This can be used to delete all sessions belonging to a user/login.
For no it just iterates over all sessions.
This could be enhanced by using a prefix for and deleting by prefix.
- call cleanup_sessions on first kvstore access
this allows to run cmdline actions without redis (and makes it faster)
- Allow development using DictStore by setting REDIS_ADDRESS to the empty string in env
- don't sign 64bit random session id as suggested by nextgens
1785: Fix bug #1660 (don't replace nested headers) r=mergify[bot] a=nextgens
## What type of PR?
bug-fix
## What does this PR do?
Don't replace nested headers (typically in forwarded/attached emails). This will ensure we don't break cryptographic signatures.
### Related issue(s)
- close#1660
## Prerequistes
Before we can consider review and merge, please make sure the following list is done and checked.
If an entry in not applicable, you can check it or remove it from the list.
- [x] Unless it's docs or a minor change: add [changelog](https://mailu.io/master/contributors/guide.html#changelog) entry file.
Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
1783: Switch to server-side sessions r=mergify[bot] a=nextgens
## What type of PR?
bug-fix
## What does this PR do?
It simplifies session management.
- it ensures that sessions will eventually expire (*)
- it implements some mitigation against session-fixation attacks
- it switches from client-side to server-side sessions (in Redis)
It doesn't prevent us from (re)-implementing a "remember_me" type of feature if that's considered useful by some.
Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
1610: add option to enforce inbound starttls r=mergify[bot] a=lub
## What type of PR?
Feature
## What does this PR do?
It implements a check in the auth_http handler to check for Auth-SSL == on and otherwise returns a 530 starttls error.
If INBOUND_TLS_ENFORCE is not set the behaviour is still the same as before, so existing installations should be unaffected.
Although there is a small difference to e.g. smtpd_tls_security_level of Postfix.
Postfix already throws a 530 after mail from, but this solution only throws it after rcpt to. auth_http is only the request after rcpt to, so it's not possible to do it earlier.
### Related issue(s)
#1328 is kinda related, although this PR doesn't solve the issue that the headers will still display ESMTP instead of ESMTPS
## Prerequistes
Before we can consider review and merge, please make sure the following list is done and checked.
If an entry in not applicable, you can check it or remove it from the list.
- [x] In case of feature or enhancement: documentation updated accordingly
- [x] Unless it's docs or a minor change: add [changelog](https://mailu.io/master/contributors/guide.html#changelog) entry file.
Co-authored-by: lub <git@lubiland.de>
1638: Remove the username from the milter_headers r=mergify[bot] a=githtz
Rspamd adds the name of the authenticated user by default. Setting add_smtp_user to false prevents the login to be leaked.
## What type of PR?
Enhancement
## What does this PR do?
This PR prevents the user login to be leaked in sent emails (for example using an alias)
### Related issue(s)
Closes https://github.com/Mailu/Mailu/issues/1465
## Prerequistes
Before we can consider review and merge, please make sure the following list is done and checked.
If an entry in not applicable, you can check it or remove it from the list.
- [x] Unless it's docs or a minor change: add [changelog](https://mailu.io/master/contributors/guide.html#changelog) entry file.
Co-authored-by: anrc <15327800+githtz@users.noreply.github.com>
- remove PASSWORD_SCHEME altogether
- introduce CREDENTIAL_ROUNDS
- migrate all old hashes to the current format
- auto-detect/enable all hash types that passlib supports
- upgrade passlib to 1.7.4 (see #1706: ldap_salted_sha512 support)
1765: Set sensible cookie flags on the admin app r=mergify[bot] a=nextgens
## What type of PR?
Bugfix
## What does this PR do?
It sets the right flags on the session cookie issued by the admin app.
This should probably be backported as the lack of secure flag on TLS-enabled setup is a high risk vulnerability.
