1/**************************************************************** 2 3The author of this software is David M. Gay. 4 5Copyright (C) 1998, 1999 by Lucent Technologies 6All Rights Reserved 7 8Permission to use, copy, modify, and distribute this software and 9its documentation for any purpose and without fee is hereby 10granted, provided that the above copyright notice appear in all 11copies and that both that the copyright notice and this 12permission notice and warranty disclaimer appear in supporting 13documentation, and that the name of Lucent or any of its entities 14not be used in advertising or publicity pertaining to 15distribution of the software without specific, written prior 16permission. 17 18LUCENT DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, 19INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. 20IN NO EVENT SHALL LUCENT OR ANY OF ITS ENTITIES BE LIABLE FOR ANY 21SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES 22WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER 23IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, 24ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF 25THIS SOFTWARE. 26 27****************************************************************/ 28 29/* Please send bug reports to David M. Gay (dmg at acm dot org, 30 * with " at " changed at "@" and " dot " changed to "."). */ 31 32#include "gdtoaimp.h" 33 34/* dtoa for IEEE arithmetic (dmg): convert double to ASCII string. 35 * 36 * Inspired by "How to Print Floating-Point Numbers Accurately" by 37 * Guy L. Steele, Jr. and Jon L. White [Proc. ACM SIGPLAN '90, pp. 112-126]. 38 * 39 * Modifications: 40 * 1. Rather than iterating, we use a simple numeric overestimate 41 * to determine k = floor(log10(d)). We scale relevant 42 * quantities using O(log2(k)) rather than O(k) multiplications. 43 * 2. For some modes > 2 (corresponding to ecvt and fcvt), we don't 44 * try to generate digits strictly left to right. Instead, we 45 * compute with fewer bits and propagate the carry if necessary 46 * when rounding the final digit up. This is often faster. 47 * 3. Under the assumption that input will be rounded nearest, 48 * mode 0 renders 1e23 as 1e23 rather than 9.999999999999999e22. 49 * That is, we allow equality in stopping tests when the 50 * round-nearest rule will give the same floating-point value 51 * as would satisfaction of the stopping test with strict 52 * inequality. 53 * 4. We remove common factors of powers of 2 from relevant 54 * quantities. 55 * 5. When converting floating-point integers less than 1e16, 56 * we use floating-point arithmetic rather than resorting 57 * to multiple-precision integers. 58 * 6. When asked to produce fewer than 15 digits, we first try 59 * to get by with floating-point arithmetic; we resort to 60 * multiple-precision integer arithmetic only if we cannot 61 * guarantee that the floating-point calculation has given 62 * the correctly rounded result. For k requested digits and 63 * "uniformly" distributed input, the probability is 64 * something like 10^(k-15) that we must resort to the Long 65 * calculation. 