Add support for forwarded messages

[Kaffeine]

Add a new message header keys:

• forward-from (u - Contact_Handle) The contact who sent the original message. If 0 or omitted, the contact who sent the message could not be determined.
• forward-from-id (s) The identifier of the contact who sent the original message, i.e. the result of calling InspectHandles on forward-from. If omitted, clients MUST fall back to looking at forward-from.
• forward-sent-timestamp (x - Unix_Timestamp64) The time the original message was sent.
• forward-received-timestamp (x - Unix_Timestamp64) The time the original message was received.

Maybe some protocols provide information about the original channel. E.g.: Message "Hello" is forwarded from Juliet in the public chat "Montecchi's veranda", but IMO it is OK to miss such capability for now. :)

Extra keys:

• forward-message-token (s) An opaque identifier of the original message, as used by the underlying protocol.

[Kaffeine]

The second option — introduce a new channel iface ofdT.Channel.Interface.Forwarding and use a new message part similar to the header part.

{ # Header
  'message-token': '3ea270cb834c',
  'message-sender': 42,
  'message-sent': 1210067943,
  'message-received': 1210067947,
  'message-type': 0,  # = Channel_Text_Message_Type_Normal
},
{ # Forward header
  'interface': 'ofdT.Channel.Interface.Forwarding'
  'message-token': 'asd234lzxc91s',
  'message-sender': 68, # The contact who sent the original message
  'message-source-id': muc12345, # The original conversation
  'message-source-type': Tp::HandleTypeRoom,
  'message-sent': 1210067243,
  'message-received': 1210067910,
  'message-sender-alias': 'Juliet'
},
{ # Body/other parts
  ...
}

The presence of the interface on a channel will also indicate the 'Message Forwarding' support.

In order to forward a message, a client should specify three values: message-token, message-source/message-source-id and message-source-type. The Connection Manager is expected to take care on all other keys (that should be presented in the MessageSent signal)

[gkiagia]

I prefer the interface approach, it looks more clean and it's also detectable if support for this exists. The headers that you propose seem reasonable.