fyquah / bril

an educational compiler intermediate representation

Home Page:https://capra.cs.cornell.edu/bril/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Bril: A Compiler Intermediate Representation for Learning

Bril (the Big Red Intermediate Language) is a compiler IR made for teaching CS 6120, a grad compilers course. It is an extremely simple instruction-based IR that is meant to be extended. Its canonical representation is JSON, which makes it easy to build tools from scratch to manipulate it.

This repository contains the documentation, including the language reference document, and some infrastructure for Bril. There are some quick-start instructions below for some of the main tools, but check out the docs for more details about what's available.

Install the Tools

TypeScript Compiler & IR Interpreter

To install the TypeScript compiler and IR interpreter, you will need Node and Yarn. Go to the bril-ts directory and do this:

$ yarn
$ yarn build
$ yarn link

The last thing will install symlinks to the two utility programs, but they may not be in a standard location. To find where these tools were installed, run yarn global bin. You probably want to add this to your $PATH.

The tools are brili, an interpreter, which takes a Bril program as JSON on stdin, and ts2bril, which compiles a TypeScript file given on the command line to Bril.

Text Format

The parser & pretty printer for the human-editable text form of Bril are written for Python 3. To install them, you need Flit, so run this:

$ pip install --user flit

Then, go to the bril-txt directory and use Flit to install symlinks to the tools:

$ flit install --symlink --user

The tools are called bril2json and bril2txt. They also take input on stdin and produce output on stdout.

Tests

There are some tests in the test/ directory. They use Turnt, which lets us write the expected output for individual commands. Install it with pip:

$ pip install --user turnt

Then run all the tests by typing make test.

About

an educational compiler intermediate representation

https://capra.cs.cornell.edu/bril/

License:MIT License


Languages

Language:Rust 39.5%Language:JavaScript 23.4%Language:OCaml 15.0%Language:Python 10.3%Language:TypeScript 6.1%Language:Vim Script 2.8%Language:Shell 1.8%Language:Makefile 1.0%Language:Standard ML 0.1%