/msgpack-javascript

MessagePack for JavaScript/TypeScript/ECMA-262 / msgpack.org[JavaScript]

Primary LanguageTypeScriptISC LicenseISC

MessagePack for JavaScript/ECMA-262

npm version Build Status codecov bundlephobia

Browser Matrix powered by Sauce Labs

This is a JavaScript/ECMA-262 implementation of MessagePack, an efficient binary serilization format:

https://msgpack.org/

This library is a universal JavaScript, meaning it is compatible with all the major browsers and NodeJS. In addition, because it is implemented in TypeScript, type definition files (d.ts) are bundled in the distribution.

Note that this is the second version of MessagePack for JavaScript. The first version, which was implemented in ES5 and was never released to npmjs.com, is tagged as classic.

Synopsis

import { deepStrictEqual } from "assert";
import { encode, decode } from "@msgpack/msgpack";

const object = {
  nil: null,
  integer: 1,
  float: Math.PI,
  string: "Hello, world!",
  binary: Uint8Array.from([1, 2, 3]),
  array: [10, 20, 30],
  map: { foo: "bar" },
  timestampExt: new Date(),
};

const encoded: Uint8Array = encode(object);

deepStrictEqual(decode(encoded), object);

Table of Contents

Install

This library is published to npmjs.com as @msgpack/msgpack.

npm install @msgpack/msgpack

API

encode(data: unknown, options?: EncodeOptions): Uint8Array

It encodes data and returns a byte array as Uint8Array, throwing errors if data is, or includes, a non-serializable object such as a function or a symbol.

for example:

import { encode } from "@msgpack/msgpack";

const encoded: Uint8Array = encode({ foo: "bar" });
console.log(encoded);

If you'd like to convert the uint8array to a NodeJS Buffer, use Buffer.from(arrayBuffer, offset, length) in order not to copy the underlying ArrayBuffer, while Buffer.from(uint8array) copies it:

import { encode } from "@msgpack/msgpack";

const encoded: Uint8Array = encode({ foo: "bar" });

// `buffer` refers the same ArrayBuffer as `encoded`.
const buffer: Buffer = Buffer.from(encoded.buffer, encoded.byteOffset, encoded.byteLength);
console.log(buffer);

EncodeOptions

Name Type Default
extensionCodec ExtensionCodec ExtensinCodec.defaultCodec
maxDepth number 100
initialBufferSize number 2048
sortKeys boolean false
forceFloat32 boolean false
ignoreUndefined boolean false
context user-defined -

decode(buffer: ArrayLike<number> | ArrayBuffer, options?: DecodeOptions): unknown

It decodes buffer encoded in MessagePack, and returns a decoded object as uknown.

buffer must be an array of bytes, which is typically Uint8Array, or ArrayBuffer.

for example:

import { decode } from "@msgpack/msgpack";

const encoded: Uint8Array;
const object = decode(encoded);
console.log(object);

NodeJS Buffer is also acceptable because it is a subclass of Uint8Array.

DecodeOptions

Name Type Default
extensionCodec ExtensionCodec ExtensinCodec.defaultCodec
maxStrLength number 4_294_967_295 (UINT32_MAX)
maxBinLength number 4_294_967_295 (UINT32_MAX)
maxArrayLength number 4_294_967_295 (UINT32_MAX)
maxMapLength number 4_294_967_295 (UINT32_MAX)
maxExtLength number 4_294_967_295 (UINT32_MAX)
context user-defined -

You can use max${Type}Length to limit the length of each type decoded.

decodeAsync(stream: AsyncIterable<ArrayLike<number>> | ReadableStream<ArrayLike<number>>, options?: DecodeAsyncOptions): Promise<unknown>

It decodes stream in an async iterable of byte arrays, and returns decoded object as uknown type, wrapped in Promise. This function works asyncronously.

DecodeAsyncOptions is the same as DecodeOptions for decode().

