]>
Commit | Line | Data |
---|---|---|
0f5d89e8 A |
1 | // © 2018 and later: Unicode, Inc. and others. |
2 | // License & terms of use: http://www.unicode.org/copyright.html | |
3 | // | |
4 | // From the double-conversion library. Original license: | |
5 | // | |
6 | // Copyright 2010 the V8 project authors. All rights reserved. | |
7 | // Redistribution and use in source and binary forms, with or without | |
8 | // modification, are permitted provided that the following conditions are | |
9 | // met: | |
10 | // | |
11 | // * Redistributions of source code must retain the above copyright | |
12 | // notice, this list of conditions and the following disclaimer. | |
13 | // * Redistributions in binary form must reproduce the above | |
14 | // copyright notice, this list of conditions and the following | |
15 | // disclaimer in the documentation and/or other materials provided | |
16 | // with the distribution. | |
17 | // * Neither the name of Google Inc. nor the names of its | |
18 | // contributors may be used to endorse or promote products derived | |
19 | // from this software without specific prior written permission. | |
20 | // | |
21 | // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS | |
22 | // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT | |
23 | // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR | |
24 | // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT | |
25 | // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, | |
26 | // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT | |
27 | // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, | |
28 | // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY | |
29 | // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT | |
30 | // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE | |
31 | // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. | |
32 | ||
33 | // ICU PATCH: ifdef around UCONFIG_NO_FORMATTING | |
34 | #include "unicode/utypes.h" | |
35 | #if !UCONFIG_NO_FORMATTING | |
36 | ||
3d1f044b | 37 | #include <cmath> |
0f5d89e8 A |
38 | |
39 | // ICU PATCH: Customize header file paths for ICU. | |
40 | ||
41 | #include "double-conversion-bignum-dtoa.h" | |
42 | ||
43 | #include "double-conversion-bignum.h" | |
44 | #include "double-conversion-ieee.h" | |
45 | ||
46 | // ICU PATCH: Wrap in ICU namespace | |
47 | U_NAMESPACE_BEGIN | |
48 | ||
49 | namespace double_conversion { | |
50 | ||
51 | static int NormalizedExponent(uint64_t significand, int exponent) { | |
52 | ASSERT(significand != 0); | |
53 | while ((significand & Double::kHiddenBit) == 0) { | |
54 | significand = significand << 1; | |
55 | exponent = exponent - 1; | |
56 | } | |
57 | return exponent; | |
58 | } | |
59 | ||
60 | ||
61 | // Forward declarations: | |
62 | // Returns an estimation of k such that 10^(k-1) <= v < 10^k. | |
63 | static int EstimatePower(int exponent); | |
64 | // Computes v / 10^estimated_power exactly, as a ratio of two bignums, numerator | |
65 | // and denominator. | |
66 | static void InitialScaledStartValues(uint64_t significand, | |
67 | int exponent, | |
68 | bool lower_boundary_is_closer, | |
69 | int estimated_power, | |
70 | bool need_boundary_deltas, | |
71 | Bignum* numerator, | |
72 | Bignum* denominator, | |
73 | Bignum* delta_minus, | |
74 | Bignum* delta_plus); | |
75 | // Multiplies numerator/denominator so that its values lies in the range 1-10. | |
76 | // Returns decimal_point s.t. | |
77 | // v = numerator'/denominator' * 10^(decimal_point-1) | |
78 | // where numerator' and denominator' are the values of numerator and | |
79 | // denominator after the call to this function. | |
80 | static void FixupMultiply10(int estimated_power, bool is_even, | |
81 | int* decimal_point, | |
82 | Bignum* numerator, Bignum* denominator, | |
83 | Bignum* delta_minus, Bignum* delta_plus); | |
