Especially useful for very long subject lines

fixes #4429

the 2.11.0 release removed support for enabling sm:2
unfortunatly sm:2 was still detected as "server supports stream managment"
down the line leading to resend loops.

fixes #4426

Since you can't install any other version while the data is present, and it's
hard to clear, it really doesn't help in the ways we hoped.

fixes #4421

fixes #4421

* tag '2.11.0': (101 commits)
  version bump to 2.11.0
  temporarily use Snikket’s build of WebRTC
  delay candidates until after session-init/accept
  pulled translations from transifex
  pulled translations from transifex
  null PeerConnection reference before disposing; otherwise getState() might be issued against disposed object
  avoid race condition when restarting ICE
  version bump to 2.11.0-beta.2
  pulled translations from transifex
  add switch to video menu item to call
  prepare JingleRtpConnection for content-adds
  trim xmpp address after user input
  add helper methods for content modification to RtpContentMap
  ensure cc-ed proceed is equivalent to accept
  rename initiateIceRestart to renegotiate to handle content adds
  take senders attr into account when converting to and from sdp
  make sure VideoSourceWrapper is stored in property
  refactor WebRTCWrapper to allow for track adds
  use plurals for missed call strings
  version bump to 2.11.0-beta
  ...

Even in tables

Approved by F-Droid for use and is up to date

Just ignore the button press is there isn't one.

We setup the buttons, but never clear them, so when a person has enough
messages in their view and one of them has a button, as they scroll
around and views get recycled the button can be attached onto any
message and not just the ones that actually specify it.

The documentation makes it a bit unclear if the adapter can be set to
null or not... so I just clear it instead. I could leave it, I suppose,
but that feels a bit weird, and the next time we go through this code
we'll just replace it with another one anyway.

Data and Field Element subclasses expect to be able to mutate the list of
children on the original Element that they are derived from using parse.  The
way the new setChildren worked resulted in not actually mutating the original.

Change this for a bindTo that makes sure all fields reference the original (and
hence any mutation on the new object mutates the original) and also a
replaceChildren that uses clear+addAll to ensure that the existing list is
mutated rather than a new list being used.

In practise, this affected room configuration and other data forms being
properly filled out.