Rationale

UCC is a hand-written implementation aiming to support the full C11 grammar specification. This is a hobby project with the goal to dive in head-first into compiler theory without any solid foundation, and see where I come out by the end.

Building & usage

The target is built with the Makefile build system:

git clone --depth=1 https://github.com/unazed/ucc && cd ucc/
make

Then the target binary is named ucc, argument usage is provided by ucc -?. Multiple compilation units are supported.

Milestones

1. Tokenization
   • Lexing
     • Raw strings (commit)
     • Strings w/ encodings (UTF8, UTF16, UTF32 and wide encodings) (commit)
       • Octal/hexadecimal sequences (commit)
       • TODO: Universal escape sequences (unlikely to do, snorefest)
     • Floating point numbers (and hexadecimal floats) (commit)
     • Decimal numbers (commit), octal numbers (commit), hexadecimal numbers (commit), with suffixes supported
     • Identifiers (commit)
     • Unary/binary operators (commit)
     • Include directives/multiline macros (commit)
     • Dumping lexer output to file (commit)
   • Rich diagnostics (commit)
   • Multiple translation units (commit)
   • Regression testing (commit)
2. Preprocessing
3. Parsing
4. Code generation (LLVM backend)