Spanda

Lesson 7 — Tasks, triggers, and concurrency

Goal: Run periodic work, react to events, and coordinate background tasks.

Examples:

Deep dives: triggers.md, concurrency.md

Tasks — periodic work with priorities

task sense every 50ms {
  let scan = lidar.read();
  if scan.nearest_distance < 0.5 m {
    emit ObstacleSeen;
  }
}

Tasks run on a scheduler with optional priority (critical, high, low) and resource budgets. Use them for control loops separate from one-shot behaviors.

Event triggers

event ObstacleSeen;

on ObstacleSeen {
  wheels.stop();
}

Emit events from tasks or behaviors:

emit ObstacleSeen;

Handlers run when the event fires. Trace them at runtime:

spanda run examples/integration/triggers_minimal.sd --trace-triggers

Concurrency primitives

For background work and message passing:

module comm.ping;

export fn tick() -> Int {
  return 1;
}

robot ConcurrentRover {
  behavior run() {
    let ch = channel();
    send(ch, 42);

    select {
      recv(ch) => wheels.stop();
    };

    spawn tick();
  }
}
Builtin Purpose
channel() Create a message channel
send / recv Non-blocking message pass
select Wait on first ready channel
spawn fn() Queue background function call
parallel { } Cooperative concurrent blocks

Full demo: examples/concurrency.sd

Try it

spanda check examples/integration/triggers_minimal.sd
spanda run examples/integration/triggers_minimal.sd --trace-triggers

spanda check examples/integration/concurrency_minimal.sd
spanda run examples/integration/concurrency_minimal.sd --trace-scheduler

spanda run examples/concurrency.sd --trace-scheduler --trace-tasks

Exercise

  1. Add an event LowBattery to your robot
  2. Add on LowBattery { wheels.stop(); }
  3. In a task, emit LowBattery when a stub condition is true (e.g. always once after 5 loop iterations in simulation)

Next: Lesson 8 — Hardware profiles and verify

This site is open source. Improve this page.