Add test-network.sh and fix IPC goroutine lifecycle bugs

test-network.sh: shell script that starts anchor + alice + bob + charlie,
joins them to a named network, exchanges five messages, then has charlie
leave — verifying the full join/chat/leave cycle with coloured per-peer output.

ipc: two fixes exposed by the test script:
- Network join context was per-IPC-client, so the join was immediately
  cancelled when the nc connection closed. Lifted join/leave state to the
  Run() level (shared across all clients, protected by a mutex) so the
  network stays connected independent of which client issued the command.
- Event-pusher goroutine could panic with "send on closed channel" when
  the command loop closed writeCh while the pusher was mid-select. Added
  defer recover() to both the pusher goroutine and the send helper, and
  removed the default arm so the select blocks cleanly on done.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

internal/ipc/ipc.go

@@ -12,6 +12,7 @@ import (
 	"fmt"
 	"log"
 	"net"
+	"sync"
 	"time"
 
 	"github.com/google/uuid"
@@ -20,10 +21,10 @@ import (
 )
 
 // JoinFunc is called when the UI issues a join_network command.
-// It should connect to the anchor and block until done or ctx is cancelled.
 type JoinFunc func(ctx context.Context, networkName string)
 
 // Run starts the IPC listener. Blocks until the listener fails.
+// Network join/leave state is daemon-scoped (shared across all IPC clients).
 func Run(m *mesh.Mesh, port int, join JoinFunc) error {
 	addr := fmt.Sprintf("127.0.0.1:%d", port)
 	ln, err := net.Listen("tcp", addr)
@@ -32,24 +33,53 @@ func Run(m *mesh.Mesh, port int, join JoinFunc) error {
 	}
 	log.Printf("ipc: listening on %s", addr)
 
+	// networkCancel is shared across all clients — any client can join/leave,
+	// and the join persists even after the commanding client disconnects.
+	var (
+		networkMu     sync.Mutex
+		networkCancel context.CancelFunc
+	)
+
+	doJoin := func(name string) {
+		networkMu.Lock()
+		if networkCancel != nil {
+			networkCancel()
+		}
+		ctx, cancel := context.WithCancel(context.Background())
+		networkCancel = cancel
+		networkMu.Unlock()
+		go join(ctx, name)
+	}
+
+	doLeave := func() {
+		networkMu.Lock()
+		if networkCancel != nil {
+			networkCancel()
+			networkCancel = nil
+		}
+		networkMu.Unlock()
+	}
+
 	for {
 		conn, err := ln.Accept()
 		if err != nil {
 			return fmt.Errorf("ipc accept: %w", err)
 		}
 		log.Printf("ipc: UI client connected from %s", conn.RemoteAddr())
-		go handleClient(conn, m, join)
+		go handleClient(conn, m, doJoin, doLeave)
 	}
 }
 
-func handleClient(conn net.Conn, m *mesh.Mesh, join JoinFunc) {
+func handleClient(conn net.Conn, m *mesh.Mesh, doJoin func(string), doLeave func()) {
 	defer conn.Close()
 
 	events := m.Subscribe()
 	defer m.Unsubscribe(events)
 
 	writeCh := make(chan []byte, 128)
+	done := make(chan struct{})
 
+	// Writer goroutine — sole owner of the write side of the connection.
 	go func() {
 		w := bufio.NewWriter(conn)
 		for line := range writeCh {
@@ -61,31 +91,51 @@ func handleClient(conn net.Conn, m *mesh.Mesh, join JoinFunc) {
 		}
 	}()
 
+	// Event pusher — forwards mesh events to the UI client.
+	// recover() guards against the rare race where writeCh is closed while a
+	// send is in flight (closed channel panics even inside select).
 	go func() {
-		for evt := range events {
+		defer func() { recover() }() //nolint:errcheck
+		for {
+			select {
+			case evt, ok := <-events:
+				if !ok {
+					return
+				}
 				line, err := json.Marshal(evt)
 				if err != nil {
 					continue
 				}
 				select {
 				case writeCh <- line:
-			default:
+				case <-done:
+					return
+				}
+			case <-done:
+				return
 			}
 		}
 	}()
 
-	send(writeCh, proto.IpcMessage{
+	send := func(msg proto.IpcMessage) {
+		defer func() { recover() }() //nolint:errcheck
+		line, err := json.Marshal(msg)
+		if err != nil {
+			return
+		}
+		select {
+		case writeCh <- line:
+		case <-done:
+		}
+	}
+
+	send(proto.IpcMessage{
 		Type:           proto.EvtStateSnapshot,
 		LocalPeer:      ptr(m.Identity.PeerInfo()),
 		ConnectedPeers: m.ConnectedPeers(),
 		Rooms:          []string{"general"},
 	})
 
