Install bnfc with Homebrew, apt, pacman

Project links

• Homepage
• Upstream docs
• Package manager page

Commands and aliases

• bnfc

Project history

BNFC started at Chalmers and the University of Gothenburg. A 2004 Haskell workshop demonstration by Markus Forsberg and Aarne Ranta says Forsberg and Ranta began development in 2002 as a tool generating Haskell.

The same source records that Michael Pellauer retargeted BNFC to C, C++, and Java in 2003, moving it from a Haskell-only generator toward a multilingual compiler-front-end generator.

Modern BNFC documentation describes the LBNF formalism independently of any one output language, and the project now advertises generation for Haskell, Agda, C, C++, Java, OCaml, XML representations, and Pygments syntax highlighters.

Adoption history

BNFC found adoption in programming-languages courses, research prototypes, and small language implementations where the grammar is the most stable artifact and hand-written front ends would be repetitive.

Its package-manager footprint is narrower than mainstream parser generators but meaningful: Homebrew, Debian/Ubuntu, Arch, and Haskell package ecosystems carry it for developers who want an executable generator rather than a library-only parser toolkit.

How it is used

A user writes an LBNF grammar, runs bnfc, and then compiles the generated lexer/parser and abstract-syntax code with the target language's usual parser tooling, such as Alex and Happy for Haskell or Flex/Bison-style tools for C-family output.

BNFC is especially useful when the same grammar should produce both implementation artifacts and a readable language specification, keeping parser code, AST constructors, pretty printers, and documentation aligned.

Why package nerds care

BNFC is package-nerd-significant because it packages an academic compiler-construction idea as a practical command-line generator. It is not just yacc with another syntax; its labeled grammar rules become typed abstract syntax across target languages.

For language-tooling nerds, BNFC sits in the interesting space between parser generators, syntax-specification formalisms, and teaching tools. It makes a grammar file the source of truth for several generated artifacts that are otherwise easy to let drift.

Its long survival in package managers matters because many compiler-course and research tools depend on old but useful grammar generators being installable without reconstructing a Chalmers-era toolchain by hand.

Timeline

• 2002: Forsberg and Ranta start BNFC as a Haskell-generating tool.
• 2003: BNFC is retargeted to C, C++, and Java by Michael Pellauer.
• 2004: BNFC is demonstrated at the ACM SIGPLAN Haskell workshop.
• 2010s: BNFC migrates into modern package managers and public GitHub maintenance.
• 2020s: BNFC documentation lists target-language support including Haskell, Agda, C, C++, Java, and OCaml.

Related projects

• Alex and Happy: Haskell lexer and parser tools often used with BNFC's Haskell backend.
• Flex and Bison: traditional lexer/parser tools related to BNFC's C/C++ style generated front ends.
• ANTLR, yacc, and Menhir: parser-generator neighbors with different grammar models and target ecosystems.
• LBNF: the Labelled BNF grammar formalism used by BNFC.

Sources

• https://bnfc.digitalgrammars.com/
• https://bnfc.digitalgrammars.com/tutorial/bnfc-tutorial.html
• https://bnfc.readthedocs.io/_/downloads/en/v2.9.6.1/pdf/
• https://bnfc.readthedocs.io/en/latest
• https://github.com/BNFC/bnfc
• https://github.com/BNFC/bnfc/blob/master/README.md
• https://www.cse.chalmers.se/alumni/markus/BNFC/BNFC_HW2004.pdf