LCOV - code coverage report
Current view: top level - src/objects - bigint.cc (source / functions) Hit Total Coverage
Test: app.info Lines: 793 854 92.9 %
Date: 2019-04-19 Functions: 86 92 93.5 %

          Line data    Source code
       1             : // Copyright 2017 the V8 project authors. All rights reserved.
       2             : // Use of this source code is governed by a BSD-style license that can be
       3             : // found in the LICENSE file.
       4             : 
       5             : // Parts of the implementation below:
       6             : 
       7             : // Copyright (c) 2014 the Dart project authors.  Please see the AUTHORS file [1]
       8             : // for details. All rights reserved. Use of this source code is governed by a
       9             : // BSD-style license that can be found in the LICENSE file [2].
      10             : //
      11             : // [1] https://github.com/dart-lang/sdk/blob/master/AUTHORS
      12             : // [2] https://github.com/dart-lang/sdk/blob/master/LICENSE
      13             : 
      14             : // Copyright 2009 The Go Authors. All rights reserved.
      15             : // Use of this source code is governed by a BSD-style
      16             : // license that can be found in the LICENSE file [3].
      17             : //
      18             : // [3] https://golang.org/LICENSE
      19             : 
      20             : #include "src/objects/bigint.h"
      21             : 
      22             : #include "src/conversions.h"
      23             : #include "src/double.h"
      24             : #include "src/heap/factory.h"
      25             : #include "src/heap/heap-write-barrier-inl.h"
      26             : #include "src/isolate-inl.h"
      27             : #include "src/objects-inl.h"
      28             : #include "src/objects/heap-number-inl.h"
      29             : #include "src/objects/instance-type-inl.h"
      30             : #include "src/objects/smi.h"
      31             : 
      32             : namespace v8 {
      33             : namespace internal {
      34             : 
      35             : // The MutableBigInt class is an implementation detail designed to prevent
      36             : // accidental mutation of a BigInt after its construction. Step-by-step
      37             : // construction of a BigInt must happen in terms of MutableBigInt, the
      38             : // final result is then passed through MutableBigInt::MakeImmutable and not
      39             : // modified further afterwards.
      40             : // Many of the functions in this class use arguments of type {BigIntBase},
      41             : // indicating that they will be used in a read-only capacity, and both
      42             : // {BigInt} and {MutableBigInt} objects can be passed in.
      43             : class MutableBigInt : public FreshlyAllocatedBigInt {
      44             :  public:
      45             :   // Bottleneck for converting MutableBigInts to BigInts.
      46             :   static MaybeHandle<BigInt> MakeImmutable(MaybeHandle<MutableBigInt> maybe);
      47             :   static Handle<BigInt> MakeImmutable(Handle<MutableBigInt> result);
      48             : 
      49             :   // Allocation helpers.
      50             :   static MaybeHandle<MutableBigInt> New(
      51             :       Isolate* isolate, int length,
      52             :       AllocationType allocation = AllocationType::kYoung);
      53             :   static Handle<BigInt> NewFromInt(Isolate* isolate, int value);
      54             :   static Handle<BigInt> NewFromDouble(Isolate* isolate, double value);
      55             :   void InitializeDigits(int length, byte value = 0);
      56             :   static Handle<MutableBigInt> Copy(Isolate* isolate,
      57             :                                     Handle<BigIntBase> source);
      58        5807 :   static Handle<BigInt> Zero(Isolate* isolate) {
      59             :     // TODO(jkummerow): Consider caching a canonical zero-BigInt.
      60       11614 :     return MakeImmutable(New(isolate, 0)).ToHandleChecked();
      61             :   }
      62             : 
      63             :   static Handle<MutableBigInt> Cast(Handle<FreshlyAllocatedBigInt> bigint) {
      64             :     SLOW_DCHECK(bigint->IsBigInt());
      65             :     return Handle<MutableBigInt>::cast(bigint);
      66             :   }
      67             :   static MutableBigInt unchecked_cast(Object o) {
      68             :     return MutableBigInt(o.ptr());
      69             :   }
      70             : 
      71             :   // Internal helpers.
      72             :   static MaybeHandle<MutableBigInt> BitwiseAnd(Isolate* isolate,
      73             :                                                Handle<BigInt> x,
      74             :                                                Handle<BigInt> y);
      75             :   static MaybeHandle<MutableBigInt> BitwiseXor(Isolate* isolate,
      76             :                                                Handle<BigInt> x,
      77             :                                                Handle<BigInt> y);
      78             :   static MaybeHandle<MutableBigInt> BitwiseOr(Isolate* isolate,
      79             :                                               Handle<BigInt> x,
      80             :                                               Handle<BigInt> y);
      81             : 
      82             :   static Handle<BigInt> TruncateToNBits(Isolate* isolate, int n,
      83             :                                         Handle<BigInt> x);
      84             :   static Handle<BigInt> TruncateAndSubFromPowerOfTwo(Isolate* isolate, int n,
      85             :                                                      Handle<BigInt> x,
      86             :                                                      bool result_sign);
      87             : 
      88             :   static MaybeHandle<BigInt> AbsoluteAdd(Isolate* isolate, Handle<BigInt> x,
      89             :                                          Handle<BigInt> y, bool result_sign);
      90             :   static Handle<BigInt> AbsoluteSub(Isolate* isolate, Handle<BigInt> x,
      91             :                                     Handle<BigInt> y, bool result_sign);
      92             :   static MaybeHandle<MutableBigInt> AbsoluteAddOne(
      93             :       Isolate* isolate, Handle<BigIntBase> x, bool sign,
      94             :       MutableBigInt result_storage = MutableBigInt());
      95             :   static Handle<MutableBigInt> AbsoluteSubOne(Isolate* isolate,
      96             :                                               Handle<BigIntBase> x);
      97             :   static MaybeHandle<MutableBigInt> AbsoluteSubOne(Isolate* isolate,
      98             :                                                    Handle<BigIntBase> x,
      99             :                                                    int result_length);
     100             : 
     101             :   enum ExtraDigitsHandling { kCopy, kSkip };
     102             :   enum SymmetricOp { kSymmetric, kNotSymmetric };
     103             :   static inline Handle<MutableBigInt> AbsoluteBitwiseOp(
     104             :       Isolate* isolate, Handle<BigIntBase> x, Handle<BigIntBase> y,
     105             :       MutableBigInt result_storage, ExtraDigitsHandling extra_digits,
     106             :       SymmetricOp symmetric,
     107             :       const std::function<digit_t(digit_t, digit_t)>& op);
     108             :   static Handle<MutableBigInt> AbsoluteAnd(
     109             :       Isolate* isolate, Handle<BigIntBase> x, Handle<BigIntBase> y,
     110             :       MutableBigInt result_storage = MutableBigInt());
     111             :   static Handle<MutableBigInt> AbsoluteAndNot(
     112             :       Isolate* isolate, Handle<BigIntBase> x, Handle<BigIntBase> y,
     113             :       MutableBigInt result_storage = MutableBigInt());
     114             :   static Handle<MutableBigInt> AbsoluteOr(
     115             :       Isolate* isolate, Handle<BigIntBase> x, Handle<BigIntBase> y,
     116             :       MutableBigInt result_storage = MutableBigInt());
     117             :   static Handle<MutableBigInt> AbsoluteXor(
     118             :       Isolate* isolate, Handle<BigIntBase> x, Handle<BigIntBase> y,
     119             :       MutableBigInt result_storage = MutableBigInt());
     120             : 
     121             :   static int AbsoluteCompare(Handle<BigIntBase> x, Handle<BigIntBase> y);
     122             : 
     123             :   static void MultiplyAccumulate(Handle<BigIntBase> multiplicand,
     124             :                                  digit_t multiplier,
     125             :                                  Handle<MutableBigInt> accumulator,
     126             :                                  int accumulator_index);
     127             :   static void InternalMultiplyAdd(BigIntBase source, digit_t factor,
     128             :                                   digit_t summand, int n, MutableBigInt result);
     129             :   void InplaceMultiplyAdd(uintptr_t factor, uintptr_t summand);
     130             : 
     131             :   // Specialized helpers for Divide/Remainder.
     132             :   static void AbsoluteDivSmall(Isolate* isolate, Handle<BigIntBase> x,
     133             :                                digit_t divisor, Handle<MutableBigInt>* quotient,
     134             :                                digit_t* remainder);
     135             :   static bool AbsoluteDivLarge(Isolate* isolate, Handle<BigIntBase> dividend,
     136             :                                Handle<BigIntBase> divisor,
     137             :                                Handle<MutableBigInt>* quotient,
     138             :                                Handle<MutableBigInt>* remainder);
     139             :   static bool ProductGreaterThan(digit_t factor1, digit_t factor2, digit_t high,
     140             :                                  digit_t low);
     141             :   digit_t InplaceAdd(Handle<BigIntBase> summand, int start_index);
     142             :   digit_t InplaceSub(Handle<BigIntBase> subtrahend, int start_index);
     143             :   void InplaceRightShift(int shift);
     144             :   enum SpecialLeftShiftMode {
     145             :     kSameSizeResult,
     146             :     kAlwaysAddOneDigit,
     147             :   };
     148             :   static MaybeHandle<MutableBigInt> SpecialLeftShift(Isolate* isolate,
     149             :                                                      Handle<BigIntBase> x,
     150             :                                                      int shift,
     151             :                                                      SpecialLeftShiftMode mode);
     152             : 
     153             :   // Specialized helpers for shift operations.
     154             :   static MaybeHandle<BigInt> LeftShiftByAbsolute(Isolate* isolate,
     155             :                                                  Handle<BigIntBase> x,
     156             :                                                  Handle<BigIntBase> y);
     157             :   static Handle<BigInt> RightShiftByAbsolute(Isolate* isolate,
     158             :                                              Handle<BigIntBase> x,
     159             :                                              Handle<BigIntBase> y);
     160             :   static Handle<BigInt> RightShiftByMaximum(Isolate* isolate, bool sign);
     161             :   static Maybe<digit_t> ToShiftAmount(Handle<BigIntBase> x);
     162             : 
     163             :   static MaybeHandle<String> ToStringBasePowerOfTwo(Isolate* isolate,
     164             :                                                     Handle<BigIntBase> x,
     165             :                                                     int radix,
     166             :                                                     ShouldThrow should_throw);
     167             :   static MaybeHandle<String> ToStringGeneric(Isolate* isolate,
     168             :                                              Handle<BigIntBase> x, int radix,
     169             :                                              ShouldThrow should_throw);
     170             : 
     171             :   static double ToDouble(Handle<BigIntBase> x);
     172             :   enum Rounding { kRoundDown, kTie, kRoundUp };
     173             :   static Rounding DecideRounding(Handle<BigIntBase> x, int mantissa_bits_unset,
     174             :                                  int digit_index, uint64_t current_digit);
     175             : 
     176             :   // Returns the least significant 64 bits, simulating two's complement
     177             :   // representation.
     178             :   static uint64_t GetRawBits(BigIntBase x, bool* lossless);
     179             : 
     180             :   // Digit arithmetic helpers.
     181             :   static inline digit_t digit_add(digit_t a, digit_t b, digit_t* carry);
     182             :   static inline digit_t digit_sub(digit_t a, digit_t b, digit_t* borrow);
     183             :   static inline digit_t digit_mul(digit_t a, digit_t b, digit_t* high);
     184             :   static inline digit_t digit_div(digit_t high, digit_t low, digit_t divisor,
     185             :                                   digit_t* remainder);
     186             :   static digit_t digit_pow(digit_t base, digit_t exponent);
     187             :   static inline bool digit_ismax(digit_t x) {
     188             :     return static_cast<digit_t>(~x) == 0;
     189             :   }
     190             : 
     191             : // Internal field setters. Non-mutable BigInts don't have these.
     192             : #include "src/objects/object-macros.h"
     193      109697 :   inline void set_sign(bool new_sign) {
     194      109697 :     int32_t bitfield = RELAXED_READ_INT32_FIELD(*this, kBitfieldOffset);
     195             :     bitfield = SignBits::update(bitfield, new_sign);
     196      109697 :     RELAXED_WRITE_INT32_FIELD(*this, kBitfieldOffset, bitfield);
     197      109697 :   }
     198       85902 :   inline void synchronized_set_length(int new_length) {
     199       85902 :     int32_t bitfield = RELAXED_READ_INT32_FIELD(*this, kBitfieldOffset);
     200             :     bitfield = LengthBits::update(bitfield, new_length);
     201       85902 :     RELEASE_WRITE_INT32_FIELD(*this, kBitfieldOffset, bitfield);
     202       85902 :   }
     203             :   inline void initialize_bitfield(bool sign, int length) {
     204      162727 :     int32_t bitfield = LengthBits::encode(length) | SignBits::encode(sign);
     205      162763 :     WRITE_INT32_FIELD(*this, kBitfieldOffset, bitfield);
     206             :   }
     207             :   inline void set_digit(int n, digit_t value) {
     208             :     SLOW_DCHECK(0 <= n && n < length());
     209   219809600 :     WRITE_UINTPTR_FIELD(*this, kDigitsOffset + n * kDigitSize, value);
     210             :   }
     211             : 
     212             :   void set_64_bits(uint64_t bits);
     213             : 
     214             :   bool IsMutableBigInt() const { return IsBigInt(); }
     215             : 
     216             :   NEVER_READ_ONLY_SPACE
     217             : 
     218             :   OBJECT_CONSTRUCTORS(MutableBigInt, FreshlyAllocatedBigInt);
     219             : };
     220             : 
     221             : OBJECT_CONSTRUCTORS_IMPL(MutableBigInt, FreshlyAllocatedBigInt)
     222             : NEVER_READ_ONLY_SPACE_IMPL(MutableBigInt)
     223             : 
     224             : #include "src/objects/object-macros-undef.h"
     225             : 
     226      118732 : MaybeHandle<MutableBigInt> MutableBigInt::New(Isolate* isolate, int length,
     227             :                                               AllocationType allocation) {
     228      118732 :   if (length > BigInt::kMaxLength) {
     229          18 :     THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntTooBig),
     230             :                     MutableBigInt);
     231             :   }
     232             :   Handle<MutableBigInt> result =
     233      118723 :       Cast(isolate->factory()->NewBigInt(length, allocation));
     234             :   result->initialize_bitfield(false, length);
     235             : #if DEBUG
     236             :   result->InitializeDigits(length, 0xBF);
     237             : #endif
     238      118723 :   return result;
     239             : }
     240             : 
     241       47735 : Handle<BigInt> MutableBigInt::NewFromInt(Isolate* isolate, int value) {
     242       47735 :   if (value == 0) return Zero(isolate);
     243       43491 :   Handle<MutableBigInt> result = Cast(isolate->factory()->NewBigInt(1));
     244       43491 :   bool sign = value < 0;
     245             :   result->initialize_bitfield(sign, 1);
     246       43491 :   if (!sign) {
     247       43061 :     result->set_digit(0, value);
     248             :   } else {
     249         430 :     if (value == kMinInt) {
     250             :       STATIC_ASSERT(kMinInt == -kMaxInt - 1);
     251             :       result->set_digit(0, static_cast<BigInt::digit_t>(kMaxInt) + 1);
     252             :     } else {
     253         430 :       result->set_digit(0, -value);
     254             :     }
     255             :   }
     256       43491 :   return MakeImmutable(result);
     257             : }
     258             : 
     259         441 : Handle<BigInt> MutableBigInt::NewFromDouble(Isolate* isolate, double value) {
     260             :   DCHECK_EQ(value, std::floor(value));
     261         441 :   if (value == 0) return Zero(isolate);
     262             : 
     263         378 :   bool sign = value < 0;  // -0 was already handled above.
     264             :   uint64_t double_bits = bit_cast<uint64_t>(value);
     265             :   int raw_exponent =
     266         378 :       static_cast<int>(double_bits >> Double::kPhysicalSignificandSize) & 0x7FF;
     267             :   DCHECK_NE(raw_exponent, 0x7FF);
     268             :   DCHECK_GE(raw_exponent, 0x3FF);
     269         378 :   int exponent = raw_exponent - 0x3FF;
     270         378 :   int digits = exponent / kDigitBits + 1;
     271         378 :   Handle<MutableBigInt> result = Cast(isolate->factory()->NewBigInt(digits));
     272             :   result->initialize_bitfield(sign, digits);
     273             : 
     274             :   // We construct a BigInt from the double {value} by shifting its mantissa
     275             :   // according to its exponent and mapping the bit pattern onto digits.
     276             :   //
     277             :   //               <----------- bitlength = exponent + 1 ----------->
     278             :   //                <----- 52 ------> <------ trailing zeroes ------>
     279             :   // mantissa:     1yyyyyyyyyyyyyyyyy 0000000000000000000000000000000
     280             :   // digits:    0001xxxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx
     281             :   //                <-->          <------>
     282             :   //          msd_topbit         kDigitBits
     283             :   //
     284             :   uint64_t mantissa =
     285         378 :       (double_bits & Double::kSignificandMask) | Double::kHiddenBit;
     286             :   const int kMantissaTopBit = Double::kSignificandSize - 1;  // 0-indexed.
     287             :   // 0-indexed position of most significant bit in the most significant digit.
     288         378 :   int msd_topbit = exponent % kDigitBits;
     289             :   // Number of unused bits in {mantissa}. We'll keep them shifted to the
     290             :   // left (i.e. most significant part) of the underlying uint64_t.
     291             :   int remaining_mantissa_bits = 0;
     292             :   // Next digit under construction.
     293             :   digit_t digit;
     294             : 
     295             :   // First, build the MSD by shifting the mantissa appropriately.
     296         378 :   if (msd_topbit < kMantissaTopBit) {
     297          81 :     remaining_mantissa_bits = kMantissaTopBit - msd_topbit;
     298          81 :     digit = mantissa >> remaining_mantissa_bits;
     299          81 :     mantissa = mantissa << (64 - remaining_mantissa_bits);
     300             :   } else {
     301             :     DCHECK_GE(msd_topbit, kMantissaTopBit);
     302         297 :     digit = mantissa << (msd_topbit - kMantissaTopBit);
     303             :     mantissa = 0;
     304             :   }
     305             :   result->set_digit(digits - 1, digit);
     306             :   // Then fill in the rest of the digits.
     307         540 :   for (int digit_index = digits - 2; digit_index >= 0; digit_index--) {
     308         162 :     if (remaining_mantissa_bits > 0) {
     309          27 :       remaining_mantissa_bits -= kDigitBits;
     310             :       if (sizeof(digit) == 4) {
     311             :         digit = mantissa >> 32;
     312             :         mantissa = mantissa << 32;
     313             :       } else {
     314             :         DCHECK_EQ(sizeof(digit), 8);
     315             :         digit = mantissa;
     316             :         mantissa = 0;
     317             :       }
     318             :     } else {
     319             :       digit = 0;
     320             :     }
     321             :     result->set_digit(digit_index, digit);
     322             :   }
     323         378 :   return MakeImmutable(result);
     324             : }
     325             : 
     326        6367 : Handle<MutableBigInt> MutableBigInt::Copy(Isolate* isolate,
     327             :                                           Handle<BigIntBase> source) {
     328             :   int length = source->length();
     329             :   // Allocating a BigInt of the same length as an existing BigInt cannot throw.