84 | // Generates digits from the left to the right and stops when the generated | |
85 | // digits yield the shortest decimal representation of v. | |
86 | static void GenerateShortestDigits(Bignum* numerator, Bignum* denominator, | |
87 | Bignum* delta_minus, Bignum* delta_plus, | |
88 | bool is_even, | |
89 | Vector<char> buffer, int* length); | |
90 | // Generates 'requested_digits' after the decimal point. | |
91 | static void BignumToFixed(int requested_digits, int* decimal_point, | |
92 | Bignum* numerator, Bignum* denominator, | |
93 | Vector<char>(buffer), int* length); | |
94 | // Generates 'count' digits of numerator/denominator. | |
95 | // Once 'count' digits have been produced rounds the result depending on the | |
96 | // remainder (remainders of exactly .5 round upwards). Might update the | |
97 | // decimal_point when rounding up (for example for 0.9999). | |
98 | static void GenerateCountedDigits(int count, int* decimal_point, | |
99 | Bignum* numerator, Bignum* denominator, | |
100 | Vector<char>(buffer), int* length); | |
101 | ||
102 | ||
103 | void BignumDtoa(double v, BignumDtoaMode mode, int requested_digits, | |
104 | Vector<char> buffer, int* length, int* decimal_point) { | |
105 | ASSERT(v > 0); | |
106 | ASSERT(!Double(v).IsSpecial()); | |
107 | uint64_t significand; | |
108 | int exponent; | |
109 | bool lower_boundary_is_closer; | |
110 | if (mode == BIGNUM_DTOA_SHORTEST_SINGLE) { | |
111 | float f = static_cast<float>(v); | |
112 | ASSERT(f == v); | |
113 | significand = Single(f).Significand(); | |
114 | exponent = Single(f).Exponent(); | |
115 | lower_boundary_is_closer = Single(f).LowerBoundaryIsCloser(); | |
116 | } else { | |
117 | significand = Double(v).Significand(); | |
118 | exponent = Double(v).Exponent(); | |
119 | lower_boundary_is_closer = Double(v).LowerBoundaryIsCloser(); | |
120 | } | |
121 | bool need_boundary_deltas = | |
122 | (mode == BIGNUM_DTOA_SHORTEST || mode == BIGNUM_DTOA_SHORTEST_SINGLE); | |
123 | ||
124 | bool is_even = (significand & 1) == 0; | |
125 | int normalized_exponent = NormalizedExponent(significand, exponent); | |
126 | // estimated_power might be too low by 1. | |
127 | int estimated_power = EstimatePower(normalized_exponent); | |
128 | ||
129 | // Shortcut for Fixed. | |
130 | // The requested digits correspond to the digits after the point. If the | |
131 | // number is much too small, then there is no need in trying to get any | |
132 | // digits. | |
133 | if (mode == BIGNUM_DTOA_FIXED && -estimated_power - 1 > requested_digits) { | |
134 | buffer[0] = '\0'; | |
135 | *length = 0; | |
136 | // Set decimal-point to -requested_digits. This is what Gay does. | |
137 | // Note that it should not have any effect anyways since the string is | |
138 | // empty. | |
139 | *decimal_point = -requested_digits; | |
140 | return; | |
141 | } | |
142 | ||
143 | Bignum numerator; | |
144 | Bignum denominator; | |
145 | Bignum delta_minus; | |
146 | Bignum delta_plus; | |
147 | // Make sure the bignum can grow large enough. The smallest double equals | |
148 | // 4e-324. In this case the denominator needs fewer than 324*4 binary digits. | |
149 | // The maximum double is 1.7976931348623157e308 which needs fewer than | |
150 | // 308*4 binary digits. | |
151 | ASSERT(Bignum::kMaxSignificantBits >= 324*4); | |
152 | InitialScaledStartValues(significand, exponent, lower_boundary_is_closer, | |
153 | estimated_power, need_boundary_deltas, | |
154 | &numerator, &denominator, | |
155 | &delta_minus, &delta_plus); | |
156 | // We now have v = (numerator / denominator) * 10^estimated_power. | |
157 | FixupMultiply10(estimated_power, is_even, decimal_point, | |
158 | &numerator, &denominator, | |
159 | &delta_minus, &delta_plus); | |
160 | // We now have v = (numerator / denominator) * 10^(decimal_point-1), and | |
161 | // 1 <= (numerator + delta_plus) / denominator < 10 | |
