Files
sublime-music/CONTRIBUTING.rst
2020-07-11 13:01:59 -06:00

205 lines
7.4 KiB
ReStructuredText

Contributing
############
Contributions are welcome! Please create an issue, submit a MR, or engage on our
`Matrix chat`_.
.. _Matrix chat: https://matrix.to/#/!veTDkgvBExJGKIBYlU:matrix.org?via=matrix.org
Issue Reporting
===============
You can report issues or propose features using the GitLab Issues feature for
this repository: https://gitlab.com/sublime-music/sublime-music/issues.
Please note that as of right now, I (Sumner) am basically the only contributor
to this project, so my response time to your issue may be anywhere from instant
to infinite.
When reporting a bug, please be as specific as possible, and include steps to
reproduce. Additionally, you can run Sublime Music with the ``-m`` flag to
enable logging at different levels. For the most verbose logging, run Sublime
Music with ``debug`` level logging::
sublime-music -m debug
This may not be necessary, and using ``info`` may also suffice.
Code
====
If you want to propose a code change, please submit a merge request. If it is
good, I will merge it in.
To get an overview of the Sublime Music code structure, I recommend taking a
look at the |docs|_.
.. |docs| replace:: ``sublime`` package documentation
.. _docs: https://sublime-music.gitlab.io/sublime-music/api/sublime.html
Requirements
------------
**WIP:** Please create an MR with any other dependencies that you had to
install to develop the app. In general, the requirements are:
- Python 3.8 (I recommend you install this via Pyenv_)
- GTK3
- GLib
Specific Requirements for Various Distros/OSes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
* **Arch Linux:** `pacman -S libnm-glib libnotify python-gobject`
* **macOS (Homebrew):** `brew install mp3 gobject-introspection pkg-config pygobject3 gtk+3 adwaita-icon-theme`
Installing
----------
This project uses a ``Pipfile`` for managing dev dependencies. Make sure that
you have Pipenv_ (and Pyenv_ if necessary) set up properly, then run::
$ pipenv install --dev
to install the development dependencies as well as install ``sublime-music``
into the virtual environment as editable.
.. _Pipenv: https://pipenv.readthedocs.io/
.. _Pyenv: https://github.com/pyenv/pyenv
Running
-------
If your Pipenv virtual environment is activated, the just run::
$ sublime-music
to run the application. If you do not want to activate the pipenv virtual
environment, you can use::
$ pipenv run sublime-music
Building the flatpak
--------------------
- A flatpak-builder environment must be setup on the build machine to do a
flatpak build. This includes ``org.gnome.SDK//3.34`` and
``org.gnome.Platform//3.34``.
- The ``flatpak`` folder contains the required files to build a flatpak package.
- The script ``flatpak_build.sh`` will run the required commands to grab the
remaining dependencies and build the flatpak.
- You can install the Flatpak using: ``flatpak install sublime-music.flatpak``
and run it using ``flatpak run app.sublimemusic.SublimeMusic``.
Code Style
----------
This project follows `PEP-8`_ **strictly**. The *only* exception is maximum line
length, which is 88 for this project (in accordance with ``black``'s defaults).
Lines that contain a single string literal are allowed to extend past the
maximum line length limit.
This project uses flake8, mypy, and black to do static analysis of the code and
to enforce a consistent (and as deterministic as possible) code style.
Although you can technically do all of the formatting yourself, it is
recommended that you use the following tools (they are automatically installed
if you are using pipenv). The CI process uses these to check all commits, so you
will probably want these so you don't have to wait for results of the build
before knowing if your code is the correct style.
* `flake8`_ is used for linting. The following additional plugins are also used:
* ``flake8-annotations``: enforce type annotations on function definitions.
* ``flake8-bugbear``: enforce a bunch of fairly opinionated styles.
* ``flake8-comprehensions``: enforce usage of comprehensions wherever
possible.
* ``flake8-importorder`` (with the ``edited`` import style): enforce ordering
of import statements.
* ``flake8-pep3101``: no ``%`` string formatting.
* `mypy`_ is used for type checking. All type errors must be resolved.
* `black`_ is used for auto-formatting. The CI process runs ``black --check`` to
make sure that you've run ``black`` on all files (or are just good at manually
formatting).
* ``TODO`` statements must include an associated issue number (in other words,
if you want to check in a change with outstanding TODOs, there must be an
issue associated with it to fix it).
* ``print`` statements are not allowed. Use the more powerful and useful
``logging`` library instead. In the rare case that you actually want to print
to the terminal (the ``--version`` flag for example), then just disable this
check with a ``# noqa`` or a ``# noqa: T001`` comment.
.. _black: https://github.com/psf/black
.. _`PEP-8`: https://www.python.org/dev/peps/pep-0008/
.. _mypy: http://mypy-lang.org/
The CI process uses Flake8 and MyPy to lint the Python code. The CI process also
checks for uses of ``print``. You can run the same checks that the lint job runs
yourself with the following commands::
$ flake8
$ mypy sublime tests/**/*.py
$ black --check .
$ ./cicd/custom_style_check.py
Testing
-------
This project uses ``pytest`` for testing. Tests can be added in the docstrings
of the methods that are being tested or in the ``tests`` directory. 100% test
coverage is **not** a goal of this project, and will never be. There is a lot of
code that just doesn't need tested, or is better if just tested manually (for
example most of the UI code).
Simulating Bad Network Conditions
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
One of the primary goals of this project is to be resilient to crappy network
conditions. If you have good internet, you can simulate bad internet with the
``REQUEST_DELAY`` environment variable. This environment variable should be two
values, separated by a ``,``: the lower and upper limit for the delay to add to
each network request. The delay will be a random number between the lower and
upper bounds. For example, the following will run Sublime Music and every
request will have an additional 3-5 seconds of latency::
REQUEST_DELAY=3,5 sublime-music
CI/CD Pipeline
--------------
This project uses a CI/CD pipeline for building, testing, and deploying the
application to PyPi. A brief description of each of the stages is as follows:
``build-containers``
* Jobs in this stage are run every month by a Job Schedule. These jobs build
the containers that some of the other jobs use.
``test``
* Lints the code using ``flake8`` and ``mypy`` and prevents the use of
``print`` except in certain cases.
* Runs unit tests and doctests and produces a code coverage report.
``build``
* Builds the Python dist tar file
* Builds the flatpak.
``deploy``
* Deploys the documentation to GitLab pages. This job only runs on
``master``.
* Deploys the dist file to PyPi. This only happens for commits tagged with a
tag of the form ``v*``.
``verify``
* Installs Sublime Music from PyPi to make sure that the raw install from
PyPi works.
``release``
Creates a new `GitLab Release`_ using the content from the most recent
section of the ``CHANGELOG``.
.. _GitLab Release: https://gitlab.com/sublime-music/sublime-music/-/releases