Part 8: Finishing

Adding a License

FreedomBox is licensed under the GNU Affero General Public License Version 3 or later. FreedomBox apps, which run as modules under FreedomBox Service (Plinth), also need to be under the same license or under a compatible license. The license of our app needs to clear for our app to be accepted by users and other developers. Let us add license headers to our application.

# SPDX-License-Identifier: AGPL-3.0-or-later

The above header needs to be present in every file of the application. It is suitable for Python files. However, in template files, we need to modify it slightly.

{% extends "base.html" %}
{% comment %}
# SPDX-License-Identifier: AGPL-3.0-or-later
{% endcomment %}

Coding standards

For readability and easy collaboration it is important to follow common coding standards. FreedomBox uses the Python coding standards and uses the pylint and flake8 tools to check if the there are any violations. Run these tools on our application and fix any errors and warnings. Better yet, integrate these tools into your favorite IDE for on-the-fly checking.

For the most part, the code we have written so far, is already compliant with the coding standards. This includes variable/method naming, indentation, document strings, comments, etc. One thing we have to add are the module documentation strings. Let us add those. In add the top:

FreedomBox app to configure Transmission.

Contributing code to FreedomBox

The and files in the FredomBox source code have tips on how to contribute code to the project. Be sure to read them if you are submitting your app for including on the project.

Here is

# Hacking

## Requirements for Development OS

FreedomBox is built as part of Debian GNU/Linux. However, you don't need to
install Debian to do development for FreedomBox. FreedomBox development is
typically done on a Virtual Machine. You can work on any operating system that
can install latest versions of Git, Vagrant and VirtualBox.

### For Debian GNU/Linux and Derivatives

1. Install Git, Vagrant and VirtualBox using apt.

   $ sudo apt install git virtualbox vagrant

### For Other GNU/Linux Distributions or *BSDs

1. Install Git, Vagrant and VirtualBox using your favourite package manager.

### For macOS

1. Install [Brew](

2. Install Git, Vagrant and VirtualBox using Brew.

   brew install git
   brew cask install vagrant
   brew cask install virtualbox

### For Windows

1. Install [Git](,
   [VirtualBox]( and
   [Vagrant]( from their respective
   download pages.

2. Tell Git to use Unix line endings by running the following in Git Bash.

   git config --global core.autocrlf input

3. Run all the following commands inside Git Bash.

## Setting Up Development Environment Using Vagrant

Vagrant is a free software command line utility for managing the life cycle of
virtual machines. The FreedomBox project provides ready-made virtual machines
(VMs) for use with Vagrant. These images make setting up an environment for
FreedomBox development rather simple: You can edit the source code on your host
and immediately see the effects in the running VM. The entire setup is automatic
and requires about 4.5 GB of disk space.

1.  Checkout FreedomBox Service (Plinth) source code using Git.

    git clone
    cd plinth

2.  To download, setup, run, and configure a VM for FreedomBox development using
    Vagrant, simply execute in your FreedomBox Service (Plinth) development

    $ vagrant up

3.  SSH into the running vagrant box with the following command:

    $ vagrant ssh

4.  Run the development version of FreedomBox Service (Plinth) from your source
    directory in the virtual machine using the following command. This command
    continuously deploys your code changes into the virtual machine providing a
    quick feedback cycle during development.

    $ sudo -u plinth /vagrant/run --develop

5.  If you have changed any system configuration files during your development,
    you will need to run the following to install those files properly on to the
    system and their changes to reflect properly.

    $ sudo ./ install

Note: This development virtual machine has automatic upgrades disabled by

## Running Tests

To run all the tests:

$ py.test-3

Another way to run tests (not recommended):

$ ./ test

To run a specific test function, test class or test module, use pytest filtering
options. See pytest documentation for further filter options.


# Run tests in a directory
$ py.test-3 plinth/tests

# Run tests in a module
$ py.test-3 plinth/tests/

# Run tests of one class in test module
$ py.test-3 plinth/tests/

# Run one test in a class or module
$ py.test-3 plinth/tests/

## Running the Test Coverage Analysis

To run the coverage tool:

$ py.test-3 --cov=plinth

To collect HTML report:

$ py.test-3 --cov=plinth --cov-report=html

Invoking this command generates a HTML report to the `htmlcov` directory.
`index.html` presents the coverage summary, broken down by module. Data columns
can be sorted by clicking on the column header. Clicking on the name of a
particular source file opens a page that displays the contents of that file,
with color-coding in the left margin to indicate which statements or branches
were executed via the tests (green) and which statements or branches were not
executed (red).

## Running Functional Tests

### Install Dependencies

#### For running tests inside the VM

Run `vagrant provision --provision-with tests`.

#### For running tests on host machine

Follow the instructions below to run the tests on host machine. If you wish
perform the tests on host machine, the host machine must be based on Debian
Buster (or later).