This function is designed to work with whatwg fetch() like this:

import { decodeAsync } from "@msgpack/msgpack";

const MSGPACK_TYPE = "application/x-msgpack";

const response = await fetch(url);
const contentType = response.headers.get("Content-Type");
if (contentType && contentType.startsWith(MSGPACK_TYPE) && response.body != null) {
  const object = await decodeAsync(response.body);
  // do something with object
} else { /* handle errors */ }

decodeArrayStream(stream: AsyncIterable<ArrayLike<number>> | ReadableStream<ArrayLike<number>>, options?: DecodeAsyncOptions): AsyncIterable<unknown>

It is alike to decodeAsync(), but only accepts an array of items as the input stream, and emits the decoded item one by one.

It throws errors when the input is not an array-family.

for example:

import { encode } from "@msgpack/msgpack";

const stream: AsyncIterator<Uint8Array>;

// in an async function:
for await (const item of decodeArrayStream(stream)) {
  console.log(item);
}

decodeStream(stream: AsyncIterable<ArrayLike<number>> | ReadableStream<ArrayLike<number>>, options?: DecodeAsyncOptions): AsyncIterable<unknown>

It is alike to decodeAsync() and decodeArrayStream(), but the input stream consists of independent MessagePack items.

In other words, it decodes an unlimited stream and emits an item one by one.

for example:

import { encode } from "@msgpack/msgpack";

const stream: AsyncIterator<Uint8Array>;

// in an async function:
for await (const item of decodeStream(stream)) {
  console.log(item);
}

Extension Types

To handle MessagePack Extension Types, this library provides ExtensionCodec class.

Here is an example to setup custom extension types that handles Map and Set classes in TypeScript:

import { encode, decode, ExtensionCodec } from "@msgpack/msgpack";

const extensionCodec = new ExtensionCodec();

// Set<T>
const SET_EXT_TYPE = 0 // Any in 0-127
extensionCodec.register({
  type: SET_EXT_TYPE,
  encode: (object: unknown): Uint8Array | null => {
    if (object instanceof Set) {
      return encode([...object]);
    } else {
      return null;
    }
  },
  decode: (data: Uint8Array) => {
    const array = decode(data) as Array<unknown>;
    return new Set(array);
  },
});

// Map<T>
const MAP_EXT_TYPE = 1; // Any in 0-127
extensionCodec.register({
  type: MAP_EXT_TYPE,
  encode: (object: unknown): Uint8Array => {
    if (object instanceof Map) {
      return encode([...object]);
    } else {
      return null;
    }
  },
  decode: (data: Uint8Array) => {
    const array = decode(data) as Array<[unknown, unknown]>;
    return new Map(array);
  },
});

// and later
import { encode, decode } from "@msgpack/msgpack";

const encoded = = encode([new Set<any>(), new Map<any, any>()], { extensionCodec });
const decoded = decode(encoded, { extensionCodec });

Not that extension types for custom objects must be [0, 127], while [-1, -128] is reserved for MessagePack itself.

Codec context

When using an extension codec, it may be necessary to keep encoding/decoding state, to keep track of which objects got encoded/re-created. To do this, pass a context to the EncodeOptions and DecodeOptions (and if using typescript, type the ExtensionCodec too). Don't forget to pass the {extensionCodec, context} along recursive encoding/decoding:

import { encode, decode, ExtensionCodec } from "@msgpack/msgpack";

class MyContext {
  track(object: any) { /*...*/ }
}

class MyType { /* ... */ }

const extensionCodec = new ExtensionCodec<MyContext>();

// MyType
const MYTYPE_EXT_TYPE = 0 // Any in 0-127
extensionCodec.register({
  type: MYTYPE_EXT_TYPE,
  encode: (object, context) => {
    if (object instanceof MyType) {
      context.track(object); // <-- like this
      return encode(object.toJSON(), { extensionCodec, context });
    } else {
      return null;
    }
  },
  decode: (data, extType, context) => {
    const decoded = decode(data, { extensionCodec, context });
    const my = new MyType(decoded);
    context.track(my); // <-- and like this
    return my;
  },
});

