1
0
mirror of https://github.com/Mailu/Mailu.git synced 2024-12-14 10:53:30 +02:00
Commit Graph

3042 Commits

Author SHA1 Message Date
Florent Daigniere
ef5f82362c Merge remote-tracking branch 'upstream/master' into policyd-mta-sts 2021-09-01 08:45:13 +02:00
Florent Daigniere
92cc664e82 Cosmetic change 2021-09-01 08:41:59 +02:00
Florent Daigniere
489520f067 forgot about alpine/lmdb 2021-09-01 08:41:39 +02:00
Florent Daigniere
9f66e2672b Use DEFER_ON_TLS_ERROR here too
We just don't know whether the lookup failed because we are under attack
or whether it's a glitch; the safe behaviour is to defer
2021-08-31 20:44:57 +02:00
Florent Daigniere
a1da4daa4c Implement the DANE-only lookup policyd
https://github.com/Snawoot/postfix-mta-sts-resolver/issues/67 for
context
2021-08-31 20:24:06 +02:00
Dimitri Huisman
5f18860669 Remove workaround. Remove deprecated url-loader. 2021-08-31 10:04:44 +00:00
Dimitri Huisman
60be06e298 Temporary workaround to get FontAwesome icons working. 2021-08-31 08:08:33 +00:00
Florent Daigniere
d607ba0ef2 Clarify that a restart may be required 2021-08-30 17:52:31 +02:00
Florent Daigniere
fb34f53493 Do operations in the right (safe) order 2021-08-30 17:18:19 +02:00
Florent Daigniere
fccb0cc57f Add a longer max_age (15days) 2021-08-30 17:16:41 +02:00
Dimitri Huisman
5da7a06675 Resolve webpack.config.js error 2021-08-30 15:01:05 +00:00
Florent Daigniere
67db72d774 Behave like documented 2021-08-30 17:00:12 +02:00
Florent Daigniere
05b57c972e remove the static policy as it will override MTA-STS and DANE 2021-08-30 14:44:13 +02:00
Florent Daigniere
a8142dabbe Introduce DEFER_ON_TLS_ERROR
This will default to True and defer emails that fail even "loose"
validation of DANE or MTA-STS

It should work most of the time but if it doesn't and you would rather
see your emails delivered, you can turn it off.
2021-08-30 14:21:28 +02:00
bors[bot]
7e86f5cb57
Merge #1959
1959: Ensure that we don't trust client headers r=mergify[bot] a=nextgens

## What type of PR?

bug-fix

## What does this PR do?

Document how REAL_IP_FROM and REAL_IP_HEADER should be used. Ensure that we strip True-Client-IP and X-Forwarded-For if neither are set.

We should also update the documentation on reverse-proxies... but that's #1958

### Related issue(s)
- #1958

## 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/workflow.html#changelog) entry file.


Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
2021-08-30 11:52:29 +00:00
Florent Daigniere
7c5dcfa025 MTA-STS is a major feature 2021-08-29 18:32:17 +02:00
Florent Daigniere
5efe35329b doh 2021-08-29 18:29:44 +02:00
Florent Daigniere
5634354911 document how to publish an MTA-STS policy 2021-08-29 18:28:56 +02:00
Florent Daigniere
a019607873 towncrier 2021-08-29 17:46:28 +02:00
Florent Daigniere
52d3a33875 Remove the domains that have a valid MTA-STS policy
gmail.com
comcast.net
mail.ru
googlemail.com
wp.pl
2021-08-29 17:41:55 +02:00
Florent Daigniere
4f96e99144 MTA-STS (use rather than publish policies) 2021-08-29 17:40:37 +02:00
Dimitri Huisman
00276d8b70
Merge branch 'master' into AdminLTE-3 2021-08-28 17:43:29 +02:00
bors[bot]
1d9850490c
Merge #1958
1958: Update the documentation on reverse proxies r=mergify[bot] a=nextgens

## What type of PR?

documentation

## What does this PR do?

Update the documentation on reverse proxies; this is mostly cosmetic (fix the links, use example.com where appropriate, ...).

It also removes the last option (run Mailu without its frontend) as that won't work with SSO and is a terrible idea anyway.

