zig/lib/std/math.zig

1095 lines
37 KiB
Zig

const std = @import("std.zig");
const assert = std.debug.assert;
const testing = std.testing;
/// Euler's number (e)
pub const e = 2.71828182845904523536028747135266249775724709369995;
/// Archimedes' constant (π)
pub const pi = 3.14159265358979323846264338327950288419716939937510;
/// Circle constant (τ)
pub const tau = 2 * pi;
/// log2(e)
pub const log2e = 1.442695040888963407359924681001892137;
/// log10(e)
pub const log10e = 0.434294481903251827651128918916605082;
/// ln(2)
pub const ln2 = 0.693147180559945309417232121458176568;
/// ln(10)
pub const ln10 = 2.302585092994045684017991454684364208;
/// 2/sqrt(π)
pub const two_sqrtpi = 1.128379167095512573896158903121545172;
/// sqrt(2)
pub const sqrt2 = 1.414213562373095048801688724209698079;
/// 1/sqrt(2)
pub const sqrt1_2 = 0.707106781186547524400844362104849039;
// From a small c++ [program using boost float128](https://github.com/winksaville/cpp_boost_float128)
pub const f128_true_min = @bitCast(f128, @as(u128, 0x00000000000000000000000000000001));
pub const f128_min = @bitCast(f128, @as(u128, 0x00010000000000000000000000000000));
pub const f128_max = @bitCast(f128, @as(u128, 0x7FFEFFFFFFFFFFFFFFFFFFFFFFFFFFFF));
pub const f128_epsilon = @bitCast(f128, @as(u128, 0x3F8F0000000000000000000000000000));
pub const f128_toint = 1.0 / f128_epsilon;
// float.h details
pub const f64_true_min = 4.94065645841246544177e-324;
pub const f64_min = 2.2250738585072014e-308;
pub const f64_max = 1.79769313486231570815e+308;
pub const f64_epsilon = 2.22044604925031308085e-16;
pub const f64_toint = 1.0 / f64_epsilon;
pub const f32_true_min = 1.40129846432481707092e-45;
pub const f32_min = 1.17549435082228750797e-38;
pub const f32_max = 3.40282346638528859812e+38;
pub const f32_epsilon = 1.1920928955078125e-07;
pub const f32_toint = 1.0 / f32_epsilon;
pub const f16_true_min = 0.000000059604644775390625; // 2**-24
pub const f16_min = 0.00006103515625; // 2**-14
pub const f16_max = 65504;
pub const f16_epsilon = 0.0009765625; // 2**-10
pub const f16_toint = 1.0 / f16_epsilon;
pub const nan_u16 = @as(u16, 0x7C01);
pub const nan_f16 = @bitCast(f16, nan_u16);
pub const qnan_u16 = @as(u16, 0x7E00);
pub const qnan_f16 = @bitCast(f16, qnan_u16);
pub const inf_u16 = @as(u16, 0x7C00);
pub const inf_f16 = @bitCast(f16, inf_u16);
pub const nan_u32 = @as(u32, 0x7F800001);
pub const nan_f32 = @bitCast(f32, nan_u32);
pub const qnan_u32 = @as(u32, 0x7FC00000);
pub const qnan_f32 = @bitCast(f32, qnan_u32);
pub const inf_u32 = @as(u32, 0x7F800000);
pub const inf_f32 = @bitCast(f32, inf_u32);
pub const nan_u64 = @as(u64, 0x7FF << 52) | 1;
pub const nan_f64 = @bitCast(f64, nan_u64);
pub const qnan_u64 = @as(u64, 0x7ff8000000000000);
pub const qnan_f64 = @bitCast(f64, qnan_u64);
pub const inf_u64 = @as(u64, 0x7FF << 52);
pub const inf_f64 = @bitCast(f64, inf_u64);
pub const nan_u128 = @as(u128, 0x7fff0000000000000000000000000001);
pub const nan_f128 = @bitCast(f128, nan_u128);
pub const qnan_u128 = @as(u128, 0x7fff8000000000000000000000000000);
pub const qnan_f128 = @bitCast(f128, qnan_u128);
pub const inf_u128 = @as(u128, 0x7fff0000000000000000000000000000);
pub const inf_f128 = @bitCast(f128, inf_u128);
pub const nan = @import("math/nan.zig").nan;
pub const snan = @import("math/nan.zig").snan;
pub const inf = @import("math/inf.zig").inf;
pub fn approxEq(comptime T: type, x: T, y: T, epsilon: T) bool {
assert(@typeInfo(T) == .Float);
return fabs(x - y) < epsilon;
}
// TODO: Hide the following in an internal module.