// and later
import { encode, decode } from "@msgpack/msgpack";

const context = new MyContext();

const encoded = = encode({myType: new MyType<any>()}, { extensionCodec, context });
const decoded = decode(encoded, { extensionCodec, context });

Handling BigInt with ExtensionCodec

This library does not handle BigInt by default, but you can handle it with ExtensionCodec like this:

import { deepStrictEqual } from "assert";
import { encode, decode, ExtensionCodec } from "@msgpack/msgpack";

const BIGINT_EXT_TYPE = 0; // Any in 0-127
const extensionCodec = new ExtensionCodec();
extensionCodec.register({
  type: BIGINT_EXT_TYPE,
  encode: (input: unknown) => {
    if (typeof input === "bigint") {
      return encode(input.toString());
    } else {
      return null;
    }
  },
  decode: (data: Uint8Array) => {
    return BigInt(decode(data));
  },
});

const value = BigInt(Number.MAX_SAFE_INTEGER) + BigInt(1);
const encoded: = encode(value, { extensionCodec });
deepStrictEqual(decode(encoded, { extensionCodec }), value);

The temporal module as timestamp extensions

There is a proposal for a new date/time representations in JavaScript:

This library maps Date to the MessagePack timestamp extension by default, but you can re-map the temporal module (or Temporal Polyfill) to the timestamp extension like this:

import { Instant } from "@std-proposal/temporal";
import { deepStrictEqual } from "assert";
import {
  encode,
  decode,
  ExtensionCodec,
  EXT_TIMESTAMP,
  encodeTimeSpecToTimestamp,
  decodeTimestampToTimeSpec,
} from "@msgpack/msgpack";

const extensionCodec = new ExtensionCodec();
extensionCodec.register({
  type: EXT_TIMESTAMP, // override the default behavior!
  encode: (input: any) => {
    if (input instanceof Instant) {
      const sec = input.seconds;
      const nsec = Number(input.nanoseconds - BigInt(sec) * BigInt(1e9));
      return encodeTimeSpecToTimestamp({ sec, nsec });
    } else {
      return null;
    }
  },
  decode: (data: Uint8Array) => {
    const timeSpec = decodeTimestampToTimeSpec(data);
    const sec = BigInt(timeSpec.sec);
    const nsec = BigInt(timeSpec.nsec);
    return Instant.fromEpochNanoseconds(sec * BigInt(1e9) + nsec);
  },
});

const instant = Instant.fromEpochMilliseconds(Date.now());
const encoded = encode(instant, { extensionCodec });
const decoded = decode(encoded, { extensionCodec });
deepStrictEqual(decoded, instant);

This will be default once the temporal module is standardizied, which is not a near-future, though.

Decoding a Blob

Blob is a binary data container provided by browsers. To read its contents, you can use Blob#arrayBuffer() or Blob#stream(). Blob#stream() is recommended if your target platform support it. This is because streaming decode should be faster for large objects. In both ways, you need to use asynchronous API.

async function decodeFromBlob(blob: Blob): unknown {
  if (blob.stream) {
    // Blob#stream(): ReadableStream<Uint8Array> (recommended)
    return await decodeAsync(blob.stream());
  } else {
    // Blob#arrayBuffer(): Promise<ArrayBuffer> (if stream() is not available)
    return decode(await blob.arrayBuffer());
  }
}

MessagePack Specification

This library is compatible with the "August 2017" revision of MessagePack specification at the point where timestamp ext was added:

  • str/bin separation, added at August 2013
  • extension types, added at August 2013
  • timestamp ext type, added at August 2017

The livinng specification is here:

https://github.com/msgpack/msgpack

Note that as of June 2019 there're no official "version" on the MessagePack specification. See msgpack/msgpack#195 for the discussions.

MessagePack Mapping Table

The following table shows how JavaScript values are mapped to MessagePack formats and vice versa.

