Files
navidrome/plugins/host_websocket.go
Deluan Quintão 205c85da55 fix(plugins): surface host service failures when loading plugins (#5756)
* fix(plugins): surface host service failures when loading plugins

When a host service failed to initialize during plugin load (e.g. the
taskqueue database could not be created because the data folder is not
writable), the error was logged and swallowed, and its host functions were
silently omitted. Instantiation then failed with a misleading error such as
'"task_createqueue" is not exported in module "extism:host/user"', which
reads as a plugin/host API mismatch and gets wrongly blamed on plugin
authors (see kgarner7/navidrome-listenbrainz-daily-playlist#26).

Host service factories now return an error, and loadPluginWithConfig fails
fast with the actual cause (e.g. 'creating Task service: creating plugin
data directory: ...'), which is also stored in the plugin's last_error.
Closers accumulated before a load failure are now closed (via a deferred
guard covering all failure paths), so partially-created services no longer
leak goroutines or database handles.

Also fix a panic in 'navidrome plugin enable': CLI commands use the plugin
Manager without calling Start, so manager.ctx was nil and
newTaskQueueService panicked in context.WithCancel. Long-lived host
services (taskqueue, kvstore, websocket) now receive their lifecycle
context explicitly in the constructor, sourced from serviceContext.baseCtx(),
which falls back to context.Background() for the unstarted-manager case.

* docs(plugins): correct websocket readLoop lifecycle comment

The comment claimed the read loop's context is always cancelled during
application shutdown, which is not true when the manager was never started
(one-shot CLI runs, where baseCtx falls back to context.Background()).
Clarify that connection closure via Close() on plugin unload is what ends
the read loop, with context cancellation as a server-shutdown backstop.

Addresses review feedback on #5756.

* test(plugins): exclude plugin-loading specs from Windows builds

The new loadPluginWithConfig specs reference test suite helpers
(testdataDir, noopMetricsRecorder) defined in plugins_suite_test.go, which
is excluded on Windows, breaking test compilation there. Move the specs to
their own file with the same build constraint, keeping the pure-function
specs in manager_loader_test.go running on Windows as before.
2026-07-10 19:15:49 -04:00

