Node.js v4.2.2 Documentation

Table of Contents

Buffer#

Stability: 2 - Stable

Pure JavaScript is Unicode friendly but not nice to binary data. When dealing with TCP streams or the file system, it's necessary to handle octet streams. Node.js has several strategies for manipulating, creating, and consuming octet streams.

Raw data is stored in instances of the Buffer class. A Buffer is similar to an array of integers but corresponds to a raw memory allocation outside the V8 heap. A Buffer cannot be resized.

The Buffer class is a global, making it very rare that one would need to ever require('buffer').

Converting between Buffers and JavaScript string objects requires an explicit encoding method. Here are the different string encodings.

  • 'ascii' - for 7 bit ASCII data only. This encoding method is very fast, and will strip the high bit if set.

  • 'utf8' - Multibyte encoded Unicode characters. Many web pages and other document formats use UTF-8.

  • 'utf16le' - 2 or 4 bytes, little endian encoded Unicode characters. Surrogate pairs (U+10000 to U+10FFFF) are supported.

  • 'ucs2' - Alias of 'utf16le'.

  • 'base64' - Base64 string encoding.

  • 'binary' - A way of encoding the buffer into a one-byte (i.e. latin-1) encoded string. The string 'latin-1' is not supported. Instead simply pass 'binary' to use 'latin-1' encoding.

  • 'hex' - Encode each byte as two hexadecimal characters.

Creating a typed array from a Buffer works with the following caveats:

  1. The buffer's memory is copied, not shared.

  2. The buffer's memory is interpreted as an array, not a byte array. That is, new Uint32Array(new Buffer([1,2,3,4])) creates a 4-element Uint32Array with elements [1,2,3,4], not a Uint32Array with a single element [0x1020304] or [0x4030201].

NOTE: Node.js v0.8 simply retained a reference to the buffer in array.buffer instead of cloning it.

While more efficient, it introduces subtle incompatibilities with the typed arrays specification. ArrayBuffer#slice() makes a copy of the slice while Buffer#slice() creates a view.

Class: Buffer#

The Buffer class is a global type for dealing with binary data directly. It can be constructed in a variety of ways.

new Buffer(size)#

  • size Number

Allocates a new buffer of size bytes. size must be less than 1,073,741,824 bytes (1 GB) on 32 bits architectures or 2,147,483,648 bytes (2 GB) on 64 bits architectures, otherwise a RangeError is thrown.

Unlike ArrayBuffers, the underlying memory for buffers is not initialized. So the contents of a newly created Buffer are unknown and could contain sensitive data. Use buf.fill(0) to initialize a buffer to zeroes.

new Buffer(array)#

  • array Array

Allocates a new buffer using an array of octets.

new Buffer(buffer)#

  • buffer Buffer

Copies the passed buffer data onto a new Buffer instance.

new Buffer(str[, encoding])#

  • str String - string to encode.
  • encoding String - encoding to use, Optional.

Allocates a new buffer containing the given str. encoding defaults to 'utf8'.

Class Method: Buffer.isEncoding(encoding)#

  • encoding String The encoding string to test

Returns true if the encoding is a valid encoding argument, or false otherwise.

Class Method: Buffer.isBuffer(obj)#

  • obj Object
  • Return: Boolean

Tests if obj is a Buffer.

Class Method: Buffer.byteLength(string[, encoding])#

  • string String
  • encoding String, Optional, Default: 'utf8'
  • Return: Number

Gives the actual byte length of a string. encoding defaults to 'utf8'. This is not the same as String.prototype.length since that returns the number of characters in a string.

Example: 

str = '\u00bd + \u00bc = \u00be';

console.log(str + ": " + str.length + " characters, " +
  Buffer.byteLength(str, 'utf8') + " bytes");

// ½ + ¼ = ¾: 9 characters, 12 bytes

Class Method: Buffer.concat(list[, totalLength])#

  • list Array List of Buffer objects to concat
  • totalLength Number Total length of the buffers in the list when concatenated

Returns a buffer which is the result of concatenating all the buffers in the list together.

If the list has no items, or if the totalLength is 0, then it returns a zero-length buffer.

If totalLength is not provided, it is read from the buffers in the list. However, this adds an additional loop to the function, so it is faster to provide the length explicitly.