162 | switch (mode) { | |
163 | case BIGNUM_DTOA_SHORTEST: | |
164 | case BIGNUM_DTOA_SHORTEST_SINGLE: | |
165 | GenerateShortestDigits(&numerator, &denominator, | |
166 | &delta_minus, &delta_plus, | |
167 | is_even, buffer, length); | |
168 | break; | |
169 | case BIGNUM_DTOA_FIXED: | |
170 | BignumToFixed(requested_digits, decimal_point, | |
171 | &numerator, &denominator, | |
172 | buffer, length); | |
173 | break; | |
174 | case BIGNUM_DTOA_PRECISION: | |
175 | GenerateCountedDigits(requested_digits, decimal_point, | |
176 | &numerator, &denominator, | |
177 | buffer, length); | |
178 | break; | |
179 | default: | |
180 | UNREACHABLE(); | |
181 | } | |
182 | buffer[*length] = '\0'; | |
183 | } | |
184 | ||
185 | ||
186 | // The procedure starts generating digits from the left to the right and stops | |
187 | // when the generated digits yield the shortest decimal representation of v. A | |
188 | // decimal representation of v is a number lying closer to v than to any other | |
189 | // double, so it converts to v when read. | |
190 | // | |
191 | // This is true if d, the decimal representation, is between m- and m+, the | |
192 | // upper and lower boundaries. d must be strictly between them if !is_even. | |
193 | // m- := (numerator - delta_minus) / denominator | |
194 | // m+ := (numerator + delta_plus) / denominator | |
195 | // | |
196 | // Precondition: 0 <= (numerator+delta_plus) / denominator < 10. | |
197 | // If 1 <= (numerator+delta_plus) / denominator < 10 then no leading 0 digit | |
198 | // will be produced. This should be the standard precondition. | |
199 | static void GenerateShortestDigits(Bignum* numerator, Bignum* denominator, | |
200 | Bignum* delta_minus, Bignum* delta_plus, | |
201 | bool is_even, | |
202 | Vector<char> buffer, int* length) { | |
203 | // Small optimization: if delta_minus and delta_plus are the same just reuse | |
204 | // one of the two bignums. | |
205 | if (Bignum::Equal(*delta_minus, *delta_plus)) { | |
206 | delta_plus = delta_minus; | |
207 | } | |
208 | *length = 0; | |
209 | for (;;) { | |
210 | uint16_t digit; | |
211 | digit = numerator->DivideModuloIntBignum(*denominator); | |
212 | ASSERT(digit <= 9); // digit is a uint16_t and therefore always positive. | |
213 | // digit = numerator / denominator (integer division). | |
214 | // numerator = numerator % denominator. | |
215 | buffer[(*length)++] = static_cast<char>(digit + '0'); | |
216 | ||
217 | // Can we stop already? | |
218 | // If the remainder of the division is less than the distance to the lower | |
219 | // boundary we can stop. In this case we simply round down (discarding the | |
220 | // remainder). | |
221 | // Similarly we test if we can round up (using the upper boundary). | |
222 | bool in_delta_room_minus; | |
223 | bool in_delta_room_plus; | |
224 | if (is_even) { | |
225 | in_delta_room_minus = Bignum::LessEqual(*numerator, *delta_minus); | |
226 | } else { | |
227 | in_delta_room_minus = Bignum::Less(*numerator, *delta_minus); | |
228 | } | |
229 | if (is_even) { | |
230 | in_delta_room_plus = | |
231 | Bignum::PlusCompare(*numerator, *delta_plus, *denominator) >= 0; | |
232 | } else { | |
233 | in_delta_room_plus = | |
234 | Bignum::PlusCompare(*numerator, *delta_plus, *denominator) > 0; | |
235 | } | |
236 | if (!in_delta_room_minus && !in_delta_room_plus) { | |
237 | // Prepare for next iteration. | |
238 | numerator->Times10(); | |
239 | delta_minus->Times10(); | |
240 | // We optimized delta_plus to be equal to delta_minus (if they share the | |
241 | // same value). So don't multiply delta_plus if they point to the same | |
242 | // object. | |
243 | if (delta_minus != delta_plus) { | |
244 | delta_plus->Times10(); | |
245 | } | |
246 | } else if (in_delta_room_minus && in_delta_room_plus) { | |
247 | // Let's see if 2*numerator < denominator. | |
