[javascript] Should I use JSLint or JSHint JavaScript validation?

I am currently validating my JavaScript against JSLint and making progress on, it's assisting me to write better JavaScript - in particular in working with the Jquery library.

I have now come across JSHint, a fork of JSLint.
So I am wondering for web applications, which are very much JavaScript was driven, which is the better or most applicable validation tool to work against:

• JSLint or JSHint?

I want to decide now on a validation mechanism and moving forward, use this for client side validation.

And Difference between jshint and jslint? Please explain in single javascript example.

Links:

1. jshint- http://www.jshint.com/

2. jslint- http://jslint.com/

I had the same question a couple of weeks ago and was evaluating both JSLint and JSHint.

Contrary to the answers in this question, my conclusion was not:

By all means use JSLint.

Or:

If you're looking for a very high standard for yourself or team, JSLint.

As you can configure almost the same rules in JSHint as in JSLint. So I would argue that there's not much difference in the rules you could achieve.

So the reasons to choose one over another are more political than technical.

We've finally decided to go with JSHint because of the following reasons:

• Seems to be more configurable that JSLint.
• Looks definitely more community-driven rather than one-man-show (no matter how cool The Man is).
• JSHint matched our code style OOTB better that JSLint.

I'd make a third suggestion, Google Closure Compiler (and also the Closure Linter). You can try it out online here.

The Closure Compiler is a tool for making JavaScript download and run faster. It is a true compiler for JavaScript. Instead of compiling from a source language to machine code, it compiles from JavaScript to better JavaScript. It parses your JavaScript, analyzes it, removes dead code and rewrites and minimizes what's left. It also checks syntax, variable references, and types, and warns about common JavaScript pitfalls.

There is also an another actively developed alternative - JSCS — JavaScript Code Style:

JSCS is a code style linter for programmatically enforcing your style guide. You can configure JSCS for your project in detail using over 150 validation rules, including presets from popular style guides like jQuery, Airbnb, Google, and more.

It comes with multiple presets that you can choose from by simply specifying the preset in the .jscsrc configuration file and customize it - override, enable or disable any rules:

{
    "preset": "jquery",
    "requireCurlyBraces": null
}

There are also plugins and extensions built for popular editors.

Also see:

• Perfect code, zero effort

Well, Instead of doing manual lint settings we can include all the lint settings at the top of our JS file itself e.g.

Declare all the global var in that file like:

/*global require,dojo,dojoConfig,alert */

Declare all the lint settings like:

/*jslint browser:true,sloppy:true,nomen:true,unparam:true,plusplus:true,indent:4 */

Hope this will help you :)

tl;dr takeaway:

If you're looking for a very high standard for yourself or team, JSLint. But its not necessarily THE standard, just A standard, some of which comes to us dogmatically from a javascript god named Doug Crockford. If you want to be a bit more flexible, or have some old pros on your team that don't buy into JSLint's opinions, or are going back and forth between JS and other C-family languages on a regular basis, try JSHint.

long version:

The reasoning behind the fork explains pretty well why JSHint exists:

http://badassjs.com/post/3364925033/jshint-an-community-driven-fork-of-jslint http://anton.kovalyov.net/2011/02/20/why-i-forked-jslint-to-jshint/

So I guess the idea is that it's "community-driven" rather than Crockford-driven. In practicality, JSHint is generally a bit more lenient (or at least configurable or agnostic) on a few stylistic and minor syntactical "opinions" that JSLint is a stickler on.

As an example, if you think both the A and B below are fine, or if you want to write code with one or more of the aspects of A that aren't available in B, JSHint is for you. If you think B is the only correct option... JSLint. I'm sure there are other differences, but this highlights a few.

