Function NormalizeAndSimplifyString

libs/indexer/search_string_utils.cpp:180–255

}

UniString NormalizeAndSimplifyString(std::string_view s)
{
  UniString uniString = MakeUniString(s);
  for (size_t i = 0; i < uniString.size(); ++i)
  {
    UniChar & c = uniString[i];
    switch (c)
    {
    // Replace "d with stroke" to simple d letter. Used in Vietnamese.
    // (unicode-compliant implementation leaves it unchanged)
    case 0x0110:
    case 0x0111: c = 'd'; break;
    // Replace small turkish dotless 'ı' with dotted 'i'.  Our own
    // invented hack to avoid well-known Turkish I-letter bug.
    case 0x0131: c = 'i'; break;
    // Replace capital turkish dotted 'İ' with dotted lowercased 'i'.
    // Here we need to handle this case manually too, because default
    // unicode-compliant implementation of MakeLowerCase converts 'İ'
    // to 'i' + 0x0307.
    case 0x0130: c = 'i'; break;
    case 0x0152:  // Œ
    case 0x0153:  // œ
      c = 'o';
      uniString.insert(uniString.begin() + (i++) + 1, 'e');
      break;
    case 0x2018:  // ‘
    case 0x2019:  // ’
      c = '\'';
      break;
    case 0x2116:  // №
      c = '#';
      break;
    }
  }

  MakeLowerCaseInplace(uniString);
  NormalizeInplace(uniString);
  TransliterateHiraganaToKatakana(uniString);

  // Remove accents that can appear after NFKD normalization.
  uniString.erase_if([](UniChar const & c)
  {
    // ̀  COMBINING GRAVE ACCENT
    // ́  COMBINING ACUTE ACCENT
    return (c == 0x0300 || c == 0x0301);
  });

  // Replace sequence of spaces with single one.
  base::Unique(uniString, [](UniChar l, UniChar r) { return (l == r && l == ' '); });

  return uniString;

  /// @todo Restore this logic to distinguish и-й in future.
  /*
  // Just after lower casing is a correct place to avoid normalization for specific chars.
  static auto const isSpecificChar = [](UniChar c) -> bool
  {
    return c == 0x0439; // й