66 */ 67 68#ifdef Honor_FLT_ROUNDS 69#undef Check_FLT_ROUNDS 70#define Check_FLT_ROUNDS 71#else 72#define Rounding Flt_Rounds 73#endif 74 75 char * 76dtoa 77#ifdef KR_headers 78 (d0, mode, ndigits, decpt, sign, rve) 79 double d0; int mode, ndigits, *decpt, *sign; char **rve; 80#else 81 (double d0, int mode, int ndigits, int *decpt, int *sign, char **rve) 82#endif 83{ 84 /* Arguments ndigits, decpt, sign are similar to those 85 of ecvt and fcvt; trailing zeros are suppressed from 86 the returned string. If not null, *rve is set to point 87 to the end of the return value. If d is +-Infinity or NaN, 88 then *decpt is set to 9999. 89 90 mode: 91 0 ==> shortest string that yields d when read in 92 and rounded to nearest. 93 1 ==> like 0, but with Steele & White stopping rule; 94 e.g. with IEEE P754 arithmetic , mode 0 gives 95 1e23 whereas mode 1 gives 9.999999999999999e22. 96 2 ==> max(1,ndigits) significant digits. This gives a 97 return value similar to that of ecvt, except 98 that trailing zeros are suppressed. 99 3 ==> through ndigits past the decimal point. This 100 gives a return value similar to that from fcvt, 101 except that trailing zeros are suppressed, and 102 ndigits can be negative. 103 4,5 ==> similar to 2 and 3, respectively, but (in 104 round-nearest mode) with the tests of mode 0 to 105 possibly return a shorter string that rounds to d. 106 With IEEE arithmetic and compilation with 107 -DHonor_FLT_ROUNDS, modes 4 and 5 behave the same 108 as modes 2 and 3 when FLT_ROUNDS != 1. 109 6-9 ==> Debugging modes similar to mode - 4: don't try 110 fast floating-point estimate (if applicable). 111 112 Values of mode other than 0-9 are treated as mode 0. 113 114 Sufficient space is allocated to the return value 115 to hold the suppressed trailing zeros. 116 */ 117 118 int bbits, b2, b5, be, dig, i, ieps, ilim, ilim0, ilim1, 119 j, j1, k, k0, k_check, leftright, m2, m5, s2, s5, 120 spec_case, try_quick; 121 Long L; 122#ifndef Sudden_Underflow 123 int denorm; 124 ULong x; 125#endif 126 Bigint *b, *b1, *delta, *mlo, *mhi, *S; 127 U d, d2, eps; 128 double ds; 129 char *s, *s0; 130#ifdef SET_INEXACT 131 int inexact, oldinexact; 132#endif 133#ifdef Honor_FLT_ROUNDS /*{*/ 134 int Rounding; 135#ifdef Trust_FLT_ROUNDS /*{{ only define this if FLT_ROUNDS really works! */ 136 Rounding = Flt_Rounds; 137#else /*}{*/ 138 Rounding = 1; 139 switch(fegetround()) { 140 case FE_TOWARDZERO: Rounding = 0; break; 141 case FE_UPWARD: Rounding = 2; break; 142 case FE_DOWNWARD: Rounding = 3; 143 } 144#endif /*}}*/ 145#endif /*}*/ 146 147#ifndef MULTIPLE_THREADS 148 if (dtoa_result) { 149 freedtoa(dtoa_result); 150 dtoa_result = 0; 151 } 152#endif 153 d.d = d0; 154 if (word0(&d) & Sign_bit) { 155 /* set sign for everything, including 0's and NaNs */ 156 *sign = 1; 157 word0(&d) &= ~Sign_bit; /* clear sign bit */ 158 } 159 else 160 *sign = 0; 161 162#if defined(IEEE_Arith) + defined(VAX) 163#ifdef IEEE_Arith 164 if ((word0(&d) & Exp_mask) == Exp_mask) 165#else 166 if (word0(&d) == 0x8000) 167#endif 168 { 169 /* Infinity or NaN */ 170 *decpt = 9999; 171#ifdef IEEE_Arith 172 if (!word1(&d) && !