A) Passes JSHint out of the box - fails JSLint

(function() {
  "use strict";
  var x=0, y=2;
  function add(val1, val2){
    return val1 + val2;
  }
  var z;
  for (var i=0; i<2; i++){
    z = add(y, x+i);
  }
})();

B) Passes Both JSHint and JSLint

(function () {
    "use strict";
    var x = 0, y = 2, i, z;
    function add(val1, val2) {
       return val1 + val2;
    }
    for (i = 0; i < 2; i += 1) {
        z = add(y, x + i);
    }
}());

Personally I find JSLint code very nice to look at, and the only hard features of it that I disagree with are its hatred of more than one var declaration in a function and of for-loop var i = 0 declarations, and some of the whitespace enforcements for function declarations.

A few of the whitespace things that JSLint enforces, I find to be not necessarily bad, but out of sync with some pretty standard whitespace conventions for other languages in the family (C, Java, Python, etc...), which are often followed as conventions in Javascript as well. Since I'm writing in various of these languages throughout the day, and working with team members who don't like Lint-style whitespace in our code, I find JSHint to be a good balance. It catches stuff that's a legitimate bug or really bad form, but doesn't bark at me like JSLint does (sometimes, in ways I can't disable) for the stylistic opinions or syntactic nitpicks that I don't care for.

A lot of good libraries aren't Lint'able, which to me demonstrates that there's some truth to the idea that some of JSLint is simply just about pushing 1 version of "good code" (which is, indeed, good code). But then again, the same libraries (or other good ones) probably aren't Hint'able either, so, touché.

There is an another mature and actively developed "player" on the javascript linting front - ESLint:

ESLint is a tool for identifying and reporting on patterns found in ECMAScript/JavaScript code. In many ways, it is similar to JSLint and JSHint with a few exceptions:

• ESLint uses Esprima for JavaScript parsing.
• ESLint uses an AST to evaluate patterns in code.
• ESLint is completely pluggable, every single rule is a plugin and you can add more at runtime.

What really matters here is that it is extendable via custom plugins/rules. There are already multiple plugins written for different purposes. Among others, there are:

• eslint-plugin-angular (enforces some of the guidelines from John Papa's Angular Style Guide)
• eslint-plugin-jasmine
• eslint-plugin-backbone

And, of course, you can use your build tool of choice to run ESLint:

• grunt-eslint
• gulp-eslint

Foreword: Well, that escalated quickly. But decided to pull it through. May this answer be helpful to you and other readers.

Code Hinting

While JSLint and JSHint are good tools to use, over the years I've come to appreciate what my friend @ugly_syntax calls:

smaller design space.

This is a general principle, much like a "zen monk", limiting the choices one has to make, one can be more productive and creative.

Therefore my current favourite zero-config JS code style:

StandardJS.

UPDATE:

Flow has improved a lot. With it, you can add types to your JS with will help you prevent a lot of bugs. But it can also stay out of your way, for instance when interfacing untyped JS. Give it a try!

Quickstart / TL;DR

Add standard as a dependency to you project

npm install --save standard

Then in package.json, add the following test script:

"scripts": {
    "test": "node_modules/.bin/standard && echo put further tests here"
},

For snazzier output while developing, npm install --global snazzy and run it instead of npm test.

Note: Type checking versus Heuristics

My friend when mentioning design space referred to Elm and I encourage you to give that language a try.

Why? JS is in fact inspired by LISP, which is a special class of languages, which happens to be untyped. Language such as Elm or Purescript are typed functional programming languages.

Type restrict your freedom in order for the compiler to be able to check and guide you when you end up violation the language or your own program's rules; regardless of the size (LOC) of your program.

We recently had a junior colleague implement a reactive interface twice: once in Elm, once in React; have a look to get some idea of what I'm talking about.

Compare Main.elm (typed) ? index.js (untyped, no tests)

(ps. note that the React code is not idiomatic and could be improved)

One final remark,

the reality is that JS is untyped. Who am I to suggest typed programming to you?

See, with JS we are in a different domain: freed from types, we can easily express things that are hard or impossible to give a proper type (which can certainly be an advantage).

But without types there is little to keep our programs in check, so we are forced to introduce tests and (to a lesser extend) code styles.

I recommend you look at LISP (e.g. ClojureScript) for inspiration and invest in testing your codes. Read The way of the substack to get an idea.

Peace.