@ohos-ports/react-native-quick-crypto 0.0.0-stage → 1.1.8-beta.0

package/deps/simdutf/include/simdutf/scalar/utf8.h

@@ -0,0 +1,326 @@
+#ifndef SIMDUTF_UTF8_H
+#define SIMDUTF_UTF8_H
+
+namespace simdutf {
+namespace scalar {
+namespace {
+namespace utf8 {
+
+// credit: based on code from Google Fuchsia (Apache Licensed)
+template <class BytePtr>
+simdutf_constexpr23 simdutf_warn_unused bool validate(BytePtr data,
+                                                      size_t len) noexcept {
+  static_assert(
+      std::is_same<typename std::decay<decltype(*data)>::type, uint8_t>::value,
+      "dereferencing the data pointer must result in a uint8_t");
+  uint64_t pos = 0;
+  uint32_t code_point = 0;
+  while (pos < len) {
+    uint64_t next_pos;
+#if SIMDUTF_CPLUSPLUS23
+    if !consteval
+#endif
+    { // check if the next 16 bytes are ascii.
+      next_pos = pos + 16;
+      if (next_pos <= len) { // if it is safe to read 16 more bytes, check
+                             // that they are ascii
+        uint64_t v1{};
+        std::memcpy(&v1, data + pos, sizeof(uint64_t));
+        uint64_t v2{};
+        std::memcpy(&v2, data + pos + sizeof(uint64_t), sizeof(uint64_t));
+        uint64_t v{v1 | v2};
+        if ((v & 0x8080808080808080) == 0) {
+          pos = next_pos;
+          continue;
+        }
+      }
+    }
+
+    unsigned char byte = data[pos];
+
+    while (byte < 0b10000000) {
+      if (++pos == len) {
+        return true;
+      }
+      byte = data[pos];
+    }
+
+    if ((byte & 0b11100000) == 0b11000000) {
+      next_pos = pos + 2;
+      if (next_pos > len) {
+        return false;
+      }
+      if ((data[pos + 1] & 0b11000000) != 0b10000000) {
+        return false;
+      }
+      // range check
+      code_point = (byte & 0b00011111) << 6 | (data[pos + 1] & 0b00111111);
+      if ((code_point < 0x80) || (0x7ff < code_point)) {
+        return false;
+      }
+    } else if ((byte & 0b11110000) == 0b11100000) {
+      next_pos = pos + 3;
+      if (next_pos > len) {
+        return false;
+      }
+      if ((data[pos + 1] & 0b11000000) != 0b10000000) {
+        return false;
+      }
+      if ((data[pos + 2] & 0b11000000) != 0b10000000) {
+        return false;
+      }
+      // range check
+      code_point = (byte & 0b00001111) << 12 |
+                   (data[pos + 1] & 0b00111111) << 6 |
+                   (data[pos + 2] & 0b00111111);
+      if ((code_point < 0x800) || (0xffff < code_point) ||
+          (0xd7ff < code_point && code_point < 0xe000)) {
+        return false;
+      }
+    } else if ((byte & 0b11111000) == 0b11110000) { // 0b11110000
+      next_pos = pos + 4;
+      if (next_pos > len) {
+        return false;
+      }
+      if ((data[pos + 1] & 0b11000000) != 0b10000000) {
+        return false;
+      }
+      if ((data[pos + 2] & 0b11000000) != 0b10000000) {
+        return false;
+      }
+      if ((data[pos + 3] & 0b11000000) != 0b10000000) {
+        return false;
+      }
+      // range check
+      code_point =
+          (byte & 0b00000111) << 18 | (data[pos + 1] & 0b00111111) << 12 |
+          (data[pos + 2] & 0b00111111) << 6 | (data[pos + 3] & 0b00111111);
+      if (code_point <= 0xffff || 0x10ffff < code_point) {
+        return false;
+      }
+    } else {
+      // we may have a continuation
+      return false;
+    }
+    pos = next_pos;
+  }
+  return true;
+}
+
+simdutf_really_inline simdutf_warn_unused bool validate(const char *buf,
+                                                        size_t len) noexcept {
+  return validate(reinterpret_cast<const uint8_t *>(buf), len);
+}
+
+template <class BytePtr>
+simdutf_constexpr23 simdutf_warn_unused result
+validate_with_errors(BytePtr data, size_t len) noexcept {
+  static_assert(
+      std::is_same<typename std::decay<decltype(*data)>::type, uint8_t>::value,
+      "dereferencing the data pointer must result in a uint8_t");
+  size_t pos = 0;
+  uint32_t code_point = 0;
+  while (pos < len) {
+    // check of the next 16 bytes are ascii.
