SPAM Mail

[troypulk]

Today so far I have received 11 spam mails.

Does the spam filter even work?

On average I get around 3 every day.

Any one know what's going on?

Thanks

[WildEnte]

@troypulk Vivaldi webmail does not have a learning spam filter. The mail client integrated in the browser also does not have a learning spam filter but only moves mail to the imap spam folder. If that server has a learning filter all is well. But if you use Vivaldi mail with the Vivaldi mail service, you don't have a spam filter unless you set up a custom filter yourself in the webmail interface

https://forum.vivaldi.net/topic/75141/add-a-spam-filter-to-webmail-vivaldi-net

[troypulk]

@WildEnte
Thanks, I'll look into it.

[becm]

@troypulk the Vivaldi mail server does at least seem to run a classifier (insertion of X-Spam-* headers).

If mail in case of X-Spam-Flag: YES is not moved to Spam automatically, you could add a Sieve rule based on that or try to fine-tune a rule based on a lower threshold than the pre-configured 6.31.

[Dancer18]

@becm You seem to be a professional. Please translate your idea for non-professionals.
Where and how can i add this Sieve rule?
What is the pre-configured 6.31?

[becm]

@Dancer18 the Sieve filter rules are a server-side configuration and can be accessed in via the webmail interface.

The required=6.31 entry in the X-Spam-Status mail header should determine when a mail gets categorized as Spam by default.
In Vivaldi Mail (browser) you can see all headers when displaying raw mail data (right-click on message content).

Evaluating the X-Spam-Flag would be an easy matter (see above) but might/should already be done internally.
Correctly using the value in X-Spam-Score is a little involved and requires additional filter features on the server (not sure if they are available).

[Dancer18]

@becm Thanks! I have also set up the obvious first with webmail Vivaldi. The German interface shows a slightly different field under Rules than yours. Please see if I made the right pre-selection (...) to then enter "X-Spam-Flag". See screenshot.

[becm]

@Dancer18 looks correct, although I'm not sure if this rule has to be defined manually at all (it should/might be builtin).

You can have a look at the X-Spam-Score of an obvious (for a human) spam mail, if this actually has an expected (high) positive value.
(Webmail Mail View → More → Show source)
If not this part of the exercise was for naught, but you may find other classifiers (origin address pattern) to pre-filter junk mail.

Additionally, those rules are applied when the mail is received, so testing might get tedious.

[troypulk]

@becm
I've looked into the score's of emails and I have adjusted my filter using the score.

In looking at the spams they all have a positive score and the non-spams have a negative score but I did see a non-spam with a score of 1.3 so I adjusted my filter score to 2.

It must of worked because I get spam everyday and they now go into the junk folder.

On a side note, does it matter if I leave the spam mail unread or not?

Thanks for you help.

[becm]

@troypulk read status of spam mails is something I'd consider a personal preference.

But it can also be used as an additional classifier (in case of even higher spam score) to only check borderline spam categorization manually.

You could also post your example filter set here (ASCII code block) for future reference.

[troypulk]

@becm
Is this what you mean? I'm not familiar with ASCII code blocks.

[becm]

@troypulk so basically a config that comiles down to (Actions→Edit filter set)

require ["fileinto","relational"];
# rule:[Spam]
if header :value "ge" "x-spam-score" "2"
{
	fileinto "Junk";
}

If this can cope with negative values correctly (e.g. -3.14→ no spam) it should be fine (never actually tried to apply this anywhere).
The approach used in the docs seems to require some additional logic here.

[troypulk]

@becm
Thanks,
I like this and would consider using except I don't know how to use it. So where would said script go? Do I make my own? how do I attach it to the mail client and get the mail client to use it? or do I use an extension that does all this for me?

The picture I posted looks like the easiest way to go about it and is less involved and it does the same thing.

I looked around at the in the docs web site but it was not set up for beginners.

Thanks again

require ["comparator-i;ascii-numeric","relational"];
if allof (
   not header :matches "x-spam-score" "-*",
   header :value "ge" :comparator "i;ascii-numeric" "x-spam-score" "10" )
{
  discard;
  stop;
}

[becm]

@troypulk your are using the config I posted (as you can see when displaying the actual code, hint ↓).

This is a pure server-side configuration but one easy option to access it is via the webmail interface you set up the original filter in.

If you want multiple rules, make sure they are applied in the order you intend to.
Use only a single (merged) require directive.

[janrif]

@becm Does setting up the X-Spam-Filter in webmail help the M3 side of things?

[becm]

@janrif the server should pre-sort Junk into the appropriate folder before M3 even touches it.

If the filter logic learns from all mails (including manually moved ones → autolearn=spam) additional manual moves will also help later on.
Not sure however how good the current state of software is regarding this and how it gets applied on Vivaldi mail servers.