ref

ref

Turn Buffer instances into "pointers"

This module is inspired by the old Pointer class from node-ffi, but with the intent of using Node's fast Buffer instances instead of a slow C++ Pointer class. These two concepts were previously very similar, but now this module brings over the functionality that Pointers had and Buffers are missing, so now Buffers are a lot more powerful.

Features:

• Get the memory address of any Buffer instance
• Read/write references to JavaScript Objects into Buffer instances
• Read/write Buffer instances' memory addresses to other Buffer instances
• Read/write int64_t and uint64_t data values (Numbers or Strings)
• A "type" convention, so that you can specify a buffer as an int *, and reference/dereference at will.
• Offers a buffer instance representing the NULL pointer

Installation

Install with npm:

$ npm install ref

Examples

referencing and derefencing

var ref = require('ref')

// so we can all agree that a buffer with the int value written
// to it could be represented as an "int *"
var buf = new Buffer(4)
buf.writeInt32LE(12345, 0)

// first, what is the memory address of the buffer?
console.log(buf.address())  // ← 140362165284824

// using `ref`, you can set the "type", and gain magic abilities!
buf.type = ref.types.int32

// now we can dereference to get the "meaningful" value
console.log(buf.deref())  // ← 12345


// you can also get references to the original buffer if you need it.
// this buffer could be thought of as an "int **"
var one = buf.ref()

// and you can dereference all the way back down to an int
console.log(one.deref().deref())  // ← 12345

Additions to Buffer.prototype

ref extends Node's core Buffer instances with some useful additions:

---

Buffer#address() → Number

Returns the memory address of the Buffer instance.

---

Buffer#isNull() → Boolean

Returns true if the Buffer's memory address is NULL, false otherwise.

---

Buffer#ref() → Buffer

Returns a new Buffer instance that is referencing this Buffer. That is, the new Buffer is "pointer" sized, and points to the memory address of this Buffer.

The returned Buffer's type property gets set properly as well, with an indirection level increased by 1.

---

Buffer#deref() → ???

Returns the dereferenced value from the Buffer instance. This depends on the type property being set to a proper "type" instance (see below).

The returned value can be another Buffer, or pretty much be anything else, depending on the get() function of the "type" instance and current indirection level of the Buffer.

---

Buffer#readObject(Number offset) → Object

Returns the JS Object that has previously been written to the Buffer at the given offset using writeObject().

---

Buffer#writeObject(Object obj, Number offset) → undefined

Writes the given JS Object to the Buffer at the given offset. Make sure that at least ref.sizeof.Object bytes are available in the Buffer after the specified offset. The object can later be retrieved using readObject().

obj gets "attached" to the buffer instance, so that the written object won't be garbage collected until the target buffer does.

---

Buffer#readPointer(Number offset, Number size) → Buffer

Returns a new Buffer instance pointing to the address specified in this Buffer at the given offset. The size is the length of the returned Buffer, which defaults to 0.

---

Buffer#writePointer(Buffer pointer, Number offset) → undefined

Writes the given Buffer's memory address to this Buffer at the given offset. Make sure that at least ref.sizeof.pointer bytes are available in the Buffer after the specified offset. The Buffer can later be retrieved again using readPointer().

pointer gets "attached" to the buffer instance, so that the written pointer won't be garbage collected until the target buffer does.

---

Buffer#readCString(Number offset) → String

Returns a JS String from read from the Buffer at the given offset. The C String is read up til the first NULL byte, which indicates the end of the C String.

This function can read beyond the length of a Buffer, and reads up until the first NULL byte regardless.

---

Buffer#writeCString(String string, Number offset, String encoding) → undefined

Writes string as a C String (i.e. NULL terminated) to this Buffer at the given offset. encoding is optional and defaults to utf8.

---

Buffer#readInt64[native-endianness](Number offset) → Number|String

Returns a Number or String representation of the 64-bit int read from this Buffer at the given offset. If the returned value will fit inside a Number without losing precision, then a Number is returned, otherwise a String is returned.

