Huge News!Announcing our $40M Series B led by Abstract Ventures.Learn More
Socket
Sign inDemoInstall
Socket

crio

Package Overview
Dependencies
Maintainers
1
Versions
60
Alerts
File Explorer

Advanced tools

Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

crio

Immutable objects and arrays in a natural way

  • 2.6.0
  • Source
  • npm
  • Socket score

Version published
Weekly downloads
62
increased by342.86%
Maintainers
1
Weekly downloads
 
Created
Source

crio

Immutable JS objects with a natural API

Jump to the API

API Documentation

Usage
// ES2015
import crio from 'crio';

// CommonJS
const crio = require('crio').default;

// UMD
const crio = window.crio;
What is immutable?

When something is described as immutable, it means that it cannot change after it has been created. In JavaScript terms, this means that any attempted change to an object results in a brand new object being created, without changing the original object.

Why do we need this in JavaScript?

The concept of immutability already exists in a lot of places in JavaScript, for example:

const two = 2;
const three = 3;
const five = two + three;

By adding together two and three you expect to get five, however you don't expect the value of two to change. You can continue working with it even after using it in an expression:

const two = 2;
const three = 3;
const five = two + three;
const four = two * two;

This is true of strings, numbers, undefined, and null, and is an expected behavior. The same idea, however, is not true for complex objects in JavaScript. For example:

const foo = ['foo'];
const fooBar = foo.push('bar');

The expectation is that you have pushed the value of "bar" into foo and created a new array bar that contains "foo, bar", however in reality this is what happens:

const foo = ['foo'];
const fooBar = foo.push('bar');

console.log(foo); // ['foo', 'bar']
console.log(fooBar); // 1

Basically, you have mutated foo so that it is no longer empty, and what the .push() method returns is actually the index of the item you just added. This double-standard of expectations creates a lot confusion from a development perspective, but also makes keeping track of the state of your application very difficult because there is no traceability of what transactions have occurred to create that state at any given point.

Enter crio

crio attempts to solve the problem by closing the "immutable loop" on collection items, meaning it applies immutability to objects that are normally mutable by nature by replacing mutating methods with immutable counterparts. As a point of reference:

Naturally immutable objects

  • Numbers
  • Strings
  • undefined
  • null

Naturally mutable objects

  • Arrays
  • Dates (not covered by crio)
  • Objects

To create a new crio object, its pretty straightforward:

const crioArray = crio([]);
const crioObject = crio({});

These are examples with empty objects, but you can pass in populated objects as well, or if you pass in nothing it will default to an object. What crio does is clone and freeze the object via Object.freeze, and stores as a custom CrioArray or CrioObject with a prototypical methods that will return a new immutable version of the object with each update. Example:

const foo = crio(['foo']);
const fooBar = foo.push('bar');

console.log(foo); // ['foo']
console.log(fooBar); // ['foo', 'bar']

The API is the same as you already know working with those objects, and includes polyfills for all ES6 and some ES7 functions, as well as a few helpful crio-specific functions. The only difference is that any setting happens via .set() rather than direct index / property setting. You can work with the objects as you normally would with other libraries (lodash, for example). There is also no change to the protoypes of native objects, so you can apply this on your existing code go-forward. Basically, you shouldn't even notice you aren't working with the native objects, save for the fact everything is immutable.

Why not just use X immutable library?

There are a bunch of ones out there, but the three that people usually gravitate towards:

Immutable is quite nice, and very highly regarded by the community, however it creates an opaque object that cannot be used with other external libraries (namely lodash) without converting back to vanilla JS. Additionally, the object itself is not truly immutable, just constructed in a way that makes it incredibly difficult to alter outside of the API. This decision was likely made for performance reasons, but can cause unintended consequences (you can totally assign foo.bar = 'baz' and no error is thrown).

mori is the seasoned veteran, having been hardened via ClosureScript, and for many is chosen specifically because it does not try to "immutabilize" the default API methods. Personal taste, I wasn't interested in relearning an entirely new API (that for a person without a ClosureScript background is obtuse). Plus, it's not written in JavaScript ... its compiled to it, which just felt wrong to a JavaScript devout like me.

seamless-immutable has some great ideas, and I thought that could be the best option because they try to retain the native operations while leveraging Object.freeze, much like crio does. That said, they do not try to replace mutable methods with immutable ones, they just throw errors when you attempt them and its up to you to figure out the "right way". As such, it fell short of my expectations.

