1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160 | // Copyright (C) 2023 Internet Systems Consortium, Inc. ("ISC")
//
// This Source Code Form is subject to the terms of the Mozilla Public
// License, v. 2.0. If a copy of the MPL was not distributed with this
// file, You can obtain one at http://mozilla.org/MPL/2.0/.
#include <config.h>
#include <testutils/gtest_utils.h>
#include <util/bigints.h>
#include <gtest/gtest.h><--- Include file: not found. Please note: Cppcheck does not need standard library headers to get proper results.
#include <limits><--- Include file: not found. Please note: Cppcheck does not need standard library headers to get proper results.
using namespace isc::util;
using namespace std;
namespace {
// C++ doesn't allow very big integer literals, so that's why some tests on big
// numbers might appear like they're unnecessarily circumventing a more obvious
// test choice.
// Checks that int128_t behaves like a signed integer should.
TEST(BigintTest, int128) {
// Check addition with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(24, int128_t(16) + int128_t(8));
});
// Check subtraction with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(48, int128_t(64) - int128_t(16));
});
// Check multiplication with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(8, int128_t(2) * int128_t(4));
});
// Check division with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(16, int128_t(64) / int128_t(4));
});
// Check rounded division with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(16, int128_t(65) / int128_t(4));
});
// Check that INT128_MIN < INT64_MIN.
EXPECT_NO_THROW({
EXPECT_LT(numeric_limits<int128_t>::min(), numeric_limits<int64_t>::min());
});
// Check that UINT64_MAX < INT128_MAX.
EXPECT_NO_THROW({
EXPECT_LT(numeric_limits<uint64_t>::max(), numeric_limits<int128_t>::max());
});
// Check that int128_t is default-initialized to zero. Not a strict
// requirement by Kea, but something that the current implementation ensures.
int128_t i128;
EXPECT_NO_THROW({
EXPECT_EQ(0, i128);
});
// Shifting to the right beyond zero does not result in an underflow error.
EXPECT_NO_THROW({
EXPECT_EQ(0, uint128_t(1) >> 2);
});
// Check that dividing by zero throws.
EXPECT_THROW_MSG(int128_t(1) / 0, overflow_error, "Division by zero.");
// Check that underflowing results in an error.
i128 = numeric_limits<int128_t>::min();
EXPECT_THROW_MSG(i128 - 1, overflow_error, "overflow in addition");
EXPECT_THROW_MSG(i128 + i128, overflow_error, "overflow in addition");
// Check that overflowing results in an error.
i128 = numeric_limits<int128_t>::max();
EXPECT_THROW_MSG(i128 + 1, overflow_error, "overflow in addition");
EXPECT_THROW_MSG(i128 + i128, overflow_error, "overflow in addition");
EXPECT_THROW_MSG(2 * i128, overflow_error, "overflow in multiplication");
EXPECT_THROW_MSG(i128 << 1, overflow_error, "Shift out of range");
}
// Checks that uint128_t behaves like an unsigned integer should.
TEST(BigintTest, uint128) {<--- syntax error
// Check addition with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(24, uint128_t(16) + uint128_t(8));
});
// Check subtraction with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(48, uint128_t(64) - uint128_t(16));
});
// Check multiplication with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(8, uint128_t(2) * uint128_t(4));
});
// Check division with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(16, uint128_t(64) / uint128_t(4));
});
// Check rounded division with small numbers.
EXPECT_NO_THROW({
EXPECT_EQ(16, uint128_t(65) / uint128_t(4));
});
// Check that UINT128_MIN is 0.
EXPECT_NO_THROW({
EXPECT_EQ(0, numeric_limits<uint128_t>::min());
});
// Check that UINT64_MAX < UINT128_MAX.
EXPECT_NO_THROW({
EXPECT_LT(numeric_limits<uint64_t>::max(), numeric_limits<int128_t>::max());
});
// Check that INT128_MAX < UINT128_MAX.
EXPECT_NO_THROW({
EXPECT_LT(numeric_limits<uint64_t>::max(), numeric_limits<int128_t>::max());
});
// Check that uint128_t is default-initialized to zero. Not a strict
// requirement by Kea, but something that the current implementation ensures.
uint128_t u128;
EXPECT_NO_THROW({
EXPECT_EQ(0, u128);
});
// Shifting to the right beyond zero does not result in an underflow error.
EXPECT_NO_THROW({
EXPECT_EQ(0, uint128_t(1) >> 2);
});
// Check that dividing by zero throws.
EXPECT_THROW_MSG(uint128_t(1) / 0, overflow_error, "Division by zero.");
// Check that underflowing results in an error.
u128 = numeric_limits<uint128_t>::min();
EXPECT_THROW_MSG(u128 - 1, range_error, "Subtraction resulted in a negative value, but the type is unsigned");
// Check that overflowing results in an error.
u128 = numeric_limits<uint128_t>::max();
EXPECT_THROW_MSG(u128 + 1, overflow_error, "overflow in addition");
EXPECT_THROW_MSG(u128 + u128, overflow_error, "overflow in addition");
EXPECT_THROW_MSG(2 * u128, overflow_error, "overflow in multiplication");
EXPECT_THROW_MSG(u128 << 1, overflow_error, "Shift out of range");
}
} // namespace
|