A user on Discord questioned whether a DNS resolution failure might be breaking reconnect logic, so I went down a rabbit hole and made some changes.

Functional change:

• Changing options while the app is trying to reconnect will attempt reconnect immediately and reset the backoff logic (users no longer have to wait ~5 minutes for the next attempt for changes to server address or port take effect)
• Clicking the network icon on the UI also resets the backoff logic, and when connecting the app using the UI, the connection 'watchdog' is now used, ensuring retries if the initial attempt fails
• Disconnecting via the API now cancels any in-progress reconnect process
• The network icon overlay changes to indicate when a retry attempt is underway (green sync icon), when we're waiting for the next attempt (yellow clock) and when the app is both disconnected and not making an attempt to reconnect (red x)

Connecting:

Waiting:

Code changes:

• Remove IConnectionWatchdog interface. Lately I'm finding C# interfaces annoying and redundant 🤷
• Add a cancellation token to the retry logic and the delays and connection logic within to ensure the loop is exited promptly
• Added ServerConnectionWatchdog to application state that includes flags to indicate when it's active, when it's in a retry loop, and the approximate timestamp of the next attempt

Also fixes what I'd consider a bug with the server state; address and IP are now null when the server is disconnected.