[Narasimha-sc]

reconnectAllServers closes the clients but left every retry loop waiting out its current interval, so "Reconnect all servers" and a network config change could take up to the maximum interval to deliver anything. It now increments the network epoch as a network change does, after closing the clients.

This is a different trigger from the network change the rest of the stack is about - it does not affect the IP-change case - so it is its own one-line PR.

---

Stack

Reducing reconnection delay after a network change. Each PR is based on the previous one and can be merged on its own once the ones below it are in.

PR		base
#1926	transport: bound unacknowledged-data failure time	master	894 s -> 95 s, no app change, independent
#1927	agent: reconnect to the servers when the network changes	#1926	the 15-minute stall
#1928	agent: fix three races closing clients makes reachable	#1927	correctness guard for #1927
#1929	agent: restart the retry loops when the network changes	#1928	retry interval -> ~3 s
#1930	agent: restart retry loops when reconnecting all servers	#1929	manual reconnect only
simplex-chat/simplex-chat#7650	android, ios: pass the chat controller to the network info command	master	crash fix, independent
simplex-chat/simplex-chat#7651	android, ios: report a network change when the network identity changes	#7650	app side of simplexmq#1927
simplex-chat/simplex-chat#7652	desktop: report network changes to the core	#7651	new capability