using System;
using Unity.Collections.LowLevel.Unsafe;

namespace Unity.Collections
{
    /// <summary>
    /// Kinds of format errors.
    /// </summary>
    public enum FormatError
    {
        /// <summary>
        /// No error.
        /// </summary>
        None,

        /// <summary>
        /// The target storage does not have sufficient capacity.
        /// Note that the format's write failed. It did not truncate.
        /// </summary>
        Overflow,

        /// <summary>
        /// The source format specifier is not itself correctly formatted, or
        /// a format specifier tokens were found outside of accepted usage.
        /// Note that the format's write failed.
        /// </summary>
        BadFormatSpecifier,
    }

    /// <summary>
    /// Kinds of parse errors.
    /// </summary>
    public enum ParseError
    {
        /// <summary>
        /// No parse error.
        /// </summary>
        None,

        /// <summary>
        /// The text parsed does not form a number.
        /// </summary>
        Syntax,

        /// <summary>
        /// The number exceeds the range of the target type.
        /// The number was either truncated, or failed to write entirely.
        /// </summary>
        Overflow,

        /// <summary>
        /// The number exceeds the precision of the target type.
        /// </summary>
        Underflow,
    }

    /// <summary>
    /// Kinds of copy errors.
    /// </summary>
    public enum CopyError
    {
        /// <summary>
        /// No copy error.
        /// </summary>
        None,

        /// <summary>
        /// The target storage does not have sufficient capacity.
        /// Unless stated in the API comment, assume that the write operation was partially applied.
        /// </summary>
        Truncation,
    }

    /// <summary>
    /// Kinds of conversion errors.
    /// </summary>
    public enum ConversionError
    {
        /// <summary>
        /// No conversion error.
        /// </summary>
        None,

        /// <summary>
        /// The target storage does not have sufficient capacity.
        /// For copy operations; the value was either truncated into the target storage, or failed to write entirely.
        /// </summary>
        Overflow,

        /// <summary>
        /// The bytes do not form a valid character.
        /// </summary>
        Encoding,

        /// <summary>
        /// The rune is not a valid code point.
        /// </summary>
        CodePoint,
    }

    /// <summary>
    /// Provides utility methods for UTF-8, UTF-16, UCS-4 (a.k.a. UTF-32), and WTF-8.
    /// </summary>
    [GenerateTestsForBurstCompatibility]
    public unsafe struct Unicode
    {
        /// <summary>
        /// Representation of a Unicode character as a code point.
        /// </summary>
        [GenerateTestsForBurstCompatibility]
        public struct Rune
        {
            /// <summary>
            /// The code point.
            /// </summary>
            /// <value>The code point.</value>
            public int value;

            /// <summary>
            /// Initializes and returns an instance of Rune.
            /// </summary>
            /// <remarks>You are responsible for the code point being valid.</remarks>
            /// <param name="codepoint">The code point.</param>
            public Rune(int codepoint)
            {
                value = codepoint;
            }

            /// <summary>
            /// Returns a rune.
            /// </summary>
            /// <remarks>Because a char is 16-bit, it can only represent the first 2^16 code points, not all 1.1 million.</remarks>
            /// <param name="codepoint">A code point.</param>
            /// <returns>A rune.</returns>
            public static implicit operator Rune(char codepoint) => new Rune { value = codepoint };

            /// <summary>
            /// Evaluates if one is equal to the other.
            /// </summary>
            /// <param name="lhs">The left-hand side</param>
            /// <param name="rhs">The right-hand side</param>
            /// <returns>True if the left-hand side's is equal to the right-hand side's.</returns>
            public static bool operator ==(Rune lhs, Rune rhs)
            {
                return lhs.value == rhs.value;
            }

            /// <summary>
            /// Returns true if the value stored in this Rune is equal to an object.
            /// </summary>
            /// <remarks>Can only be equal if the object is itself a Rune.</remarks>
            /// <param name="obj">An object to compare with.</param>
            /// <returns>True if the value stored in this Rune is equal to the object.</returns>
            [ExcludeFromBurstCompatTesting("Takes managed object")]
            public override bool Equals(object obj)
            {
                if (obj is Rune)
                {
                    return value == ((Rune)obj).value;
                }

                return false;
            }

            /// <summary>
            /// A hash used for comparisons.
            /// </summary>
            /// <returns>A unique hash code.</returns>
            public override int GetHashCode()
            {
                return value;
            }