248 | // If yes, then the next digit would be < 5 and we can round down. | |
249 | int compare = Bignum::PlusCompare(*numerator, *numerator, *denominator); | |
250 | if (compare < 0) { | |
251 | // Remaining digits are less than .5. -> Round down (== do nothing). | |
252 | } else if (compare > 0) { | |
253 | // Remaining digits are more than .5 of denominator. -> Round up. | |
254 | // Note that the last digit could not be a '9' as otherwise the whole | |
255 | // loop would have stopped earlier. | |
256 | // We still have an assert here in case the preconditions were not | |
257 | // satisfied. | |
258 | ASSERT(buffer[(*length) - 1] != '9'); | |
259 | buffer[(*length) - 1]++; | |
260 | } else { | |
261 | // Halfway case. | |
262 | // TODO(floitsch): need a way to solve half-way cases. | |
263 | // For now let's round towards even (since this is what Gay seems to | |
264 | // do). | |
265 | ||
266 | if ((buffer[(*length) - 1] - '0') % 2 == 0) { | |
267 | // Round down => Do nothing. | |
268 | } else { | |
269 | ASSERT(buffer[(*length) - 1] != '9'); | |
270 | buffer[(*length) - 1]++; | |
271 | } | |
272 | } | |
273 | return; | |
274 | } else if (in_delta_room_minus) { | |
275 | // Round down (== do nothing). | |
276 | return; | |
277 | } else { // in_delta_room_plus | |
278 | // Round up. | |
279 | // Note again that the last digit could not be '9' since this would have | |
280 | // stopped the loop earlier. | |
281 | // We still have an ASSERT here, in case the preconditions were not | |
282 | // satisfied. | |
283 | ASSERT(buffer[(*length) -1] != '9'); | |
284 | buffer[(*length) - 1]++; | |
285 | return; | |
286 | } | |
287 | } | |
288 | } | |
289 | ||
290 | ||
291 | // Let v = numerator / denominator < 10. | |
292 | // Then we generate 'count' digits of d = x.xxxxx... (without the decimal point) | |
293 | // from left to right. Once 'count' digits have been produced we decide wether | |
294 | // to round up or down. Remainders of exactly .5 round upwards. Numbers such | |
295 | // as 9.999999 propagate a carry all the way, and change the | |
296 | // exponent (decimal_point), when rounding upwards. | |
297 | static void GenerateCountedDigits(int count, int* decimal_point, | |
298 | Bignum* numerator, Bignum* denominator, | |
299 | Vector<char> buffer, int* length) { | |
300 | ASSERT(count >= 0); | |
301 | for (int i = 0; i < count - 1; ++i) { | |
302 | uint16_t digit; | |
303 | digit = numerator->DivideModuloIntBignum(*denominator); | |
304 | ASSERT(digit <= 9); // digit is a uint16_t and therefore always positive. | |
305 | // digit = numerator / denominator (integer division). | |
306 | // numerator = numerator % denominator. | |
307 | buffer[i] = static_cast<char>(digit + '0'); | |
308 | // Prepare for next iteration. | |
309 | numerator->Times10(); | |
310 | } | |
311 | // Generate the last digit. | |
312 | uint16_t digit; | |
313 | digit = numerator->DivideModuloIntBignum(*denominator); | |
314 | if (Bignum::PlusCompare(*numerator, *numerator, *denominator) >= 0) { | |
315 | digit++; | |
316 | } | |
317 | ASSERT(digit <= 10); | |
318 | buffer[count - 1] = static_cast<char>(digit + '0'); | |
319 | // Correct bad digits (in case we had a sequence of '9's). Propagate the | |
320 | // carry until we hat a non-'9' or til we reach the first digit. | |
321 | for (int i = count - 1; i > 0; --i) { | |
322 | if (buffer[i] != '0' + 10) break; | |
323 | buffer[i] = '0'; | |
324 | buffer[i - 1]++; | |
325 | } | |
326 | if (buffer[0] == '0' + 10) { | |
327 | // Propagate a carry past the top place. | |
328 | buffer[0] = '1'; | |
329 | (*decimal_point)++; | |
330 | } | |
331 | *length = count; | |
332 | } | |
333 | ||
334 | ||
335 | // Generates 'requested_digits' after the decimal point. It might omit | |
336 | // trailing '0's. If the input number is too small then no digits at all are | |
337 | // generated (ex.: 2 fixed digits for 0.00001). | |