pub fn forceEval(value: anytype) void {
const T = @TypeOf(value);
switch (T) {
f16 => {
var x: f16 = undefined;
const p = @ptrCast(*volatile f16, &x);
p.* = x;
},
f32 => {
var x: f32 = undefined;
const p = @ptrCast(*volatile f32, &x);
p.* = x;
},
f64 => {
var x: f64 = undefined;
const p = @ptrCast(*volatile f64, &x);
p.* = x;
},
f128 => {
var x: f128 = undefined;
const p = @ptrCast(*volatile f128, &x);
p.* = x;
},
else => {
@compileError("forceEval not implemented for " ++ @typeName(T));
},
}
}
pub fn raiseInvalid() void {
// Raise INVALID fpu exception
}
pub fn raiseUnderflow() void {
// Raise UNDERFLOW fpu exception
}
pub fn raiseOverflow() void {
// Raise OVERFLOW fpu exception
}
pub fn raiseInexact() void {
// Raise INEXACT fpu exception
}
pub fn raiseDivByZero() void {
// Raise INEXACT fpu exception
}
pub const isNan = @import("math/isnan.zig").isNan;
pub const isSignalNan = @import("math/isnan.zig").isSignalNan;
pub const fabs = @import("math/fabs.zig").fabs;
pub const ceil = @import("math/ceil.zig").ceil;
pub const floor = @import("math/floor.zig").floor;
pub const trunc = @import("math/trunc.zig").trunc;
pub const round = @import("math/round.zig").round;
pub const frexp = @import("math/frexp.zig").frexp;
pub const frexp32_result = @import("math/frexp.zig").frexp32_result;
pub const frexp64_result = @import("math/frexp.zig").frexp64_result;
pub const modf = @import("math/modf.zig").modf;
pub const modf32_result = @import("math/modf.zig").modf32_result;
pub const modf64_result = @import("math/modf.zig").modf64_result;
pub const copysign = @import("math/copysign.zig").copysign;
pub const isFinite = @import("math/isfinite.zig").isFinite;
pub const isInf = @import("math/isinf.zig").isInf;
pub const isPositiveInf = @import("math/isinf.zig").isPositiveInf;
pub const isNegativeInf = @import("math/isinf.zig").isNegativeInf;
pub const isNormal = @import("math/isnormal.zig").isNormal;
pub const signbit = @import("math/signbit.zig").signbit;
pub const scalbn = @import("math/scalbn.zig").scalbn;
pub const pow = @import("math/pow.zig").pow;
pub const powi = @import("math/powi.zig").powi;
pub const sqrt = @import("math/sqrt.zig").sqrt;
pub const cbrt = @import("math/cbrt.zig").cbrt;
pub const acos = @import("math/acos.zig").acos;
pub const asin = @import("math/asin.zig").asin;
pub const atan = @import("math/atan.zig").atan;
pub const atan2 = @import("math/atan2.zig").atan2;
pub const hypot = @import("math/hypot.zig").hypot;
pub const exp = @import("math/exp.zig").exp;
pub const exp2 = @import("math/exp2.zig").exp2;
pub const expm1 = @import("math/expm1.zig").expm1;
pub const ilogb = @import("math/ilogb.zig").ilogb;
pub const ln = @import("math/ln.zig").ln;
pub const log = @import("math/log.zig").log;
pub const log2 = @import("math/log2.zig").log2;
pub const log10 = @import("math/log10.zig").log10;
pub const log1p = @import("math/log1p.zig").log1p;
pub const fma = @import("math/fma.zig").fma;
pub const asinh = @import("math/asinh.zig").asinh;
pub const acosh = @import("math/acosh.zig").acosh;
pub const atanh = @import("math/atanh.zig").atanh;
pub const sinh = @import("math/sinh.zig").sinh;
pub const cosh = @import("math/cosh.zig").cosh;
pub const tanh = @import("math/tanh.zig").tanh;
pub const cos = @import("math/cos.zig").cos;
pub const sin = @import("math/sin.zig").sin;
pub const tan = @import("math/tan.zig").tan;
pub const complex = @import("math/complex.zig");
pub const Complex = complex.Complex;
pub const big = @import("math/big.zig");
test "" {
std.meta.refAllDecls(@This());
}
pub fn floatMantissaBits(comptime T: type) comptime_int {
assert(@typeInfo(T) == .Float);
return switch (T.bit_count) {
16 => 10,
32 => 23,
64 => 52,
80 => 64,
128 => 112,
else => @compileError("unknown floating point type " ++ @typeName(T)),
};
}
pub fn floatExponentBits(comptime T: type) comptime_int {
assert(@typeInfo(T) == .Float);
return switch (T.bit_count) {
16 => 5,
32 => 8,
64 => 11,
80 => 15,
128 => 15,
else => @compileError("unknown floating point type " ++ @typeName(T)),
};
}
/// Given two types, returns the smallest one which is capable of holding the
/// full range of the minimum value.
pub fn Min(comptime A: type, comptime B: type) type {
switch (@typeInfo(A)) {
.Int => |a_info| switch (@typeInfo(B)) {
.Int => |b_info| if (!a_info.is_signed and !b_info.is_signed) {
if (a_info.bits < b_info.bits) {
return A;
} else {
return B;
}
},
else => {},
},
else => {},
}
return @TypeOf(@as(A, 0) + @as(B, 0));
}
/// Returns the smaller number. When one of the parameter's type's full range fits in the other,
/// the return type is the smaller type.
pub fn min(x: anytype, y: anytype) Min(@TypeOf(x), @TypeOf(y)) {
const Result = Min(@TypeOf(x), @TypeOf(y));
if (x < y) {
// TODO Zig should allow this as an implicit cast because x is immutable and in this
// scope it is known to fit in the return type.
switch (@typeInfo(Result)) {
.Int => return @intCast(Result, x),
else => return x,
}
} else {
// TODO Zig should allow this as an implicit cast because y is immutable and in this
// scope it is known to fit in the return type.