393 lines
11 KiB
Go

package plugins
import (
"context"
"errors"
"fmt"
"maps"
"net/http"
"net/url"
"strings"
"sync"
"time"
"github.com/gorilla/websocket"
"github.com/navidrome/navidrome/log"
"github.com/navidrome/navidrome/model/id"
"github.com/navidrome/navidrome/plugins/capabilities"
"github.com/navidrome/navidrome/plugins/host"
)
// CapabilityWebSocket indicates the plugin can receive WebSocket callbacks.
// Detected when the plugin exports any of the WebSocket callback functions.
const CapabilityWebSocket Capability = "WebSocket"
// webSocketCallbackTimeout is the maximum duration allowed for a WebSocket callback.
const webSocketCallbackTimeout = 30 * time.Second
// WebSocket callback function names
const (
FuncWebSocketOnTextMessage = "nd_websocket_on_text_message"
FuncWebSocketOnBinaryMessage = "nd_websocket_on_binary_message"
FuncWebSocketOnError = "nd_websocket_on_error"
FuncWebSocketOnClose = "nd_websocket_on_close"
)
func init() {
registerCapability(
CapabilityWebSocket,
FuncWebSocketOnTextMessage,
FuncWebSocketOnBinaryMessage,
FuncWebSocketOnError,
FuncWebSocketOnClose,
)
}
// wsConnection represents an active WebSocket connection.
type wsConnection struct {
conn *websocket.Conn
done chan struct{}
closeMu sync.Mutex
isClosed bool
}
// webSocketServiceImpl implements host.WebSocketService.
// It provides plugins with WebSocket communication capabilities.
type webSocketServiceImpl struct {
baseCtx context.Context // bounds the read loops, which outlive the Connect() call
pluginName string
manager *Manager
requiredHosts []string
mu sync.RWMutex
connections map[string]*wsConnection
}
// newWebSocketService creates a new WebSocketService for a plugin.
func newWebSocketService(ctx context.Context, pluginName string, manager *Manager, permission *WebSocketPermission) *webSocketServiceImpl {
return &webSocketServiceImpl{
baseCtx: ctx,
pluginName: pluginName,
manager: manager,
requiredHosts: permission.RequiredHosts,
connections: make(map[string]*wsConnection),
}
}
func (s *webSocketServiceImpl) Connect(ctx context.Context, urlStr string, headers map[string]string, connectionID string) (string, error) {
// Parse and validate URL
parsedURL, err := url.Parse(urlStr)
if err != nil {
return "", fmt.Errorf("invalid URL: %w", err)
}
// Validate scheme
if parsedURL.Scheme != "ws" && parsedURL.Scheme != "wss" {
return "", fmt.Errorf("invalid URL scheme: must be ws:// or wss://")
}
// Validate host against allowed hosts
if !s.isHostAllowed(parsedURL.Host) {
return "", fmt.Errorf("host %q is not allowed", parsedURL.Host)
}
// Generate connection ID if not provided
if connectionID == "" {
connectionID = id.NewRandom()
}
s.mu.Lock()
if _, exists := s.connections[connectionID]; exists {
s.mu.Unlock()
return "", fmt.Errorf("connection ID %q already exists", connectionID)
}
s.mu.Unlock()
// Create HTTP headers for handshake
httpHeaders := http.Header{}
for k, v := range headers {
httpHeaders.Set(k, v)
}
// Establish WebSocket connection
dialer := websocket.Dialer{
HandshakeTimeout: 30 * time.Second,
}
conn, resp, err := dialer.DialContext(ctx, urlStr, httpHeaders)
if resp != nil && resp.Body != nil {
_ = resp.Body.Close()
}
if err != nil {
return "", fmt.Errorf("failed to connect: %w", err)
}
wsConn := &wsConnection{
conn: conn,
done: make(chan struct{}),
}
s.mu.Lock()
s.connections[connectionID] = wsConn
s.mu.Unlock()
// Start read goroutine with the service's base context instead of the
// caller's ctx, because the readLoop must outlive the Connect() call.
// Connections are closed by Close() when the plugin is unloaded, which ends
// the readLoop; the base context is a backstop that also ends it on server
// shutdown (it is never cancelled in one-shot CLI runs).
go s.readLoop(s.baseCtx, connectionID, wsConn)
log.Debug(ctx, "WebSocket connected", "plugin", s.pluginName, "connectionID", connectionID, "url", urlStr)
return connectionID, nil
}
func (s *webSocketServiceImpl) SendText(ctx context.Context, connectionID, message string) error {
wsConn, err := s.getConnection(connectionID)
if err != nil {
return err
}
if err := wsConn.conn.WriteMessage(websocket.TextMessage, []byte(message)); err != nil {
return fmt.Errorf("failed to send text message: %w", err)
}
return nil
}
func (s *webSocketServiceImpl) SendBinary(ctx context.Context, connectionID string, data []byte) error {
wsConn, err := s.getConnection(connectionID)
if err != nil {
return err
}
if err := wsConn.conn.WriteMessage(websocket.BinaryMessage, data); err != nil {
return fmt.Errorf("failed to send binary message: %w", err)
}
return nil
}
func (s *webSocketServiceImpl) CloseConnection(ctx context.Context, connectionID string, code int32, reason string) error {
s.mu.Lock()
wsConn, exists := s.connections[connectionID]
if !exists {
s.mu.Unlock()
return fmt.Errorf("connection ID %q not found", connectionID)
}
delete(s.connections, connectionID)
s.mu.Unlock()
// Mark as closed to prevent callback
wsConn.closeMu.Lock()
wsConn.isClosed = true
wsConn.closeMu.Unlock()
// Send close message
closeMsg := websocket.FormatCloseMessage(int(code), reason)
_ = wsConn.conn.WriteControl(websocket.CloseMessage, closeMsg, time.Now().Add(5*time.Second))
_ = wsConn.conn.Close()
// Signal read goroutine to stop
close(wsConn.done)
// Invoke close callback
s.invokeOnClose(ctx, connectionID, code, reason)
log.Debug(ctx, "WebSocket connection closed", "plugin", s.pluginName, "connectionID", connectionID, "code", code)
return nil
}
