[Rspamd-Users] Speaking of feature requests ...
rspamd at vlh.dk
rspamd at vlh.dk
Mon Mar 13 19:58:23 UTC 2023
> -----Original Message-----
> From: Users <users-bounces at lists.rspamd.com> On Behalf Of Håkon
> Alstadheim via Users
> Sent: 12. marts 2023 00:59
> To: users at lists.rspamd.com
> Cc: Håkon Alstadheim <hakon at alstadheim.priv.no>
> Subject: Re: [Rspamd-Users] Speaking of feature requests ...
> Den 10.03.2023 17:33, skrev Phil Stracchino:
> > ... One of the really *INCREDIBLY NICE* things about DSpam, before it
> > succumbed to terminal bitrot, was that in the event of a false
> > positive or false negative, each message listed in history had a
> > button at the end of the line that was either "Retrain as HAM" or
> > "Retrain as SPAM" depending on how the message had been scored. All
> > you had to do was click "Retrain as SPAM", and that message's cached
> > flags and tokens would be retrained until it scored as spam, with no
> > need to manually copy, paste and inject the message for retraining.
> > Is this something that Rspamd could perhaps do in the future?
> If you're running dovecot it's simple to set up with
> https://docs.iredmail.org/dovecot.imapsieve.html . You just use your MUA
> to move messages into/out of your spam folder, and have imapsieve on the
> server side pipe the message through a script that does
> "/usr/bin/rspamc learn_spam".
> With the right MUA (I use Thunderbird), there will be a single click that flags a
> message as spam (or not-spam) and simultaneously does the refiling to/from
> the spam folder.
> This is way better, because you have the interface right in your MUA.
This rely on users to correctly identify spam though 😉- I know people moving stuff to junk that they signed up for.
(Also really couldn't get it to work on my test-system, complained about broken pipe/permissions - but I guess that's not for here really 😊)
More information about the Users