yajily
Yet Another Javascript Implement of Lexer and Yacc
features & usage
The source in src folder is written in ES6, thus can only run on latest Firefox. If you want to use it in other browsers, please include the script in the build folder.
The lexer is based on RegExps. If you define an action and return a token, it should have a 'type' property or the parser will not work.
var tokens = lex(input, [
[/\^|\+|-|\*|\/|%|\(|\)/, function(m) {
return { type:m, value:m }
}],
[/\d+/, function(m) {
return { type:'NUM', value:parseInt(m) }
}],
[/\s+/],
])It takes some time for the parser to generate the LALR(1) table from the grammars. You may want to save the table and pass it to the the parser later.
var table = yajily.parse.build([
['S', ['E']],
['E', ['(', 'E', ')'], function(_l, e, _r){ return e }],
['E', ['-', 'E'], function(_s, e ){ return [_s, e] }],
['E', ['E', '+', 'E'], function(e1, _, e2){ return [_, e1, e2] }],
['E', ['E', '-', 'E'], function(e1, _, e2){ return [_, e1, e2] }],
['E', ['E', '*', 'E'], function(e1, _, e2){ return [_, e1, e2] }],
['E', ['E', '/', 'E'], function(e1, _, e2){ return [_, e1, e2] }],
['E', ['E', '%', 'E'], function(e1, _, e2){ return [_, e1, e2] }],
['E', ['E', '^', 'E'], function(e1, _, e2){ return [_, e1, e2] }],
['E', ['NUM']],
])Ambiguous grammars are ok. The table saves all the edges, and shift-reduce conflicts can be resolved with the precedence definitions when parsing.
var tree = yajily.parse(tokens, grammars, table, {
'+': [11, 'left'],
'-': [11, 'left'],
'*': [12, 'left'],
'/': [12, 'left'],
'%': [12, 'left'],
'^': [20, 'right'],
})See example.html for more details. Also see the following project for advanced uses.
todo
- add a method to resolve reduce-reduce conflicts
- add error handlers to recover from errors
license
GPL