src/objects/intl-objects.tq

// Copyright 2019 the V8 project authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

#include 'src/objects/js-objects.h'
#include 'src/objects/intl-objects.h'

extern macro IntlAsciiCollationWeightsL1(): RawPtr<uint8>;
extern macro IntlAsciiCollationWeightsL3(): RawPtr<uint8>;
const kIntlAsciiCollationWeightsLength:
    constexpr int31 generates 'Intl::kAsciiCollationWeightsLength';

macro IntlAsciiCollationWeightL1(c: char8): uint8 labels _Bailout {
  static_assert(kIntlAsciiCollationWeightsLength == 256);
  return IntlAsciiCollationWeightsL1()[Convert<intptr>(c)];
}
macro IntlAsciiCollationWeightL1(c: char16): uint8 labels Bailout {
  if (Convert<uint32>(c) >= kIntlAsciiCollationWeightsLength) goto Bailout;
  return IntlAsciiCollationWeightsL1()[Convert<intptr>(c)];
}

macro IntlAsciiCollationWeightL3(c: char8): uint8 labels _Bailout {
  static_assert(kIntlAsciiCollationWeightsLength == 256);
  return IntlAsciiCollationWeightsL3()[Convert<intptr>(c)];
}
macro IntlAsciiCollationWeightL3(c: char16): uint8 labels Bailout {
  if (Convert<uint32>(c) >= kIntlAsciiCollationWeightsLength) goto Bailout;
  return IntlAsciiCollationWeightsL3()[Convert<intptr>(c)];
}

macro CheckEmptyOr1Byte(
    _it: torque_internal::SliceIterator<char8, const &char8>):
    void labels _Bailout {
  // char8 is always within 0xFF.
}
macro CheckEmptyOr1Byte(
    it: torque_internal::SliceIterator<char16, const &char16>):
    void labels Bailout {
  let it = it;
  if ((it.Next() otherwise return) > 0xFF) goto Bailout;
}

// This fast path works for ASCII-only strings and is based on the assumption
// that most strings are either bytewise equal or differ on L1 (i.e., not just
// in capitalization). So we first compare the strings on L1 and only afterwards
// consider L3. This makes use of the 256-entry L1 and L3 tables defined in
// src/objects/intl-objects.cc.
macro LocaleCompareFastPath<T1: type, T2: type>(
    left: ConstSlice<T1>, right: ConstSlice<T2>): Number labels Bailout {
  if (EqualContent(left, right)) return 0;
  let leftIt = left.Iterator();
  let rightIt = right.Iterator();
  while (true) {
    try {
      const lChar = leftIt.Next() otherwise goto LeftExhausted;
      const leftWeight = IntlAsciiCollationWeightL1(lChar) otherwise Bailout;
      if (leftWeight == 0) goto Bailout;
      // If rightIt is exhausted, we already checked that the next char of the
      // left string has non-zero weight, so it cannot be ignorable or a
      // combining character.
      // Return 1 because right string is shorter and L1 is equal.
      const rChar = rightIt.Next() otherwise return 1;
      const rightWeight = IntlAsciiCollationWeightL1(rChar) otherwise Bailout;
      if (rightWeight == 0) goto Bailout;
      if (leftWeight == rightWeight) continue;
      // The result is only valid if the last processed character is not
      // followed by a unicode combining character (we are overly strict and
      // restrict to code points up to 0xFF).
      CheckEmptyOr1Byte(leftIt) otherwise Bailout;
      CheckEmptyOr1Byte(rightIt) otherwise Bailout;
      if (leftWeight < rightWeight) return -1;
      return 1;
    } label LeftExhausted {
      const rChar = rightIt.Next() otherwise break;
      const rightWeight = IntlAsciiCollationWeightL1(rChar) otherwise Bailout;
      // If the following character might be ignorable or a combining character,
      // we bail out because the strings might still be considered equal.
      if (rightWeight == 0) goto Bailout;
      // Return -1 because left string is shorter and L1 is equal.
      return -1;
    }
  }
  leftIt = left.Iterator();
  rightIt = right.Iterator();
  while (true) {
    const lChar = leftIt.Next() otherwise unreachable;
    const leftWeight = IntlAsciiCollationWeightL3(lChar) otherwise unreachable;
    dcheck(leftWeight != 0);
    const rChar = rightIt.Next() otherwise unreachable;
    const rightWeight = IntlAsciiCollationWeightL3(rChar) otherwise unreachable;
    dcheck(rightWeight != 0);
    dcheck(
        IntlAsciiCollationWeightL1(lChar) otherwise unreachable ==
        IntlAsciiCollationWeightL1(rChar) otherwise unreachable);
    if (leftWeight == rightWeight) continue;
    if (leftWeight < rightWeight) return -1;
    return 1;
  }
  VerifiedUnreachable();
}

transitioning builtin StringFastLocaleCompare(
    implicit context: Context)(localeCompareFn: JSFunction, left: JSAny,
    right: JSAny, locales: JSAny): JSAny {
  try {
    const left = Cast<String>(left) otherwise Bailout;
    if (TaggedEqual(left, right)) return SmiConstant(0);
    StringToSlice(left) otherwise LeftOneByte, LeftTwoByte;
  } label LeftOneByte(leftSlice: ConstSlice<char8>) {
    try {
      const right = Cast<String>(right) otherwise Bailout;
      StringToSlice(right) otherwise RightOneByte, RightTwoByte;
    } label RightOneByte(rightSlice: ConstSlice<char8>) {
      return LocaleCompareFastPath(leftSlice, rightSlice) otherwise Bailout;
    } label RightTwoByte(rightSlice: ConstSlice<char16>) {
      return LocaleCompareFastPath(leftSlice, rightSlice) otherwise Bailout;
    }
  } label LeftTwoByte(leftSlice: ConstSlice<char16>) {
    try {
      const right = Cast<String>(right) otherwise Bailout;
      StringToSlice(right) otherwise RightOneByte, RightTwoByte;
    } label RightOneByte(rightSlice: ConstSlice<char8>) {
      return LocaleCompareFastPath(leftSlice, rightSlice) otherwise Bailout;
    } label RightTwoByte(rightSlice: ConstSlice<char16>) {
      return LocaleCompareFastPath(leftSlice, rightSlice) otherwise Bailout;
    }
  } label Bailout deferred {
    return Call(context, localeCompareFn, left, right, locales);
  }
}