line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
// Copyright 2019 Ulf Adams |
2
|
|
|
|
|
|
|
// |
3
|
|
|
|
|
|
|
// The contents of this file may be used under the terms of the Apache License, |
4
|
|
|
|
|
|
|
// Version 2.0. |
5
|
|
|
|
|
|
|
// |
6
|
|
|
|
|
|
|
// (See accompanying file LICENSE-Apache or copy at |
7
|
|
|
|
|
|
|
// http://www.apache.org/licenses/LICENSE-2.0) |
8
|
|
|
|
|
|
|
// |
9
|
|
|
|
|
|
|
// Alternatively, the contents of this file may be used under the terms of |
10
|
|
|
|
|
|
|
// the Boost Software License, Version 1.0. |
11
|
|
|
|
|
|
|
// (See accompanying file LICENSE-Boost or copy at |
12
|
|
|
|
|
|
|
// https://www.boost.org/LICENSE_1_0.txt) |
13
|
|
|
|
|
|
|
// |
14
|
|
|
|
|
|
|
// Unless required by applicable law or agreed to in writing, this software |
15
|
|
|
|
|
|
|
// is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY |
16
|
|
|
|
|
|
|
// KIND, either express or implied. |
17
|
|
|
|
|
|
|
|
18
|
|
|
|
|
|
|
/* The location of the headers, relative to this file, has been changed by |
19
|
|
|
|
|
|
|
* Sisyphus */ |
20
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
#include "ryu_headers/ryu_parse.h" |
22
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
#include |
24
|
|
|
|
|
|
|
#include |
25
|
|
|
|
|
|
|
#include |
26
|
|
|
|
|
|
|
#include |
27
|
|
|
|
|
|
|
#include |
28
|
|
|
|
|
|
|
|
29
|
|
|
|
|
|
|
#ifdef RYU_DEBUG |
30
|
|
|
|
|
|
|
#include |
31
|
|
|
|
|
|
|
#include |
32
|
|
|
|
|
|
|
#endif |
33
|
|
|
|
|
|
|
|
34
|
|
|
|
|
|
|
#include "ryu_headers/common.h" |
35
|
|
|
|
|
|
|
#include "ryu_headers/d2s_intrinsics.h" |
36
|
|
|
|
|
|
|
|
37
|
|
|
|
|
|
|
#if defined(RYU_OPTIMIZE_SIZE) |
38
|
|
|
|
|
|
|
#include "ryu_headers/d2s_small_table.h" |
39
|
|
|
|
|
|
|
#else |
40
|
|
|
|
|
|
|
#include "ryu_headers/d2s_full_table.h" |
41
|
|
|
|
|
|
|
#endif |
42
|
|
|
|
|
|
|
|
43
|
|
|
|
|
|
|
#define DOUBLE_MANTISSA_BITS 52 |
44
|
|
|
|
|
|
|
#define DOUBLE_EXPONENT_BITS 11 |
45
|
|
|
|
|
|
|
#define DOUBLE_EXPONENT_BIAS 1023 |
46
|
|
|
|
|
|
|
|
47
|
|
|
|
|
|
|
#if defined(_MSC_VER) |
48
|
|
|
|
|
|
|
#include |
49
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
static inline uint32_t floor_log2(const uint64_t value) { |
51
|
|
|
|
|
|
|
#if !defined(_WIN64) /* _BitScanReverse64 unavailable */ |
52
|
|
|
|
|
|
|
long index = 0; |
53
|
|
|
|
|
|
|
uint64_t v = value; |
54
|
|
|
|
|
|
|
if(!v) return 64; |
55
|
|
|
|
|
|
|
while(v >>= 1) index += 1; |
56
|
|
|
|
|
|
|
return index; |
57
|
|
|
|
|
|
|
#else |
58
|
|
|
|
|
|
|
long index; |
59
|
|
|
|
|
|
|
return _BitScanReverse64(&index, value) ? index : 64; |
60
|
|
|
|
|
|
|
#endif |
61
|
|
|
|
|
|
|
} |
62
|
|
|
|
|
|
|
|
63
|
|
|
|
|
|
|
#else |
64
|
|
|
|
|
|
|
|
65
|
0
|
|
|
|
|
|
