/*
  ==============================================================================

   This file is part of the JUCE library.
   Copyright (c) 2015 - ROLI Ltd.

   Permission is granted to use this software under the terms of either:
   a) the GPL v2 (or any later version)
   b) the Affero GPL v3

   Details of these licenses can be found at: www.gnu.org/licenses

   JUCE is distributed in the hope that it will be useful, but WITHOUT ANY
   WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
   A PARTICULAR PURPOSE.  See the GNU General Public License for more details.

   ------------------------------------------------------------------------------

   To release a closed-source product which uses JUCE, commercial licenses are
   available: visit www.juce.com for more information.

  ==============================================================================
*/

#ifndef JUCE_CPLUSPLUSCODETOKENISERFUNCTIONS_H_INCLUDED
#define JUCE_CPLUSPLUSCODETOKENISERFUNCTIONS_H_INCLUDED


//==============================================================================
/** Class containing some basic functions for simple tokenising of C++ code.
*/
struct CppTokeniserFunctions
{
    static bool isIdentifierStart (const juce_wchar c) noexcept
    {
        return CharacterFunctions::isLetter (c)
                || c == '_' || c == '@';
    }

    static bool isIdentifierBody (const juce_wchar c) noexcept
    {
        return CharacterFunctions::isLetterOrDigit (c)
                || c == '_' || c == '@';
    }

    static bool isReservedKeyword (String::CharPointerType token, const int tokenLength) noexcept
    {
        static const char* const keywords2Char[] =
            { "if", "do", "or", nullptr };

        static const char* const keywords3Char[] =
            { "for", "int", "new", "try", "xor", "and", "asm", "not", nullptr };

        static const char* const keywords4Char[] =
            { "bool", "void", "this", "true", "long", "else", "char",
              "enum", "case", "goto", "auto", nullptr };

        static const char* const keywords5Char[] =
            { "float", "const", "while", "break", "false", "catch", "class", "bitor",
              "compl", "or_eq", "short", "throw", "union", "using", "final", nullptr };

        static const char* const keywords6Char[] =
            { "return", "and_eq", "bitand", "delete", "double", "export", "extern",
              "friend", "inline", "not_eq", "public", "signed", "sizeof", "static",
              "struct", "switch", "typeid", "xor_eq", nullptr };

        static const char* const keywords7Char[] =
            { "nullptr", "alignas", "alignof", "default", "mutable", "private",
              "typedef", "virtual", "wchar_t", "__cdecl", "_Pragma", "uint8_t", nullptr };

        static const char* const keywordsOther[] =
            { "char16_t", "char32_t", "const_cast", "constexpr", "continue", "decltype", "dynamic_cast",
              "explicit", "namespace", "noexcept", "operator", "protected", "register", "reinterpret_cast",
              "static_assert", "static_cast", "template", "thread_local", "typename", "unsigned", "volatile",
              "@class", "@dynamic", "@end", "@implementation", "@interface", "@public", "@private",
              "@protected", "@property", "@synthesize", "__fastcall", "__stdcall", nullptr };

        const char* const* k;

        switch (tokenLength)
        {
            case 2:     k = keywords2Char; break;
            case 3:     k = keywords3Char; break;
            case 4:     k = keywords4Char; break;
            case 5:     k = keywords5Char; break;
            case 6:     k = keywords6Char; break;
            case 7:     k = keywords7Char; break;

            default:
                if (tokenLength < 2 || tokenLength > 16)
                    return false;

                k = keywordsOther;
                break;
        }

        for (int i = 0; k[i] != 0; ++i)
            if (token.compare (CharPointer_ASCII (k[i])) == 0)
                return true;

        return false;
    }

    template <typename Iterator>
    static int parseIdentifier (Iterator& source) noexcept
    {
        int tokenLength = 0;
        String::CharPointerType::CharType possibleIdentifier [100];
        String::CharPointerType possible (possibleIdentifier);

        while (isIdentifierBody (source.peekNextChar()))
        {
            const juce_wchar c = source.nextChar();

            if (tokenLength < 20)
                possible.write (c);

            ++tokenLength;
        }

        if (tokenLength > 1 && tokenLength <= 16)
        {
            possible.writeNull();

            if (isReservedKeyword (String::CharPointerType (possibleIdentifier), tokenLength))
                return CPlusPlusCodeTokeniser::tokenType_keyword;
        }

        return CPlusPlusCodeTokeniser::tokenType_identifier;
    }

    template <typename Iterator>
    static bool skipNumberSuffix (Iterator& source)
    {
        const juce_wchar c = source.peekNextChar();
        if (c == 'l' || c == 'L' || c == 'u' || c == 'U')
            source.skip();

        if (CharacterFunctions::isLetterOrDigit (source.peekNextChar()))
            return false;

        return true;
    }

    static bool isHexDigit (const juce_wchar c) noexcept
    {
        return (c >= '0' && c <= '9')
                || (c >= 'a' && c <= 'f')
                || (c >= 'A' && c <= 'F');
    }

    template <typename Iterator>
    static bool parseHexLiteral (Iterator& source) noexcept
    {
        if (source.peekNextChar() == '-')
            source.skip();

