npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details


  • User packages



Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.


Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner




Translation between JavaScript values and Buffers





NPM version Build Status Coverage Status

buffer-layout is a utility module implemented in pure JavaScript that supports translations between JavaScript values and Buffers. It is made available through github and released under the MIT license.

Layout support is provided for these types of data:

  • Signed and unsigned integral values from 1 to 6 bytes in length, in little-endian or big-endian format;
  • Signed and unsigned 64-bit integral values decoded as integral Numbers;
  • Float and double values (also little-endian or big-endian);
  • Sequences of instances of an arbitrary layout, with constant or data-dependent length;
  • Structures with named fields containing arbitrary layouts;
  • Unions of variant layouts where the type of data is recorded in a prefix value, another layout element, or provided externally;
  • Bit fields within 8, 16, 24, or 32-bit unsigned integers, numbering from the least or most significant bit;
  • NUL-terminated C strings;
  • Blobs of fixed or variable-length raw data.


Development and testing is done using Node.js, supporting versions 4.5 and later. Install with npm install buffer-layout.


All examples are from the test/examples.js unit test and assume the following context:

const assert = require('assert');
const util = require('util');
const lo = require('buffer-layout');

The examples give only a taste of what can be done. Structures, unions, and sequences can nest; union discriminators can be within the union or external to it; sequence and blob lengths may be fixed or read from the buffer.

For full details see the documentation.

Four-element array of 16-bit signed little-endian integers

The C definition:

int16_t arr[4] = { 1, -1, 3, -3 };

The buffer-layout way:

const ds = lo.seq(lo.s16(), 4);
const b = Buffer.alloc(8);
assert.equal(ds.encode([1, -1, 3, -3], b), 4 * 2);
assert.equal(Buffer.from('0100ffff0300fdff', 'hex').compare(b), 0);
assert.deepEqual(ds.decode(b), [1, -1, 3, -3]);

See Int and Sequence.

A native C struct on a 32-bit little-endian machine

The C definition:

struct ds {
  uint8_t v;
  uint32_t u32;
} st;

The buffer-layout way:

const ds = lo.struct([lo.u8('v'),
                    lo.seq(lo.u8(), 3), // alignment padding
assert.equal(ds.offsetOf('u32'), 4);
const b = Buffer.alloc(8);
assert.equal(ds.encode({v: 1, u32: 0x12345678}, b), 1 + 3 + 4);
assert.equal(Buffer.from('01bdbdbd78563412', 'hex').compare(b), 0);
assert.deepEqual(ds.decode(b), {v: 1, u32: 0x12345678});

Note that the C language requires padding which must be explicitly added in the buffer-layout structure definition. Since the padding is not accessible, the corresponding layout has no property.

See Structure.

A packed C struct on a 32-bit little-endian machine

The C definition:

struct ds {
  uint8_t v;
  uint32_t u32;
} __attribute__((__packed__)) st;

The buffer-layout way:

const ds = lo.struct([lo.u8('v'),
assert.equal(ds.offsetOf('u32'), 1);
const b = Buffer.alloc(5);
assert.equal(ds.encode({v: 1, u32: 0x12345678}, b), 1 + 4);
assert.equal(Buffer.from('0178563412', 'hex').compare(b), 0);
assert.deepEqual(ds.decode(b), {v: 1, u32: 0x12345678});

A tagged union of 4-byte values

Assume a 5-byte packed structure where the interpretation of the last four bytes depends on the first byte. The C definition:

struct {
  uint8_t t;
  union ds {
    uint8_t u8[4];  // default interpretation
    int16_t s16[2]; // when t is 'h'
    uint32_t u32;   // when t is 'w'
    float f32;      // when t is 'f'
  } u;
} __attribute__((__packed__)) un;

The buffer-layout way:

const t = lo.u8('t');
const un = lo.union(t, lo.seq(lo.u8(), 4, 'u8'));
const nul = un.addVariant('n'.charCodeAt(0), 'nul');
const u32 = un.addVariant('w'.charCodeAt(0), lo.u32(), 'u32');
const s16 = un.addVariant('h'.charCodeAt(0), lo.seq(lo.s16(), 2), 's16');
const f32 = un.addVariant('f'.charCodeAt(0), lo.f32(), 'f32');
const b = Buffer.alloc(un.span);
assert.deepEqual(un.decode(b), {t: 0, u8: [0, 0, 0, 0]});
assert.deepEqual(un.decode(Buffer.from('6e01020304', 'hex')),
                 {nul: true});
assert.deepEqual(un.decode(Buffer.from('7778563412', 'hex')),
                 {u32: 0x12345678});
assert.deepEqual(un.decode(Buffer.from('660000bd41', 'hex')),
                 {f32: 23.625});
assert.deepEqual(un.decode(Buffer.from('a5a5a5a5a5', 'hex')),
                 {t: 0xa5, u8: [0xa5, 0xa5, 0xa5, 0xa5]});
assert.equal(s16.encode({s16: [123, -123]}, b), 1 + 2 * 2);
assert.equal(Buffer.from('687b0085ff', 'hex').compare(b), 0);

See Union.

Packed bit fields on a little-endian machine

The C definition:

struct ds {
  unsigned int b00l03: 3;
  unsigned int flg03: 1;
  unsigned int b04l18: 24;
  unsigned int b1Cl04: 4;
} st;

The buffer-layout way:

const ds = lo.bits(lo.u32());
const b = Buffer.alloc(4);
ds.addField(3, 'b00l03');
ds.addField(24, 'b04l18');
ds.addField(4, 'b1Cl04');
assert.equal(ds.encode({b00l03: 3, b04l18: 24, b1Cl04: 4}, b), 4);
assert.equal(Buffer.from('8b010040', 'hex').compare(b), 0);
                 {b00l03: 3, flg03: true, b04l18: 24, b1Cl04: 4});

See BitStructure.