     330       12734 :   Handle<MutableBigInt> result = New(isolate, length).ToHandleChecked();
     331       19101 :   memcpy(reinterpret_cast<void*>(result->address() + BigIntBase::kHeaderSize),
     332        6367 :          reinterpret_cast<void*>(source->address() + BigIntBase::kHeaderSize),
     333        6367 :          BigInt::SizeFor(length) - BigIntBase::kHeaderSize);
     334        6367 :   return result;
     335             : }
     336             : 
     337           0 : void MutableBigInt::InitializeDigits(int length, byte value) {
     338       57759 :   memset(reinterpret_cast<void*>(ptr() + kDigitsOffset - kHeapObjectTag), value,
     339       57759 :          length * kDigitSize);
     340           0 : }
     341             : 
     342        7932 : MaybeHandle<BigInt> MutableBigInt::MakeImmutable(
     343             :     MaybeHandle<MutableBigInt> maybe) {
     344             :   Handle<MutableBigInt> result;
     345        7932 :   if (!maybe.ToHandle(&result)) return MaybeHandle<BigInt>();
     346        7932 :   return MakeImmutable(result);
     347             : }
     348             : 
     349      160766 : Handle<BigInt> MutableBigInt::MakeImmutable(Handle<MutableBigInt> result) {
     350             :   // Check if we need to right-trim any leading zero-digits.
     351             :   int old_length = result->length();
     352             :   int new_length = old_length;
     353     1667121 :   while (new_length > 0 && result->digit(new_length - 1) == 0) new_length--;
     354      160766 :   int to_trim = old_length - new_length;
     355      160766 :   if (to_trim != 0) {
     356       85902 :     int size_delta = to_trim * kDigitSize;
     357       85902 :     Address new_end = result->address() + BigInt::SizeFor(new_length);
     358             :     Heap* heap = result->GetHeap();
     359       85902 :     if (!heap->IsLargeObject(*result)) {
     360             :       // We do not create a filler for objects in large object space.
     361             :       // TODO(hpayer): We should shrink the large object page if the size
     362             :       // of the object changed significantly.
     363       85884 :       heap->CreateFillerObjectAt(new_end, size_delta, ClearRecordedSlots::kNo);
     364             :     }
     365       85902 :     result->synchronized_set_length(new_length);
     366             : 
     367             :     // Canonicalize -0n.
     368       85902 :     if (new_length == 0) {
     369         370 :       result->set_sign(false);
     370             :       // TODO(jkummerow): If we cache a canonical 0n, return that here.
     371             :     }
     372             :   }
     373             :   DCHECK_IMPLIES(result->length() > 0,
     374             :                  result->digit(result->length() - 1) != 0);  // MSD is non-zero.
     375      160766 :   return Handle<BigInt>(result.location());
     376             : }
     377             : 
     378        1175 : Handle<BigInt> BigInt::Zero(Isolate* isolate) {
     379        1359 :   return MutableBigInt::Zero(isolate);
     380             : }
     381             : 
     382        6401 : Handle<BigInt> BigInt::UnaryMinus(Isolate* isolate, Handle<BigInt> x) {
     383             :   // Special case: There is no -0n.
     384        6401 :   if (x->is_zero()) {
     385          34 :     return x;
     386             :   }
     387        6367 :   Handle<MutableBigInt> result = MutableBigInt::Copy(isolate, x);
     388       12734 :   result->set_sign(!x->sign());
     389        6367 :   return MutableBigInt::MakeImmutable(result);
     390             : }
     391             : 
     392         342 : MaybeHandle<BigInt> BigInt::BitwiseNot(Isolate* isolate, Handle<BigInt> x) {
     393             :   MaybeHandle<MutableBigInt> result;
     394         342 :   if (x->sign()) {
     395             :     // ~(-x) == ~(~(x-1)) == x-1
     396         162 :     result = MutableBigInt::AbsoluteSubOne(isolate, x, x->length());
     397             :   } else {
     398             :     // ~x == -x-1 == -(x+1)
     399         180 :     result = MutableBigInt::AbsoluteAddOne(isolate, x, true);
     400             :   }
     401         342 :   return MutableBigInt::MakeImmutable(result);
     402             : }
     403             : 
     404         633 : MaybeHandle<BigInt> BigInt::Exponentiate(Isolate* isolate, Handle<BigInt> base,
     405             :                                          Handle<BigInt> exponent) {
     406             :   // 1. If exponent is < 0, throw a RangeError exception.
     407         633 :   if (exponent->sign()) {
     408          18 :     THROW_NEW_ERROR(isolate,
     409             :                     NewRangeError(MessageTemplate::kBigIntNegativeExponent),
     410             :                     BigInt);
     411             :   }
     412             :   // 2. If base is 0n and exponent is 0n, return 1n.
     413         624 :   if (exponent->is_zero()) {
     414          54 :     return MutableBigInt::NewFromInt(isolate, 1);
     415             :   }
     416             :   // 3. Return a BigInt representing the mathematical value of base raised
     417             :   //    to the power exponent.
     418         570 :   if (base->is_zero()) return base;
     419        1095 :   if (base->length() == 1 && base->digit(0) == 1) {
     420             :     // (-1) ** even_number == 1.
     421         117 :     if (base->sign() && (exponent->digit(0) & 1) == 0) {
     422          27 :       return UnaryMinus(isolate, base);
     423             :     }
     424             :     // (-1) ** odd_number == -1; 1 ** anything == 1.
     425          36 :     return base;
     426             :   }
     427             :   // For all bases >= 2, very large exponents would lead to unrepresentable
     428             :   // results.
     429             :   STATIC_ASSERT(kMaxLengthBits < std::numeric_limits<digit_t>::max());
     430         489 :   if (exponent->length() > 1) {
     431           0 :     THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntTooBig),
     432             :                     BigInt);
     433             :   }
     434             :   digit_t exp_value = exponent->digit(0);
     435         489 :   if (exp_value == 1) return base;
     436         453 :   if (exp_value >= kMaxLengthBits) {
     437          54 :     THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntTooBig),
     438             :                     BigInt);
     439             :   }
     440             :   STATIC_ASSERT(kMaxLengthBits <= kMaxInt);
     441         426 :   int n = static_cast<int>(exp_value);
     442         843 :   if (base->length() == 1 && base->digit(0) == 2) {
     443             :     // Fast path for 2^n.
     444         317 :     int needed_digits = 1 + (n / kDigitBits);
     445             :     Handle<MutableBigInt> result;
     446         634 :     if (!MutableBigInt::New(isolate, needed_digits).ToHandle(&result)) {
     447           0 :       return MaybeHandle<BigInt>();
     448             :     }
     449             :     result->InitializeDigits(needed_digits);
     450             :     // All bits are zero. Now set the n-th bit.
     451         317 :     digit_t msd = static_cast<digit_t>(1) << (n % kDigitBits);
     452             :     result->set_digit(needed_digits - 1, msd);
     453             :     // Result is negative for odd powers of -2n.
     454         353 :     if (base->sign()) result->set_sign((n & 1) != 0);
     455         317 :     return MutableBigInt::MakeImmutable(result);
     456             :   }
     457             :   Handle<BigInt> result;
     458             :   Handle<BigInt> running_square = base;
     459             :   // This implicitly sets the result's sign correctly.
     460         109 :   if (n & 1) result = base;
     461         109 :   n >>= 1;
     462         809 :   for (; n != 0; n >>= 1) {
     463             :     MaybeHandle<BigInt> maybe_result =
     464         364 :         Multiply(isolate, running_square, running_square);
     465         364 :     if (!maybe_result.ToHandle(&running_square)) return maybe_result;
     466         350 :     if (n & 1) {
     467         169 :       if (result.is_null()) {
     468             :         result = running_square;
     469             :       } else {
     470         115 :         maybe_result = Multiply(isolate, result, running_square);
     471         115 :         if (!maybe_result.ToHandle(&result)) return maybe_result;
     472             :       }
     473             :     }
     474             :   }
     475          95 :   return result;
     476             : }
     477             : 
     478       46838 : MaybeHandle<BigInt> BigInt::Multiply(Isolate* isolate, Handle<BigInt> x,
     479             :                                      Handle<BigInt> y) {
     480       46838 :   if (x->is_zero()) return x;
     481       46208 :   if (y->is_zero()) return y;
     482       46208 :   int result_length = x->length() + y->length();
     483             :   Handle<MutableBigInt> result;
     484       92416 :   if (!MutableBigInt::New(isolate, result_length).ToHandle(&result)) {
     485           9 :     return MaybeHandle<BigInt>();
     486             :   }
     487             :   result->InitializeDigits(result_length);
     488             :   uintptr_t work_estimate = 0;
     489      434093 :   for (int i = 0; i < x->length(); i++) {
     490      193952 :     MutableBigInt::MultiplyAccumulate(y, x->digit(i), result, i);
     491             : 
     492             :     // Multiplication can take a long time. Check for interrupt requests
     493             :     // every now and then (roughly every 10-20 of milliseconds -- rarely
     494             :     // enough not to create noticeable overhead, frequently enough not to
     495             :     // appear frozen).
     496      193952 :     work_estimate += y->length();
     497      193952 :     if (work_estimate > 5000000) {
     498             :       work_estimate = 0;
     499             :       StackLimitCheck interrupt_check(isolate);
     500          31 :       if (interrupt_check.InterruptRequested() &&
     501           6 :           isolate->stack_guard()->HandleInterrupts()->IsException(isolate)) {
     502           5 :         return MaybeHandle<BigInt>();
     503             :       }
     504             :     }
     505             :   }
     506       92388 :   result->set_sign(x->sign() != y->sign());
     507       46194 :   return MutableBigInt::MakeImmutable(result);
     508             : }
     509             : 
     510         392 : MaybeHandle<BigInt> BigInt::Divide(Isolate* isolate, Handle<BigInt> x,
     511             :                                    Handle<BigInt> y) {
     512             :   // 1. If y is 0n, throw a RangeError exception.
     513         392 :   if (y->is_zero()) {
     514          18 :     THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntDivZero),
     515             :                     BigInt);
     516             :   }
     517             :   // 2. Let quotient be the mathematical value of x divided by y.
     518             :   // 3. Return a BigInt representing quotient rounded towards 0 to the next
     519             :   //    integral value.
     520         383 :   if (MutableBigInt::AbsoluteCompare(x, y) < 0) {
     521          15 :     return Zero(isolate);
     522             :   }
     523             :   Handle<MutableBigInt> quotient;
     524         368 :   bool result_sign = x->sign() != y->sign();
     525         368 :   if (y->length() == 1) {
     526             :     digit_t divisor = y->digit(0);
     527         255 :     if (divisor == 1) {
     528          51 :       return result_sign == x->sign() ? x : UnaryMinus(isolate, x);
     529             :     }
     530             :     digit_t remainder;
     531         204 :     MutableBigInt::AbsoluteDivSmall(isolate, x, divisor, &quotient, &remainder);
     532             :   } else {
     533         113 :     if (!MutableBigInt::AbsoluteDivLarge(isolate, x, y, &quotient, nullptr)) {
     534           5 :       return MaybeHandle<BigInt>();
     535             :     }
     536             :   }
     537         624 :   quotient->set_sign(x->sign() != y->sign());
     538         312 :   return MutableBigInt::MakeImmutable(quotient);
     539             : }
     540             : 
     541         387 : MaybeHandle<BigInt> BigInt::Remainder(Isolate* isolate, Handle<BigInt> x,
     542             :                                       Handle<BigInt> y) {
     543             :   // 1. If y is 0n, throw a RangeError exception.
     544         387 :   if (y->is_zero()) {
     545          18 :     THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntDivZero),
     546             :                     BigInt);
     547             :   }
     548             :   // 2. Return the BigInt representing x modulo y.
     549             :   // See https://github.com/tc39/proposal-bigint/issues/84 though.
     550         378 :   if (MutableBigInt::AbsoluteCompare(x, y) < 0) return x;
     551             :   Handle<MutableBigInt> remainder;
     552         345 :   if (y->length() == 1) {
     553             :     digit_t divisor = y->digit(0);
     554         288 :     if (divisor == 1) return Zero(isolate);
     555             :     digit_t remainder_digit;
     556         186 :     MutableBigInt::AbsoluteDivSmall(isolate, x, divisor, nullptr,
     557         186 :                                     &remainder_digit);
     558         186 :     if (remainder_digit == 0) {
     559         118 :       return Zero(isolate);
     560             :     }
     561         136 :     remainder = MutableBigInt::New(isolate, 1).ToHandleChecked();
     562          68 :     remainder->set_digit(0, remainder_digit);
     563             :   } else {
     564         108 :     if (!MutableBigInt::AbsoluteDivLarge(isolate, x, y, nullptr, &remainder)) {
     565           0 :       return MaybeHandle<BigInt>();
     566             :     }
     567             :   }
     568         352 :   remainder->set_sign(x->sign());
     569         176 :   return MutableBigInt::MakeImmutable(remainder);
     570             : }
     571             : 
     572       46548 : MaybeHandle<BigInt> BigInt::Add(Isolate* isolate, Handle<BigInt> x,
     573             :                                 Handle<BigInt> y) {
     574             :   bool xsign = x->sign();
     575       46548 :   if (xsign == y->sign()) {
     576             :     // x + y == x + y
     577             :     // -x + -y == -(x + y)
     578       46420 :     return MutableBigInt::AbsoluteAdd(isolate, x, y, xsign);
     579             :   }
     580             :   // x + -y == x - y == -(y - x)
     581             :   // -x + y == y - x == -(x - y)
     582         128 :   if (MutableBigInt::AbsoluteCompare(x, y) >= 0) {
     583          82 :     return MutableBigInt::AbsoluteSub(isolate, x, y, xsign);
     584             :   }
     585          46 :   return MutableBigInt::AbsoluteSub(isolate, y, x, !xsign);
     586             : }
     587             : 
     588         810 : MaybeHandle<BigInt> BigInt::Subtract(Isolate* isolate, Handle<BigInt> x,
     589             :                                      Handle<BigInt> y) {
     590             :   bool xsign = x->sign();
     591         810 :   if (xsign != y->sign()) {
     592             :     // x - (-y) == x + y
     593             :     // (-x) - y == -(x + y)
     594         137 :     return MutableBigInt::AbsoluteAdd(isolate, x, y, xsign);
     595             :   }
     596             :   // x - y == -(y - x)
     597             :   // (-x) - (-y) == y - x == -(x - y)
     598         673 :   if (MutableBigInt::AbsoluteCompare(x, y) >= 0) {
     599         560 :     return MutableBigInt::AbsoluteSub(isolate, x, y, xsign);
     600             :   }
     601         113 :   return MutableBigInt::AbsoluteSub(isolate, y, x, !xsign);
     602             : }
     603             : 
     604         423 : MaybeHandle<BigInt> BigInt::LeftShift(Isolate* isolate, Handle<BigInt> x,
     605             :                                       Handle<BigInt> y) {
     606         828 :   if (y->is_zero() || x->is_zero()) return x;
     607         483 :   if (y->sign()) return MutableBigInt::RightShiftByAbsolute(isolate, x, y);
     608         309 :   return MutableBigInt::LeftShiftByAbsolute(isolate, x, y);
     609             : }
     610             : 
     611         342 : MaybeHandle<BigInt> BigInt::SignedRightShift(Isolate* isolate, Handle<BigInt> x,
     612             :                                              Handle<BigInt> y) {
     613         675 :   if (y->is_zero() || x->is_zero()) return x;
     614         447 :   if (y->sign()) return MutableBigInt::LeftShiftByAbsolute(isolate, x, y);
     615         219 :   return MutableBigInt::RightShiftByAbsolute(isolate, x, y);
     616             : }
     617             : 
     618         117 : MaybeHandle<BigInt> BigInt::UnsignedRightShift(Isolate* isolate,
     619             :                                                Handle<BigInt> x,
     620             :                                                Handle<BigInt> y) {
     621         234 :   THROW_NEW_ERROR(isolate, NewTypeError(MessageTemplate::kBigIntShr), BigInt);
     622             : }
     623             : 
     624             : namespace {
     625             : 
     626             : // Produces comparison result for {left_negative} == sign(x) != sign(y).
     627             : ComparisonResult UnequalSign(bool left_negative) {
     628             :   return left_negative ? ComparisonResult::kLessThan
     629         266 :                        : ComparisonResult::kGreaterThan;
     630             : }
     631             : 
     632             : // Produces result for |x| > |y|, with {both_negative} == sign(x) == sign(y);
     633             : ComparisonResult AbsoluteGreater(bool both_negative) {
     634             :   return both_negative ? ComparisonResult::kLessThan
     635         850 :                        : ComparisonResult::kGreaterThan;
     636             : }
     637             : 
     638             : // Produces result for |x| < |y|, with {both_negative} == sign(x) == sign(y).
     639             : ComparisonResult AbsoluteLess(bool both_negative) {
     640             :   return both_negative ? ComparisonResult::kGreaterThan
     641         687 :                        : ComparisonResult::kLessThan;
     642             : }
     643             : 
     644             : }  // namespace
     645             : 
     646             : // (Never returns kUndefined.)
     647        1737 : ComparisonResult BigInt::CompareToBigInt(Handle<BigInt> x, Handle<BigInt> y) {
     648             :   bool x_sign = x->sign();
     649        1943 :   if (x_sign != y->sign()) return UnequalSign(x_sign);
     650             : 
     651        1531 :   int result = MutableBigInt::AbsoluteCompare(x, y);
     652        2154 :   if (result > 0) return AbsoluteGreater(x_sign);
     653        1351 :   if (result < 0) return AbsoluteLess(x_sign);
     654             :   return ComparisonResult::kEqual;
     655             : }
     656             : 
     657       15507 : bool BigInt::EqualToBigInt(BigInt x, BigInt y) {
     658       15507 :   if (x->sign() != y->sign()) return false;
     659       15458 :   if (x->length() != y->length()) return false;
     660       51760 :   for (int i = 0; i < x->length(); i++) {
     661       18683 :     if (x->digit(i) != y->digit(i)) return false;
     662             :   }
     663             :   return true;
     664             : }
     665             : 
     666         369 : MaybeHandle<BigInt> BigInt::BitwiseAnd(Isolate* isolate, Handle<BigInt> x,
     667             :                                        Handle<BigInt> y) {
     668         369 :   return MutableBigInt::MakeImmutable(MutableBigInt::BitwiseAnd(isolate, x, y));
     669             : }
     670             : 
     671         369 : MaybeHandle<MutableBigInt> MutableBigInt::BitwiseAnd(Isolate* isolate,
     672             :                                                      Handle<BigInt> x,
     673             :                                                      Handle<BigInt> y) {
     674         651 :   if (!x->sign() && !y->sign()) {
     675         236 :     return AbsoluteAnd(isolate, x, y);
     676         220 :   } else if (x->sign() && y->sign()) {
     677          41 :     int result_length = Max(x->length(), y->length()) + 1;
     678             :     // (-x) & (-y) == ~(x-1) & ~(y-1) == ~((x-1) | (y-1))
     679             :     // == -(((x-1) | (y-1)) + 1)
     680             :     Handle<MutableBigInt> result;
     681          82 :     if (!AbsoluteSubOne(isolate, x, result_length).ToHandle(&result)) {
     682           0 :       return MaybeHandle<MutableBigInt>();
     683             :     }
     684          41 :     Handle<MutableBigInt> y_1 = AbsoluteSubOne(isolate, y);
     685          41 :     result = AbsoluteOr(isolate, result, y_1, *result);
     686          41 :     return AbsoluteAddOne(isolate, result, true, *result);
     687             :   } else {
     688             :     DCHECK(x->sign() != y->sign());
     689             :     // Assume that x is the positive BigInt.
