Contributing

All contributions to this project will be released under the CC0 public domain dedication. By submitting a pull request or filing a bug, issue, or feature request, you are agreeing to comply with this waiver of copyright interest. Details can be found in our TERMS and LICENSE.


There are two primary ways to help:

  • Using the issue tracker, and
  • Changing the codebase.

Using the issue tracker

Use the issue tracker to suggest feature requests, report bugs, and ask questions. This is also a great way to connect with the developers of the project as well as others who are interested in this solution.

Use the issue tracker to find ways to contribute. Find a bug or a feature, mention in the issue that you will take on that effort, then follow the Changing the codebase guidance below.

Changing the codebase

We work off feature branches from the master branch. After you’ve edited a component, open a Pull Request to merge your feature branch back into master.

For example, if you wanted to change cf-buttons and use it in cfgov-refresh, here’s what you’d do:

  1. git clone git@github.com:cfpb/capital-framework.git, if you haven’t already.
  2. cd capital-framework
  3. git checkout master && git pull to ensure you’re on the latest changes (this step is not necessary when cloning for the first time).
  4. yarn to install dependencies and set up workspaces
  5. git checkout -b button-fix to create a new branch for your changes.
  6. Edit file(s) in /packages/cf-buttons/ however you want.
  7. Run gulp docs to update the documentation.
  8. yarn run cf-link to link your local CF components.
  9. cd ~/wherever/cfgov-refresh/ to navigate to another project where you’d like to test your buttons changes (in this case, cfgov-refresh).
  10. yarn link cf-buttons to link cf-buttons to cfgov-refresh.
  11. gulp build in cfgov-refresh to compile your stylesheets.
  12. Start cfgov-refresh and navigate to a page with buttons to view your cf-buttons changes.
  13. When you’re pleased with your changes, cd back to your capital-framework repo and commit your changes: git commit -am "Fix button border radius"
  14. git push origin button-fix to push your branch up to GitHub.
  15. Go to https://github.com/cfpb/capital-framework and open a pull request to merge button-fix into master.

If you are not a current contributor to Capital Framework, use forks by first clicking the fork button on top of the repository and cloning your fork in step 1. In the final step, go to https://github.com/cfpb/capital-framework and file a pull request by clicking the link to compare changes across forks.

Updating Documentation

Documentation is maintained in each component’s folder’s usage.md file. Changes to documentation are made on the master branch using the workflow above and pulled in the docs directory using the gulp docs command. The usage.md files are then rendered as separated pages on Capital Framework’s website (See this example from cf-core).

Testing components locally

Automated tests

Automated tests can be run with the command yarn test.

Testing in the documentation site

If you’re hacking on a component and want to test it in the documentation site follow the following steps.

  1. Ensure that you have the Bundler Ruby gem installed by running gem install bundler. (more info).
  2. Check out a working branch.
  3. Run gulp docs.
  4. Go into the documentation directory with cd docs.
  5. Run bundle install if you haven’t before (also run bundle update periodically).
  6. Run bundle exec jekyll serve watch and visit http://localhost:4000/.

Browser support

We configure Autoprefixer and Babel to support the following list of browsers.

  • Latest 2 releases of all browsers including:
    • Chrome
    • Firefox
    • Safari
    • Internet Explorer
    • Edge
    • Opera
    • iOS Safari
    • Opera Mini
    • Android Browser
    • BlackBerry Browser
    • Opera Mobile
    • Chrome for Android
    • Firefox for Android
    • Samsung Internet
  • Internet Explorer 9

http://browserl.ist/?q=last+2+versions%2C+Explorer+%3E%3D+9

As well as additional Autoprefixer support for:

  • Internet Explorer 8

http://browserl.ist/?q=last+2+versions%2C+Explorer+%3E%3D+8

What this means to the end-user is we’ve added a level of backward compatability for modern features as much as possible. This doesn’t necessarily mean feature parity. Where it’s impossible or impractical to implement a modern feature, we fallback to standard practices for that browser. For example, we do not deliver interactive scripting for Internet Explorer 8, but we do ensure that default browser features continue to work so users that can’t or don’t want to upgrade continue to have access to the site and our content.

Browser Testing

We have automated tests that use a headless version of Chrome to ensure the majority of the site is working as expected. For manual testing, we realistically test this project locally or in a virtual environment with the following list of browsers:

  • Chrome
  • Firefox
  • Safari
  • Internet Explorer 8, 9, 10, and 11
  • Edge
  • iOS Safari
  • Chrome for Android

Autoprefixer

Autoprefixer parses our CSS and adds vendor prefixes to rules where necessary using reported feature support by Can I Use. For more information visit the [Autoprefixer documentation site] (https://autoprefixer.github.io/).

Babel

Babel compiles our ES6 JavaScript where necessary for the browsers that either don’t support or have limited support of ES6 features. For more information visit the [Babel documentation site] (https://babeljs.io/).

Known feature differences

  • JavaScript: We do not serve interactive scripting to IE 8 but we do deliver analytics via JavaScript.
  • Icons: We currently use icon fonts to deliver scalable icons. Browsers that do not support icon fonts unfortunately do not receive backups but we try to always pair icons with text.

Resources

  • https://developer.microsoft.com/en-us/microsoft-edge/tools/vms/
  • https://saucelabs.com/beta/dashboard/tests
  • http://developer.samsung.com/remotetestlab/rtlDeviceList.action#

Release management

Ready to publish changes to npm?

  1. Ensure you’re on master and git pull to confirm you’re up-to-date.
  2. Export a personal access token called GITHUB_AUTH.
  3. Run yarn run changelog and open CHANGELOG.md to see a preview of new changelog entries. We use a tool that scans our Pull Requests for specific labels so if you see a PR missing from the changelog, ensure it has been labeled breaking, enhancement, bug, documentation or internal.
  4. If you’re pleased with the changelog preview, git checkout CHANGELOG.md to reset it.
  5. Run yarn run release to start the release.
  6. If all packages are shown as having been published, run git push to push changes to the remote master branch.

Lerna will update the changelog, ask for a new version number, create a git tag, push to GitHub and publish to npm.

Notes

The CSS and JavaScript files that are generated during the build task are only used for testing. Because this project doesn’t necessarily produce a final product it is up to the projects that use it to generate and maintain their own browser support config. An example of this is consumerfinance.gov, which generates specific IE 8 and 9 stylesheets, whereas the Capital Framework build task bundles those with the main stylesheet. Despite the differences in delivery, the output and support are the same.

Adhere to any linting errors or warnings

The gulp lint linting tasks that are set up within the build processes are there to promote consistency. When contributing code please publicly track that there are no linting errors or warnings using the testing checklist in the pull request description.

Follow our CSS naming conventions

We are using a customized BEM format

.block-name
.block-name_element-name
.block-name__block-modifier
.block-name_element-name__element-modifier

Avoid creating elements of modifiers

Appending an element name to a modifier class can result in a confusing class name like .list__space_item. Avoid this in favor of using a descendant, like this: .list__spaced .list_item.

Shoot for mobile first declarations

In most cases styles should be declared mobile first, then enhanced with min-width media queries. By doing this we create a base experience that all devices can use and one that does not require media query support.