- Compile the project (described in the next section)
- Copy all .dll and .exe files from
build/
to your game directory - Copy the contents of
data/ship/
to your game directory
-
With Docker:
Make sure to install Docker and just, then run
just
. The binaries should appear in thebuild/
directory. To see list of possible build targets, runjust -l
. -
Without Docker:
This scenario is not officially supported, but you can see how it's done by examining the files in the
tools/docker/
directory for the external dependencies andmeson.build
for the local files, then tailoring your system to match the process.
-
Using WSL:
Install WSL (video guide: https://www.youtube.com/watch?v=5RTSlby-l9w)
- Run Powershell as Administrator
- Copy and paste the following command:
Enable-WindowsOptionalFeature -Online -FeatureName Microsoft-Windows-Subsystem-Linux
- Restart the computer
- Go to Windows App Store
- Install Ubuntu
Run WSL and continue with the instructions from the
Compiling on Ubuntu
section.
Please be advised that any build systems that are not the one we use for automating releases (= mingw-w64) come at user's own risk. They might crash or even refuse to compile.
- Compatibility with the original game's look and feel
- Maintainability
- Automation where possible
- Documentation (git history and GitHub issues are great for this purpose)
This project uses pre-commit to make sure the code
is formatted the right way. This tool has additional external dependencies:
clang-format
for automatic code formatting. To install pre-commit:
python3 -m pip install --user pre-commit
pre-commit install
To install required external dependencies on Ubuntu:
apt-get install -y clang-format-18
After this, each time you make a commit a hook should trigger to automatically
format your changes. Additionally, in order to trigger this process manually,
you can run just lint-format
. This doesn't include the slowest checks that
would hinder productivity – to run the full process, you can run just lint
.
If for any reason you can't install the above software on your machine, our CI
pipeline will still show what needs to be changed in case of mistakes.
While the existing source code does not conform to the rules because it uses the original Core Design's naming, new code should adhere to the following guidelines:
- Variables are
lower_snake_case
- Global variables are
g_PascalCase
- Module variables are
m_PascalCase
- Function names are
Module_PascalCase
- Macros are
UPPER_SNAKE_CASE
- Struct names are
UPPER_SNAKE_CASE
- Struct members are
lower_snake_case
- Enum names are
UPPER_SNAKE_CASE
- Enum members are
UPPER_SNAKE_CASE
Try to avoid global variables, if possible declare them as static
in the
module you're using them. Changes to original game functionality most often
should be configurable.
Other things:
-
We use clang-format to automatically format the code
-
We do not omit
{
and}
-
We use K&R brace style
-
We condense
if
expressions into one, so:if (a && b) { }
and not:
if (a) { if (b) { } }
If the expressions are extraordinarily complex, we refactor these into smaller conditions or functions.
We commit via pull requests and avoid committing directly to develop
, which
is a protected branch. Each pull request gets peer-reviewed and should have at
least one approval from the code developer team before merging. We never merge
until all discussions are marked as resolved and generally try to test things
before merging. When a remark on the code review is trivial and the PR author
has pushed a fix for it, it should be resolved by the pull request author.
Otherwise we don't mark the discussions as resolved and give a chance for the
reviewer to reply. Once all change requests are addressed, we should re-request
a review from the interested parties.
We keep a changelog in CHANGELOG.md
. Anything other than an internal change
or refactor needs an entry there. Furthermore, new features and OG bugfixes
should be documented in README as well.
Either you can make a lot of throwaway commits such as 'test' 'continue testing' 'fix 1' 'fix 2' 'fix of the fix' and then squash your pull request as a whole, or you can craft a nice history with proper commit messages and then merge-rebase. The first case is mostly acceptable for isolated features, but in general we prefer the latter approach. As a principle, refactors should made in separate commits. Code review changes are best made incrementally and then squashed prior to merging, for easing the review process.
The most important thing to remember: bug fixes and feature implementations
should always include the phrase Resolves #123
. If there's no ticket and the
pull request you're making contains player-facing changes, a ticket needs
to be created first – no exceptions.
Anything else is just for consistency and general neatness. Our commit messages aim to respect the 50/72 rule and have the following form:
module-prefix: description in an imperative mood (max 50 characters)
Longer description of what happens that can span multiple lines. Each
line should be maximally 72 characters long, with the exceptions of
code/log dumps.
The prefix should describe the module that the pull request touches the most.
In general this is the name of the .c
or .h
file with the most changes.
Note that this includes the folder names which are separated with /
. Avoid
underscores (_
) in favor of dashes (-
).
The description should be as concise as possible; any details should be given
in the commit message body. Use simple, to the point words like add
, fix
,
remove
, improve
.
Good:
ui: improve resolution changing
Added the ability for the player to switch resolutions directly from
the game ui.
Resolves #123.
Great:
log: fix varargs for Log_Message()
On Linux, the engine crashes when printing the log messages. This
happens because the current code re-uses the same va_list variable on
two calls to vprintf() and vfprintf(). Actually, this is not allowed.
For using the same information on multiple formatting functions, it is
needed to create a copy of the primary va_list to a second one, by using
va_copy(). After rewriting properly the Log_Message() function, the
segmentation fault is gone. Tested on both Linux and Windows builds.
- This has no ticket number, but it was an internal change improving support for a platform unsupported at that time, which made it acceptable.
Bad:
ui: implemented the ability to switch resolutions from the ui
- the subject doesn't use imperative mood
- the subject is too long
- missing ticket number
Bad:
dart: added dart emitters to the savegame (#779)
dart: added dart emitters to the savegame
Add function for checking legacy savegame save flags
Resolves #774.
- it duplicates the subject in the message body
- the subject doesn't use imperative mood
When merging via squash, it's OK to have GitHub append the pull request number, but pay special attention to the body field which often gets filled with garbage.
We have two branches: develop
and stable
. develop
is where all changes
about to be published in the next release land. stable
is the latest release.
We avoid creating merge commits between these two – they should always point to
the same HEAD when applicable. This means that any hotfixes that need to be
released ahead of unpublished work in develop
are merged directly to
stable
, and develop
needs to be then rebased on top of the now-patched
stable
.
We try to code all of our internal tools in a reasonably recent version of Python. Avoid bash, shell and other similar languages.
New version releases happen automatically whenever a new tag is pushed to the
stable
branch with the help of GitHub actions. In general this is accompanied
with a special commit docs: release X.Y.Z
that assigns unreleased changes to
a specific version. See git history for details.
- OG: original game
- UK Box: the version released on discs in the UK
- Multipatch: the version released on Steam
- PS: PlayStation version of the game