SameSite is hardening / helps against CSRF on modern browsers
HTTPOnly is hardening / helps reduce the impact of XSS
Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
1763: show flash messages again r=mergify[bot] a=lub
## What type of PR?
bug-fix
## What does this PR do?
This basically restores the behaviour, that got removed in
ecdf0c25b3 during refactoring.
### Related issue(s)
- noticed it while reviewing #1756
## Prerequistes
Before we can consider review and merge, please make sure the following list is done and checked.
If an entry in not applicable, you can check it or remove it from the list.
- [-] In case of feature or enhancement: documentation updated accordingly
- [-] Unless it's docs or a minor change: add [changelog](https://mailu.io/master/contributors/guide.html#changelog) entry file.
Co-authored-by: lub <git@lubiland.de>
1618: add OCSP stapling to nginx.conf r=mergify[bot] a=lub
It's not added in tls.conf, because apparently the mail ssl module
doesnt' support OCSP stapling.
https://nginx.org/en/docs/http/ngx_http_ssl_module.html#ssl_stapling
^ exists
https://nginx.org/en/docs/mail/ngx_mail_ssl_module.html#ssl_stapling
^ missing
When the configured certificate doesn't have OCSP information, it'll
just log a warning during startup.
## What type of PR?
enhancement
## What does this PR do?
It enables OCSP stapling for the http server. OCSP stapling reduces roundtrips for the client and reduces load on OCSP responders.
### Related issue(s)
- fixes #1616
## Prerequistes
Before we can consider review and merge, please make sure the following list is done and checked.
If an entry in not applicable, you can check it or remove it from the list.
- [x] Unless it's docs or a minor change: add [changelog](https://mailu.io/master/contributors/guide.html#changelog) entry file.
Co-authored-by: lub <git@lubiland.de>
1607: _FILE variables for Docker swarm secrets r=mergify[bot] a=lub
## What type of PR?
enhancement
## What does this PR do?
This PR enables usage of DB_PW_FILE and SECRET_KEY_FILE instead of DB_PW and SECRET_KEY to load these values from files instead of supplying them directly. That way it's possible to use Docker secrets.
### Related issue(s)
## Prerequistes
Before we can consider review and merge, please make sure the following list is done and checked.
If an entry in not applicable, you can check it or remove it from the list.
- [x] In case of feature or enhancement: documentation updated accordingly
- [x] Unless it's docs or a minor change: add [changelog](https://mailu.io/master/contributors/guide.html#changelog) entry file.
Co-authored-by: lub <git@lubiland.de>
With `CHUNKING`set as a capability, nginx advertises this capability to
clients at a stage where the SMTP dialog does not seem to be forwarded
to the proxy-target (postfix) yet. Nginx' SMTP parser itself does not
support the `BDAT` command issued as part of a chunke-d dialog. This makes
Nginx respond with a `250 2.0.0 OK` and close the connection, after the
mail-data got sent by the client — without forwarding this to the
proxy-target.
With this, users mail can be lost.
Furthermore, when a user uses a sieve filter to forward mail, dovecot
sometimes chunks the forwarded mail when sending it through `front`.
These forwards then fail.
Removing `CHUNKING` from the capabilities fixes this behavior.
- yaml-import is now logged via schema.Logger
- iremoved relative imports - not used in other mailu modules
- removed develepment comments
- added Mailconfig.check method to check for duplicate domain names
- converted .format() to .format_map() where possible
- switched to yaml multiline dump for dkim_key
- converted dkim_key import from regex to string functions
- automatically unhide/unexclude explicitly specified attributes on dump
- use field order when loading to stabilize import
- fail when using 'hash_password' without 'password'
- fixed logging of dkim_key
- fixed pruning and deleting of lists
- modified error messages
- added debug flag and two verbosity levels
- idna.encode does not encode upper-case letters,
so .lower() has to be called on value not on result
- split email-address on '@' only once
- converted '*'.format(*) to f-strings
- added docstrings
- removed from_dict method
- code cleanup/style (list concat, exceptions, return&else, line-length)
- added TODO comments on possible future changes
- revived original config-update function for backwards compability
- renamed config-dump to config-export to be in line with config-import
- converted '*'.format(*) to f-strings
- converted string-concatenation to f-strings
renamed single letter variables (m => match)
renamed classmethod arguments to cls (model)
removed shadowing of variables (hash, context)
shortened unneeded lambda functions (id)
converted type ... is to isinstance(...)
removed unneded imports (flask)
Updated ConfigManager to only modify app.config and not replace it.
