danger

What is danger?

Danger is a tool that helps automate code review chores. It allows you to create rules that can be enforced during the code review process, ensuring that certain standards and practices are followed.

What are danger's main functionalities?

Automate Code Review

This feature allows you to automate code review tasks by defining rules that can fail or warn if certain conditions are not met. For example, you can fail a pull request if it does not have a description or warn if it does not have any labels.

danger({
  fail: 'This PR does not have a description.',
  warn: 'This PR does not have any labels.'
});

Check for Changelog Updates

This feature checks if the changelog has been updated in the pull request. If not, it can warn the user to add a changelog entry for their changes.

if (!danger.git.modified_files.includes('CHANGELOG.md')) {
  warn('Please add a changelog entry for your changes.');
}

Enforce PR Size Limits

This feature allows you to enforce size limits on pull requests. If a pull request exceeds a certain number of additions and deletions, it can warn the user to consider breaking it down into smaller PRs.

const bigPRThreshold = 500;
if (danger.github.pr.additions + danger.github.pr.deletions > bigPRThreshold) {
  warn('This PR is too large. Consider breaking it down into smaller PRs.');
}

Other packages similar to danger

eslint

ESLint is a tool for identifying and reporting on patterns found in ECMAScript/JavaScript code. While it focuses on linting and enforcing coding standards, it can be used in conjunction with Danger to ensure code quality.

prettier

Prettier is an opinionated code formatter that enforces a consistent style by parsing your code and re-printing it. It complements Danger by ensuring code style consistency, which can be part of the automated code review process.

husky

Husky is a tool that allows you to run scripts before committing or pushing code. It can be used to enforce pre-commit hooks, ensuring that certain checks (like those defined in Danger) are run before code is committed.

README

Danger on Node, wonder what's going on? see VISION.md

Welcome!

So, what's the deal? Well, right now Danger JS does a lot of the simpler parts of the Ruby version. You can look at Git metadata, or GitHub metadata on Travis, Circle, Semaphore, Jenkins, Docker Cloud, or Codeship.

Danger can fail your build, write a comment on GitHub, edit it as your PR changes and then delete it once you've passed review. Perfect.

Getting set up in your Project

# with yarn
yarn add danger --dev

# or with npm
npm install --save-dev danger

If using NPM, add a run command to your package.json

{
  "scripts": {
    "danger": "danger"  
  }
}

Then create a dangerfile.js in the project root with some rules:

import { danger, fail, warn } from "danger"
import fs from "fs"

// Make sure there are changelog entries
const hasChangelog = danger.git.modified_files.includes("changelog.md")
if (!hasChangelog) { fail("No Changelog changes!") }

const jsFiles = danger.git.created_files.filter(path => path.endsWith("js"))

// new js files should have `@flow` at the top
const unFlowedFiles = jsFiles.filter(filepath => {
  const content = fs.readFileSync(filepath)
  return !content.includes("@flow")
})

if (unFlowedFiles.length > 0) {
  warn(`These new JS files do not have Flow enabled: ${unFlowedFiles.join(", ")}`)
}

Then you add yarn run danger to the end of your CI run, and Danger will run. Here's an example. 👍

Want to see some existing examples? Check out:

• Apps - Artsy/Emission and Artsy/metaphysics.
• Libraries - Facebook/Jest, styled-components/styled-components and ReactiveX/rxjs.

I'd love PRs adding more.

Getting set up on CI

For now, to get set up I'd recommend looking at the setup guide for the Ruby version. All the environment vars are the exact same between versions.

You will need to create a bot account, and set up CI to run danger.

If you are using Docker Cloud, make sure to set the following blank ENV vars in your docker-compose.test.yml file so they are carried forward from the build environment:

sut:
  build: .
  environment:
    - DANGER_GITHUB_API_TOKEN
    - DOCKER_REPO
    - PULL_REQUEST_URL
    - SOURCE_REPOSITORY_URL

Running/Testing manually against a repo

There are two ways to do this:

Using danger pr

The command danger pr expects an argument of a PR url, e.g. danger pr https://github.com/danger/danger-js/pull/100.

This will use your local dangerfile.js against the metadata of that PR. Danger will then output the results as JSON, instead of on the PR itself.

Using danger

If you create an appropriately scoped temporary api token for your github account, this can be a good way to see if danger is suitable for you before integrating it into your CI system.

You can manually trigger danger against a pull request on the command line by setting the following environmental variables:

export DANGER_FAKE_CI="YEP"
export DANGER_GITHUB_API_TOKEN='xxxxxxxxxx'  # a github api token
export DANGER_TEST_REPO='username/reponame'

Then you can run against a local branch that is attached to a pull-request, by running the following

git checkout branch-for-pr-1234
DANGER_TEST_PR='1234' npm run danger

assuming that your local file-system matches up to that branch on github, this will be a good approximation of how danger will work when you integrate it into your CI system.

Note: this will leave a comment on the PR.

Advice

• You can have Danger read build logs if you use tee in your CI process: yarn run lint | tee linter_output.txt. This can then be picked up with readFileSync at linter_output.txt in your Dangerfile later.

Known issues

• We're still figuring out how to handle async code correctly (you can use await with no problem though)
• Codeship support does not support fork to fork GitHub PRs.

This thing is broken, I should help improve it!

Awesommmmee.

git clone https://github.com/danger/danger-js.git
cd danger-js

# if you don't have yarn installed
npm install -g yarn

yarn install

You can then verify your install by running the tests, and the linters:

yarn test
yarn lint

---

Dev Life

Tips:

• You can run the danger command globally from your dev build by running yarn run link.
• If you're using VS Code, press Run to start an example run, should go through most of the process with a debugger attached. Either use breakpoints, or add debugger to get a repl and context.

What is the TODO?

Check the issues, I try and keep my short term perspective there. Long term is in the VISION.md.

Changelog

0.11.3 - 0.11.5

• Internal changes for usage with Peril - [@orta]

• Add danger pr --repl, which drops into a Node.js REPL after evaluating the dangerfile - macklinu

• Add support for Codeship - deecewan