     690          92 :     if (x->sign()) std::swap(x, y);
     691             :     // x & (-y) == x & ~(y-1) == x &~ (y-1)
     692          92 :     Handle<MutableBigInt> y_1 = AbsoluteSubOne(isolate, y);
     693          92 :     return AbsoluteAndNot(isolate, x, y_1);
     694             :   }
     695             : }
     696             : 
     697         342 : MaybeHandle<BigInt> BigInt::BitwiseXor(Isolate* isolate, Handle<BigInt> x,
     698             :                                        Handle<BigInt> y) {
     699         342 :   return MutableBigInt::MakeImmutable(MutableBigInt::BitwiseXor(isolate, x, y));
     700             : }
     701             : 
     702         342 : MaybeHandle<MutableBigInt> MutableBigInt::BitwiseXor(Isolate* isolate,
     703             :                                                      Handle<BigInt> x,
     704             :                                                      Handle<BigInt> y) {
     705         570 :   if (!x->sign() && !y->sign()) {
     706         164 :     return AbsoluteXor(isolate, x, y);
     707         292 :   } else if (x->sign() && y->sign()) {
     708             :     int result_length = Max(x->length(), y->length());
     709             :     // (-x) ^ (-y) == ~(x-1) ^ ~(y-1) == (x-1) ^ (y-1)
     710             :     Handle<MutableBigInt> result =
     711         136 :         AbsoluteSubOne(isolate, x, result_length).ToHandleChecked();
     712          68 :     Handle<MutableBigInt> y_1 = AbsoluteSubOne(isolate, y);
     713          68 :     return AbsoluteXor(isolate, result, y_1, *result);
     714             :   } else {
     715             :     DCHECK(x->sign() != y->sign());
     716         110 :     int result_length = Max(x->length(), y->length()) + 1;
     717             :     // Assume that x is the positive BigInt.
     718         110 :     if (x->sign()) std::swap(x, y);
     719             :     // x ^ (-y) == x ^ ~(y-1) == ~(x ^ (y-1)) == -((x ^ (y-1)) + 1)
     720             :     Handle<MutableBigInt> result;
     721         220 :     if (!AbsoluteSubOne(isolate, y, result_length).ToHandle(&result)) {
     722           0 :       return MaybeHandle<MutableBigInt>();
     723             :     }
     724         110 :     result = AbsoluteXor(isolate, result, x, *result);
     725         110 :     return AbsoluteAddOne(isolate, result, true, *result);
     726             :   }
     727             : }
     728             : 
     729         342 : MaybeHandle<BigInt> BigInt::BitwiseOr(Isolate* isolate, Handle<BigInt> x,
     730             :                                       Handle<BigInt> y) {
     731         342 :   return MutableBigInt::MakeImmutable(MutableBigInt::BitwiseOr(isolate, x, y));
     732             : }
     733             : 
     734         342 : MaybeHandle<MutableBigInt> MutableBigInt::BitwiseOr(Isolate* isolate,
     735             :                                                     Handle<BigInt> x,
     736             :                                                     Handle<BigInt> y) {
     737             :   int result_length = Max(x->length(), y->length());
     738         570 :   if (!x->sign() && !y->sign()) {
     739         191 :     return AbsoluteOr(isolate, x, y);
     740         265 :   } else if (x->sign() && y->sign()) {
     741             :     // (-x) | (-y) == ~(x-1) | ~(y-1) == ~((x-1) & (y-1))
     742             :     // == -(((x-1) & (y-1)) + 1)
     743             :     Handle<MutableBigInt> result =
     744         118 :         AbsoluteSubOne(isolate, x, result_length).ToHandleChecked();
     745          59 :     Handle<MutableBigInt> y_1 = AbsoluteSubOne(isolate, y);
     746          59 :     result = AbsoluteAnd(isolate, result, y_1, *result);
     747          59 :     return AbsoluteAddOne(isolate, result, true, *result);
     748             :   } else {
     749             :     DCHECK(x->sign() != y->sign());
     750             :     // Assume that x is the positive BigInt.
     751          92 :     if (x->sign()) std::swap(x, y);
     752             :     // x | (-y) == x | ~(y-1) == ~((y-1) &~ x) == -(((y-1) &~ x) + 1)
     753             :     Handle<MutableBigInt> result =
     754         184 :         AbsoluteSubOne(isolate, y, result_length).ToHandleChecked();
     755          92 :     result = AbsoluteAndNot(isolate, result, x, *result);
     756          92 :     return AbsoluteAddOne(isolate, result, true, *result);
     757             :   }
     758             : }
     759             : 
     760         482 : MaybeHandle<BigInt> BigInt::Increment(Isolate* isolate, Handle<BigInt> x) {
     761         482 :   if (x->sign()) {
     762         216 :     Handle<MutableBigInt> result = MutableBigInt::AbsoluteSubOne(isolate, x);
     763         216 :     result->set_sign(true);
     764         216 :     return MutableBigInt::MakeImmutable(result);
     765             :   } else {
     766             :     return MutableBigInt::MakeImmutable(
     767         266 :         MutableBigInt::AbsoluteAddOne(isolate, x, false));
     768             :   }
     769             : }
     770             : 
     771         482 : MaybeHandle<BigInt> BigInt::Decrement(Isolate* isolate, Handle<BigInt> x) {
     772             :   MaybeHandle<MutableBigInt> result;
     773         482 :   if (x->sign()) {
     774         225 :     result = MutableBigInt::AbsoluteAddOne(isolate, x, true);
     775         257 :   } else if (x->is_zero()) {
     776             :     // TODO(jkummerow): Consider caching a canonical -1n BigInt.
     777          18 :     return MutableBigInt::NewFromInt(isolate, -1);
     778             :   } else {
     779         239 :     result = MutableBigInt::AbsoluteSubOne(isolate, x);
     780             :   }
     781         464 :   return MutableBigInt::MakeImmutable(result);
     782             : }
     783             : 
     784         378 : ComparisonResult BigInt::CompareToString(Isolate* isolate, Handle<BigInt> x,
     785             :                                          Handle<String> y) {
     786             :   // a. Let ny be StringToBigInt(y);
     787         378 :   MaybeHandle<BigInt> maybe_ny = StringToBigInt(isolate, y);
     788             :   // b. If ny is NaN, return undefined.
     789             :   Handle<BigInt> ny;
     790         378 :   if (!maybe_ny.ToHandle(&ny)) {
     791             :     DCHECK(!isolate->has_pending_exception());
     792             :     return ComparisonResult::kUndefined;
     793             :   }
     794             :   // c. Return BigInt::lessThan(x, ny).
     795         306 :   return CompareToBigInt(x, ny);
     796             : }
     797             : 
     798         495 : bool BigInt::EqualToString(Isolate* isolate, Handle<BigInt> x,
     799             :                            Handle<String> y) {
     800             :   // a. Let n be StringToBigInt(y).
     801         495 :   MaybeHandle<BigInt> maybe_n = StringToBigInt(isolate, y);
     802             :   // b. If n is NaN, return false.
     803             :   Handle<BigInt> n;
     804         495 :   if (!maybe_n.ToHandle(&n)) {
     805             :     DCHECK(!isolate->has_pending_exception());
     806             :     return false;
     807             :   }
     808             :   // c. Return the result of x == n.
     809         315 :   return EqualToBigInt(*x, *n);
     810             : }
     811             : 
     812        1458 : bool BigInt::EqualToNumber(Handle<BigInt> x, Handle<Object> y) {
     813             :   DCHECK(y->IsNumber());
     814             :   // a. If x or y are any of NaN, +∞, or -∞, return false.
     815             :   // b. If the mathematical value of x is equal to the mathematical value of y,
     816             :   //    return true, otherwise return false.
     817        1458 :   if (y->IsSmi()) {
     818             :     int value = Smi::ToInt(*y);
     819        1278 :     if (value == 0) return x->is_zero();
     820             :     // Any multi-digit BigInt is bigger than a Smi.
     821             :     STATIC_ASSERT(sizeof(digit_t) >= sizeof(value));
     822        2844 :     return (x->length() == 1) && (x->sign() == (value < 0)) &&
     823         828 :            (x->digit(0) ==
     824         828 :             static_cast<digit_t>(std::abs(static_cast<int64_t>(value))));
     825             :   }
     826             :   DCHECK(y->IsHeapNumber());
     827             :   double value = Handle<HeapNumber>::cast(y)->value();
     828         180 :   return CompareToDouble(x, value) == ComparisonResult::kEqual;
     829             : }
     830             : 
     831         900 : ComparisonResult BigInt::CompareToNumber(Handle<BigInt> x, Handle<Object> y) {
     832             :   DCHECK(y->IsNumber());
     833         900 :   if (y->IsSmi()) {
     834             :     bool x_sign = x->sign();
     835             :     int y_value = Smi::ToInt(*y);
     836         594 :     bool y_sign = (y_value < 0);
     837         648 :     if (x_sign != y_sign) return UnequalSign(x_sign);
     838             : 
     839         540 :     if (x->is_zero()) {
     840             :       DCHECK(!y_sign);
     841             :       return y_value == 0 ? ComparisonResult::kEqual
     842          36 :                           : ComparisonResult::kLessThan;
     843             :     }
     844             :     // Any multi-digit BigInt is bigger than a Smi.
     845             :     STATIC_ASSERT(sizeof(digit_t) >= sizeof(y_value));
     846         504 :     if (x->length() > 1) return AbsoluteGreater(x_sign);
     847             : 
     848        1008 :     digit_t abs_value = std::abs(static_cast<int64_t>(y_value));
     849             :     digit_t x_digit = x->digit(0);
     850         666 :     if (x_digit > abs_value) return AbsoluteGreater(x_sign);
     851         504 :     if (x_digit < abs_value) return AbsoluteLess(x_sign);
     852             :     return ComparisonResult::kEqual;
     853             :   }
     854             :   DCHECK(y->IsHeapNumber());
     855             :   double value = Handle<HeapNumber>::cast(y)->value();
     856         306 :   return CompareToDouble(x, value);
     857             : }
     858             : 
     859         525 : ComparisonResult BigInt::CompareToDouble(Handle<BigInt> x, double y) {
     860         525 :   if (std::isnan(y)) return ComparisonResult::kUndefined;
     861         434 :   if (y == V8_INFINITY) return ComparisonResult::kLessThan;
     862         379 :   if (y == -V8_INFINITY) return ComparisonResult::kGreaterThan;
     863             :   bool x_sign = x->sign();
     864             :   // Note that this is different from the double's sign bit for -0. That's
     865             :   // intentional because -0 must be treated like 0.
     866         324 :   bool y_sign = (y < 0);
     867         330 :   if (x_sign != y_sign) return UnequalSign(x_sign);
     868         318 :   if (y == 0) {
     869             :     DCHECK(!x_sign);
     870             :     return x->is_zero() ? ComparisonResult::kEqual
     871         165 :                         : ComparisonResult::kGreaterThan;
     872             :   }
     873         153 :   if (x->is_zero()) {
     874             :     DCHECK(!y_sign);
     875             :     return ComparisonResult::kLessThan;
     876             :   }
     877             :   uint64_t double_bits = bit_cast<uint64_t>(y);
     878             :   int raw_exponent =
     879         151 :       static_cast<int>(double_bits >> Double::kPhysicalSignificandSize) & 0x7FF;
     880         151 :   uint64_t mantissa = double_bits & Double::kSignificandMask;
     881             :   // Non-finite doubles are handled above.
     882             :   DCHECK_NE(raw_exponent, 0x7FF);
     883         151 :   int exponent = raw_exponent - 0x3FF;
     884         151 :   if (exponent < 0) {
     885             :     // The absolute value of the double is less than 1. Only 0n has an
     886             :     // absolute value smaller than that, but we've already covered that case.
     887             :     DCHECK(!x->is_zero());
     888           2 :     return AbsoluteGreater(x_sign);
     889             :   }
     890             :   int x_length = x->length();
     891         149 :   digit_t x_msd = x->digit(x_length - 1);
     892         149 :   int msd_leading_zeros = base::bits::CountLeadingZeros(x_msd);
     893         149 :   int x_bitlength = x_length * kDigitBits - msd_leading_zeros;
     894         149 :   int y_bitlength = exponent + 1;
     895         152 :   if (x_bitlength < y_bitlength) return AbsoluteLess(x_sign);
     896         148 :   if (x_bitlength > y_bitlength) return AbsoluteGreater(x_sign);
     897             : 
     898             :   // At this point, we know that signs and bit lengths (i.e. position of
     899             :   // the most significant bit in exponent-free representation) are identical.
     900             :   // {x} is not zero, {y} is finite and not denormal.
     901             :   // Now we virtually convert the double to an integer by shifting its
     902             :   // mantissa according to its exponent, so it will align with the BigInt {x},
     903             :   // and then we compare them bit for bit until we find a difference or the
     904             :   // least significant bit.
     905             :   //                    <----- 52 ------> <-- virtual trailing zeroes -->
     906             :   // y / mantissa:     1yyyyyyyyyyyyyyyyy 0000000000000000000000000000000
     907             :   // x / digits:    0001xxxx xxxxxxxx xxxxxxxx ...
     908             :   //                    <-->          <------>
     909             :   //              msd_topbit         kDigitBits
     910             :   //
     911         144 :   mantissa |= Double::kHiddenBit;
     912             :   const int kMantissaTopBit = 52;  // 0-indexed.
     913             :   // 0-indexed position of {x}'s most significant bit within the {msd}.
     914         144 :   int msd_topbit = kDigitBits - 1 - msd_leading_zeros;
     915             :   DCHECK_EQ(msd_topbit, (x_bitlength - 1) % kDigitBits);
     916             :   // Shifted chunk of {mantissa} for comparing with {digit}.
     917             :   digit_t compare_mantissa;
     918             :   // Number of unprocessed bits in {mantissa}. We'll keep them shifted to
     919             :   // the left (i.e. most significant part) of the underlying uint64_t.
     920             :   int remaining_mantissa_bits = 0;
     921             : 
     922             :   // First, compare the most significant digit against the beginning of
     923             :   // the mantissa.
     924         144 :   if (msd_topbit < kMantissaTopBit) {
     925         140 :     remaining_mantissa_bits = (kMantissaTopBit - msd_topbit);
     926         140 :     compare_mantissa = mantissa >> remaining_mantissa_bits;
     927         140 :     mantissa = mantissa << (64 - remaining_mantissa_bits);
     928             :   } else {
     929             :     DCHECK_GE(msd_topbit, kMantissaTopBit);
     930           4 :     compare_mantissa = mantissa << (msd_topbit - kMantissaTopBit);
     931             :     mantissa = 0;
     932             :   }
     933         204 :   if (x_msd > compare_mantissa) return AbsoluteGreater(x_sign);
     934          86 :   if (x_msd < compare_mantissa) return AbsoluteLess(x_sign);
     935             : 
     936             :   // Then, compare additional digits against any remaining mantissa bits.
     937          83 :   for (int digit_index = x_length - 2; digit_index >= 0; digit_index--) {
     938           3 :     if (remaining_mantissa_bits > 0) {
     939           3 :       remaining_mantissa_bits -= kDigitBits;
     940             :       if (sizeof(mantissa) != sizeof(x_msd)) {
     941             :         compare_mantissa = mantissa >> (64 - kDigitBits);
     942             :         // "& 63" to appease compilers. kDigitBits is 32 here anyway.
     943             :         mantissa = mantissa << (kDigitBits & 63);
     944             :       } else {
     945             :         compare_mantissa = mantissa;
     946             :         mantissa = 0;
     947             :       }
     948             :     } else {
     949             :       compare_mantissa = 0;
     950             :     }
     951             :     digit_t digit = x->digit(digit_index);
     952           4 :     if (digit > compare_mantissa) return AbsoluteGreater(x_sign);
     953           3 :     if (digit < compare_mantissa) return AbsoluteLess(x_sign);
     954             :   }
     955             : 
     956             :   // Integer parts are equal; check whether {y} has a fractional part.