            /// <summary>
            /// Evaluates if one is not equal to the other.
            /// </summary>
            /// <param name="lhs">The left-hand side</param>
            /// <param name="rhs">The right-hand side</param>
            /// <returns>True if the left-hand side's is not equal to the right-hand side's.</returns>
            public static bool operator !=(Rune lhs, Rune rhs)
            {
                return lhs.value != rhs.value;
            }

            /// <summary>
            /// Returns true if a rune is a numerical digit character.
            /// </summary>
            /// <param name="r">The rune.</param>
            /// <returns>True if the rune is a numerical digit character.</returns>
            public static bool IsDigit(Rune r)
            {
                return r.IsDigit();
            }

            internal bool IsAscii()
            {
                return value < 0x80;
            }

            internal bool IsLatin1()
            {
                return value < 0x100;
            }

            internal bool IsDigit()
            {
                return value >= '0' && value <= '9';
            }

            internal bool IsWhiteSpace()
            {
                // https://en.wikipedia.org/wiki/Whitespace_character#Unicode

                if (IsLatin1())
                {
                    return value == ' '
                        || (value >= 0x9 && value <= 0xD) // CHARACTER TABULATION (U+0009), LINE FEED (U+000A), LINE TABULATION (U+000B), FORM FEED (U+000C), CARRIAGE RETURN (U+000D)
                        || value == 0xA0 // NO-BREAK SPACE
                        || value == 0x85 // NEXT LINE
                        ;
                }

                return value == 0x1680 // OGHAM SPACE MARK
                    || (value >= 0x2000 && value <= 0x200A) // EN QUAD(U+2000)
                                                            // EM QUAD(U+2001)
                                                            // EN SPACE(U+2002)
                                                            // EM SPACE(U+2003)
                                                            // THREE - PER - EM SPACE(U + 2004)
                                                            // FOUR - PER - EM SPACE(U + 2005)
                                                            // SIX - PER - EM SPACE(U + 2006)
                                                            // FIGURE SPACE(U+2007)
                                                            // PUNCTUATION SPACE(U+2008)
                                                            // THIN SPACE(U+2009)
                                                            // HAIR SPACE(U+200A)
                    || value == 0x2028 // LINE SEPARATOR
                    || value == 0x2029 // PARAGRAPH SEPARATOR
                    || value == 0x202F // NARROW NO-BREAK SPACE
                    || value == 0x205F // MEDIUM MATHEMATICAL SPACE
                    || value == 0x3000 // IDEOGRAPHIC SPACE
                    ;
            }

            internal Rune ToLowerAscii()
            {
                return new Rune(value + (((uint)(value - 'A') <= ('Z' - 'A')) ? 0x20 : 0));
            }

            internal Rune ToUpperAscii()
            {
                return new Rune(value - (((uint)(value - 'a') <= ('z' - 'a')) ? 0x20 : 0));
            }

            /// <summary>
            /// Returns the number of bytes required to encode this rune as UTF-8.
            /// </summary>
            /// <returns>The number of bytes required to encode this rune as UTF-8. If the rune's codepoint
            /// is invalid, returns 4 (the maximum possible encoding length).</returns>
            public int LengthInUtf8Bytes()
            {
                if (value < 0)
                    return 4; // invalid codepoint
                if (value <= 0x7F)
                    return 1;
                if (value <= 0x7FF)
                    return 2;
                if (value <= 0xFFFF)
                    return 3;
                if (value <= 0x1FFFFF)
                    return 4;
                // invalid codepoint, max size.
                return 4;
            }
        }

        /// <summary>The maximum value of a valid UNICODE code point</summary>
        public const int kMaximumValidCodePoint = 0x10FFFF;

