Chrome fakes

The fakes are installed by the Vitest preset or by calling installChromeFakes() directly. Each namespace is wrapped in a Proxy that throws on access to unmodeled members.

`runtime`

Import: import { createRuntimeFake } from 'extforge/testing';

What it models

Member	Type	Notes
`chrome.runtime.id`	`string`	Fixed to `'extforge-test-extension-id'`
`chrome.runtime.onInstalled`	event	`.addListener` / `.removeListener`
`chrome.runtime.onStartup`	event	`.addListener` / `.removeListener`
`chrome.runtime.onMessage`	event	`.addListener` / `.removeListener`
`chrome.runtime.sendMessage`	spy	Resolves `undefined` by default
`chrome.runtime.reload`	spy	No-op

What it does not model

getURL, connect, onConnect, getManifest, openOptionsPage, and all other runtime.* members throw the not-modeled trap error.

Test-side controls

Control	Description
`fakes.runtime.fireOnInstalled(details?)`	Calls all `onInstalled` listeners. Default `details`: `{ reason: 'install' }`.
`fakes.runtime.fireOnStartup()`	Calls all `onStartup` listeners.
`fakes.runtime.fireOnMessage(message, sender?)`	Calls all `onMessage` listeners and returns a Promise that resolves to the `sendResponse` argument. If no listener calls `sendResponse`, resolves `undefined`.
`fakes.runtime.reset()`	Clears all listeners and resets spy call counts.

Example

import { fakes } from 'extforge/testing/vitest';
import { expect, it } from 'vitest';

it('fires onInstalled with reason update', () => {
  let receivedReason = '';
  chrome.runtime.onInstalled.addListener(({ reason }) => {
    receivedReason = reason;
  });
  fakes.runtime.fireOnInstalled({ reason: 'update' });
  expect(receivedReason).toBe('update');
});

`storage`

Import: import { createStorageFake } from 'extforge/testing';

What it models

Three storage areas: chrome.storage.local, chrome.storage.sync, and chrome.storage.session. Each area models:

Method	Behavior
`get(keys?)`	Returns matching keys from in-memory state. Pass `null` or omit for all keys.
`set(items)`	Merges `items` into in-memory state.
`remove(keys)`	Deletes keys from state.
`clear()`	Empties the state object.

All methods are spies with .calls and .reset().

What it does not model

chrome.storage.managed, chrome.storage.onChanged, and any method other than the four above.

Test-side controls

Control	Description
`fakes.storage.chrome.local.__state()`	Returns a snapshot of the current local storage state. Same for `sync` and `session`.
`fakes.storage.reset()`	Clears state and resets call counts for all three areas.

Example

import { fakes } from 'extforge/testing/vitest';
import { expect, it } from 'vitest';

it('persists storage across calls within a test', async () => {
  await chrome.storage.local.set({ theme: 'dark' });
  await chrome.storage.local.set({ font: 'mono' });
  const all = fakes.storage.chrome.local.__state();
  expect(all).toEqual({ theme: 'dark', font: 'mono' });
});

`tabs`

Import: import { createTabsFake } from 'extforge/testing';

What it models

Method	Behavior
`chrome.tabs.query(info)`	Filters the seeded tab list by `active` and `url`.
`chrome.tabs.sendMessage(tabId, message)`	Spy; resolves `undefined`.
`chrome.tabs.create(props)`	Adds a new tab to the internal list; returns the tab record.
`chrome.tabs.reload(tabId)`	Spy; no-op.

What it does not model

get, update, remove, onActivated, onUpdated, onRemoved, and all other tabs.* members.

Test-side controls

Control	Description
`fakes.tabs.__seed(tabs)`	Replaces the tab list. Each entry: `{ id, url, active? }`. Repeated calls overwrite.
`fakes.tabs.reset()`	Clears the tab list and resets all spy call counts.

