Skip to content

johannst/llvm-kaleidoscope-rs

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

32 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

llvm-kaleidoscope-rs

Rust Rustdoc

The purpose of this repository is to learn about the llvm compiler infrastructure and practice some rust-lang.

To reach the goals set, we follow the official llvm tutorial Kaleidoscope: Implementing a Language with LLVM. This tutorial is written in C++ and structured in multiple chapters, we will try to follow along and implement every chapter in rust.

The topics of the chapters are as follows:

The implementation after each chapter can be compiled and executed by checking out the corresponding tag for the chapter.

> git tag -l
chapter1
chapter2
chapter3
chapter4
chapter5

Names of variables and functions as well as the structure of the functions are mainly kept aligned with the official tutorial. This aims to make it easy to map the rust implementation onto the C++ implementation when following the tutorial.

One further note on the llvm API, instead of using the llvm C++ API we are going to use the llvm C API and build our own safe wrapper specialized for this tutorial. The wrapper offers a similar interface as the C++ API and is implemented in src/llvm/

Demo

# Run kaleidoscope program from file.
cargo run ks/<file>

# Run REPL loop, parsing from stdin.
cargo run

Documentation

Rustdoc for this crate is available at johannst.github.io/llvm-kaleidoscope-rs.

Build with provided container file

The provided Dockerfile documents the required dependencies for an ubuntu based system and serves as a build environment with the correct llvm version as specified in the Cargo.toml file.

## Either user podman ..

# Build the image *ks-rs*. Depending on the downlink this may take some minutes.
make -C container

podman run --rm -it -v $PWD:/work -w /work ks-rs
# Drops into a shell in the container, just use cargo build / run ...

## .. or docker.

# Build the image *ks-rs*. Depending on the downlink this may take some minutes.
make -C container docker

docker run --rm -it -v $PWD:/work -w /work ks-rs
# Drops into a shell in the container, just use cargo build / run ...

License

This project is licensed under the MIT license.