338 | // | |
339 | // Input verifies: 1 <= (numerator + delta) / denominator < 10. | |
340 | static void BignumToFixed(int requested_digits, int* decimal_point, | |
341 | Bignum* numerator, Bignum* denominator, | |
342 | Vector<char>(buffer), int* length) { | |
343 | // Note that we have to look at more than just the requested_digits, since | |
344 | // a number could be rounded up. Example: v=0.5 with requested_digits=0. | |
345 | // Even though the power of v equals 0 we can't just stop here. | |
346 | if (-(*decimal_point) > requested_digits) { | |
347 | // The number is definitively too small. | |
348 | // Ex: 0.001 with requested_digits == 1. | |
349 | // Set decimal-point to -requested_digits. This is what Gay does. | |
350 | // Note that it should not have any effect anyways since the string is | |
351 | // empty. | |
352 | *decimal_point = -requested_digits; | |
353 | *length = 0; | |
354 | return; | |
355 | } else if (-(*decimal_point) == requested_digits) { | |
356 | // We only need to verify if the number rounds down or up. | |
357 | // Ex: 0.04 and 0.06 with requested_digits == 1. | |
358 | ASSERT(*decimal_point == -requested_digits); | |
359 | // Initially the fraction lies in range (1, 10]. Multiply the denominator | |
360 | // by 10 so that we can compare more easily. | |
361 | denominator->Times10(); | |
362 | if (Bignum::PlusCompare(*numerator, *numerator, *denominator) >= 0) { | |
363 | // If the fraction is >= 0.5 then we have to include the rounded | |
364 | // digit. | |
365 | buffer[0] = '1'; | |
366 | *length = 1; | |
367 | (*decimal_point)++; | |
368 | } else { | |
369 | // Note that we caught most of similar cases earlier. | |
370 | *length = 0; | |
371 | } | |
372 | return; | |
373 | } else { | |
374 | // The requested digits correspond to the digits after the point. | |
375 | // The variable 'needed_digits' includes the digits before the point. | |
376 | int needed_digits = (*decimal_point) + requested_digits; | |
377 | GenerateCountedDigits(needed_digits, decimal_point, | |
378 | numerator, denominator, | |
379 | buffer, length); | |
380 | } | |
381 | } | |
382 | ||
383 | ||
384 | // Returns an estimation of k such that 10^(k-1) <= v < 10^k where | |
385 | // v = f * 2^exponent and 2^52 <= f < 2^53. | |
386 | // v is hence a normalized double with the given exponent. The output is an | |
387 | // approximation for the exponent of the decimal approimation .digits * 10^k. | |
388 | // | |
389 | // The result might undershoot by 1 in which case 10^k <= v < 10^k+1. | |
390 | // Note: this property holds for v's upper boundary m+ too. | |
391 | // 10^k <= m+ < 10^k+1. | |
392 | // (see explanation below). | |
393 | // | |
394 | // Examples: | |
395 | // EstimatePower(0) => 16 | |
396 | // EstimatePower(-52) => 0 | |
397 | // | |
398 | // Note: e >= 0 => EstimatedPower(e) > 0. No similar claim can be made for e<0. | |
399 | static int EstimatePower(int exponent) { | |
400 | // This function estimates log10 of v where v = f*2^e (with e == exponent). | |
401 | // Note that 10^floor(log10(v)) <= v, but v <= 10^ceil(log10(v)). | |
402 | // Note that f is bounded by its container size. Let p = 53 (the double's | |
403 | // significand size). Then 2^(p-1) <= f < 2^p. | |
404 | // | |
405 | // Given that log10(v) == log2(v)/log2(10) and e+(len(f)-1) is quite close | |
406 | // to log2(v) the function is simplified to (e+(len(f)-1)/log2(10)). | |
407 | // The computed number undershoots by less than 0.631 (when we compute log3 | |
408 | // and not log10). | |
409 | // | |
410 | // Optimization: since we only need an approximated result this computation | |
411 | // can be performed on 64 bit integers. On x86/x64 architecture the speedup is | |
412 | // not really measurable, though. | |
413 | // | |
414 | // Since we want to avoid overshooting we decrement by 1e10 so that | |
