2021-05-05 20:56:44 +00:00
|
|
|
#include <gtest/gtest.h>
|
2021-05-06 22:37:23 +00:00
|
|
|
|
|
|
|
#include <cstdint>
|
|
|
|
#include <limits>
|
|
|
|
#include <type_traits>
|
|
|
|
#include <initializer_list>
|
|
|
|
|
|
|
|
#include <Core/Types.h>
|
2021-05-05 20:56:44 +00:00
|
|
|
#include <IO/WriteHelpers.h>
|
2021-05-07 21:47:23 +00:00
|
|
|
#include <IO/ReadHelpers.h>
|
2021-10-02 07:13:14 +00:00
|
|
|
#include <base/demangle.h>
|
2021-05-06 22:37:23 +00:00
|
|
|
|
|
|
|
|
|
|
|
static_assert(is_signed_v<Int128>);
|
|
|
|
static_assert(!is_unsigned_v<Int128>);
|
2021-09-11 20:14:34 +00:00
|
|
|
static_assert(is_integer<Int128>);
|
2021-05-06 22:37:23 +00:00
|
|
|
static_assert(sizeof(Int128) == 16);
|
|
|
|
|
|
|
|
static_assert(is_signed_v<Int256>);
|
|
|
|
static_assert(!is_unsigned_v<Int256>);
|
2021-09-11 20:14:34 +00:00
|
|
|
static_assert(is_integer<Int256>);
|
2021-05-06 22:37:23 +00:00
|
|
|
static_assert(sizeof(Int256) == 32);
|
|
|
|
|
|
|
|
static_assert(!is_signed_v<UInt128>);
|
|
|
|
static_assert(is_unsigned_v<UInt128>);
|
2021-09-11 20:14:34 +00:00
|
|
|
static_assert(is_integer<UInt128>);
|
2021-05-06 22:37:23 +00:00
|
|
|
static_assert(sizeof(UInt128) == 16);
|
|
|
|
|
|
|
|
static_assert(!is_signed_v<UInt256>);
|
|
|
|
static_assert(is_unsigned_v<UInt256>);
|
2021-09-11 20:14:34 +00:00
|
|
|
static_assert(is_integer<UInt256>);
|
2021-05-06 22:37:23 +00:00
|
|
|
static_assert(sizeof(UInt256) == 32);
|
|
|
|
|
2021-05-05 20:56:44 +00:00
|
|
|
|
|
|
|
using namespace DB;
|
|
|
|
|
|
|
|
|
2021-05-05 21:20:55 +00:00
|
|
|
GTEST_TEST(WideInteger, Conversions)
|
2021-05-05 20:56:44 +00:00
|
|
|
{
|
|
|
|
ASSERT_EQ(toString(UInt128(12345678901234567890ULL)), "12345678901234567890");
|
|
|
|
ASSERT_EQ(toString(UInt256(12345678901234567890ULL)), "12345678901234567890");
|
|
|
|
|
2021-05-06 22:37:23 +00:00
|
|
|
Int128 minus_one = -1;
|
|
|
|
ASSERT_EQ(minus_one.items[0], -1);
|
|
|
|
ASSERT_EQ(minus_one.items[1], -1);
|
|
|
|
|
|
|
|
ASSERT_EQ(0, memcmp(&minus_one, "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(minus_one)));
|
|
|
|
|
|
|
|
ASSERT_EQ(minus_one, -1);
|
|
|
|
ASSERT_EQ(minus_one, -1LL);
|
|
|
|
ASSERT_EQ(minus_one, Int8(-1));
|
|
|
|
ASSERT_EQ(minus_one, Int16(-1));
|
|
|
|
ASSERT_EQ(minus_one, Int32(-1));
|
|
|
|
ASSERT_EQ(minus_one, Int64(-1));
|
|
|
|
|
|
|
|
ASSERT_LT(minus_one, 0);
|
|
|
|
|
|
|
|
Int128 zero = 0;
|
|
|
|
zero += -1;
|
|
|
|
ASSERT_EQ(zero, -1);
|
|
|
|
ASSERT_EQ(zero, minus_one);
|
|
|
|
|
|
|
|
zero += minus_one;
|
|
|
|
ASSERT_EQ(0, memcmp(&zero, "\xFE\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(zero)));
|
|
|
|
|
|
|
|
zero += 2;
|
|
|
|
ASSERT_EQ(zero, 0);
|
|
|
|
|
|
|
|
ASSERT_EQ(toString(Int128(-1)), "-1");
|
|
|
|
