We welcome all types of contribution.
Need a feature or found a bug? Please create an issue.
Have a question or suggestion? Please create a discussion.
Want to contribute by creating a pull request? Great! Fork iTwinUI-react to get started.
To clone and build iTwinUI-react, you'll need Git, Node 16, and Yarn 1 installed on your computer.
- Create a local clone of your forked repository. You can do this from the command line or using the Github Desktop app.
- Go to the directory where you cloned iTwinUI-react. e.g.
cd iTwinUI-react
. - Run
yarn install
from that directory.
If using vscode, our prettier and editor configs will be used. Please make sure to install all recommended extensions in extensions.json.
We use Turborepo as our monorepo tool to improve the experience of yarn workspaces. It allows running commands in parallel and caches build outputs.
The root package.json includes a few commands that use turbo run
to run the corresponding command in all workspaces.
e.g. to build all workspaces together, run the following command from the root:
yarn build
To start the development server for all workspaces, run the following command from the root.
yarn dev
This will automatically build anything that's not already built, and run the dev
script for every workspace in parallel, watching for file changes.
By default, it will open the storybook and docs website in your browser. Any examples and playgrounds will need to be manually opened. The urls/ports are as follows (assuming they are available):
- docs website:
http://localhost:1700
- vite playground:
http://localhost:1701
- next playground:
http://localhost:1702
- astro playground:
http://localhost:1703
- storybook:
http://localhost:6006
(storybook default)
If you only need to run this task for a specific workspace, you can specify turborepo's --filter
argument. For example, if you only want to start storybook, you could run yarn dev --filter=storybook
.
If a script is not available in the root package.json or if you need to pass workspace-specific cli args, then you can specify the workspace as follows:
# passing Alert as a cli arg to the `test` command in itwinui-react
yarn workspace @itwin/itwinui-react test Alert
...or you can simply run the command normally from inside the workspace folder instead of the monorepo root.
# run this from inside packages/iTwinUI-react/ for the same result
yarn test Alert
Note that this bypasses the turborepo pipeline, so you will need to manually run any dependent tasks first. For example, if the build
command of storybook
relies on the build
command of @itwin/itwinui-react
, then you will need to manually run the build
commands in the right order.
yarn workspace @itwin/itwinui-react build
yarn workspace storybook build
This is why it's recommended to use the turbo --filter
syntax whenever possible.
yarn build --filter=storybook
Before developing, please read our style guide.
If you are creating a new component, use this script:
yarn createComponent
It ensures all needed imports are added and files are created.
Note: Every component needs to use
useTheme()
hook to ensure styling (theming) works fine. ThecreateComponent
script mentioned above adds it automatically.
For a component named Alert
, the createComponent
script will add/modify the following files:
Directory structure
packages/iTwinUI-react
|
+ - src
| |
| + - core
| |
| + - Alert
| | + - > Alert.test.tsx
| | + - > Alert.tsx
| | + - > index.ts
| |
| + - > index.ts
|
apps/storybook
| |
| + - src
| |
| + - > Alert.stories.tsx
|. + - > Alert.test.ts
src/core/Alert/Alert.tsx
- Implements the React component and exports it, both named and by default.
src/core/Alert/Alert.test.tsx
- Contains all test cases for the component.
src/core/Alert/index.ts
export { Alert } from './Alert';
export type { AlertProps } from './Alert';
export default './Alert';
src/core/index.ts
---
export { Alert } from './Alert';
export type { AlertProps } from './Alert';
---
Note: The script will add the exports to index.ts but you will need to manually sort them alphabetically.
apps/storybook/src/Alert.stories.tsx
- Contains common demo states and examples for the component.
apps/storybook/src/Alert.test.tsx
- Contains the Cypress test for the component.
- Manually add the
<svg>
component toutils/icons
andutils/icons/index.ts
- Import it from
utils
e.g.
import { SvgClose, SvgInfoCircular } from '../utils';
- Import it from
@itwin/itwinui-icons-react
e.g.
import { SvgChevronRightDouble, SvgFolder } from '@itwin/itwinui-icons-react';
We use JSDoc (not TSDoc) to write documentation for our code.
