#include "simdutf.h"

#include <array>
#include <vector>

#include <tests/reference/encode_utf16.h>
#include <tests/helpers/random_int.h>
#include <tests/helpers/random_utf8.h>
#include <tests/helpers/random_utf16.h>
#include <tests/helpers/test.h>
#include <tests/helpers/utf16.h>

namespace {
std::array<size_t, 7> input_size{8, 16, 12, 64, 68, 128, 256};

template <typename Pointer> bool has_bom(Pointer *data, size_t size) {
  return (simdutf::BOM::check_bom(data, size) !=
          simdutf::encoding_type::unspecified);
}

template <typename RandomGenerator>
std::vector<uint8_t> generate_utf8(RandomGenerator random, size_t size) {
  // This is quite ugly workaround, but we expect almost none repeats
  while (true) {
    const auto generated = random.generate(size);
    if (!has_bom(generated.data(), generated.size())) {
      return generated;
    }
  }
}

template <typename RandomGenerator>
std::vector<char16_t> generate_utf16_le(RandomGenerator random, size_t size) {
  // This is quite ugly workaround, but we expect almost none repeats
  while (true) {
    const auto generated = random.generate_le(size);
    if (!has_bom(reinterpret_cast<const char *>(generated.data()),
                 generated.size())) {
      return generated;
    }
  }
}

template <typename RandomGenerator>
std::vector<uint16_t> generate_u16(RandomGenerator &random, size_t count) {
  std::vector<uint16_t> result;
  result.reserve(count);

  union {
    uint16_t word[2];
    char string[4];
  } first;

  static_assert(sizeof(first) == 4, "union got an unexpected size");

  // Make sure our random data does not start with a BOM marker.
  do {
    first.word[0] = random();
    first.word[1] = random();
  } while (has_bom(first.string, 4));

  result.push_back(first.word[0]);
  result.push_back(first.word[1]);
  for (size_t i = 2; i < count; i++) {
    result.push_back(random());
  }

  return result;
}

template <typename RandomGenerator>
std::vector<uint32_t> generate_u32(RandomGenerator &random, size_t count) {
  std::vector<uint32_t> result;
  result.reserve(count);

  union {
    uint32_t word;
    char string[4];
  } first;

  static_assert(sizeof(first) == 4, "union got an unexpected size");

  // Make sure our random data does not start with a BOM marker.
  do {
    first.word = random();
  } while (has_bom(first.string, 4));

  result.push_back(first.word);
  for (size_t i = 1; i < count; i++) {
    result.push_back(random());
  }

  return result;
}

} // namespace

TEST(issue_519) {
  // implementation icelake gave 0
  // implementation haswell gave 0
  // implementation westmere gave 0
  // implementation fallback gave 1
  std::vector<unsigned char> data{
      32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32,  32,  32,
      32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 223, 164, 32,
      32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32,  32,  32,
      32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32,  32,  32,
  };
  const auto r =
      implementation.detect_encodings((const char *)data.data(), data.size());
  ASSERT_EQUAL(r, 1);
}

TEST(issue_516) {
  std::vector<unsigned char> data{0x20, 0xd8, 0x00, 0x00};
  const auto r =
      implementation.detect_encodings((const char *)data.data(), data.size());
  ASSERT_EQUAL(r, 0);
}

TEST(issue818) {
  std::string data = "\xEF\xBB\xBF";
  const auto r = simdutf::BOM::check_bom(data.data(), data.size());
  ASSERT_EQUAL(r, simdutf::encoding_type::UTF8);
}

TEST(boommmmm) {
  const char *utf8_bom = "\xef\xbb\xbf";
  const char *utf16be_bom = "\xfe\xff";
  const char *utf16le_bom = "\xff\xfe";
  ASSERT_EQUAL(implementation.detect_encodings(utf8_bom, 3),
               simdutf::encoding_type::UTF8);
  ASSERT_EQUAL(implementation.detect_encodings(utf16be_bom, 2),
               simdutf::encoding_type::UTF16_BE);
  ASSERT_EQUAL(implementation.detect_encodings(utf16le_bom, 2),
               simdutf::encoding_type::UTF16_LE);
}

#if !SIMDUTF_IS_BIG_ENDIAN
TEST(issue_627) {
  std::vector<unsigned char> data{
      251, 219, 37,  222, 0,   199, 218, 0,   157, 0,   0, 255, 8,   8,
      8,   8,   227, 8,   8,   8,   8,   8,   8,   255, 0, 0,   248, 0,
      255, 8,   8,   8,   8,   10,  8,   8,   8,   8,   8, 8,   8,   0,
      219, 0,   0,   122, 0,   255, 0,   5,   0,   0,   0, 0,   0,   0,
      0,   255, 243, 159, 129, 172, 1,   219, 37,  222,
  };
  const auto r1 =
      implementation.detect_encodings((const char *)data.data(), data.size());
  ASSERT_EQUAL(r1, 2);

  std::vector<unsigned char> data2{
      0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,   0, 0,
      0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,   0, 0,
      0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 219,
  };
  const auto r2 =
      implementation.detect_encodings((const char *)data2.data(), data2.size());
  ASSERT_EQUAL(r2, 0);
}
#endif

TEST_LOOP(pure_utf8_ASCII) {
  simdutf::tests::helpers::random_utf8 random(seed, 1, 0, 0, 0);

  for (size_t size : input_size) {
    const auto generated = generate_utf8(random, size);
    auto expected =
        simdutf::encoding_type::UTF8 | simdutf::encoding_type::UTF16_LE;
    auto actual = implementation.detect_encodings(
        reinterpret_cast<const char *>(generated.data()), size);
    ASSERT_EQUAL(actual, expected);
  }
}