ASSERT_EQ(toString(Int256(-1)), "-1");
|
|
|
|
|
|
|
|
ASSERT_EQ(toString(Int128(-1LL)), "-1");
|
|
|
|
ASSERT_EQ(toString(Int256(-1LL)), "-1");
|
|
|
|
|
2021-05-05 20:56:44 +00:00
|
|
|
ASSERT_EQ(toString(Int128(-1234567890123456789LL)), "-1234567890123456789");
|
|
|
|
ASSERT_EQ(toString(Int256(-1234567890123456789LL)), "-1234567890123456789");
|
|
|
|
|
|
|
|
ASSERT_EQ(UInt64(UInt128(12345678901234567890ULL)), 12345678901234567890ULL);
|
|
|
|
ASSERT_EQ(UInt64(UInt256(12345678901234567890ULL)), 12345678901234567890ULL);
|
|
|
|
|
|
|
|
ASSERT_EQ(__uint128_t(UInt128(12345678901234567890ULL)), 12345678901234567890ULL);
|
|
|
|
ASSERT_EQ(__uint128_t(UInt256(12345678901234567890ULL)), 12345678901234567890ULL);
|
|
|
|
|
2021-05-05 21:20:55 +00:00
|
|
|
ASSERT_EQ(__int128_t(Int128(-1234567890123456789LL)), -1234567890123456789LL);
|
|
|
|
ASSERT_EQ(__int128_t(Int256(-1234567890123456789LL)), -1234567890123456789LL);
|
|
|
|
|
|
|
|
ASSERT_EQ(toString(Int128(-1)), "-1");
|
2021-05-05 20:56:44 +00:00
|
|
|
ASSERT_EQ(toString(Int256(-1)), "-1");
|
2021-05-05 21:20:55 +00:00
|
|
|
|
|
|
|
ASSERT_EQ(toString(UInt128(123.456)), "123");
|
|
|
|
ASSERT_EQ(toString(UInt256(123.456)), "123");
|
|
|
|
ASSERT_EQ(toString(Int128(-123.456)), "-123");
|
|
|
|
ASSERT_EQ(toString(Int256(-123.456)), "-123");
|
|
|
|
|
|
|
|
ASSERT_EQ(toString(UInt128(123.456f)), "123");
|
|
|
|
ASSERT_EQ(toString(UInt256(123.456f)), "123");
|
|
|
|
ASSERT_EQ(toString(Int128(-123.456f)), "-123");
|
|
|
|
ASSERT_EQ(toString(Int256(-123.456f)), "-123");
|
|
|
|
|
|
|
|
ASSERT_EQ(toString(UInt128(1) * 1000000000 * 1000000000 * 1000000000 * 1000000000), "1000000000000000000000000000000000000");
|
|
|
|
ASSERT_EQ(Float64(UInt128(1) * 1000000000 * 1000000000 * 1000000000 * 1000000000), 1e36);
|
|
|
|
|
|
|
|
ASSERT_EQ(toString(UInt256(1) * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000),
|
|
|
|
"1000000000000000000000000000000000000000000000000000000000000000000000000");
|
|
|
|
ASSERT_EQ(Float64(UInt256(1) * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000), 1e72);
|
2021-05-07 21:47:23 +00:00
|
|
|
|
|
|
|
EXPECT_EQ(toString(parse<Int128>("148873535527910577765226390751398592640")), "148873535527910577765226390751398592640");
|
|
|
|
EXPECT_EQ(toString(parse<UInt128>("148873535527910577765226390751398592640")), "148873535527910577765226390751398592640");
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
template <typename T>
|
|
|
|
static T divide(T & numerator, T && denominator)
|
|
|
|
{
|
|
|
|
if (!denominator)
|
|
|
|
throwError("Division by zero");
|
|
|
|
|
|
|
|
T & n = numerator;
|
|
|
|
T & d = denominator;
|
|
|
|
T x = 1;
|
|
|
|
T quotient = 0;
|
|
|
|
|
|
|
|
/// Multiply d to the power of two until it will be greater than n.