static inline uint32_t floor_log2(const uint64_t value) { |
66
|
0
|
|
|
|
|
|
return 63 - __builtin_clzll(value); |
67
|
|
|
|
|
|
|
} |
68
|
|
|
|
|
|
|
|
69
|
|
|
|
|
|
|
#endif |
70
|
|
|
|
|
|
|
|
71
|
|
|
|
|
|
|
// The max function is already defined on Windows. |
72
|
0
|
|
|
|
|
|
static inline int32_t max32(int32_t a, int32_t b) { |
73
|
0
|
|
|
|
|
|
return a < b ? b : a; |
74
|
|
|
|
|
|
|
} |
75
|
|
|
|
|
|
|
|
76
|
0
|
|
|
|
|
|
static inline double int64Bits2Double(uint64_t bits) { |
77
|
|
|
|
|
|
|
double f; |
78
|
0
|
|
|
|
|
|
memcpy(&f, &bits, sizeof(double)); |
79
|
0
|
|
|
|
|
|
return f; |
80
|
|
|
|
|
|
|
} |
81
|
|
|
|
|
|
|
|
82
|
0
|
|
|
|
|
|
enum Status s2d_n(const char * buffer, const int len, double * result) { |
83
|
0
|
0
|
|
|
|
|
if (len == 0) { |
84
|
0
|
|
|
|
|
|
return INPUT_TOO_SHORT; |
85
|
|
|
|
|
|
|
} |
86
|
0
|
|
|
|
|
|
int m10digits = 0; |
87
|
0
|
|
|
|
|
|
int e10digits = 0; |
88
|
0
|
|
|
|
|
|
int dotIndex = len; |
89
|
0
|
|
|
|
|
|
int eIndex = len; |
90
|
0
|
|
|
|
|
|
uint64_t m10 = 0; |
91
|
0
|
|
|
|
|
|
int32_t e10 = 0; |
92
|
0
|
|
|
|
|
|
bool signedM = false; |
93
|
0
|
|
|
|
|
|
bool signedE = false; |
94
|
0
|
|
|
|
|
|
int i = 0; |
95
|
0
|
0
|
|
|
|
|
if (buffer[i] == '-') { |
96
|
0
|
|
|
|
|
|
signedM = true; |
97
|
0
|
|
|
|
|
|
i++; |
98
|
|
|
|
|
|
|
} |
99
|
0
|
0
|
|
|
|
|
for (; i < len; i++) { |
100
|
0
|
|
|
|
|
|
char c = buffer[i]; |
101
|
0
|
0
|
|
|
|
|
if (c == '.') { |
102
|
0
|
0
|
|
|
|
|
if (dotIndex != len) { |
103
|
0
|
|
|
|
|
|
return MALFORMED_INPUT; |
104
|
|
|
|
|
|
|
} |
105
|
0
|
|
|
|
|
|
dotIndex = i; |
106
|
0
|
|
|
|
|
|
continue; |
107
|
|
|
|
|
|
|
} |
108
|
0
|
0
|
|
|
|
|
if ((c < '0') || (c > '9')) { |
|
|
0
|
|
|
|
|
|
109
|
|
|
|
|
|
|
break; |
110
|
|
|
|
|
|
|
} |
111
|
0
|
0
|
|
|
|
|
if (m10digits >= 17) { |
112
|
0
|
|
|
|
|
|
return INPUT_TOO_LONG; |
113
|
|
|
|
|
|
|
} |
114
|
0
|
|
|
|
|
|
m10 = 10 * m10 + (c - '0'); |
115
|
0
|
0
|
|
|
|
|
if (m10 != 0) { |
116
|
0
|
|
|
|
|
|
m10digits++; |
117
|
|
|
|
|
|
|
} |
118
|
|
|
|
|
|
|
} |
119
|
0
|
0
|
|
|
|
|
if (i < len && ((buffer[i] == 'e') || (buffer[i] == 'E'))) { |
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
120
|
0
|
|
|
|
|
|
eIndex = i; |
121
|
0
|
|
|
|
|
|
i++; |
122
|
0
|
0
|
|
|
|
|
if (i < len && ((buffer[i] == '-') || (buffer[i] == '+'))) { |
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
123
|
0
|
|
|
|
|
|
signedE = buffer[i] == '-'; |
124
|
0
|
|
|
|
|
|
i++; |
125
|
|
|
|
|
|
|
} |
126
|
0
|
0
|
|
|
|
|
for (; i < len; i++) { |
127
|
0
|
|
|
|
|
|
char c = buffer[i]; |
128
|
0
|
0
|
|
|
|
|
if ((c < '0') || (c > '9')) { |
|
|
0
|
|
|
|
|
|
129
|
0
|
|
|
|
|
|
return MALFORMED_INPUT; |
130
|
|
|
|
|
|
|
} |
131
|
0
|
0
|
|
|
|
|
if (e10digits > 3) { |
132
|
