jitsi-meet/CONTRIBUTING.md

How to contribute

We would love to have your help. Before you start working however, please read and follow this short guide.

Reporting Issues

Provide as much information as possible. Mention the version of Jitsi Meet, Jicofo and JVB you are using, and explain (as detailed as you can) how the problem can be reproduced.

Code contributions

Found a bug and know how to fix it? Great! Please read on.

Contributor License Agreement

While the Jitsi projects are released under the Apache License 2.0, the copyright holder and principal creator is 8x8. To ensure that we can continue making these projects available under an Open Source license, we need you to sign our Apache-based contributor license agreement as either a corporation or an individual. If you cannot accept the terms laid out in the agreement, unfortunately, we cannot accept your contribution.

Creating Pull Requests

• Make sure your code passes the linter rules beforehand. The linter is executed automatically when committing code.
• Perform one logical change per pull request.
• Maintain a clean list of commits, squash them if necessary.
• Rebase your topic branch on top of the master branch before creating the pull request.

Coding style

Comments

• Comments documenting the source code are required.

  • Comments from which documentation is automatically generated are not subject to case-by-case decisions. Such comments are used, for example, on types and their members. Examples of tools which automatically generate documentation from such comments include JSDoc, Javadoc, Doxygen.

  • Comments which are not automatically processed are strongly encouraged. They are subject to case-by-case decisions. Such comments are often observed in function bodies.

• Comments should be formatted as proper English sentences. Such formatting pays attention to, for example, capitalization and punctuation.

Duplication

• Don't copy-paste source code. Reuse it.

Formatting

• Line length is limited to 120 characters.

• Sort by alphabetical order in order to make the addition of new entities as easy as looking a word up in a dictionary. Otherwise, one risks duplicate entries (with conflicting values in the cases of key-value pairs). For example:

  • Within an import of multiple names from a module, sort the names in alphabetical order. (Of course, the default name stays first as required by the import syntax.)

    import {
        DOMINANT_SPEAKER_CHANGED,
        JITSI_CLIENT_CONNECTED,
        JITSI_CLIENT_CREATED,
        JITSI_CLIENT_DISCONNECTED,
        JITSI_CLIENT_ERROR,
        JITSI_CONFERENCE_JOINED,
        MODERATOR_CHANGED,
        PEER_JOINED,
        PEER_LEFT,
        RTC_ERROR
    } from './actionTypes';
    

  • Within a group of imports (e.g. groups of imports delimited by an empty line may be: third-party modules, then project modules, and eventually the private files of a module), sort the module names in alphabetical order.

    import React, { Component } from 'react';
    import { connect } from 'react-redux';
    

Indentation

• Align switch and case/default. Don't indent the case/default more than its switch.

  switch (i) {
  case 0:
      ...
      break;
  default:
      ...
  }
  

Naming

• An abstraction should have one name within the project and across multiple projects. For example:

  • The instance of lib-jitsi-meet's JitsiConnection type should be named connection or jitsiConnection in jitsi-meet, not client.

  • The class ReducerRegistry should be defined in ReducerRegistry.js and its imports in other files should use the same name. Don't define the class Registry in ReducerRegistry.js and then import it as Reducers in other files.

• The names of global constants (including ES6 module-global constants) should be written in uppercase with underscores to separate words. For example, BACKGROUND_COLOR.

• The underscore character at the beginning of a name signals that the respective variable, function, property is non-public i.e. private, protected, or internal. In contrast, the lack of an underscore at the beginning of a name signals public API.

Feature layout

When adding a new feature, this would be the usual layout.

react/features/sample/
├── actionTypes.ts
├── actions.js
├── components
│   ├── AnotherComponent.js
│   ├── OneComponent.js
│   └── index.js
├── middleware.js
└── reducer.js

The middleware must be imported in react/features/app/ specifically in middlewares.any.ts, middlewares.native.ts or middlewares.web.ts where appropriate. Likewise for the reducer.

An index.js file must not be provided for exporting actions, action types and component. Features / files requiring those must import them explicitly.

This has not always been the case and the entire codebase hasn't been migrated to this model but new features should follow this new layout.

When working on an old feature, adding the necessary changes to migrate to the new model is encouraged.

Avoiding bundle bloat

When adding a new feature it's possible that it triggers a build failure due to the increased bundle size. We have safeguards inplace to avoid bundles growing disproportionatelly. While there are legit reasons for increasing the limits, please analyze the bundle first to make sure no unintended dependencies have been included, causing the increase in size.

First, make a production build with bundle-analysis enabled:

npx webpack -p --analyze-bundle

Then open the interactive bundle analyzer tool:

npx webpack-bundle-analyzer build/app-stats.json