Example: build a single buffer from a list of three buffers: 

var buf1 = new Buffer(10);
var buf2 = new Buffer(14);
var buf3 = new Buffer(18);

buf1.fill(0);
buf2.fill(0);
buf3.fill(0);

var buffers = [buf1, buf2, buf3];

var totalLength = 0;
for (var i = 0; i < buffers.length; i++) {
  totalLength += buffers[i].length;
}

console.log(totalLength);
var bufA = Buffer.concat(buffers, totalLength);
console.log(bufA);
console.log(bufA.length);

// 42
// <Buffer 00 00 00 00 ...>
// 42

Class Method: Buffer.compare(buf1, buf2)#

  • buf1 Buffer
  • buf2 Buffer

The same as buf1.compare(buf2). Useful for sorting an Array of Buffers: 

var arr = [Buffer('1234'), Buffer('0123')];
arr.sort(Buffer.compare);

buf.length#

  • Number

The size of the buffer in bytes. Note that this is not necessarily the size of the contents. length refers to the amount of memory allocated for the buffer object. It does not change when the contents of the buffer are changed. 

buf = new Buffer(1234);

console.log(buf.length);
buf.write("some string", 0, "ascii");
console.log(buf.length);

// 1234
// 1234

While the length property is not immutable, changing the value of length can result in undefined and inconsistent behavior. Applications that wish to modify the length of a buffer should therefore treat length as read-only and use buf.slice to create a new buffer. 

buf = new Buffer(10);
buf.write("abcdefghj", 0, "ascii");
console.log(buf.length); // 10
buf = buf.slice(0,5);
console.log(buf.length); // 5

buf.write(string[, offset][, length][, encoding])#

  • string String - data to be written to buffer
  • offset Number, Optional, Default: 0
  • length Number, Optional, Default: buffer.length - offset
  • encoding String, Optional, Default: 'utf8'

Writes string to the buffer at offset using the given encoding. offset defaults to 0, encoding defaults to 'utf8'. length is the number of bytes to write. Returns number of octets written. If buffer did not contain enough space to fit the entire string, it will write a partial amount of the string. length defaults to buffer.length - offset. The method will not write partial characters. 

buf = new Buffer(256);
len = buf.write('\u00bd + \u00bc = \u00be', 0);
console.log(len + " bytes: " + buf.toString('utf8', 0, len));

buf.writeUIntLE(value, offset, byteLength[, noAssert])#

buf.writeUIntBE(value, offset, byteLength[, noAssert])#

buf.writeIntLE(value, offset, byteLength[, noAssert])#

buf.writeIntBE(value, offset, byteLength[, noAssert])#

  • value {Number} Bytes to be written to buffer
  • offset {Number} 0 <= offset <= buf.length
  • byteLength {Number} 0 < byteLength <= 6
  • noAssert {Boolean} Default: false
  • Return: {Number}

Writes value to the buffer at the specified offset and byteLength. Supports up to 48 bits of accuracy. For example: 

var b = new Buffer(6);
b.writeUIntBE(0x1234567890ab, 0, 6);
// <Buffer 12 34 56 78 90 ab>

Set noAssert to true to skip validation of value and offset. Defaults to false.

buf.readUIntLE(offset, byteLength[, noAssert])#

buf.readUIntBE(offset, byteLength[, noAssert])#

buf.readIntLE(offset, byteLength[, noAssert])#

buf.readIntBE(offset, byteLength[, noAssert])#

  • offset {Number} 0 <= offset <= buf.length
  • byteLength {Number} 0 < byteLength <= 6
  • noAssert {Boolean} Default: false
  • Return: {Number}

A generalized version of all numeric read methods. Supports up to 48 bits of accuracy. For example: 

var b = new Buffer(6);
b.writeUInt16LE(0x90ab, 0);
b.writeUInt32LE(0x12345678, 2);
b.readUIntLE(0, 6).toString(16);  // Specify 6 bytes (48 bits)
// output: '1234567890ab'

Set noAssert to true to skip validation of offset. This means that offset may be beyond the end of the buffer. Defaults to false.

buf.toString([encoding][, start][, end])#

  • encoding String, Optional, Default: 'utf8'
  • start Number, Optional, Default: 0
  • end Number, Optional, Default: buffer.length

Decodes and returns a string from buffer data encoded using the specified character set encoding. If encoding is undefined or null, then encoding defaults to 'utf8'. The start and end parameters default to 0 and buffer.length when undefined. 