Every component should have a multiline description and at least one example.
/**
* A small box to quickly grab user attention and communicate a brief message.
* @example
* <Alert>This is a basic alert.</Alert>
*/
export const Alert = (props: AlertProps) => {
...
Examples can be captioned. This is especially helpful when there are multiple examples.
/**
* Footer element with all needed legal and info links.
* Be sure to place it manually at the bottom of your page.
* You can use position 'absolute' with relative body or set the height of the content and place footer at the end.
* @example <caption>Appending custom element after default elements</caption>
* <Footer customElements={[{title: 'Bentley', url: 'https://www.bentley.com/'}]} />
* @example <caption>Returning only custom elements</caption>
* <Footer customElements={() => newFooterElements)} />
...
*/
export const Footer = (props: FooterProps) => {
...
Every prop should have a multiline description with relevant informational tags.
export type AlertProps = {
/**
* Type of the alert.
* @default 'informational'
*/
type?: 'positive' | 'warning' | 'negative' | 'informational';
/**
* Action handler for the clickable text.
* @deprecated `clickableTextProps` should be used instead.
*/
onClick?: () => void;
...
More examples can be found in the style guide.
Each component has a corresponding jest test inside of its directory. Be sure to cover your added code with tests.
Use yarn test
to run the tests. Add --watch
flag if you want tests to rerun after changes.
We usually do not use describe
block and our test case should start with 'should'.
it('should be visible', () => {
const { getByText } = render(
<Tooltip parentId='container' content='some text' isVisible>
<div>Visible!</div>
</Tooltip>,
);
getByText('some text');
});
We reuse our stories for visual tests by taking screenshots of the story iframes in Cypress.
- Make sure you have Docker installed and running.
- From the monorepo root, run
yarn test --filter=storybook
. This will build storybook and run all cypress tests in docker.- If you only need to run tests for a specific component, you can do so by passing the
--spec
argument to cypress. e.g. for testingAlert
, you can runyarn workspace storybook test --spec="**/Alert.*"
. Don't forget to build storybook first (yarn build --filter=storybook).
- If you only need to run tests for a specific component, you can do so by passing the
- Once the tests finish running, you can approve any failing test images using
yarn workspace storybook test:approve
.
Inside the apps/storybook
workspace, the src/
directory has a set of -.stories.tsx
files, each of which is accompanied by a -.test.ts
file. Here's what a typical test file should look like:
describe('Alert', () => {
const storyPath = 'Core/Alert';
const tests = [
'Positive',
'Negative',
'Warning',
'Informational',
];
tests.forEach((testName) => {
it(testName, () => {
const id = Cypress.storyId(storyPath, testName);
cy.visit('iframe', { qs: { id } });
cy.compareSnapshot(testName);
});
});
});
Notice how we do all of these things manually:
- defining the names of all the stories that need to be tested and excluding the ones that don't.
- specifying the story iframe url using the custom
storyId
helper. - visiting the iframe using
cy.visit
. - taking and comparing the screenshot using
cy.compareSnapshot
fromcypress-image-diff-js
.
We have full access to the Cypress API so any additional interactions or custom logic can be easily added.
Before creating a pull request, make sure your changes address a specific issue. Do a search to see if there are any existing issues that are still open. If you don't find one, you can create one.
To enable us to quickly review and accept your pull requests, always create one pull request per issue. Never merge multiple requests in one unless they have the same root cause. Be sure to follow best practices and keep code changes as small as possible. Avoid pure formatting changes or random "fixes" that are unrelated to the linked issue.
- Component added or modified using guidelines above.
- All required files and exports added.
- Proper inline documentation added.
- Code follows style guide and has no linting errors (pre-commit hook will run linter).
- Tests added for all new code.
- All existing and new tests should pass.
- Stories added to demonstrate new features.
Verify that your changes are ready, then create a pull request from your fork. Make sure that the name of your pull request follows the Conventional Commits spec, and that you have a proper description with screenshots and a closing keyword.
Your pull request will be reviewed by one or more maintainers who might leave some comments/suggestions to help improve the quality and consistency of your code. Once approved, your changes will be accepted into the repository.