Define and macro by setting as environment variable:
EJABBERD_MACRO_ + macro name
For example, if you configure in ejabberd.yml:
define_macro:
LOGLEVEL: 4
loglevel: LOGLEVEL
You can define (and overwrite) that macro definition when starting ejabberd:
EJABBERD_MACRO_LOGLEVEL=5 make relive
Until now it was possible to create a shared roster group with name
"Group1", and it was a different group that "group1".
From now on, new group names will be stored lowercase,
just like the username in a Jabber ID.
This only affects commands srg_add and srg_create.
All the other commands are still case sensitive, to allow admins of
existing databases with case-sensitive groups manage them.
Until now it created all the rooms in the storage of the first vhost
listed in the ejabberd configuration file. Similarly, it used only
the default room options defined for the first vhost.
Allow to specify an Elixir module name in `auth_method`.
If the referenced module, `M`, cannot be loaded as `ejabberd_auth_M`,
try to load it as `Elixir.M`.
The problem appeared when there are several vhosts,
the first vhost uses SQL for persistent data (and RAM for volatile),
and another vhost wants to use Mnesia
Example config to trigger the problem:
hosts:
- mysql.localhost
- localhost
host_config:
mysql.localhost:
default_db: sql
In that case, ejabberd crashed at start with an error like:
[critical] Internal error of module mod_muc has occurred during start:
...
** exception exit: {aborted,
{no_exists,
[muc_room,
[{{muc_room,{'_',<<"conference.localhost">>},'_'},
[],
['$_']}]]}}
The CI workflow uses ubuntu-24.04 and setup-beam action,
so it can test only Erlang/OTP 24 and higher.
To ensure Dialyzer is also ran with older Erlang versions,
let's add Dialyzer testing to the Runtime workflow,
which uses the erlang container, and that allows
to run Erlang/OTP 20 in ubuntu-24.04.
STUN via UDP can easily be abused for reflection/amplification DDoS
attacks. Suggest a non-standard port to make it harder for attackers to
discover the service.
Modern XMPP clients discover the port via XEP-0215, so there's no
advantage in sticking to the standard port.
Some users reported ejabberd being restarted by systemd due to missing
watchdog pings despite the actual service operating just fine. So far,
we weren't able to track down the issue, so we'll no longer enable the
watchdog in our example service unit.
Closes#4054.
