Testing a pull request
When submitting a pull request to Sourcegraph, there are a few tests you’ll need to pass before the request can be reviewed. These tests run logical tests that we author to make sure our code runs as expected.
What are the tests?
- Prettier is responsible for mostly visual choices like if we have a line of code that’s longer than 70 characters, it will fail. This has nothing to do with the correctness of the code, more of a stylistic choice we make as a team to make sure we all use the same standard.
- eslint is similar but more so focused on the logic aspect of the code. We use eslint to avoid bad patterns which otherwise stylistically make sense but lead to potentially buggy code
- Husky (link) and Semantic Pull Requests (link) ensure your commit messages have enough semantic information to be able to trigger a release. For example,
feat:tag is for new features.
Manual formatting and linting tools
If you don’t want to wait for CI to find out whether you’ve made a mistake, you can either use an editor or IDE that hooks itself up with prettier and eslint, or use the following commands manually before submitting your pull request:
yarn format: applies prettier to your code (takes about 30s to run)
yarn format:check: checks if your code is passing the prettier checks (takes about 30s to run)
yarn eslint: checks if your code is passing the eslint checks (takes about 2s to run)
sg lint: can run any prettier and lint checks. (complete docs) The benefit is that it runs them the same way that CI runs them (On Buildkite, it’s called “Linters and static analysis” → “Run sg lint”). It’s also fast. The downside might be that it doesn’t necessarily run on all files, e.g. Prettier skipped
client/jetbrainsduring some testing.