Linux - ServerThis forum is for the discussion of Linux Software used in a server related context.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
Regarding SpamAssassin using USER_IN_WELCOMLIST instead of USER_IN_WHITELIST, in my email X-Spam-Report I get:
Code:
* -0.0 USER_IN_WELCOMELIST User is listed in 'welcomelist_from'
whereas with the USER_IN_WHITELIST I would get:
Code:
* -100 USER_IN_WHITELIST DEPRECATED: See USER_IN_WELCOMELIST
The -100 was apparently the 'default' score for USER_IN_WHITELIST. Is there no such default for USER_IN_WELCOMLIST? The default appears to be 0.0. Do have to explicitly set a score in local.cif or does enabling "shortcircuit USER_IN_WELCOMLIST on" take care of that?
BTW - not being a perl expert, how would I determine that 'plugin Mail::SpamAssassin::Plugin::Shortcircuit' is enabled?
Is SpamAssassin still in Perl?? I thought that had been rewritten as python .
I think you should have a negative score attached to the whilelist. Also, you need to up the (positive) scores of Spam-identifying rules markedly. The way it's set up, most spam will get through. A rule like "Composed in m$ Front Page" had a score of 1.5, with a fail mark of 5. Who writes mail in M$ Front Page? I don't even know if Front Page exists still - a HTML creator thing. It was a 100% indicator of spam. I gave it a score of +5.2.
IIRC, the plugins were rulesets? I must confess to using spamc/spamd because they were binary and I didn't have the ram for several simultaneous instances of some perl behemoth. Procmail ran me out of ram also.
I had to be a member of an Electronic Hardware mailing list back 20 years ago. I found the spam that list attracted was particularly offensive. Spammers had to join to post, but they targeted hundreds with just the one email(before they were banned). To catch spam my mail went through
Vipul's Razor, which Caught ~50% of spam.
Distributed Checksum Clearinghouse (DCC) which caught the rest of the bulk mail.
SpamAssassin which got 30% on average.
My efficiency eventually was nudged up to 105%. So I had to trawl spam for real mail occasionally. One of my brothers always landed himself in the spam, for instance. But I didn't whitelist him - he belonged in the Spam .
You will find a mailing list/forum or irc channel attached to the SpamAssassin site which is/was well supported by these guys who could instantly express any unclean thoughts you may be privvy to in Perl Regular Expressions. They are always welcoming of anyone they could talk at(!), and competed for the most exotic or innovative answers to your questions. They might prove to be of assistance.
I found it useful to keep a (mbox style ascii) mailbox full of spam. I could test my settings by simply
Code:
cat mbox >127.0.0.1:25
which bounced them at postfix & delivered them to me. It allowed me to refine the settings.
Last edited by business_kid; 05-07-2024 at 01:04 PM.
I think you should have a negative score attached to the whilelist. Also, you need to up the (positive) scores of Spam-identifying rules markedly. The way it's set up, most spam will get through. A rule like "Composed in m$ Front Page" had a score of 1.5, with a fail mark of 5. Who writes mail in M$ Front Page? I don't even know if Front Page exists still - a HTML creator thing. It was a 100% indicator of spam. I gave it a score of +5.2.
Actually, I've been using SpamAssassin for years and have a pretty robust set of rules, and I do Bayes training periodically. I'm good there. I was just wondering why this new "WELCOMELIST" rule doesn't have a preset default like the deprecated "WHITELIST" did. I have gone ahead and set it to -100, like WHITELIST used to be. I suppose that'll do unless some SpamAssassin gurus have deeper insight.
Quote:
IIRC, the plugins were rulesets? I must confess to using spamc/spamd because they were binary and I didn't have the ram for several simultaneous instances of some perl behemoth. Procmail ran me out of ram also.
I am using spamd and spamass-milter as a mail filter. I believe they're part of the package.
Quote:
You will find a mailing list/forum or irc channel attached to the SpamAssassin site which is/was well supported by these guys who could instantly express any unclean thoughts you may be privvy to in Perl Regular Expressions. They are always welcoming of anyone they could talk at(!), and competed for the most exotic or innovative answers to your questions. They might prove to be of assistance.
Yes, I believe I've used that list in the past. I generally try the experts at LQ before resorting to lists.
I'll leave this question open a bit longer in case anyone else wants to weigh-in with a thought.
Am I the only one who replied? My experience is less than your own and 20 years old! Spamc was one instance per mail iirc. They didn't have spamass-milter then. Vipul's Razor was another thing and I found it very good. You took a checksum, tried them, and they told you if the mail was spam or not. You were expected to report spam. If you reported spam that was actually spam, as reported by others, your reputation was good. If you reported spam as clean and good mail as spam, you were a twit, and they didn't take your reports seriously. In the early 2000s, /vipul's Razor was free for <1000 emails/day. The guy is linux-friendly. The site seems down but this link works https://sourceforge.net/projects/raz...atest/download
The DCC was good too, and free for <1000 emails/day. But if anyone signed up for marketing emails, they were rejected unless you whitelisted the address.
Python or binary is the only was to run SpamAssassin, imho. Guys would roll these regexes searching for 50 terms on one line, and everything would grind to a halt with perl. The spam that got through typically had one or two lines and a url. I tried to keep searches to the Subject, not the body.
Since this is related, I'll post this here. If no response I'll start a new thread.
I've changed all my local.cf "whitelist_from" entries to "welcomelist_from". However, when restarting spamd I got 111 of the following type of message:\
Code:
Thu May 30 23:04:26 2024 [7350] info: config: failed to parse line, skipping, in "/etc/mail/spamassassin/local.cf": welcomelist_from subscription@symantec.rsys2.com
Curiously, one that does not error is "welcomelist_from *myorg.org". Ones without the *, or with *@ all generated the above message.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.