switch (@typeInfo(Result)) {
.Int => return @intCast(Result, y),
else => return y,
}
}
}
test "math.min" {
testing.expect(min(@as(i32, -1), @as(i32, 2)) == -1);
{
var a: u16 = 999;
var b: u32 = 10;
var result = min(a, b);
testing.expect(@TypeOf(result) == u16);
testing.expect(result == 10);
}
{
var a: f64 = 10.34;
var b: f32 = 999.12;
var result = min(a, b);
testing.expect(@TypeOf(result) == f64);
testing.expect(result == 10.34);
}
{
var a: i8 = -127;
var b: i16 = -200;
var result = min(a, b);
testing.expect(@TypeOf(result) == i16);
testing.expect(result == -200);
}
{
const a = 10.34;
var b: f32 = 999.12;
var result = min(a, b);
testing.expect(@TypeOf(result) == f32);
testing.expect(result == 10.34);
}
}
pub fn max(x: anytype, y: anytype) @TypeOf(x, y) {
return if (x > y) x else y;
}
test "math.max" {
testing.expect(max(@as(i32, -1), @as(i32, 2)) == 2);
}
pub fn clamp(val: anytype, lower: anytype, upper: anytype) @TypeOf(val, lower, upper) {
assert(lower <= upper);
return max(lower, min(val, upper));
}
test "math.clamp" {
// Within range
testing.expect(std.math.clamp(@as(i32, -1), @as(i32, -4), @as(i32, 7)) == -1);
// Below
testing.expect(std.math.clamp(@as(i32, -5), @as(i32, -4), @as(i32, 7)) == -4);
// Above
testing.expect(std.math.clamp(@as(i32, 8), @as(i32, -4), @as(i32, 7)) == 7);
// Floating point
testing.expect(std.math.clamp(@as(f32, 1.1), @as(f32, 0.0), @as(f32, 1.0)) == 1.0);
testing.expect(std.math.clamp(@as(f32, -127.5), @as(f32, -200), @as(f32, -100)) == -127.5);
// Mix of comptime and non-comptime
var i: i32 = 1;
testing.expect(std.math.clamp(i, 0, 1) == 1);
}
pub fn mul(comptime T: type, a: T, b: T) (error{Overflow}!T) {
var answer: T = undefined;
return if (@mulWithOverflow(T, a, b, &answer)) error.Overflow else answer;
}
pub fn add(comptime T: type, a: T, b: T) (error{Overflow}!T) {
var answer: T = undefined;
return if (@addWithOverflow(T, a, b, &answer)) error.Overflow else answer;
}
pub fn sub(comptime T: type, a: T, b: T) (error{Overflow}!T) {
var answer: T = undefined;
return if (@subWithOverflow(T, a, b, &answer)) error.Overflow else answer;
}
pub fn negate(x: anytype) !@TypeOf(x) {
return sub(@TypeOf(x), 0, x);
}
pub fn shlExact(comptime T: type, a: T, shift_amt: Log2Int(T)) !T {
var answer: T = undefined;
return if (@shlWithOverflow(T, a, shift_amt, &answer)) error.Overflow else answer;
}
/// Shifts left. Overflowed bits are truncated.
/// A negative shift amount results in a right shift.
pub fn shl(comptime T: type, a: T, shift_amt: anytype) T {
const abs_shift_amt = absCast(shift_amt);
const casted_shift_amt = if (abs_shift_amt >= T.bit_count) return 0 else @intCast(Log2Int(T), abs_shift_amt);
if (@TypeOf(shift_amt) == comptime_int or @TypeOf(shift_amt).is_signed) {
if (shift_amt < 0) {
return a >> casted_shift_amt;
}
}
return a << casted_shift_amt;
}
test "math.shl" {
testing.expect(shl(u8, 0b11111111, @as(usize, 3)) == 0b11111000);
testing.expect(shl(u8, 0b11111111, @as(usize, 8)) == 0);
testing.expect(shl(u8, 0b11111111, @as(usize, 9)) == 0);
testing.expect(shl(u8, 0b11111111, @as(isize, -2)) == 0b00111111);
testing.expect(shl(u8, 0b11111111, 3) == 0b11111000);
testing.expect(shl(u8, 0b11111111, 8) == 0);
testing.expect(shl(u8, 0b11111111, 9) == 0);
testing.expect(shl(u8, 0b11111111, -2) == 0b00111111);
}
/// Shifts right. Overflowed bits are truncated.
/// A negative shift amount results in a left shift.
pub fn shr(comptime T: type, a: T, shift_amt: anytype) T {
const abs_shift_amt = absCast(shift_amt);
const casted_shift_amt = if (abs_shift_amt >= T.bit_count) return 0 else @intCast(Log2Int(T), abs_shift_amt);
if (@TypeOf(shift_amt) == comptime_int or @TypeOf(shift_amt).is_signed) {
if (shift_amt >= 0) {
return a >> casted_shift_amt;
} else {
return a << casted_shift_amt;
}
}
return a >> casted_shift_amt;
}
test "math.shr" {
testing.expect(shr(u8, 0b11111111, @as(usize, 3)) == 0b00011111);
testing.expect(shr(u8, 0b11111111, @as(usize, 8)) == 0);
testing.expect(shr(u8, 0b11111111, @as(usize, 9)) == 0);
testing.expect(shr(u8, 0b11111111, @as(isize, -2)) == 0b11111100);
testing.expect(shr(u8, 0b11111111, 3) == 0b00011111);
testing.expect(shr(u8, 0b11111111, 8) == 0);
testing.expect(shr(u8, 0b11111111, 9) == 0);
testing.expect(shr(u8, 0b11111111, -2) == 0b11111100);
}
/// Rotates right. Only unsigned values can be rotated.