(word0(&d) & 0xfffff)) 173 return nrv_alloc("Infinity", rve, 8); 174#endif 175 return nrv_alloc("NaN", rve, 3); 176 } 177#endif 178#ifdef IBM 179 dval(&d) += 0; /* normalize */ 180#endif 181 if (!dval(&d)) { 182 *decpt = 1; 183 return nrv_alloc("0", rve, 1); 184 } 185 186#ifdef SET_INEXACT 187 try_quick = oldinexact = get_inexact(); 188 inexact = 1; 189#endif 190#ifdef Honor_FLT_ROUNDS 191 if (Rounding >= 2) { 192 if (*sign) 193 Rounding = Rounding == 2 ? 0 : 2; 194 else 195 if (Rounding != 2) 196 Rounding = 0; 197 } 198#endif 199 200 b = d2b(dval(&d), &be, &bbits); 201#ifdef Sudden_Underflow 202 i = (int)(word0(&d) >> Exp_shift1 & (Exp_mask>>Exp_shift1)); 203#else 204 if (( i = (int)(word0(&d) >> Exp_shift1 & (Exp_mask>>Exp_shift1)) )!=0) { 205#endif 206 dval(&d2) = dval(&d); 207 word0(&d2) &= Frac_mask1; 208 word0(&d2) |= Exp_11; 209#ifdef IBM 210 if (( j = 11 - hi0bits(word0(&d2) & Frac_mask) )!=0) 211 dval(&d2) /= 1 << j; 212#endif 213 214 /* log(x) ~=~ log(1.5) + (x-1.5)/1.5 215 * log10(x) = log(x) / log(10) 216 * ~=~ log(1.5)/log(10) + (x-1.5)/(1.5*log(10)) 217 * log10(&d) = (i-Bias)*log(2)/log(10) + log10(&d2) 218 * 219 * This suggests computing an approximation k to log10(&d) by 220 * 221 * k = (i - Bias)*0.301029995663981 222 * + ( (d2-1.5)*0.289529654602168 + 0.176091259055681 ); 223 * 224 * We want k to be too large rather than too small. 225 * The error in the first-order Taylor series approximation 226 * is in our favor, so we just round up the constant enough 227 * to compensate for any error in the multiplication of 228 * (i - Bias) by 0.301029995663981; since |i - Bias| <= 1077, 229 * and 1077 * 0.30103 * 2^-52 ~=~ 7.2e-14, 230 * adding 1e-13 to the constant term more than suffices. 231 * Hence we adjust the constant term to 0.1760912590558. 232 * (We could get a more accurate k by invoking log10, 233 * but this is probably not worthwhile.) 234 */ 235 236 i -= Bias; 237#ifdef IBM 238 i <<= 2; 239 i += j; 240#endif 241#ifndef Sudden_Underflow 242 denorm = 0; 243 } 244 else { 245 /* d is denormalized */ 246 247 i = bbits + be + (Bias + (P-1) - 1); 248 x = i > 32 ? word0(&d) << (64 - i) | word1(&d) >> (i - 32) 249 : word1(&d) << (32 - i); 250 dval(&d2) = x; 251 word0(&d2) -= 31*Exp_msk1; /* adjust exponent */ 252 i -= (Bias + (P-1) - 1) + 1; 253 denorm = 1; 254 } 255#endif 256 ds = (dval(&d2)-1.5)*0.289529654602168 + 0.1760912590558 + i*0.301029995663981; 257 k = (int)ds; 258 if (ds < 0. && ds != k) 259 k--; /* want k = floor(ds) */ 260 k_check = 1; 261 if (k >= 0 && k <= Ten_pmax) { 262 if (dval(&d) < tens[k]) 263 k--; 264 k_check = 0; 265 } 266 j = bbits - i - 1; 267 if (j >= 0) { 268 b2 = 0; 269 s2 = j; 270 } 271 else { 272 b2 = -j; 273 s2 = 0; 274 } 275 if (k >= 0) { 276 b5 = 0; 277 s5 = k; 278 s2 += k; 279 } 280 else { 281 b2 -= k; 282 b5 = -k; 283 s5 = 0; 284 } 285 if (mode < 0 || mode > 9) 286 mode = 0; 287 288#ifndef SET_INEXACT 289#ifdef Check_FLT_ROUNDS 290 try_quick = Rounding == 1; 291#else 292 try_quick = 1; 293#endif 294#endif /*SET_INEXACT*/ 295 296 if (mode > 5) { 297 mode -= 4; 298 try_quick = 0; 299 } 300 leftright = 1; 301 