[Telepathy-commits] merge: telepathy-gabble

[Merge Monkey]

Merge to:

  telepathy-gabble

From:

  /home/robot101/public_html/telepathy-gabble-rob

Changes:

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    build MUC self_jid with the alias for our self_handle, not just the local part of the JID (note it's guaranteed not to be NULL, as _get_cached_alias falls back to the local part of the JID anyway)

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    make gabble_presence_as_message accept NULL as a shorthand for "first resource", as self_presence only has one resource

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    invert sense of the if (priv->invited)

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    make create_room_identity return self_handle directly

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    don't queue MUC invite reasons as a text message, we've got the LocalPendingMembers property

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    set self_jid directly from create_room_identity (and a nice assertion which uncovered the double-setting bug)

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    when we're invited, mark the channel as ready in the constructor (emitting a signal is pointless before the constructor has returned)

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    store MUC state booleans in bit fields

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    remove resource argument from gabble_presence_as_message entirely, as it's only ever useful on your self_presence which only has one resource

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    fix MUC test to actually finish the function

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    inline _handle_invited function into the constructor

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    rename contact_handle_to_room_identity to create_room_identity and remove needless self_handle argument

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    push out new MUC presences when the global presence is changed

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    rename MUC "invite-self" property to "invited" with the inverse sense, because I have no idea what "invite-self" is meant to mean

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    add a "to" argument to gabble_presence_as_message, as you can't change the recipient of an LmMessage after it's been created

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    add test for MUC alias coming from our connection alias

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    fix repeated creation (and leaking) of self_jid when invited to a MUC

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    remove needless assertions (including removing the ability to cause an assertion error by sending an old-style MUC invite with no body)

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    test that closing the MUC results in sending unavailable presence to it

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    generate MUC presence messages using the real connection-wide presence
With a pedantic interpretation, this is arguable a spec violation because it
exposes your presence to handles who are not on your "publish" list. However, I
think that given other protocols (essentially, anything but XMPP) exist where
joining a chat will always expose your presence, I don't think it's reasonable
for clients or users to rely on this additional exposure not happening anyway.
Consider the absurdity of expecting that you can be invisible even when you've
joined a group chat. If someone wants the XMPP-specific semantics badly, we can
add a RoomPresence interface or somesuch, and they can set a property in order
to publish a per-room presence.

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    remove unused code (the function is only ever called after assert priv->initiator != 0)

  Robert McQueen <robert.mcqueen at collabora.co.uk>
    rename ready_emitted to ready