-	// Track an active network join so we can cancel it on leave_network.
-	var (
-		networkCancel context.CancelFunc
-	)
-
 	scanner := bufio.NewScanner(conn)
 	for scanner.Scan() {
 		var cmd proto.IpcMessage
@@ -115,24 +165,16 @@ func handleClient(conn net.Conn, m *mesh.Mesh, join JoinFunc) {
 
 		case proto.CmdJoinNetwork:
 			if cmd.NetworkName == "" {
-				send(writeCh, errMsg("join_network: network_name is required"))
+				send(errMsg("join_network: network_name is required"))
 				continue
 			}
-			if networkCancel != nil {
+			doJoin(cmd.NetworkName)
-				networkCancel() // leave any previous network
-			}
-			ctx, cancel := context.WithCancel(context.Background())
-			networkCancel = cancel
-			go join(ctx, cmd.NetworkName)
 
 		case proto.CmdLeaveNetwork:
-			if networkCancel != nil {
+			doLeave()
-				networkCancel()
-				networkCancel = nil
-			}
 
 		case proto.CmdGetState:
-			send(writeCh, proto.IpcMessage{
+			send(proto.IpcMessage{
 				Type:           proto.EvtStateSnapshot,
 				LocalPeer:      ptr(m.Identity.PeerInfo()),
 				ConnectedPeers: m.ConnectedPeers(),
@@ -140,31 +182,18 @@ func handleClient(conn net.Conn, m *mesh.Mesh, join JoinFunc) {
 			})
 
 		case proto.CmdSendFile:
-			send(writeCh, errMsg("file transfer not yet implemented"))
+			send(errMsg("file transfer not yet implemented"))
 
 		default:
-			send(writeCh, errMsg(fmt.Sprintf("unknown command: %s", cmd.Type)))
+			send(errMsg(fmt.Sprintf("unknown command: %s", cmd.Type)))
 		}
 	}
 
-	if networkCancel != nil {
+	close(done)
-		networkCancel()
-	}
 	close(writeCh)
 	log.Printf("ipc: UI client disconnected")
 }
 
-func send(ch chan<- []byte, msg proto.IpcMessage) {
-	line, err := json.Marshal(msg)
-	if err != nil {
-		return
-	}
-	select {
-	case ch <- line:
-	default:
-	}
-}
-
 func errMsg(s string) proto.IpcMessage {
 	return proto.IpcMessage{Type: proto.EvtError, ErrorMessage: s}
 }

