Skip to content

Latest commit

 

History

History
296 lines (207 loc) · 13 KB

README.md

File metadata and controls

296 lines (207 loc) · 13 KB

Testing

Thanks for writing tests! Here's a quick run-down on our current setup.

Getting started

  1. Add a unit test to packages/*/src/TheUnitInQuestion/TheUnitInQuestion.test.js or an integration test packages/*/test/.
  2. Run pnpm t TheUnitInQuestion.
  3. Implement the tested behavior
  4. Open a PR once the test passes or if you want somebody to review your work

Tools we use

Writing tests

For all unit tests, please use the return value from @mui/internal-test-utils/createRenderer. It prepares the test suite and returns a function with the same interface as render from @testing-library/react.

describe('test suite', () => {
  const { render } = createRenderer();

  test('first', () => {
    render(<input />);
  });
});

For new tests please use expect from the BDD testing approach. Prefer to use as expressive matchers as possible. This keeps the tests readable, and, more importantly, the message if they fail as descriptive as possible.

In addition to the core matchers from chai we also use matchers from chai-dom.

Deciding where to put a test is (like naming things) a hard problem:

  • When in doubt, put the new test case directly in the unit test file for that component, for example packages/mui-material/src/Button/Button.test.js.
  • If your test requires multiple components from the library create a new integration test.
  • If you find yourself using a lot of data-testid attributes or you're accessing a lot of styles consider adding a component (that doesn't require any interaction) to test/regressions/tests/, for example test/regressions/tests/List/ListWithSomeStyleProp
  • If you have to dispatch and compose many different DOM events prefer end-to-end tests (Checkout the end-to-end testing readme for more information.)

Unexpected calls to console.error or console.warn

By default, our test suite fails if any test recorded console.error or console.warn calls that are unexpected.

The failure message includes the full test name (suite names + test name). This should help locating the test in case the top of the stack can't be read due to excessive error messages. The error includes the logged message as well as the stacktrace of that message.

You can explicitly expect no console calls for when you're adding a regression test. This makes the test more readable and properly fails the test in watchmode if the test had unexpected console calls.

Writing a test for console.error or console.warn

If you add a new warning via console.error or console.warn you should add tests that expect this message. For tests that expect a call you can use our custom toWarnDev or toErrorDev matchers. The expected messages must be a subset of the actual messages and match the casing. The order of these messages must match as well.

Example:

function SomeComponent({ variant }) {
  if (process.env.NODE_ENV !== 'production') {
    if (variant === 'unexpected') {
      console.error("That variant doesn't make sense.");
    }
    if (variant !== undefined) {
      console.error('`variant` is deprecated.');
    }
  }

  return <div />;
}
expect(() => {
  render(<SomeComponent variant="unexpected" />);
}).toErrorDev(["That variant doesn't make sense.", '`variant` is deprecated.']);
function SomeComponent({ variant }) {
  if (process.env.NODE_ENV !== 'production') {
    if (variant === 'unexpected') {
      console.error("That variant doesn't make sense.");
    }
    if (variant !== undefined) {
      console.error('`variant` is deprecated.');
    }
  }

  return <div />;
}
expect(() => {
  render(<SomeComponent />);
}).not.toErrorDev();

Commands

We uses a wide range of tests approach as each of them comes with a different trade-off, mainly completeness vs. speed.

React API level

Debugging tests

If you want to debug tests with the, for example Chrome inspector (chrome://inspect) you can run pnpm t <testFilePattern> --debug. Note that the test will not get executed until you start code execution in the inspector.

We have a dedicated task to use VSCode's integrated debugger to debug the currently opened test file. Open the test you want to run and press F5 (launch "Test Current File").

Run the core mocha unit/integration test suite

To run all of the unit and integration tests run pnpm test:unit

If you want to grep for certain tests add -g STRING_TO_GREP though for development we recommend pnpm t <testFilePattern>.

Watch the core mocha unit/integration test suite

pnpm t <testFilePattern>

First, we have the unit test suite. It uses mocha and a thin wrapper around @testing-library/react. Here is an example with the Dialog component.

Next, we have the integration tests. They are mostly used for components that act as composite widgets like Select or Menu. Here is an example with the Menu component.

Create HTML coverage reports

pnpm test:coverage:html

When running this command you should get under coverage/index.html a full coverage report in HTML format. This is created using Istanbul's HTML reporter and gives good data such as line, branch and function coverage.

DOM API level

Run the mocha test suite using the karma runner

pnpm test:karma

Testing the components at the React level isn't enough; we need to make sure they will behave as expected with a real DOM. To solve that problem we use Karma, which is almost a drop-in replacement of jsdom. Our tests run on different browsers to increase the coverage:

In development mode, if pnpm test:karma fails with this error "Cannot start ChromeHeadless. Can not find the binary", you can solve it by installing the missing headless browsers: pnpm playwright install --with-deps.