415 | // floating-point imprecisions don't affect us. | |
416 | // | |
417 | // Explanation for v's boundary m+: the computation takes advantage of | |
418 | // the fact that 2^(p-1) <= f < 2^p. Boundaries still satisfy this requirement | |
419 | // (even for denormals where the delta can be much more important). | |
420 | ||
421 | const double k1Log10 = 0.30102999566398114; // 1/lg(10) | |
422 | ||
423 | // For doubles len(f) == 53 (don't forget the hidden bit). | |
424 | const int kSignificandSize = Double::kSignificandSize; | |
425 | double estimate = ceil((exponent + kSignificandSize - 1) * k1Log10 - 1e-10); | |
426 | return static_cast<int>(estimate); | |
427 | } | |
428 | ||
429 | ||
430 | // See comments for InitialScaledStartValues. | |
431 | static void InitialScaledStartValuesPositiveExponent( | |
432 | uint64_t significand, int exponent, | |
433 | int estimated_power, bool need_boundary_deltas, | |
434 | Bignum* numerator, Bignum* denominator, | |
435 | Bignum* delta_minus, Bignum* delta_plus) { | |
436 | // A positive exponent implies a positive power. | |
437 | ASSERT(estimated_power >= 0); | |
438 | // Since the estimated_power is positive we simply multiply the denominator | |
439 | // by 10^estimated_power. | |
440 | ||
441 | // numerator = v. | |
442 | numerator->AssignUInt64(significand); | |
443 | numerator->ShiftLeft(exponent); | |
444 | // denominator = 10^estimated_power. | |
445 | denominator->AssignPowerUInt16(10, estimated_power); | |
446 | ||
447 | if (need_boundary_deltas) { | |
448 | // Introduce a common denominator so that the deltas to the boundaries are | |
449 | // integers. | |
450 | denominator->ShiftLeft(1); | |
451 | numerator->ShiftLeft(1); | |
452 | // Let v = f * 2^e, then m+ - v = 1/2 * 2^e; With the common | |
453 | // denominator (of 2) delta_plus equals 2^e. | |
454 | delta_plus->AssignUInt16(1); | |
455 | delta_plus->ShiftLeft(exponent); | |
456 | // Same for delta_minus. The adjustments if f == 2^p-1 are done later. | |
457 | delta_minus->AssignUInt16(1); | |
458 | delta_minus->ShiftLeft(exponent); | |
459 | } | |
460 | } | |
461 | ||
462 | ||
463 | // See comments for InitialScaledStartValues | |
464 | static void InitialScaledStartValuesNegativeExponentPositivePower( | |
465 | uint64_t significand, int exponent, | |
466 | int estimated_power, bool need_boundary_deltas, | |
467 | Bignum* numerator, Bignum* denominator, | |
468 | Bignum* delta_minus, Bignum* delta_plus) { | |
469 | // v = f * 2^e with e < 0, and with estimated_power >= 0. | |
470 | // This means that e is close to 0 (have a look at how estimated_power is | |
471 | // computed). | |
472 | ||
473 | // numerator = significand | |
474 | // since v = significand * 2^exponent this is equivalent to | |
475 | // numerator = v * / 2^-exponent | |
476 | numerator->AssignUInt64(significand); | |
477 | // denominator = 10^estimated_power * 2^-exponent (with exponent < 0) | |
478 | denominator->AssignPowerUInt16(10, estimated_power); | |
479 | denominator->ShiftLeft(-exponent); | |
480 | ||
481 | if (need_boundary_deltas) { | |
482 | // Introduce a common denominator so that the deltas to the boundaries are | |
483 | // integers. | |
484 | denominator->ShiftLeft(1); | |
485 | numerator->ShiftLeft(1); | |
486 | // Let v = f * 2^e, then m+ - v = 1/2 * 2^e; With the common | |
487 | // denominator (of 2) delta_plus equals 2^e. | |
488 | // Given that the denominator already includes v's exponent the distance | |
489 | // to the boundaries is simply 1. | |
490 | delta_plus->AssignUInt16(1); | |
491 | // Same for delta_minus. The adjustments if f == 2^p-1 are done later. | |
492 | delta_minus->AssignUInt16(1); | |
493 | } | |
494 | } | |
495 | ||
496 | ||
497 | // See comments for InitialScaledStartValues | |
498 | static void InitialScaledStartValuesNegativeExponentNegativePower( | |