|
|
|
|
/// The factor will be collected in x.
|
|
|
|
while (d <= n && ((d >> (sizeof(T) * 8 - 1)) & 1) == 0)
|
|
|
|
{
|
|
|
|
x <<= 1;
|
|
|
|
d <<= 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
std::cerr << toString(x) << ", " << toString(d) << "\n";
|
|
|
|
|
|
|
|
while (x)
|
|
|
|
{
|
|
|
|
if (d <= n)
|
|
|
|
{
|
|
|
|
n -= d;
|
|
|
|
quotient |= x;
|
|
|
|
}
|
|
|
|
|
|
|
|
x >>= 1;
|
|
|
|
d >>= 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return quotient;
|
2021-05-05 20:56:44 +00:00
|
|
|
}
|
|
|
|
|
2021-05-05 21:20:55 +00:00
|
|
|
|
|
|
|
GTEST_TEST(WideInteger, Arithmetic)
|
|
|
|
{
|
2021-05-06 22:37:23 +00:00
|
|
|
Int128 minus_one = -1;
|
|
|
|
Int128 zero = 0;
|
|
|
|
|
|
|
|
zero += -1;
|
|
|
|
ASSERT_EQ(zero, -1);
|
|
|
|
ASSERT_EQ(zero, minus_one);
|
|
|
|
|
|
|
|
zero += minus_one;
|
|
|
|
ASSERT_EQ(0, memcmp(&zero, "\xFE\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(zero)));
|
|
|
|
|
|
|
|
zero += 2;
|
|
|
|
ASSERT_EQ(zero, 0);
|
|
|
|
|
2021-05-05 21:20:55 +00:00
|
|
|
ASSERT_EQ(UInt256(12345678901234567890ULL) * 12345678901234567890ULL / 12345678901234567890ULL, 12345678901234567890ULL);
|
|
|
|
ASSERT_EQ(UInt256(12345678901234567890ULL) * UInt256(12345678901234567890ULL) / 12345678901234567890ULL, 12345678901234567890ULL);
|
|
|
|
ASSERT_EQ(UInt256(12345678901234567890ULL) * 12345678901234567890ULL / UInt256(12345678901234567890ULL), 12345678901234567890ULL);
|
|
|
|
ASSERT_EQ(UInt256(12345678901234567890ULL) * 12345678901234567890ULL / 12345678901234567890ULL, UInt256(12345678901234567890ULL));
|
|
|
|
ASSERT_EQ(UInt128(12345678901234567890ULL) * 12345678901234567890ULL / UInt128(12345678901234567890ULL), 12345678901234567890ULL);
|
|
|
|
ASSERT_EQ(UInt256(12345678901234567890ULL) * UInt128(12345678901234567890ULL) / 12345678901234567890ULL, 12345678901234567890ULL);
|
2021-05-06 22:37:23 +00:00
|
|
|
|
|
|
|
ASSERT_EQ(Int128(0) + Int32(-1), Int128(-1));
|
2021-05-07 21:47:23 +00:00
|
|
|
|
|
|
|
Int128 x(parse<Int128>("148873535527910577765226390751398592640"));
|
|
|
|
Int128 dividend = x / 10;
|
|
|
|
ASSERT_EQ(toString(dividend), "14887353552791057776522639075139859264");
|
2021-05-06 22:37:23 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
GTEST_TEST(WideInteger, DecimalArithmetic)
|
|
|
|
{
|
|
|
|
Decimal128 zero{};
|
|
|
|
Decimal32 addend = -1000;
|
|
|
|
|
|
|
|
zero += Decimal128(addend);
|
|
|
|
ASSERT_EQ(zero.value, -1000);
|
|
|
|
|
|
|
|
zero += addend;
|
|
|
|
ASSERT_EQ(zero.value, -2000);
|
2021-05-05 21:20:55 +00:00
|
|
|
}
|
2021-05-07 17:12:14 +00:00
|
|
|
|
|
|
|
|
|
|
|
GTEST_TEST(WideInteger, FromDouble)
|
|
|
|
{
|
|
|
|
/// Check that we are being able to convert double to big integer without the help of floating point instructions.