|
|
|
|
|
|
// TODO: Be more lenient. Return +/-Infinity or +/-0 instead. |
133
|
0
|
|
|
|
|
|
return INPUT_TOO_LONG; |
134
|
|
|
|
|
|
|
} |
135
|
0
|
|
|
|
|
|
e10 = 10 * e10 + (c - '0'); |
136
|
0
|
0
|
|
|
|
|
if (e10 != 0) { |
137
|
0
|
|
|
|
|
|
e10digits++; |
138
|
|
|
|
|
|
|
} |
139
|
|
|
|
|
|
|
} |
140
|
|
|
|
|
|
|
} |
141
|
0
|
0
|
|
|
|
|
if (i < len) { |
142
|
0
|
|
|
|
|
|
return MALFORMED_INPUT; |
143
|
|
|
|
|
|
|
} |
144
|
0
|
0
|
|
|
|
|
if (signedE) { |
145
|
0
|
|
|
|
|
|
e10 = -e10; |
146
|
|
|
|
|
|
|
} |
147
|
0
|
0
|
|
|
|
|
e10 -= dotIndex < eIndex ? eIndex - dotIndex - 1 : 0; |
148
|
0
|
0
|
|
|
|
|
if (m10 == 0) { |
149
|
0
|
0
|
|
|
|
|
*result = signedM ? -0.0 : 0.0; |
150
|
0
|
|
|
|
|
|
return SUCCESS; |
151
|
|
|
|
|
|
|
} |
152
|
|
|
|
|
|
|
|
153
|
|
|
|
|
|
|
#ifdef RYU_DEBUG |
154
|
0
|
|
|
|
|
|
printf("Input=%s\n", buffer); |
155
|
0
|
|
|
|
|
|
printf("m10digits = %d\n", m10digits); |
156
|
0
|
|
|
|
|
|
printf("e10digits = %d\n", e10digits); |
157
|
0
|
|
|
|
|
|
printf("m10 * 10^e10 = %" PRIu64 " * 10^%d\n", m10, e10); |
158
|
|
|
|
|
|
|
#endif |
159
|
|
|
|
|
|
|
|
160
|
0
|
0
|
|
|
|
|
if ((m10digits + e10 <= -324) || (m10 == 0)) { |
|
|
0
|
|
|
|
|
|
161
|
|
|
|
|
|
|
// Number is less than 1e-324, which should be rounded down to 0; return +/-0.0. |
162
|
0
|
|
|
|
|
|
uint64_t ieee = ((uint64_t) signedM) << (DOUBLE_EXPONENT_BITS + DOUBLE_MANTISSA_BITS); |
163
|
0
|
|
|
|
|
|
*result = int64Bits2Double(ieee); |
164
|
0
|
|
|
|
|
|
return SUCCESS; |
165
|
|
|
|
|
|
|
} |
166
|
0
|
0
|
|
|
|
|
if (m10digits + e10 >= 310) { |
167
|
|
|
|
|
|
|
// Number is larger than 1e+309, which should be rounded to +/-Infinity. |
168
|
0
|
|
|
|
|
|
uint64_t ieee = (((uint64_t) signedM) << (DOUBLE_EXPONENT_BITS + DOUBLE_MANTISSA_BITS)) | (0x7ffull << DOUBLE_MANTISSA_BITS); |
169
|
0
|
|
|
|
|
|
*result = int64Bits2Double(ieee); |
170
|
0
|
|
|
|
|
|
return SUCCESS; |
171
|
|
|
|
|
|
|
} |
172
|
|
|
|
|
|
|
|
173
|
|
|
|
|
|
|
// Convert to binary float m2 * 2^e2, while retaining information about whether the conversion |
174
|
|
|
|
|
|
|
// was exact (trailingZeros). |
175
|
|
|
|
|
|
|
int32_t e2; |
176
|
|
|
|
|
|
|
uint64_t m2; |
177
|
|
|
|
|
|
|
bool trailingZeros; |
178
|
0
|
0
|
|
|
|
|
if (e10 >= 0) { |
179
|
|
|
|
|
|
|
// The length of m * 10^e in bits is: |
180
|
|
|
|
|
|
|
// log2(m10 * 10^e10) = log2(m10) + e10 log2(10) = log2(m10) + e10 + e10 * log2(5) |
181
|
|
|
|
|
|
|
// |
182
|
|
|
|
|
|
|
// We want to compute the DOUBLE_MANTISSA_BITS + 1 top-most bits (+1 for the implicit leading |
183
|
|
|
|
|
|
|
// one in IEEE format). We therefore choose a binary output exponent of |
184
|
|
|
|
|
|
|
// log2(m10 * 10^e10) - (DOUBLE_MANTISSA_BITS + 1). |
185
|
|
|
|
|
|
|
// |
186
|
|
|
|
|
|
|
