Basic Sieve Usage and Configuration
Setting up Sieve Access
Sieve is a server side protocol for filtering mail. The benefit over sorting mail client side is that rules will be applied regardless of your MUA. Sieve scripts can be edited locally or remotely using the managesieve service. I am using the Pigeonhole implementation of sieve which is bundled with dovecot 2.x. I will be going over in a later article how to setup dovecot.
Testing locally
To test that everything is working correctly create a .sieve/ folder in your home directory and create your sieve scripts within it. Call it anything you’d like but it has to end in a .sieve extension. I suggest you copy my basic config that I have listed below and make the appropriate changes.
When you are all done editing create a symlink in your home directory pointing to the file you just made and call it .sieve.conf. This will be your active sieve script that Pigeonhole uses.
Send an email that would test your script. If there are any issues with the script you wrote it be logged to ~/.sieve.conf.log.
Accessing remotely with Thunderbird
As of writing this article the sieve plugin for Thunderbird available for download through the built in store does not work. It broke during an API change and a new version has not been submitted. Instead a nightly release has to be downloaded from the projects Github page. You can find the nightly version I used here.
Install the xpi file and you should be able to connect to your sieve server. Issues can be worked through with the builtin debugger.
When you get access either open your existing rule set or create a new one. Thunderbird will run though it and make sure there are no syntax errors. Make sure you activate the script which will create the symlink for you if you haven’t already. The plugin also has some good resources of sieve script writing.
Basic Sieve Usage
The important notes about the config listed below are that the special commands like fileinto and addflag need to be included with require if they are used. The other thing to note is that the stop command is required unless you want multiple actions applies to a single envelope.
Basic Sieve Config
require ["fileinto", "imap4flags"];
# This rule is for spamassassin headers.
if header :contains "X-Spam-Flag" "YES" {
fileinto "Spam";
stop;
}
# Apply the Personal label to emails addressed to me. Continue executing the script.
if address :is ["to","cc"] "ehouse@ehouse.io"{
addflag "label3";
}
# File openbsd misc mailing list into openbsd/misc folder.
if address :is ["to","cc"] ["misc@openbsd.org","misc@cvs.openbsd.org"]{
fileinto "openbsd.misc";
stop;
}
# File into folder oss-security.
if address :is ["to","cc"] "oss-security@lists.openwall.com"{
fileinto "oss-security";
stop;
}
# File system logs into logs folder.
if address :is "from" "root@ehouse.io"{
fileinto "logs";
stop;
}Additional sieve resources
You can check out here for a full list of options for sieve or here for some great sieve examples. It’s a little weird at first but in no time you’ll be writing new filters rules without looking at the docs.