Javascript's Number data type is insufficient for arithmetic that requires specific byte-width accuracy. Bitwise operations are only permitted on values up to 32 bits, and integer precision is lost for values larger than 2^53-1. This module provides a FixedInt data type for cases where more precise behavior is required for integers of 8, 16, 32, or 64 bits.
| FixedInt method | Return type | Description |
|---|---|---|
FixedInt(size, lo, hi) |
FixedInt |
Construct a new FixedInt of size from the Number lo, (and hi if size is 8) |
FixedInt(size, buf, offset, le) |
FixedInt |
Construct a new FixedInt by reading size bytes from the ArrayBuffer wrapped by buf, starting at offset |
FixedInt(other) |
FixedInt |
Construct a new FixedInt form another FixedInt or object containing properites size, lo, (and hi) |
isNegative() |
boolean |
Return true if this object is negative when interpreted as signed |
isOdd() |
boolean |
Return true if the number represented by this object is odd |
isSafeInteger() |
boolean |
Return true if this number is less than 2^53-1 |
isLessThan(that) |
boolean |
Return true if the number referenced by that is greater than this, when both are interpretted as unsigned |
equals(that) |
boolean |
Return true if this integer has the same value as that |
clone() |
FixedInt |
Return a new FixedInt with the same value and size as this |
toSize(size, signExtend) |
FixedInt |
Return a new FixedInt of size size with the same value as this, truncated if size < this.size, or zero or sign extended if not, according to signExtend |
toBuffer(buf, offset) |
ArrayBuffer |
Write this to a value |
valueOf(signed) |
Number |
Return a Number with the same value as this, or the Number with the value closest to this if it is 2^53 or greater. |
toString(radix, signed) |
String |
Return a string representation of this, in base radix. If radix = 10, decode this as a signed integer according to signed |
The other class exported by FixedInt is the ALU. All of the ALU methods are static and accept FixedInts for both arguments. If the second argument is an integer, it will be coerced to a FixedInt of the same size as the first.
| ALU Method | Description |
|---|---|
add(a, b) |
Add b to a |
sub(a, b) |
Subtract b from a |
mul(a, b) |
Multiply a by b. If a has size, then the lower size bytes of the result are returned, and the upper size bytes are stored on ALU.aux |
div(a, b) |
Divide a by b. The remainder/modulus is stored on ALU.aux |
sar(a, b) |
Shift a right by b bits, preserving the sign of a |
shr(a, b) |
Shift a right by b bits, filling with zeros |
shl(a, b) |
Shift a left by b bits, shifting in zeros |
and(a, b) |
Return the bitwise and of a and b, i.e. a & b |
xor(a, b) |
Return the bitwise exclusive or of a and b, i.e. a ^ b |
or(a, b) |
Return the bitwise or of a and b, i.e. `a |
not(a) |
Return the 1s complement of a, i.e. ~a |
neg(a) |
Return the 2s complement of a, i.e. -a = ~a + 1 |
The ALU also exposes 5 properties which are set according to the results of particular operations.
| Property | Description |
|---|---|
ALU.aux |
Auxiliary value. Is set to the remainder after a call to div, or the upper half of the result after mul. |
ALU.CF |
Carry flag, indicates that the preceding operation caused a carry out of (or borrow into) the most significant bit. |
ALU.OF |
Overflow flag, indicates that the carry into and out of the sign bits were not equal. |
ALU.ZF |
Zero flag, indicates that the result of the last operation was zero |
ALU.SF |
Sign flag, indicates that the result of the last operation was negative, when interpretted as signed. |