// Close closes all connections for this plugin.
// This is called when the plugin is unloaded.
func (s *webSocketServiceImpl) Close() error {
s.mu.Lock()
connections := make(map[string]*wsConnection, len(s.connections))
maps.Copy(connections, s.connections)
s.connections = make(map[string]*wsConnection)
s.mu.Unlock()
ctx := context.Background()
for connID, wsConn := range connections {
wsConn.closeMu.Lock()
wsConn.isClosed = true
wsConn.closeMu.Unlock()
closeMsg := websocket.FormatCloseMessage(websocket.CloseGoingAway, "plugin unloaded")
err := wsConn.conn.WriteControl(websocket.CloseMessage, closeMsg, time.Now().Add(2*time.Second))
if err != nil {
log.Warn("Failed to send WebSocket close message on plugin unload", "plugin", s.pluginName, "connectionID", connID, "error", err)
}
err = wsConn.conn.Close()
if err != nil {
log.Warn("Failed to close WebSocket connection on plugin unload", "plugin", s.pluginName, "connectionID", connID, "error", err)
}
close(wsConn.done)
s.invokeOnClose(ctx, connID, websocket.CloseGoingAway, "plugin unloaded")
log.Debug("WebSocket connection closed on plugin unload", "plugin", s.pluginName, "connectionID", connID)
}
return nil
}
func (s *webSocketServiceImpl) getConnection(connectionID string) (*wsConnection, error) {
s.mu.RLock()
defer s.mu.RUnlock()
wsConn, exists := s.connections[connectionID]
if !exists {
return nil, fmt.Errorf("connection ID %q not found", connectionID)
}
return wsConn, nil
}
func (s *webSocketServiceImpl) isHostAllowed(host string) bool {
// Strip port from host if present
hostWithoutPort := host
if idx := strings.LastIndex(host, ":"); idx != -1 {
hostWithoutPort = host[:idx]
}
for _, pattern := range s.requiredHosts {
if matchHostPattern(pattern, hostWithoutPort) {
return true
}
}
return false
}
// matchHostPattern matches a host against a pattern.
// Supports "*" (allow all) and wildcards like "*.example.com".
func matchHostPattern(pattern, host string) bool {
if pattern == "*" {
return true
}
if pattern == host {
return true
}
// Handle wildcard patterns like *.example.com
if strings.HasPrefix(pattern, "*.") {
suffix := pattern[1:] // Get .example.com
return strings.HasSuffix(host, suffix)
}
return false
}
func (s *webSocketServiceImpl) readLoop(ctx context.Context, connectionID string, wsConn *wsConnection) {
defer func() {
// Remove connection if still present
s.mu.Lock()
delete(s.connections, connectionID)
s.mu.Unlock()
}()
for {
select {
case <-wsConn.done:
return
default:
}
messageType, data, err := wsConn.conn.ReadMessage()
if err != nil {
wsConn.closeMu.Lock()
isClosed := wsConn.isClosed
wsConn.closeMu.Unlock()
if isClosed {
return
}
// Check if it's a close error
if websocket.IsCloseError(err, websocket.CloseNormalClosure, websocket.CloseGoingAway, websocket.CloseNoStatusReceived) {
closeCode := websocket.CloseNoStatusReceived
closeReason := ""
if ce, ok := errors.AsType[*websocket.CloseError](err); ok {
closeCode = ce.Code
closeReason = ce.Text
}
s.invokeOnClose(ctx, connectionID, int32(closeCode), closeReason)
return
}
// Other read error
s.invokeOnError(ctx, connectionID, err.Error())
return
}
switch messageType {
case websocket.TextMessage:
s.invokeOnTextMessage(ctx, connectionID, string(data))
case websocket.BinaryMessage:
s.invokeOnBinaryMessage(ctx, connectionID, data)
}
}
}
// invokeWebSocketCallback is a generic helper that handles the common callback invocation pattern.
func invokeWebSocketCallback[I any](ctx context.Context, s *webSocketServiceImpl, funcName string, input I, callbackName string, connectionID string) {
instance := s.getPluginInstance()
if instance == nil {
return
}
callbackCtx, cancel := context.WithTimeout(ctx, webSocketCallbackTimeout)
defer cancel()
start := time.Now()
err := callPluginFunctionNoOutput(callbackCtx, instance, funcName, input)
if err != nil {
if !errors.Is(errFunctionNotFound, err) {
log.Error(ctx, "WebSocket "+callbackName+" callback failed", "plugin", s.pluginName, "connectionID", connectionID, "duration", time.Since(start), err)
}
}
}
func (s *webSocketServiceImpl) invokeOnTextMessage(ctx context.Context, connectionID, message string) {
invokeWebSocketCallback(ctx, s, FuncWebSocketOnTextMessage, capabilities.OnTextMessageRequest{
ConnectionID: connectionID,
Message: message,
}, "text message", connectionID)
}
func (s *webSocketServiceImpl) invokeOnBinaryMessage(ctx context.Context, connectionID string, data []byte) {
invokeWebSocketCallback(ctx, s, FuncWebSocketOnBinaryMessage, capabilities.OnBinaryMessageRequest{
ConnectionID: connectionID,
Data: data,
}, "binary message", connectionID)
}
func (s *webSocketServiceImpl) invokeOnError(ctx context.Context, connectionID, errorMsg string) {
invokeWebSocketCallback(ctx, s, FuncWebSocketOnError, capabilities.OnErrorRequest{
ConnectionID: connectionID,
Error: errorMsg,
}, "error", connectionID)
}
func (s *webSocketServiceImpl) invokeOnClose(ctx context.Context, connectionID string, code int32, reason string) {
invokeWebSocketCallback(ctx, s, FuncWebSocketOnClose, capabilities.OnCloseRequest{
ConnectionID: connectionID,
Code: code,
Reason: reason,
}, "close", connectionID)
}
func (s *webSocketServiceImpl) getPluginInstance() *plugin {
s.manager.mu.RLock()
instance, ok := s.manager.plugins[s.pluginName]
s.manager.mu.RUnlock()
if !ok {
log.Warn("Plugin not loaded for WebSocket callback", "plugin", s.pluginName)
return nil
}
return instance
}
// Verify interface implementation
var _ host.WebSocketService = (*webSocketServiceImpl)(nil)