$ pip3 install splinter
$ pip3 install pytest-splinter
$ sudo apt install python3-pytest-bdd
$ sudo apt install xvfb python3-pytest-xvfb  # optional, to avoid opening browser windows
$ sudo apt install firefox
$ sudo apt install smbclient # optional, to test samba

- Install the latest version of geckodriver. It is usually a single binary which
  you can place at /usr/local/bin/geckodriver . Geckodriver will use whichever
  binary is named 'firefox' for launching the browser and interacting with it.

### Run FreedomBox Service

*Warning*: Functional tests will change the configuration of the system
 under test, including changing the hostname and users. Therefore you
 should run the tests using FreedomBox running on a throw-away VM.

The VM should have NAT port-forwarding enabled so that 4430 on the
host forwards to 443 on the guest. From where the tests are running, the web
interface of FreedomBox should be accessible at https://localhost:4430/.

To run samba tests, port 4450 on the host should be forwarded to port 445
on the guest.

### Setup FreedomBox Service for tests

Via Plinth, create a new user as follows:

* Username: tester
* Password: testingtesting

This step is optional if a fresh install of Plinth is being tested. Functional
tests will create the required user using FreedomBox's first boot process.

### Run Functional Tests

**When inside a VM you will need to target the guest VM**

export FREEDOMBOX_URL=https://localhost FREEDOMBOX_SAMBA_PORT=445

You will be running `py.test-3`.

$ py.test-3 --include-functional

The full test suite can take a long time to run (more than an hour). You can
also specify which tests to run, by specifying a mark:

$ py.test-3 -m essential --include-functional
$ py.test-3 -m mediawiki --include-functional

If xvfb is installed and you still want to see browser windows, use the
`--no-xvfb` command-line argument.

$ py.test-3 --no-xvfb -m mediawiki --include-functional

## Building the Documentation Separately

FreedomBox Service (Plinth) man page is built from DocBook source in the `doc/`
directory. FreedomBox manual is downloaded from the wiki is also available
there. Both these are build during the installation process.

To build the documentation separately, run:

$ make -C doc

## Repository

FreedomBox Service (Plinth) is available from

## Bugs & TODO

You can report bugs on FreedomBox Service's (Plinth's) [issue

See for information how to best contribute code.

## Internationalization

To mark text for translation, FreedomBox Service (Plinth) uses Django's
translation strings. A module should e.g. `from django.utils.translation import
ugettext as _` and wrap user-facing text with `_()`. Use it like this:

message = _('Application successfully installed and configured.')

## Translations

The easiest way to start translating is with your browser, by using
Your changes will automatically get pushed to the code repository.

Alternatively, you can directly edit the `.po` file in your language directory
`Plinth/plinth/locale/` and create a pull request (see
In that case, consider introducing yourself on #freedombox IRC (,
because some work may have been done already on the [Debian translators
discussion lists](
or the Weblate localization platform.

For more information on translations:

## Application Icons

When adding a new App into FreedomBox, an icon is needed to represent the app in
the application view and for shortcuts in the front page. The following the
guidelines for creating an app icon:

- Use SVG format.
- Keep the size and complexity of the SVG minimal. Simplify the graphic if
- Units for the entire document should be in pixels.
- View area should be 512x512 pixels.
- Background should be transparent.
- Leave no margins and prefer a square icon. If the icon is wide, leave top and
  bottom margins. If the icon is tall, leave left and right margins.

Here is

# Contributing
Here are some contributing guidelines for authors and reviewers of code changes.
The goal is a readable log of code changes, to enhance transparency of their
purpose and simplify debugging. Consider these guidelines as best practices, not
as absolute rules - we're all learning by doing, and imperfect changes and
commits are much better than none at all. For an introduction how to edit and
test the code, have look into the file. Note that you need some basic
understanding of Git to contribute; there are many tutorials in the Internet
that we cannot repeat here.

Naming conventions:
* 'Code change', 'patch', and 'commit' are used interchangeably.
* 'Author' and 'contributor' are used interchangeably.
* Git 'log' and 'history' are used interchangeably.
* PR, MR - pull request and merge request, used interchangeably.
* 'Merging' often means 'applying a patch to git history' in a general sense,
  not literal execution of the command `git merge`.

