Skip to content

Latest commit

 

History

History
660 lines (445 loc) · 48.1 KB

CONTRIBUTING.md

File metadata and controls

660 lines (445 loc) · 48.1 KB

Contributing to Cypress

Thanks for taking the time to contribute! 😄

Once you learn how to use Cypress, you can contribute in many ways:

Want to dive deeper into how Cypress works? There are several ways you can help with the development of Cypress:

Table of Contents

Code of Conduct

All contributors are expecting to abide by our Code of Conduct.

Opening Issues

The most important things to do are:

Finally, if you are up to date, supported, have collected information about the problem, and have the best reproduction instructions you can give, you are ready to open an issue.

Update Cypress

Before filing a bug, make sure you are up to date. Your issue may have already been fixed. Even if you do not see the issue described as resolved in a newer version, a newer version may help in the process of debugging your issue by giving more helpful error messages.

See our document on installing cypress

Getting more information

For some issues, there are places you can check for more information. This may help you resolve the issue yourself. Even if it doesn't, this information can help us figure out and resolve an issue.

  • For issues in the web browser, check the JavaScript console and your Network tab in your DevTools.
  • Click on any command in the Command Log where the failure occurred, this will log more information about the error to the JavaScript console.
  • Use Cypress debug or pause commands to step through your commands.
  • Ask other Cypress users for help in our chat or Discord.
  • Try more advanced troubleshooting from troubleshooting Cypress doc.

Fill out our Issue Template

When opening an issue, there is a provided issue template. Fill out the information according to the template. This is information needed for Cypress to continue forward with your problem. Any issues that don't fill out the issue template will be closed.

Describe Problems

When you file a feature request or bug, it's best to describe the problem you are facing first, not just your desired solution.

Often, your problem may have a lot in common with other similar problems. If we understand your use case, we can compare it to other use cases and sometimes find a more powerful or more general solution which solves several problems at once. Understanding the root issue can let us merge and contextualize things. Sometimes there's already a way to solve your problem that might just not be obvious.

Also, your proposed solution may not be compatible with the direction we want to take the product, but we may be able to come up with another solution which has approximately the same effect and does fit into the product direction.

Reproducibility

It is nearly impossible for us to resolve issues if we can not reproduce them. Your best chance of getting a bug looked at quickly is to provide a repository with a reproducible bug that can be cloned and run.

Common issues

Label Description Issues
browser detection Local browser is not detected open, closed
cross-origin Getting cross-origin error open, closed
cy.request Issues related to cy.request command open, closed
fixtures Fixture loading and usage
hooks Issues related to hooks open, closed
iframes Working with iframes open, closed
installation Cypress cannot be downloaded or installed open, closed
network Controlling network requests open, closed
performance Slow loading, slow network, etc open, closed
screenshots Taking image screenshots open, closed
scrolling Scrolling elements into view open, closed
spec execution Running all specs or some specs in some specific order open
test execution Running tests inside a single spec open, closed
typescript Transpiling or bundling TypeScript open, closed
video Problems with video recordings open, closed

Triaging Issues

When an issue is opened in cypress, we need to evaluate the issue to determine what steps should be taken next. So, when approaching new issues, there are some steps that should be taken.

Is this a question?

Some opened issues are questions, not bug reports or feature requests. Issues are reserved for potential bugs or feature requests only. If this is the case, you should:

  • Explain that issues in our GitHub repo are reserved for potential bugs or feature requests and that the issue will be closed since it appears to be neither a bug nor a feature request.
  • Guide them to existing resources where their questions can be asked like our Discussions, community chat, Discord, or Stack Overflow.
  • Cypress offers support via email when signing up for any of our paid plans, so remind them that this is an option if they already have a paid account.
  • Move the issue to Discussions.

Does this issue belong in this repository?

