Contributions to Exodus are welcomed! Please ensure that the code you're contributing fits with the Design Philosophy of Exodus. If you're making a significant modification or addition, it might be worthwhile making contact on the forums first to check if your changes might overlap or be affected by other upcoming changes. The basic steps for making a code submission are as follows:
- Create an account on Bitbucket if you don't already have one
- Agree to the Contributor License Agreement, ensuring you enter your Bitbucket username. You only need to do this once, not per submission.
- Follow the standard procedure for Forking, Modifying, and Submitting a Pull Request on Bitbucket
Any code contributions to the Exodus project are greatly appreciated. This project can only meet its intended goals of supporting a wide variety of platforms through the contributions of others.
What is forking and how should I use it?
The easiest way to contribute code changes to Exodus is to "fork" the Exodus repository, and make the changes you want to make in that fork. When you're happy with the changes you've made, and you'd like them to be merged into the official repository, you submit a "pull request", and your changes can then be reviewed, and merged in once they pass the review. When you hear the word "fork", you may think that implies that you're breaking away from the Exodus project and starting your own independent project that will run forever into the future. This actually isn't how forks are intended to be used with hosted source repositories. The purpose of forks is to keep a relationship between the source repository and your fork of it, so that you can eventually merge the two together again through a pull request. Forks can be thought of as essentially being the same as "branches", but branches designed to work better with independent and disconnected programmers or programming groups.
Here are a few guidelines on how to make working with forks easier, and for your changes to have a better chance of being merged:
- Only do a single job per fork. Don't make multiple independent changes in a single fork, and then submit a pull request, because then all those changes need to be reviewed and passed together, or none of them can be accepted into the main repository. If you want to work on several changes in parallel, you can create multiple forks, one per job
- Take your time to test and verify your changes before submitting a pull request. You should make sure what you've written actually works the way it was intended.
- Don't start working on major changes that affect fundamental design aspects of the system, or involve extensive changes to large portions of the codebase, without discussing the changes first. There may be future plans, conflicting changes already in development, or concerns about the nature of the changes being made, which would affect the likelihood of your changes being merged into the main repository. Communication at the start will help avoid wasted effort and frustration.
- Make sure your changes fit the general coding conventions. There isn't a strict set of coding conventions for Exodus, but if you are adding a small portion of code to an existing file or library, please ensure the naming and formatting of your code is consistent with the area you're working in. This helps keep the codebase clean and readable.
- Merge often. Code changes may be submitted into the main repository after you create your fork. If your code changes require some time to complete, you should merge in changes from the main repository into your fork frequently, to ensure you're working on a current version of the code. If you don't merge as you go, it can make the merge process much more difficult when you want to submit your changes back to the main repository.
What is the purpose of the Contributor License Agreement?
A Contributor License Agreement is something that's vital for open source development, and there is an increasing awareness of its importance today. A Contributor License Agreement, or CLA, is designed to protect open source projects, and the contributors to those projects, from possible legal consequences and dilemmas that can and do arise. Without a CLA in place, there are very real liabilities that can have major consequences for contributors and open source projects themselves. Many of these issues are rare, and require malicious intent from someone deliberately trying to harm a project, but there are real life examples of many of these issues arising.
1. Liabilities for contributors
When someone submits code to an open source project without a CLA, the code they submit remains their personal responsibility. When the code is shared or distributed in source or binary form, they remain personally responsible for their contribution. If the code is distributed under some form of license agreement, every contributor of code to that project is entering into a direct and individual agreement with every user of that code. This also means they are personally responsible for any legal ramifications of that code. If the code they have written infringes on any patents, or if any individual alleges that it has infringed on any patents, or if there is a dispute about breach of copyright, or any other form of legal claim of damage or misuse, a legal dispute can be directed at the licensor (the author) by the licensee (the user). This means if an organization or individual had cause to file a legal dispute, they could, and in fact must by law, direct that dispute to individual code contributors to that project. In the event those contributors lost a legal challenge, they would be personally responsible for any damages awarded as a result.
2. Liabilites for the project
You may assume that when someone submits code changes to an open-source project, they've automatically given permission to the people maintaining that project to actually use the code they sent them, and integrate it in the main source repository, build a compiled program from it, modify it, and so on. Legally they haven't, unless that contributor made a formal statement to that effect. You may also assume that because someone contributes code to an open source project, the person who made the submission has given that project the right to use that code forever into the future. Legally, that also isn't the case unless they stated so. Without a CLA, 10 years after making a submission to a project, a contributor could decide he no longer wants his code to be included in the project, and formally demand it be removed. Legally it would be his right to do so. In this kind of case, his code contributions, and all derived works (things build from or using that code) must be stripped out of the code repository, and purged from its history. All releases of the software to include his code must be removed from distribution. Any sharing of that code as source or binary is no longer permitted. This could literally destroy open source projects if a significant former contributor turned against the project.
There is another liability for open source projects without a CLA, and that is, as described above, when the code is distributed in source or binary form, without a CLA it is actually each author entering an individual license agreement with each user. This also means that if there is a breach of the license agreement, it is the responsibility of each individual author to pursue legal action to enforce the license. Nobody else has the right to do this. Most contributors would not be willing to do this individually, and in the case of authors that are no longer involved in a project, or no longer contactable, this effectively gives anyone the ability to violate the license agreement for any code those authors have submitted without consequence, because nobody except the original authors can enforce the license.
The fundamental issue here is one of copyright law. Copyright law gives the original author of a work the rights to decide who is allowed to copy or alter their work, and the original author has the exclusive right to make those decisions. The original author also has the exclusive responsibility to ensure his work is used within the rights that he has granted to others. This works well where there is only single author for a work, but where an author is contributing their work as part of a larger work that is formed from the contributions of many others, it is essential that the group overseeing that larger work be granted some rights under copyright law. A Contributor License Agreement is the means by which a contributor to a project grants some rights and responsibilities for their contribution to the group managing that project. In any other place where authors publish their own work through some form of formal journal, paper, book, or similar media, they must agree to assign or grant some irrevocable rights to the publisher, or their submission will not be accepted. Open source software projects must have the same requirements if they are to protect themselves and their contributors legally.
The Contributor License Agreement that the Exodus project has adopted is the same license used by Apache and Google to govern contributions to their open source projects. It is unmodified except to reference the Exodus project where appropriate. This agreement in no way affects your rights to do whatever you want with the code you have written. What it does is ensure that once you've submitted code to this project, you can't change your mind about it later, nor are you legally responsible for what the project does with your code afterwards. It also ensures that the project is free to release your contributed code to other people as source and binary releases. Basically, it's formalizing exactly what you probably thought was already implied by submitting code changes to this project.