Summary

Implements cooperative idle server cleanup for sbtn (issue #8610). When a client disconnects from an sbt server, that server notifies all
other registered servers to shut down if they've been idle long enough and have no connected clients. This prevents accumulation of idle
background JVMs across projects.

fixes #8610

Problem

Each sbtn session leaves a background JVM server running with a 7-day idle timeout. Users working across many projects accumulate idle
servers that waste system resources (memory, file descriptors, threads).

Solution

• Each server registers itself via a <PID>.json file in a shared proc directory (~/.cache/sbt/v2/proc/ on Linux,
  ~/Library/Caches/sbt/v2/proc/ on macOS).
• When a client disconnects, the server sends a sbt/dropIfIdle JSON-RPC notification to all other registered servers.
• A server receiving dropIfIdle shuts down if idle >= sbt.server.secondaryIdleTimeout (default 600s) AND has no initialized
  NetworkChannel clients.
• Stale proc files (unreachable servers) are cleaned up on contact failure.
• Anti-cascading: only initialized channels (real clients that completed the initialize handshake) trigger notifyOtherServers on
  disconnect. Temporary dropIfIdle connections never initialize, so their removal won't cascade.

Testing

• Manual: started server with -Dsbt.server.secondaryIdleTimeout=5, verified proc file created, sent dropIfIdle notification, confirmed
  server drops when idle with no clients and stays alive when a client is connected.
• Verified proc file cleanup on shutdown.