Jason WebSocket reconnection (#47)

[evdokimovs]

Part of #47

Required for #51

Synopsis

Currently, when connection between server and client is lost, nothing happen. But we need mechanism of reconnection with some deadline and delay between reconnection tries.

Note that in this PR will be implemented only WebSocket reconnection without state synchronization. Synchronization of the Jason state will be implemented in #51.

Solution

• 1. Send Event::RpcSettingsUpdated with idle_timeout and reconnection_timeout from a server
• 2. Implement reconnection in RpcTransport
• 3. Implement reconnection in RpcClient
• 4. Add delay for RpcClient reconnection tries
• 5. Add deadline for RpcClient reconnection (which will be calculated based on Event::RpcSettingsUpdated)
• 6. Add tests for this functional

Checklist

• Created PR:
  • In draft mode
  • Name contains WIP: prefix
  • Name contains issue reference
  • Has k:: labels applied
  • Has assignee
• Documentation is updated (if required)
• Tests are updated (if required)
• Changes conform code style
• CHANGELOG entry is added (if required)
• FCM (final commit message) is posted
  • and approved
• Review is completed and changes are approved
• Before merge:
  • Milestone is set
  • PR's name and description are correct and up-to-date
  • WIP: prefix is removed
  • All temporary labels are removed

[evdokimovs]

FCM

Jason WebSocket reconnection (#75, #47)

- implement `RpcClient`reconnection
- add 'ServerMsg::RpcSettings' and this message sending and processing
- add 'Room.on_connection_loss' callback which fires when 'RpcClient' loses connection
- add 'ReconnectorHandle' which is used on JS side for reconnection starting (or not)
- implement reconnection with backoff and simple reconnection with delay
- add connection loss notification with button for manual reconnection to the 'e2e-demo'
- add WebSocket connection state indicator to the 'demo'
- remove 'RpcTransport::on_close` and use `RpcTransport::on_state_change` instead of it

Additionally:
- rewrite 'satisfies_by_device_id' macro and 'console_error' macro as functions
- remove 'weak_map!' macro and add 'upgrade_or_detached!' instead of it
- add 'new_js_error!' macro for creating 'JasonError' with 'tracerr' information for JS side
- reverse `ping-pong` mechanism: server sends 'Ping' and expects 'Pongs' from clients now
- add 'rpc.ping_interval' configuration
- use 'fakerator' instead of 'faker' for generating random male usernames in demos

[evdokimovs]

FCM

Jason WebSocket reconnection (#75, #47)

- impl RpcClient reconnection
- add ServerMsg::RpcSettings and its sending/processing
- add Room.on_connection_loss callback which fires when RpcClient loses connection
- add ReconnectHandle for JS side to start reconnection
- impl reconnection with backoff and simple reconnection with a constant retry delay
- add connection loss notification with button for manual reconnection to 'e2e-demo' app
- add WebSocket connection state indicator to 'demo' app
- remove RpcTransport::on_close and use RpcTransport::on_state_change instead

Additionally:
- rewrite 'satisfies_by_device_id!' and 'console_error!' macros as functions
- remove 'weak_map!' macro and add 'upgrade_or_detached!' instead
- add 'new_js_error!' macro for creating JasonError with 'tracerr' information for JS side
- reverse ping/pong mechanism: server sends Ping and expects Pong from client
- add 'rpc.ping_interval' configuration option for Medea
- use 'fakerator' instead of 'faker' for generating random male usernames in demos

[alexlapa]

@evdokimovs ,
LGTM.
Retest, update FCM. If everything is ok, then send to tyranron.