buf = new Buffer(26);
for (var i = 0 ; i < 26 ; i++) {
  buf[i] = i + 97; // 97 is ASCII a
}
buf.toString('ascii'); // outputs: abcdefghijklmnopqrstuvwxyz
buf.toString('ascii',0,5); // outputs: abcde
buf.toString('utf8',0,5); // outputs: abcde
buf.toString(undefined,0,5); // encoding defaults to 'utf8', outputs abcde

See buffer.write() example, above.

buf.toJSON()#

Returns a JSON-representation of the Buffer instance. JSON.stringify implicitly calls this function when stringifying a Buffer instance.

Example: 

var buf = new Buffer('test');
var json = JSON.stringify(buf);

console.log(json);
// '{"type":"Buffer","data":[116,101,115,116]}'

var copy = JSON.parse(json, function(key, value) {
    return value && value.type === 'Buffer'
      ? new Buffer(value.data)
      : value;
  });

console.log(copy);
// <Buffer 74 65 73 74>

buf[index]#

Get and set the octet at index. The values refer to individual bytes, so the legal range is between 0x00 and 0xFF hex or 0 and 255.

Example: copy an ASCII string into a buffer, one byte at a time: 

str = "Node.js";
buf = new Buffer(str.length);

for (var i = 0; i < str.length ; i++) {
  buf[i] = str.charCodeAt(i);
}

console.log(buf);

// Node.js

buf.equals(otherBuffer)#

  • otherBuffer Buffer

Returns a boolean of whether this and otherBuffer have the same bytes.

buf.compare(otherBuffer)#

  • otherBuffer Buffer

Returns a number indicating whether this comes before or after or is the same as the otherBuffer in sort order.

buf.copy(targetBuffer[, targetStart][, sourceStart][, sourceEnd])#

  • targetBuffer Buffer object - Buffer to copy into
  • targetStart Number, Optional, Default: 0
  • sourceStart Number, Optional, Default: 0
  • sourceEnd Number, Optional, Default: buffer.length

Copies data from a region of this buffer to a region in the target buffer even if the target memory region overlaps with the source. If undefined the targetStart and sourceStart parameters default to 0 while sourceEnd defaults to buffer.length.

Example: build two Buffers, then copy buf1 from byte 16 through byte 19 into buf2, starting at the 8th byte in buf2. 

buf1 = new Buffer(26);
buf2 = new Buffer(26);

for (var i = 0 ; i < 26 ; i++) {
  buf1[i] = i + 97; // 97 is ASCII a
  buf2[i] = 33; // ASCII !
}

buf1.copy(buf2, 8, 16, 20);
console.log(buf2.toString('ascii', 0, 25));

// !!!!!!!!qrst!!!!!!!!!!!!!

Example: Build a single buffer, then copy data from one region to an overlapping region in the same buffer 

buf = new Buffer(26);

for (var i = 0 ; i < 26 ; i++) {
  buf[i] = i + 97; // 97 is ASCII a
}

buf.copy(buf, 0, 4, 10);
console.log(buf.toString());

// efghijghijklmnopqrstuvwxyz

buf.slice([start[, end]])#

  • start Number, Optional, Default: 0
  • end Number, Optional, Default: buffer.length

Returns a new buffer which references the same memory as the old, but offset and cropped by the start (defaults to 0) and end (defaults to buffer.length) indexes. Negative indexes start from the end of the buffer.

Modifying the new buffer slice will modify memory in the original buffer!

Example: build a Buffer with the ASCII alphabet, take a slice, then modify one byte from the original Buffer. 

var buf1 = new Buffer(26);

for (var i = 0 ; i < 26 ; i++) {
  buf1[i] = i + 97; // 97 is ASCII a
}

var buf2 = buf1.slice(0, 3);
console.log(buf2.toString('ascii', 0, buf2.length));
buf1[0] = 33;
console.log(buf2.toString('ascii', 0, buf2.length));

// abc
// !bc

buf.indexOf(value[, byteOffset])#

  • value String, Buffer or Number
  • byteOffset Number, Optional, Default: 0
  • Return: Number

Operates similar to Array#indexOf(). Accepts a String, Buffer or Number. Strings are interpreted as UTF8. Buffers will use the entire buffer. So in order to compare a partial Buffer use Buffer#slice(). Numbers can range from 0 to 255.

buf.readUInt8(offset[, noAssert])#

  • offset Number
  • noAssert Boolean, Optional, Default: false
  • Return: Number

Reads an unsigned 8 bit integer from the buffer at the specified offset.

