Hi, I'm Ethan. I got interested in designing programming langauges since reading Bob Nystorm's Crafting Interpreters many years ago. This project is a complete overhaul/rewrite of my compiler infrastructure for my language pulsar. Feel free to explore!

I am building this compiler from scratch, that is,

using only POSIX and two libraries of my own making (mtrack and libtermcolor)
without parser/lexer generators
without LLVM
in C

In other words, my aim is to gain as much experience with the workings of every component in a compiler, NOT to create a useful language for others.

This rewrite is still very much a work in progress. Much of the code has yet to be written.

Overview

This project is not the complete compiler. Rather, it is merely the frontend-to-optimized-IR pipeline. libpulsar can load various backends (called "modules") that take in this optimized IR and produce output of varying formats. Obviously, the first thing I will do is write a VM to interpret the IR, but the main goal is to produce object files natively. (This is nontrivial on macOS; see my posts on how I managed to figure this out through reverse engineering.)

Project Components

The Frontend contains the lexer, parser, and static analyzer
The Backend handles IR generation optimization
The Data Structures include hash maps and [wip]
The Utilities manage memory and program state
The Error system enables reporting and displaying errors
The Testing Infrastructure ensures that the compiler is working properly

Website Credits

This website was built using Doxygen. The theme was adapted from Joe Wong's doxygen_theme_flat_design. It is licensed under the MIT license. My modified source is available here.