bison

Generate parsers from grammar descriptions

TLDR

Generate parser

$ bison [grammar.y]

Generate with verbose output

$ bison -v [grammar.y]

Specify output file

$ bison -o [parser.c] [grammar.y]

Generate header file

$ bison -d [grammar.y]

bison is a general-purpose parser generator that converts grammar descriptions (in .y files) into C, C++, or Java parsers. It's compatible with yacc and used for building compilers, interpreters, and other language processors.
The tool is a GNU replacement for yacc with additional features and better error handling.

PARAMETERS

-o, --output=file

Output file name

-d, --defines

Generate header file with token definitions

-v, --verbose

Create .output file with parser states

-t, --debug

Enable debug output in parser

-g, --graph

Generate VCG graph of parser

-r, --report=things

Generate report (state, itemset, lookahead)

-W, --warnings

Enable warnings

-l, --no-lines

Don't generate #line directives

GRAMMAR FILE FORMAT

$ %{
#include <stdio.h>
%}

%token NUMBER
%%

expr: NUMBER '+' NUMBER { $$ = $1 + $3; }
;

%%

FEATURES

- LALR(1) parser generation
- GLR parsing support
- C++, Java output
- Location tracking
- Error recovery
- Precedence declarations
- Semantic actions

WORKFLOW

$ # Generate parser
bison -d calculator.y

# Compile with flex lexer
flex lexer.l
gcc lex.yy.c calculator.tab.c -o calculator

# View parser details
bison -v grammar.y
cat grammar.output

CAVEATS

Learning curve for grammar syntax. Shift/reduce and reduce/reduce conflicts possible. Generated code can be large. Error messages sometimes cryptic. Usually paired with flex for lexical analysis.

HISTORY

Bison was written by Robert Corbett in 1985 as a free replacement for Unix yacc, later adopted by the GNU Project.