     957          80 :   if (mantissa != 0) {
     958             :     DCHECK_GT(remaining_mantissa_bits, 0);
     959          76 :     return AbsoluteLess(x_sign);
     960             :   }
     961             :   return ComparisonResult::kEqual;
     962             : }
     963             : 
     964       11759 : MaybeHandle<String> BigInt::ToString(Isolate* isolate, Handle<BigInt> bigint,
     965             :                                      int radix, ShouldThrow should_throw) {
     966       11759 :   if (bigint->is_zero()) {
     967         819 :     return isolate->factory()->NewStringFromStaticChars("0");
     968             :   }
     969       10940 :   if (base::bits::IsPowerOfTwo(radix)) {
     970             :     return MutableBigInt::ToStringBasePowerOfTwo(isolate, bigint, radix,
     971         180 :                                                  should_throw);
     972             :   }
     973       10760 :   return MutableBigInt::ToStringGeneric(isolate, bigint, radix, should_throw);
     974             : }
     975             : 
     976       47946 : MaybeHandle<BigInt> BigInt::FromNumber(Isolate* isolate,
     977             :                                        Handle<Object> number) {
     978             :   DCHECK(number->IsNumber());
     979       47946 :   if (number->IsSmi()) {
     980       47406 :     return MutableBigInt::NewFromInt(isolate, Smi::ToInt(*number));
     981             :   }
     982             :   double value = HeapNumber::cast(*number)->value();
     983         540 :   if (!std::isfinite(value) || (DoubleToInteger(value) != value)) {
     984         198 :     THROW_NEW_ERROR(isolate,
     985             :                     NewRangeError(MessageTemplate::kBigIntFromNumber, number),
     986             :                     BigInt);
     987             :   }
     988         441 :   return MutableBigInt::NewFromDouble(isolate, value);
     989             : }
     990             : 
     991        4757 : MaybeHandle<BigInt> BigInt::FromObject(Isolate* isolate, Handle<Object> obj) {
     992        4757 :   if (obj->IsJSReceiver()) {
     993         770 :     ASSIGN_RETURN_ON_EXCEPTION(
     994             :         isolate, obj,
     995             :         JSReceiver::ToPrimitive(Handle<JSReceiver>::cast(obj),
     996             :                                 ToPrimitiveHint::kNumber),
     997             :         BigInt);
     998             :   }
     999             : 
    1000        4757 :   if (obj->IsBoolean()) {
    1001         198 :     return MutableBigInt::NewFromInt(isolate, obj->BooleanValue(isolate));
    1002             :   }
    1003        4559 :   if (obj->IsBigInt()) {
    1004        3916 :     return Handle<BigInt>::cast(obj);
    1005             :   }
    1006         643 :   if (obj->IsString()) {
    1007             :     Handle<BigInt> n;
    1008         534 :     if (!StringToBigInt(isolate, Handle<String>::cast(obj)).ToHandle(&n)) {
    1009         126 :       THROW_NEW_ERROR(isolate,
    1010             :                       NewSyntaxError(MessageTemplate::kBigIntFromObject, obj),
    1011             :                       BigInt);
    1012             :     }
    1013         204 :     return n;
    1014             :   }
    1015             : 
    1016         752 :   THROW_NEW_ERROR(
    1017             :       isolate, NewTypeError(MessageTemplate::kBigIntFromObject, obj), BigInt);
    1018             : }
    1019             : 
    1020         468 : Handle<Object> BigInt::ToNumber(Isolate* isolate, Handle<BigInt> x) {
    1021         477 :   if (x->is_zero()) return Handle<Smi>(Smi::zero(), isolate);
    1022         819 :   if (x->length() == 1 && x->digit(0) < Smi::kMaxValue) {
    1023         279 :     int value = static_cast<int>(x->digit(0));
    1024         279 :     if (x->sign()) value = -value;
    1025         279 :     return Handle<Smi>(Smi::FromInt(value), isolate);
    1026             :   }
    1027         180 :   double result = MutableBigInt::ToDouble(x);
    1028         180 :   return isolate->factory()->NewHeapNumber(result);
    1029             : }
    1030             : 
    1031         180 : double MutableBigInt::ToDouble(Handle<BigIntBase> x) {
    1032         180 :   if (x->is_zero()) return 0.0;
    1033             :   int x_length = x->length();
    1034         180 :   digit_t x_msd = x->digit(x_length - 1);
    1035         180 :   int msd_leading_zeros = base::bits::CountLeadingZeros(x_msd);
    1036         180 :   int x_bitlength = x_length * kDigitBits - msd_leading_zeros;
    1037         198 :   if (x_bitlength > 1024) return x->sign() ? -V8_INFINITY : V8_INFINITY;
    1038         162 :   uint64_t exponent = x_bitlength - 1;
    1039             :   // We need the most significant bit shifted to the position of a double's
    1040             :   // "hidden bit". We also need to hide that MSB, so we shift it out.
    1041             :   uint64_t current_digit = x_msd;
    1042             :   int digit_index = x_length - 1;
    1043         162 :   int shift = msd_leading_zeros + 1 + (64 - kDigitBits);
    1044             :   DCHECK_LE(1, shift);
    1045             :   DCHECK_LE(shift, 64);
    1046         162 :   uint64_t mantissa = (shift == 64) ? 0 : current_digit << shift;
    1047         162 :   mantissa >>= 12;
    1048         162 :   int mantissa_bits_unset = shift - 12;
    1049             :   // If not all mantissa bits are defined yet, get more digits as needed.
    1050         162 :   if (mantissa_bits_unset >= kDigitBits && digit_index > 0) {
    1051           0 :     digit_index--;
    1052             :     current_digit = static_cast<uint64_t>(x->digit(digit_index));
    1053           0 :     mantissa |= (current_digit << (mantissa_bits_unset - kDigitBits));
    1054             :     mantissa_bits_unset -= kDigitBits;
    1055             :   }
    1056         162 :   if (mantissa_bits_unset > 0 && digit_index > 0) {
    1057             :     DCHECK_LT(mantissa_bits_unset, kDigitBits);
    1058          45 :     digit_index--;
    1059             :     current_digit = static_cast<uint64_t>(x->digit(digit_index));
    1060          45 :     mantissa |= (current_digit >> (kDigitBits - mantissa_bits_unset));
    1061          45 :     mantissa_bits_unset -= kDigitBits;
    1062             :   }
    1063             :   // If there are unconsumed digits left, we may have to round.
    1064             :   Rounding rounding =
    1065         162 :       DecideRounding(x, mantissa_bits_unset, digit_index, current_digit);
    1066         162 :   if (rounding == kRoundUp || (rounding == kTie && (mantissa & 1) == 1)) {
    1067          54 :     mantissa++;
    1068             :     // Incrementing the mantissa can overflow the mantissa bits. In that case
    1069             :     // the new mantissa will be all zero (plus hidden bit).
    1070          54 :     if ((mantissa >> Double::kPhysicalSignificandSize) != 0) {
    1071             :       mantissa = 0;
    1072          36 :       exponent++;
    1073             :       // Incrementing the exponent can overflow too.
    1074          36 :       if (exponent > 1023) {
    1075           9 :         return x->sign() ? -V8_INFINITY : V8_INFINITY;
    1076             :       }
    1077             :     }
    1078             :   }
    1079             :   // Assemble the result.
    1080         153 :   uint64_t sign_bit = x->sign() ? (static_cast<uint64_t>(1) << 63) : 0;
    1081         153 :   exponent = (exponent + 0x3FF) << Double::kPhysicalSignificandSize;
    1082         153 :   uint64_t double_bits = sign_bit | exponent | mantissa;
    1083         153 :   return bit_cast<double>(double_bits);
    1084             : }
    1085             : 
    1086             : // This is its own function to keep control flow sane. The meaning of the
    1087             : // parameters is defined by {ToDouble}'s local variable usage.
    1088         162 : MutableBigInt::Rounding MutableBigInt::DecideRounding(Handle<BigIntBase> x,
    1089             :                                                       int mantissa_bits_unset,
    1090             :                                                       int digit_index,
    1091             :                                                       uint64_t current_digit) {
    1092         162 :   if (mantissa_bits_unset > 0) return kRoundDown;
    1093             :   int top_unconsumed_bit;
    1094         135 :   if (mantissa_bits_unset < 0) {
    1095             :     // There are unconsumed bits in {current_digit}.
    1096         126 :     top_unconsumed_bit = -mantissa_bits_unset - 1;
    1097             :   } else {
    1098             :     DCHECK_EQ(mantissa_bits_unset, 0);
    1099             :     // {current_digit} fit the mantissa exactly; look at the next digit.
    1100           9 :     if (digit_index == 0) return kRoundDown;
    1101           0 :     digit_index--;
    1102             :     current_digit = static_cast<uint64_t>(x->digit(digit_index));
    1103             :     top_unconsumed_bit = kDigitBits - 1;
    1104             :   }
    1105             :   // If the most significant remaining bit is 0, round down.
    1106         126 :   uint64_t bitmask = static_cast<uint64_t>(1) << top_unconsumed_bit;
    1107         126 :   if ((current_digit & bitmask) == 0) {
    1108             :     return kRoundDown;
    1109             :   }
    1110             :   // If any other remaining bit is set, round up.
    1111          72 :   bitmask -= 1;
    1112          72 :   if ((current_digit & bitmask) != 0) return kRoundUp;
    1113         216 :   while (digit_index > 0) {
    1114         162 :     digit_index--;
    1115         162 :     if (x->digit(digit_index) != 0) return kRoundUp;
    1116             :   }
    1117             :   return kTie;
    1118             : }
    1119             : 
    1120           0 : void BigInt::BigIntShortPrint(std::ostream& os) {
    1121           0 :   if (sign()) os << "-";
    1122             :   int len = length();
    1123           0 :   if (len == 0) {
    1124           0 :     os << "0";
    1125           0 :     return;
    1126             :   }
    1127           0 :   if (len > 1) os << "...";
    1128             :   os << digit(0);
    1129             : }
    1130             : 
    1131             : // Internal helpers.
    1132             : 
    1133       47421 : MaybeHandle<BigInt> MutableBigInt::AbsoluteAdd(Isolate* isolate,
    1134             :                                                Handle<BigInt> x,
    1135             :                                                Handle<BigInt> y,
    1136             :                                                bool result_sign) {
    1137       47421 :   if (x->length() < y->length()) return AbsoluteAdd(isolate, y, x, result_sign);
    1138       46557 :   if (x->is_zero()) {
    1139             :     DCHECK(y->is_zero());
    1140           0 :     return x;
    1141             :   }
    1142       46557 :   if (y->is_zero()) {
    1143        4923 :     return result_sign == x->sign() ? x : BigInt::UnaryMinus(isolate, x);
    1144             :   }
    1145             :   Handle<MutableBigInt> result;
    1146       83268 :   if (!New(isolate, x->length() + 1).ToHandle(&result)) {
    1147           0 :     return MaybeHandle<BigInt>();
    1148             :   }
    1149             :   digit_t carry = 0;
    1150             :   int i = 0;
    1151      125406 :   for (; i < y->length(); i++) {
    1152             :     digit_t new_carry = 0;
    1153             :     digit_t sum = digit_add(x->digit(i), y->digit(i), &new_carry);
    1154             :     sum = digit_add(sum, carry, &new_carry);
    1155             :     result->set_digit(i, sum);
    1156             :     carry = new_carry;
    1157             :   }
    1158      194346 :   for (; i < x->length(); i++) {
    1159             :     digit_t new_carry = 0;
    1160             :     digit_t sum = digit_add(x->digit(i), carry, &new_carry);
    1161             :     result->set_digit(i, sum);
    1162             :     carry = new_carry;
    1163             :   }
    1164             :   result->set_digit(i, carry);
    1165       83268 :   result->set_sign(result_sign);
    1166       41634 :   return MakeImmutable(result);
    1167             : }
    1168             : 
    1169         801 : Handle<BigInt> MutableBigInt::AbsoluteSub(Isolate* isolate, Handle<BigInt> x,
    1170             :                                           Handle<BigInt> y, bool result_sign) {
    1171             :   DCHECK(x->length() >= y->length());
    1172             :   SLOW_DCHECK(AbsoluteCompare(x, y) >= 0);
    1173         801 :   if (x->is_zero()) {
    1174             :     DCHECK(y->is_zero());
    1175           0 :     return x;
    1176             :   }
    1177         801 :   if (y->is_zero()) {
    1178           9 :     return result_sign == x->sign() ? x : BigInt::UnaryMinus(isolate, x);
    1179             :   }
    1180        1584 :   Handle<MutableBigInt> result = New(isolate, x->length()).ToHandleChecked();
    1181             :   digit_t borrow = 0;
    1182             :   int i = 0;
    1183     1183266 :   for (; i < y->length(); i++) {
    1184             :     digit_t new_borrow = 0;
    1185             :     digit_t difference = digit_sub(x->digit(i), y->digit(i), &new_borrow);
    1186             :     difference = digit_sub(difference, borrow, &new_borrow);
    1187             :     result->set_digit(i, difference);
    1188             :     borrow = new_borrow;
    1189             :   }
    1190     1181178 :   for (; i < x->length(); i++) {
    1191             :     digit_t new_borrow = 0;
    1192             :     digit_t difference = digit_sub(x->digit(i), borrow, &new_borrow);
    1193             :     result->set_digit(i, difference);
    1194             :     borrow = new_borrow;
    1195             :   }
    1196             :   DCHECK_EQ(0, borrow);
    1197        1584 :   result->set_sign(result_sign);
    1198         792 :   return MakeImmutable(result);
    1199             : }
    1200             : 
    1201             : // Adds 1 to the absolute value of {x} and sets the result's sign to {sign}.
    1202             : // {result_storage} is optional; if present, it will be used to store the
    1203             : // result, otherwise a new BigInt will be allocated for the result.
    1204             : // {result_storage} and {x} may refer to the same BigInt for in-place
    1205             : // modification.
    1206         996 : MaybeHandle<MutableBigInt> MutableBigInt::AbsoluteAddOne(
    1207             :     Isolate* isolate, Handle<BigIntBase> x, bool sign,
    1208             :     MutableBigInt result_storage) {
    1209             :   int input_length = x->length();
    1210             :   // The addition will overflow into a new digit if all existing digits are
    1211             :   // at maximum.
    1212             :   bool will_overflow = true;
    1213         996 :   for (int i = 0; i < input_length; i++) {
    1214         969 :     if (!digit_ismax(x->digit(i))) {
    1215             :       will_overflow = false;
    1216             :       break;
    1217             :     }
    1218             :   }
    1219         996 :   int result_length = input_length + will_overflow;
    1220             :   Handle<MutableBigInt> result(result_storage, isolate);
    1221         996 :   if (result_storage.is_null()) {
    1222        1342 :     if (!New(isolate, result_length).ToHandle(&result)) {
    1223           0 :       return MaybeHandle<MutableBigInt>();
    1224             :     }
    1225             :   } else {
    1226             :     DCHECK(result->length() == result_length);
    1227             :   }
    1228             :   digit_t carry = 1;
    1229        5612 :   for (int i = 0; i < input_length; i++) {
    1230             :     digit_t new_carry = 0;
    1231             :     result->set_digit(i, digit_add(x->digit(i), carry, &new_carry));
    1232             :     carry = new_carry;
    1233             :   }
    1234         996 :   if (result_length > input_length) {
    1235             :     result->set_digit(input_length, carry);
    1236             :   } else {
    1237             :     DCHECK_EQ(carry, 0);
    1238             :   }
    1239        1992 :   result->set_sign(sign);
    1240         996 :   return result;
    1241             : }
    1242             : 
    1243             : // Subtracts 1 from the absolute value of {x}. {x} must not be zero.
    1244         715 : Handle<MutableBigInt> MutableBigInt::AbsoluteSubOne(Isolate* isolate,
    1245             :                                                     Handle<BigIntBase> x) {
    1246             :   DCHECK(!x->is_zero());
    1247             :   // Requesting a result length identical to an existing BigInt's length
    1248             :   // cannot overflow the limit.
    1249        1430 :   return AbsoluteSubOne(isolate, x, x->length()).ToHandleChecked();
    1250             : }
    1251             : 
    1252             : // Like the above, but you can specify that the allocated result should have
    1253             : // length {result_length}, which must be at least as large as {x->length()}.
    1254        1247 : MaybeHandle<MutableBigInt> MutableBigInt::AbsoluteSubOne(Isolate* isolate,
    1255             :                                                          Handle<BigIntBase> x,
    1256             :                                                          int result_length) {
    1257             :   DCHECK(!x->is_zero());
    1258             :   DCHECK(result_length >= x->length());
    1259             :   Handle<MutableBigInt> result;
    1260        2494 :   if (!New(isolate, result_length).ToHandle(&result)) {
    1261           0 :     return MaybeHandle<MutableBigInt>();
    1262             :   }
    1263             :   int length = x->length();
    1264             :   digit_t borrow = 1;
    1265        6567 :   for (int i = 0; i < length; i++) {
    1266             :     digit_t new_borrow = 0;
    1267             :     result->set_digit(i, digit_sub(x->digit(i), borrow, &new_borrow));
    1268             :     borrow = new_borrow;
    1269             :   }
    1270             :   DCHECK_EQ(borrow, 0);
    1271        1945 :   for (int i = length; i < result_length; i++) {
    1272             :     result->set_digit(i, borrow);
    1273             :   }
    1274        1247 :   return result;
    1275             : }
    1276             : 
    1277             : // Helper for Absolute{And,AndNot,Or,Xor}.
    1278             : // Performs the given binary {op} on digit pairs of {x} and {y}; when the
    1279             : // end of the shorter of the two is reached, {extra_digits} configures how
    1280             : // remaining digits in the longer input (if {symmetric} == kSymmetric, in
    1281             : // {x} otherwise) are handled: copied to the result or ignored.
    1282             : // If {result_storage} is non-nullptr, it will be used for the result and
    1283             : // any extra digits in it will be zeroed out, otherwise a new BigInt (with
    1284             : // the same length as the longer input) will be allocated.
    1285             : // {result_storage} may alias {x} or {y} for in-place modification.
    1286             : // Example:
    1287             : //              y:             [ y2 ][ y1 ][ y0 ]
    1288             : //              x:       [ x3 ][ x2 ][ x1 ][ x0 ]
    1289             : //                          |     |     |     |
    1290             : //                      (kCopy)  (op)  (op)  (op)
    1291             : //                          |     |     |     |
    1292             : //                          v     v     v     v
    1293             : // result_storage: [  0 ][ x3 ][ r2 ][ r1 ][ r0 ]
    1294        1053 : inline Handle<MutableBigInt> MutableBigInt::AbsoluteBitwiseOp(
    1295             :     Isolate* isolate, Handle<BigIntBase> x, Handle<BigIntBase> y,
    1296             :     MutableBigInt result_storage, ExtraDigitsHandling extra_digits,
    1297             :     SymmetricOp symmetric, const std::function<digit_t(digit_t, digit_t)>& op) {
    1298             :   int x_length = x->length();
    1299             :   int y_length = y->length();
    1300             :   int num_pairs = y_length;
    1301        1053 :   if (x_length < y_length) {
    1302             :     num_pairs = x_length;
    1303         108 :     if (symmetric == kSymmetric) {
    1304             :       std::swap(x, y);
    1305             :       std::swap(x_length, y_length);
    1306             :     }
    1307             :   }
    1308             :   DCHECK(num_pairs == Min(x_length, y_length));
    1309             :   Handle<MutableBigInt> result(result_storage, isolate);
    1310        1053 :   int result_length = extra_digits == kCopy ? x_length : num_pairs;
    1311        1053 :   if (result_storage.is_null()) {
    1312        1366 :     result = New(isolate, result_length).ToHandleChecked();
    1313             :   } else {
    1314             :     DCHECK(result_storage->length() >= result_length);
    1315             :     result_length = result_storage->length();
    1316             :   }
    1317             :   int i = 0;
    1318        4725 :   for (; i < num_pairs; i++) {
    1319             :     result->set_digit(i, op(x->digit(i), y->digit(i)));
    1320             :   }
    1321        1053 :   if (extra_digits == kCopy) {
    1322        1762 :     for (; i < x_length; i++) {
    1323             :       result->set_digit(i, x->digit(i));
    1324             :     }
    1325             :   }
    1326        1053 :   for (; i < result_length; i++) {
    1327             :     result->set_digit(i, 0);
    1328             :   }
    1329        1053 :   return result;
    1330             : }
    1331             : 
    1332             : // If {result_storage} is non-nullptr, it will be used for the result,
    1333             : // otherwise a new BigInt of appropriate length will be allocated.
    1334             : // {result_storage} may alias {x} or {y} for in-place modification.