        /// <summary>
        /// Returns true if a code point is valid.
        /// </summary>
        /// <param name="codepoint">A code point.</param>
        /// <returns>True if a code point is valid.</returns>
        public static bool IsValidCodePoint(int codepoint)
        {
            if (codepoint > kMaximumValidCodePoint) // maximum valid code point
                return false;
//            if (codepoint >= 0xD800 && codepoint <= 0xDFFF) // surrogate pair
//                return false;
            if (codepoint < 0) // negative?
                return false;
            return true;
        }

        /// <summary>
        /// Returns true if the byte is not the last byte of a UTF-8 character.
        /// </summary>
        /// <param name="b">The byte.</param>
        /// <returns>True if the byte is not the last byte of a UTF-8 character.</returns>
        public static bool NotTrailer(byte b)
        {
            return (b & 0xC0) != 0x80;
        }

        /// <summary>
        /// The Unicode character �.
        /// </summary>
        /// <remarks>This character is used to stand-in for characters that can't be rendered.</remarks>
        /// <value>The Unicode character �.</value>
        public static Rune ReplacementCharacter => new Rune { value = 0xFFFD };

        /// <summary>
        /// The null rune value.
        /// </summary>
        /// <remarks>In this package, the "bad rune" is used as a null character. It represents no valid code point.</remarks>
        /// <value>The null rune value.</value>
        public static Rune BadRune => new Rune { value = 0 };

        /// <summary>
        /// Reads a UTF-8 encoded character from a buffer.
        /// </summary>
        /// <param name="rune">Outputs the character read. If the read fails, outputs <see cref="ReplacementCharacter"/>.</param>
        /// <param name="buffer">The buffer of bytes to read.</param>
        /// <param name="index">Reference to a byte index into the buffer. If the read succeeds, index is incremented by the
        /// size in bytes of the character read. If the read fails, index is incremented by 1.</param>
        /// <param name="capacity">The size in bytes of the buffer. Used to check that the read is in bounds.</param>
        /// <returns><see cref="ConversionError.None"/> if the read succeeds. Otherwise, returns <see cref="ConversionError.Overflow"/> or <see cref="ConversionError.Encoding"/>.</returns>
        public static ConversionError Utf8ToUcs(out Rune rune, byte* buffer, ref int index, int capacity)
        {
            int code = 0;
            rune = ReplacementCharacter;
            if (index + 1 > capacity)
            {
                return ConversionError.Overflow;
            }

            if ((buffer[index] & 0b10000000) == 0b00000000) // if high bit is 0, 1 byte
            {
                rune.value = buffer[index + 0];
                index += 1;
                return ConversionError.None;
            }

            if ((buffer[index] & 0b11100000) == 0b11000000) // if high 3 bits are 110, 2 bytes
            {
                if (index + 2 > capacity)
                {
                    index += 1;
                    return ConversionError.Overflow;
                }
                code = (buffer[index + 0] & 0b00011111);
                code = (code << 6) | (buffer[index + 1] & 0b00111111);
                if (code < (1 << 7) || NotTrailer(buffer[index + 1]))
                {
                    index += 1;
                    return ConversionError.Encoding;
                }
                rune.value = code;
                index += 2;
                return ConversionError.None;
            }

            if ((buffer[index] & 0b11110000) == 0b11100000) // if high 4 bits are 1110, 3 bytes
            {
                if (index + 3 > capacity)
                {
                    index += 1;
                    return ConversionError.Overflow;
                }
                code = (buffer[index + 0] & 0b00001111);
                code = (code << 6) | (buffer[index + 1] & 0b00111111);
                code = (code << 6) | (buffer[index + 2] & 0b00111111);
                if (code < (1 << 11) || !IsValidCodePoint(code) || NotTrailer(buffer[index + 1]) || NotTrailer(buffer[index + 2]))
                {
                    index += 1;
                    return ConversionError.Encoding;
                }
                rune.value = code;
                index += 3;
                return ConversionError.None;
            }

            if ((buffer[index] & 0b11111000) == 0b11110000) // if high 5 bits are 11110, 4 bytes
            {
                if (index + 4 > capacity)
                {
                    index += 1;
                    return ConversionError.Overflow;
                }
                code = (buffer[index + 0] & 0b00000111);
                code = (code << 6) | (buffer[index + 1] & 0b00111111);
                code = (code << 6) | (buffer[index + 2] & 0b00111111);
                code = (code << 6) | (buffer[index + 3] & 0b00111111);
                if (code < (1 << 16) || !IsValidCodePoint(code) || NotTrailer(buffer[index + 1]) || NotTrailer(buffer[index + 2]) || NotTrailer(buffer[index + 3]))
                {
                    index += 1;
                    return ConversionError.Encoding;
                }
                rune.value = code;
                index += 4;
                return ConversionError.None;
            }

