Plugin Controller Lifecycle

Source file: packages/core/src/journey-machine/plugin-controller.ts

This file is the adapter between the core machine and optional plugins.

Its job is deliberately narrow: run plugin.setup(...) once, keep the resulting hooks organized, and call those hooks at the right points in the machine lifecycle.

How It Works

1. During setup, the controller calls every plugin's setup(...) function with the same setup context. If a plugin throws, the file rolls back already-initialized plugins by calling their dispose() hooks in reverse order, then wraps the original setup error with the plugin name so failures are attributable.
2. hydrateSnapshot(...) reduces over the plugin hook list, letting each plugin transform the starting snapshot before the runtime owns it.
3. onSnapshotChange(...) fans out every committed snapshot change, including the previous snapshot and the machine-level reason for the write.
4. extendMachine(...) lets plugins add extra methods or properties to the public machine. The controller refuses collisions so a plugin cannot silently override existing machine API.
5. dispose() forwards to every plugin hook so storage listeners, timers, or other resources can clean up.

This is how Journey keeps the base runtime small while still supporting features such as persistence and execution path analysis.

Recommended Reading

• Read createJourneyMachine to see when plugin setup runs.
• Read Plugins Overview for the public extension model.
• Read Persistence and Execution Paths Plugin for concrete examples of these hooks in use.