mcpField
S
signals-preact-dart

signals-preact-dart

github.com/rodydavisDevelopment82475 installsUpdated 12w ago
Repository

About signals-preact-dart

A guide to reactive programming with the `preact_signals` package in Dart, for developers using signals to manage application state. It explains core primitives, dependency tracking, batching, and cycle-avoidance rules.

Key features

  1. 01Mutable and read-only signals
  2. 02Lazy, memoized computed values
  3. 03Effects with dynamic dependency tracking
  4. 04Batching multiple state writes
  5. 05Guidance on dependency cycles and re-entrancy

Use cases

  1. 01Modeling reactive application state
  2. 02Deriving values from multiple signals
  3. 03Running side effects when tracked state changes
  4. 04Grouping state updates into one batch

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/rodydavis/signals.dart --skill signals-preact-dart

After installing, type/signals-preact-dartin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

CONTEXT.md
core/
batch.md
computed.md
CONTEXT.md
effect.md
readonly.md
signal.md

Details

From the project README

Preact Signals in Dart

This skill covers the core reactive programming primitives and implementation patterns of the preact_signals package.

Related Documentation & Items

Related FileDescription
signal.mdStandard writeable reactive state primitive containing value accessors, mutation mechanics, and basic custom options.
computed.mdLazy, memoized read-only derived state signal tracking reactive dependencies dynamically.
effect.mdSynchronous observer managing active subscription loops, dynamic dependency updates, and lifecycle teardown.
readonly.mdRead-only signal views ensuring unidirectional data access flows.
batch.mdTransactional 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 isInitialized or check the initialization status before forcing a .peek() inside observers to avoid re-entrancy cycle crashes.

Comments

Markdown supported