Neural Processing Mechanisms
Miss any of our Open RFC calls?Watch the recordings here! »


0.3.5 • Public • Published


Unpack multibyte binary values from buffers and streams. (Maintenance fork of substack's binary)

You can specify the endianness and signedness of the fields to be unpacked too.

This module is a cleaner and more complete version of bufferlist's binary module that runs on pre-allocated buffers instead of a linked list.

build status


npm install binary


See the docs/examples/ directory.


var binary = require('binary-pmb')

var b = binary()

Return a new writable stream b that has the chainable methods documented below for buffering binary input.


Parse a static buffer in one pass. Returns a chainable interface with the methods below plus a vars field to get at the variable stash as the last item in a chain.

In parse mode, methods will set their keys to null if the buffer isn't big enough except buffer() and scan() which read up up to the end of the buffer and stop.


See binary.parse.


Parse bytes in the buffer or stream given:

  • number of bits
  • endianness ( l : little, b : big ),
  • signedness ( u and e : unsigned, s : signed )

These functions won't start parsing until all previous parser functions have run and the data is available.

The result of the parse goes into the variable stash at key. If key has dots (.s), it refers to a nested address. If parent container values don't exist they will be created automatically, so for instance you can assign into dst.addr and dst.port and the dst key in the variable stash will be { addr : x, port : y } afterwards.

b.buffer(key, size)

Take size bytes directly off the buffer stream, putting the resulting buffer slice in the variable stash at key.

vars lookup: If size is a string, use the value at vars[size]. The key follows the same dotted address rules as the word functions.

b.str(key, size[, encoding])

Same as .buffer but decode the data as a string. Default encoding is latin1.

b.utf8(key, size)

Same as .buffer but decode the data as a UTF-8 string.


Jump dist bytes ahead. The "vars lookup" feature from .buffer applies to dist.

b.scan(key, buffer)

Search for buffer in the stream and store all the intervening data in the stash at at key, excluding the search buffer. If buffer passed as a string, it will be converted into a Buffer internally.

For example, to read in a line you can just do:

var b = binary()
    .scan('line', new Buffer('\r\n'))
    .tap(function (vars) {


The callback cb is provided with the variable stash from all the previous actions once they've all finished.

You can nest additional actions onto this inside the callback.

b.into(key, cb)

Like .tap(), except all nested actions will assign into a key in the vars stash.


Loop, each time calling cb(end, vars) for function end and the variable stash with this set to a new chain for nested parsing. The loop terminates once end is called.


Clear the variable stash entirely.

Known issues

  • The word64 functions will only return approximations since javascript uses IEEE floating point for all number types. Mind the loss of precision.





npm i binary-pmb

DownloadsWeekly Downloads






Unpacked Size

47.9 kB

Total Files


Last publish


  • avatar