Can't create forwarding address

chad

Hi, I'm trying to create a forwarding address, but I don't seem to be able to. The "Create Email Address" button is missing from the dialog. The button appears if I add a mailbox, but I don't want the address linked to a mailbox, I just want forwarding.

sean

All addresses require a mailbox.

We do not forward spam from our MX servers, so without a mailbox any false-positive spam would be lost forever.

If you don't care about that and don't want the assigned mailbox to fill up, then you can add a procmail rule to the mailbox like:

:0:
* 
/dev/null

This will cause all mail that arrives in the attached mailbox to be deleted.

chad

Thank you sean, that makes sense.

Is there a technical reason not to forward spam? If not, it'd be nice if there were an option to forward everything, including spam, and eliminate the need for a local mailbox.

Is there a procmail rule that would keep messages for a certain period of time before deleting them? Say, for 1 year.

ttavoy

sean All addresses require a mailbox.

But it seems possible to create an address which delivers to a single mailbox, and then remove that mailbox from the destinations list with '/api/v1/address/update/'

If that's valid, it creates a more minimalist version of what the OP is asking for.

sean

chad We don't forward spam because doing so will cause our MX servers to become blacklisted.

Procmail acts on incoming messages when they arrive, it doesn't do anything with messages that have already been delivered. You'll need to do this with an IMAP client instead.

Here's an example of how to do that with Python's imaplib, you could run this script daily via cron:

#!/bin/env python3

import imaplib


mailbox = 'xxx'
password = 'yyy'

mailserver = 'mail.us.opalstack.com'

age_in_days = 365

cutoff_date = (datetime.date.today() - datetime.timedelta(age_in_days)).strftime("%d-%b-%Y")

M = imaplib.IMAP4_SSL(mailserver)
M.login(mailbox, password)
M.select('INBOX')
typ, data = M.search(None, f'BEFORE {cutoff_date}')

for msgid in data[0].split():
    M.store(msgid, '+FLAGS', '(\\Deleted)')
M.expunge()
M.close()
M.logout()

chad

sean Thank you. This is very generous. I'm going to see if the volume of spam is actually an issue before I do this.

HelenaGwyn

sean I tried this but can't seem to login.

#!/bin/env python3

import datetime
import imaplib
import os

mailbox = os.path.splitext(os.path.basename(__file__))[0]
password = open(os.environ['HOME']+"/.local/etc/"+mailbox,"r").readlines()[0]
mailserver = 'mail.de.opalstack.com'
age_in_days = 30

cutoff_date = (datetime.date.today() - datetime.timedelta(age_in_days)).strftime("%d-%b-%Y")

M = imaplib.IMAP4(mailserver)
M.login(mailbox, password)
M.select('INBOX')
typ, data = M.search(None, f'BEFORE {cutoff_date}')

for msgid in data[0].split():
    M.store(msgid, '+FLAGS', '(\\Deleted)')
M.expunge()
M.close()
M.logout()

chad

I checked the mailbox that's associated with my forwarding address, using Opalstack's webmail, and it appears that all mail is being delivered to that mailbox, not just the spam/non-forwarded messages. This makes it difficult to look for false positives.

I don't want to store the messages that get forwarded. Can this be achieved with a procmail rule? If not, is there any other way to identify which messages have been forwarded and which haven't?

sean

chad At this moment, everything is forwarded. This part of the system is not complete yet but eventually it will work by preventing forwarding of any message with a spam score higher than N (to be determined).

If you want to emulate that now then you can create a procmail rule to delete anything with a spam score of less than N.

For example, if you want procmail to delete everything with a spam score of less than or equal to 7:

    :0:
    * ^x-spam-level: \*\*\*\*\*\*\*$
    /dev/null

sean

HelenaGwyn sorry about this, the IMAP connection needs to be made with SSL, like so:

M = imaplib.IMAP4_SSL(mailserver)

I've updated the original example.

sean

ttavoy we removed the mailbox requirement for forwarding addresses several months ago so workarounds like this should no longer be needed.