Bottom line, I support each one of these projects to the fullest because they are trying to create immutability in JavaScript, just with different approaches.

Browser support

The only requirement is that your browser has a proper ES5 environment, which you can mostly shim with something like es5-shim, however certain things like Object.defineProperty cannot be shimmed, so IE8 and below cannot be supported.

crio has been tested on the following browsers:

  • Chrome
  • Firefox
  • Edge
  • IE11

This is only because of the youth of the project, as the intended support should include Safari, Opera, and IE versions back to 9. Theoretically all of these browsers should work out of the box, I just have not verified it. Please report any issues that you encounter.

Performance

There has been a lot of performance tuning (and hopefully more to come), however because new objects are being instantiated with each creation then inevitably things will be slower than the native methods. Additionally, because objects are frozen upon creation, the only way to produce a new object is to clone the existing object, so doing a bunch of assignment operations in a loop can add up. We're still talking milliseconds here (and it probably won't be noticeable to you anyway), but who knows ... maybe 51ms to create an array of 1000 unique objects is way too slow for your application.

Basically, if you're noticing a perceivable slowdown, check the implementation method. The majority of processing time is spent in the construction of the CrioArray / CrioObject, so optimizing for that will keep things performant.

A silly and unrealistic example that micro-optimizers love to use:


let crioArray = crio([]),
    index = -1;
    
while (++index < 100000) {
    crioArray = crioArray.push(index);
}

This will take a while, but a small tweak makes it much faster:


let array = [],
    index = -1;
    
while (++index < 100000) {
    array.push(index);
}

const crioArray = crio(array);

The difference here is we created the CrioArray in a single shot, whereas before a new CrioArray was instantiated with each .push(). Focusing your code on optimizing for the creation will keep things nice and snappy.

Additionally, if you plan to do a bunch of manipulations to it, you can always use the .mutate() method:

const crioObject = crioArray.mutate((array) => {
    let object = {};
    
    array.forEach((item, index) => {
        object[index] = item;
    });
    
    return object;
});

This is a simple example, but mutate will allow you to work with the standard (mutable) JS objects and maximize performance.

Gotchas

Recursive objects are not allowed

Immutable objects with recursive values are basically impossible, and trying them will cause a stack overflow, so be mindful of that!

Development

Standard stuff, clone the repo and npm install dependencies. The npm scripts available:

  • build => run webpack to build crio.js with NODE_ENV=development
  • build-minifed => run webpack to build crio.min.js with NODE_ENV=production
  • dev => run webpack dev server to run example app (playground!)
  • lint => run ESLint against all files in the src folder
  • prepublish => run lint, test, transpile, build, and build-minified
  • test => run AVA test functions with NODE_ENV=test
  • test-production => run AVA test functions with NODE_ENV=production
  • test:watch => same as test, but runs persistent watcher
  • test-production:watch => same as test, but runs persistent watcher
  • transpile => run babel against all files in src to create files in lib

Keywords

FAQs

Package last updated on 15 Jul 2016

Did you know?

Socket

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.

Install

Related posts

SocketSocket SOC 2 Logo

Product

  • Package Alerts
  • Integrations
  • Docs
  • Pricing
  • FAQ
  • Roadmap
  • Changelog

Packages

npm

Stay in touch

Get open source security insights delivered straight into your inbox.


  • Terms
  • Privacy
  • Security

Made with ⚡️ by Socket Inc