What you'll learn

• How directives separate state transitions from side effects
• How to schedule delayed or recurring signals cleanly
• How to avoid runaway loops in time-based workflows
• How to reason about clock, retries, and cancellation

Prerequisites

• You understand action return tuples and signal routing
• You can reason about eventually delivered messages
• You are comfortable debugging asynchronous behavior

Lesson Breakdown

1. Directive fundamentals: emit, schedule, and compose side effect instructions.
2. Scheduling model: one-shot schedules and self-rescheduling loops.
3. Loop controls: state flags and termination conditions.
4. Failure handling: retry windows, jitter, and dead-letter strategies.
5. Testing time: assert directives rather than sleeping in tests.

Hands-on Exercise

Extend a demand tracker with auto-decay:

1. Toggle auto-decay with listing.demand.auto_decay.toggle.
2. Emit listing.demand.changed after each decay.
3. Schedule a future listing.demand.tick while auto mode is enabled.
4. Disable auto mode and confirm no future tick is scheduled.
5. Add tests asserting emitted directives and schedule payloads.

Validation Checklist

• [ ] Actions return directives instead of executing side effects directly.
• [ ] Recurring behavior is state-controlled, not unconditional.
• [ ] Tests assert directive content and timing intent.
• [ ] Loop stops correctly when guard state flips.

Next Module

Continue with LiveView + Jido Integration Patterns.