# For authors of patches
* If you would like to contribute, but are unsure what to do, just ask. There
  are usually also issues tagged as 'Contributor Invite' or 'beginner', which
  might be a good starting point to work on and have a known solution. Also,
  other developers are ready to guide you on the implementation for such tasks.
  Feel free to pickup a task from the issue by announcing it on the issue or by
  creating a new issue for whatever task you are going to work on.
* To get your changes included, you must open a merge request (MR) and get them
  reviewed. Briefly, fork the repository to your account, create a branch, edit
  code, commit and push there. Then you can create a merge request on the main
* Before committing your changes ensure that your code conforms to base code
  hygiene by running tests (see and the automatic code formatting
  tools `yapf` and `isort`.
* Please include one single feature per merge request, to keep the review simple
  and focused on one topic. (This might still mean hundreds of lines of code.)
  Use a branch other than `master`, so you can create multiple merge requests
  and still keep merging from `master`. Depending on the complexity of your
  merge request, it may take a while until it is reviewed and merged.
* Keep your commits organized logically and as small as possible. If commit B
  fixes a mistake in commit A, both of which are part of the same merge request,
  combine them into a single commit. If commit A introduces a single logical
  change but breaks existing functionality and then commit B rectifies it, then
  also combine the two commits. This is to ensure that the source code can be
  checked out at any revision and used (such as during git bisect). If there are
  two unrelated changes in the same commit, split them to into separate commits.
  See Git documentation on how to merge, split and reorder commits.
* Please create meaningful commit messages, by following common guidelines:
    * Multiple lines are allowed if it makes the message clearer.
    * Separate the first subject line from the text body with a blank line.
	* Add the component you changed as the first word in the subject line.
    * Wrap the text at 72 characters.
    * Use the body to explain what your changes do, and maybe why and how it is
      achieved (the main idea).
	* Look into the git log to get an idea.
	* If it exists, mention the issue number.
	* End the message with a "Signed-off-by", see next entry.
* Consider adding `Signed-off-by: YOUR NAME <YOUR EMAIL>` into your commit
  message. With this, you explicitly certify that you have the rights to submit
  your work under the project's license (see LICENSES file) and that you agree
  to a [Developer Certificate of Origin](
* If (part of) your code changes were inspired or plainly copied from another
  source, please indicate this in the merge request, so the reviewer can handle
* If your merge request is not ready for merging yet, the title of your merge
  request must start with
* If a reviewer asks for changes to your merge request, perform the changes as
  requested or provide clarification. Close the discussion threads so that the
  reviewer knows that it is ready for another round of review.
* Newer changes addressing review comments should go into the old commits which
  are being changed. For example, if there is a security problem with one of
  your commits, the commit should be edited instead of introducing a new commit
  with fix for it. After a merge, if a developer checks out any revision, it
  could not contain serious problem. See Git documentation on how to merge,
  split, reorder commits and how to force push your branches after altering
* Have fun contributing :)

# For reviewers of patches

## How to review the work of others
* Be nice to contributors and give them opportunities to learn. Explain the
  reasons if you ask for changes instead of silently changing things yourself
  (unless trivial). This also saves you time in the future.
* Reviewers are expected to ensure that a contributor's work:
    * Does not break the current code. The code base should always be in a
	  usable state, without throwing (non-handled) errors. We also strive to
	  keep the code base in a release-worthy state.
    * Has no security issues.
    * Follows coding standards of the project
	  (mainly [PEP-8](
    * Is properly internationalized.
* Your main job is to make sure that the work runs as expected, by thoroughly
  testing the patch. New authors are usually not familiar with all areas of
  impact and may not have tested all cases. It is okay to rely on tests done by
  trusted authors, if they specify the specific cases they tested.
* When merging work from others, add this line to the commit message:
  `Reviewed-by: YOUR NAME <YOUR EMAIL>`.
* Some patches require knowledge of multiple technologies. If you are not
  familiar with all of them, it's fine to review only the portions you
  understand (and indicate them clearly). Then ask others for further review.
* For major architectural changes/decisions, consult others in the project
  before merging.
* You may make some minimal or obvious changes to the work before merging. If
  so, tell the contributor (and others) about your edits.
* In case more fundamental changes are necessary, or if the contributor is new,
  try to encourage them to make changes by giving appropriate feedback. This is
  a major way how we mentor new contributors.
* Any merge request whose title starts with `WIP:` cannot be merged. Communicate
  with the author on what the pending changes are. Get the author to complete
  them or complete them yourself in case of an emergency.
* Have fun reviewing :)

## Use of Git
* Avoid plain `git merge`, and use rebasing instead, to avoid merge commits.
  This keeps the history much more readable for others to understand, review,
  bisect and rollback.
* When merging patches with multiple commits, try to make each patch meaningful.
  For example, fixes to patches should be squashed into the original patch.
  Security issues and other serious issues should be avoided in intermediate
  patches – even if they are fixed in later patches.

## Use of GPG
* Sign all commits with GPG. This means avoiding GitLab's fancy merge and rebase
  buttons and doing it locally, where your private key is.
* In case a contributor signed with GPG, rebasing will strip it away. To
  compensate, put your GPG signature on the rebased commits. Given that we have
  to actually verify the signatures on each commit and the contributor may not
  be in our web of trust (as we allow anonymous contributions). In such cases,
  GPG signatures of the reviewers are more important.
* Get your GPG key signed by other reviewers or Debian Developers, to enable
  verification of your signed commits.

## (When) can I commit without review?
* Even if you have commit access, you should get your patches reviewed by the
  other reviewers.
* Certain patches do not require review. These include updating the manual, typo
  fixes, and creating new locales.