I wonder if we should just get rid of that section

### Related issue(s)
- #1528
- #1422
- #1038
- #1879

Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
2021-08-28 11:24:24 +00:00
Florent Daigniere
8106892ee8 towncrier 2021-08-28 10:06:22 +02:00
Florent Daigniere
394c2fe22c Document REAL_IP_HEADER and REAL_IP_FROM
Fix a security vulnerability whereby we were not clearing other headers
2021-08-28 10:03:18 +02:00
Florent Daigniere
6bba0cecfc Strip the Forwarded header since nothing is compatible with it yet 2021-08-28 09:02:52 +02:00
Florent Daigniere
0e45bb3ae5 use example.com 2021-08-27 18:49:08 +02:00
Florent Daigniere
d65993886a Fix the links 2021-08-27 18:42:11 +02:00
Florent Daigniere
9e306bf255 use example.com 2021-08-27 18:25:37 +02:00
Florent Daigniere
5ed77750f2 clarify 2021-08-27 18:20:21 +02:00
Florent Daigniere
13e0b56a0d This breaks SSO 2021-08-27 18:17:42 +02:00
bors[bot]
6e32092abd
Merge #1873
1873: Completed Hebrew translation r=mergify[bot] a=yarons

The Hebrew translation is incomplete so I've completed it.

Co-authored-by: Yaron Shahrabani <sh.yaron@gmail.com>
2021-08-27 14:37:54 +00:00
bors[bot]
4c52eb3e0e
Merge #1957
1957: BugFix 1952 - use punycode encoding in HTTP headers for webmail/radicale r=mergify[bot] a=Diman0

## What type of PR?

Bug fix

## What does this PR do?

Fixes a bug introduced by the SSO implementation and an already existing bug for radicale.
In auth.py we did not use punycode (ACE) encoding for the domain part of an email. 
Since we pass the user name in the HTTP header to webmail/radicale, we would sometime pass non-ascii. E.g. user@exämple.io.
This is illegal. HTTP headers may only contain ASCII. The domain part of the user name therefore now uses punycode encoding.

I tested that I can log in with the form user@exämple.io and user@xn--exmple-cua.io for
- admin
- roundcube (also tested sending emails of course)
- rainloop (also tested sending emails of course)
- radicale (webdav)
- thunderbird - sending/receiving emails and accessing/modifying the webdav calendar added in radicale.
  - for the calendar you can use the normal and punnycode notation
  - for email you can only use punnycode. This is a limitation of thunderbird. It does not accept email addresses with non-ascii in the domain part of an email address.

### Related issue(s)
- closes  #1952 

## 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/workflow.html#changelog) entry file.


Co-authored-by: Dimitri Huisman <diman@huisman.xyz>
2021-08-27 09:07:48 +00:00
Dimitri Huisman
169a540692 Use punycode for HTTP header for radicale and create changelog 2021-08-27 08:20:52 +00:00
Dimitri Huisman
4f5cb0974e Make sure HTTP header only contains ASCII 2021-08-26 15:11:35 +00:00
bors[bot]
ecb39fed04
Merge #1947
1947: k8s is helm-charts only r=mergify[bot] a=nextgens

## What type of PR?

documentation

## What does this PR do?

Remove the k8s documentation templates and document that helm charts is the supported way to do it.

### Related issue(s)

- #1451
- closes #1329
- closes #1191 
- closes #1823
- closes #1433
- closes #1285
and
- closes #1420
- closes #1826
- closes #1919


Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
2021-08-26 08:21:04 +00:00
Florent Daigniere
e742c5432b simplify 2021-08-24 18:49:27 +02:00
Florent Daigniere
0a6f3448ec k8s is helm-chart only 2021-08-24 18:42:51 +02:00
bors[bot]
ecaaf25dcb
Merge #1939
1939: Ensure that we don't do multiple DNS lookups in the sieve script r=mergify[bot] a=nextgens

## What type of PR?

bug-fix

## What does this PR do?

It ensures that DNS lookups don't introduce inconsistent state. We may want to go further and actually check the return codes of rspamc too.

I haven't tested it but it should work.

### Related issue(s)
- #1938



Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
2021-08-24 12:42:17 +00:00
Florent Daigniere
368b40b4fd doh 2021-08-24 09:24:14 +02:00
bors[bot]
db0370858e
Merge #1941
1941: Fix a bug whereby adding new HOSTNAMEs won't necessarily lead to certificates being renewed. r=mergify[bot] a=nextgens

## What type of PR?

bug-fix

## What does this PR do?

Fix a bug whereby adding new HOSTNAMEs won't necessarily lead to certificates being renewed.

certbot's defaut behaviour has changed when --renew-with-new-domains was introduced

