After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 309181 - Spam filter does not seem to work
Spam filter does not seem to work
Status: RESOLVED INCOMPLETE
Product: evolution
Classification: Applications
Component: Mailer
2.2.x (obsolete)
Other Linux
: High enhancement
: ---
Assigned To: evolution-mail-maintainers
Evolution QA team
: 311328 (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2005-06-28 12:18 UTC by Martin Probst
Modified: 2013-09-10 14:04 UTC
See Also:
GNOME target: ---
GNOME version: 2.9/2.10



Description Martin Probst 2005-06-28 12:18:39 UTC
Distribution/Version: Ubuntu

1. Install Evolution 2.2.1.1
2. Install SpamAssassin
3. Check mail and look for spam
4. Mark email as spam

Expected: spam mail trains spam assassin

Actual Result: nothing special happens with spam mail

My email account is an IMAP account, spam filtering is turned on in both of the
locations.

Or in more detail:
SpamAssassin is installed. When I mark an email as spam, Evolution shows
"Learning Spam" with a percentage progress meter in the task bar (but only
showing 0% or 100% for a short time). I can also see the CPU load spiking for a
short time.

But tailing the log files (/var/log/mail.(log|info)) does not show anything
happening. Those files seem to be used as I find lot's of entries where
evolution opens connections and checks emails, and also some where evolution
trains an email as ham (surprising, does it do that somewhat automatically
without user input?). But there is not a single line about learning that
something is spam, even though I've marked several messages as spam.

Is it actually possible to get spam filtering to work at this time? I tried spam
filtering on 3 computers (2x Ubuntu, 1x Gentoo), on none of them it worked.

Is there a way to get more information about what evolution is trying to do?
"evolution --debug=file" doesn't help.
Comment 1 André Klapper 2005-07-02 22:37:16 UTC
...if i remember correctly you have to mark at least 200 messages by hand until 
spamassassin and its bayes algorithm kicks in. so mark messages as junk and try 
later again... ;-)
no, seriously: have you already marked at least 200 messages as spam?
Comment 2 Martin Probst 2005-07-03 17:33:38 UTC
I figured that out, yes. Also it seems you have to mark 200 messages as not junk
too, to get spam filtering to work.

But my actual problem seems to have been different. I get lots of spam from some
stock advertisement company (gnarf) that looks a lot like legitimate email -
correct header fields etc., only the HELO is spoofed. So SA classifies it as
spam level 0.1 from it's rules (without bayesian filtering). And the default
setting is to learn email as "ham" to the bayesian filter if it has a level of
0.1 or below.

So all the spam I get is already classified as ham and learned into the bayesian
filter, and clicking "Junk" in evolution does not seem to have any effect as the
email is already trained - you have to call sa-learn with --forget to retrain
messages.

Maybe the second thing can be fixed by forcing SA to learn? Also the thing with
the 200 emails spam/ham is quite annoying and not very obvious to novice users.
At least I thought it would not work at all...
Comment 3 Luis Villa 2005-07-25 13:30:04 UTC
See also bug 311328, 305056, and 272926. Lots of people seem to be finding this
inefficient at best.
Comment 4 C Shilpa 2005-07-28 10:04:32 UTC
confirming the bug based on the discussion with partha.
Comment 5 vivek jain 2005-08-09 05:52:29 UTC
Probably we have to write an alternative to spamassasin. Another plugin.
Comment 6 Lionel Dricot 2006-03-25 10:21:25 UTC
*** Bug 311328 has been marked as a duplicate of this bug. ***
Comment 7 Quim Gil 2006-03-27 07:56:28 UTC
My small contribution to this (big) problem. From a Ubuntu Dapper user perspective, since it's the distribution where I'm using Evolution (2.6.0).


- Evolution is the default email tool provided. The User is going to use it.

- The User starts tagging spam emails as "Junk". No dialog will tell her that this functionality works after activating Edit > Preferences > Mail Preferences > Junk > "Check incoming mail for junk". (not checked by default, as far as I remember). The emails selected go to Junk, but Evolution never learns. An idea could be to show a dialog advertising this with a checkbox to Don't tell me this again. Unab ling the Junk button until the option  is activated in the Preferences would be another solution. The real solution is activating Junk detection by defauiilt (and making sure it works, as the Evolution competitors achieve).

- One day The User discovers somehow that SpamAssassin or an alternative plugin needs to be installed. No dialog in the application told her anything about this. I Ubuntu she needs to activate the universe package if it's not activated yet. She needs to install SpamAssassin and dependencies. But by doing this only selected email will go to Junk. Evolution never learns.

- The problem is that SpamAssassin is installed but the daemon is not activated by default. You only know this if you open the terminal dialog in Synaptic during the installation and read the messages provided. At least a pop up configuration dialog could have told The User to choose if she wants the daemon on/off. To activate it you need to sudo and edit a text file in /etc/default/spamasssassin. (I know it's a SA problem but let me report it here as part of the frustrating Evo user experience)

And I wonder which average user is going to complete this process, specially if coming from Outlook or Thunderbird, where spam detection works out of the box.
Comment 8 Omer Akram 2010-04-14 17:29:31 UTC
Its been a long while since any last activity happened on this one. This should be closed as wont fix?
Comment 9 Matthew Barnes 2010-04-14 17:51:43 UTC
Closing as incomplete.  Junk filtering with Bogofilter works fine in 2.30.