    1335         295 : Handle<MutableBigInt> MutableBigInt::AbsoluteAnd(Isolate* isolate,
    1336             :                                                  Handle<BigIntBase> x,
    1337             :                                                  Handle<BigIntBase> y,
    1338             :                                                  MutableBigInt result_storage) {
    1339             :   return AbsoluteBitwiseOp(isolate, x, y, result_storage, kSkip, kSymmetric,
    1340        1074 :                            [](digit_t a, digit_t b) { return a & b; });
    1341             : }
    1342             : 
    1343             : // If {result_storage} is non-nullptr, it will be used for the result,
    1344             : // otherwise a new BigInt of appropriate length will be allocated.
    1345             : // {result_storage} may alias {x} or {y} for in-place modification.
    1346         184 : Handle<MutableBigInt> MutableBigInt::AbsoluteAndNot(
    1347             :     Isolate* isolate, Handle<BigIntBase> x, Handle<BigIntBase> y,
    1348             :     MutableBigInt result_storage) {
    1349             :   return AbsoluteBitwiseOp(isolate, x, y, result_storage, kCopy, kNotSymmetric,
    1350         714 :                            [](digit_t a, digit_t b) { return a & ~b; });
    1351             : }
    1352             : 
    1353             : // If {result_storage} is non-nullptr, it will be used for the result,
    1354             : // otherwise a new BigInt of appropriate length will be allocated.
    1355             : // {result_storage} may alias {x} or {y} for in-place modification.
    1356         232 : Handle<MutableBigInt> MutableBigInt::AbsoluteOr(Isolate* isolate,
    1357             :                                                 Handle<BigIntBase> x,
    1358             :                                                 Handle<BigIntBase> y,
    1359             :                                                 MutableBigInt result_storage) {
    1360             :   return AbsoluteBitwiseOp(isolate, x, y, result_storage, kCopy, kSymmetric,
    1361         795 :                            [](digit_t a, digit_t b) { return a | b; });
    1362             : }
    1363             : 
    1364             : // If {result_storage} is non-nullptr, it will be used for the result,
    1365             : // otherwise a new BigInt of appropriate length will be allocated.
    1366             : // {result_storage} may alias {x} or {y} for in-place modification.
    1367         342 : Handle<MutableBigInt> MutableBigInt::AbsoluteXor(Isolate* isolate,
    1368             :                                                  Handle<BigIntBase> x,
    1369             :                                                  Handle<BigIntBase> y,
    1370             :                                                  MutableBigInt result_storage) {
    1371             :   return AbsoluteBitwiseOp(isolate, x, y, result_storage, kCopy, kSymmetric,
    1372        1359 :                            [](digit_t a, digit_t b) { return a ^ b; });
    1373             : }
    1374             : 
    1375             : // Returns a positive value if abs(x) > abs(y), a negative value if
    1376             : // abs(x) < abs(y), or zero if abs(x) == abs(y).
    1377        3093 : int MutableBigInt::AbsoluteCompare(Handle<BigIntBase> x, Handle<BigIntBase> y) {
    1378        3093 :   int diff = x->length() - y->length();
    1379        3093 :   if (diff != 0) return diff;
    1380        2328 :   int i = x->length() - 1;
    1381        5197 :   while (i >= 0 && x->digit(i) == y->digit(i)) i--;
    1382        2328 :   if (i < 0) return 0;
    1383        1629 :   return x->digit(i) > y->digit(i) ? 1 : -1;
    1384             : }
    1385             : 
    1386             : // Multiplies {multiplicand} with {multiplier} and adds the result to
    1387             : // {accumulator}, starting at {accumulator_index} for the least-significant
    1388             : // digit.
    1389             : // Callers must ensure that {accumulator} is big enough to hold the result.
    1390      193952 : void MutableBigInt::MultiplyAccumulate(Handle<BigIntBase> multiplicand,
    1391             :                                        digit_t multiplier,
    1392             :                                        Handle<MutableBigInt> accumulator,
    1393             :                                        int accumulator_index) {
    1394             :   // This is a minimum requirement; the DCHECK in the second loop below
    1395             :   // will enforce more as needed.
    1396             :   DCHECK(accumulator->length() > multiplicand->length() + accumulator_index);
    1397      193952 :   if (multiplier == 0L) return;
    1398             :   digit_t carry = 0;
    1399             :   digit_t high = 0;
    1400   330796924 :   for (int i = 0; i < multiplicand->length(); i++, accumulator_index++) {
    1401             :     digit_t acc = accumulator->digit(accumulator_index);
    1402             :     digit_t new_carry = 0;
    1403             :     // Add last round's carryovers.
    1404             :     acc = digit_add(acc, high, &new_carry);
    1405             :     acc = digit_add(acc, carry, &new_carry);
    1406             :     // Compute this round's multiplication.
    1407             :     digit_t m_digit = multiplicand->digit(i);
    1408             :     digit_t low = digit_mul(multiplier, m_digit, &high);
    1409             :     acc = digit_add(acc, low, &new_carry);
    1410             :     // Store result and prepare for next round.
    1411             :     accumulator->set_digit(accumulator_index, acc);
    1412             :     carry = new_carry;
    1413             :   }
    1414      426708 :   for (; carry != 0 || high != 0; accumulator_index++) {
    1415             :     DCHECK(accumulator_index < accumulator->length());
    1416             :     digit_t acc = accumulator->digit(accumulator_index);
    1417             :     digit_t new_carry = 0;
    1418             :     acc = digit_add(acc, high, &new_carry);
    1419             :     high = 0;
    1420             :     acc = digit_add(acc, carry, &new_carry);
    1421             :     accumulator->set_digit(accumulator_index, acc);
    1422             :     carry = new_carry;
    1423             :   }
    1424             : }
    1425             : 
    1426             : // Multiplies {source} with {factor} and adds {summand} to the result.
    1427             : // {result} and {source} may be the same BigInt for inplace modification.
    1428       37671 : void MutableBigInt::InternalMultiplyAdd(BigIntBase source, digit_t factor,
    1429             :                                         digit_t summand, int n,
    1430             :                                         MutableBigInt result) {
    1431             :   DCHECK(source->length() >= n);
    1432             :   DCHECK(result->length() >= n);
    1433             :   digit_t carry = summand;
    1434             :   digit_t high = 0;
    1435    50268639 :   for (int i = 0; i < n; i++) {
    1436             :     digit_t current = source->digit(i);
    1437             :     digit_t new_carry = 0;
    1438             :     // Compute this round's multiplication.
    1439             :     digit_t new_high = 0;
    1440             :     current = digit_mul(current, factor, &new_high);
    1441             :     // Add last round's carryovers.
    1442             :     current = digit_add(current, high, &new_carry);
    1443             :     current = digit_add(current, carry, &new_carry);
    1444             :     // Store result and prepare for next round.
    1445             :     result->set_digit(i, current);
    1446             :     carry = new_carry;
    1447             :     high = new_high;
    1448             :   }
    1449       37671 :   if (result->length() > n) {
    1450        3489 :     result->set_digit(n++, carry + high);
    1451             :     // Current callers don't pass in such large results, but let's be robust.
    1452        3489 :     while (n < result->length()) {
    1453           0 :       result->set_digit(n++, 0);
    1454             :     }
    1455             :   } else {
    1456       68364 :     CHECK_EQ(carry + high, 0);
    1457             :   }
    1458       37671 : }
    1459             : 
    1460             : // Multiplies {x} with {factor} and then adds {summand} to it.
    1461       34182 : void BigInt::InplaceMultiplyAdd(Handle<FreshlyAllocatedBigInt> x,
    1462             :                                 uintptr_t factor, uintptr_t summand) {
    1463             :   STATIC_ASSERT(sizeof(factor) == sizeof(digit_t));
    1464             :   STATIC_ASSERT(sizeof(summand) == sizeof(digit_t));
    1465             :   Handle<MutableBigInt> bigint = MutableBigInt::Cast(x);
    1466       68364 :   MutableBigInt::InternalMultiplyAdd(*bigint, factor, summand, bigint->length(),
    1467       34182 :                                      *bigint);
    1468       34182 : }
    1469             : 
    1470             : // Divides {x} by {divisor}, returning the result in {quotient} and {remainder}.
    1471             : // Mathematically, the contract is:
    1472             : // quotient = (x - remainder) / divisor, with 0 <= remainder < divisor.
    1473             : // If {quotient} is an empty handle, an appropriately sized BigInt will be
    1474             : // allocated for it; otherwise the caller must ensure that it is big enough.
    1475             : // {quotient} can be the same as {x} for an in-place division. {quotient} can
    1476             : // also be nullptr if the caller is only interested in the remainder.
    1477        3584 : void MutableBigInt::AbsoluteDivSmall(Isolate* isolate, Handle<BigIntBase> x,
    1478             :                                      digit_t divisor,
    1479             :                                      Handle<MutableBigInt>* quotient,
    1480             :                                      digit_t* remainder) {
    1481             :   DCHECK_NE(divisor, 0);
    1482             :   DCHECK(!x->is_zero());  // Callers check anyway, no need to handle this.
    1483        3584 :   *remainder = 0;
    1484             :   int length = x->length();
    1485        3584 :   if (quotient != nullptr) {
    1486        3398 :     if ((*quotient).is_null()) {
    1487        2428 :       *quotient = New(isolate, length).ToHandleChecked();
    1488             :     }
    1489     2622830 :     for (int i = length - 1; i >= 0; i--) {
    1490     2619432 :       digit_t q = digit_div(*remainder, x->digit(i), divisor, remainder);
    1491             :       (*quotient)->set_digit(i, q);
    1492             :     }
    1493             :   } else {
    1494         534 :     for (int i = length - 1; i >= 0; i--) {
    1495         348 :       digit_div(*remainder, x->digit(i), divisor, remainder);
    1496             :     }
    1497             :   }
    1498        3584 : }
    1499             : 
    1500             : // Divides {dividend} by {divisor}, returning the result in {quotient} and
    1501             : // {remainder}. Mathematically, the contract is:
    1502             : // quotient = (dividend - remainder) / divisor, with 0 <= remainder < divisor.
    1503             : // Both {quotient} and {remainder} are optional, for callers that are only
    1504             : // interested in one of them.
    1505             : // See Knuth, Volume 2, section 4.3.1, Algorithm D.
    1506         221 : bool MutableBigInt::AbsoluteDivLarge(Isolate* isolate,
    1507             :                                      Handle<BigIntBase> dividend,
    1508             :                                      Handle<BigIntBase> divisor,
    1509             :                                      Handle<MutableBigInt>* quotient,
    1510             :                                      Handle<MutableBigInt>* remainder) {
    1511             :   DCHECK_GE(divisor->length(), 2);
    1512             :   DCHECK(dividend->length() >= divisor->length());
    1513             :   // The unusual variable names inside this function are consistent with
    1514             :   // Knuth's book, as well as with Go's implementation of this algorithm.
    1515             :   // Maintaining this consistency is probably more useful than trying to
    1516             :   // come up with more descriptive names for them.
    1517             :   int n = divisor->length();
    1518         221 :   int m = dividend->length() - n;
    1519             : 
    1520             :   // The quotient to be computed.
    1521             :   Handle<MutableBigInt> q;
    1522         334 :   if (quotient != nullptr) q = New(isolate, m + 1).ToHandleChecked();
    1523             :   // In each iteration, {qhatv} holds {divisor} * {current quotient digit}.
    1524             :   // "v" is the book's name for {divisor}, "qhat" the current quotient digit.
    1525             :   Handle<MutableBigInt> qhatv;
    1526         442 :   if (!New(isolate, n + 1).ToHandle(&qhatv)) return false;
    1527             : 
    1528             :   // D1.
    1529             :   // Left-shift inputs so that the divisor's MSB is set. This is necessary
    1530             :   // to prevent the digit-wise divisions (see digit_div call below) from
    1531             :   // overflowing (they take a two digits wide input, and return a one digit
    1532             :   // result).
    1533         442 :   int shift = base::bits::CountLeadingZeros(divisor->digit(n - 1));
    1534         221 :   if (shift > 0) {
    1535         442 :     divisor = SpecialLeftShift(isolate, divisor, shift, kSameSizeResult)
    1536             :                   .ToHandleChecked();
    1537             :   }
    1538             :   // Holds the (continuously updated) remaining part of the dividend, which
    1539             :   // eventually becomes the remainder.
    1540             :   Handle<MutableBigInt> u;
    1541         442 :   if (!SpecialLeftShift(isolate, dividend, shift, kAlwaysAddOneDigit)
    1542             :            .ToHandle(&u)) {
    1543             :     return false;
    1544             :   }
    1545             : 
    1546             :   // D2.
    1547             :   // Iterate over the dividend's digit (like the "grad school" algorithm).
    1548             :   // {vn1} is the divisor's most significant digit.
    1549             :   digit_t vn1 = divisor->digit(n - 1);
    1550             :   uintptr_t work_estimate = 0;
    1551        7189 :   for (int j = m; j >= 0; j--) {
    1552             :     // D3.
    1553             :     // Estimate the current iteration's quotient digit (see Knuth for details).
    1554             :     // {qhat} is the current quotient digit.
    1555             :     digit_t qhat = std::numeric_limits<digit_t>::max();
    1556             :     // {ujn} is the dividend's most significant remaining digit.
    1557        3489 :     digit_t ujn = u->digit(j + n);
    1558        3489 :     if (ujn != vn1) {
    1559             :       // {rhat} is the current iteration's remainder.
    1560             :       digit_t rhat = 0;
    1561             :       // Estimate the current quotient digit by dividing the most significant
    1562             :       // digits of dividend and divisor. The result will not be too small,
    1563             :       // but could be a bit too large.
    1564        3489 :       qhat = digit_div(ujn, u->digit(j + n - 1), vn1, &rhat);
    1565             : 
    1566             :       // Decrement the quotient estimate as needed by looking at the next
    1567             :       // digit, i.e. by testing whether
    1568             :       // qhat * v_{n-2} > (rhat << kDigitBits) + u_{j+n-2}.
    1569        3489 :       digit_t vn2 = divisor->digit(n - 2);
    1570        3489 :       digit_t ujn2 = u->digit(j + n - 2);
    1571        4803 :       while (ProductGreaterThan(qhat, vn2, rhat, ujn2)) {
    1572        2048 :         qhat--;
    1573             :         digit_t prev_rhat = rhat;
    1574        2048 :         rhat += vn1;
    1575             :         // v[n-1] >= 0, so this tests for overflow.
    1576        2048 :         if (rhat < prev_rhat) break;
    1577             :       }
    1578             :     }
    1579             : 
    1580             :     // D4.
    1581             :     // Multiply the divisor with the current quotient digit, and subtract
    1582             :     // it from the dividend. If there was "borrow", then the quotient digit
    1583             :     // was one too high, so we must correct it and undo one subtraction of
    1584             :     // the (shifted) divisor.
    1585        3489 :     InternalMultiplyAdd(*divisor, qhat, 0, n, *qhatv);
    1586        3489 :     digit_t c = u->InplaceSub(qhatv, j);
    1587        3489 :     if (c != 0) {
    1588           0 :       c = u->InplaceAdd(divisor, j);
    1589           0 :       u->set_digit(j + n, u->digit(j + n) + c);
    1590           0 :       qhat--;
    1591             :     }
    1592             : 
    1593        3489 :     if (quotient != nullptr) q->set_digit(j, qhat);
    1594             : 
    1595             :     // Division can take a long time. Check for interrupt requests every
    1596             :     // now and then (roughly every 10-20 of milliseconds -- rarely enough
    1597             :     // not to create noticeable overhead, frequently enough not to appear
    1598             :     // frozen).
    1599        3489 :     work_estimate += n;
    1600        3489 :     if (work_estimate > 5000000) {
    1601             :       work_estimate = 0;
    1602             :       StackLimitCheck interrupt_check(isolate);
    1603          10 :       if (interrupt_check.InterruptRequested() &&
    1604           5 :           isolate->stack_guard()->HandleInterrupts()->IsException(isolate)) {
    1605           5 :         return false;
    1606             :       }
    1607             :     }
    1608             :   }
    1609         216 :   if (quotient != nullptr) {
    1610         108 :     *quotient = q;  // Caller will right-trim.
    1611             :   }
    1612         216 :   if (remainder != nullptr) {
    1613         108 :     u->InplaceRightShift(shift);
    1614         108 :     *remainder = u;
    1615             :   }
    1616             :   return true;
    1617             : }
    1618             : 
    1619             : // Returns whether (factor1 * factor2) > (high << kDigitBits) + low.
    1620           0 : bool MutableBigInt::ProductGreaterThan(digit_t factor1, digit_t factor2,
    1621             :                                        digit_t high, digit_t low) {
    1622             :   digit_t result_high;
    1623             :   digit_t result_low = digit_mul(factor1, factor2, &result_high);
    1624        4803 :   return result_high > high || (result_high == high && result_low > low);
    1625             : }
    1626             : 
    1627             : // Adds {summand} onto {this}, starting with {summand}'s 0th digit
    1628             : // at {this}'s {start_index}'th digit. Returns the "carry" (0 or 1).
    1629           0 : BigInt::digit_t MutableBigInt::InplaceAdd(Handle<BigIntBase> summand,
    1630             :                                           int start_index) {
    1631             :   digit_t carry = 0;
    1632             :   int n = summand->length();
    1633             :   DCHECK(length() >= start_index + n);
    1634           0 :   for (int i = 0; i < n; i++) {
    1635             :     digit_t new_carry = 0;
    1636             :     digit_t sum =
    1637           0 :         digit_add(digit(start_index + i), summand->digit(i), &new_carry);
    1638             :     sum = digit_add(sum, carry, &new_carry);
    1639             :     set_digit(start_index + i, sum);
    1640             :     carry = new_carry;
    1641             :   }
    1642           0 :   return carry;
    1643             : }
    1644             : 
    1645             : // Subtracts {subtrahend} from {this}, starting with {subtrahend}'s 0th digit
    1646             : // at {this}'s {start_index}-th digit. Returns the "borrow" (0 or 1).
    1647        3489 : BigInt::digit_t MutableBigInt::InplaceSub(Handle<BigIntBase> subtrahend,
    1648             :                                           int start_index) {
    1649             :   digit_t borrow = 0;
    1650             :   int n = subtrahend->length();
    1651             :   DCHECK(length() >= start_index + n);
    1652    50044203 :   for (int i = 0; i < n; i++) {
    1653             :     digit_t new_borrow = 0;
    1654             :     digit_t difference =
    1655    25020357 :         digit_sub(digit(start_index + i), subtrahend->digit(i), &new_borrow);
    1656             :     difference = digit_sub(difference, borrow, &new_borrow);
    1657             :     set_digit(start_index + i, difference);
    1658             :     borrow = new_borrow;
    1659             :   }
    1660        3489 :   return borrow;
    1661             : }
    1662             : 
    1663         108 : void MutableBigInt::InplaceRightShift(int shift) {
    1664             :   DCHECK_GE(shift, 0);
    1665             :   DCHECK_LT(shift, kDigitBits);
    1666             :   DCHECK_GT(length(), 0);
    1667             :   DCHECK_EQ(digit(0) & ((static_cast<digit_t>(1) << shift) - 1), 0);
    1668         108 :   if (shift == 0) return;
    1669         108 :   digit_t carry = digit(0) >> shift;
    1670         108 :   int last = length() - 1;
    1671         954 :   for (int i = 0; i < last; i++) {
    1672         423 :     digit_t d = digit(i + 1);
    1673         423 :     set_digit(i, (d << (kDigitBits - shift)) | carry);
    1674         423 :     carry = d >> shift;
    1675             :   }
    1676             :   set_digit(last, carry);
    1677             : }
    1678             : 
    1679             : // Always copies the input, even when {shift} == 0.