Example

import { fakes } from 'extforge/testing/vitest';
import { expect, it, beforeEach } from 'vitest';

beforeEach(() => {
  fakes.tabs.__seed([
    { id: 100, url: 'https://example.com/', active: true },
    { id: 101, url: 'https://other.com/', active: false },
  ]);
});

it('finds the active tab', async () => {
  const [tab] = await chrome.tabs.query({ active: true });
  expect(tab!.id).toBe(100);
});

`action`

Import: import { createActionFake } from 'extforge/testing';

What it models

Method	Behavior
`chrome.action.setBadgeText({ text, tabId? })`	Stores `text` keyed by `tabId` (or `'global'`).
`chrome.action.getBadgeText({ tabId? })`	Returns stored text for that key.
`chrome.action.setIcon(details)`	Spy; no-op.
`chrome.action.enable(tabId?)`	Spy; no-op.
`chrome.action.disable(tabId?)`	Spy; no-op.

What it does not model

setPopup, getPopup, setBadgeBackgroundColor, openPopup, and all other action.* members.

Test-side controls

Control	Description
`fakes.action.reset()`	Clears badge state and resets spy call counts.

Example

import { fakes } from 'extforge/testing/vitest';
import { expect, it } from 'vitest';

it('sets and reads badge text', async () => {
  await chrome.action.setBadgeText({ text: '5' });
  const text = await chrome.action.getBadgeText({});
  expect(text).toBe('5');
});

`scripting`

Import: import { createScriptingFake } from 'extforge/testing';

What it models

Method	Behavior
`chrome.scripting.executeScript(injection)`	Returns `[{ result: undefined, frameId: 0 }]` by default. Use `__nextResult` to queue a specific return value.

What it does not model

insertCSS, removeCSS, registerContentScripts, getRegisteredContentScripts, and all other scripting.* members.

Test-side controls

Control	Description
`fakes.scripting.__nextResult(value)`	Queues `value` as the `result` field of the next `executeScript` call. Multiple calls queue up in order.
`fakes.scripting.reset()`	Clears the queue and resets spy call counts.

Example

import { fakes } from 'extforge/testing/vitest';
import { expect, it } from 'vitest';

it('returns the queued executeScript result', async () => {
  fakes.scripting.__nextResult(42);
  const [frame] = await chrome.scripting.executeScript({
    target: { tabId: 1 },
    func: () => 42,
  });
  expect(frame!.result).toBe(42);
});

The not-modeled trap

Accessing any member not listed above throws:

chrome.<ns>.<method> is not modeled by extforge/testing v1;
supply your own mock or extend the fake.
Docs: https://extforge.arshadshah.com/testing#unmodeled

To add a stub for an unmodeled method in a test file:

// Extend globally for the duration of the test file
(globalThis as any).chrome.history = {
  search: vi.fn().mockResolvedValue([]),
};

Or, use the per-namespace factory and spread your method before passing the fake to your code:

import { createTabsFake } from 'extforge/testing';

const tabs = createTabsFake();
const extendedChrome = {
  ...tabs.chrome,
  get: vi.fn().mockResolvedValue({ id: 1, url: 'https://example.com' }),
};

Exports

extforge/testing exports the installer plus a factory and a type per namespace.

Export	Kind	Notes
`installChromeFakes()`	function	Install a fresh `ChromeFakes` on `globalThis.chrome`. Returns the handle.
`resetChromeFakes()`	function	Reset state on the currently installed fakes (per-test cleanup).
`createChromeFakes()`	function	Build a `ChromeFakes` handle without touching globals.
`createRuntimeFake()` · `createStorageFake()` · `createTabsFake()` · `createActionFake()` · `createScriptingFake()`	functions	Per-namespace factories (see each section above).
`ChromeFakes`	type	The aggregate handle returned by `installChromeFakes` / `createChromeFakes`.
`RuntimeFake` · `StorageFake` · `StorageAreaFake` · `TabsFake` · `TabRecord` · `ActionFake` · `ScriptingFake` · `ExecuteScriptInjection`	types	Handles and record shapes for the individual namespace fakes.