Source Value MessagePack Format Value Decoded
null, undefined nil null (*1)
boolean (true, false) bool family boolean (true, false)
number (53-bit int) int family number (53-bit int)
number (64-bit float) float family number (64-bit float)
string str family string
ArrayBufferView bin family Uint8Array (*2)
Array array family Array
Object map family Object (*3)
Date timestamp ext family Date (*4)
  • *1 Both null and undefined are mapped to nil (0xC0) type, and are decoded into null
  • *2 Any ArrayBufferViews including NodeJS's Buffer are mapped to bin family, and are decoded into Uint8Array
  • *3 In handling Object, it is regarded as Record<string, unknown> in terms of TypeScript
  • *4 MessagePack timestamps may have nanoseconds, which will lost when it is decoded into JavaScript Date. This behavior can be overridden by registering -1 for the extension codec.

Prerequsites

This is a universal JavaScript library that supports major browsers and NodeJS.

ECMA-262

  • ES5 language features
  • ES2018 standard library, including:
    • Typed arrays (ES2015)
    • Async iterations (ES2018)
    • Features added in ES2015-ES2018

ES2018 standard library used in this library can be polyfilled with core-js.

If you support IE11, import core-js in your application entrypoints, as this library does in testing for browsers.

NodeJS

NodeJS v10 is required, but NodeJS v12 or later is recommended because it includes the V8 feature of Improving DataView performance in V8.

NodeJS before v10 will work by importing @msgpack/msgpack/dist.es5/msgpack.

Benchmark

Run-time performance is not the only reason to use MessagePack, but a benchmark suite is provided to monitor the performance of this library continuously.

V8's built-in JSON has been improved for years, esp. JSON.parse() is significantly improved in V8/7.6, it is the fastest deserializer as of 2019, as the benchmark result bellow suggests.

However, MessagePack can handles binary data effectively, actual performance depends on situations. You'd better take benchmark on your own use-case if performance matters.

Benchmark on NodeJS/v12.16.2 (v8/7.8)

operation op ms op/s
buf = Buffer.from(JSON.stringify(obj)); 768400 5000 153680
obj = JSON.parse(buf); 1557500 5000 311500
buf = require("msgpack-lite").encode(obj); 588100 5000 117620
obj = require("msgpack-lite").decode(buf); 315800 5001 63147
buf = require("@msgpack/msgpack").encode(obj); 948200 5000 189640
obj = require("@msgpack/msgpack").decode(buf); 778900 5000 155780

Note that Buffer.from() for JSON.stringify() is necessary to emulate I/O where a JavaScript string must be converted into a byte array encoded in UTF-8, whereas MessagePack's encode() returns a byte array.

Distribution

NPM / npmjs.com

The NPM package distributed in npmjs.com includes both ES2015+ and ES5 files:

  • dist/ is compiled into ES2015+
  • dist.es5/ is compiled into ES5 and bundled to singile file
    • dist.es5/msgpack.min.js - the default, minified file (UMD)
    • dist.es5/msgpack.js - an optional, non-minified file (UMD)

If you use NodeJS and/or webpack, their module resolvers use the suitable one automatically.

CDN / unpkg.com

This library is availble via CDN:

<script crossorigin src="https://unpkg.com/@msgpack/msgpack"></script>

It loads MessagePack module to the global object.

Maintenance

Testing

For simple testing:

npm run test

Continuous Integration

This library uses Travis CI.

test matrix:

  • TypeScript targets
    • target=es2019 / target=es5
  • JavaScript engines
    • NodeJS, borwsers (Chrome, Firefox, Safari, IE11, and so on)

See test:* in package.json and .travis.yml for details.

Release Engineering

# run tests on NodeJS, Chrome, and Firefox
make test-all

# edit the changelog
code CHANGELOG.md

# bump version
npm version patch|minor|major

# run the publishing task
make publish

Updating Dependencies

npm run update-dependencies

Big Thanks

Cross-browser Testing Platform and Open Source <3 Provided by Sauce Labs.

Sauce Labs

License

Copyright 2019 The MessagePack community.

This software uses the ISC license:

https://opensource.org/licenses/ISC

See LICENSE for details.