Other open source repos

Issues may be opened about wanting changes to our documentation, our example-kitchensink app, or another repository. In this case you should:

  • Thank them for their contribution.
  • Explain that this repo is only for bugs or feature requests of the Cypress App.
  • If you have permission to 'Transfer the issue', do so. If not, explain that they can open an issue in our other repository and link to the repository.
  • Close the issue (if not already transferred).

Cypress Dashboard

Issues may be opened about wanting features in our Dashboard Service. In this case you should:

  • Thank them for opening an issue.
  • Add the external: dashboard label.

Component Testing

Issues may be opened about wanting features in Component Testing. In this case you should:

  • Thank them for opening an issue.
  • Add the component testing label.

Is this already an open issue?

Search all issues for keywords from the issue to ensure there isn't already an issue open for this. GitHub has some search tips that may help you better find the relevant issue.

If an issue already exists you should:

  • Thank them for their contribution.
  • Explain that this issue is a duplicate of another issue, linking to the relevant issue (#1234).
  • Add the type: duplicate label to the issue.
  • Close the issue.

Does the issue provide all the information from our issue template?

When opening an issue, there is a provided issue template based on the type of issue. If the opened issue does not provide enough information asked from the issue template you should:

  • Explain that we require new issues follow our provided issue template and that issues that are opened without this information are automatically closed per our contributing guidelines.
  • Close the issue.

Are they running the current version of Cypress?

If they listed an older version of Cypress in their issue. We don't want to spend the time to set up a reproducible project (which can be time consuming) only to find that bumping the Cypress version fixes it. You should:

  • Ask them to update to the newest version of Cypress and comment about the results.
  • Add the stage: awaiting response label to the issue.

Is the fix or feature within our vision for Cypress?

There will inevitably be suggestions that will not fit within the scope of Cypress's vision for our product. If an issue or pull request falls under this category you should:

  • Thank them for their contribution.
  • Explain why it doesn't fit into the scope at Cypress, and offer clear suggestions for improvement, if you're able. Be kind, but firm.
  • Link to relevant documentation, if there is any. If you notice repeated requests for things that are not within scope, add them into the documentation to avoid repeating yourself.
  • Add the stage: wontfix label to the issue.
  • Close the issue/pull request.

Is what they're describing actually happening?

The best way to determine the validity of a bug is to recreate it yourself. Follow the directions or information provided to recreate the bug that is described. Did they provide a repository that demonstrates the bug? Great - fork it and run the project and steps required. If they didn't provide a repository, the best way to reproduce the issue is to have a 'sandbox' project up and running locally for Cypress. This is just a simple project with Cypress installed where you can freely edit the application under test and the tests themselves to recreate the problem.

Attempting to recreate the bug will lead to a few scenarios:

1. You can't recreate the bug

If you can't recreate the situation happening you should:

  • Thank them for their contribution.
  • Explain that there isn't enough information to reproduce the bug. Provide information on how you went about recreating the scenario, if you're able. Note your OS, Browser, Cypress version and any other information.
  • Note that if no reproducible example is provided, we will unfortunately have to close the issue.
  • Add the stage: needs information label to the issue.

2. You can recreate the bug

If you can recreate the bug you should:

  • Thank them for their contribution.
  • Explain that you're able to recreate the bug. Provide the exact test code ran and the versions of Cypress, OS, and browser you used to recreate it.
  • If you know where the code is that could possibly fix this issue - link to the file or line of code from the cypress repo and remind the user that we are open source and that we gladly accept PRs, even if they are a work in progress.
  • Add the stage: ready for work label to the issue.

3. You can tell the problem is a user error

In recreating the issue, you may realize that they had a typo or used the Cypress API incorrectly, etc. In this case you should:

  • Leave a comment informing the user of their error.
  • Link to relevant documentation, if there is any. If you notice repeated user errors for the same situation, add them into the documentation to avoid repeating yourself.
  • Close the issue.

Has the issue gone stale?

Some issues are opened and sadly forgotten about by the person originally opening the issue.

