Lifecycle

Spry is easier to reason about when you know the request order.

Request flow

For an incoming request, Spry works through these layers:

1. check public/ for matching GET or HEAD assets
2. match the route handler from routes/
3. create the request-scoped Event
4. run global and scoped middleware
5. run the matched handler or fallback
6. if something throws, run scoped _error.dart handlers

This is the part people actually care about during development: where behavior can intercept a request.

For websocket routes, the handshake still follows this flow. After the runtime commits the websocket upgrade, the session leaves the normal HTTP middleware and error pipeline. That means _error.dart can still shape handshake failures, but not post-upgrade websocket session exceptions. See WebSockets.

Lifecycle hooks

Use hooks.dart for process-level lifecycle work:

import 'package:spry/spry.dart';

Future<void> onStart(ServerLifecycleContext context) async {
  print('Spry dart_vm example started: ${context.runtime.name}');
}

This is a good place for:

• startup logging
• runtime warm-up
• teardown work

Mental model

Middleware shapes request flow. Hooks shape process lifecycle. Route handlers shape the response.