The problem with #roster{} is that every new record entry is also stored
in the mnesia roster table. Adding the mix_participant_id there makes
no sense because the normal roster items are no MIX channels. Using
\#roster_item{} for the hook and #roster{} for storing the normal items
seems to be a better idea.
If we use _`whatever`_ here in ejabberd man pages,
it is converted to *`whatever`* in markdown,
and docs.ejabberd.im/Makefile converts to the proper links
This caused garbled text in some places in webadmin when using language
that used characters > 128.
Thanks to chengshq for noticing this and providing preliminary patch.
Sometimes we can observer combinations of subscription/ask that shouldn't
happen normally, but can be generated with api calls, let's try to handle
that gracefully instead of crashing.
Several documentation callbacks (doc/0 and mod_doc/0) are implemented
and `ejabberdctl man` command is added to generate a man page. Note
that the command requires a2x to be installed (which is a part of
asciidoc package).
By calling erlang:get_stacktrace() inside a lager function
we obtain actually a stacktrace of the lager function, not
the one we got during exception. This is not a problem for
newest Erlang versions though.
Now every such string MUST be encapsulated into ?T() macro.
The macro itself is defined in include/translate.hrl.
Example:
-module(foo).
-export([bar/1]).
-include("translate.hrl").
bar(Lang) ->
translate:translate(Lang, ?T("baz")).
Since we now require R19, we shouldn't need that anymore.
There are still couple places where p1_time_compat:unique_timestamp() is
used as there is no direct equivalent.
The header consisted of too many unrelated stuff and macros misuse.
Some stuff is moved into scram.hrl and type_compat.hrl.
All macros have been replaced with the corresponding function calls.
TODO: probably type_compat.hrl is not even needed anymore since
we support only Erlang >= OTP 17.5
Since we got rid of all bottle-neck processes and we have
a connection pool for every database, the option is no longer
needed and in fact is detrimental: in practice what you get
is just a bunch of overloaded processes in the IQ handlers pool
no matter how much you increase the `iqdisc` value.
Given that there are close to zero operators understanding
the meaning of the option and, hence, not using it all,
it's not simply deprecated but completely removed.
The commit also deprecates the following functions:
- gen_iq_handler:add_iq_handler/6
- gen_iq_handler:handle/5
- gen_iq_handler:iqdisc/1