+    size_t next_pos = pos + 16;
+    if (next_pos <=
+        len) { // if it is safe to read 16 more bytes, check that they are ascii
+      uint64_t v1;
+      std::memcpy(&v1, data + pos, sizeof(uint64_t));
+      uint64_t v2;
+      std::memcpy(&v2, data + pos + sizeof(uint64_t), sizeof(uint64_t));
+      uint64_t v{v1 | v2};
+      if ((v & 0x8080808080808080) == 0) {
+        pos = next_pos;
+        continue;
+      }
+    }
+    unsigned char byte = data[pos];
+
+    while (byte < 0b10000000) {
+      if (++pos == len) {
+        return result(error_code::SUCCESS, len);
+      }
+      byte = data[pos];
+    }
+
+    if ((byte & 0b11100000) == 0b11000000) {
+      next_pos = pos + 2;
+      if (next_pos > len) {
+        return result(error_code::TOO_SHORT, pos);
+      }
+      if ((data[pos + 1] & 0b11000000) != 0b10000000) {
+        return result(error_code::TOO_SHORT, pos);
+      }
+      // range check
+      code_point = (byte & 0b00011111) << 6 | (data[pos + 1] & 0b00111111);
+      if ((code_point < 0x80) || (0x7ff < code_point)) {
+        return result(error_code::OVERLONG, pos);
+      }
+    } else if ((byte & 0b11110000) == 0b11100000) {
+      next_pos = pos + 3;
+      if (next_pos > len) {
+        return result(error_code::TOO_SHORT, pos);
+      }
+      if ((data[pos + 1] & 0b11000000) != 0b10000000) {
+        return result(error_code::TOO_SHORT, pos);
+      }
+      if ((data[pos + 2] & 0b11000000) != 0b10000000) {
+        return result(error_code::TOO_SHORT, pos);
+      }
+      // range check
+      code_point = (byte & 0b00001111) << 12 |
+                   (data[pos + 1] & 0b00111111) << 6 |
+                   (data[pos + 2] & 0b00111111);
+      if ((code_point < 0x800) || (0xffff < code_point)) {
+        return result(error_code::OVERLONG, pos);
+      }
+      if (0xd7ff < code_point && code_point < 0xe000) {
+        return result(error_code::SURROGATE, pos);
+      }
+    } else if ((byte & 0b11111000) == 0b11110000) { // 0b11110000
+      next_pos = pos + 4;
+      if (next_pos > len) {
+        return result(error_code::TOO_SHORT, pos);
+      }
+      if ((data[pos + 1] & 0b11000000) != 0b10000000) {
+        return result(error_code::TOO_SHORT, pos);
+      }
+      if ((data[pos + 2] & 0b11000000) != 0b10000000) {
+        return result(error_code::TOO_SHORT, pos);
+      }
+      if ((data[pos + 3] & 0b11000000) != 0b10000000) {
+        return result(error_code::TOO_SHORT, pos);
+      }
+      // range check
+      code_point =
+          (byte & 0b00000111) << 18 | (data[pos + 1] & 0b00111111) << 12 |
+          (data[pos + 2] & 0b00111111) << 6 | (data[pos + 3] & 0b00111111);
+      if (code_point <= 0xffff) {
+        return result(error_code::OVERLONG, pos);
+      }
+      if (0x10ffff < code_point) {
+        return result(error_code::TOO_LARGE, pos);
+      }
+    } else {
+      // we either have too many continuation bytes or an invalid leading byte
+      if ((byte & 0b11000000) == 0b10000000) {
+        return result(error_code::TOO_LONG, pos);
+      } else {
+        return result(error_code::HEADER_BITS, pos);
+      }
+    }
+    pos = next_pos;
+  }
+  return result(error_code::SUCCESS, len);
+}
+
+simdutf_really_inline simdutf_warn_unused result
+validate_with_errors(const char *buf, size_t len) noexcept {
+  return validate_with_errors(reinterpret_cast<const uint8_t *>(buf), len);
+}
+
+// Finds the previous leading byte starting backward from buf and validates with
+// errors from there Used to pinpoint the location of an error when an invalid
+// chunk is detected We assume that the stream starts with a leading byte, and
+// to check that it is the case, we ask that you pass a pointer to the start of
+// the stream (start).