Swagger does not play well, when app.config is not a real dict and
it is not necessary to keep ConfigManager around after init.
Also added "API" flag to config (default: disabled).
1654: Ensure that the rendered file ends with newline in order to make `pos… r=mergify[bot] a=tremlin
## What type of PR?
Bugfix
## What does this PR do?
This fixes#1580
### Related issue(s)
- closes#1580
## Prerequistes
Before we can consider review and merge, please make sure the following list is done and checked.
If an entry in not applicable, you can check it or remove it from the list.
Co-authored-by: Thomas Rehn <thomas.rehn@initos.com>
Adds a 'none' env option to WEBROOT_REDIRECT so that no `location /`
configuration is written to nginx.conf.
This is useful for setting up Mailu and Mailman where we override the
root to proxy to the mailing list server instead. Without this change
the nginx container will not start, or for 1.7 users can set their
WEBMAIL_PATH to / with no webmail to get the same results.
This fix means that future users don't have to choose between webmail
and a root override and makes the configuration intention clear.
1611: Adds own server on port 80 for letsencrypt and redirect r=mergify[bot] a=elektro-wolle
## What type of PR?
Bugfix
## What does this PR do?
Handle letsencrypt route to `.well-known` by own server configuration within nginx.
### Related issue(s)
closes#1564
## Prerequistes
Before we can consider review and merge, please make sure the following list is done and checked.
If an entry in not applicable, you can check it or remove it from the list.
- [x] Unless it's docs or a minor change: add [changelog](https://mailu.io/master/contributors/guide.html#changelog) entry file.
Co-authored-by: Wolfgang Jung <w.jung@polyas.de>
1635: Add support for AUTH LOGIN authentication mechanism for relaying emai… r=mergify[bot] a=Diman0
…l via smart hosts.
## What type of PR?
Feature
## What does this PR do?
This PR adds support to postfix for AUTH LOGIN authentication mechanism. This enables using smart hosts which only offer AUTH LOGIN.
### Related issue(s)
- Auto close an issue like: closes#1633
## Prerequistes
Before we can consider review and merge, please make sure the following list is done and checked.
If an entry in not applicable, you can check it or remove it from the list.
- [n/a] In case of feature or enhancement: documentation updated accordingly
- [x] Unless it's docs or a minor change: add [changelog](https://mailu.io/master/contributors/guide.html#changelog) entry file.
Co-authored-by: Dimitri Huisman <diman@huisman.xyz>
Because basically every MTA out there uses opportunistic TLS _in
the best case_, it's actually counter productive to use such strict
settings.
The alternative to a handshake error is often an unencrypted submission,
which is basically the opposite of what strict ssl_protocols and
ssl_ciphers tries to achieve.
Even big and established providers like Amazon SES are incompatible with the current
settings.
This reverts commit 2ddf46ad2b.
this enables usage of e.g. docker swarm secrets instead of exposing the
passwords directly via environment variables
just use DB_PW_FILE and SECRET_KEY_FILE instead of DB_PW and SECRET_KEY
- stabilized CommaSeparatedList by sorting values
- CommaSeparatedList can now handle list and set input
- from_dict now handles mapped keys
- from_dict now handles null values
- class Domain: handle dkim-key None correctly
- class User: delete obsolete keys after converting
- class Alias: now uses Email._dict_input
Alexander Graf
lub