Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
Fixes #3376.
|
|
- nick colors are not random by default
- /close is a useful command
|
|
|
|
leading to tracebacks and shenanigans
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Fixes #3213
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
(and fix some small bugs that went unnoticed for years)
|
|
|
|
|
|
Make it slightly more declarative and readable (and longer), and put
it at the end of the file.
|
|
|
|
|
|
|
|
add a log because it should not happen
|
|
Keep all presence before status code=110 in a buffer, then batch-process
them when receiving our initial join.
|
|
|
|
|
|
|
|
|
|
When a user isn't in a room, but due to a component bug we receive a
unavailable presence, we were showing a join.
|
|
|
|
exit the rooms properly, call functions instead of full-blown commands,
do not randomly focus stabs, etc
|
|
|
|
unused variables, imports, name shadowing
|
|
Make hide_status_change work there as well, with per-jid configuration
|
|
|
|
|
|
|
|
|
|
this call is actually quite expensive, so it will help in specific cases
where we receive a lot of presence stanzas (e.g. a netsplit in an irc
gateway, or a +vvvvvvvvvvvvvvvvvvvvvvvvv, etc) that are not joins
|