Skip to main content

no-atomics-mutation

Disallow mutating shared memory via Atomics write operations.

Targeted pattern scope​

This rule targets mutating static calls on Atomics that write into shared typed array memory.

What this rule reports​

  • Atomics.store(...)
  • Atomics.add(...), Atomics.sub(...)
  • Atomics.and(...), Atomics.or(...), Atomics.xor(...)
  • Atomics.exchange(...), Atomics.compareExchange(...)

Why this rule exists​

Atomics write operations mutate shared memory directly, introducing global mutable state across workers/threads.

This rule helps enforce immutable concurrency design by preventing in-place shared-memory writes at lint time.

❌ Incorrect​

const view = new Int32Array(new SharedArrayBuffer(8));
Atomics.store(view, 0, 1);

✅ Correct​

const view = new Int32Array(new SharedArrayBuffer(8));
Atomics.load(view, 0);

Additional examples​

// ❌ Mutates shared memory in place
const view = new Int32Array(new SharedArrayBuffer(8));
Atomics.compareExchange(view, 0, 1, 2);

// ✅ Read-only synchronization primitive
const view = new Int32Array(new SharedArrayBuffer(8));
Atomics.wait(view, 0, 0);

ESLint flat config example​

import immutable from "eslint-plugin-immutable-2";

export default [
    {
        files: ["**/*.{js,mjs,cjs,ts,mts,cts,tsx}"],
        plugins: { immutable },
        rules: {
            "immutable/no-atomics-mutation": "error",
        },
    },
];

When not to use it​

If your system intentionally uses shared-memory mutation via Atomics for low-level synchronization internals, this rule may be too strict for those specialized modules.

Rule catalog ID: R923

Further reading​