+inline simdutf_warn_unused result rewind_and_validate_with_errors(
+    const char *start, const char *buf, size_t len) noexcept {
+  // First check that we start with a leading byte
+  if ((*start & 0b11000000) == 0b10000000) {
+    return result(error_code::TOO_LONG, 0);
+  }
+  size_t extra_len{0};
+  // A leading byte cannot be further than 4 bytes away
+  for (int i = 0; i < 5; i++) {
+    unsigned char byte = *buf;
+    if ((byte & 0b11000000) != 0b10000000) {
+      break;
+    } else {
+      buf--;
+      extra_len++;
+    }
+  }
+
+  result res = validate_with_errors(buf, len + extra_len);
+  res.count -= extra_len;
+  return res;
+}
+
+template <typename InputPtr>
+#if SIMDUTF_CPLUSPLUS20
+  requires simdutf::detail::indexes_into_byte_like<InputPtr>
+#endif
+simdutf_constexpr23 size_t count_code_points(InputPtr data, size_t len) {
+  size_t counter{0};
+  for (size_t i = 0; i < len; i++) {
+    // -65 is 0b10111111, anything larger in two-complement's should start a new
+    // code point.
+    if (int8_t(data[i]) > -65) {
+      counter++;
+    }
+  }
+  return counter;
+}
+
+template <typename InputPtr>
+#if SIMDUTF_CPLUSPLUS20
+  requires simdutf::detail::indexes_into_byte_like<InputPtr>
+#endif
+simdutf_constexpr23 size_t utf16_length_from_utf8(InputPtr data, size_t len) {
+  size_t counter{0};
+  for (size_t i = 0; i < len; i++) {
+    if (int8_t(data[i]) > -65) {
+      counter++;
+    }
+    if (uint8_t(data[i]) >= 240) {
+      counter++;
+    }
+  }
+  return counter;
+}
+
+template <typename InputPtr>
+#if SIMDUTF_CPLUSPLUS20
+  requires simdutf::detail::indexes_into_byte_like<InputPtr>
+#endif
+simdutf_warn_unused simdutf_constexpr23 size_t
+trim_partial_utf8(InputPtr input, size_t length) {
+  if (length < 3) {
+    switch (length) {
+    case 2:
+      if (uint8_t(input[length - 1]) >= 0xc0) {
+        return length - 1;
+      } // 2-, 3- and 4-byte characters with only 1 byte left
+      if (uint8_t(input[length - 2]) >= 0xe0) {
+        return length - 2;
+      } // 3- and 4-byte characters with only 2 bytes left
+      return length;
+    case 1:
+      if (uint8_t(input[length - 1]) >= 0xc0) {
+        return length - 1;
+      } // 2-, 3- and 4-byte characters with only 1 byte left
+      return length;
+    case 0:
+      return length;
+    }
+  }
+  if (uint8_t(input[length - 1]) >= 0xc0) {
+    return length - 1;
+  } // 2-, 3- and 4-byte characters with only 1 byte left
+  if (uint8_t(input[length - 2]) >= 0xe0) {
+    return length - 2;
+  } // 3- and 4-byte characters with only 1 byte left
+  if (uint8_t(input[length - 3]) >= 0xf0) {
+    return length - 3;
+  } // 4-byte characters with only 3 bytes left
+  return length;
+}
+
+} // namespace utf8
+} // unnamed namespace
+} // namespace scalar
+} // namespace simdutf
+
+#endif

package/deps/simdutf/include/simdutf/scalar/utf8_to_latin1/utf8_to_latin1.h

@@ -0,0 +1,225 @@
+#ifndef SIMDUTF_UTF8_TO_LATIN1_H
+#define SIMDUTF_UTF8_TO_LATIN1_H
+
+namespace simdutf {
+namespace scalar {
+namespace {
+namespace utf8_to_latin1 {
+
+template <typename InputPtr, typename OutputPtr>
+#if SIMDUTF_CPLUSPLUS20
+  requires(simdutf::detail::indexes_into_byte_like<InputPtr> &&
+           simdutf::detail::indexes_into_byte_like<OutputPtr>)
+#endif
+simdutf_constexpr23 size_t convert(InputPtr data, size_t len,
+                                   OutputPtr latin_output) {
+  size_t pos = 0;
+  auto start = latin_output;
+
+  while (pos < len) {
+#if SIMDUTF_CPLUSPLUS23
+    if !consteval
+#endif
+    {
+      // try to convert the next block of 16 ASCII bytes
+      if (pos + 16 <= len) { // if it is safe to read 16 more bytes, check that
+                             // they are ascii
