If you are looking for issues to work on, have a look at the milestones and see which one is the closest to release then look at the tickets targeted at this release.
Don't forget that Lutris is not only a desktop client, there are also a lot of issues to work on on the website and also in the build scripts repository where you can submit bash scripts for various open source games and engines we do not already have.
Another area where users can help is confirming some issues that can't be reproduced on the developers setup. Other issues, tagged need help might be a bit more technical to resolve but you can always have a look and see if they fit your area of expertise. Also, while not fully ready, we do appreciate receiving translations for other languages, support for i18n will come in a future update.
Note that Lutris is not a playground or a toy project. One cannot submit new features that aren't on the roadmap and submit a pull request for them without agreeing on a design first with the development team. You can submit such pull requests but they will have to be updated with necessary changes from code reviews of the development team. This is not only about writing good code but also about providing good design for the overall product. Make sure to post all the relevant information in a ticket or on the pull request. New features must at all times have a valid use case based on an actual game, be very specific about why you are implementing a feature otherwise it will get rejected. Avoid adding options in the GUI or introducing new installer directives for things that can be automated. Lutris focuses heavily on automation and on doing the right thing by default. Only introduce new option when absolutely necessary.
Contributors are welcome to suggest architectural changes or better code design if they feel like the current implementation should be improved but please take note that we're trying to stay as lean as possible. Requests introducing complex architectural changes for the sake of "modularity", "Unix pureness" or subjective aspects might not be received warmly. There are no plans for any rewrite in another language or switching to another toolkit.
Running Lutris from a local git repository is easy, it only requires cloning the repository and executing Lutris from there.
git clone https://github.com/lutris/lutris
cd lutris
./bin/lutris -d
Make sure you have all necessary dependencies installed. It is recommended that
you keep a copy of the stable version installed with your package manager to
ensure that all dependencies are available.
If you are working on newly written code that might introduce
new dependencies, check in the package configuration files for new packages to
install. Debian based distros will have their dependencies listed
in debian/control and RPM based ones in lutris.spec.
Under NO circumstances should you use a virtualenv or install dependencies with pip. The PyGOject introspection libraries are not regular python packages and it is not possible for pip to install them or use them from a virtualenv. Make sure to always use PyGOject from your distribution's package manager. Also install the necessary GObject bindings as described in the INSTALL file.
To ensure getting your contributions getting merged faster and to avoid other developers from going back and fixing your code, please make sure your code passes pylint checks. We highly recommend that you install a pylint plugin in your code editor. Once you have pylint set up to check the code, you can configure it to use 120 characters max per line instead of 80.
You can help fixing formatting issues or other code smells by having a look at the CodeFactor page: https://www.codefactor.io/repository/github/lutris/lutris
When writing docstrings, you should follow the Google style (See: https://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html) You should always provide docstrings, otherwise your code wouldn't pass a Pylint check.
Do not add type annotations, those are not supported in Python 3.4.
If your patch does not require interactions with a GUI or external processes,
please consider adding unit tests for your code. Have a look at the existing
test suite in the tests folder to see what kind of features are tested.
Be sure to test your changes thoroughly, never submit changes without running
the code. At the very least, run the test suite and check that nothing broke.
You can run the test suite by typing make test in the source directory.
In order to run the test, you'll need to install nosetests and flake8:
pip3 install nose flake8
It is very important that any of your changes be tested manually, especially if you didn't add unit tests for the patch. Even trivial changes should be tested as they could potentially introduce breaking changes from a simple oversight.
Make a new git branch based of master in most cases, or next if you want to
target a future release. Send a pull request through GitHub describing what
issue the patch solves. If the PR is related to and existing bug report, you
can add (Closes #nnn) or (Fixes #nnn) to your PR title or message, where
nnn is the ticket number you're fixing. If you have been fixing your PR with
several commits, please consider squashing those commits into one with git rebase -i.
Lutris uses Python 3 and GObject / Gtk+ 3 as its core stack, here are some links to some resources that can help you familiarize yourself with the project's code base.
- Python 3 documentation
- PyGObject documentation
- Python Gtk 3 tutorial
- Fakegir GObject code completion
[root]-+ Config files and READMEs
|
+-[bin] Main lutris executable script
+-[debian] Debian / Ubuntu packaging configuration
+-[docs] User documentation
+-[lutris]-+ Source folder
| |
| +-[gui] Gtk UI code
| +-[installer] Install script interpreter
| +-[migrations] Migration scripts for user side changes
| +-[runners] Runner code, detailing launch options and settings
| +-[services] External services (Steam, GOG, ...)
| +-[util] Generic utilities
|
+-[po] Translation files
+-[share] Lutris resources like icons, ui files, scripts
+-[tests] Unit tests