/// Negative shift values results in shift modulo the bit count.
pub fn rotr(comptime T: type, x: T, r: anytype) T {
if (T.is_signed) {
@compileError("cannot rotate signed integer");
} else {
const ar = @mod(r, T.bit_count);
return shr(T, x, ar) | shl(T, x, T.bit_count - ar);
}
}
test "math.rotr" {
testing.expect(rotr(u8, 0b00000001, @as(usize, 0)) == 0b00000001);
testing.expect(rotr(u8, 0b00000001, @as(usize, 9)) == 0b10000000);
testing.expect(rotr(u8, 0b00000001, @as(usize, 8)) == 0b00000001);
testing.expect(rotr(u8, 0b00000001, @as(usize, 4)) == 0b00010000);
testing.expect(rotr(u8, 0b00000001, @as(isize, -1)) == 0b00000010);
}
/// Rotates left. Only unsigned values can be rotated.
/// Negative shift values results in shift modulo the bit count.
pub fn rotl(comptime T: type, x: T, r: anytype) T {
if (T.is_signed) {
@compileError("cannot rotate signed integer");
} else {
const ar = @mod(r, T.bit_count);
return shl(T, x, ar) | shr(T, x, T.bit_count - ar);
}
}
test "math.rotl" {
testing.expect(rotl(u8, 0b00000001, @as(usize, 0)) == 0b00000001);
testing.expect(rotl(u8, 0b00000001, @as(usize, 9)) == 0b00000010);
testing.expect(rotl(u8, 0b00000001, @as(usize, 8)) == 0b00000001);
testing.expect(rotl(u8, 0b00000001, @as(usize, 4)) == 0b00010000);
testing.expect(rotl(u8, 0b00000001, @as(isize, -1)) == 0b10000000);
}
pub fn Log2Int(comptime T: type) type {
// comptime ceil log2
comptime var count = 0;
comptime var s = T.bit_count - 1;
inline while (s != 0) : (s >>= 1) {
count += 1;
}
return std.meta.Int(false, count);
}
pub fn IntFittingRange(comptime from: comptime_int, comptime to: comptime_int) type {
assert(from <= to);
if (from == 0 and to == 0) {
return u0;
}
const is_signed = from < 0;
const largest_positive_integer = max(if (from < 0) (-from) - 1 else from, to); // two's complement
const base = log2(largest_positive_integer);
const upper = (1 << base) - 1;
var magnitude_bits = if (upper >= largest_positive_integer) base else base + 1;
if (is_signed) {
magnitude_bits += 1;
}
return std.meta.Int(is_signed, magnitude_bits);
}
test "math.IntFittingRange" {
testing.expect(IntFittingRange(0, 0) == u0);
testing.expect(IntFittingRange(0, 1) == u1);
testing.expect(IntFittingRange(0, 2) == u2);
testing.expect(IntFittingRange(0, 3) == u2);
testing.expect(IntFittingRange(0, 4) == u3);
testing.expect(IntFittingRange(0, 7) == u3);
testing.expect(IntFittingRange(0, 8) == u4);
testing.expect(IntFittingRange(0, 9) == u4);
testing.expect(IntFittingRange(0, 15) == u4);
testing.expect(IntFittingRange(0, 16) == u5);
testing.expect(IntFittingRange(0, 17) == u5);
testing.expect(IntFittingRange(0, 4095) == u12);
testing.expect(IntFittingRange(2000, 4095) == u12);
testing.expect(IntFittingRange(0, 4096) == u13);
testing.expect(IntFittingRange(2000, 4096) == u13);
testing.expect(IntFittingRange(0, 4097) == u13);
testing.expect(IntFittingRange(2000, 4097) == u13);
testing.expect(IntFittingRange(0, 123456789123456798123456789) == u87);
testing.expect(IntFittingRange(0, 123456789123456798123456789123456789123456798123456789) == u177);
testing.expect(IntFittingRange(-1, -1) == i1);
testing.expect(IntFittingRange(-1, 0) == i1);
testing.expect(IntFittingRange(-1, 1) == i2);
testing.expect(IntFittingRange(-2, -2) == i2);
testing.expect(IntFittingRange(-2, -1) == i2);
testing.expect(IntFittingRange(-2, 0) == i2);
testing.expect(IntFittingRange(-2, 1) == i2);
testing.expect(IntFittingRange(-2, 2) == i3);
testing.expect(IntFittingRange(-1, 2) == i3);
testing.expect(IntFittingRange(-1, 3) == i3);
testing.expect(IntFittingRange(-1, 4) == i4);
testing.expect(IntFittingRange(-1, 7) == i4);
testing.expect(IntFittingRange(-1, 8) == i5);
testing.expect(IntFittingRange(-1, 9) == i5);
testing.expect(IntFittingRange(-1, 15) == i5);
testing.expect(IntFittingRange(-1, 16) == i6);
testing.expect(IntFittingRange(-1, 17) == i6);
testing.expect(IntFittingRange(-1, 4095) == i13);
testing.expect(IntFittingRange(-4096, 4095) == i13);
testing.expect(IntFittingRange(-1, 4096) == i14);
testing.expect(IntFittingRange(-4097, 4095) == i14);
testing.expect(IntFittingRange(-1, 4097) == i14);
testing.expect(IntFittingRange(-1, 123456789123456798123456789) == i88);
testing.expect(IntFittingRange(-1, 123456789123456798123456789123456789123456798123456789) == i178);
}
test "math overflow functions" {
testOverflow();
comptime testOverflow();
}
fn testOverflow() void {
testing.expect((mul(i32, 3, 4) catch unreachable) == 12);
testing.expect((add(i32, 3, 4) catch unreachable) == 7);
testing.expect((sub(i32, 3, 4) catch unreachable) == -1);
testing.expect((shlExact(i32, 0b11, 4) catch unreachable) == 0b110000);
}
pub fn absInt(x: anytype) !@TypeOf(x) {
const T = @TypeOf(x);
comptime assert(@typeInfo(T) == .Int); // must pass an integer to absInt
comptime assert(T.is_signed); // must pass a signed integer to absInt
