feat: Sign bounce messages#874
Conversation
missytake
left a comment
There was a problem hiding this comment.
I don't think this approach would work, see my comment. In any case, this PR needs a test.
2e2e6cd to
0bfc671
Compare
0bfc671 to
ada0060
Compare
ada0060 to
9dac959
Compare
9dac959 to
43590d3
Compare
43590d3 to
b197edf
Compare
b197edf to
726837f
Compare
ae0ef53 to
40ee20f
Compare
40ee20f to
d0d5635
Compare
CI fails because ci-chatmail runs old filtermail version, it has to be upgraded. Other than that this is ready. |
I upgraded ci-chatmail.testrun.org to main :) weirdly, I can't restart the failed CI jobs, maybe you can rebase on main to re-trigger the CI? |
89bfab6 to
3dea2a1
Compare
3dea2a1 to
91e8d2a
Compare
Closes #873 Signed-off-by: Jagoda Ślązak <jslazak@jslazak.com>
91e8d2a to
5cd887f
Compare
|
|
ah yes, it is - it runs the tests with |
|
I wonder if this will require further changes after #917 |
|
Closing, since we should no longer send bounce messages anyway |
|
We still do send bounce messages in some cases (e.g. mailbox doesn't exist), and that they are not DKIM signed causes non-delivered bounces: May 05 13:33:34 nine postfix/lmtp[3828308]: F3B411902838: to=<zagkkunah@nine.testrun.org>, relay=nine.testrun.org[private/dovecot-lmtp], delay=0.05, delays=0.01/0/0/0.04, dsn=5.1.1, status=bounced (host nine.testrun.org[private/dovecot-lmtp] said: 550 5.1.1 <zagkkunah@nine.testrun.org> User doesn't exist: zagkkunah@nine.testrun.org (in reply to RCPT TO command)) |
|
re-did and amended this PR into #964 |
This is based on Jagoda's #874 but comes with a simpler and more robust test. TODO: requires chatmail/filtermail#149
This was originally based on Jagoda's #874 but then the postfix config was simplified, and it comes with a simpler and more robust test.
This was originally based on Jagoda's #874 but then the postfix config was simplified, and it comes with a simpler and more robust test.
This was originally based on Jagoda's #874 but then the postfix config was simplified, and it comes with a simpler and more robust test.
Closes #873