// We use floor(log2(5^e10)) so that we get at least this many bits; better to |
187
|
|
|
|
|
|
|
// have an additional bit than to not have enough bits. |
188
|
0
|
|
|
|
|
|
e2 = floor_log2(m10) + e10 + log2pow5(e10) - (DOUBLE_MANTISSA_BITS + 1); |
189
|
|
|
|
|
|
|
|
190
|
|
|
|
|
|
|
// We now compute [m10 * 10^e10 / 2^e2] = [m10 * 5^e10 / 2^(e2-e10)]. |
191
|
|
|
|
|
|
|
// To that end, we use the DOUBLE_POW5_SPLIT table. |
192
|
0
|
|
|
|
|
|
int j = e2 - e10 - ceil_log2pow5(e10) + DOUBLE_POW5_BITCOUNT; |
193
|
0
|
0
|
|
|
|
|
assert(j >= 0); |
194
|
|
|
|
|
|
|
#if defined(RYU_OPTIMIZE_SIZE) |
195
|
|
|
|
|
|
|
uint64_t pow5[2]; |
196
|
|
|
|
|
|
|
double_computePow5(e10, pow5); |
197
|
|
|
|
|
|
|
m2 = mulShift64(m10, pow5, j); |
198
|
|
|
|
|
|
|
#else |
199
|
0
|
0
|
|
|
|
|
assert(e10 < DOUBLE_POW5_TABLE_SIZE); |
200
|
0
|
|
|
|
|
|
m2 = mulShift64(m10, DOUBLE_POW5_SPLIT[e10], j); |
201
|
|
|
|
|
|
|
#endif |
202
|
|
|
|
|
|
|
// We also compute if the result is exact, i.e., |
203
|
|
|
|
|
|
|
// [m10 * 10^e10 / 2^e2] == m10 * 10^e10 / 2^e2. |
204
|
|
|
|
|
|
|
// This can only be the case if 2^e2 divides m10 * 10^e10, which in turn requires that the |
205
|
|
|
|
|
|
|
// largest power of 2 that divides m10 + e10 is greater than e2. If e2 is less than e10, then |
206
|
|
|
|
|
|
|
// the result must be exact. Otherwise we use the existing multipleOfPowerOf2 function. |
207
|
0
|
0
|
|
|
|
|
trailingZeros = e2 < e10 || (e2 - e10 < 64 && multipleOfPowerOf2(m10, e2 - e10)); |
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
208
|
|
|
|
|
|
|
} else { |
209
|
0
|
|
|
|
|
|
e2 = floor_log2(m10) + e10 - ceil_log2pow5(-e10) - (DOUBLE_MANTISSA_BITS + 1); |
210
|
0
|
|
|
|
|
|
int j = e2 - e10 + ceil_log2pow5(-e10) - 1 + DOUBLE_POW5_INV_BITCOUNT; |
211
|
|
|
|
|
|
|
#if defined(RYU_OPTIMIZE_SIZE) |
212
|
|
|
|
|
|
|
uint64_t pow5[2]; |
213
|
|
|
|
|
|
|
double_computeInvPow5(-e10, pow5); |
214
|
|
|
|
|
|
|
m2 = mulShift64(m10, pow5, j); |
215
|
|
|
|
|
|
|
#else |
216
|
0
|
0
|
|
|
|
|
assert(-e10 < DOUBLE_POW5_INV_TABLE_SIZE); |
217
|
0
|
|
|
|
|
|
m2 = mulShift64(m10, DOUBLE_POW5_INV_SPLIT[-e10], j); |
218
|
|
|
|
|
|
|
#endif |
219
|
0
|
|
|
|
|
|
trailingZeros = multipleOfPowerOf5(m10, -e10); |
220
|
|
|
|
|
|
|
} |
221
|
|
|
|
|
|
|
|
222
|
|
|
|
|
|
|
#ifdef RYU_DEBUG |
223
|
0
|
|
|
|
|
|
printf("m2 * 2^e2 = %" PRIu64 " * 2^%d\n", m2, e2); |
224
|
|
|
|
|
|
|
#endif |
225
|
|
|
|
|
|
|
|
226
|
|
|
|
|
|
|
// Compute the final IEEE exponent. |
227
|
0
|
|
|
|
|
|
uint32_t ieee_e2 = (uint32_t) max32(0, e2 + DOUBLE_EXPONENT_BIAS + floor_log2(m2)); |
228
|
|
|
|
|
|
|
|
229
|
0
|
0
|
|
|
|
|
if (ieee_e2 > 0x7fe) { |
230
|
|
|
|
|
|
|
// Final IEEE exponent is larger than the maximum representable; return +/-Infinity. |
231
|
0
|
|
|