Set noAssert to true to skip validation of offset. This means that offset may be beyond the end of the buffer. Defaults to false.

Example: 

var buf = new Buffer(4);

buf[0] = 0x3;
buf[1] = 0x4;
buf[2] = 0x23;
buf[3] = 0x42;

for (ii = 0; ii < buf.length; ii++) {
  console.log(buf.readUInt8(ii));
}

// 0x3
// 0x4
// 0x23
// 0x42

buf.readUInt16LE(offset[, noAssert])#

buf.readUInt16BE(offset[, noAssert])#

  • offset Number
  • noAssert Boolean, Optional, Default: false
  • Return: Number

Reads an unsigned 16 bit integer from the buffer at the specified offset with specified endian format.

Set noAssert to true to skip validation of offset. This means that offset may be beyond the end of the buffer. Defaults to false.

Example: 

var buf = new Buffer(4);

buf[0] = 0x3;
buf[1] = 0x4;
buf[2] = 0x23;
buf[3] = 0x42;

console.log(buf.readUInt16BE(0));
console.log(buf.readUInt16LE(0));
console.log(buf.readUInt16BE(1));
console.log(buf.readUInt16LE(1));
console.log(buf.readUInt16BE(2));
console.log(buf.readUInt16LE(2));

// 0x0304
// 0x0403
// 0x0423
// 0x2304
// 0x2342
// 0x4223

buf.readUInt32LE(offset[, noAssert])#

buf.readUInt32BE(offset[, noAssert])#

  • offset Number
  • noAssert Boolean, Optional, Default: false
  • Return: Number

Reads an unsigned 32 bit integer from the buffer at the specified offset with specified endian format.

Set noAssert to true to skip validation of offset. This means that offset may be beyond the end of the buffer. Defaults to false.

Example: 

var buf = new Buffer(4);

buf[0] = 0x3;
buf[1] = 0x4;
buf[2] = 0x23;
buf[3] = 0x42;

console.log(buf.readUInt32BE(0));
console.log(buf.readUInt32LE(0));

// 0x03042342
// 0x42230403

buf.readInt8(offset[, noAssert])#

  • offset Number
  • noAssert Boolean, Optional, Default: false
  • Return: Number

Reads a signed 8 bit integer from the buffer at the specified offset.

Set noAssert to true to skip validation of offset. This means that offset may be beyond the end of the buffer. Defaults to false.

Works as buffer.readUInt8, except buffer contents are treated as two's complement signed values.

buf.readInt16LE(offset[, noAssert])#

buf.readInt16BE(offset[, noAssert])#

  • offset Number
  • noAssert Boolean, Optional, Default: false
  • Return: Number

Reads a signed 16 bit integer from the buffer at the specified offset with specified endian format.

Set noAssert to true to skip validation of offset. This means that offset may be beyond the end of the buffer. Defaults to false.

Works as buffer.readUInt16*, except buffer contents are treated as two's complement signed values.

buf.readInt32LE(offset[, noAssert])#

buf.readInt32BE(offset[, noAssert])#

  • offset Number
  • noAssert Boolean, Optional, Default: false
  • Return: Number

Reads a signed 32 bit integer from the buffer at the specified offset with specified endian format.

Set noAssert to true to skip validation of offset. This means that offset may be beyond the end of the buffer. Defaults to false.

Works as buffer.readUInt32*, except buffer contents are treated as two's complement signed values.

buf.readFloatLE(offset[, noAssert])#

buf.readFloatBE(offset[, noAssert])#

  • offset Number
  • noAssert Boolean, Optional, Default: false
  • Return: Number

Reads a 32 bit float from the buffer at the specified offset with specified endian format.

Set noAssert to true to skip validation of offset. This means that offset may be beyond the end of the buffer. Defaults to false.