+        uint64_t v1;
+        ::memcpy(&v1, data + pos, sizeof(uint64_t));
+        uint64_t v2;
+        ::memcpy(&v2, data + pos + sizeof(uint64_t), sizeof(uint64_t));
+        uint64_t v{v1 | v2}; // We are only interested in these bits: 1000 1000
+                             // 1000 1000 .... etc
+        if ((v & 0x8080808080808080) ==
+            0) { // if NONE of these are set, e.g. all of them are zero, then
+                 // everything is ASCII
+          size_t final_pos = pos + 16;
+          while (pos < final_pos) {
+            *latin_output++ = char(data[pos]);
+            pos++;
+          }
+          continue;
+        }
+      }
+    }
+
+    // suppose it is not an all ASCII byte sequence
+    uint8_t leading_byte = data[pos]; // leading byte
+    if (leading_byte < 0b10000000) {
+      // converting one ASCII byte !!!
+      *latin_output++ = char(leading_byte);
+      pos++;
+    } else if ((leading_byte & 0b11100000) ==
+               0b11000000) { // the first three bits indicate:
+      // We have a two-byte UTF-8
+      if (pos + 1 >= len) {
+        return 0;
+      } // minimal bound checking
+      if ((data[pos + 1] & 0b11000000) != 0b10000000) {
+        return 0;
+      } // checks if the next byte is a valid continuation byte in UTF-8. A
+        // valid continuation byte starts with 10.
+      // range check -
+      uint32_t code_point =
+          (leading_byte & 0b00011111) << 6 |
+          (data[pos + 1] &
+           0b00111111); // assembles the Unicode code point from the two bytes.
+                        // It does this by discarding the leading 110 and 10
+                        // bits from the two bytes, shifting the remaining bits
+                        // of the first byte, and then combining the results
+                        // with a bitwise OR operation.
+      if (code_point < 0x80 || 0xFF < code_point) {
+        return 0; // We only care about the range 129-255 which is Non-ASCII
+                  // latin1 characters. A code_point beneath 0x80 is invalid as
+                  // it is already covered by bytes whose leading bit is zero.
+      }
+      *latin_output++ = char(code_point);
+      pos += 2;
+    } else {
+      return 0;
+    }
+  }
+  return latin_output - start;
+}
+
+template <typename InputPtr>
+#if SIMDUTF_CPLUSPLUS20
+  requires simdutf::detail::indexes_into_byte_like<InputPtr>
+#endif
+simdutf_constexpr23 result convert_with_errors(InputPtr data, size_t len,
+                                               char *latin_output) {
+  size_t pos = 0;
+  char *start{latin_output};
+
+  while (pos < len) {
+#if SIMDUTF_CPLUSPLUS23
+    if !consteval
+#endif
+    {
+      // try to convert the next block of 16 ASCII bytes
+      if (pos + 16 <= len) { // if it is safe to read 16 more bytes, check that
+                             // they are ascii
+        uint64_t v1;
+        ::memcpy(&v1, data + pos, sizeof(uint64_t));
+        uint64_t v2;
+        ::memcpy(&v2, data + pos + sizeof(uint64_t), sizeof(uint64_t));
+        uint64_t v{v1 | v2}; // We are only interested in these bits: 1000 1000
+                             // 1000 1000...etc
+        if ((v & 0x8080808080808080) ==
+            0) { // if NONE of these are set, e.g. all of them are zero, then
+                 // everything is ASCII
+          size_t final_pos = pos + 16;
+          while (pos < final_pos) {
+            *latin_output++ = char(data[pos]);
+            pos++;
+          }
+          continue;
+        }
+      }
+    }
+    // suppose it is not an all ASCII byte sequence
+    uint8_t leading_byte = data[pos]; // leading byte
+    if (leading_byte < 0b10000000) {
+      // converting one ASCII byte !!!