test-network.sh

@@ -0,0 +1,251 @@
+#!/usr/bin/env bash
+# test-network.sh — spin up anchor + 3 peers, connect them, exchange messages.
+# Usage: ./test-network.sh
+# Requires: go, nc, jq
+
+set -euo pipefail
+
+# ── colours ───────────────────────────────────────────────────────────────────
+RED='\033[0;31m'; GREEN='\033[0;32m'; YELLOW='\033[1;33m'
+CYAN='\033[0;36m'; MAGENTA='\033[0;35m'; BLUE='\033[0;34m'
+BOLD='\033[1m'; DIM='\033[2m'; RESET='\033[0m'
+
+ANCHOR_COLOR=$YELLOW
+ALICE_COLOR=$CYAN
+BOB_COLOR=$MAGENTA
+CHARLIE_COLOR=$GREEN
+
+# ── config ────────────────────────────────────────────────────────────────────
+ANCHOR_PORT=19339
+ALICE_IPC=19337
+BOB_IPC=19340
+CHARLIE_IPC=19343
+NETWORK_NAME="test-$(date +%s)"
+DATA_ROOT=$(mktemp -d /tmp/waste-test-XXXXXX)
+
+# ── cleanup ───────────────────────────────────────────────────────────────────
+PIDS=()
+cleanup() {
+    echo ""
+    echo -e "${DIM}── cleaning up ──────────────────────────────────────────${RESET}"
+    for pid in "${PIDS[@]:-}"; do
+        kill "$pid" 2>/dev/null || true
+    done
+    wait 2>/dev/null || true
+    rm -rf "$DATA_ROOT"
+    echo -e "${DIM}done.${RESET}"
+}
+trap cleanup EXIT INT TERM
+
+# ── helpers ───────────────────────────────────────────────────────────────────
+log() {
+    local color="$1"; shift
+    local label="$1"; shift
+    echo -e "${color}${BOLD}[${label}]${RESET} $*"
+}
+
+# Pretty-print a raw IPC JSON line
+pretty() {
+    local color="$1"
+    local label="$2"
+    local line="$3"
+    local type
+    type=$(echo "$line" | jq -r '.type // "?"' 2>/dev/null)
+    case "$type" in
+        state_snapshot)
+            local id nick
+            id=$(echo "$line" | jq -r '.local_peer.id[:16]' 2>/dev/null | sed 's/.\{4\}/& /g')
+            nick=$(echo "$line" | jq -r '.local_peer.alias' 2>/dev/null)
+            echo -e "${color}${BOLD}[${label}]${RESET} ${DIM}state_snapshot${RESET} id=${CYAN}${id}${RESET} alias=${BOLD}${nick}${RESET}"
+            ;;
+        session_ready)
+            local pid nick
+            pid=$(echo "$line" | jq -r 'if .peer_id then .peer_id[:16] else "?" end' 2>/dev/null | sed 's/.\{4\}/& /g')
+            nick=$(echo "$line" | jq -r '.nick // "?"' 2>/dev/null)
+            echo -e "${color}${BOLD}[${label}]${RESET} ${GREEN}✓ session_ready${RESET} peer=${BOLD}${nick}${RESET} (${DIM}${pid}${RESET})"
+            ;;
+        peer_connected)
+            local nick
+            nick=$(echo "$line" | jq -r '.peer.alias // "?"' 2>/dev/null)
+            echo -e "${color}${BOLD}[${label}]${RESET} ${GREEN}→ peer_connected${RESET} ${BOLD}${nick}${RESET}"
+            ;;
+        peer_disconnected)
+            local pid
+            pid=$(echo "$line" | jq -r 'if .peer_id then .peer_id[:8] else "?" end' 2>/dev/null)
+            echo -e "${color}${BOLD}[${label}]${RESET} ${RED}← peer_disconnected${RESET} ${DIM}${pid}${RESET}"
+            ;;
+        message_received)
+            local from body room
+            from=$(echo "$line" | jq -r '.message.from[:8] // "?"' 2>/dev/null)
+            body=$(echo "$line" | jq -r '.message.body // ""' 2>/dev/null)
+            room=$(echo "$line" | jq -r '.message.room // ""' 2>/dev/null)
+            echo -e "${color}${BOLD}[${label}]${RESET} ${BOLD}💬 #${room}${RESET} ${DIM}<${from}…>${RESET} ${body}"
+            ;;
+        error)
+            local msg
+            msg=$(echo "$line" | jq -r '.error_message // .' 2>/dev/null)
+            echo -e "${color}${BOLD}[${label}]${RESET} ${RED}✗ error:${RESET} ${msg}"
+            ;;
+        *)
+            echo -e "${color}${BOLD}[${label}]${RESET} ${DIM}${line}${RESET}"
+            ;;
+    esac
+}
+
+# Subscribe to a peer's IPC events in the background, pretty-printing each line.
+subscribe() {
+    local color="$1"
+    local label="$2"
+    local port="$3"
+    (
+        set +e
+        nc 127.0.0.1 "$port" 2>/dev/null | while IFS= read -r line; do
+            pretty "$color" "$label" "$line"
+        done
+    ) &
+    PIDS+=($!)
+}
+
+# Send a single JSON command to a peer's IPC port (fire-and-forget).
+ipc() {
+    local port="$1"
+    local json="$2"
+    echo "$json" | nc -q 0 127.0.0.1 "$port" >/dev/null 2>&1 || true
+}
+
+# Wait until a TCP port accepts connections (up to N seconds).
+wait_port() {
+    local port="$1"
+    local label="$2"
+    local n=0
+    while ! nc -z 127.0.0.1 "$port" 2>/dev/null; do
+        sleep 0.1
+        n=$(( n + 1 ))
+        if [ "$n" -gt 80 ]; then
+            echo -e "${RED}timeout waiting for ${label} on :${port}${RESET}" >&2
+            exit 1
+        fi
+    done
+}
+