if (x == minInt(@TypeOf(x))) {
return error.Overflow;
} else {
@setRuntimeSafety(false);
return if (x < 0) -x else x;
}
}
test "math.absInt" {
testAbsInt();
comptime testAbsInt();
}
fn testAbsInt() void {
testing.expect((absInt(@as(i32, -10)) catch unreachable) == 10);
testing.expect((absInt(@as(i32, 10)) catch unreachable) == 10);
}
pub const absFloat = fabs;
test "math.absFloat" {
testAbsFloat();
comptime testAbsFloat();
}
fn testAbsFloat() void {
testing.expect(absFloat(@as(f32, -10.05)) == 10.05);
testing.expect(absFloat(@as(f32, 10.05)) == 10.05);
}
pub fn divTrunc(comptime T: type, numerator: T, denominator: T) !T {
@setRuntimeSafety(false);
if (denominator == 0) return error.DivisionByZero;
if (@typeInfo(T) == .Int and T.is_signed and numerator == minInt(T) and denominator == -1) return error.Overflow;
return @divTrunc(numerator, denominator);
}
test "math.divTrunc" {
testDivTrunc();
comptime testDivTrunc();
}
fn testDivTrunc() void {
testing.expect((divTrunc(i32, 5, 3) catch unreachable) == 1);
testing.expect((divTrunc(i32, -5, 3) catch unreachable) == -1);
testing.expectError(error.DivisionByZero, divTrunc(i8, -5, 0));
testing.expectError(error.Overflow, divTrunc(i8, -128, -1));
testing.expect((divTrunc(f32, 5.0, 3.0) catch unreachable) == 1.0);
testing.expect((divTrunc(f32, -5.0, 3.0) catch unreachable) == -1.0);
}
pub fn divFloor(comptime T: type, numerator: T, denominator: T) !T {
@setRuntimeSafety(false);
if (denominator == 0) return error.DivisionByZero;
if (@typeInfo(T) == .Int and T.is_signed and numerator == minInt(T) and denominator == -1) return error.Overflow;
return @divFloor(numerator, denominator);
}
test "math.divFloor" {
testDivFloor();
comptime testDivFloor();
}
fn testDivFloor() void {
testing.expect((divFloor(i32, 5, 3) catch unreachable) == 1);
testing.expect((divFloor(i32, -5, 3) catch unreachable) == -2);
testing.expectError(error.DivisionByZero, divFloor(i8, -5, 0));
testing.expectError(error.Overflow, divFloor(i8, -128, -1));
testing.expect((divFloor(f32, 5.0, 3.0) catch unreachable) == 1.0);
testing.expect((divFloor(f32, -5.0, 3.0) catch unreachable) == -2.0);
}
pub fn divExact(comptime T: type, numerator: T, denominator: T) !T {
@setRuntimeSafety(false);
if (denominator == 0) return error.DivisionByZero;
if (@typeInfo(T) == .Int and T.is_signed and numerator == minInt(T) and denominator == -1) return error.Overflow;
const result = @divTrunc(numerator, denominator);
if (result * denominator != numerator) return error.UnexpectedRemainder;
return result;
}
test "math.divExact" {
testDivExact();
comptime testDivExact();
}
fn testDivExact() void {
testing.expect((divExact(i32, 10, 5) catch unreachable) == 2);
testing.expect((divExact(i32, -10, 5) catch unreachable) == -2);
testing.expectError(error.DivisionByZero, divExact(i8, -5, 0));
testing.expectError(error.Overflow, divExact(i8, -128, -1));
testing.expectError(error.UnexpectedRemainder, divExact(i32, 5, 2));
testing.expect((divExact(f32, 10.0, 5.0) catch unreachable) == 2.0);
testing.expect((divExact(f32, -10.0, 5.0) catch unreachable) == -2.0);
testing.expectError(error.UnexpectedRemainder, divExact(f32, 5.0, 2.0));
}
pub fn mod(comptime T: type, numerator: T, denominator: T) !T {
@setRuntimeSafety(false);
if (denominator == 0) return error.DivisionByZero;
if (denominator < 0) return error.NegativeDenominator;
return @mod(numerator, denominator);
}
test "math.mod" {
testMod();
comptime testMod();
}
fn testMod() void {
testing.expect((mod(i32, -5, 3) catch unreachable) == 1);
testing.expect((mod(i32, 5, 3) catch unreachable) == 2);
testing.expectError(error.NegativeDenominator, mod(i32, 10, -1));
testing.expectError(error.DivisionByZero, mod(i32, 10, 0));
testing.expect((mod(f32, -5, 3) catch unreachable) == 1);
testing.expect((mod(f32, 5, 3) catch unreachable) == 2);
testing.expectError(error.NegativeDenominator, mod(f32, 10, -1));
testing.expectError(error.DivisionByZero, mod(f32, 10, 0));
}
pub fn rem(comptime T: type, numerator: T, denominator: T) !T {
@setRuntimeSafety(false);
if (denominator == 0) return error.DivisionByZero;
if (denominator < 0) return error.NegativeDenominator;
return @rem(numerator, denominator);
}
test "math.rem" {
testRem();
comptime testRem();
}
fn testRem() void {
testing.expect((rem(i32, -5, 3) catch unreachable) == -2);
testing.expect((rem(i32, 5, 3) catch unreachable) == 2);
testing.expectError(error.NegativeDenominator, rem(i32, 10, -1));
testing.expectError(error.DivisionByZero, rem(i32, 10, 0));
testing.expect((rem(f32, -5, 3) catch unreachable) == -2);
testing.expect((rem(f32, 5, 3) catch unreachable) == 2);
testing.expectError(error.NegativeDenominator, rem(f32, 10, -1));
testing.expectError(error.DivisionByZero, rem(f32, 10, 0));
}
/// Returns the absolute value of the integer parameter.