        if (source.nextChar() != '0')
            return false;

        juce_wchar c = source.nextChar();
        if (c != 'x' && c != 'X')
            return false;

        int numDigits = 0;
        while (isHexDigit (source.peekNextChar()))
        {
            ++numDigits;
            source.skip();
        }

        if (numDigits == 0)
            return false;

        return skipNumberSuffix (source);
    }

    static bool isOctalDigit (const juce_wchar c) noexcept
    {
        return c >= '0' && c <= '7';
    }

    template <typename Iterator>
    static bool parseOctalLiteral (Iterator& source) noexcept
    {
        if (source.peekNextChar() == '-')
            source.skip();

        if (source.nextChar() != '0')
            return false;

        if (! isOctalDigit (source.nextChar()))
            return false;

        while (isOctalDigit (source.peekNextChar()))
            source.skip();

        return skipNumberSuffix (source);
    }

    static bool isDecimalDigit (const juce_wchar c) noexcept
    {
        return c >= '0' && c <= '9';
    }

    template <typename Iterator>
    static bool parseDecimalLiteral (Iterator& source) noexcept
    {
        if (source.peekNextChar() == '-')
            source.skip();

        int numChars = 0;
        while (isDecimalDigit (source.peekNextChar()))
        {
            ++numChars;
            source.skip();
        }

        if (numChars == 0)
            return false;

        return skipNumberSuffix (source);
    }

    template <typename Iterator>
    static bool parseFloatLiteral (Iterator& source) noexcept
    {
        if (source.peekNextChar() == '-')
            source.skip();

        int numDigits = 0;

        while (isDecimalDigit (source.peekNextChar()))
        {
            source.skip();
            ++numDigits;
        }

        const bool hasPoint = (source.peekNextChar() == '.');

        if (hasPoint)
        {
            source.skip();

            while (isDecimalDigit (source.peekNextChar()))
            {
                source.skip();
                ++numDigits;
            }
        }

        if (numDigits == 0)
            return false;

        juce_wchar c = source.peekNextChar();
        const bool hasExponent = (c == 'e' || c == 'E');

        if (hasExponent)
        {
            source.skip();

            c = source.peekNextChar();
            if (c == '+' || c == '-')
                source.skip();

            int numExpDigits = 0;
            while (isDecimalDigit (source.peekNextChar()))
            {
                source.skip();
                ++numExpDigits;
            }

            if (numExpDigits == 0)
                return false;
        }

        c = source.peekNextChar();
        if (c == 'f' || c == 'F')
            source.skip();
        else if (! (hasExponent || hasPoint))
            return false;

        return true;
    }

    template <typename Iterator>
    static int parseNumber (Iterator& source)
    {
        const Iterator original (source);

        if (parseFloatLiteral (source))    return CPlusPlusCodeTokeniser::tokenType_float;
        source = original;

        if (parseHexLiteral (source))      return CPlusPlusCodeTokeniser::tokenType_integer;
        source = original;

        if (parseOctalLiteral (source))    return CPlusPlusCodeTokeniser::tokenType_integer;
        source = original;

        if (parseDecimalLiteral (source))  return CPlusPlusCodeTokeniser::tokenType_integer;
        source = original;

        return CPlusPlusCodeTokeniser::tokenType_error;
    }

    template <typename Iterator>
    static void skipQuotedString (Iterator& source) noexcept
    {
        const juce_wchar quote = source.nextChar();

        for (;;)
        {
            const juce_wchar c = source.nextChar();

            if (c == quote || c == 0)
                break;

            if (c == '\\')
                source.skip();
        }
    }

    template <typename Iterator>
    static void skipComment (Iterator& source) noexcept
    {
        bool lastWasStar = false;

        for (;;)
        {
            const juce_wchar c = source.nextChar();

            if (c == 0 || (c == '/' && lastWasStar))
                break;

            lastWasStar = (c == '*');
        }
    }

    template <typename Iterator>
    static void skipPreprocessorLine (Iterator& source) noexcept
    {
        bool lastWasBackslash = false;

        for (;;)
        {
            const juce_wchar c = source.peekNextChar();

            if (c == '"')
            {
                skipQuotedString (source);
                continue;
            }

            if (c == '/')
            {
                Iterator next (source);
                next.skip();
                const juce_wchar c2 = next.peekNextChar();

                if (c2 == '/' || c2 == '*')
                    return;
            }

            if (c == 0)
                break;

            if (c == '\n' || c == '\r')
            {
                source.skipToEndOfLine();

                if (lastWasBackslash)
                    skipPreprocessorLine (source);

                break;
            }

            lastWasBackslash = (c == '\\');
            source.skip();
        }
    }

    template <typename Iterator>
    static void skipIfNextCharMatches (Iterator& source, const juce_wchar c) noexcept
    {
        if (source.peekNextChar() == c)
            source.skip();
    }

    template <typename Iterator>
    static void skipIfNextCharMatches (Iterator& source, const juce_wchar c1, const juce_wchar c2) noexcept
    {
        const juce_wchar c = source.peekNextChar();