+# ── build ─────────────────────────────────────────────────────────────────────
+echo ""
+echo -e "${BOLD}waste-go network test${RESET}"
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+echo -e "${DIM}network : ${BOLD}${NETWORK_NAME}${RESET}"
+echo -e "${DIM}data    : ${DATA_ROOT}${RESET}"
+echo ""
+
+echo -e "${DIM}building binaries…${RESET}"
+go build -o "$DATA_ROOT/bin/waste-anchor" ./cmd/anchor
+go build -o "$DATA_ROOT/bin/waste-daemon" ./cmd/daemon
+echo -e "${GREEN}✓ built${RESET}"
+echo ""
+
+# ── anchor ────────────────────────────────────────────────────────────────────
+log "$ANCHOR_COLOR" "anchor" "starting on :${ANCHOR_PORT}"
+"$DATA_ROOT/bin/waste-anchor" -bind "127.0.0.1:${ANCHOR_PORT}" \
+    2> >(while IFS= read -r l; do echo -e "${ANCHOR_COLOR}${DIM}[anchor] ${l}${RESET}"; done) &
+PIDS+=($!)
+wait_port "$ANCHOR_PORT" "anchor"
+log "$ANCHOR_COLOR" "anchor" "ready"
+echo ""
+
+# ── daemons ───────────────────────────────────────────────────────────────────
+ANCHOR_URL="ws://127.0.0.1:${ANCHOR_PORT}/ws"
+
+log "$ALICE_COLOR" "alice" "starting daemon (ipc :${ALICE_IPC})"
+"$DATA_ROOT/bin/waste-daemon" \
+    -alias alice \
+    -data-dir "$DATA_ROOT/alice" \
+    -ipc-port "$ALICE_IPC" \
+    -anchor "$ANCHOR_URL" \
+    2> >(while IFS= read -r l; do echo -e "${ALICE_COLOR}${DIM}[alice]  ${l}${RESET}"; done) &
+PIDS+=($!)
+
+log "$BOB_COLOR" "bob" "starting daemon (ipc :${BOB_IPC})"
+"$DATA_ROOT/bin/waste-daemon" \
+    -alias bob \
+    -data-dir "$DATA_ROOT/bob" \
+    -ipc-port "$BOB_IPC" \
+    -anchor "$ANCHOR_URL" \
+    2> >(while IFS= read -r l; do echo -e "${BOB_COLOR}${DIM}[bob]    ${l}${RESET}"; done) &
+PIDS+=($!)
+
+log "$CHARLIE_COLOR" "charlie" "starting daemon (ipc :${CHARLIE_IPC})"
+"$DATA_ROOT/bin/waste-daemon" \
+    -alias charlie \
+    -data-dir "$DATA_ROOT/charlie" \
+    -ipc-port "$CHARLIE_IPC" \
+    -anchor "$ANCHOR_URL" \
+    2> >(while IFS= read -r l; do echo -e "${CHARLIE_COLOR}${DIM}[charlie]${l}${RESET}"; done) &
+PIDS+=($!)
+
+wait_port "$ALICE_IPC"   "alice"
+wait_port "$BOB_IPC"     "bob"
+wait_port "$CHARLIE_IPC" "charlie"
+echo ""
+
+# ── subscribe ─────────────────────────────────────────────────────────────────
+echo -e "${DIM}subscribing to IPC event streams…${RESET}"
+subscribe "$ALICE_COLOR"   "alice  " "$ALICE_IPC"
+subscribe "$BOB_COLOR"     "bob    " "$BOB_IPC"
+subscribe "$CHARLIE_COLOR" "charlie" "$CHARLIE_IPC"
+sleep 0.3   # let state_snapshot lines arrive
+
+# ── join network ──────────────────────────────────────────────────────────────
+echo ""
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+echo -e "joining all peers to network: ${BOLD}${NETWORK_NAME}${RESET}"
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+
+JOIN=$(printf '{"type":"join_network","network_name":"%s"}' "$NETWORK_NAME")
+ipc "$ALICE_IPC"   "$JOIN"
+sleep 0.2
+ipc "$BOB_IPC"     "$JOIN"
+sleep 0.2
+ipc "$CHARLIE_IPC" "$JOIN"
+
+echo -e "${DIM}waiting for ICE / DataChannel setup (up to 10s)…${RESET}"
+sleep 6
+
+# ── chat ──────────────────────────────────────────────────────────────────────
+echo ""
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+echo -e "sending messages"
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+
+ipc "$ALICE_IPC" '{"type":"send_message","room":"general","body":"hey everyone, alice here"}'
+sleep 0.4
+ipc "$BOB_IPC" '{"type":"send_message","room":"general","body":"hi alice! bob checking in"}'
+sleep 0.4
+ipc "$CHARLIE_IPC" '{"type":"send_message","room":"general","body":"charlie online. the mesh works!"}'
+sleep 0.4
+ipc "$ALICE_IPC" '{"type":"send_message","room":"general","body":"nice — three-way chat over WebRTC 🎉"}'
+sleep 0.4
+ipc "$BOB_IPC" '{"type":"send_message","room":"general","body":"no central server, fully peer-to-peer"}'
+sleep 0.4
+
+# ── state snapshot ────────────────────────────────────────────────────────────
+echo ""
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+echo -e "requesting state snapshots"
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+ipc "$ALICE_IPC"   '{"type":"get_state"}'
+ipc "$BOB_IPC"     '{"type":"get_state"}'
+ipc "$CHARLIE_IPC" '{"type":"get_state"}'
+sleep 0.5
+
+# ── leave ─────────────────────────────────────────────────────────────────────
+echo ""
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+echo -e "charlie leaves the network"
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+ipc "$CHARLIE_IPC" '{"type":"leave_network"}'
+sleep 1
+
+echo ""
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+echo -e "${GREEN}${BOLD}test complete${RESET}"
+echo -e "${DIM}────────────────────────────────────────────────────────${RESET}"
+sleep 0.5