ilim = ilim1 = -1; /* Values for cases 0 and 1; done here to */ 302 /* silence erroneous "gcc -Wall" warning. */ 303 switch(mode) { 304 case 0: 305 case 1: 306 i = 18; 307 ndigits = 0; 308 break; 309 case 2: 310 leftright = 0; 311 /* no break */ 312 case 4: 313 if (ndigits <= 0) 314 ndigits = 1; 315 ilim = ilim1 = i = ndigits; 316 break; 317 case 3: 318 leftright = 0; 319 /* no break */ 320 case 5: 321 i = ndigits + k + 1; 322 ilim = i; 323 ilim1 = i - 1; 324 if (i <= 0) 325 i = 1; 326 } 327 s = s0 = rv_alloc(i); 328 329#ifdef Honor_FLT_ROUNDS 330 if (mode > 1 && Rounding != 1) 331 leftright = 0; 332#endif 333 334 if (ilim >= 0 && ilim <= Quick_max && try_quick) { 335 336 /* Try to get by with floating-point arithmetic. */ 337 338 i = 0; 339 dval(&d2) = dval(&d); 340 k0 = k; 341 ilim0 = ilim; 342 ieps = 2; /* conservative */ 343 if (k > 0) { 344 ds = tens[k&0xf]; 345 j = k >> 4; 346 if (j & Bletch) { 347 /* prevent overflows */ 348 j &= Bletch - 1; 349 dval(&d) /= bigtens[n_bigtens-1]; 350 ieps++; 351 } 352 for(; j; j >>= 1, i++) 353 if (j & 1) { 354 ieps++; 355 ds *= bigtens[i]; 356 } 357 dval(&d) /= ds; 358 } 359 else if (( j1 = -k )!=0) { 360 dval(&d) *= tens[j1 & 0xf]; 361 for(j = j1 >> 4; j; j >>= 1, i++) 362 if (j & 1) { 363 ieps++; 364 dval(&d) *= bigtens[i]; 365 } 366 } 367 if (k_check && dval(&d) < 1. && ilim > 0) { 368 if (ilim1 <= 0) 369 goto fast_failed; 370 ilim = ilim1; 371 k--; 372 dval(&d) *= 10.; 373 ieps++; 374 } 375 dval(&eps) = ieps*dval(&d) + 7.; 376 word0(&eps) -= (P-1)*Exp_msk1; 377 if (ilim == 0) { 378 S = mhi = 0; 379 dval(&d) -= 5.; 380 if (dval(&d) > dval(&eps)) 381 goto one_digit; 382 if (dval(&d) < -dval(&eps)) 383 goto no_digits; 384 goto fast_failed; 385 } 386#ifndef No_leftright 387 if (leftright) { 388 /* Use Steele & White method of only 389 * generating digits needed. 390 */ 391 dval(&eps) = 0.5/tens[ilim-1] - dval(&eps); 392 for(i = 0;;) { 393 L = dval(&d); 394 dval(&d) -= L; 395 *s++ = '0' + (int)L; 396 if (dval(&d) < dval(&eps)) 397 goto ret1; 398 if (1. - dval(&d) < dval(&eps)) 399 goto bump_up; 400 if (++i >= ilim) 401 break; 402 dval(&eps) *= 10.; 403 dval(&d) *= 10.; 404 } 405 } 406 else { 407#endif 408 /* Generate ilim digits, then fix them up. */ 409 dval(&eps) *= tens[ilim-1]; 410 for(i = 1;; i++, dval(&d) *= 10.) { 411 L = (Long)(dval(&d)); 412 if (!(dval(&d) -= L)) 413 ilim = i; 414 *s++ = '0' + (int)L; 415 if (i == ilim) { 416 if (dval(&d) > 0.5 + dval(&eps)) 417 goto bump_up; 418 else if (dval(&d) < 0.5 - dval(&eps)) { 419 while(*--s == '0'); 420 s++; 421 goto ret1; 422 } 423 break; 424 } 425 } 426#ifndef No_leftright 427 } 428#endif 429 fast_failed: 430 s = s0; 431 dval(&d) = dval(&d2); 432 k = k0; 433 ilim = ilim0; 434 } 435 436 /* Do we have a "small" integer? */ 437 438 if (be >= 0 && k <= Int_max) { 439 /* Yes. */ 440 ds = tens[k]; 441 if (ndigits < 0 && ilim <= 0) { 442 S = mhi = 0; 443 if (ilim < 0 || dval(&d) <= 5*ds) 444 goto no_digits; 445 goto one_digit; 446 } 447 for(i = 1;; i++, dval(&d) *= 10.) { 448 L = (Long)(dval(&d) / ds); 449 dval(&d) -= L*ds; 450#ifdef Check_FLT_ROUNDS 451 /* If FLT_ROUNDS == 2, L will usually be high by 1 */ 452 if (dval(&d) < 0) { 453 L--; 454 dval(&d) += ds; 455 } 456#endif 457 *s++ = '0' + (int)L; 458 if (!dval(&d)) { 459#ifdef SET_INEXACT 460 inexact = 0; 461#endif 462 break; 463 } 464 if (i == ilim) { 465#ifdef Honor_FLT_ROUNDS 466 if (mode > 1) 467 switch(Rounding) { 468 case 0: goto ret1; 469 case 2: goto bump_up; 470 } 471#endif 472 dval(&d) += dval(&d); 473#ifdef ROUND_BIASED 474 if (dval(&d) >= ds) 475#else 476 if (dval(&d) > ds || (dval(&d) == ds && L & 1)) 477#endif 478 { 479 bump_up: 480 while(*--s == '9') 481 if (s == s0) { 482 k++; 483 *s = '0'; 484 break; 485 } 486 ++*s++; 487 } 488 break; 489 } 490 } 491 goto ret1; 492 } 493 494 m2 = b2; 495 m5 = b5; 496 mhi = mlo = 0; 497 if (leftright) { 498 i = 499#ifndef Sudden_Underflow 500 denorm ? be + (Bias + (P-1) - 1 + 1) : 501#endif 502#ifdef IBM 503 1 + 4*P - 3 - bbits + ((bbits + be - 1) & 3); 504#else 505 1 + P - bbits; 506#endif 507 b2 += i; 508 s2 += i; 509 mhi = i2b(1); 510 } 511 if (m2 > 0 && s2 > 0) { 512 i = m2 < s2 ? m2 : s2; 513 b2 -= i; 514 m2 -= i; 515 s2 -= i; 516 } 517 if (b5 > 0) { 518 if (leftright) { 519 if (m5 > 0) { 520 mhi = pow5mult(mhi, m5); 521 b1 = mult(mhi, b); 522 Bfree(b); 523 b = b1; 524 } 525 if (( j = b5 - m5 )!=0) 526 b = pow5mult(b, j); 527 } 528 else 529 b = pow5mult(b, b5); 530 } 531 S = i2b(1); 532 if (s5 > 0) 533 S = pow5mult(S, s5); 534 535 /* Check for special case that d is a normalized power of 2. */ 536 537 spec_case = 0; 538 if ((mode < 2 || leftright) 539#ifdef Honor_FLT_ROUNDS 540 && Rounding == 1 541#endif 542 ) { 543 if (!word1(&d) && !(word0(&d) & Bndry_mask) 544#ifndef Sudden_Underflow 545 && word0(&d) & (Exp_mask & ~Exp_msk1) 546#endif 547 ) { 548 /* The special case */ 549 b2 += Log2P; 550 s2 += Log2P; 551 spec_case = 1; 552 } 553 } 554 555 /* Arrange for convenient computation of quotients: 556 * shift left if necessary so divisor has 4 leading 0 bits. 557 * 558 * Perhaps we should just compute leading 28 bits of S once 559 * and for all and pass them and a shift to quorem, so it 560 * can do shifts and ors to compute the numerator for q. 561 */ 562#ifdef Pack_32 563 if (( i = ((s5 ? 32 - hi0bits(S->x[S->wds-1]) : 1) + s2) & 0x1f )!=0) 564 i = 32 - i; 565#else 566 if (( i = ((s5 ? 32 - hi0bits(S->x[S->wds-1]) : 1) + s2) & 0xf )!=0) 567 i = 16 - i; 568#endif 569 if (i > 4) { 570 i -= 4; 571 b2 += i; 572 m2 += i; 573 s2 += i; 574 } 575 else if (i < 4) { 576 i += 28; 577 b2 += i; 578 m2 += i; 579 s2 += i; 580 } 581 if (b2 > 0) 582 b = lshift(b, b2); 583 if (s2 > 0) 584 S = lshift(S, s2); 585 if (k_check) { 586 if (cmp(b,S) < 0) { 587 k--; 588 b = multadd(b, 10, 0); /* we botched the k estimate */ 589 if (leftright) 590 mhi = multadd(mhi, 10, 0); 591 ilim = ilim1; 592 } 593 } 594 if (ilim <= 0 && (mode == 3 || mode == 5)) { 595 if (ilim < 0 || cmp(b,S = multadd(S,5,0)) <= 0) { 596 /* no digits, fcvt style */ 597 no_digits: 598 k = -1 - ndigits; 599 goto ret; 600 } 601 one_digit: 602 *s++ = '1'; 603 k++; 604 goto ret; 605 } 606 if (leftright) { 607 if (m2 > 0) 608 mhi = lshift(mhi, m2); 609 610 /* Compute mlo -- check for special case 611 * that d is a normalized power of 2. 