BrowserStack

We only use BrowserStack for non-PR commits to save resources. BrowserStack rarely reports actual issues so we only use it as a stop-gap for releases not merges.

To force a run of BrowserStack on a PR you have to run the pipeline with browserstack-force set to true. For example, you've opened a PR with the number 64209 and now after everything is green you want to make sure the change passes all browsers:

curl --request POST \
  --url https://circleci.com/api/v2/project/gh/mui/material-ui/pipeline \
  --header 'content-type: application/json' \
  --header 'Circle-Token: $CIRCLE_TOKEN' \
  --data-raw '{"branch":"pull/64209/head","parameters":{"browserstack-force":true}}'

Browser API level

In the end, components are going to be used in a real browser. The DOM is just one dimension of that environment, so we also need to take into account the rendering engine.

Visual regression tests

Check out the visual regression testing readme for more information.

end-to-end tests

Checkout the end-to-end testing readme for more information.

Development

When working on the visual regression tests you can run pnpm test:regressions:dev in the background to constantly rebuild the views used for visual regression testing. To actually take the screenshots you can then run pnpm test:regressions:run. You can pass the same arguments as you could to mocha. For example, pnpm test:regressions:run --watch --grep "docs-system-basic" to take new screenshots of every demo in docs/src/pages/system/basic. You can view the screenshots in test/regressions/screenshots/chrome.

Alternatively, you might want to open http://localhost:5001 (while pnpm test:regressions:dev is running) to view individual views separately.

For testing the components using Pigment CSS, you can run pnpm test:regressions-pigment-css:dev. You can then open http://localhost:5001/fixtures to vide individual views separately.

Caveats

Accessibility tree exclusion

Our tests also explicitly document which parts of the queried element are included in the accessibility (a11y) tree and which are excluded. This check is fairly expensive which is why it is disabled when tests are run locally by default. The rationale being that in almost all cases including or excluding elements from a query-set depending on their a11y-tree membership makes no difference.

The queries where this does make a difference explicitly include checking for a11y tree inclusion, for example getByRole('button', { hidden: false }) (see byRole documentation for more information). To see if your test (test:karma or test:unit) behaves the same between CI and local environment, set the environment variable CI to 'true'.

Not considering a11y tree exclusion is a common cause of "Unable to find an accessible element with the role" or "Found multiple elements with the role".

Performance monitoring

We have a dedicated CI task that profiles our core test suite. Since this task is fairly expensive and not relevant to most day-to-day work it has to be started manually. The CircleCI docs explain how to start a pipeline manually in detail. Example: With an environment variable $CIRCLE_TOKEN containing a CircleCI personal access token.

The following command triggers the profile workflow for the pull request #24289.

curl --request POST \
  --url https://circleci.com/api/v2/project/gh/mui/material-ui/pipeline \
  --header 'content-type: application/json' \
  --header 'Circle-Token: $CIRCLE_TOKEN' \
  --data-raw '{"branch":"pull/24289/head","parameters":{"workflow":"profile"}}'

To analyze this profile run you can use https://mui-dashboard.netlify.app/test-profile/:job-number.

To find out the job number you can start with the response of the previous CircleCI API request which includes the created pipeline id. You then have to search in the CircleCI UI for the job number of test_profile that is part of the started pipeline. The job number can be extracted from the URL of a particular CircleCI job.

For example, in https://app.circleci.com/pipelines/github/mui/material-ui/32796/workflows/23f946de-328e-49b7-9c94-bfe0a0248a12/jobs/211258 jobs/211258 points to the job number which is in this case 211258 which means you want to visit https://mui-dashboard.netlify.app/test-profile/211258 to analyze the profile.

Testing multiple versions of React

You can check integration of different versions of React (for example different release channels or PRs to React) by running:

pnpm use-react-version <version>

Possible values for version:

  • default: stable (minimum supported React version)
  • a tag on npm, for example next, experimental or latest
  • an older version, for example ^17.0.0

CI

Circle CI web interface

There are two workflows that can be triggered for any given PR manually in the CircleCI web interface:

  • react-next
  • react-17

Follow these steps:

  1. Go to https://app.circleci.com/pipelines/github/mui/material-ui?branch=pull/PR_NUMBER/head and replace PR_NUMBER with the PR number you want to test.

  2. Click Trigger Pipeline button.

  3. Expand Add parameters (optional) and add the following parameter:

    Parameter type Name Value
    string workflow react-next or react-17
  4. Click Trigger Pipeline button.

API request

You can pass the same version to our CircleCI pipeline as well:

With the following API request we're triggering a run of the default workflow in PR #24289 for react@next

curl --request POST \
  --url https://circleci.com/api/v2/project/gh/mui/material-ui/pipeline \
  --header 'content-type: application/json' \
  --header 'Circle-Token: $CIRCLE_TOKEN' \
  --data-raw '{"branch":"pull/24289/head","parameters":{"react-version":"next"}}'