It's been considered too disruptive and will often misfire especially
if there are issues with the data channels.

It's an evolution of audio-only mode, where we also allow for receiving a remote
screen-share.

Diving deeper: this basically sets last N to 1 or 0 depending on the
availability of a screen-share.

Updating react-native-fast-image brings a couple of interesting changes:

- onLoad is not called for cached images (reported and ignored upstream)
- load progress not working if component not displayed (on Android)

In order to fix this, a combination of 2 approaches was used:

- onLoadEnd / onError are used to detect if the image is loaded
- off-screen rendering is used on Android to get progress events

While implementing the above, yours truly noticed the complexity was increasing
way too much, so some extra refactoring was also performed:

- componentWillReceiveProps is dropped
- an auxiliary component (AvatarContent) is used for the actual content of the
  Avatar, with the former passing the key prop to the latter

Using the key prop ensures AvatarContent will be recreated if the URI changes,
which is not a bad idea anyway, since the new image needs to be downloaded.

Use react-native-fastimage, which uses 2 full-native image impleentations using
well known and mature (native) libraries.

This gets us rid of 2 libraries which were observerd as a source of bugs and
created trouble with dependencies: react-native-fetch-blob and
react-native-img-cache. They are also no longer well maintained.

The Image adds a fade effect without asking, so lets explicitly disable
it. More info here:
https://github.com/facebook/react-native/issues/10194

Use it unless the connection is not ACTIVE. We don't really care if it's
recovering or whatever, if it's not active it has problems, so that's that.

This fixes a potential edge case in which the connection remains in RESTORING
state for some time.

As part of the work on fixing the problem with the multiplying
thumbnails, we've associated remote participant w/ JitsiConference.
However, there are periods of time when multiple JitsiConferences are in
the redux state (and that period is going to be shorted by
StateListenerRegistry). In order to give more control to the feature
base/participants, reduce the occurrences of direct access to the
features/base/participants redux state and utilize the feature's
existing read access functions. Which will allow us in the future to
enhance these functions to access participants which are relevant to the
current conference of interest to the user only.

Since the main conference container is no longer "clickable" there must
be a way for clicking on the "large video". A clickable TestHint nested
in ParticipantView makes it easier for dealing with the fact that the
click handler is not always on the same component (required for the
pinch and zoom feature to work correctly).

TouchableWithoutFeedback and TouchableHighlight interfere with the
implementation of 'pinch to zoom' to come. We prepare for it by driving
the onClick/onPress handler(s) out of Conference, through LargeVideo and
ParticipantView into Video itself where the bulk of 'pinch to zoom' will
be implemented.

In preparation for "pinch to zoom" support in desktop streams on mobile, make
certain Views not intervene in touch event handling. While the modification is
necessary for "pinch to zoom" which is coming later, it really makes sense for
the modified Views to not be involved in touching because they're used to aid
layout and/or animations and are to behave to the user as if they're not there.

Flow caught an incorrect function call.

When in PiP mode the LargeView will not be large enough to hold the avatar (for
those interested in the details, our avatar's size is 200, and in PiP mode the
app is resized to about 150).

In order to solve it, this PR refactors how the avatar style is passed along,
reducing it to a single "size" prop. With this only prop, the Avatar compononent
will compute the width, height and borderRadius, plus deal with some Android
shenanigans.

In addition, the LargeView component now uses DimensionsDetector to check its
own size and adjust the size prop passed to the Avatar component as needed.

The video will switch to the avatar and be tinted with gray. On the large view,
a text message indicating the user has connectivity issues will be shown.

ESLint 4.8.0 discovers a lot of error related to formatting. While I
tried to fix as many of them as possible, a portion of them actually go
against our coding style. In such a case, I've disabled the indent rule
which effectively leaves it as it was before ESLint 4.8.0.

Additionally, remove jshint because it's becoming a nuisance with its
lack of understanding of ES2015+.

* Javadoc introduced @code as a replacement of <code> and <tt> which is
  better aligned with other javadoc tags such as @link. Use it in the
  Java source code. If we switch to Kotlin, then we'll definitely use
  Markdown.

* There are more uses of @code in the JavaScript source code than <tt>
  so use @code for the sake of consistency. Eventually, I'd rather we
  switch to Markdown because it's easier on my eyes.

* Xcode is plain confused by @code and @link. The Internet says that
  Xcode supports the backquote character to denote the beginning and end
  of a string of characters which should be formatted for display as
  code but it doesn't work for me. <tt> is not rendered at all. So use
  the backquote which is rendered itself. Hopefully, if we switch to
  Markdown, then it'll be common between JavaScript and Objective-C
  source code.

Prepare for React 16.

Refactors the previous "[RN] Cache avatars and provide a default in
case load fails" for the purposes of simplification but also modifies
its functionality at the same time. For example:

- Always displays the default avatar immediately which may be seen if
  the remote avatar needs to be downloaded.
- Does not use random colors.
- Uses a default avatar image which is not transparent and ugly but at
  least it's the same image that's used on Web. I've started talks to
  have images/avatar2.png replaced with a transparent and beautiful
  so that will land later on and we'll see the automagic colors in all
  their glory then.

Avatars are cached to the filesystem and loaded from there when requested again.
The cache is cleaned after a conference ends and on application startup
(defensive move).

In addition, implement a fully local avatar system, which is used as a fallback
when loading a remote avatar fails. It can also be forced using a prop.

The fully local avatars use a user icon as a mask and apply a background color
qhich is picked by hashing the URI passed to the avatar. If no URI is passed a
random color is chosen.

A grace period of 1 second is also implemented so a default local avatar will be
rendered if an Avatar component is mounted but has no URI. If a URI is specified
later on, it will be loaded and displayed. In case loading the remote avatar
fails, the locally generated one will be used.

The files styles.js are used (pretty much) on React Native only and each
of them exports 1 name. Export it as default to cut down on source code.

When entering audio-only mode, VideoBridge is instructed to stop sending
remote videos. However, if the instruction fails because DataChannels do
not work, for example, then the app continues to display the remote
videos. Even though they're received in the case of such a failure, no
videos are to be displayed in audio-only mode.

Introduce certain React Components which may be used to write
cross-platform source code such as Audio like Web's audio, Container
like Web's div, Text like Web's p, etc.

Also move (native) avatar style to film-strip styles, since  that's where it
applies. This is analogous to how the large-view avatar is styled.

This reverts commit 739298c782.

They better represent if a participant has video available or not. There are
cases when even a participant in the last N set would not have video because it
disconnected momentarily, for example.