Understanding mock.restoreAll() and mock.reset() in Node.js Test Runner

[iatsiuk]

I wont to ensure that the call count is correctly reset between test cases when using the Node.js Test Runner.

Here is my code:

const assert = require("node:assert/strict");
const { describe, afterEach, it, mock } = require("node:test");

describe("", async () => {
  let testFn = mock.fn();

  afterEach(() => {
    mock.reset();
    // or should we use mock.restoreAll() instead?
  });

  it("", () => {
    testFn();
  });

  it("", () => {
    testFn();

    assert.strictEqual(testFn.mock.calls.length, 1);
  });
});

Using mock.reset() should theoretically reset the call count for the testFn mock between test cases. However, it seems that the Node.js Test Runner's mock implementation may have a bug or behave differently than expected in this case:

    AssertionError [ERR_ASSERTION]: Expected values to be strictly equal:

    2 !== 1

Here is another example, that using sinon.js fake():

const assert = require("node:assert/strict");
const { describe, afterEach, it } = require("node:test");
const sinon = require("sinon");

describe("", async () => {
  let testFn = sinon.fake();

  afterEach(() => {
    sinon.reset();
  });

  it("", () => {
    testFn();
  });

  it("", () => {
    testFn();

    assert.strictEqual(testFn.callCount, 1);
  });
});

In the second code example, using sinon.reset() from the Sinon library correctly resets the call count for the testFn fake between test cases.

How to correctly reset the number of mock calls using the Node.js test runner. What should mock.restoreAll() and mock.reset() be used for?

[cjihrig]

You probably want testFn.mock.resetCalls(); instead of mock.reset();

[iatsiuk]

@cjihrig Thank you for response!

I understand that it is possible to reset the state of each mock individually or even re-assign the variable that holds it. My question was related to the cases where it is appropriate to use mock.restoreAll() and mock.reset() from the global scope. I expect that calling these methods SHOULD reset the call count of the mock, otherwise, what other purpose do they serve?

[cjihrig]

testFn.mock is an instance of MockFunctionContext. Each individual mocked function has one of these, and this is the only way to get information about that particular mock.

The top level mock, and the per test context t.mock, are instances of MockTracker. This is used to create and restore mocks. This manages mock functions, methods, modules, etc. Restoring a mock does not change the behavior of the MockFunctionContext instances. Instead, it makes sure that mocks are restored to their original behavior. That is why if you create mocks with t.mock, the test runner is able to unmock things after the test ends.