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