README-DEVELOPMENT.adoc

Development

The parser can be generated in any language supported as ANTLR4 target.

java -jar ~/.m2/repository/org/antlr/antlr4/4.7.3-SNAPSHOT/antlr4-4.7.3-SNAPSHOT-complete.jar \
    -listener \
    -visitor \
    -Dlanguage=JS \
    Express.g4

Generated Parser and Lexer classes can be used as follows:

const { CharStreams, CommonTokenStream } = require('antlr4');
const { ExpressLexer } = require('./ExpressLexer');
const { ExpressParser } = require('./ExpressParser');
const fs = require('fs');

const input = fs.readFileSync(0, 'utf-8');

const charStream = CharStreams.fromString(input);
const lexer = new ExpressLexer(charStream);
const tokenStream = new CommonTokenStream(lexer);
const parser = new ExpressParser(tokenStream);

const parseTree = parser.syntax();

Generated Visitor classes can be used for mapping from ANTLR4 parse tree classes to custom classes.

Ruby parser

Generating a pure Ruby parser requires to compile ANTLR4 fork from @twalmsley, ruby_dev branch. It is compatible with antlr4-ruby-runtime gem.

java -jar ~/.m2/repository/org/antlr/antlr4/4.7.3-SNAPSHOT/antlr4-4.7.3-SNAPSHOT-complete.jar \
    -listener \
    -visitor \
    -Dlanguage=Ruby \
    Express.g4

Generated Parser and Lexer classes can be used as follows:

require './ExpressListener'
require './ExpressVisitor'
require './ExpressLexer'
require './ExpressParser'

input = STDIN.read

char_stream = Antlr4::Runtime::CharStreams.from_string(input, 'String')
lexer = Express::ExpressLexer.new(char_stream)
token_stream = Antlr4::Runtime::CommonTokenStream.new(lexer)
parser = Express::ExpressParser.new(token_stream)

parse_tree = parser.syntax()

However, Ruby performance shows it’s not a good fit for parsing long files. The other possibility is to generate C++ parser with Ruby FFI binding. See MODLanguage/antlr4-ruby-runtime#9 for details.