64-bit values as Numbers

The C definition:

uint64_t v = 0x0102030405060708ULL;

The buffer-layout way:

const ds = lo.nu64be();
const b = Buffer.from('0102030405060708', 'hex');
const v = 72623859790382856;
const nv = v - 6;
assert.equal(v, nv);
assert.equal(ds.decode(b), nv);

Note that because the exact value is not less than 2^53 it cannot be represented as a JavaScript Number, and is instead approximated by a nearby representable integer that is equivalent within Numbers.

See NearUInt64.

A NUL-terminated C string

The C definition:

const char str[] = "hi!";

The buffer-layout way:

const ds = lo.cstr();
const b = Buffer.alloc(8);
assert.equal(ds.encode('hi!', b), 3 + 1);
const slen = ds.getSpan(b);
assert.equal(slen, 4);
assert.equal(Buffer.from('68692100', 'hex').compare(b.slice(0, slen)), 0);
assert.equal(ds.decode(b), 'hi!');

See CString.

A fixed-length block of data offset within a buffer

The buffer-layout way:

const ds = lo.blob(4);
const b = Buffer.from('0102030405060708', 'hex');
assert.equal(Buffer.from('03040506', 'hex').compare(ds.decode(b, 2)), 0);

See Blob.

A variable-length array of pairs of C strings

The buffer-layout way:

const pr = lo.seq(lo.cstr(), 2);
const n = lo.u8('n');
const vla = lo.seq(pr, lo.offset(n, -1), 'a');
const st = lo.struct([n, vla], 'st');
const b = Buffer.alloc(32);
const arr = [['k1', 'v1'], ['k2', 'v2'], ['k3', 'etc']];
assert.equal(st.encode({a: arr}, b),
             1 + (2 * ((2 + 1) + (2 + 1)) + (2 + 1) + (3 + 1)));
const span = st.getSpan(b);
assert.equal(span, 20);
assert.equal(Buffer.from('036b31007631006b32007632006b330065746300', 'hex')
             .compare(b.slice(0, span)), 0);
assert.deepEqual(st.decode(b), {n: 3, a: arr});

See OffsetLayout.

A C flexible array member with implicit length

When data is obtained over a packetized interface the length of the packet can provide implicit limits on the last field.

The C definition:

struct ds {
  uint8_t prop;
  uint16_t data[];

The buffer-layout way:

const st = lo.struct([lo.u8('prop'),
const b = Buffer.from('21010002030405', 'hex');
assert.deepEqual(st.decode(b), {prop: 33, data: [0x0001, 0x0302, 0x0504]});
assert.equal(st.encode({prop: 9, data: [5, 6]}, b), 1 + 2 * 2);
assert.equal(Buffer.from('0905000600FFFF', 'hex').compare(b), 0);

Tagged values, or variable-length unions

Storing arbitrary data using a leading byte to identify the content then a value that takes up only as much room as is necessary.

The example also shows how to extend the variant recognition API to support abitrary constant without consuming space for them in the encoded union. This could be used to make something similar to BSON.

Here's the code that defines the union, the variants, and the recognition of true and false values for b as distinct variants:

const un = lo.union(lo.u8('t'));
const u8 = un.addVariant('B'.charCodeAt(0), lo.u8(), 'u8');
const s16 = un.addVariant('h'.charCodeAt(0), lo.s16(), 's16');
const s48 = un.addVariant('Q'.charCodeAt(0), lo.s48(), 's48');
const cstr = un.addVariant('s'.charCodeAt(0), lo.cstr(), 'str');
const tr = un.addVariant('T'.charCodeAt(0), lo.const(true), 'b');
const fa = un.addVariant('F'.charCodeAt(0), lo.const(false), 'b');
const b = Buffer.alloc(1 + 6);
un.configGetSourceVariant(function(src) {
  if (src.hasOwnProperty('b')) {
    return src.b ? tr : fa;
  return this.defaultGetSourceVariant(src);

And here are examples of encoding, checking the encoded length, and decoding each of the alternatives:

assert.equal(un.encode({u8: 1}, b), 1 + 1);
assert.equal(un.getSpan(b), 2);
assert.equal(Buffer.from('4201ffffffffff', 'hex').compare(b), 0);
assert.equal(un.decode(b).u8, 1);

assert.equal(un.encode({s16: -32000}, b), 1 + 2);
assert.equal(un.getSpan(b), 3);
assert.equal(Buffer.from('680083ffffffff', 'hex').compare(b), 0);
assert.equal(un.decode(b).s16, -32000);

const v48 = Math.pow(2, 47) - 1;
assert.equal(un.encode({s48: v48}, b), 1 + 6);
assert.equal(un.getSpan(b), 7);
assert.equal(Buffer.from('51ffffffffff7f', 'hex').compare(b), 0);
assert.equal(un.decode(b).s48, v48);

assert.equal(un.encode({b: true}, b), 1);
assert.equal(un.getSpan(b), 1);
assert.equal(Buffer.from('54ffffffffffff', 'hex').compare(b), 0);
assert.strictEqual(un.decode(b).b, true);

assert.equal(un.encode({b: false}, b), 1);
assert.equal(un.getSpan(b), 1);
assert.equal(Buffer.from('46ffffffffffff', 'hex').compare(b), 0);
assert.strictEqual(un.decode(b).b, false);

NOTE This code tickles a long-standing bug in Buffer.writeInt{L,B}E; if you are using Node prior to 4.2.4 or 5.2.0 you should update.