Jest.js 假计时器不适用于最新版本的用户事件?

hl0ma9xz  于 6个月前  发布在  Jest
关注(0)|答案(3)|浏览(98)

下面是一个我想测试的自定义钩子:

import { useEffect, useState } from "react";

export const useAlert = () => {
  const [alert, setAlert] = useState(null);

  useEffect(() => {
    let timerId = setTimeout(() => {
      console.log("Timeout, removing alert from DOM");
      setAlert(null);
    }, 200);

    return () => clearTimeout(timerId);
  }, [alert]);

  return {
    renderAlert: alert ? alert : null,
    setAlert,
  };
};

字符串
它只允许组件设置警报,并在300毫秒后自动清除警报。
这是对上述挂钩的工作测试

import { render, screen, waitFor } from "@testing-library/react";
import userEvent from "@testing-library/user-event";
import { useAlert } from "../components/useAlert";

// jest.useFakeTimers();

function FakeComponent() {
  const { renderAlert, setAlert } = useAlert();

  return (
    <div>
      <button onClick={() => setAlert("fake alert")}>Set Alert</button>
      <p>{renderAlert}</p>
    </div>
  );
}

test("testing", async () => {
  const user = userEvent.setup();
  render(<FakeComponent />);
  const button = screen.getByRole("button", { name: /set alert/i });

  await user.click(button);
  expect(screen.getByText(/fake alert/i)).toBeInTheDocument();

  await waitFor(() => {
    expect(screen.queryByText(/fake alert/i)).not.toBeInTheDocument();
  });
});


我的疑问是,我想在测试中使用jest的假计时器,但如果我取消注解jest.useFakeTimers()行,测试就会中断,说测试超时,因为默认的超时值是5000 ms。
我不明白为什么会这样,请帮帮忙!

eivgtgni

eivgtgni1#

要在@testing-library/user-event的最新版本中使用假计时器,您需要在设置userEvent时配置advanceTimers选项(文档):

const user = userEvent.setup({
  advanceTimers: () => jest.runOnlyPendingTimers(),
});

字符串
以下是您的更新测试:

import { render, screen, waitFor } from "@testing-library/react";
import userEvent from "@testing-library/user-event";
import { useAlert } from "../components/useAlert";

jest.useFakeTimers();

function FakeComponent() {
  const { renderAlert, setAlert } = useAlert();

  return (
    <div>
      <button onClick={() => setAlert("fake alert")}>Set Alert</button>
      <p>{renderAlert}</p>
    </div>
  );
}

test("testing", async () => {
  const user = userEvent.setup({
    advanceTimers: () => jest.runOnlyPendingTimers(),
  });
  render(<FakeComponent />);
  const button = screen.getByRole("button", { name: /set alert/i });

  await user.click(button);
  expect(screen.getByText(/fake alert/i)).toBeInTheDocument();

  act(() => jest.runAllTimers());
  expect(screen.queryByText(/fake alert/i)).toBeNull();
});

hec6srdp

hec6srdp2#

可以将delay属性作为null传递给userEvent.setup调用

test("testing", async () => {
  const user = userEvent.setup({ delay: null });

  render(<FakeComponent />);
  const button = screen.getByRole("button", { name: /set alert/i });

  await user.click(button);
});

字符串

oyt4ldly

oyt4ldly3#

@som-shekhar-mukherjee的解决方案是好的,但不严格,尽管它设法为这个非常特殊的情况下工作。
但是,我建议在userEventadvanceTimers配置中使用advanceTimersToNextTimer而不是runOnlyPendingTimers
这样做可以确保jest只运行userEvent的内部方法设置的计时器。如果userEvent触发了代码中使用计时器的内容,runOnlyPendingTimers将耗尽这些内容。advanceTimersToNextTimer只会提前userEvent所导致的计时器。
基本上,您的setTimeout回调需要在测试中运行两次。

  • 一次:挂载元件后至少200毫秒(因为挂载元件时会呼叫useEffect)。
  • 两次:设置"fake alert"后至少200 ms。

@som-shekhar-mukherjee提出的解决方案之所以有效,是因为一旦测试到达并执行await user.click(button);行,runOnlyPendingTimers将耗尽所有挂起的计时器:

  • 安装测试组件时useEffect设置的定时器;
  • user.click调用设置的计时器。

它恰好适用于您的情况,但却阻止您控制正在测试的计时器何时发生。(你可能不知道)提前了你的测试组件的计时器。这是不正确的,因为,在真实的的执行中,它不会只在单击按钮时被耗尽。这个
所以我建议如下:

beforeEach(() => {
  jest.useFakeTimers();
});

afterEach(() => {
  jest.runOnlyPendingTimers();
  jest.useRealTimers();
});

test("testing", async () => {
  const user = userEvent.setup({
    advanceTimers: () => jest.advanceTimersToNextTimer(),
  });

  render(<FakeComponent />);

  // Basically allows the callback scheduled via the component mount to run.
  // You could control more granularly what timers you want to advance here.
  jest.runOnlyPendingTimers();

  const button = screen.getByRole("button", { name: /set alert/i });

  // `advanceTimersToNextTimer` will prevent this from blocking indefinitely
  // and that's what, primarily, solves your question.
  await user.click(button);

  expect(screen.getByText(/fake alert/i)).toBeInTheDocument();

  // Allows the callback scheduled via the set of "fake alert" to run.
  // Again, you could control more granularly what timers you want to advance here.
  act(() => jest.runOnlyPendingTimers());

  expect(screen.queryByText(/fake alert/i)).not.toBeInTheDocument();
});

字符串
总而言之,在userEvent配置上使用advanceTimersToNextTimer解决了阻塞问题,并允许您在测试期间更正确地控制所测试组件的计时器发生的情况。测试挂钩是由React Testing Library docs建议的。顺便说一句,我使用您的组件和上面的挂钩仔细测试了我提出的解决方案。如果您对我的答案有疑问,请告诉我。

相关问题