            index += 1;
            return ConversionError.Encoding;
        }

        static int FindUtf8CharStartInReverse(byte* ptr, ref int index)
        {
            do
            {
                if (index <= 0)
                {
                    return 0;
                }

                --index;

            } while ((ptr[index] & 0xC0) == 0x80);

            return index;
        }

        internal static ConversionError Utf8ToUcsReverse(out Rune rune, byte* buffer, ref int index, int capacity)
        {
            var prev = index;
            --index;

            index = FindUtf8CharStartInReverse(buffer, ref index);

            if (index == prev)
            {
                rune = ReplacementCharacter;
                return ConversionError.Overflow;
            }

            var ignore = index;
            return Utf8ToUcs(out rune, buffer, ref ignore, capacity);
        }

        /// <summary>
        /// Returns true if a char is a Unicode leading surrogate.
        /// </summary>
        /// <param name="c">The char.</param>
        /// <returns>True if the char is a Unicode leading surrogate.</returns>
        static bool IsLeadingSurrogate(char c)
        {
            return c >= 0xD800 && c <= 0xDBFF;
        }

        /// <summary>
        /// Returns true if a char is a Unicode trailing surrogate.
        /// </summary>
        /// <param name="c">The char.</param>
        /// <returns>True if the char is a Unicode trailing surrogate.</returns>
        static bool IsTrailingSurrogate(char c)
        {
            return c >= 0xDC00 && c <= 0xDFFF;
        }

        /// <summary>
        /// Reads a UTF-16 encoded character from a buffer.
        /// </summary>
        /// <param name="rune">Outputs the character read. If the read fails, rune is not set.</param>
        /// <param name="buffer">The buffer of chars to read.</param>
        /// <param name="index">Reference to a char index into the buffer. If the read succeeds, index is incremented by the
        /// size in chars of the character read. If the read fails, index is not incremented.</param>
        /// <param name="capacity">The size in chars of the buffer. Used to check that the read is in bounds.</param>
        /// <returns><see cref="ConversionError.None"/> if the read succeeds. Otherwise, returns <see cref="ConversionError.Overflow"/>.</returns>
        public static ConversionError Utf16ToUcs(out Rune rune, char* buffer, ref int index, int capacity)
        {
            int code = 0;
            rune = ReplacementCharacter;
            if (index + 1 > capacity)
                return ConversionError.Overflow;
            if (!IsLeadingSurrogate(buffer[index]) || (index + 2 > capacity))
            {
                rune.value = buffer[index];
                index += 1;
                return ConversionError.None;
            }
            code =                (buffer[index + 0] & 0x03FF);
            char next = buffer[index + 1];
            if (!IsTrailingSurrogate(next))
            {
                rune.value = buffer[index];
                index += 1;
                return ConversionError.None;
            }
            code = (code << 10) | (buffer[index + 1] & 0x03FF);
            code += 0x10000;
            rune.value = code;
            index += 2;
            return ConversionError.None;
        }

        internal static ConversionError UcsToUcs(out Rune rune, Rune* buffer, ref int index, int capacity)
        {
            rune = ReplacementCharacter;
            if (index + 1 > capacity)
                return ConversionError.Overflow;
            rune = buffer[index];
            index += 1;
            return ConversionError.None;
        }