|
|
|
uint64_t ieee = (((uint64_t) signedM) << (DOUBLE_EXPONENT_BITS + DOUBLE_MANTISSA_BITS)) | (0x7ffull << DOUBLE_MANTISSA_BITS); |
232
|
0
|
|
|
|
|
|
*result = int64Bits2Double(ieee); |
233
|
0
|
|
|
|
|
|
return SUCCESS; |
234
|
|
|
|
|
|
|
} |
235
|
|
|
|
|
|
|
|
236
|
|
|
|
|
|
|
// We need to figure out how much we need to shift m2. The tricky part is that we need to take |
237
|
|
|
|
|
|
|
// the final IEEE exponent into account, so we need to reverse the bias and also special-case |
238
|
|
|
|
|
|
|
// the value 0. |
239
|
0
|
0
|
|
|
|
|
int32_t shift = (ieee_e2 == 0 ? 1 : ieee_e2) - e2 - DOUBLE_EXPONENT_BIAS - DOUBLE_MANTISSA_BITS; |
240
|
0
|
0
|
|
|
|
|
assert(shift >= 0); |
241
|
|
|
|
|
|
|
#ifdef RYU_DEBUG |
242
|
0
|
|
|
|
|
|
printf("ieee_e2 = %d\n", ieee_e2); |
243
|
0
|
|
|
|
|
|
printf("shift = %d\n", shift); |
244
|
|
|
|
|
|
|
#endif |
245
|
|
|
|
|
|
|
|
246
|
|
|
|
|
|
|
// We need to round up if the exact value is more than 0.5 above the value we computed. That's |
247
|
|
|
|
|
|
|
// equivalent to checking if the last removed bit was 1 and either the value was not just |
248
|
|
|
|
|
|
|
// trailing zeros or the result would otherwise be odd. |
249
|
|
|
|
|
|
|
// |
250
|
|
|
|
|
|
|
// We need to update trailingZeros given that we have the exact output exponent ieee_e2 now. |
251
|
0
|
|
|
|
|
|
trailingZeros &= (m2 & ((1ull << (shift - 1)) - 1)) == 0; |
252
|
0
|
|
|
|
|
|
uint64_t lastRemovedBit = (m2 >> (shift - 1)) & 1; |
253
|
0
|
0
|
|
|
|
|
bool roundUp = (lastRemovedBit != 0) && (!trailingZeros || (((m2 >> shift) & 1) != 0)); |
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
254
|
|
|
|
|
|
|
|
255
|
|
|
|
|
|
|
#ifdef RYU_DEBUG |
256
|
0
|
|
|
|
|
|
printf("roundUp = %d\n", roundUp); |
257
|
0
|
|
|
|
|
|
printf("ieee_m2 = %" PRIu64 "\n", (m2 >> shift) + roundUp); |
258
|
|
|
|
|
|
|
#endif |
259
|
0
|
|
|
|
|
|
uint64_t ieee_m2 = (m2 >> shift) + roundUp; |
260
|
0
|
0
|
|
|
|
|
assert(ieee_m2 <= (1ull << (DOUBLE_MANTISSA_BITS + 1))); |
261
|
0
|
|
|
|
|
|
ieee_m2 &= (1ull << DOUBLE_MANTISSA_BITS) - 1; |
262
|
0
|
0
|
|
|
|
|
if (ieee_m2 == 0 && roundUp) { |
|
|
0
|
|
|
|
|
|
263
|
|
|
|
|
|
|
// Due to how the IEEE represents +/-Infinity, we don't need to check for overflow here. |
264
|
0
|
|
|
|
|
|
ieee_e2++; |
265
|
|
|
|
|
|
|
} |
266
|
|
|
|
|
|
|
|
267
|
0
|
|
|
|
|
|
uint64_t ieee = (((((uint64_t) signedM) << DOUBLE_EXPONENT_BITS) | (uint64_t)ieee_e2) << DOUBLE_MANTISSA_BITS) | ieee_m2; |
268
|
0
|
|
|
|
|
|
*result = int64Bits2Double(ieee); |
269
|
0
|
|
|
|
|
|
return SUCCESS; |
270
|
|
|
|
|
|
|
} |
271
|
|
|
|
|
|
|
|
272
|
0
|
|
|
|
|
|
enum Status s2d(const char * buffer, double * result) { |
273
|
0
|
|
|
|
|
|
return s2d_n(buffer, strlen(buffer), result); |
274
|
|
|
|
|
|
|
} |