<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 07.12.2016 12:00,
<a class="moz-txt-link-abbreviated" href="mailto:info-cyrus-request@lists.andrew.cmu.edu">info-cyrus-request@lists.andrew.cmu.edu</a> wrote:<br>
</div>
<blockquote
cite="mid:mailman.747.1481130018.34762.info-cyrus@lists.andrew.cmu.edu"
type="cite">
<pre wrap="">I use this to feed spamassassin automatically:
<a moz-do-not-send="true" class="moz-txt-link-freetext" href="https://packages.debian.org/jessie/sa-learn-cyrus">https://packages.debian.org/jessie/sa-learn-cyrus</a>
</pre>
</blockquote>
<p>Do you even need that? Each message can be found on the
filesystem already in format directly understandable by sa-learn.
I have the following cron-job running on my server as the <tt>cyrus</tt>-user:</p>
<blockquote>
<p><tt>find /var/spool/imap/user/*/spam*/ -type f -name '[0-9]*.'
| xargs sa-learn --quiet --spam</tt></p>
</blockquote>
<p>As you can see, it is not even IMAP-aware. You may need it to be,
if it needs to run on a machine separate from Cyrus-server --
because that's where your incoming mail is being filtered.</p>
<p>But even in that case, it would be faster to run the learning on
the Cyrus-system -- dealing with local files is always faster,
than going for the same files via IMAP -- and then copy the
updated database over to the incoming-mail server.</p>
<blockquote>
<p>-mi<br>
</p>
</blockquote>
</body>
</html>