Not enough information ever provided

Sometimes we request more information to be provided (label stage: needs information) for an open issue, but no one is able to provide a reproducible example or they simply never respond. This does not mean that we don't believe that there is a bug! We just, unfortunately, don't have a path forward to fix it without this information. In this case you should:

  • Add a comment reminding them or our request for more information and that the issue will be closed if it is not provided. Sometimes issues get forgotten about, and all the person needs is a gentle reminder.
  • If there is still no response after a weeks time, explain that you are closing the issue due to not enough information or inactivity and that they can comment in the issue with a reproducible example and we will reopen the issue.
  • Close the issue.

They already solved their issue

Some issues are resolved by the community, by giving some guidance or a workaround, but the original opener of the issue forgets to close the issue. In this case you should:

  • Explain that you are closing the issue as resolved and that they can comment if they are still having the issue and we will consider reopening it.
  • Close the issue.

Writing Documentation

Cypress documentation lives in a separate repository with its own dependencies and build tools. See Documentation Contributing Guideline.

Writing code

Working on your first Pull Request? You can learn how from this free series How to Contribute to an Open Source Project on GitHub

What you need to know before getting started

Cypress and Packages

Cypress is a large open source project. When you want to contribute to Cypress, you may be unsure which part of the project to work within.

Cypress uses a monorepo, which means there are many independent packages in this repository. There are two main types of packages: private and public.

Private packages generally live within the packages directory and are in the @packages/ namespace. These packages are combined to form the main Cypress app that you get when you npm install cypress. They are discrete modules with different responsibilities, but each is necessary for the Cypress app and is not necessarily useful outside of the Cypress app. Since these modules are all compiled and bundled into a binary upon release, they are sometimes collectively referred to as the Cypress binary.

Here is a list of the core packages in this repository with a short description, located within the packages directory:

Folder Name Package Name Purpose
cli cypress The command-line tool that is packaged as an npm module.
desktop-gui @packages/desktop-gui The front-end code for the Cypress Desktop GUI.
driver @packages/driver The code that is used to drive the behavior of the API commands.
electron @packages/electron The Cypress implementation of Electron.
example @packages/example Our example kitchen-sink application.
extension @packages/extension The Cypress Chrome browser extension
https-proxy @packages/https-proxy This does https proxy for handling http certs and traffic.
net-stubbing @packages/net-stubbing Contains server side code for Cypress' network stubbing features.
network @packages/network Various utilities related to networking.
proxy @packages/proxy Code for Cypress' network proxy layer.
launcher @packages/launcher Finds and launches browsers installed on your system.
reporter @packages/reporter The reporter shows the running results of the tests (The Command Log UI).
root @packages/root Dummy package pointing at the root of the repository.
runner @packages/runner The runner is the minimal "chrome" around the user's application under test.
runner-ct @packages/runner-ct The runner for component testing
runner-shared @packages/runner-shared The shared components between the runner and the runner-ct packages
server @packages/server The <3 of Cypress. This orchestrates everything. The backend node process.
server-ct @packages/server-ct Some Component Testing specific overrides. Mostly extends functionality from @packages/server
socket @packages/socket A wrapper around socket.io to provide common libraries.
ts @packages/ts A centralized version of typescript.

Public packages live within the npm folder and are standalone modules that get independently published to npm under the @cypress/ namespace. These packages generally contain extensions, plugins, or other packages that are complementary to, yet independent of, the main Cypress app.

Here is a list of the npm packages in this repository:

Folder Name Package Name Purpose
angular @cypress/angular Cypress component testing for Angular.
create-cypress-tests @cypress/create-cypress-tests Tooling to scaffold Cypress configuration and demo test files.
eslint-plugin-dev @cypress/eslint-plugin-dev Eslint plugin for internal development.
mount-utils @cypress/mount-utils Common functionality for Vue/React/Angular adapters.
react @cypress/react Cypress component testing for React.
vite-dev-server @cypress/vite-dev-server Vite powered dev server for Component Testing.
webpack-preprocessor @cypress/webpack-preprocessor Cypress preprocessor for bundling JavaScript via webpack.
webpack-dev-server @cypress/webpack-dev-server Webpack powered dev server for Component Testing.
vue @cypress/vue Cypress component testing for Vue.