    1680             : // {shift} must be less than kDigitBits, {x} must be non-zero.
    1681         442 : MaybeHandle<MutableBigInt> MutableBigInt::SpecialLeftShift(
    1682             :     Isolate* isolate, Handle<BigIntBase> x, int shift,
    1683             :     SpecialLeftShiftMode mode) {
    1684             :   DCHECK_GE(shift, 0);
    1685             :   DCHECK_LT(shift, kDigitBits);
    1686             :   DCHECK_GT(x->length(), 0);
    1687             :   int n = x->length();
    1688         442 :   int result_length = mode == kAlwaysAddOneDigit ? n + 1 : n;
    1689             :   Handle<MutableBigInt> result;
    1690         884 :   if (!New(isolate, result_length).ToHandle(&result)) {
    1691           0 :     return MaybeHandle<MutableBigInt>();
    1692             :   }
    1693         442 :   if (shift == 0) {
    1694           0 :     for (int i = 0; i < n; i++) result->set_digit(i, x->digit(i));
    1695           0 :     if (mode == kAlwaysAddOneDigit) result->set_digit(n, 0);
    1696           0 :     return result;
    1697             :   }
    1698             :   DCHECK_GT(shift, 0);
    1699             :   digit_t carry = 0;
    1700      433058 :   for (int i = 0; i < n; i++) {
    1701             :     digit_t d = x->digit(i);
    1702      216308 :     result->set_digit(i, (d << shift) | carry);
    1703      216308 :     carry = d >> (kDigitBits - shift);
    1704             :   }
    1705         442 :   if (mode == kAlwaysAddOneDigit) {
    1706             :     result->set_digit(n, carry);
    1707             :   } else {
    1708             :     DCHECK_EQ(mode, kSameSizeResult);
    1709             :     DCHECK_EQ(carry, 0);
    1710             :   }
    1711         442 :   return result;
    1712             : }
    1713             : 
    1714         423 : MaybeHandle<BigInt> MutableBigInt::LeftShiftByAbsolute(Isolate* isolate,
    1715             :                                                        Handle<BigIntBase> x,
    1716             :                                                        Handle<BigIntBase> y) {
    1717         423 :   Maybe<digit_t> maybe_shift = ToShiftAmount(y);
    1718         423 :   if (maybe_shift.IsNothing()) {
    1719           0 :     THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntTooBig),
    1720             :                     BigInt);
    1721             :   }
    1722             :   digit_t shift = maybe_shift.FromJust();
    1723         423 :   int digit_shift = static_cast<int>(shift / kDigitBits);
    1724         423 :   int bits_shift = static_cast<int>(shift % kDigitBits);
    1725             :   int length = x->length();
    1726         801 :   bool grow = bits_shift != 0 &&
    1727         756 :               (x->digit(length - 1) >> (kDigitBits - bits_shift)) != 0;
    1728         423 :   int result_length = length + digit_shift + grow;
    1729         423 :   if (result_length > kMaxLength) {
    1730           0 :     THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntTooBig),
    1731             :                     BigInt);
    1732             :   }
    1733             :   Handle<MutableBigInt> result;
    1734         846 :   if (!New(isolate, result_length).ToHandle(&result)) {
    1735           0 :     return MaybeHandle<BigInt>();
    1736             :   }
    1737         423 :   if (bits_shift == 0) {
    1738             :     int i = 0;
    1739        1485 :     for (; i < digit_shift; i++) result->set_digit(i, 0ul);
    1740         135 :     for (; i < result_length; i++) {
    1741          45 :       result->set_digit(i, x->digit(i - digit_shift));
    1742             :     }
    1743             :   } else {
    1744             :     digit_t carry = 0;
    1745       32508 :     for (int i = 0; i < digit_shift; i++) result->set_digit(i, 0ul);
    1746        1800 :     for (int i = 0; i < length; i++) {
    1747             :       digit_t d = x->digit(i);
    1748         711 :       result->set_digit(i + digit_shift, (d << bits_shift) | carry);
    1749         711 :       carry = d >> (kDigitBits - bits_shift);
    1750             :     }
    1751         378 :     if (grow) {
    1752             :       result->set_digit(length + digit_shift, carry);
    1753             :     } else {
    1754             :       DCHECK_EQ(carry, 0);
    1755             :     }
    1756             :   }
    1757         846 :   result->set_sign(x->sign());
    1758         423 :   return MakeImmutable(result);
    1759             : }
    1760             : 
    1761         306 : Handle<BigInt> MutableBigInt::RightShiftByAbsolute(Isolate* isolate,
    1762             :                                                    Handle<BigIntBase> x,
    1763             :                                                    Handle<BigIntBase> y) {
    1764             :   int length = x->length();
    1765             :   bool sign = x->sign();
    1766         306 :   Maybe<digit_t> maybe_shift = ToShiftAmount(y);
    1767         306 :   if (maybe_shift.IsNothing()) {
    1768           0 :     return RightShiftByMaximum(isolate, sign);
    1769             :   }
    1770             :   digit_t shift = maybe_shift.FromJust();
    1771         306 :   int digit_shift = static_cast<int>(shift / kDigitBits);
    1772         306 :   int bits_shift = static_cast<int>(shift % kDigitBits);
    1773         306 :   int result_length = length - digit_shift;
    1774         306 :   if (result_length <= 0) {
    1775          96 :     return RightShiftByMaximum(isolate, sign);
    1776             :   }
    1777             :   // For negative numbers, round down if any bit was shifted out (so that e.g.
    1778             :   // -5n >> 1n == -3n and not -2n). Check now whether this will happen and
    1779             :   // whether it can cause overflow into a new digit. If we allocate the result
    1780             :   // large enough up front, it avoids having to do a second allocation later.
    1781             :   bool must_round_down = false;
    1782         210 :   if (sign) {
    1783          28 :     const digit_t mask = (static_cast<digit_t>(1) << bits_shift) - 1;
    1784          28 :     if ((x->digit(digit_shift) & mask) != 0) {
    1785             :       must_round_down = true;
    1786             :     } else {
    1787           5 :       for (int i = 0; i < digit_shift; i++) {
    1788           0 :         if (x->digit(i) != 0) {
    1789             :           must_round_down = true;
    1790             :           break;
    1791             :         }
    1792             :       }
    1793             :     }
    1794             :   }
    1795             :   // If bits_shift is non-zero, it frees up bits, preventing overflow.
    1796         210 :   if (must_round_down && bits_shift == 0) {
    1797             :     // Overflow cannot happen if the most significant digit has unset bits.
    1798           0 :     digit_t msd = x->digit(length - 1);
    1799             :     bool rounding_can_overflow = digit_ismax(msd);
    1800           0 :     if (rounding_can_overflow) result_length++;
    1801             :   }
    1802             : 
    1803             :   DCHECK_LE(result_length, length);
    1804         420 :   Handle<MutableBigInt> result = New(isolate, result_length).ToHandleChecked();
    1805         210 :   if (bits_shift == 0) {
    1806           0 :     for (int i = digit_shift; i < length; i++) {
    1807           0 :       result->set_digit(i - digit_shift, x->digit(i));
    1808             :     }
    1809             :   } else {
    1810         210 :     digit_t carry = x->digit(digit_shift) >> bits_shift;
    1811         210 :     int last = length - digit_shift - 1;
    1812         408 :     for (int i = 0; i < last; i++) {
    1813          99 :       digit_t d = x->digit(i + digit_shift + 1);
    1814          99 :       result->set_digit(i, (d << (kDigitBits - bits_shift)) | carry);
    1815          99 :       carry = d >> bits_shift;
    1816             :     }
    1817             :     result->set_digit(last, carry);
    1818             :   }
    1819             : 
    1820         210 :   if (sign) {
    1821          28 :     result->set_sign(true);
    1822          28 :     if (must_round_down) {
    1823             :       // Since the result is negative, rounding down means adding one to
    1824             :       // its absolute value. This cannot overflow.
    1825          46 :       result = AbsoluteAddOne(isolate, result, true, *result).ToHandleChecked();
    1826             :     }
    1827             :   }
    1828         210 :   return MakeImmutable(result);
    1829             : }
    1830             : 
    1831          96 : Handle<BigInt> MutableBigInt::RightShiftByMaximum(Isolate* isolate, bool sign) {
    1832          96 :   if (sign) {
    1833             :     // TODO(jkummerow): Consider caching a canonical -1n BigInt.
    1834          59 :     return NewFromInt(isolate, -1);
    1835             :   } else {
    1836          37 :     return Zero(isolate);
    1837             :   }
    1838             : }
    1839             : 
    1840             : // Returns the value of {x} if it is less than the maximum bit length of
    1841             : // a BigInt, or Nothing otherwise.
    1842         729 : Maybe<BigInt::digit_t> MutableBigInt::ToShiftAmount(Handle<BigIntBase> x) {
    1843         729 :   if (x->length() > 1) return Nothing<digit_t>();
    1844             :   digit_t value = x->digit(0);
    1845             :   STATIC_ASSERT(kMaxLengthBits < std::numeric_limits<digit_t>::max());
    1846         729 :   if (value > kMaxLengthBits) return Nothing<digit_t>();
    1847             :   return Just(value);
    1848             : }
    1849             : 
    1850             : // Lookup table for the maximum number of bits required per character of a
    1851             : // base-N string representation of a number. To increase accuracy, the array
    1852             : // value is the actual value multiplied by 32. To generate this table:
    1853             : // for (var i = 0; i <= 36; i++) { print(Math.ceil(Math.log2(i) * 32) + ","); }
    1854             : constexpr uint8_t kMaxBitsPerChar[] = {
    1855             :     0,   0,   32,  51,  64,  75,  83,  90,  96,  // 0..8
    1856             :     102, 107, 111, 115, 119, 122, 126, 128,      // 9..16
    1857             :     131, 134, 136, 139, 141, 143, 145, 147,      // 17..24
    1858             :     149, 151, 153, 154, 156, 158, 159, 160,      // 25..32
    1859             :     162, 163, 165, 166,                          // 33..36
    1860             : };
    1861             : 
    1862             : static const int kBitsPerCharTableShift = 5;
    1863             : static const size_t kBitsPerCharTableMultiplier = 1u << kBitsPerCharTableShift;
    1864             : 
    1865       11252 : MaybeHandle<FreshlyAllocatedBigInt> BigInt::AllocateFor(
    1866             :     Isolate* isolate, int radix, int charcount, ShouldThrow should_throw,
    1867             :     AllocationType allocation) {
    1868             :   DCHECK(2 <= radix && radix <= 36);
    1869             :   DCHECK_GE(charcount, 0);
    1870       11252 :   size_t bits_per_char = kMaxBitsPerChar[radix];
    1871       11252 :   size_t chars = static_cast<size_t>(charcount);
    1872             :   const int roundup = kBitsPerCharTableMultiplier - 1;
    1873       11252 :   if (chars <= (std::numeric_limits<size_t>::max() - roundup) / bits_per_char) {
    1874       11252 :     size_t bits_min = bits_per_char * chars;
    1875             :     // Divide by 32 (see table), rounding up.
    1876       11252 :     bits_min = (bits_min + roundup) >> kBitsPerCharTableShift;
    1877       11252 :     if (bits_min <= static_cast<size_t>(kMaxInt)) {
    1878             :       // Divide by kDigitsBits, rounding up.
    1879       11252 :       int length = static_cast<int>((bits_min + kDigitBits - 1) / kDigitBits);
    1880       11252 :       if (length <= kMaxLength) {
    1881             :         Handle<MutableBigInt> result =
    1882       22486 :             MutableBigInt::New(isolate, length, allocation).ToHandleChecked();
    1883             :         result->InitializeDigits(length);
    1884       11243 :         return result;
    1885             :       }
    1886             :     }
    1887             :   }
    1888             :   // All the overflow/maximum checks above fall through to here.
    1889           9 :   if (should_throw == kThrowOnError) {
    1890           0 :     THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntTooBig),
    1891             :                     FreshlyAllocatedBigInt);
    1892             :   } else {
    1893           9 :     return MaybeHandle<FreshlyAllocatedBigInt>();
    1894             :   }
    1895             : }
    1896             : 
    1897       11081 : Handle<BigInt> BigInt::Finalize(Handle<FreshlyAllocatedBigInt> x, bool sign) {
    1898             :   Handle<MutableBigInt> bigint = MutableBigInt::Cast(x);
    1899       22162 :   bigint->set_sign(sign);
    1900       11081 :   return MutableBigInt::MakeImmutable(bigint);
    1901             : }
    1902             : 
    1903             : // The serialization format MUST NOT CHANGE without updating the format
    1904             : // version in value-serializer.cc!
    1905           6 : uint32_t BigInt::GetBitfieldForSerialization() const {
    1906             :   // In order to make the serialization format the same on 32/64 bit builds,
    1907             :   // we convert the length-in-digits to length-in-bytes for serialization.
    1908             :   // Being able to do this depends on having enough LengthBits:
    1909             :   STATIC_ASSERT(kMaxLength * kDigitSize <= LengthBits::kMax);
    1910           6 :   int bytelength = length() * kDigitSize;
    1911           6 :   return SignBits::encode(sign()) | LengthBits::encode(bytelength);
    1912             : }
    1913             : 
    1914          17 : int BigInt::DigitsByteLengthForBitfield(uint32_t bitfield) {
    1915          17 :   return LengthBits::decode(bitfield);
    1916             : }
    1917             : 
    1918             : // The serialization format MUST NOT CHANGE without updating the format
    1919             : // version in value-serializer.cc!
    1920           6 : void BigInt::SerializeDigits(uint8_t* storage) {
    1921             :   void* digits =
    1922           6 :       reinterpret_cast<void*>(ptr() + kDigitsOffset - kHeapObjectTag);
    1923             : #if defined(V8_TARGET_LITTLE_ENDIAN)
    1924           6 :   int bytelength = length() * kDigitSize;
    1925           6 :   memcpy(storage, digits, bytelength);
    1926             : #elif defined(V8_TARGET_BIG_ENDIAN)
    1927             :   digit_t* digit_storage = reinterpret_cast<digit_t*>(storage);
    1928             :   const digit_t* digit = reinterpret_cast<const digit_t*>(digits);
    1929             :   for (int i = 0; i < length(); i++) {
    1930             :     *digit_storage = ByteReverse(*digit);
    1931             :     digit_storage++;
    1932             :     digit++;
    1933             :   }
    1934             : #endif  // V8_TARGET_BIG_ENDIAN
    1935           6 : }
    1936             : 
    1937             : // The serialization format MUST NOT CHANGE without updating the format
    1938             : // version in value-serializer.cc!
    1939          11 : MaybeHandle<BigInt> BigInt::FromSerializedDigits(
    1940             :     Isolate* isolate, uint32_t bitfield, Vector<const uint8_t> digits_storage,
    1941             :     AllocationType allocation) {
    1942             :   int bytelength = LengthBits::decode(bitfield);
    1943             :   DCHECK(digits_storage.length() == bytelength);
    1944             :   bool sign = SignBits::decode(bitfield);
    1945          11 :   int length = (bytelength + kDigitSize - 1) / kDigitSize;  // Round up.
    1946             :   Handle<MutableBigInt> result =
    1947          11 :       MutableBigInt::Cast(isolate->factory()->NewBigInt(length, allocation));
    1948             :   result->initialize_bitfield(sign, length);
    1949             :   void* digits =
    1950          11 :       reinterpret_cast<void*>(result->ptr() + kDigitsOffset - kHeapObjectTag);
    1951             : #if defined(V8_TARGET_LITTLE_ENDIAN)
    1952          11 :   memcpy(digits, digits_storage.start(), bytelength);
    1953             :   void* padding_start =
    1954          11 :       reinterpret_cast<void*>(reinterpret_cast<Address>(digits) + bytelength);
    1955          11 :   memset(padding_start, 0, length * kDigitSize - bytelength);
    1956             : #elif defined(V8_TARGET_BIG_ENDIAN)
    1957             :   digit_t* digit = reinterpret_cast<digit_t*>(digits);
    1958             :   const digit_t* digit_storage =
    1959             :       reinterpret_cast<const digit_t*>(digits_storage.start());
    1960             :   for (int i = 0; i < bytelength / kDigitSize; i++) {
    1961             :     *digit = ByteReverse(*digit_storage);
    1962             :     digit_storage++;
    1963             :     digit++;
    1964             :   }
    1965             :   if (bytelength % kDigitSize) {
    1966             :     *digit = 0;
    1967             :     byte* digit_byte = reinterpret_cast<byte*>(digit);
    1968             :     digit_byte += sizeof(*digit) - 1;
    1969             :     const byte* digit_storage_byte =
    1970             :         reinterpret_cast<const byte*>(digit_storage);
    1971             :     for (int i = 0; i < bytelength % kDigitSize; i++) {
    1972             :       *digit_byte = *digit_storage_byte;
    1973             :       digit_byte--;
    1974             :       digit_storage_byte++;
    1975             :     }
    1976             :   }
    1977             : #endif  // V8_TARGET_BIG_ENDIAN
    1978          11 :   return MutableBigInt::MakeImmutable(result);
    1979             : }
    1980             : 
    1981             : static const char kConversionChars[] = "0123456789abcdefghijklmnopqrstuvwxyz";
    1982             : 
    1983         180 : MaybeHandle<String> MutableBigInt::ToStringBasePowerOfTwo(
    1984             :     Isolate* isolate, Handle<BigIntBase> x, int radix,
    1985             :     ShouldThrow should_throw) {
    1986             :   STATIC_ASSERT(base::bits::IsPowerOfTwo(kDigitBits));
    1987             :   DCHECK(base::bits::IsPowerOfTwo(radix));
    1988             :   DCHECK(radix >= 2 && radix <= 32);
    1989             :   DCHECK(!x->is_zero());
    1990             : 
    1991             :   const int length = x->length();
    1992             :   const bool sign = x->sign();
    1993         180 :   const int bits_per_char = base::bits::CountTrailingZeros(radix);
    1994         180 :   const int char_mask = radix - 1;
    1995             :   // Compute the length of the resulting string: divide the bit length of the
    1996             :   // BigInt by the number of bits representable per character (rounding up).