/// Result is an unsigned integer.
pub fn absCast(x: anytype) switch (@typeInfo(@TypeOf(x))) {
.ComptimeInt => comptime_int,
.Int => |intInfo| std.meta.Int(false, intInfo.bits),
else => @compileError("absCast only accepts integers"),
} {
switch (@typeInfo(@TypeOf(x))) {
.ComptimeInt => {
if (x < 0) {
return -x;
} else {
return x;
}
},
.Int => |intInfo| {
const Uint = std.meta.Int(false, intInfo.bits);
if (x < 0) {
return ~@bitCast(Uint, x +% -1);
} else {
return @intCast(Uint, x);
}
},
else => unreachable,
}
}
test "math.absCast" {
testing.expectEqual(@as(u1, 1), absCast(@as(i1, -1)));
testing.expectEqual(@as(u32, 999), absCast(@as(i32, -999)));
testing.expectEqual(@as(u32, 999), absCast(@as(i32, 999)));
testing.expectEqual(@as(u32, -minInt(i32)), absCast(@as(i32, minInt(i32))));
testing.expectEqual(999, absCast(-999));
}
/// Returns the negation of the integer parameter.
/// Result is a signed integer.
pub fn negateCast(x: anytype) !std.meta.Int(true, @TypeOf(x).bit_count) {
if (@TypeOf(x).is_signed) return negate(x);
const int = std.meta.Int(true, @TypeOf(x).bit_count);
if (x > -minInt(int)) return error.Overflow;
if (x == -minInt(int)) return minInt(int);
return -@intCast(int, x);
}
test "math.negateCast" {
testing.expect((negateCast(@as(u32, 999)) catch unreachable) == -999);
testing.expect(@TypeOf(negateCast(@as(u32, 999)) catch unreachable) == i32);
testing.expect((negateCast(@as(u32, -minInt(i32))) catch unreachable) == minInt(i32));
testing.expect(@TypeOf(negateCast(@as(u32, -minInt(i32))) catch unreachable) == i32);
testing.expectError(error.Overflow, negateCast(@as(u32, maxInt(i32) + 10)));
}
/// Cast an integer to a different integer type. If the value doesn't fit,
/// return an error.
pub fn cast(comptime T: type, x: anytype) (error{Overflow}!T) {
comptime assert(@typeInfo(T) == .Int); // must pass an integer
comptime assert(@typeInfo(@TypeOf(x)) == .Int); // must pass an integer
if (maxInt(@TypeOf(x)) > maxInt(T) and x > maxInt(T)) {
return error.Overflow;
} else if (minInt(@TypeOf(x)) < minInt(T) and x < minInt(T)) {
return error.Overflow;
} else {
return @intCast(T, x);
}
}
test "math.cast" {
testing.expectError(error.Overflow, cast(u8, @as(u32, 300)));
testing.expectError(error.Overflow, cast(i8, @as(i32, -200)));
testing.expectError(error.Overflow, cast(u8, @as(i8, -1)));
testing.expectError(error.Overflow, cast(u64, @as(i8, -1)));
testing.expect((try cast(u8, @as(u32, 255))) == @as(u8, 255));
testing.expect(@TypeOf(try cast(u8, @as(u32, 255))) == u8);
}
pub const AlignCastError = error{UnalignedMemory};
/// Align cast a pointer but return an error if it's the wrong alignment
pub fn alignCast(comptime alignment: u29, ptr: anytype) AlignCastError!@TypeOf(@alignCast(alignment, ptr)) {
const addr = @ptrToInt(ptr);
if (addr % alignment != 0) {
return error.UnalignedMemory;
}
return @alignCast(alignment, ptr);
}
pub fn isPowerOfTwo(v: anytype) bool {
assert(v != 0);
return (v & (v - 1)) == 0;
}
pub fn floorPowerOfTwo(comptime T: type, value: T) T {
var x = value;
comptime var i = 1;
inline while (T.bit_count > i) : (i *= 2) {
x |= (x >> i);
}
return x - (x >> 1);
}
test "math.floorPowerOfTwo" {
testFloorPowerOfTwo();
comptime testFloorPowerOfTwo();
}
fn testFloorPowerOfTwo() void {
testing.expect(floorPowerOfTwo(u32, 63) == 32);
testing.expect(floorPowerOfTwo(u32, 64) == 64);
testing.expect(floorPowerOfTwo(u32, 65) == 64);
testing.expect(floorPowerOfTwo(u4, 7) == 4);
testing.expect(floorPowerOfTwo(u4, 8) == 8);
testing.expect(floorPowerOfTwo(u4, 9) == 8);
}
/// Returns the next power of two (if the value is not already a power of two).
/// Only unsigned integers can be used. Zero is not an allowed input.
