Evolution spam filtering of HTML content
 

I have just started using Ximian Evolution v 1.2.2 that came installed with Red Hat 9. I have set up some filters for spam but have noticed that some of them do not work. If the message is in plain text, the filters work great but if the message is in HTML then they completely ignore the message.

Has anyone been able to work around this problem?

Given the date on your post, you've probably solved this already. However, others may have the same question (as I did) so I'm posting this for their sake. I looked into this trying to figure out why my filters only worked sometimes and found that the spammers are putting html tags in the middle of words hence a filter must eliminate everything between < and > in order to to match keywords. The spammers also have a lot of mismatched < > pairs further complicating identification of tags. This I found in just one day's review of spam (who knows how many other tricks the spammers are using), so I quickly concluded that it's best to put in a bit of effort to install a spam elimination package rather than re-invent the wheel by coding filters myself. Spam-Assassin looks like the most likely candidate at this point (just installed and started using myself).

No, I haven't solved my problem but your reason makes sense.

I am using Spam-Assassin but am having mixed results. It catches a lot of junk but probably only half of it. Plus I am having a weird problem--every now and then Spam Assassin seems to have trouble or something because only part of my email is downloaded and the next time I check my mail it downloads the same messages again without clearing them off the server or downloading all the messages. If I remove the Spam Assassin filter then the email downloads properly. Let me know how it works for you.

Ok, I'll post in a week or so how it's working. My results for the first day are about the same as yours, half went to spam half stayed in inbox, although 2 LQ.org mails went to spam so I need to do some tweaking.

As a follow-up to my previous post, I can say after 2 weeks that my results are similar to yours. However, to eliminate more spam I lowered the score (required hits) from 5 to 2 and that helped. I still get some spam in inbox but most goes to the spam folder. I also have the problem where it hangs occasionally but I just hit cancel on the send&receive popup window and then redo it and then it works ok but with one duplicate email message. I had one time where things got totally hosed and I had to kill the background processes and restart evolution, hopefully that won't happen very often. Thus far, it seems to be working well enough for my purposes, I think it's a keeper until something better comes along.

Thanks for the update.

I am again using SpamAssassin (SA). I installed the newest version plus I also installed Razor. After doing some experimenting, I concluded that the reason all my email wasn't being downloaded some of time was probably due to a slow connection (occasionally) combined with SA's long process time which resulted in losing the POP connection.

To correct this, I installed qMail as my MTA and configured fetchmail to download my mail at regular intervals. Now Evolution just reads in the new emails from a file on my HD and runs the SA tests. Since setting up things this way, SpamAssassin hasn't caused any trouble.

I'm not sure if it was the newer version of SA or the addition of Razor but I am seeing fewer spam emails in my inbox. I studies the documentation a bit further and read how to train SA. Now I am keeping all Spam that get's past SA in a special folder which I will use for training once it has over 100 messages or so.