    1997         180 :   const digit_t msd = x->digit(length - 1);
    1998         180 :   const int msd_leading_zeros = base::bits::CountLeadingZeros(msd);
    1999         180 :   const size_t bit_length = length * kDigitBits - msd_leading_zeros;
    2000             :   const size_t chars_required =
    2001         180 :       (bit_length + bits_per_char - 1) / bits_per_char + sign;
    2002             : 
    2003         180 :   if (chars_required > String::kMaxLength) {
    2004           0 :     if (should_throw == kThrowOnError) {
    2005           0 :       THROW_NEW_ERROR(isolate, NewInvalidStringLengthError(), String);
    2006             :     } else {
    2007           0 :       return MaybeHandle<String>();
    2008             :     }
    2009             :   }
    2010             : 
    2011             :   Handle<SeqOneByteString> result =
    2012             :       isolate->factory()
    2013         360 :           ->NewRawOneByteString(static_cast<int>(chars_required))
    2014             :           .ToHandleChecked();
    2015             :   DisallowHeapAllocation no_gc;
    2016             :   uint8_t* buffer = result->GetChars(no_gc);
    2017             :   // Print the number into the string, starting from the last position.
    2018         180 :   int pos = static_cast<int>(chars_required - 1);
    2019             :   digit_t digit = 0;
    2020             :   // Keeps track of how many unprocessed bits there are in {digit}.
    2021             :   int available_bits = 0;
    2022         720 :   for (int i = 0; i < length - 1; i++) {
    2023             :     digit_t new_digit = x->digit(i);
    2024             :     // Take any leftover bits from the last iteration into account.
    2025         270 :     int current = (digit | (new_digit << available_bits)) & char_mask;
    2026         270 :     buffer[pos--] = kConversionChars[current];
    2027         270 :     int consumed_bits = bits_per_char - available_bits;
    2028         270 :     digit = new_digit >> consumed_bits;
    2029         270 :     available_bits = kDigitBits - consumed_bits;
    2030       11250 :     while (available_bits >= bits_per_char) {
    2031        5490 :       buffer[pos--] = kConversionChars[digit & char_mask];
    2032        5490 :       digit >>= bits_per_char;
    2033        5490 :       available_bits -= bits_per_char;
    2034             :     }
    2035             :   }
    2036             :   // Take any leftover bits from the last iteration into account.
    2037         180 :   int current = (digit | (msd << available_bits)) & char_mask;
    2038         180 :   buffer[pos--] = kConversionChars[current];
    2039         180 :   digit = msd >> (bits_per_char - available_bits);
    2040        3762 :   while (digit != 0) {
    2041        1791 :     buffer[pos--] = kConversionChars[digit & char_mask];
    2042        1791 :     digit >>= bits_per_char;
    2043             :   }
    2044         180 :   if (sign) buffer[pos--] = '-';
    2045             :   DCHECK_EQ(pos, -1);
    2046         180 :   return result;
    2047             : }
    2048             : 
    2049       10760 : MaybeHandle<String> MutableBigInt::ToStringGeneric(Isolate* isolate,
    2050             :                                                    Handle<BigIntBase> x,
    2051             :                                                    int radix,
    2052             :                                                    ShouldThrow should_throw) {
    2053             :   DCHECK(radix >= 2 && radix <= 36);
    2054             :   DCHECK(!x->is_zero());
    2055             :   Heap* heap = isolate->heap();
    2056             : 
    2057             :   const int length = x->length();
    2058             :   const bool sign = x->sign();
    2059             : 
    2060             :   // Compute (an overapproximation of) the length of the resulting string:
    2061             :   // Divide bit length of the BigInt by bits representable per character.
    2062             :   const size_t bit_length =
    2063       21520 :       length * kDigitBits - base::bits::CountLeadingZeros(x->digit(length - 1));
    2064             :   // Maximum number of bits we can represent with one character. We'll use this
    2065             :   // to find an appropriate chunk size below.
    2066       10760 :   const uint8_t max_bits_per_char = kMaxBitsPerChar[radix];
    2067             :   // For estimating result length, we have to be pessimistic and work with
    2068             :   // the minimum number of bits one character can represent.
    2069       10760 :   const uint8_t min_bits_per_char = max_bits_per_char - 1;
    2070             :   // Perform the following computation with uint64_t to avoid overflows.
    2071             :   uint64_t chars_required = bit_length;
    2072       10760 :   chars_required *= kBitsPerCharTableMultiplier;
    2073       10760 :   chars_required += min_bits_per_char - 1;  // Round up.
    2074       10760 :   chars_required /= min_bits_per_char;
    2075       10760 :   chars_required += sign;
    2076             : 
    2077       10760 :   if (chars_required > String::kMaxLength) {
    2078           0 :     if (should_throw == kThrowOnError) {
    2079           0 :       THROW_NEW_ERROR(isolate, NewInvalidStringLengthError(), String);
    2080             :     } else {
    2081           0 :       return MaybeHandle<String>();
    2082             :     }
    2083             :   }
    2084             :   Handle<SeqOneByteString> result =
    2085             :       isolate->factory()
    2086       21520 :           ->NewRawOneByteString(static_cast<int>(chars_required))
    2087             :           .ToHandleChecked();
    2088             : 
    2089             : #if DEBUG
    2090             :   // Zap the string first.
    2091             :   {
    2092             :     DisallowHeapAllocation no_gc;
    2093             :     uint8_t* chars = result->GetChars(no_gc);
    2094             :     for (int i = 0; i < static_cast<int>(chars_required); i++) chars[i] = '?';
    2095             :   }
    2096             : #endif
    2097             : 
    2098             :   // We assemble the result string in reverse order, and then reverse it.
    2099             :   // TODO(jkummerow): Consider building the string from the right, and
    2100             :   // left-shifting it if the length estimate was too large.
    2101             :   int pos = 0;
    2102             : 
    2103             :   digit_t last_digit;
    2104       10760 :   if (length == 1) {
    2105             :     last_digit = x->digit(0);
    2106             :   } else {
    2107             :     int chunk_chars =
    2108        1010 :         kDigitBits * kBitsPerCharTableMultiplier / max_bits_per_char;
    2109        1010 :     digit_t chunk_divisor = digit_pow(radix, chunk_chars);
    2110             :     // By construction of chunk_chars, there can't have been overflow.
    2111             :     DCHECK_NE(chunk_divisor, 0);
    2112             :     int nonzero_digit = length - 1;
    2113             :     DCHECK_NE(x->digit(nonzero_digit), 0);
    2114             :     // {rest} holds the part of the BigInt that we haven't looked at yet.
    2115             :     // Not to be confused with "remainder"!
    2116             :     Handle<MutableBigInt> rest;
    2117             :     // In the first round, divide the input, allocating a new BigInt for
    2118             :     // the result == rest; from then on divide the rest in-place.
    2119             :     Handle<BigIntBase>* dividend = &x;
    2120             :     uintptr_t work_estimate = 0;
    2121             :     do {
    2122             :       digit_t chunk;
    2123        3194 :       AbsoluteDivSmall(isolate, *dividend, chunk_divisor, &rest, &chunk);
    2124             :       DCHECK(!rest.is_null());
    2125             :       dividend = reinterpret_cast<Handle<BigIntBase>*>(&rest);
    2126             :       DisallowHeapAllocation no_gc;
    2127             :       uint8_t* chars = result->GetChars(no_gc);
    2128      105468 :       for (int i = 0; i < chunk_chars; i++) {
    2129       51137 :         chars[pos++] = kConversionChars[chunk % radix];
    2130       51137 :         chunk /= radix;
    2131             :       }
    2132             :       DCHECK_EQ(chunk, 0);
    2133        3194 :       if (rest->digit(nonzero_digit) == 0) nonzero_digit--;
    2134             :       // We can never clear more than one digit per iteration, because
    2135             :       // chunk_divisor is smaller than max digit value.
    2136             :       DCHECK_GT(rest->digit(nonzero_digit), 0);
    2137             : 
    2138             :       // String formatting can take a long time. Check for interrupt requests
    2139             :       // every now and then (roughly every 10-20 of milliseconds -- rarely
    2140             :       // enough not to create noticeable overhead, frequently enough not to
    2141             :       // appear frozen).
    2142        3194 :       work_estimate += length;
    2143        3194 :       if (work_estimate > 500000) {
    2144             :         work_estimate = 0;
    2145             :         StackLimitCheck interrupt_check(isolate);
    2146           5 :         if (interrupt_check.InterruptRequested()) {
    2147             :           {
    2148             :             AllowHeapAllocation might_throw;
    2149          10 :             if (isolate->stack_guard()->HandleInterrupts()->IsException(
    2150             :                     isolate)) {
    2151           5 :               return MaybeHandle<String>();
    2152             :             }
    2153             :           }
    2154             :           // If there was an interrupt request but no termination, reload
    2155             :           // the raw characters pointer (as the string might have moved).
    2156             :           chars = result->GetChars(no_gc);
    2157             :         }
    2158           0 :         if (interrupt_check.InterruptRequested() &&
    2159           0 :             isolate->stack_guard()->HandleInterrupts()->IsException(isolate)) {
    2160           0 :           return MaybeHandle<String>();
    2161             :         }
    2162             :       }
    2163        3189 :     } while (nonzero_digit > 0);
    2164             :     last_digit = rest->digit(0);
    2165             :   }
    2166             :   DisallowHeapAllocation no_gc;
    2167             :   uint8_t* chars = result->GetChars(no_gc);
    2168             :   do {
    2169       38310 :     chars[pos++] = kConversionChars[last_digit % radix];
    2170       38310 :     last_digit /= radix;
    2171       38310 :   } while (last_digit > 0);
    2172             :   DCHECK_GE(pos, 1);
    2173             :   DCHECK(pos <= static_cast<int>(chars_required));
    2174             :   // Remove leading zeroes.
    2175       10755 :   while (pos > 1 && chars[pos - 1] == '0') pos--;
    2176       10755 :   if (sign) chars[pos++] = '-';
    2177             :   // Trim any over-allocation (which can happen due to conservative estimates).
    2178       10755 :   if (pos < static_cast<int>(chars_required)) {
    2179             :     result->synchronized_set_length(pos);
    2180             :     int string_size =
    2181             :         SeqOneByteString::SizeFor(static_cast<int>(chars_required));
    2182             :     int needed_size = SeqOneByteString::SizeFor(pos);
    2183         554 :     if (needed_size < string_size) {
    2184          36 :       Address new_end = result->address() + needed_size;
    2185             :       heap->CreateFillerObjectAt(new_end, (string_size - needed_size),
    2186          36 :                                  ClearRecordedSlots::kNo);
    2187             :     }
    2188             :   }
    2189             :   // Reverse the string.
    2190       51898 :   for (int i = 0, j = pos - 1; i < j; i++, j--) {
    2191       41143 :     uint8_t tmp = chars[i];
    2192       41143 :     chars[i] = chars[j];
    2193       41143 :     chars[j] = tmp;
    2194             :   }
    2195             : #if DEBUG
    2196             :   // Verify that all characters have been written.
    2197             :   DCHECK(result->length() == pos);
    2198             :   for (int i = 0; i < pos; i++) DCHECK_NE(chars[i], '?');
    2199             : #endif
    2200       10755 :   return result;
    2201             : }
    2202             : 
    2203        1134 : Handle<BigInt> BigInt::AsIntN(Isolate* isolate, uint64_t n, Handle<BigInt> x) {
    2204        1134 :   if (x->is_zero()) return x;
    2205        1017 :   if (n == 0) return MutableBigInt::Zero(isolate);
    2206         972 :   uint64_t needed_length = (n + kDigitBits - 1) / kDigitBits;
    2207         972 :   uint64_t x_length = static_cast<uint64_t>(x->length());
    2208             :   // If {x} has less than {n} bits, return it directly.
    2209         972 :   if (x_length < needed_length) return x;
    2210             :   DCHECK_LE(needed_length, kMaxInt);
    2211         792 :   digit_t top_digit = x->digit(static_cast<int>(needed_length) - 1);
    2212         792 :   digit_t compare_digit = static_cast<digit_t>(1) << ((n - 1) % kDigitBits);
    2213         792 :   if (x_length == needed_length && top_digit < compare_digit) return x;
    2214             :   // Otherwise we have to truncate (which is a no-op in the special case
    2215             :   // of x == -2^(n-1)), and determine the right sign. We also might have
    2216             :   // to subtract from 2^n to simulate having two's complement representation.
    2217             :   // In most cases, the result's sign is x->sign() xor "(n-1)th bit present".
    2218             :   // The only exception is when x is negative, has the (n-1)th bit, and all
    2219             :   // its bits below (n-1) are zero. In that case, the result is the minimum
    2220             :   // n-bit integer (example: asIntN(3, -12n) => -4n).
    2221         459 :   bool has_bit = (top_digit & compare_digit) == compare_digit;
    2222             :   DCHECK_LE(n, kMaxInt);
    2223         459 :   int N = static_cast<int>(n);
    2224         459 :   if (!has_bit) {
    2225         144 :     return MutableBigInt::TruncateToNBits(isolate, N, x);
    2226             :   }
    2227         315 :   if (!x->sign()) {
    2228         162 :     return MutableBigInt::TruncateAndSubFromPowerOfTwo(isolate, N, x, true);
    2229             :   }
    2230             :   // Negative numbers must subtract from 2^n, except for the special case
    2231             :   // described above.
    2232         153 :   if ((top_digit & (compare_digit - 1)) == 0) {
    2233          45 :     for (int i = static_cast<int>(needed_length) - 2; i >= 0; i--) {
    2234           0 :       if (x->digit(i) != 0) {
    2235             :         return MutableBigInt::TruncateAndSubFromPowerOfTwo(isolate, N, x,
    2236           0 :                                                            false);
    2237             :       }
    2238             :     }
    2239             :     // Truncation is no-op if x == -2^(n-1).
    2240          45 :     if (x_length == needed_length && top_digit == compare_digit) return x;
    2241          18 :     return MutableBigInt::TruncateToNBits(isolate, N, x);
    2242             :   }
    2243         108 :   return MutableBigInt::TruncateAndSubFromPowerOfTwo(isolate, N, x, false);
    2244             : }
    2245             : 
    2246        1152 : MaybeHandle<BigInt> BigInt::AsUintN(Isolate* isolate, uint64_t n,
    2247             :                                     Handle<BigInt> x) {
    2248        1152 :   if (x->is_zero()) return x;
    2249        1035 :   if (n == 0) return MutableBigInt::Zero(isolate);
    2250             :   // If {x} is negative, simulate two's complement representation.
    2251         990 :   if (x->sign()) {
    2252         441 :     if (n > kMaxLengthBits) {
    2253           0 :       THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntTooBig),
    2254             :                       BigInt);
    2255             :     }
    2256             :     return MutableBigInt::TruncateAndSubFromPowerOfTwo(
    2257         441 :         isolate, static_cast<int>(n), x, false);
    2258             :   }
    2259             :   // If {x} is positive and has up to {n} bits, return it directly.
    2260         549 :   if (n >= kMaxLengthBits) return x;
    2261             :   STATIC_ASSERT(kMaxLengthBits < kMaxInt - kDigitBits);
    2262         513 :   int needed_length = static_cast<int>((n + kDigitBits - 1) / kDigitBits);
    2263         513 :   if (x->length() < needed_length) return x;
    2264         441 :   int bits_in_top_digit = n % kDigitBits;
    2265         441 :   if (x->length() == needed_length) {
    2266         423 :     if (bits_in_top_digit == 0) return x;
    2267         405 :     digit_t top_digit = x->digit(needed_length - 1);
    2268         405 :     if ((top_digit >> bits_in_top_digit) == 0) return x;
    2269             :   }
    2270             :   // Otherwise, truncate.
    2271             :   DCHECK_LE(n, kMaxInt);
    2272         189 :   return MutableBigInt::TruncateToNBits(isolate, static_cast<int>(n), x);
    2273             : }
    2274             : 
    2275         351 : Handle<BigInt> MutableBigInt::TruncateToNBits(Isolate* isolate, int n,
    2276             :                                               Handle<BigInt> x) {
    2277             :   // Only call this when there's something to do.
    2278             :   DCHECK_NE(n, 0);
    2279             :   DCHECK_GT(x->length(), n / kDigitBits);
    2280             : 
    2281         351 :   int needed_digits = (n + (kDigitBits - 1)) / kDigitBits;
    2282             :   DCHECK_LE(needed_digits, x->length());
    2283         702 :   Handle<MutableBigInt> result = New(isolate, needed_digits).ToHandleChecked();
    2284             : 
    2285             :   // Copy all digits except the MSD.
    2286         351 :   int last = needed_digits - 1;
    2287         351 :   for (int i = 0; i < last; i++) {
    2288             :     result->set_digit(i, x->digit(i));
    2289             :   }
    2290             : 
    2291             :   // The MSD might contain extra bits that we don't want.
    2292             :   digit_t msd = x->digit(last);
    2293         351 :   if (n % kDigitBits != 0) {
    2294         333 :     int drop = kDigitBits - (n % kDigitBits);
    2295         333 :     msd = (msd << drop) >> drop;
    2296             :   }
    2297             :   result->set_digit(last, msd);
    2298         702 :   result->set_sign(x->sign());
    2299         351 :   return MakeImmutable(result);
    2300             : }
    2301             : 
    2302             : // Subtracts the least significant n bits of abs(x) from 2^n.
    2303         711 : Handle<BigInt> MutableBigInt::TruncateAndSubFromPowerOfTwo(Isolate* isolate,
    2304             :                                                            int n,
    2305             :                                                            Handle<BigInt> x,
    2306             :                                                            bool result_sign) {
    2307             :   DCHECK_NE(n, 0);
    2308             :   DCHECK_LE(n, kMaxLengthBits);
    2309             : 
    2310         711 :   int needed_digits = (n + (kDigitBits - 1)) / kDigitBits;
    2311             :   DCHECK_LE(needed_digits, kMaxLength);  // Follows from n <= kMaxLengthBits.
    2312        1422 :   Handle<MutableBigInt> result = New(isolate, needed_digits).ToHandleChecked();
    2313             : 
    2314             :   // Process all digits except the MSD.
    2315             :   int i = 0;
    2316         711 :   int last = needed_digits - 1;
    2317             :   int x_length = x->length();
    2318             :   digit_t borrow = 0;
    2319             :   // Take digits from {x} unless its length is exhausted.
    2320             :   int limit = Min(last, x_length);
    2321         855 :   for (; i < limit; i++) {
    2322             :     digit_t new_borrow = 0;
    2323             :     digit_t difference = digit_sub(0, x->digit(i), &new_borrow);
    2324             :     difference = digit_sub(difference, borrow, &new_borrow);
    2325             :     result->set_digit(i, difference);
    2326             :     borrow = new_borrow;
    2327             :   }
    2328             :   // Then simulate leading zeroes in {x} as needed.
    2329         747 :   for (; i < last; i++) {
    2330             :     digit_t new_borrow = 0;
    2331             :     digit_t difference = digit_sub(0, borrow, &new_borrow);
    2332             :     result->set_digit(i, difference);
    2333             :     borrow = new_borrow;
    2334             :   }
    2335             : 
    2336             :   // The MSD might contain extra bits that we don't want.
