Savi is a fast language for programmers who are passionate about their craft.
Savi will change the way you think about programming and equip you with the skills and tools you need to tackle more ambitious technical challenges, and have fun doing it.
Savi is here to help you write blazing-fast, concurrent software that is memory-safe and data-race-free.
Savi is an actor-oriented programming language using the Pony runtime. Like Pony, Savi has a unique type system that enforces concurrency-safety and memory-safety properties at compile time.
Like many other modern compiled languages, Savi uses LLVM to compile to a wide variety of native targets.
Our goal is to make Savi approachable and fun, as well as powerful and extensible.
We are a small team of passionate volunteers working to bring this project to full fruition. If this vision sounds interesting to you, we'd love for you to reach out in our chat and get involved.
The easiest way to try Savi is to use asdf version manager, which allows you to install and manage versions for a variety of languages/runtimes, including Savi. After installing asdf, run the following commands to get Savi up and running:
# Add the Savi plugin to asdf. asdf plugin add savi https://github.com/savi-lang/asdf-savi.git # Download and install the latest version of Savi. asdf install savi latest # Select the latest version of Savi as the one to use globally in your shell. asdf global savi latest # Prove the installation with a simple program. savi eval 'env.out.print("Savi is installed!")'
Once Savi is installed, you can enter any directory with a
manifest.savi file in it and run the
savi command there to compile one of the manifests defined in it, resulting in an executable binary program in a
./bin folder with the same name as that targeted manifest.
Alternatively, you can invoke
savi run to both compile and immediately run the program.
savi --help to learn about more available commands.
If you're a VS Code user, you may be interested to install our language extension for that editor, which includes both syntax highlighting and some Intellisense features via using the docker image as an LSP server.
We also have a vim/nvim extension as well.
Finally, if you want to contribute to Savi, read on through the next two sections for information on how to find work, as well some of the basic development commands.
Looking for ways to help? Here's a link that shows issue tickets filtered by those that should be ready to work on.
You can also filter by complexity label to try to find something that's the right level of challenge and time commitment for you.
If you're a new contributor looking for some guidance, please reach out to us in this chat room to introduce yourself and even schedule pairing sessions where we can help you understand parts of the compiler or language that you're interested in learning more about.
We think you'll find it exciting to join us at this stage of our work, where there is already enough working that you can get things done and produce a tangible result, but it's early enough that you can have a strong impact on the future of the language and the community.
We are excited to welcome all contributors that bring a positive attitude, regardless of their level of experience. Join us!
To work on this project and build the Savi compiler from source, you'll need
crystal installed (look at
shard.yml to see the correct version of Crystal that is officially supported). You may also want to have
lldb for debugging.
To get started, clone the project to your development machine, then run one of the following commands within the project working directory, or consult the
Makefile for more available commands. Note that the first time you run one of these commands it will take longer, as it will need to download some static dependencies and build the compiler for the first time.
make spec.allto run the full test suite, or refer the Makefile for information about runnning more specific sets of specs.
make format.checkto check
*.savisource files for formatting rule violations, or
make formatto fix them automatically.
make example dir="/opt/code/examples/adventofcode/2018to compile and run from the sources in
./examples/adventofcode/2018directory (or similarly for any other example code directory).
Alternatively, you can develop within the confines of a docker container, so that no development dependencies are needed other than
docker itself. To do so, first run
docker/make ready to run up a development container, then use
docker/make instead of
make to run any of the commands mentioned above (for example,
docker/make test), which will run that command inside of the development container.