        /// <summary>
        /// Writes a rune to a buffer as a UTF-8 encoded character.
        /// </summary>
        /// <param name="rune">The rune to encode.</param>
        /// <param name="buffer">The buffer to write to.</param>
        /// <param name="index">Reference to a byte index into the buffer. If the write succeeds, index is incremented by the
        /// size in bytes of the character written. If the write fails, index is not incremented.</param>
        /// <param name="capacity">The size in bytes of the buffer. Used to check that the write is in bounds.</param>
        /// <returns><see cref="ConversionError.None"/> if the write succeeds. Otherwise, returns <see cref="ConversionError.CodePoint"/>, <see cref="ConversionError.Overflow"/>, or <see cref="ConversionError.Encoding"/>.</returns>
        public static ConversionError UcsToUtf8(byte* buffer, ref int index, int capacity, Rune rune)
        {
            if (!IsValidCodePoint(rune.value))
            {
                return ConversionError.CodePoint;
            }

            if (index + 1 > capacity)
            {
                return ConversionError.Overflow;
            }

            if (rune.value <= 0x7F)
            {
                buffer[index++] = (byte)rune.value;
                return ConversionError.None;
            }

            if (rune.value <= 0x7FF)
            {
                if (index + 2 > capacity)
                {
                    return ConversionError.Overflow;
                }

                buffer[index++] = (byte)(0xC0 | (rune.value >> 6));
                buffer[index++] = (byte)(0x80 | ((rune.value >> 0) & 0x3F));
                return ConversionError.None;
            }

            if (rune.value <= 0xFFFF)
            {
                if (index + 3 > capacity)
                {
                    return ConversionError.Overflow;
                }

                buffer[index++] = (byte)(0xE0 | (rune.value >> 12));
                buffer[index++] = (byte)(0x80 | ((rune.value >> 6) & 0x3F));
                buffer[index++] = (byte)(0x80 | ((rune.value >> 0) & 0x3F));
                return ConversionError.None;
            }

            if (rune.value <= 0x1FFFFF)
            {
                if (index + 4 > capacity)
                {
                    return ConversionError.Overflow;
                }

                buffer[index++] = (byte)(0xF0 | (rune.value >> 18));
                buffer[index++] = (byte)(0x80 | ((rune.value >> 12) & 0x3F));
                buffer[index++] = (byte)(0x80 | ((rune.value >> 6) & 0x3F));
                buffer[index++] = (byte)(0x80 | ((rune.value >> 0) & 0x3F));
                return ConversionError.None;
            }

            return ConversionError.Encoding;
        }

        /// <summary>
        /// Writes a rune to a buffer as a UTF-16 encoded character.
        /// </summary>
        /// <param name="rune">The rune to encode.</param>
        /// <param name="buffer">The buffer of chars to write to.</param>
        /// <param name="index">Reference to a char index into the buffer. If the write succeeds, index is incremented by the
        /// size in chars of the character written. If the write fails, index is not incremented.</param>
        /// <param name="capacity">The size in chars of the buffer. Used to check that the write is in bounds.</param>
        /// <returns><see cref="ConversionError.None"/> if the write succeeds. Otherwise, returns <see cref="ConversionError.CodePoint"/>, <see cref="ConversionError.Overflow"/>, or <see cref="ConversionError.Encoding"/>.</returns>
        public static ConversionError UcsToUtf16(char* buffer, ref int index, int capacity, Rune rune)
        {
            if (!IsValidCodePoint(rune.value))
            {
                return ConversionError.CodePoint;
            }

            if (index + 1 > capacity)
            {
                return ConversionError.Overflow;
            }

            if (rune.value >= 0x10000)
            {
                if (index + 2 > capacity)
                {
                    return ConversionError.Overflow;
                }

                int code = rune.value - 0x10000;
                if (code >= (1 << 20))
                {
                    return ConversionError.Encoding;
                }

                buffer[index++] = (char)(0xD800 | (code >> 10));
                buffer[index++] = (char)(0xDC00 | (code & 0x3FF));
                return ConversionError.None;
            }

            buffer[index++] = (char)rune.value;
            return ConversionError.None;
        }

