name: signals-preact-dart description: Core reactive programming best practices and primitive definitions for preact_signals in Dart.
Preact Signals in Dart
This skill covers the core reactive programming primitives and implementation patterns of the preact_signals package.
Related Documentation & Items
| Related File | Description |
|---|---|
| signal.md | Standard writeable reactive state primitive containing value accessors, mutation mechanics, and basic custom options. |
| computed.md | Lazy, memoized read-only derived state signal tracking reactive dependencies dynamically. |
| effect.md | Synchronous observer managing active subscription loops, dynamic dependency updates, and lifecycle teardown. |
| readonly.md | Read-only signal views ensuring unidirectional data access flows. |
| batch.md | Transactional state mutation blocks optimizing reactive computations and preventing rendering/recompute churn. |
Core Primitives Quick Start Guide
The signals library exposes five core functions which are the building blocks to model any reactive business logic.
1. signal(initialValue)
Creates a new mutable signal container. You read a signal's value or subscribe to updates by accessing .value.
final counter = signal(0);
print(counter.value); // 0
counter.value = 1; // Mutates value and schedules dependent updates
.peek()
Reads a signal's current value without subscribing to its mutations.
final counter = signal(0);
final logCount = signal(0);
effect(() {
print(counter.value);
// Read using peek to avoid subscribing/triggering loop
logCount.value = logCount.peek() + 1;
});
2. untracked(fn)
Executes a callback that reads signals without subscribing to any of them.
final counter = signal(0);
final count = signal(0);
effect(() {
print(counter.value);
count.value = untracked(() => count.value + 1);
});
3. computed(fn)
Combines the values of multiple signals into a lazy, memoized derived signal.
final first = signal('John');
final last = signal('Doe');
final fullName = computed(() => '${first.value} ${last.value}');
print(fullName.value); // John Doe
4. effect(fn)
Orchestrates immediate synchronous side effects by running a callback and subscribing to any signals read within it.
final name = signal('Jane');
final dispose = effect(() => print('Hello $name'));
dispose(); // Clean up subscription
5. batch(fn)
Groups multiple signal writes into a single transaction, executing all dependent computed evaluations and effects exactly once at the end.
final a = signal(0);
final b = signal(0);
batch(() {
a.value = 1;
b.value = 2;
});
1. Core Primitives Details
A. Signal<T>
A state primitive representing a mutable value that can be observed.
final count = signal(0);
count.value = 1; // Notifies all active subscribers/effects
B. Computed<T>
A derived state primitive representing a read-only value calculated based on other signals. It is:
- Lazy: Evaluated only when accessed.
- Memoized: Caches its value and only recompute when its source dependencies change.
final count = signal(2);
final double = computed(() => count.value * 2);
C. Effect
An observer that executes a callback immediately and tracks read signals. It re-runs whenever any of the tracked signals change.
final dispose = effect(() {
print('Current count: ${count.value}');
});
dispose(); // Clean up subscription
2. Dependency Tracking & Re-entrancy Rules
- Synchronous Context: Dependencies are tracked dynamically based on which signals are read during the execution of the callback.
- Cycle Detection: Reading a computed signal inside its own execution flow or setting dependencies inside computed callbacks causes cycles. Avoid side-effects inside computed getters.
- Lazy Initial Value: Lazy computed signals are not evaluated when created. Always call
isInitializedor check the initialization status before forcing a.peek()inside observers to avoid re-entrancy cycle crashes.