We try to tag all issues with a pkg/ or npm/ tag describing the appropriate package the work is required in. For public packages, we use their qualified package name: For example, issues relating to the webpack preprocessor are tagged under npm: @cypress/webpack-preprocessor label and issues related to the driver package are tagged with the pkg/driver label.

Requirements

You must have the following installed on your system to contribute locally:

  • Node.js (See the root .node-version file for minimum version requirements. You can use avn to automatically switch to the right version of Node.js for this repo.)
  • yarn
  • python (since we use node-gyp. See their repo for Python version requirements.)

Getting Started

The project utilizes yarn workspaces and leverages lerna to orchestrate running within the context of one or more packages. While it is not as important to understand lerna or yarn workspaces, it is important to note that running scripts or installing new dependencies should always happen from the repo's root directory.

⚠ Running on Windows?

Many of the NPM scripts used during development use commands designed for a Linux-like shell.If you are running a Windows operating system, you may encounter many commands that are not working. To fix this behavior, you have to set a Linux-like shell as the default npm script shell. If you have Git for Windows installed, you can set Git Bash as the default script shell by using the following command:

yarn config set script-shell "C:\\Program Files\\git\\bin\\bash.exe"

Install all dependencies:

yarn

This will install all the dependencies for the repo and perform a preliminary build.

Next, start the app:

yarn start

If there are errors building the packages, prefix the commands with DEBUG=cypress:* to see more details.

This outputs a lot of debugging lines. To focus on an individual module, run with DEBUG=cypress:launcher for instance.

When running yarn start this routes through the CLI and eventually calls yarn dev with the proper arguments. This enables Cypress day-to-day development to match the logic of the built binary + CLI integration.

If you want to bypass the CLI entirely, you can use the yarn dev task and pass arguments directly. For example, to headlessly run a project in a given folder, while trying to record to the Dashboard

yarn dev --run-project /project/folder --record --key <key>

Adding new Dependencies

⚠️ There is a bug in yarn that may cause issues adding a new dependency to a workspace. You can avoid this by downgrading yarn to 1.19.1 (temporarily downgrade using npx [email protected] workspace @packages/server add my-new-dep1).

# add a new dep to the root of the repo
$ yarn add -W my-new-dep1

# add a new dep to a specific package
$ yarn workspace @packages/server add my-new-dep1
$ yarn workspace @packages/server add --dev my-new-dep1

Alternatively, you can directly add the dependency to the corresponding package.json and run yarn.

Tasks

Scripts are intended to be run from the root of the repo. Do not install dependencies or run scripts from within a sub-directory.

Common Top Level Tasks

By default, top level tasks will execute for all packages. However, most scripts can be provided one or more scopes. Providing a scope will execute tasks within the provided packages. Scope values are based on package names and not the directory structure.

Task Purpose
build Compile non-node code (typescript)
start Open Cypress in dev and global mode
watch Auto-rebuild on file changes
clean Remove build artifacts
clean-deps Remove all installed dependencies (in root and in every package)
test Run the default set of tests (may be package dependent)
test-debug Run unit/integration tests with inspect node CLI flags
test-unit Run unit tests
test-integration Run integration tests
test-e2e Run end-to-end tests
test-system Run system tests
test-watch Run unit tests and rebuild/rerun on file changes

Most of the time you will only want to run a task within a specific package; this can be done by providing the package name as a scope to the top level task.

# Run test-unit only within cypress package (./cli)
$ yarn test-unit --scope cypress

# Run test-unit only within the cypress and server packages (./cli & ./packages/server)
$ yarn test-unit --scope cypress --scope @packages/server

# Run test-unit in all packages with the name starting with `@packages/`
$ yarn test-unit --scope @packages/*
Package-Level Scripts

Although scripts are meant to be run from the root of the repo; they typically delegate to the scripts within the packages.