|
|
|
|
/// (a prototype of a function that we may need)
|
|
|
|
|
|
|
|
double f = -123.456;
|
|
|
|
UInt64 u;
|
|
|
|
memcpy(&u, &f, sizeof(f));
|
|
|
|
|
|
|
|
bool is_negative = u >> 63;
|
|
|
|
uint16_t exponent = (u >> 52) & (((1ull << 12) - 1) >> 1);
|
|
|
|
int16_t normalized_exponent = exponent - 1023;
|
|
|
|
UInt64 mantissa = u & ((1ull << 52) - 1);
|
|
|
|
|
|
|
|
// std::cerr << is_negative << ", " << normalized_exponent << ", " << mantissa << "\n";
|
|
|
|
|
|
|
|
/// x = sign * (2 ^ normalized_exponent + mantissa * 2 ^ (normalized_exponent - mantissa_bits))
|
|
|
|
|
|
|
|
Int128 res = 0;
|
|
|
|
|
|
|
|
if (normalized_exponent >= 128)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
res = mantissa;
|
|
|
|
if (normalized_exponent > 52)
|
|
|
|
res <<= (normalized_exponent - 52);
|
|
|
|
else
|
|
|
|
res >>= (52 - normalized_exponent);
|
|
|
|
|
|
|
|
if (normalized_exponent > 0)
|
|
|
|
res += Int128(1) << normalized_exponent;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (is_negative)
|
|
|
|
res = -res;
|
|
|
|
|
|
|
|
ASSERT_EQ(toString(res), "-123");
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
GTEST_TEST(WideInteger, Shift)
|
|
|
|
{
|
|
|
|
Int128 x = 1;
|
|
|
|
|
|
|
|
auto y = x << 64;
|
|
|
|
ASSERT_EQ(0, memcmp(&y, "\x00\x00\x00\x00\x00\x00\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00", sizeof(Int128)));
|
|
|
|
|
|
|
|
auto z = y << 11;
|
|
|
|
ASSERT_EQ(toString(z), "37778931862957161709568");
|
|
|
|
|
|
|
|
auto a = x << 11;
|
|
|
|
ASSERT_EQ(a, 2048);
|
|
|
|
|
|
|
|
z >>= 64;
|
|
|
|
ASSERT_EQ(z, a);
|
|
|
|
|
|
|
|
x = -1;
|
|
|
|
y = x << 16;
|
|
|
|
|
|
|
|
ASSERT_EQ(0, memcmp(&y, "\x00\x00\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(Int128)));
|
|
|
|
|
|
|
|
y >>= 16;
|
|
|
|
ASSERT_EQ(0, memcmp(&y, "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(Int128)));
|
|
|
|
|
|
|
|
y <<= 64;
|
|
|
|
ASSERT_EQ(0, memcmp(&y, "\x00\x00\x00\x00\x00\x00\x00\x00\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(Int128)));
|
|
|
|
|
|
|
|
y >>= 32;
|
|
|
|
ASSERT_EQ(0, memcmp(&y, "\x00\x00\x00\x00\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(Int128)));
|
|
|
|
|
|
|
|
y <<= 64;
|
|
|
|
ASSERT_EQ(0, memcmp(&y, "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\xFF\xFF\xFF\xFF", sizeof(Int128)));
|
|
|
|
}
|
2021-05-07 21:47:23 +00:00
|
|
|
|
|
|
|
|
|
|
|
GTEST_TEST(WideInteger, DecimalFormatting)
|
|
|
|
{
|
|
|
|
Decimal128 x(parse<Int128>("148873535527910577765226390751398592640"));
|
|
|
|
|
|
|
|
EXPECT_EQ(toString(x.value), "148873535527910577765226390751398592640");
|
|
|
|
EXPECT_EQ(toString(x.value / 10), "14887353552791057776522639075139859264");
|
|
|
|
EXPECT_EQ(toString(x.value % 10), "0");
|
|
|
|
|
|
|
|
Int128 fractional = DecimalUtils::getFractionalPart(x, 2);
|
|
|
|
|
|
|
|
EXPECT_EQ(fractional, 40);
|
|
|
|
}
|