/// Result is a type with 1 more bit than the input type.
pub fn ceilPowerOfTwoPromote(comptime T: type, value: T) std.meta.Int(T.is_signed, T.bit_count + 1) {
comptime assert(@typeInfo(T) == .Int);
comptime assert(!T.is_signed);
assert(value != 0);
comptime const PromotedType = std.meta.Int(T.is_signed, T.bit_count + 1);
comptime const shiftType = std.math.Log2Int(PromotedType);
return @as(PromotedType, 1) << @intCast(shiftType, T.bit_count - @clz(T, value - 1));
}
/// Returns the next power of two (if the value is not already a power of two).
/// Only unsigned integers can be used. Zero is not an allowed input.
/// If the value doesn't fit, returns an error.
pub fn ceilPowerOfTwo(comptime T: type, value: T) (error{Overflow}!T) {
comptime assert(@typeInfo(T) == .Int);
comptime assert(!T.is_signed);
comptime const PromotedType = std.meta.Int(T.is_signed, T.bit_count + 1);
comptime const overflowBit = @as(PromotedType, 1) << T.bit_count;
var x = ceilPowerOfTwoPromote(T, value);
if (overflowBit & x != 0) {
return error.Overflow;
}
return @intCast(T, x);
}
test "math.ceilPowerOfTwoPromote" {
testCeilPowerOfTwoPromote();
comptime testCeilPowerOfTwoPromote();
}
fn testCeilPowerOfTwoPromote() void {
testing.expectEqual(@as(u33, 1), ceilPowerOfTwoPromote(u32, 1));
testing.expectEqual(@as(u33, 2), ceilPowerOfTwoPromote(u32, 2));
testing.expectEqual(@as(u33, 64), ceilPowerOfTwoPromote(u32, 63));
testing.expectEqual(@as(u33, 64), ceilPowerOfTwoPromote(u32, 64));
testing.expectEqual(@as(u33, 128), ceilPowerOfTwoPromote(u32, 65));
testing.expectEqual(@as(u6, 8), ceilPowerOfTwoPromote(u5, 7));
testing.expectEqual(@as(u6, 8), ceilPowerOfTwoPromote(u5, 8));
testing.expectEqual(@as(u6, 16), ceilPowerOfTwoPromote(u5, 9));
testing.expectEqual(@as(u5, 16), ceilPowerOfTwoPromote(u4, 9));
}
test "math.ceilPowerOfTwo" {
try testCeilPowerOfTwo();
comptime try testCeilPowerOfTwo();
}
fn testCeilPowerOfTwo() !void {
testing.expectEqual(@as(u32, 1), try ceilPowerOfTwo(u32, 1));
testing.expectEqual(@as(u32, 2), try ceilPowerOfTwo(u32, 2));
testing.expectEqual(@as(u32, 64), try ceilPowerOfTwo(u32, 63));
testing.expectEqual(@as(u32, 64), try ceilPowerOfTwo(u32, 64));
testing.expectEqual(@as(u32, 128), try ceilPowerOfTwo(u32, 65));
testing.expectEqual(@as(u5, 8), try ceilPowerOfTwo(u5, 7));
testing.expectEqual(@as(u5, 8), try ceilPowerOfTwo(u5, 8));
testing.expectEqual(@as(u5, 16), try ceilPowerOfTwo(u5, 9));
testing.expectError(error.Overflow, ceilPowerOfTwo(u4, 9));
}
pub fn log2_int(comptime T: type, x: T) Log2Int(T) {
assert(x != 0);
return @intCast(Log2Int(T), T.bit_count - 1 - @clz(T, x));
}
pub fn log2_int_ceil(comptime T: type, x: T) Log2Int(T) {
assert(x != 0);
const log2_val = log2_int(T, x);
if (@as(T, 1) << log2_val == x)
return log2_val;
return log2_val + 1;
}
test "std.math.log2_int_ceil" {
testing.expect(log2_int_ceil(u32, 1) == 0);
testing.expect(log2_int_ceil(u32, 2) == 1);
testing.expect(log2_int_ceil(u32, 3) == 2);
testing.expect(log2_int_ceil(u32, 4) == 2);
testing.expect(log2_int_ceil(u32, 5) == 3);
testing.expect(log2_int_ceil(u32, 6) == 3);
testing.expect(log2_int_ceil(u32, 7) == 3);
testing.expect(log2_int_ceil(u32, 8) == 3);
testing.expect(log2_int_ceil(u32, 9) == 4);
testing.expect(log2_int_ceil(u32, 10) == 4);
}
pub fn lossyCast(comptime T: type, value: anytype) T {
switch (@typeInfo(@TypeOf(value))) {
.Int => return @intToFloat(T, value),
.Float => return @floatCast(T, value),
.ComptimeInt => return @as(T, value),
.ComptimeFloat => return @as(T, value),
else => @compileError("bad type"),
}
}
test "math.f64_min" {
const f64_min_u64 = 0x0010000000000000;
const fmin: f64 = f64_min;
testing.expect(@bitCast(u64, fmin) == f64_min_u64);
}
pub fn maxInt(comptime T: type) comptime_int {
const info = @typeInfo(T);
const bit_count = info.Int.bits;
if (bit_count == 0) return 0;
return (1 << (bit_count - @boolToInt(info.Int.is_signed))) - 1;
}
pub fn minInt(comptime T: type) comptime_int {
const info = @typeInfo(T);
const bit_count = info.Int.bits;
if (!info.Int.is_signed) return 0;
if (bit_count == 0) return 0;
return -(1 << (bit_count - 1));
}
test "minInt and maxInt" {
testing.expect(maxInt(u0) == 0);
testing.expect(maxInt(u1) == 1);