Each package is responsible for building itself and testing itself and can do so using whatever tools are appropriate, but each conforms to a standard set of scripts so that building, watching, testing, etc. can be orchestrated from the root of this repo. Here are the scripts supported and what they mean:

Task Purpose
build Build the package
build-prod Build all assets for production (if makes sense)
start Run a server for serving files
watch Watch source files and build development assets when they are saved. This may also run a server for serving files and run tests related to a saved file.
clean Remove any assets created by build-dev or build-prod
clean-deps Remove any dependencies installed (usually by yarn)
test Runs all tests once (this usually means running unit tests; via yarn test-unit)
test-unit Run all unit tests within the package; exit 0 if N/A
test-integration Run all integration tests within the package; exit 0 if N/A
test-watch Run all unit tests in the package in watch mode

Debugging

Some packages use debug to log debug messages to the console. The naming scheme should be cypress:<package name>; where package name is without the @packages scope. For example to see launcher messages during unit tests start it using

$ DEBUG=cypress:launcher yarn test --scope @packages/launcher

If you want to see log messages from all Cypress projects use wild card

$ DEBUG=cypress:*

Or for an individual package:

DEBUG=cypress:cli
DEBUG=cypress:server
DEBUG=cypress:launcher

Coding Style

We use eslint to lint all JavaScript code and follow rules specified in @cypress/eslint-plugin-dev plugin.

When you edit files, you can quickly fix all changed files before you commit using

$ yarn lint-changed --fix

When committing files, we run a Git pre-commit hook to lint the staged JS files. See the lint-staged project. If this command fails, you may need to run yarn lint-changed --fix and commit those changes.

We DO NOT use Prettier to format code. You can find .prettierignore file that ignores all files in this repository. To ensure this file is loaded, please always open the root repository folder in your text editor, otherwise your code formatter might execute, reformatting lots of source files.

Adding links within code

When adding links to outside resources within the Cypress Test Runner (including links to Cypress's own documentation), we utilize our cypress-on service for all links.

This is to ensure that links do not go dead in older versions of Cypress when the location of the link has changed. To add a new link:

  • Make up a new slug for the linked resource like https://on.cypress.io/my-special-link.
  • Open a PR adding the new slug in links.yml with the href of the resource it should redirect to. Note: this requires access to the internal cypress-services repo which is only granted to Cypress employees. If you're an outside contributor and need a link reroute added, please comment in the relevant PR asking for assistance.
  • Wait for the PR to be reviewed and deployed from cypress-services. This is required before your changes can be merged into the cypress project.

Tests

For most packages there are typically unit and integration tests.

Please refer to each packages' README.md which documents how to run tests. It is not feasible to try to run all of the tests together. We run our entire test fleet across over a dozen containers in CI.

There are also a set of system tests in system-tests which attempt to test the entire Cypress App as close to real world as possible. See the README for more information.

Additionally, we test the code by running it against various other example projects in CI. See CI badges and links at the top of this document.

If you're curious how we manage all of these tests in CI check out our circle.yml file found in the root cypress directory.

Docker

Sometimes tests pass locally, but fail in CI. Our CI environment is dockerized. In order to run the image used in CI locally:

  1. Install Docker and get it running on your machine.
  2. Run the following command from the root of the project:
$ yarn docker

There is a script scripts/run-docker-local.sh that runs the cypress image (see circle.yml for the current image name).

The image will start and will map the root of the repository to /cypress inside the image. Now you can modify the files using your favorite environment and rerun tests inside the docker environment.

Docker for built binary

You can also use Docker to simulate and debug the built binary. In a temporary folder (for example from the folder /tmp/test-folder/) start a Docker image:

$ docker run -it -w /app -v $PWD:/app cypress/base:8 /bin/bash

Point the installation at a specific beta binary and NPM package archive (if needed) and set local cache folder to unzip the downloaded binary into a subfolder.

