token-scope

A very fast approximation of lexical-scope. It takes the source code and returns a list of all identifiers that have the potential to be references to global variables. This just parses the token stream, so it is much less thorough than lexical-scope or using uglify-js to do the same thing. It has the advantage of being vastly faster though. It is good enough to detect that jQuery does not contain require or any of the node.js globals.

Installation

npm install token-scope

Usage

var detect = require('token-scope')
var inspectForRequire = detect(src)

if (inspectForRequire.indexOf('require') != -1) {
  // do more time consuming checks to search for require properly
}

Tests:

var detect = require('token-scope')

describe('detect(src, identifiers)', function () {
  it('returns an array of identifiers that might represent global variable references', function () {
    assert(detect('var x = require("foo")').indexOf('require') != -1)
    assert(detect('var x = foo("require")').indexOf('require') == -1)
  })
})

Benchmarking

To run the benchmarks, download this repository then do:

$ npm install
$ node bench/run.js

This runs 3 different scoping mechanisms on a minified copy of jQuery and report the total time for 10 iterations. They then also check that the result does not include any of the node.js globals. On my machine, this results in the output (note how token-scope results in a lot more potential globals, but still turns out to be good enough to rule out node.js globals):

name | time | output ---------------|--------|--------------- token-scope | 747ms | a,b,cy,f,cv,ck,… ugly-scope | 1457ms | setTimeout,par… lexical-scope | 2252ms | setTimeout,par…

License

MIT