Example: 

var buf = new Buffer(4);

buf[0] = 0x00;
buf[1] = 0x00;
buf[2] = 0x80;
buf[3] = 0x3f;

console.log(buf.readFloatLE(0));

// 0x01

buf.readDoubleLE(offset[, noAssert])#

buf.readDoubleBE(offset[, noAssert])#

  • offset Number
  • noAssert Boolean, Optional, Default: false
  • Return: Number

Reads a 64 bit double from the buffer at the specified offset with specified endian format.

Set noAssert to true to skip validation of offset. This means that offset may be beyond the end of the buffer. Defaults to false.

Example: 

var buf = new Buffer(8);

buf[0] = 0x55;
buf[1] = 0x55;
buf[2] = 0x55;
buf[3] = 0x55;
buf[4] = 0x55;
buf[5] = 0x55;
buf[6] = 0xd5;
buf[7] = 0x3f;

console.log(buf.readDoubleLE(0));

// 0.3333333333333333

buf.writeUInt8(value, offset[, noAssert])#

  • value Number
  • offset Number
  • noAssert Boolean, Optional, Default: false

Writes value to the buffer at the specified offset. Note, value must be a valid unsigned 8 bit integer.

Set noAssert to true to skip validation of value and offset. This means that value may be too large for the specific function and offset may be beyond the end of the buffer leading to the values being silently dropped. This should not be used unless you are certain of correctness. Defaults to false.

Example: 

var buf = new Buffer(4);
buf.writeUInt8(0x3, 0);
buf.writeUInt8(0x4, 1);
buf.writeUInt8(0x23, 2);
buf.writeUInt8(0x42, 3);

console.log(buf);

// <Buffer 03 04 23 42>

buf.writeUInt16LE(value, offset[, noAssert])#

buf.writeUInt16BE(value, offset[, noAssert])#

  • value Number
  • offset Number
  • noAssert Boolean, Optional, Default: false

Writes value to the buffer at the specified offset with specified endian format. Note, value must be a valid unsigned 16 bit integer.

Set noAssert to true to skip validation of value and offset. This means that value may be too large for the specific function and offset may be beyond the end of the buffer leading to the values being silently dropped. This should not be used unless you are certain of correctness. Defaults to false.

Example: 

var buf = new Buffer(4);
buf.writeUInt16BE(0xdead, 0);
buf.writeUInt16BE(0xbeef, 2);

console.log(buf);

buf.writeUInt16LE(0xdead, 0);
buf.writeUInt16LE(0xbeef, 2);

console.log(buf);

// <Buffer de ad be ef>
// <Buffer ad de ef be>

buf.writeUInt32LE(value, offset[, noAssert])#

buf.writeUInt32BE(value, offset[, noAssert])#

  • value Number
  • offset Number
  • noAssert Boolean, Optional, Default: false

Writes value to the buffer at the specified offset with specified endian format. Note, value must be a valid unsigned 32 bit integer.

Set noAssert to true to skip validation of value and offset. This means that value may be too large for the specific function and offset may be beyond the end of the buffer leading to the values being silently dropped. This should not be used unless you are certain of correctness. Defaults to false.

Example: 

var buf = new Buffer(4);
buf.writeUInt32BE(0xfeedface, 0);

console.log(buf);

buf.writeUInt32LE(0xfeedface, 0);

console.log(buf);

// <Buffer fe ed fa ce>
// <Buffer ce fa ed fe>

buf.writeInt8(value, offset[, noAssert])#

  • value Number
  • offset Number
  • noAssert Boolean, Optional, Default: false

Writes value to the buffer at the specified offset. Note, value must be a valid signed 8 bit integer.

Set noAssert to true to skip validation of value and offset. This means that value may be too large for the specific function and offset may be beyond the end of the buffer leading to the values being silently dropped. This should not be used unless you are certain of correctness. Defaults to false.

Works as buffer.writeUInt8, except value is written out as a two's complement signed integer into buffer.

buf.writeInt16LE(value, offset[, noAssert])#

buf.writeInt16BE(value, offset[, noAssert])#

  • value Number
  • offset Number
  • noAssert Boolean, Optional, Default: false

Writes value to the buffer at the specified offset with specified endian format. Note, value must be a valid signed 16 bit integer.

Set noAssert to true to skip validation of value and offset. This means that value may be too large for the specific function and offset may be beyond the end of the buffer leading to the values being silently dropped. This should not be used unless you are certain of correctness. Defaults to false.