$ export CYPRESS_INSTALL_BINARY=https://cdn.cypress.io/beta/.../cypress.zip
$ export CYPRESS_CACHE_FOLDER=./cypress-cache
$ yarn add https://cdn.cypress.io/beta/npm/.../cypress.tgz

Note that unzipping the Linux binary inside a Docker container onto a mapped volume drive is slow. But once this is done you can modify the application resource folder in the local folder /tmp/test-folder/node_modules/cypress/cypress-cache/3.3.0/Cypress/resources/app to debug issues.

Packages

Generally when making contributions, you are typically making them to a small number of packages. Most of your local development work will be inside a single package at a time.

Each package documents how to best work with it, so consult the README.md of each package.

They will outline development and test procedures. When in doubt just look at the scripts of each package.json file. Everything we do at Cypress is contained there.

Committing Code

Branches

The repository is setup with two main (protected) branches.

  • master is the code already published, both for the main Cypress app and independent npm packages.
  • develop is the current latest "pre-release" code. This branch is set as the default branch, and all pull requests that update the main Cypress binary should be made against this branch.

In general, we want to publish our standalone npm packages continuously as new features are added. Therefore, any pull requests that only change independent @cypress/ packages in the npm directory should be made directly off the master branch. We use semantic-release to automatically publish these packages to npm when a PR is merged directly into master.

When updating the main Cypress app, pull requests should be made against the develop branch. We do not continuously deploy the Cypress binary, so develop contains all of the new features and fixes that are staged to go out in the next update of the main Cypress app. In addition, if you make changes to an npm package that can't be published until the binary is also updated, you should make a pull request against the develop branch.

Essentially, if you only change files within the npm folder, then you should make a pull request against master. Otherwise, make it against develop.

All updates to master are automatically merged into develop, so develop always has the latest version of every package.

Workflow Diagrams

Independent Packages CI Workflow

Independent packages are automatically released when code is merged into master and the entire build passes.

Pull Requests

  • Break down pull requests into the smallest necessary parts to address the original issue or feature. This helps you get a timely review and helps the reviewer clearly understand which pieces of the code changes are relevant.
  • When opening a PR for a specific issue already open, please name the branch you are working on using the convention issue-[issue number]. For example, if your PR fixes Issue #803, name your branch issue-803. If the PR is a larger issue, you can add more context like issue-803-new-scrollable-area. If there's not an associated open issue, create an issue.
  • PR's can be opened before all the work is finished. In fact we encourage this! Please create a Draft Pull Request if your PR is not ready for review. Mark the PR as Ready for Review when you're ready for a Cypress team member to review the PR.
  • Prefix the title of the Pull Request using semantic-release's format as defined here. For example, if your PR is fixing a bug, you should prefix the PR title with fix:.
  • Fill out the Pull Request Template completely within the body of the PR. If you feel some areas are not relevant add N/A as opposed to deleting those sections. PR's will not be reviewed if this template is not filled in.
  • If the PR is a user facing change and you're a Cypress team member that has logged into ZenHub and downloaded the ZenHub for GitHub extension, set the release the PR is intended to ship in from the sidebar of the PR. Follow semantic versioning to select the intended release. This is used to generate the changelog for the release. If you don't tag a PR for release, it won't be mentioned in the changelog. Select release for PR
  • Please check the "Allow edits from maintainers" checkbox when submitting your PR. This will make it easier for the maintainers to make minor adjustments, to help with tests or any other changes we may need. Allow edits from maintainers checkbox
  • After the PR is approved, the original contributor can merge the PR (if the original contributor has access).
  • When you merge a PR into develop, select Squash and merge. This will squash all commits into a single commit. The only exception to squashing is when converting files to another language and there is a clear commit history needed to maintain from the file conversion.

Dependencies

We use RenovateBot to automatically upgrade our dependencies. The bot uses the settings in renovate.json to maintain our Update Dependencies issue and open PRs. You can manually select a package to open a PR from our Update Dependencies issue.