---

Buffer#writeInt64[native-endianness](Number|String value, Number offset) → undefined

Writes an value as a int64_t to this Buffer at the given offset. value may be either a Number or a String representing the 64-bit int value. Ensure that at least ref.sizeof.int64 (always 8) bytes are available in the Buffer after the given offset.

---

Buffer#readUInt64[native-endianness](Number offset) → Number|String

Returns a Number or String representation of the 64-bit unsigned int read from this Buffer at the given offset. If the returned value will fit inside a Number without losing precision, then a Number is returned, otherwise a String is returned.

---

Buffer#writeUInt64[native-endianness](Number|String value, Number offset) → undefined

Writes an value as a int64_t to this Buffer at the given offset. value may be either a Number or a String representing the 64-bit unsigned int value. Ensure that at least ref.sizeof.uint64 (always 8) bytes are available in the Buffer after the given offset.

---

Buffer#reinterpret(Number size) → Buffer

Returns a new Buffer instance with the exact same memory address as the target buffer, only you can specifiy the size of the returned buffer as well.

The original buffer instance gets "attached" to the new buffer instance, so that the original buffer won't be garbage collected until the new buffer does.

Warning: This function is potentially dangerous! There are only a small few use-cases where it really needs to be used (i.e. resizing a Buffer returned from an FFI'd malloc() call), but otherwise, try to avoid it!

Built-in "types"

ref comes with all the basic fixed-size C types that you are probably familiar with:

Name	Description
void	A void type. Derefs to null
int8	Signed 8-bit Integer
uint8	Unsigned 8-bit Integer
int16	Signed 16-bit Integer
uint16	Unsigned 16-bit Integer
int32	Signed 32-bit Integer
uint32	Unsigned 32-bit Integer
int64	Signed 64-bit Integer
uint64	Unsigned 64-bit Integer
float	Single Precision Floating Point Number (float)
double	Double Precision Floating Point Number (double)
Object	A type capable of reading/writing references to JS objects
Utf8String	NULL-terminated String (char *)

In addition to the basic types, there are type aliases for common C types.

Name	Description
bool	bool. Returns/accepts JS true/false values
byte	unsigned char
char	char
uchar	unsigned char
short	short
ushort	unsigned short
int	int
uint	unsigned int
long	long
ulong	unsigned long
longlong	long long
ulonglong	unsigned long long
size_t	platform-dependent, usually pointer size

The "type" interface

You can easily define your own "type" objects at attach to Buffer instances. It just needs to be a regular JavaScript Object that contains the following properties:

Name	Data Type	Description
size	Number	The size in bytes required to hold this type.
indirection	Number	The current level of indirection of the buffer. Usually this would be 1, and gets incremented on Buffers from ref() calls. A value of less than or equal to 0 is invalid.
get	Function (buffer, offset)	The function to invoke when dereferencing this type when the indirection level is 1.
set	Function (buffer, offset, value)	The function to invoke when setting a value to a buffer instance.
alignment	Number	(optional) The alignment of this type when placed in a struct. Defaults to the type's size.

For example, you could define a "bigint" type that dereferences into a bigint instance:

var ref = require('ref')
var bigint = require('bigint')

// define the "type" instance according to the spec
var BigintType = {
    size: ref.sizeof.int64
  , indirection: 1
  , get: function (buffer, offset) {
      // return a bigint instance from the buffer
      return bigint.fromBuffer(buffer)
    }
  , set: function (buffer, offset, value) {
      // 'value' would be a bigint instance
      var val = value.toString()
      return ref.writeInt64(buffer, offset || 0, val)
    }
}

// now we can create instances of the type from existing buffers.
// "buf" is some Buffer instance returned from some external data
// source, which should contain "bigint" binary data.
buf.type = BigintType

// and now you can create "bigint" instances using this generic "types" API
var val = buf.deref()
            .add('1234')
            .sqrt()
            .shiftLeft(5)

License

(The MIT License)

Copyright (c) 2012 Nathan Rajlich <[e-mail]>

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.