Works as buffer.writeUInt16*, except value is written out as a two's complement signed integer into buffer.

buf.writeInt32LE(value, offset[, noAssert])#

buf.writeInt32BE(value, offset[, noAssert])#

  • value Number
  • offset Number
  • noAssert Boolean, Optional, Default: false

Writes value to the buffer at the specified offset with specified endian format. Note, value must be a valid signed 32 bit integer.

Set noAssert to true to skip validation of value and offset. This means that value may be too large for the specific function and offset may be beyond the end of the buffer leading to the values being silently dropped. This should not be used unless you are certain of correctness. Defaults to false.

Works as buffer.writeUInt32*, except value is written out as a two's complement signed integer into buffer.

buf.writeFloatLE(value, offset[, noAssert])#

buf.writeFloatBE(value, offset[, noAssert])#

  • value Number
  • offset Number
  • noAssert Boolean, Optional, Default: false

Writes value to the buffer at the specified offset with specified endian format. Note, behavior is unspecified if value is not a 32 bit float.

Set noAssert to true to skip validation of value and offset. This means that value may be too large for the specific function and offset may be beyond the end of the buffer leading to the values being silently dropped. This should not be used unless you are certain of correctness. Defaults to false.

Example: 

var buf = new Buffer(4);
buf.writeFloatBE(0xcafebabe, 0);

console.log(buf);

buf.writeFloatLE(0xcafebabe, 0);

console.log(buf);

// <Buffer 4f 4a fe bb>
// <Buffer bb fe 4a 4f>

buf.writeDoubleLE(value, offset[, noAssert])#

buf.writeDoubleBE(value, offset[, noAssert])#

  • value Number
  • offset Number
  • noAssert Boolean, Optional, Default: false

Writes value to the buffer at the specified offset with specified endian format. Note, value must be a valid 64 bit double.

Set noAssert to true to skip validation of value and offset. This means that value may be too large for the specific function and offset may be beyond the end of the buffer leading to the values being silently dropped. This should not be used unless you are certain of correctness. Defaults to false.

Example: 

var buf = new Buffer(8);
buf.writeDoubleBE(0xdeadbeefcafebabe, 0);

console.log(buf);

buf.writeDoubleLE(0xdeadbeefcafebabe, 0);

console.log(buf);

// <Buffer 43 eb d5 b7 dd f9 5f d7>
// <Buffer d7 5f f9 dd b7 d5 eb 43>

buf.fill(value[, offset][, end])#

  • value
  • offset Number, Optional
  • end Number, Optional

Fills the buffer with the specified value. If the offset (defaults to 0) and end (defaults to buffer.length) are not given it will fill the entire buffer. 

var b = new Buffer(50);
b.fill("h");

buffer.values()#

Creates iterator for buffer values (bytes). This function is called automatically when buffer is used in a for..of statement.

buffer.keys()#

Creates iterator for buffer keys (indices).

buffer.entries()#

Creates iterator for [index, byte] arrays.

buffer.INSPECT_MAX_BYTES#

  • Number, Default: 50

How many bytes will be returned when buffer.inspect() is called. This can be overridden by user modules.

Note that this is a property on the buffer module returned by require('buffer'), not on the Buffer global, or a buffer instance.

ES6 iteration#

Buffers can be iterated over using for..of syntax: 

var buf = new Buffer([1, 2, 3]);

for (var b of buf)
  console.log(b)

// 1
// 2
// 3

Additionally, buffer.values(), buffer.keys() and buffer.entries() methods can be used to create iterators.

Class: SlowBuffer#

Returns an un-pooled Buffer.

In order to avoid the garbage collection overhead of creating many individually allocated Buffers, by default allocations under 4KB are sliced from a single larger allocated object. This approach improves both performance and memory usage since v8 does not need to track and cleanup as many Persistent objects.

In the case where a developer may need to retain a small chunk of memory from a pool for an indeterminate amount of time it may be appropriate to create an un-pooled Buffer instance using SlowBuffer and copy out the relevant bits. 

// need to keep around a few small chunks of memory
var store = [];

socket.on('readable', function() {
  var data = socket.read();
  // allocate for retained data
  var sb = new SlowBuffer(10);
  // copy the data into the new allocation
  data.copy(sb, 0, 0, 10);
  store.push(sb);
});

Though this should be used sparingly and only be a last resort after a developer has actively observed undue memory retention in their applications.