    2337         711 :   digit_t msd = last < x_length ? x->digit(last) : 0;
    2338         711 :   int msd_bits_consumed = n % kDigitBits;
    2339             :   digit_t result_msd;
    2340         711 :   if (msd_bits_consumed == 0) {
    2341             :     digit_t new_borrow = 0;
    2342             :     result_msd = digit_sub(0, msd, &new_borrow);
    2343             :     result_msd = digit_sub(result_msd, borrow, &new_borrow);
    2344             :   } else {
    2345         675 :     int drop = kDigitBits - msd_bits_consumed;
    2346         675 :     msd = (msd << drop) >> drop;
    2347         675 :     digit_t minuend_msd = static_cast<digit_t>(1) << (kDigitBits - drop);
    2348             :     digit_t new_borrow = 0;
    2349             :     result_msd = digit_sub(minuend_msd, msd, &new_borrow);
    2350             :     result_msd = digit_sub(result_msd, borrow, &new_borrow);
    2351             :     DCHECK_EQ(new_borrow, 0);  // result < 2^n.
    2352             :     // If all subtracted bits were zero, we have to get rid of the
    2353             :     // materialized minuend_msd again.
    2354         675 :     result_msd &= (minuend_msd - 1);
    2355             :   }
    2356             :   result->set_digit(last, result_msd);
    2357        1422 :   result->set_sign(result_sign);
    2358         711 :   return MakeImmutable(result);
    2359             : }
    2360             : 
    2361         133 : Handle<BigInt> BigInt::FromInt64(Isolate* isolate, int64_t n) {
    2362         133 :   if (n == 0) return MutableBigInt::Zero(isolate);
    2363             :   STATIC_ASSERT(kDigitBits == 64 || kDigitBits == 32);
    2364             :   int length = 64 / kDigitBits;
    2365             :   Handle<MutableBigInt> result =
    2366         124 :       MutableBigInt::Cast(isolate->factory()->NewBigInt(length));
    2367         124 :   bool sign = n < 0;
    2368             :   result->initialize_bitfield(sign, length);
    2369             :   uint64_t absolute;
    2370         124 :   if (!sign) {
    2371          98 :     absolute = static_cast<uint64_t>(n);
    2372             :   } else {
    2373          26 :     if (n == std::numeric_limits<int64_t>::min()) {
    2374             :       absolute = static_cast<uint64_t>(std::numeric_limits<int64_t>::max()) + 1;
    2375             :     } else {
    2376          10 :       absolute = static_cast<uint64_t>(-n);
    2377             :     }
    2378             :   }
    2379             :   result->set_64_bits(absolute);
    2380         124 :   return MutableBigInt::MakeImmutable(result);
    2381             : }
    2382             : 
    2383          36 : Handle<BigInt> BigInt::FromUint64(Isolate* isolate, uint64_t n) {
    2384          36 :   if (n == 0) return MutableBigInt::Zero(isolate);
    2385             :   STATIC_ASSERT(kDigitBits == 64 || kDigitBits == 32);
    2386             :   int length = 64 / kDigitBits;
    2387             :   Handle<MutableBigInt> result =
    2388          36 :       MutableBigInt::Cast(isolate->factory()->NewBigInt(length));
    2389             :   result->initialize_bitfield(false, length);
    2390             :   result->set_64_bits(n);
    2391          36 :   return MutableBigInt::MakeImmutable(result);
    2392             : }
    2393             : 
    2394          30 : MaybeHandle<BigInt> BigInt::FromWords64(Isolate* isolate, int sign_bit,
    2395             :                                         int words64_count,
    2396             :                                         const uint64_t* words) {
    2397          30 :   if (words64_count < 0 || words64_count > kMaxLength / (64 / kDigitBits)) {
    2398          30 :     THROW_NEW_ERROR(isolate, NewRangeError(MessageTemplate::kBigIntTooBig),
    2399             :                     BigInt);
    2400             :   }
    2401          15 :   if (words64_count == 0) return MutableBigInt::Zero(isolate);
    2402             :   STATIC_ASSERT(kDigitBits == 64 || kDigitBits == 32);
    2403             :   int length = (64 / kDigitBits) * words64_count;
    2404             :   DCHECK_GT(length, 0);
    2405             :   if (kDigitBits == 32 && words[words64_count - 1] <= (1ULL << 32)) length--;
    2406             : 
    2407             :   Handle<MutableBigInt> result;
    2408          20 :   if (!MutableBigInt::New(isolate, length).ToHandle(&result)) {
    2409           0 :     return MaybeHandle<BigInt>();
    2410             :   }
    2411             : 
    2412          20 :   result->set_sign(sign_bit);
    2413             :   if (kDigitBits == 64) {
    2414          50 :     for (int i = 0; i < length; ++i) {
    2415          20 :       result->set_digit(i, static_cast<digit_t>(words[i]));
    2416             :     }
    2417             :   } else {
    2418             :     for (int i = 0; i < length; i += 2) {
    2419             :       digit_t lo = static_cast<digit_t>(words[i / 2]);
    2420             :       digit_t hi = static_cast<digit_t>(words[i / 2] >> 32);
    2421             :       result->set_digit(i, lo);
    2422             :       if (i + 1 < length) result->set_digit(i + 1, hi);
    2423             :     }
    2424             :   }
    2425             : 
    2426          10 :   return MutableBigInt::MakeImmutable(result);
    2427             : }
    2428             : 
    2429          15 : int BigInt::Words64Count() {
    2430             :   STATIC_ASSERT(kDigitBits == 64 || kDigitBits == 32);
    2431             :   return length() / (64 / kDigitBits) +
    2432          15 :          (kDigitBits == 32 && length() % 2 == 1 ? 1 : 0);
    2433             : }
    2434             : 
    2435          10 : void BigInt::ToWordsArray64(int* sign_bit, int* words64_count,
    2436             :                             uint64_t* words) {
    2437             :   DCHECK_NE(sign_bit, nullptr);
    2438             :   DCHECK_NE(words64_count, nullptr);
    2439          10 :   *sign_bit = sign();
    2440          10 :   int available_words = *words64_count;
    2441          10 :   *words64_count = Words64Count();
    2442          10 :   if (available_words == 0) return;
    2443             :   DCHECK_NE(words, nullptr);
    2444             : 
    2445             :   int len = length();
    2446             :   if (kDigitBits == 64) {
    2447          50 :     for (int i = 0; i < len && i < available_words; ++i) words[i] = digit(i);
    2448             :   } else {
    2449             :     for (int i = 0; i < len && available_words > 0; i += 2) {
    2450             :       uint64_t lo = digit(i);
    2451             :       uint64_t hi = (i + 1) < len ? digit(i + 1) : 0;
    2452             :       words[i / 2] = lo | (hi << 32);
    2453             :       available_words--;
    2454             :     }
    2455             :   }
    2456             : }
    2457             : 
    2458        1802 : uint64_t MutableBigInt::GetRawBits(BigIntBase x, bool* lossless) {
    2459        1802 :   if (lossless != nullptr) *lossless = true;
    2460        1802 :   if (x->is_zero()) return 0;
    2461             :   int len = x->length();
    2462             :   STATIC_ASSERT(kDigitBits == 64 || kDigitBits == 32);
    2463        1743 :   if (lossless != nullptr && len > 64 / kDigitBits) *lossless = false;
    2464             :   uint64_t raw = static_cast<uint64_t>(x->digit(0));
    2465             :   if (kDigitBits == 32 && len > 1) {
    2466             :     raw |= static_cast<uint64_t>(x->digit(1)) << 32;
    2467             :   }
    2468             :   // Simulate two's complement. MSVC dislikes "-raw".
    2469        1743 :   return x->sign() ? ((~raw) + 1u) : raw;
    2470             : }
    2471             : 
    2472        1045 : int64_t BigInt::AsInt64(bool* lossless) {
    2473        1045 :   uint64_t raw = MutableBigInt::GetRawBits(*this, lossless);
    2474        1045 :   int64_t result = static_cast<int64_t>(raw);
    2475        1291 :   if (lossless != nullptr && (result < 0) != sign()) *lossless = false;
    2476        1045 :   return result;
    2477             : }
    2478             : 
    2479         757 : uint64_t BigInt::AsUint64(bool* lossless) {
    2480         757 :   uint64_t result = MutableBigInt::GetRawBits(*this, lossless);
    2481        1003 :   if (lossless != nullptr && sign()) *lossless = false;
    2482         757 :   return result;
    2483             : }
    2484             : 
    2485             : // Digit arithmetic helpers.
    2486             : 
    2487             : #if V8_TARGET_ARCH_32_BIT
    2488             : #define HAVE_TWODIGIT_T 1
    2489             : using twodigit_t = uint64_t;
    2490             : #elif defined(__SIZEOF_INT128__)
    2491             : // Both Clang and GCC support this on x64.
    2492             : #define HAVE_TWODIGIT_T 1
    2493             : using twodigit_t = __uint128_t;
    2494             : #endif
    2495             : 
    2496             : // {carry} must point to an initialized digit_t and will either be incremented
    2497             : // by one or left alone.
    2498             : inline BigInt::digit_t MutableBigInt::digit_add(digit_t a, digit_t b,
    2499             :                                                 digit_t* carry) {
    2500             : #if HAVE_TWODIGIT_T
    2501   546530618 :   twodigit_t result = static_cast<twodigit_t>(a) + static_cast<twodigit_t>(b);
    2502   546530618 :   *carry += result >> kDigitBits;
    2503             :   return static_cast<digit_t>(result);
    2504             : #else
    2505             :   digit_t result = a + b;
    2506             :   if (result < a) *carry += 1;
    2507             :   return result;
    2508             : #endif
    2509             : }
    2510             : 
    2511             : // {borrow} must point to an initialized digit_t and will either be incremented
    2512             : // by one or left alone.
    2513             : inline BigInt::digit_t MutableBigInt::digit_sub(digit_t a, digit_t b,
    2514             :                                                 digit_t* borrow) {
    2515             : #if HAVE_TWODIGIT_T
    2516    51816203 :   twodigit_t result = static_cast<twodigit_t>(a) - static_cast<twodigit_t>(b);
    2517    51816203 :   *borrow += (result >> kDigitBits) & 1;
    2518             :   return static_cast<digit_t>(result);
    2519             : #else
    2520             :   digit_t result = a - b;
    2521             :   if (result > a) *borrow += 1;
    2522             :   return static_cast<digit_t>(result);
    2523             : #endif
    2524             : }
    2525             : 
    2526             : // Returns the low half of the result. High half is in {high}.
    2527             : inline BigInt::digit_t MutableBigInt::digit_mul(digit_t a, digit_t b,
    2528             :                                                 digit_t* high) {
    2529             : #if HAVE_TWODIGIT_T
    2530   190421773 :   twodigit_t result = static_cast<twodigit_t>(a) * static_cast<twodigit_t>(b);
    2531   190421773 :   *high = result >> kDigitBits;
    2532   190421773 :   return static_cast<digit_t>(result);
    2533             : #else
    2534             :   // Multiply in half-pointer-sized chunks.
    2535             :   // For inputs [AH AL]*[BH BL], the result is:
    2536             :   //
    2537             :   //            [AL*BL]  // r_low
    2538             :   //    +    [AL*BH]     // r_mid1
    2539             :   //    +    [AH*BL]     // r_mid2
    2540             :   //    + [AH*BH]        // r_high
    2541             :   //    = [R4 R3 R2 R1]  // high = [R4 R3], low = [R2 R1]
    2542             :   //
    2543             :   // Where of course we must be careful with carries between the columns.
    2544             :   digit_t a_low = a & kHalfDigitMask;
    2545             :   digit_t a_high = a >> kHalfDigitBits;
    2546             :   digit_t b_low = b & kHalfDigitMask;
    2547             :   digit_t b_high = b >> kHalfDigitBits;
    2548             : 
    2549             :   digit_t r_low = a_low * b_low;
    2550             :   digit_t r_mid1 = a_low * b_high;
    2551             :   digit_t r_mid2 = a_high * b_low;
    2552             :   digit_t r_high = a_high * b_high;
    2553             : 
    2554             :   digit_t carry = 0;
    2555             :   digit_t low = digit_add(r_low, r_mid1 << kHalfDigitBits, &carry);
    2556             :   low = digit_add(low, r_mid2 << kHalfDigitBits, &carry);
    2557             :   *high =
    2558             :       (r_mid1 >> kHalfDigitBits) + (r_mid2 >> kHalfDigitBits) + r_high + carry;
    2559             :   return low;
    2560             : #endif
    2561             : }
    2562             : 
    2563             : // Returns the quotient.
    2564             : // quotient = (high << kDigitBits + low - remainder) / divisor
    2565             : BigInt::digit_t MutableBigInt::digit_div(digit_t high, digit_t low,
    2566             :                                          digit_t divisor, digit_t* remainder) {
    2567             :   DCHECK(high < divisor);
    2568             : #if V8_TARGET_ARCH_X64 && (__GNUC__ || __clang__)
    2569             :   digit_t quotient;
    2570             :   digit_t rem;
    2571             :   __asm__("divq  %[divisor]"
    2572             :           // Outputs: {quotient} will be in rax, {rem} in rdx.
    2573             :           : "=a"(quotient), "=d"(rem)
    2574             :           // Inputs: put {high} into rdx, {low} into rax, and {divisor} into
    2575             :           // any register or stack slot.
    2576     2623269 :           : "d"(high), "a"(low), [divisor] "rm"(divisor));
    2577     2619780 :   *remainder = rem;
    2578             :   return quotient;
    2579             : #elif V8_TARGET_ARCH_IA32 && (__GNUC__ || __clang__)
    2580             :   digit_t quotient;
    2581             :   digit_t rem;
    2582             :   __asm__("divl  %[divisor]"
    2583             :           // Outputs: {quotient} will be in eax, {rem} in edx.
    2584             :           : "=a"(quotient), "=d"(rem)
    2585             :           // Inputs: put {high} into edx, {low} into eax, and {divisor} into
    2586             :           // any register or stack slot.
    2587             :           : "d"(high), "a"(low), [divisor] "rm"(divisor));
    2588             :   *remainder = rem;
    2589             :   return quotient;
    2590             : #else
    2591             :   static const digit_t kHalfDigitBase = 1ull << kHalfDigitBits;
    2592             :   // Adapted from Warren, Hacker's Delight, p. 152.
    2593             :   int s = base::bits::CountLeadingZeros(divisor);
    2594             :   DCHECK_NE(s, kDigitBits);  // {divisor} is not 0.
    2595             :   divisor <<= s;
    2596             : 
    2597             :   digit_t vn1 = divisor >> kHalfDigitBits;
    2598             :   digit_t vn0 = divisor & kHalfDigitMask;
    2599             :   // {s} can be 0. {low >> kDigitBits} would be undefined behavior, so
    2600             :   // we mask the shift amount with {kShiftMask}, and the result with
    2601             :   // {s_zero_mask} which is 0 if s == 0 and all 1-bits otherwise.
    2602             :   STATIC_ASSERT(sizeof(intptr_t) == sizeof(digit_t));
    2603             :   const int kShiftMask = kDigitBits - 1;
    2604             :   digit_t s_zero_mask =
    2605             :       static_cast<digit_t>(static_cast<intptr_t>(-s) >> (kDigitBits - 1));
    2606             :   digit_t un32 =
    2607             :       (high << s) | ((low >> ((kDigitBits - s) & kShiftMask)) & s_zero_mask);
    2608             :   digit_t un10 = low << s;
    2609             :   digit_t un1 = un10 >> kHalfDigitBits;
    2610             :   digit_t un0 = un10 & kHalfDigitMask;
    2611             :   digit_t q1 = un32 / vn1;
    2612             :   digit_t rhat = un32 - q1 * vn1;
    2613             : 
    2614             :   while (q1 >= kHalfDigitBase || q1 * vn0 > rhat * kHalfDigitBase + un1) {
    2615             :     q1--;
    2616             :     rhat += vn1;
    2617             :     if (rhat >= kHalfDigitBase) break;
    2618             :   }
    2619             : 
    2620             :   digit_t un21 = un32 * kHalfDigitBase + un1 - q1 * divisor;
    2621             :   digit_t q0 = un21 / vn1;
    2622             :   rhat = un21 - q0 * vn1;
    2623             : 
    2624             :   while (q0 >= kHalfDigitBase || q0 * vn0 > rhat * kHalfDigitBase + un0) {
    2625             :     q0--;
    2626             :     rhat += vn1;
    2627             :     if (rhat >= kHalfDigitBase) break;
    2628             :   }
    2629             : 
    2630             :   *remainder = (un21 * kHalfDigitBase + un0 - q0 * divisor) >> s;
    2631             :   return q1 * kHalfDigitBase + q0;
    2632             : #endif
    2633             : }
    2634             : 
    2635             : // Raises {base} to the power of {exponent}. Does not check for overflow.
    2636           0 : BigInt::digit_t MutableBigInt::digit_pow(digit_t base, digit_t exponent) {
    2637             :   digit_t result = 1ull;
    2638        5736 :   while (exponent > 0) {
    2639        4726 :     if (exponent & 1) {
    2640        2814 :       result *= base;
    2641             :     }
    2642        4726 :     exponent >>= 1;
    2643        4726 :     base *= base;
    2644             :   }
    2645           0 :   return result;
    2646             : }
    2647             : 
    2648             : #undef HAVE_TWODIGIT_T
    2649             : 
    2650           0 : void MutableBigInt::set_64_bits(uint64_t bits) {
    2651             :   STATIC_ASSERT(kDigitBits == 64 || kDigitBits == 32);
    2652             :   if (kDigitBits == 64) {
    2653             :     set_digit(0, static_cast<digit_t>(bits));
    2654             :   } else {
    2655             :     set_digit(0, static_cast<digit_t>(bits & 0xFFFFFFFFu));
    2656             :     set_digit(1, static_cast<digit_t>(bits >> 32));
    2657             :   }
    2658           0 : }
    2659             : 
    2660             : #ifdef OBJECT_PRINT
    2661             : void BigInt::BigIntPrint(std::ostream& os) {
    2662             :   DisallowHeapAllocation no_gc;
    2663             :   PrintHeader(os, "BigInt");
    2664             :   int len = length();
    2665             :   os << "\n- length: " << len;
    2666             :   os << "\n- sign: " << sign();
    2667             :   if (len > 0) {
    2668             :     os << "\n- digits:";
    2669             :     for (int i = 0; i < len; i++) {
    2670             :       os << "\n    0x" << std::hex << digit(i);
    2671             :     }
    2672             :   }
    2673             :   os << std::dec << "\n";
    2674             : }
    2675             : #endif  // OBJECT_PRINT
    2676             : 
    2677             : }  // namespace internal
    2678      122036 : }  // namespace v8

Generated by: LCOV version 1.10