents that have one. Some have asked why we couldn't make the reference code hidden in an Email header. The reason is because the mail is generated in your client by a mailto URL mechanism. This is the same mechanism used to launch your Email client when you click on an Email Address on a web page. The RFC for mailto (http://www.faqs.org/rfcs/rfc2368.html) does not allow for hidden information in such a message. The RFC states that for security reasons all information being sent must be clearly visible to the sender.

It is important to note that just because this reference is in the subject doesn't mean it will be imported. IT IS THE USER'S RESPONSIBILITY TO CONFIGURE THEIR MAIL CLIENT AND SYNCHRONIZATION SCHEME TO ENSURE THESE MESSAGES ARE IMPORTED, because these messages are ultimately being sent by your mail client, not xTuple. The reference tagging just helps xTuple sort out the relationships IF the mail gets imported.

The only messages that are guaranteed to be imported regardless of configuration are ones actually generated by xTuple Connect itself (NOT your Email client) such as Sales Order Acknowledgements, Invoices, reports to run through xTuple Connect (AKA Batch Manager processor) or any mail generated via an EDI profile.

They won't. Both the synchronizer and xTuple connect processors calculate a cryptographic hash on the message body that creates a unique ID that is stored with the message in the database. They compare to see if there are any messages with the same hash ID already in there before attempting to insert a new copy. If one is already found, it will skip that one.

That is correct.