After a PR has been opened for a dependency update, our cypress-bot will comment on the PR detailing the guidelines to be used to review the dependency update. Please read these guidelines carefully and make any updates where you see the PR may not be meeting the quality of these guidelines.

Reviewing Code

Some rules about Code Review

  1. The contributor opening the pull request may not approve their own PR.
  2. The PR will not be merged if some reviewers have requested changes.

If any of the Pull Request Review guidelines can't be met, a comment should be left by the reviewer with 'Request changes'. The original contributor is responsible for making any updates and request re-review once those changes are addressed.

Steps to take during Code Review

  • Run the code and use it as the end user would.
  • Double check the issue and PR description to ensure it is meeting the original requirements.
  • Read through every line of changed code (Yes, we know this could be a LOT).
  • If you don't understand why some piece of code is required, ask for clarification! Likely the contributor had a reason and can provide the answer quicker than investigating yourself.

Code Review Checklist

Below are guidelines to help during code review. If any of the following requirements can't be met, leave a comment in the review selecting 'Request changes', otherwise 'Approve'.

User Experience

  • The feature/bugfix is self-documenting from within the product.
  • The change provides the end user with a way to fix their problem (no dead ends).
  • If a breaking change or a change to a commonly used API, the proposed changes have been discussed and agreed upon in the weekly team meeting (or a separate meeting if a larger change).

Functionality

  • The code works and performs its intended function with the correct logic.
  • Performance has been factored in (for example, the code cleans up after itself to not cause memory leaks).
  • The code guards against edge cases and invalid input and has tests to cover it.

Maintainability

  • The code is readable (too many nested 'if's are a bad sign).
  • Names used for variables, methods, etc, clearly describe their function.
  • The code is easy to understood and there are relevant comments explaining.
  • New algorithms are documented in the code with link(s) to external docs (flowcharts, w3c, chrome, firefox).
  • There are comments containing link(s) to the addressed issue (in tests and code).

Quality

  • The change does not reimplement code.
  • There's not a module from the ecosystem that should be used instead.
  • There is no redundant or duplicate code.
  • There are no irrelevant comments left in the code.
  • There is no irrelevant code to the issue being addressed. If there is, ask the contributor to break the work out into a separate PR.
  • Tests are testing the code's intended functionality in the best way possible.

Internal

  • The original issue has been tagged with a release in ZenHub.

Code Review of Dependency Updates

Below are some guidelines Cypress uses when reviewing dependency updates.

Dependency Update Instructions

  • Read through the entire changelog of the dependency's changes. If a changelog is not available, check every commit made to the dependency. NOTE - do not rely on semver to indicate breaking changes - every product does not follow this standard.
  • Add a PR review comment noting any relevant changes in the dependency.
  • If any of the following requirements cannot be met, leave a comment in the review selecting 'Request changes', otherwise 'Approve'.

Dependency Updates Checklist

  • Code using the dependency has been updated to accommodate any breaking changes
  • The dependency still supports the version of Node that the package requires.
  • The PR been tagged with a release in ZenHub.
  • Appropriate labels have been added to the PR (for example: label type: breaking change if it is a breaking change)

Deployment

We will try to review and merge pull requests quickly. If you want to know our build process or build your own Cypress binary, read the "Release Process" guide.

Independent packages are deployed immediately upon being merged into master. You can read more above.

Known problems

ENFILE or EMFILE

If you get ENFILE: file table overflow, ENFILE: too many open files or any other ENFILE or EMFILE errors on Mac, that means you are doing synchronous file system operations. Cypress should NEVER do them. Instead we should use async file system operations and let graceful-fs retry them. Find the place where the synchronous fs operation is done from the stacktrace and make it async.

lock file

You can rebuild the lock file using the latest develop version in a clean isolated environment using Docker. From the current branch:

git checkout develop
git pull
git checkout -
git checkout origin/develop -- yarn.lock
# remove all unknown files
git clean -xfd
yarn