GitLab

Thanks terminals, for your wonderful standardisation!

Previously it behaved as if the response was empty, which was confusing.

self-ping: do not /cycle on timeout, log only

See merge request poezio/poezio!16

Invite impromptu

See merge request poezio/poezio!2

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

I was originally panning to generate a user-friendly room name with
them, but this can/should be done at runtime (in case it changes).
Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

See #3447 instead. This can be generated at runtime.
Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Adding missing MAM bits, removing the allowpm restriction, as clients
should already be using realjids when available.

I am leaving allowinvites for now as I think that should be in there,
but we can align ourselves later on if necessary, I'm not totally
against the idea.
Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Add a command that invites people to a newly created room, with a random
localpart. The muc component is currently static. The interface
for the command might also change later on.
Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Allow additional information elements in MucInfoWin same as ConversationInfoWin

See merge request poezio/poezio!17

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

Signed-off-by: Maxime “pep” Buquet <pep@bouah.net>

on getting Link Mauve to ever run tests

This also changes from git:// to https:// because many networks like to
block non-standard (read non-HTTP) ports, especially in companies or
universities.

Fixes a regression introduced in
cce1a409 when displaying data forms.

It is relied on in quite a few places which don’t check for None before
calling it on messages which have no nick.

If we ignore it, it’s just noise.

We need to decide:
- fix the code to be like yapf wants to be (I’m against this)
- change the rules, to make them way more flexible, and then apply them
- don’t run yapf at all

tentative fix