499 | uint64_t significand, int exponent, | |
500 | int estimated_power, bool need_boundary_deltas, | |
501 | Bignum* numerator, Bignum* denominator, | |
502 | Bignum* delta_minus, Bignum* delta_plus) { | |
503 | // Instead of multiplying the denominator with 10^estimated_power we | |
504 | // multiply all values (numerator and deltas) by 10^-estimated_power. | |
505 | ||
506 | // Use numerator as temporary container for power_ten. | |
507 | Bignum* power_ten = numerator; | |
508 | power_ten->AssignPowerUInt16(10, -estimated_power); | |
509 | ||
510 | if (need_boundary_deltas) { | |
511 | // Since power_ten == numerator we must make a copy of 10^estimated_power | |
512 | // before we complete the computation of the numerator. | |
513 | // delta_plus = delta_minus = 10^estimated_power | |
514 | delta_plus->AssignBignum(*power_ten); | |
515 | delta_minus->AssignBignum(*power_ten); | |
516 | } | |
517 | ||
518 | // numerator = significand * 2 * 10^-estimated_power | |
519 | // since v = significand * 2^exponent this is equivalent to | |
520 | // numerator = v * 10^-estimated_power * 2 * 2^-exponent. | |
521 | // Remember: numerator has been abused as power_ten. So no need to assign it | |
522 | // to itself. | |
523 | ASSERT(numerator == power_ten); | |
524 | numerator->MultiplyByUInt64(significand); | |
525 | ||
526 | // denominator = 2 * 2^-exponent with exponent < 0. | |
527 | denominator->AssignUInt16(1); | |
528 | denominator->ShiftLeft(-exponent); | |
529 | ||
530 | if (need_boundary_deltas) { | |
531 | // Introduce a common denominator so that the deltas to the boundaries are | |
532 | // integers. | |
533 | numerator->ShiftLeft(1); | |
534 | denominator->ShiftLeft(1); | |
535 | // With this shift the boundaries have their correct value, since | |
536 | // delta_plus = 10^-estimated_power, and | |
537 | // delta_minus = 10^-estimated_power. | |
538 | // These assignments have been done earlier. | |
539 | // The adjustments if f == 2^p-1 (lower boundary is closer) are done later. | |
540 | } | |
541 | } | |
542 | ||
543 | ||
544 | // Let v = significand * 2^exponent. | |
545 | // Computes v / 10^estimated_power exactly, as a ratio of two bignums, numerator | |
546 | // and denominator. The functions GenerateShortestDigits and | |
547 | // GenerateCountedDigits will then convert this ratio to its decimal | |
548 | // representation d, with the required accuracy. | |
549 | // Then d * 10^estimated_power is the representation of v. | |
550 | // (Note: the fraction and the estimated_power might get adjusted before | |
551 | // generating the decimal representation.) | |
552 | // | |
553 | // The initial start values consist of: | |
554 | // - a scaled numerator: s.t. numerator/denominator == v / 10^estimated_power. | |
555 | // - a scaled (common) denominator. | |
556 | // optionally (used by GenerateShortestDigits to decide if it has the shortest | |
557 | // decimal converting back to v): | |
558 | // - v - m-: the distance to the lower boundary. | |
559 | // - m+ - v: the distance to the upper boundary. | |
560 | // | |
561 | // v, m+, m-, and therefore v - m- and m+ - v all share the same denominator. | |
562 | // | |
563 | // Let ep == estimated_power, then the returned values will satisfy: | |
564 | // v / 10^ep = numerator / denominator. | |
565 | // v's boundarys m- and m+: | |
566 | // m- / 10^ep == v / 10^ep - delta_minus / denominator | |
567 | // m+ / 10^ep == v / 10^ep + delta_plus / denominator | |
568 | // Or in other words: | |
569 | // m- == v - delta_minus * 10^ep / denominator; | |
570 | // m+ == v + delta_plus * 10^ep / denominator; | |
571 | // | |
572 | // Since 10^(k-1) <= v < 10^k (with k == estimated_power) | |
573 | // or 10^k <= v < 10^(k+1) | |
574 | // we then have 0.1 <= numerator/denominator < 1 | |
575 | // or 1 <= numerator/denominator < 10 | |
576 | // | |