+      *latin_output++ = char(leading_byte);
+      pos++;
+    } else if ((leading_byte & 0b11100000) ==
+               0b11000000) { // the first three bits indicate:
+      // We have a two-byte UTF-8
+      if (pos + 1 >= len) {
+        return result(error_code::TOO_SHORT, pos);
+      } // minimal bound checking
+      if ((data[pos + 1] & 0b11000000) != 0b10000000) {
+        return result(error_code::TOO_SHORT, pos);
+      } // checks if the next byte is a valid continuation byte in UTF-8. A
+        // valid continuation byte starts with 10.
+      // range check -
+      uint32_t code_point =
+          (leading_byte & 0b00011111) << 6 |
+          (data[pos + 1] &
+           0b00111111); // assembles the Unicode code point from the two bytes.
+                        // It does this by discarding the leading 110 and 10
+                        // bits from the two bytes, shifting the remaining bits
+                        // of the first byte, and then combining the results
+                        // with a bitwise OR operation.
+      if (code_point < 0x80) {
+        return result(error_code::OVERLONG, pos);
+      }
+      if (0xFF < code_point) {
+        return result(error_code::TOO_LARGE, pos);
+      } // We only care about the range 129-255 which is Non-ASCII latin1
+        // characters
+      *latin_output++ = char(code_point);
+      pos += 2;
+    } else if ((leading_byte & 0b11110000) == 0b11100000) {
+      // We have a three-byte UTF-8
+      return result(error_code::TOO_LARGE, pos);
+    } else if ((leading_byte & 0b11111000) == 0b11110000) { // 0b11110000
+      // we have a 4-byte UTF-8 word.
+      return result(error_code::TOO_LARGE, pos);
+    } else {
+      // we either have too many continuation bytes or an invalid leading byte
+      if ((leading_byte & 0b11000000) == 0b10000000) {
+        return result(error_code::TOO_LONG, pos);
+      }
+
+      return result(error_code::HEADER_BITS, pos);
+    }
+  }
+  return result(error_code::SUCCESS, latin_output - start);
+}
+
+inline result rewind_and_convert_with_errors(size_t prior_bytes,
+                                             const char *buf, size_t len,
+                                             char *latin1_output) {
+  size_t extra_len{0};
+  // We potentially need to go back in time and find a leading byte.
+  // In theory '3' would be sufficient, but sometimes the error can go back
+  // quite far.
+  size_t how_far_back = prior_bytes;
+  // size_t how_far_back = 3; // 3 bytes in the past + current position
+  // if(how_far_back >= prior_bytes) { how_far_back = prior_bytes; }
+  bool found_leading_bytes{false};
+  // important: it is i <= how_far_back and not 'i < how_far_back'.
+  for (size_t i = 0; i <= how_far_back; i++) {
+    unsigned char byte = buf[-static_cast<std::ptrdiff_t>(i)];
+    found_leading_bytes = ((byte & 0b11000000) != 0b10000000);
+    if (found_leading_bytes) {
+      if (i > 0 && byte < 128) {
+        // If we had to go back and the leading byte is ascii
+        // then we can stop right away.
+        return result(error_code::TOO_LONG, 0 - i + 1);
+      }
+      buf -= i;
+      extra_len = i;
+      break;
+    }
+  }
+  //
+  // It is possible for this function to return a negative count in its result.
+  // C++ Standard Section 18.1 defines size_t is in <cstddef> which is described
+  // in C Standard as <stddef.h>. C Standard Section 4.1.5 defines size_t as an
+  // unsigned integral type of the result of the sizeof operator
+  //
+  // An unsigned type will simply wrap round arithmetically (well defined).