TEST_LOOP(pure_utf16_ASCII) {
  simdutf::tests::helpers::RandomInt random(0, 127, seed);

  for (size_t size : input_size) {
    const auto generated = generate_u16(random, size / 2);
    auto expected =
        simdutf::encoding_type::UTF8 | simdutf::encoding_type::UTF16_LE;
    auto actual = implementation.detect_encodings(
        reinterpret_cast<const char *>(generated.data()), size);
    ASSERT_TRUE((actual & expected) ==
                expected); // Must be at least UTF8 and UTF16_LE.
  }
}

TEST_LOOP(pure_utf32_ASCII) {
  simdutf::tests::helpers::RandomInt random(0, 0x7f, seed);

  for (size_t size : input_size) {
    const auto generated = generate_u32(random, size / 4);
    auto expected = simdutf::encoding_type::UTF8 |
                    simdutf::encoding_type::UTF16_LE |
                    simdutf::encoding_type::UTF32_LE;
    auto actual = implementation.detect_encodings(
        reinterpret_cast<const char *>(generated.data()), size);
    ASSERT_TRUE((actual & expected) ==
                expected); // Must be at least UTF8 and UTF16_LE and UTF32_LE.
  }
}

#if 0 // XXX
TEST_LOOP(no_utf8_bytes_no_surrogates) {
  simdutf::tests::helpers::RandomIntRanges random(
      {{0x007f, 0xd800 - 1}, {0xe000, 0xffff}}, seed);

  for (size_t size : input_size) {
    auto generated = generate_u32(random, size / 4);
    if (!simdutf::match_system(simdutf::endianness::LITTLE)) {
        for (auto& val: generated) {
            val = ((val & 0xff) << 8) | (val >> 8);
        }
    }

    auto expected =
        simdutf::encoding_type::UTF16_LE | simdutf::encoding_type::UTF32_LE;
    auto actual = implementation.detect_encodings(
        reinterpret_cast<const char *>(generated.data()), size);
    ASSERT_EQUAL((actual & expected),
                 expected); // Must be at least UTF16_LE and UTF32_LE.
  }
}
#endif

TEST_LOOP(two_utf8_bytes) {
  simdutf::tests::helpers::random_utf8 random(seed, 0, 1, 0, 0);

  for (size_t size : input_size) {
    const auto generated = generate_utf8(random, size);
    auto expected =
        simdutf::encoding_type::UTF8 | simdutf::encoding_type::UTF16_LE;
    auto actual = implementation.detect_encodings(
        reinterpret_cast<const char *>(generated.data()), size);
    if (actual != expected) {
      if ((actual & simdutf::encoding_type::UTF8) == 0) {
        puts("failed to detect valid UTF-8.");
      }
      if ((actual & simdutf::encoding_type::UTF16_LE) == 0) {
        puts("failed to detect valid UTF-16LE.");
      }
    }
    ASSERT_TRUE((actual & expected) ==
                expected); // Must be at least UTF8 and UTF16_LE.
  }
}

TEST_LOOP(utf_16_surrogates) {
  simdutf::tests::helpers::random_utf16 random(seed, 0, 1);

  for (size_t size : input_size) {
    const auto generated = generate_utf16_le(random, size / 2);
    auto expected = simdutf::encoding_type::UTF16_LE;
    auto actual = implementation.detect_encodings(
        reinterpret_cast<const char *>(generated.data()), size);
    ASSERT_TRUE((actual & expected) == expected); // Must be at least UTF16_LE.
  }
}

TEST_LOOP(utf32_surrogates) {
  simdutf::tests::helpers::RandomInt random_prefix(0x10000, 0x10ffff, seed);
  simdutf::tests::helpers::RandomInt random_suffix(0xd800, 0xdfff, seed);

  for (size_t size : input_size) {
    std::vector<uint32_t> generated;
    for (unsigned int i = 0; i < size / 4; i++) {
      generated.push_back((random_prefix() & 0xffff0000) + random_suffix());
    }
    auto expected = simdutf::encoding_type::UTF32_LE;
    auto actual = implementation.detect_encodings(
        reinterpret_cast<const char *>(generated.data()), size);
    ASSERT_TRUE((actual & expected) == expected); // Must be at least UTF32_LE.
  }
}

TEST_LOOP(edge_surrogate) {
  simdutf::tests::helpers::RandomInt random(0x10000, 0x10ffff, seed);

  const size_t size = 512;
  std::vector<uint16_t> generated(size / 2, 0);
  unsigned int i = 31;
  while (i + 32 < (size / 2) - 1) {
    char16_t W1;
    char16_t W2;
    ASSERT_EQUAL(simdutf::tests::reference::utf16::encode(random(), W1, W2), 2);
    generated[i + 0] = to_utf16le(W1);
    generated[i + 1] = to_utf16le(W2);
    i += 32;
  }
  auto expected = simdutf::encoding_type::UTF16_LE;
  auto actual = implementation.detect_encodings(
      reinterpret_cast<const char *>(generated.data()), size);
  ASSERT_TRUE((actual & expected) == expected); // Must be at least UTF16_LE.
}

TEST_LOOP(tail_utf8) {
  simdutf::tests::helpers::random_utf8 random(seed, 0, 0, 1, 0);
  std::array<size_t, 5> multiples_three{12, 54, 66, 126, 252};
  for (size_t size : multiples_three) {
    const auto generated = generate_utf8(random, size);
    auto expected =
        simdutf::encoding_type::UTF8 | simdutf::encoding_type::UTF16_LE;
    auto actual = implementation.detect_encodings(
        reinterpret_cast<const char *>(generated.data()), size);
    ASSERT_TRUE((actual & expected) ==
                expected); // Must be at least UTF8 and UTF16_LE.
  }
}

TEST_MAIN