577 | // It is then easy to kickstart the digit-generation routine. | |
578 | // | |
579 | // The boundary-deltas are only filled if the mode equals BIGNUM_DTOA_SHORTEST | |
580 | // or BIGNUM_DTOA_SHORTEST_SINGLE. | |
581 | ||
582 | static void InitialScaledStartValues(uint64_t significand, | |
583 | int exponent, | |
584 | bool lower_boundary_is_closer, | |
585 | int estimated_power, | |
586 | bool need_boundary_deltas, | |
587 | Bignum* numerator, | |
588 | Bignum* denominator, | |
589 | Bignum* delta_minus, | |
590 | Bignum* delta_plus) { | |
591 | if (exponent >= 0) { | |
592 | InitialScaledStartValuesPositiveExponent( | |
593 | significand, exponent, estimated_power, need_boundary_deltas, | |
594 | numerator, denominator, delta_minus, delta_plus); | |
595 | } else if (estimated_power >= 0) { | |
596 | InitialScaledStartValuesNegativeExponentPositivePower( | |
597 | significand, exponent, estimated_power, need_boundary_deltas, | |
598 | numerator, denominator, delta_minus, delta_plus); | |
599 | } else { | |
600 | InitialScaledStartValuesNegativeExponentNegativePower( | |
601 | significand, exponent, estimated_power, need_boundary_deltas, | |
602 | numerator, denominator, delta_minus, delta_plus); | |
603 | } | |
604 | ||
605 | if (need_boundary_deltas && lower_boundary_is_closer) { | |
606 | // The lower boundary is closer at half the distance of "normal" numbers. | |
607 | // Increase the common denominator and adapt all but the delta_minus. | |
608 | denominator->ShiftLeft(1); // *2 | |
609 | numerator->ShiftLeft(1); // *2 | |
610 | delta_plus->ShiftLeft(1); // *2 | |
611 | } | |
612 | } | |
613 | ||
614 | ||
615 | // This routine multiplies numerator/denominator so that its values lies in the | |
616 | // range 1-10. That is after a call to this function we have: | |
617 | // 1 <= (numerator + delta_plus) /denominator < 10. | |
618 | // Let numerator the input before modification and numerator' the argument | |
619 | // after modification, then the output-parameter decimal_point is such that | |
620 | // numerator / denominator * 10^estimated_power == | |
621 | // numerator' / denominator' * 10^(decimal_point - 1) | |
622 | // In some cases estimated_power was too low, and this is already the case. We | |
623 | // then simply adjust the power so that 10^(k-1) <= v < 10^k (with k == | |
624 | // estimated_power) but do not touch the numerator or denominator. | |
625 | // Otherwise the routine multiplies the numerator and the deltas by 10. | |
626 | static void FixupMultiply10(int estimated_power, bool is_even, | |
627 | int* decimal_point, | |
628 | Bignum* numerator, Bignum* denominator, | |
629 | Bignum* delta_minus, Bignum* delta_plus) { | |
630 | bool in_range; | |
631 | if (is_even) { | |
632 | // For IEEE doubles half-way cases (in decimal system numbers ending with 5) | |
633 | // are rounded to the closest floating-point number with even significand. | |
634 | in_range = Bignum::PlusCompare(*numerator, *delta_plus, *denominator) >= 0; | |
635 | } else { | |
636 | in_range = Bignum::PlusCompare(*numerator, *delta_plus, *denominator) > 0; | |
637 | } | |
638 | if (in_range) { | |
639 | // Since numerator + delta_plus >= denominator we already have | |
640 | // 1 <= numerator/denominator < 10. Simply update the estimated_power. | |
641 | *decimal_point = estimated_power + 1; | |
642 | } else { | |
643 | *decimal_point = estimated_power; | |
644 | numerator->Times10(); | |
645 | if (Bignum::Equal(*delta_minus, *delta_plus)) { | |
646 | delta_minus->Times10(); | |
647 | delta_plus->AssignBignum(*delta_minus); | |
648 | } else { | |
649 | delta_minus->Times10(); | |
650 | delta_plus->Times10(); | |
651 | } | |
652 | } | |
653 | } | |
654 | ||
655 | } // namespace double_conversion | |
656 | ||
657 | // ICU PATCH: Close ICU namespace | |
658 | U_NAMESPACE_END | |
659 | #endif // ICU PATCH: close #if !UCONFIG_NO_FORMATTING |