        /// <summary>
        /// Copies UTF-16 characters from one buffer to another buffer as UTF-8.
        /// </summary>
        /// <remarks>Assumes the source data is valid UTF-16.</remarks>
        /// <param name="utf16Buffer">The source buffer.</param>
        /// <param name="utf16Length">The number of chars to read from the source.</param>
        /// <param name="utf8Buffer">The destination buffer.</param>
        /// <param name="utf8Length">Outputs the number of bytes written to the destination.</param>
        /// <param name="utf8Capacity">The size in bytes of the destination buffer.</param>
        /// <returns><see cref="ConversionError.None"/> if the copy fully completes. Otherwise, returns <see cref="ConversionError.Overflow"/>.</returns>
        public static ConversionError Utf16ToUtf8(char* utf16Buffer, int utf16Length, byte* utf8Buffer, out int utf8Length, int utf8Capacity)
        {
            utf8Length = 0;
            for (var utf16Offset = 0; utf16Offset < utf16Length;)
            {
                Utf16ToUcs(out var ucs, utf16Buffer, ref utf16Offset, utf16Length);
                if (UcsToUtf8(utf8Buffer, ref utf8Length, utf8Capacity, ucs) == ConversionError.Overflow)
                    return ConversionError.Overflow;
            }
            return ConversionError.None;
        }

        /// <summary>
        /// Copies UTF-8 characters from one buffer to another.
        /// </summary>
        /// <remarks>Assumes the source data is valid UTF-8.</remarks>
        /// <param name="srcBuffer">The source buffer.</param>
        /// <param name="srcLength">The number of bytes to read from the source.</param>
        /// <param name="destBuffer">The destination buffer.</param>
        /// <param name="destLength">Outputs the number of bytes written to the destination.</param>
        /// <param name="destCapacity">The size in bytes of the destination buffer.</param>
        /// <returns><see cref="ConversionError.None"/> if the copy fully completes. Otherwise, returns <see cref="ConversionError.Overflow"/>.</returns>
        public static ConversionError Utf8ToUtf8(byte* srcBuffer, int srcLength, byte* destBuffer, out int destLength, int destCapacity)
        {
            if (destCapacity >= srcLength)
            {
                UnsafeUtility.MemCpy(destBuffer, srcBuffer, srcLength);
                destLength = srcLength;
                return ConversionError.None;
            }
            // TODO even in this case, it's possible to MemCpy all but the last 3 bytes that fit, and then by looking at only
            // TODO the high bits of the last 3 bytes that fit, decide how many of the 3 to append. but that requires a
            // TODO little UNICODE presence of mind that nobody has today.
            destLength = 0;
            for (var srcOffset = 0; srcOffset < srcLength;)
            {
                Utf8ToUcs(out var ucs, srcBuffer, ref srcOffset, srcLength);
                if (UcsToUtf8(destBuffer, ref destLength, destCapacity, ucs) == ConversionError.Overflow)
                    return ConversionError.Overflow;
            }
            return ConversionError.None;
        }

        /// <summary>
        /// Copies UTF-8 characters from one buffer to another as UTF-16.
        /// </summary>
        /// <remarks>Assumes the source data is valid UTF-8.</remarks>
        /// <param name="utf8Buffer">The source buffer.</param>
        /// <param name="utf8Length">The number of bytes to read from the source.</param>
        /// <param name="utf16Buffer">The destination buffer.</param>
        /// <param name="utf16Length">Outputs the number of chars written to the destination.</param>
        /// <param name="utf16Capacity">The size in chars of the destination buffer.</param>
        /// <returns><see cref="ConversionError.None"/> if the copy fully completes. Otherwise, <see cref="ConversionError.Overflow"/>.</returns>
        public static ConversionError Utf8ToUtf16(byte* utf8Buffer, int utf8Length, char* utf16Buffer, out int utf16Length, int utf16Capacity)
        {
            utf16Length = 0;
            for (var utf8Offset
                = 0; utf8Offset < utf8Length;)
            {
                Utf8ToUcs(out var ucs, utf8Buffer, ref utf8Offset, utf8Length);
                if (UcsToUtf16(utf16Buffer, ref utf16Length, utf16Capacity, ucs) == ConversionError.Overflow)
                    return ConversionError.Overflow;
            }
            return ConversionError.None;
        }

        static int CountRunes(byte* utf8Buffer, int utf8Length, int maxRunes = int.MaxValue)
        {
            var numRunes = 0;

            for (var i = 0; numRunes < maxRunes && i < utf8Length; ++i)
            {
                if ((utf8Buffer[i] & 0xC0) != 0x80)
                    numRunes++;
            }

            return numRunes;
        }
    }
}