The Windows client sends a best-effort disconnect control message after the
runtime loop exits. That message previously always said `client shutting down`,
even when the loop ended because TAP I/O, relay event handling, or relay-route
verification failed.
Preserve the returned error, but include its chain in the disconnect message so
the relay log records why the client left during manual MVP testing. Clean exits
still report the original shutdown text.
Test Plan:
- cargo fmt --check
- cargo test -p lanparty-client-win
- cargo test --workspace
- cargo clippy --workspace --all-targets -- -D warnings
- git diff --check
- git diff --cached --check
Refs: MVP Windows client diagnostics
ddidderr
5f567b4320