js-tokens

What is js-tokens?

The js-tokens npm package is a lightweight, regex-based lexer for JavaScript tokens. It can tokenize JavaScript code without the overhead of a full parser, making it useful for tasks like syntax highlighting or extracting specific tokens from code.

What are js-tokens's main functionalities?

Tokenizing JavaScript code

This feature allows you to tokenize a string of JavaScript code into an array of token objects. Each token object describes a syntactic element of the code, such as a keyword, identifier, number, string, or operator.

const jsTokens = require('js-tokens');
const tokens = [...jsTokens('var x = 42;')];
console.log(tokens);

Other packages similar to js-tokens

acorn

Acorn is a full JavaScript parser that can parse source code into an abstract syntax tree (AST). While js-tokens only tokenizes the code, Acorn can parse it and provide more detailed information about the structure of the code.

esprima

Esprima is another JavaScript parser that can convert code into an AST. It offers similar functionality to Acorn but has different API and extension points. Compared to js-tokens, Esprima provides a more comprehensive analysis of the code.

babylon

Babylon is the parser used by Babel, and it's capable of handling modern JavaScript features. It's more feature-rich than js-tokens, offering AST generation and the ability to handle experimental syntax through plugins.

README

Overview

A regex that tokenizes JavaScript.

var jsTokens = require("js-tokens")

var jsString = "var foo=opts.foo;\n..."

jsString.match(jsTokens)
// ["var", " ", "foo", "=", "opts", ".", "foo", ";", "\n", ...]

Installation

• npm install js-tokens

var jsTokens = require("js-tokens")

Usage

jsTokens

A regex with the g flag that matches JavaScript tokens.

The regex always matches, even invalid JavaScript and the empty string.

The next match is always directly after the previous.

var token = jsTokens.matchToToken(match)

Takes a match returned by jsTokens.exec(string), and returns a {type: String, value: String} object. The following types are available:

• string
• comment
• regex
• number
• name
• operator
• punctuation
• whitespace
• invalid

Multi-line comments and strings also have a closed property indicating if the token was closed or not (see below).

Comments and strings both come in two flavors. To distinguish them, check if the token starts with //, /*, ' or ".

For example usage, please see this gist.

Invalid code handling

Unterminated strings are still matched as strings. JavaScript strings cannot contain (unescaped) newlines, so unterminated strings simply end at the end of the line.

Unterminated multi-line comments are also still matched as comments. They simply go on to the end of the string.

Unterminated regex literals are likely matched as division and whatever is inside the regex.

Invalid ASCII characters have their own capturing group.

Invalid non-ASCII characters are treated as names, to simplify the matching of names (except unicode spaces which are treated as whitespace).

Regex literals may contain invalid regex syntax. They are still matched as regex literals. They may also contain repeated regex flags, to keep the regex simple.

Strings may contain invalid escape sequences.

Limitations

Tokenizing JavaScript using regexes—in fact, one single regex—won’t be perfect. But that’s not the point either.

Division and regex literals collision

Consider this example:

var g = 9.82
var number = bar / 2/g

var regex = / 2/g

A human can easily understand that in the number line we’re dealing with division, and in the regex line we’re dealing with a regex literal. How come? Because humans can look at the whole code to put the / characters in context. A JavaScript regex cannot. It only sees forwards.

When the jsTokens regex scans throught the above, it will see the following at the end of both the number and regex rows:

/ 2/g

It is then impossible to know if that is a regex literal, or part of an expression dealing with division.

Here is a similar case:

foo /= 2/g
foo(/= 2/g)

The first line divides the foo variable with 2/g. The second line calls the foo function with the regex literal /= 2/g. Again, since jsTokens only sees forwards, it cannot tell the two cases apart.

There are some cases where we can tell division and regex literals apart, though.

First off, we have the simple cases where there’s only one slash in the line:

var foo = 2/g
foo /= 2

Regex literals cannot contain newlines, so the above cases are correctly identified as division. Things are only problematic when there are more than one non-comment slash in a single line.

Secondly, not every character is a valid regex flag.

var number = bar / 2/e

The above example is also correctly identified as division, because e is not a valid regex flag. I initially wanted to future-proof by allowing [a-zA-Z]* (any letter) as flags, but it is not worth it since it increases the amount of ambigous cases. So only the standard g, m and i flags, as well as the y flag supported by Firefox 3.6+, are allowed. This means that the above example will be identified as division as long as you don’t rename the e variable to some permutation of gmiy 1 to 4 characters long.

Lastly, we can look forward for information.

• If the token following what looks like a regex literal is not valid after a regex literal, but is valid in a division expression, then the regex literal is treated as division instead. For example, a flagless regex cannot be followed by a string, number or name, but all of those three can be the denominator of a division.
• Generally, if what looks like a regex literal is followed by an operator, the regex literal is treated as division instead. This is because regexes are seldomly used with operators (such as +, *, && and ==), but division could likely be part of such an expression.

Please consult the regex source and the test cases for precise information on when regex or division is matched (should you need to know). In short, you could sum it up as:

If the end of a statement looks like a regex literal (even if it isn’t), it will be treated as one. Otherwise it should work as expected (if you write sane code).

License

The X11 (“MIT”) License.

Changelog

Version 0.3.1 (2015-01-06)

• Fixed: Support for unicode spaces. They used to be allowed in names (which is very confusing), and some unicode newlines were wrongly allowed in strings and regexes.