612 */ 613 614 mlo = mhi; 615 if (spec_case) { 616 mhi = Balloc(mhi->k); 617 Bcopy(mhi, mlo); 618 mhi = lshift(mhi, Log2P); 619 } 620 621 for(i = 1;;i++) { 622 dig = quorem(b,S) + '0'; 623 /* Do we yet have the shortest decimal string 624 * that will round to d? 625 */ 626 j = cmp(b, mlo); 627 delta = diff(S, mhi); 628 j1 = delta->sign ? 1 : cmp(b, delta); 629 Bfree(delta); 630#ifndef ROUND_BIASED 631 if (j1 == 0 && mode != 1 && !(word1(&d) & 1) 632#ifdef Honor_FLT_ROUNDS 633 && Rounding >= 1 634#endif 635 ) { 636 if (dig == '9') 637 goto round_9_up; 638 if (j > 0) 639 dig++; 640#ifdef SET_INEXACT 641 else if (!b->x[0] && b->wds <= 1) 642 inexact = 0; 643#endif 644 *s++ = dig; 645 goto ret; 646 } 647#endif 648 if (j < 0 || (j == 0 && mode != 1 649#ifndef ROUND_BIASED 650 && !(word1(&d) & 1) 651#endif 652 )) { 653 if (!b->x[0] && b->wds <= 1) { 654#ifdef SET_INEXACT 655 inexact = 0; 656#endif 657 goto accept_dig; 658 } 659#ifdef Honor_FLT_ROUNDS 660 if (mode > 1) 661 switch(Rounding) { 662 case 0: goto accept_dig; 663 case 2: goto keep_dig; 664 } 665#endif /*Honor_FLT_ROUNDS*/ 666 if (j1 > 0) { 667 b = lshift(b, 1); 668 j1 = cmp(b, S); 669#ifdef ROUND_BIASED 670 if (j1 >= 0 /*)*/ 671#else 672 if ((j1 > 0 || (j1 == 0 && dig & 1)) 673#endif 674 && dig++ == '9') 675 goto round_9_up; 676 } 677 accept_dig: 678 *s++ = dig; 679 goto ret; 680 } 681 if (j1 > 0) { 682#ifdef Honor_FLT_ROUNDS 683 if (!Rounding) 684 goto accept_dig; 685#endif 686 if (dig == '9') { /* possible if i == 1 */ 687 round_9_up: 688 *s++ = '9'; 689 goto roundoff; 690 } 691 *s++ = dig + 1; 692 goto ret; 693 } 694#ifdef Honor_FLT_ROUNDS 695 keep_dig: 696#endif 697 *s++ = dig; 698 if (i == ilim) 699 break; 700 b = multadd(b, 10, 0); 701 if (mlo == mhi) 702 mlo = mhi = multadd(mhi, 10, 0); 703 else { 704 mlo = multadd(mlo, 10, 0); 705 mhi = multadd(mhi, 10, 0); 706 } 707 } 708 } 709 else 710 for(i = 1;; i++) { 711 *s++ = dig = quorem(b,S) + '0'; 712 if (!b->x[0] && b->wds <= 1) { 713#ifdef SET_INEXACT 714 inexact = 0; 715#endif 716 goto ret; 717 } 718 if (i >= ilim) 719 break; 720 b = multadd(b, 10, 0); 721 } 722 723 /* Round off last digit */ 724 725#ifdef Honor_FLT_ROUNDS 726 switch(Rounding) { 727 case 0: goto trimzeros; 728 case 2: goto roundoff; 729 } 730#endif 731 b = lshift(b, 1); 732 j = cmp(b, S); 733#ifdef ROUND_BIASED 734 if (j >= 0) 735#else 736 if (j > 0 || (j == 0 && dig & 1)) 737#endif 738 { 739 roundoff: 740 while(*--s == '9') 741 if (s == s0) { 742 k++; 743 *s++ = '1'; 744 goto ret; 745 } 746 ++*s++; 747 } 748 else { 749#ifdef Honor_FLT_ROUNDS 750 trimzeros: 751#endif 752 while(*--s == '0'); 753 s++; 754 } 755 ret: 756 Bfree(S); 757 if (mhi) { 758 if (mlo && mlo != mhi) 759 Bfree(mlo); 760 Bfree(mhi); 761 } 762 ret1: 763#ifdef SET_INEXACT 764 if (inexact) { 765 if (!oldinexact) { 766 word0(&d) = Exp_1 + (70 << Exp_shift); 767 word1(&d) = 0; 768 dval(&d) += 1.; 769 } 770 } 771 else if (!oldinexact) 772 clear_inexact(); 773#endif 774 Bfree(b); 775 *s = 0; 776 *decpt = k + 1; 777 if (rve) 778 *rve = s; 779 return s0; 780 } 781