writing-a-c-compiler

This repo contains the code for the C compiler project for the Writing a C Compiler book by Nora Sandler.

Cloning

You can clone this without the book's tests with

git clone git@github.com:shanebishop/writing-a-c-compiler.git

Or, with an appropriately modern version of Git, you can also clone the book's tests with

git clone --recurse-submodules git@github.com:shanebishop/writing-a-c-compiler.git

To pull submodules after an intial clone, run

git submodule update --init --recursive

Set up Git hooks

Git hooks should be set up right after cloning. To do so, run the following command in the root directory:

git config --local core.hooksPath .githooks/

Run tests

To run unit tests, run cargo test.

To run the book's tests, run ./run_book_tests.sh.

To run fuzzing, run cargo +nightly fuzz run fuzz-tests -- -max_total_time=2m. Adjust the fuzzing total time as desired.

Code organization

The code has been organized into the following crates:

• driver - the binary crate that has the compiler driver
• lexer - the lexer library
• parser - the parser library

This code organization has the following benefits:

• The different parts can be reused as libraries in an application that wants to perform compilation or parts of compilation without invoking a child process.
• Faster incremental build times. (Each Rust crate is compiled as a separate translation unit. This means if a change is internal to a particular crate, only that crate needs to be rebuilt. This has the consequence that some optimizations cannot be done at the compilation stage (but could be done at the link stage), but I favour the tradeoff for compile time, as this is a project for learning and not for production use.)