### Related issue(s)
- close #1270

Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
2021-08-23 19:58:25 +00:00
Florent Daigniere
3e676e232a fix #1270 2021-08-23 19:41:44 +02:00
bors[bot]
7efce99769
Merge #1902
1902: Make smtp_tls_policy_maps easily configurable r=mergify[bot] a=nextgens

## What type of PR?

Feature

## What does this PR do?

- Make smtp_tls_policy_maps easily configurable. This is useful to force TLS verification of specific destinations (or relays).
We should probably discuss what's on the list by default. I have found a top100 list online, ran it through a script to check all the records and found 90 destinations we could use.
- disable TLS session tickets (this reduces the PFS window from 1day to 1h)
- enable system CAs by default (to allow for OUTBOUND_TLS_LEVEL above encrypt without additional overrides)

### Related issue(s)
- closes #1558 
- #707

## 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/workflow.html#changelog) entry file.


Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
2021-08-22 17:11:40 +00:00
Florent Daigniere
ae8db08bdf Ensure that we don't do multiple DNS lookups in the sieve script 2021-08-21 17:14:40 +02:00
bors[bot]
53fe567d06
Merge #1925
1925: Optimize Rainloop: Change to NGINX r=mergify[bot] a=Erriez

## What type of PR?

- Reduce build time.
- Reduce image size.
- Faster user response using CGI.

## What does this PR do?

### Related issue(s)
- Mention an issue like: #1830, #1200 and #1924
- Auto close an issue like: closes #1924

## Prerequistes

Documentation updates TBD (requires some guidance):
- [ ] In case of feature or enhancement: documentation updated accordingly
- [ ] Unless it's docs or a minor change: add [changelog](https://mailu.io/master/contributors/workflow.html#changelog) entry file.

## Technical details

- Image from `php:7.4-apache` to `nginx:1.21-alpine` followed by PHP7 installation.
- Move `.ini` files to directory `defaults/`.
- Move files `sso.php and include.php` to directory `login/`.
- NGINX configuration:
  - `access_log off;` as this is handled by front.
  - `error_log /dev/stderr err; configured to reduce output. The logging in `start.py` is useless.
  - PHP configuration optimized for CGI usage and requires a config file `config/php-rainloop.conf` -> `/etc/php7/php-fpm.d/rainloop.conf`.
  - `.ini` files are parsed / substituted by `socrate` Python module.

Further optimization is possible by completely removing Python. This is only used to parse the `.ini` files and can be done via Bash scripts. This saves more build time and image size can be reduced to 112MB.

## Reviewing

This PR requires multiple reviewers and extensive testing before merging into master. Data/settings are compatible with previous images.

Co-authored-by: Erriez <Erriez@users.noreply.github.com>
2021-08-21 09:52:01 +00:00
Florent Daigniere
65a27b1c7f add additional options to make DANE easier 2021-08-20 14:18:07 +02:00
Florent Daigniere
fb8d52ceb2 Merge branch 'master' of https://github.com/Mailu/Mailu into tls_policy_map 2021-08-20 14:17:34 +02:00
bors[bot]
a461f5fa7c
Merge #1904
1904: Allow specific users to send email from any address r=mergify[bot] a=nextgens

## What type of PR?

Feature

## What does this PR do?

Allow specific users to send email from any address using the WILDCARD_SENDERS configuration variable.

### Related issue(s)
- closes #1096

## 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/workflow.html#changelog) entry file.


Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
Co-authored-by: David Fairbrother <DavidFair@users.noreply.github.com>
Co-authored-by: Diman0 <diman@huisman.xyz>
Co-authored-by: Dimitri Huisman <52963853+Diman0@users.noreply.github.com>
Co-authored-by: Erriez <Erriez@users.noreply.github.com>
2021-08-19 14:48:46 +00:00
bors[bot]
5062ee58dc
Merge #1935
1935: Fix bug #1934: logs flooded with "unbound udp connect failed: Address not available for" r=mergify[bot] a=nextgens

## What type of PR?

bug-fix

## What does this PR do?

Revert back to alpine 1.12 for the resolver/unbound container. The official fix is at:
08968baec1
but alpine doesn't ship it yet:
https://pkgs.alpinelinux.org/packages?name=unbound&branch=v3.14

### Related issue(s)
- closes #1934 


Co-authored-by: Florent Daigniere <nextgens@freenetproject.org>
2021-08-19 14:21:27 +00:00
Florent Daigniere
e1ddbb6eec Rollback to alpine 1.12
it ships unbound 1.10 that doesn't have the bug I think
08968baec1
2021-08-19 15:33:26 +02:00