JavaScript Numbers are represented as IEEE 754 double-precision floats. Unfortunately, this means they lose integer precision for values beyond +/- 2^^53. For projects that need to accurately handle 64-bit ints, such as node-thrift, a performant, Number-like class is needed. Int64 is that class.
Int64 instances look and feel much like JS-native Numbers. By way of example ...
// First, let's illustrate the problem ...
> (0x123456789).toString(16)
'123456789' // <- what we expect.
> (0x123456789abcdef0).toString(16)
'123456789abcdf00' // <- Ugh! JS doesn't do big ints. :(
// So let's create a couple Int64s using the above values ...
// Require, of course
> Int64 = require('node-int64')
// x's value is what we expect (the decimal value of 0x123456789)
> x = new Int64(0x123456789)
[Int64 value:4886718345 octets:00 00 00 01 23 45 67 89]
// y's value is Infinity because it's outside the range of integer
// precision. But that's okay - it's still useful because it's internal
// representation (octets) is what we passed in
> y = new Int64('123456789abcdef0')
[Int64 value:Infinity octets:12 34 56 78 9a bc de f0]
// Let's do some math. Int64's behave like Numbers. (Sorry, Int64 isn't
// for doing 64-bit integer arithmetic (yet) - it's just for carrying
// around int64 values
> x + 1
4886718346
> y + 1
Infinity
// Int64 string operations ...
> 'value: ' + x
'value: 4886718345'
> 'value: ' + y
'value: Infinity'
> x.toString(2)
'100100011010001010110011110001001'
> y.toString(2)
'Infinity'
// Use JS's isFinite() method to see if the Int64 value is in the
// integer-precise range of JS values
> isFinite(x)
true
> isFinite(y)
false
// Get an octet string representation. (Yay, y is what we put in!)
> x.toOctetString()
'0000000123456789'
> y.toOctetString()
'123456789abcdef0'
// Finally, some other ways to create Int64s ...
// Pass hi/lo words
> new Int64(0x12345678, 0x9abcdef0)
[Int64 value:Infinity octets:12 34 56 78 9a bc de f0]
// Pass a Buffer
> new Int64(new Buffer([0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc, 0xde, 0xf0]))
[Int64 value:Infinity octets:12 34 56 78 9a bc de f0]
// Pass a Buffer and offset
> new Int64(new Buffer([0,0,0,0,0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc, 0xde, 0xf0]), 4)
[Int64 value:Infinity octets:12 34 56 78 9a bc de f0]
// Pull out into a buffer
> new Int64(new Buffer([0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc, 0xde, 0xf0])).toBuffer()
<Buffer 12 34 56 78 9a bc de f0>
// Or copy into an existing one (at an offset)
> var buf = new Buffer(1024);
> new Int64(new Buffer([0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc, 0xde, 0xf0])).copy(buf, 512);
JavaScript Numbers are represented as [IEEE 754 double-precision floats](http://steve.hollasch.net/cgindex/coding/ieeefloat.html). Unfortunately, this means they lose integer precision for values beyond +/- 2^^53. For projects that need to accurately handle 64-bit ints, such as [node-thrift](https://github.com/wadey/node-thrift), a performant, Number-like class is needed. Int64 is that class. Int64 instances look and feel much like JS-native Numbers. By way of example ... ```js // First, let's illustrate the problem ... > (0x123456789).toString(16) '123456789' // <- what we expect. > (0x123456789abcdef0).toString(16) '123456789abcdf00' // <- Ugh! JS doesn't do big ints. :( // So let's create a couple Int64s using the above values ... // Require, of course > Int64 = require('node-int64') // x's value is what we expect (the decimal value of 0x123456789) > x = new Int64(0x123456789) [Int64 value:4886718345 octets:00 00 00 01 23 45 67 89] // y's value is Infinity because it's outside the range of integer // precision. But that's okay - it's still useful because it's internal // representation (octets) is what we passed in > y = new Int64('123456789abcdef0') [Int64 value:Infinity octets:12 34 56 78 9a bc de f0] // Let's do some math. Int64's behave like Numbers. (Sorry, Int64 isn't // for doing 64-bit integer arithmetic (yet) - it's just for carrying // around int64 values > x + 1 4886718346 > y + 1 Infinity // Int64 string operations ... > 'value: ' + x 'value: 4886718345' > 'value: ' + y 'value: Infinity' > x.toString(2) '100100011010001010110011110001001' > y.toString(2) 'Infinity' // Use JS's isFinite() method to see if the Int64 value is in the // integer-precise range of JS values > isFinite(x) true > isFinite(y) false // Get an octet string representation. (Yay, y is what we put in!) > x.toOctetString() '0000000123456789' > y.toOctetString() '123456789abcdef0' // Finally, some other ways to create Int64s ... // Pass hi/lo words > new Int64(0x12345678, 0x9abcdef0) [Int64 value:Infinity octets:12 34 56 78 9a bc de f0] // Pass a Buffer > new Int64(new Buffer([0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc, 0xde, 0xf0])) [Int64 value:Infinity octets:12 34 56 78 9a bc de f0] // Pass a Buffer and offset > new Int64(new Buffer([0,0,0,0,0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc, 0xde, 0xf0]), 4) [Int64 value:Infinity octets:12 34 56 78 9a bc de f0] // Pull out into a buffer > new Int64(new Buffer([0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc, 0xde, 0xf0])).toBuffer() <Buffer 12 34 56 78 9a bc de f0> // Or copy into an existing one (at an offset) > var buf = new Buffer(1024); > new Int64(new Buffer([0x12, 0x34, 0x56, 0x78, 0x9a, 0xbc, 0xde, 0xf0])).copy(buf, 512); ```