+  //
+  if (!found_leading_bytes) {
+    // If how_far_back == 3, we may have four consecutive continuation bytes!!!
+    // [....] [continuation] [continuation] [continuation] | [buf is
+    // continuation] Or we possibly have a stream that does not start with a
+    // leading byte.
+    return result(error_code::TOO_LONG, 0 - how_far_back);
+  }
+  result res = convert_with_errors(buf, len + extra_len, latin1_output);
+  if (res.error) {
+    res.count -= extra_len;
+  }
+  return res;
+}
+
+} // namespace utf8_to_latin1
+} // unnamed namespace
+} // namespace scalar
+} // namespace simdutf
+
+#endif

package/deps/simdutf/include/simdutf/scalar/utf8_to_latin1/valid_utf8_to_latin1.h

@@ -0,0 +1,87 @@
+#ifndef SIMDUTF_VALID_UTF8_TO_LATIN1_H
+#define SIMDUTF_VALID_UTF8_TO_LATIN1_H
+
+namespace simdutf {
+namespace scalar {
+namespace {
+namespace utf8_to_latin1 {
+
+template <typename InputPtr>
+#if SIMDUTF_CPLUSPLUS20
+  requires simdutf::detail::indexes_into_byte_like<InputPtr>
+#endif
+simdutf_constexpr23 size_t convert_valid(InputPtr data, size_t len,
+                                         char *latin_output) {
+
+  size_t pos = 0;
+  char *start{latin_output};
+
+  while (pos < len) {
+#if SIMDUTF_CPLUSPLUS23
+    if !consteval
+#endif
+    {
+      // try to convert the next block of 16 ASCII bytes
+      if (pos + 16 <= len) { // if it is safe to read 16 more bytes, check that
+                             // they are ascii
+        uint64_t v1;
+        ::memcpy(&v1, data + pos, sizeof(uint64_t));
+        uint64_t v2;
+        ::memcpy(&v2, data + pos + sizeof(uint64_t), sizeof(uint64_t));
+        uint64_t v{v1 |
+                   v2}; // We are only interested in these bits: 1000 1000 1000
+                        // 1000, so it makes sense to concatenate everything
+        if ((v & 0x8080808080808080) ==
+            0) { // if NONE of these are set, e.g. all of them are zero, then
+                 // everything is ASCII
+          size_t final_pos = pos + 16;
+          while (pos < final_pos) {
+            *latin_output++ = uint8_t(data[pos]);
+            pos++;
+          }
+          continue;
+        }
+      }
+    }
+
+    // suppose it is not an all ASCII byte sequence
+    auto leading_byte = uint8_t(data[pos]); // leading byte
+    if (leading_byte < 0b10000000) {
+      // converting one ASCII byte !!!
+      *latin_output++ = char(leading_byte);
+      pos++;
+    } else if ((leading_byte & 0b11100000) ==
+               0b11000000) { // the first three bits indicate:
+      // We have a two-byte UTF-8
+      if (pos + 1 >= len) {
+        break;
+      } // minimal bound checking
+      if ((uint8_t(data[pos + 1]) & 0b11000000) != 0b10000000) {
+        return 0;
+      } // checks if the next byte is a valid continuation byte in UTF-8. A
+        // valid continuation byte starts with 10.
+      // range check -
+      uint32_t code_point =
+          (leading_byte & 0b00011111) << 6 |
+          (uint8_t(data[pos + 1]) &
+           0b00111111); // assembles the Unicode code point from the two bytes.
+                        // It does this by discarding the leading 110 and 10
+                        // bits from the two bytes, shifting the remaining bits
+                        // of the first byte, and then combining the results
+                        // with a bitwise OR operation.
+      *latin_output++ = char(code_point);
+      pos += 2;
+    } else {
+      // we may have a continuation but we do not do error checking
+      return 0;
+    }
+  }
+  return latin_output - start;
+}
+
+} // namespace utf8_to_latin1
+} // unnamed namespace
+} // namespace scalar
+} // namespace simdutf
+
+#endif