Capstone ProjectAdvanced

Build a Tiny C Compiler

This capstone project takes you through the complete journey of building a C compiler for a subset of C89. You will learn lexing, recursive-descent parsing, type checking with symbol tables, intermediate representation design, and x86-64 code generation. By the end, you will have a working compiler that takes C source files and produces executable binaries.

6-8 weeks 5 milestones 20 build steps, 20 linked tasks

The whole plan below is free to read: every milestone, its build steps, how to verify it, and which tasks and visualizations to use. Tracking your progress, the completion certificate and the graded tasks it links to need a plan.

Related Tracks

Language FoundationsCompilers

Prerequisites

C Programming Deep Dive
Lexical Analysis
Syntax Analysis
Semantic Analysis
Code Generation

Milestones

Each milestone opens into the build order, a way to check your work, the tasks that drill its pieces, and the visualizations that show the machinery. Work top to bottom.

Your compiler reads a C file and prints its syntax tree.

Build it in this order

1Define a token: kind, lexeme, line and column. Keep the location on every token, because every later error message needs it.
2Write next_token() as a hand-rolled scanner. Skip whitespace and comments, then dispatch on the first character; recognize keywords by looking identifiers up in a fixed table.
3Parse declarations and statements by recursive descent, and expressions by precedence climbing, so you do not need a separate rule per precedence level.
4Allocate AST nodes from an arena and write a pretty printer for the tree.

How you know it works

Pretty-print the AST of a 50-line C file, re-parse that output, and print it again: the two printouts are identical.

Practise the pieces first

Basic Lexer ImplementationFree Token Struct DefinitionFree Recursive Descent ParserFree AST Node StructuresFree

See how it works

Tokenizer Recursive descent parser Abstract syntax tree

Build it in your own repo

There is no upload step. You build this on your own machine, in your own repository, and tick each milestone off here as it works. Finish all 5 and you get a certificate with a verification id; connect GitHub first and a gist artifact is created alongside it.

Start with the first task Browse the prerequisite tracks