testing.expect(maxInt(u8) == 255);
testing.expect(maxInt(u16) == 65535);
testing.expect(maxInt(u32) == 4294967295);
testing.expect(maxInt(u64) == 18446744073709551615);
testing.expect(maxInt(u128) == 340282366920938463463374607431768211455);
testing.expect(maxInt(i0) == 0);
testing.expect(maxInt(i1) == 0);
testing.expect(maxInt(i8) == 127);
testing.expect(maxInt(i16) == 32767);
testing.expect(maxInt(i32) == 2147483647);
testing.expect(maxInt(i63) == 4611686018427387903);
testing.expect(maxInt(i64) == 9223372036854775807);
testing.expect(maxInt(i128) == 170141183460469231731687303715884105727);
testing.expect(minInt(u0) == 0);
testing.expect(minInt(u1) == 0);
testing.expect(minInt(u8) == 0);
testing.expect(minInt(u16) == 0);
testing.expect(minInt(u32) == 0);
testing.expect(minInt(u63) == 0);
testing.expect(minInt(u64) == 0);
testing.expect(minInt(u128) == 0);
testing.expect(minInt(i0) == 0);
testing.expect(minInt(i1) == -1);
testing.expect(minInt(i8) == -128);
testing.expect(minInt(i16) == -32768);
testing.expect(minInt(i32) == -2147483648);
testing.expect(minInt(i63) == -4611686018427387904);
testing.expect(minInt(i64) == -9223372036854775808);
testing.expect(minInt(i128) == -170141183460469231731687303715884105728);
}
test "max value type" {
const x: u32 = maxInt(i32);
testing.expect(x == 2147483647);
}
pub fn mulWide(comptime T: type, a: T, b: T) std.meta.Int(T.is_signed, T.bit_count * 2) {
const ResultInt = std.meta.Int(T.is_signed, T.bit_count * 2);
return @as(ResultInt, a) * @as(ResultInt, b);
}
test "math.mulWide" {
testing.expect(mulWide(u8, 5, 5) == 25);
testing.expect(mulWide(i8, 5, -5) == -25);
testing.expect(mulWide(u8, 100, 100) == 10000);
}
/// See also `CompareOperator`.
pub const Order = enum {
/// Less than (`<`)
lt,
/// Equal (`==`)
eq,
/// Greater than (`>`)
gt,
pub fn invert(self: Order) Order {
return switch (self) {
.lt => .gt,
.eq => .eq,
.gt => .gt,
};
}
pub fn compare(self: Order, op: CompareOperator) bool {
return switch (self) {
.lt => switch (op) {
.lt => true,
.lte => true,
.eq => false,
.gte => false,
.gt => false,
.neq => true,
},
.eq => switch (op) {
.lt => false,
.lte => true,
.eq => true,
.gte => true,
.gt => false,
.neq => false,
},
.gt => switch (op) {
.lt => false,
.lte => false,
.eq => false,
.gte => true,
.gt => true,
.neq => true,
},
};
}
};
/// Given two numbers, this function returns the order they are with respect to each other.
pub fn order(a: anytype, b: anytype) Order {
if (a == b) {
return .eq;
} else if (a < b) {
return .lt;
} else if (a > b) {
return .gt;
} else {
unreachable;
}
}
/// See also `Order`.
pub const CompareOperator = enum {
/// Less than (`<`)
lt,
/// Less than or equal (`<=`)
lte,
/// Equal (`==`)
eq,
/// Greater than or equal (`>=`)
gte,
/// Greater than (`>`)
gt,
/// Not equal (`!=`)
neq,
};
/// This function does the same thing as comparison operators, however the
/// operator is a runtime-known enum value. Works on any operands that
/// support comparison operators.
pub fn compare(a: anytype, op: CompareOperator, b: anytype) bool {
return switch (op) {
.lt => a < b,
.lte => a <= b,
.eq => a == b,
.neq => a != b,
.gt => a > b,
.gte => a >= b,
};
}
test "compare between signed and unsigned" {
testing.expect(compare(@as(i8, -1), .lt, @as(u8, 255)));
testing.expect(compare(@as(i8, 2), .gt, @as(u8, 1)));
testing.expect(!compare(@as(i8, -1), .gte, @as(u8, 255)));
testing.expect(compare(@as(u8, 255), .gt, @as(i8, -1)));
testing.expect(!compare(@as(u8, 255), .lte, @as(i8, -1)));
testing.expect(compare(@as(i8, -1), .lt, @as(u9, 255)));
testing.expect(!compare(@as(i8, -1), .gte, @as(u9, 255)));
testing.expect(compare(@as(u9, 255), .gt, @as(i8, -1)));
testing.expect(!compare(@as(u9, 255), .lte, @as(i8, -1)));
testing.expect(compare(@as(i9, -1), .lt, @as(u8, 255)));
testing.expect(!compare(@as(i9, -1), .gte, @as(u8, 255)));
testing.expect(compare(@as(u8, 255), .gt, @as(i9, -1)));
testing.expect(!compare(@as(u8, 255), .lte, @as(i9, -1)));
testing.expect(compare(@as(u8, 1), .lt, @as(u8, 2)));
testing.expect(@bitCast(u8, @as(i8, -1)) == @as(u8, 255));
testing.expect(!compare(@as(u8, 255), .eq, @as(i8, -1)));
testing.expect(compare(@as(u8, 1), .eq, @as(u8, 1)));
}