        if (c == c1 || c == c2)
            source.skip();
    }

    template <typename Iterator>
    static int readNextToken (Iterator& source)
    {
        source.skipWhitespace();

        const juce_wchar firstChar = source.peekNextChar();

        switch (firstChar)
        {
        case 0:
            break;

        case '0':   case '1':   case '2':   case '3':   case '4':
        case '5':   case '6':   case '7':   case '8':   case '9':
        case '.':
        {
            int result = parseNumber (source);

            if (result == CPlusPlusCodeTokeniser::tokenType_error)
            {
                source.skip();

                if (firstChar == '.')
                    return CPlusPlusCodeTokeniser::tokenType_punctuation;
            }

            return result;
        }

        case ',':
        case ';':
        case ':':
            source.skip();
            return CPlusPlusCodeTokeniser::tokenType_punctuation;

        case '(':   case ')':
        case '{':   case '}':
        case '[':   case ']':
            source.skip();
            return CPlusPlusCodeTokeniser::tokenType_bracket;

        case '"':
        case '\'':
            skipQuotedString (source);
            return CPlusPlusCodeTokeniser::tokenType_string;

        case '+':
            source.skip();
            skipIfNextCharMatches (source, '+', '=');
            return CPlusPlusCodeTokeniser::tokenType_operator;

        case '-':
        {
            source.skip();
            int result = parseNumber (source);

            if (result == CPlusPlusCodeTokeniser::tokenType_error)
            {
                skipIfNextCharMatches (source, '-', '=');
                return CPlusPlusCodeTokeniser::tokenType_operator;
            }

            return result;
        }

        case '*':   case '%':
        case '=':   case '!':
            source.skip();
            skipIfNextCharMatches (source, '=');
            return CPlusPlusCodeTokeniser::tokenType_operator;

        case '/':
        {
            source.skip();
            juce_wchar nextChar = source.peekNextChar();

            if (nextChar == '/')
            {
                source.skipToEndOfLine();
                return CPlusPlusCodeTokeniser::tokenType_comment;
            }

            if (nextChar == '*')
            {
                source.skip();
                skipComment (source);
                return CPlusPlusCodeTokeniser::tokenType_comment;
            }

            if (nextChar == '=')
                source.skip();

            return CPlusPlusCodeTokeniser::tokenType_operator;
        }

        case '?':
        case '~':
            source.skip();
            return CPlusPlusCodeTokeniser::tokenType_operator;

        case '<':   case '>':
        case '|':   case '&':   case '^':
            source.skip();
            skipIfNextCharMatches (source, firstChar);
            skipIfNextCharMatches (source, '=');
            return CPlusPlusCodeTokeniser::tokenType_operator;

        case '#':
            skipPreprocessorLine (source);
            return CPlusPlusCodeTokeniser::tokenType_preprocessor;

        default:
            if (isIdentifierStart (firstChar))
                return parseIdentifier (source);

            source.skip();
            break;
        }

        return CPlusPlusCodeTokeniser::tokenType_error;
    }

    /** A class that can be passed to the CppTokeniserFunctions functions in order to
        parse a String.
    */
    struct StringIterator
    {
        StringIterator (const String& s) noexcept            : t (s.getCharPointer()), numChars (0) {}
        StringIterator (String::CharPointerType s) noexcept  : t (s), numChars (0) {}

        juce_wchar nextChar() noexcept      { if (isEOF()) return 0; ++numChars; return t.getAndAdvance(); }
        juce_wchar peekNextChar()noexcept   { return *t; }
        void skip() noexcept                { if (! isEOF()) { ++t; ++numChars; } }
        void skipWhitespace() noexcept      { while (t.isWhitespace()) skip(); }
        void skipToEndOfLine() noexcept     { while (*t != '\r' && *t != '\n' && *t != 0) skip(); }
        bool isEOF() const noexcept         { return t.isEmpty(); }

        String::CharPointerType t;
        int numChars;
    };

    //==============================================================================
    /** Takes a UTF8 string and writes it to a stream using standard C++ escape sequences for any
        non-ascii bytes.

        Although not strictly a tokenising function, this is still a function that often comes in
        handy when working with C++ code!

        Note that addEscapeChars() is easier to use than this function if you're working with Strings.

        @see addEscapeChars
    */
    static void writeEscapeChars (OutputStream& out, const char* utf8, const int numBytesToRead,
                                  const int maxCharsOnLine, const bool breakAtNewLines,
                                  const bool replaceSingleQuotes, const bool allowStringBreaks)
    {
        int charsOnLine = 0;
        bool lastWasHexEscapeCode = false;

        for (int i = 0; i < numBytesToRead || numBytesToRead < 0; ++i)
        {
            const unsigned char c = (unsigned char) utf8[i];
            bool startNewLine = false;

            switch (c)
            {
                case '\t':  out << "\\t";  lastWasHexEscapeCode = false; charsOnLine += 2; break;
                case '\r':  out << "\\r";  lastWasHexEscapeCode = false; charsOnLine += 2; break;
                case '\n':  out << "\\n";  lastWasHexEscapeCode = false; charsOnLine += 2; startNewLine = breakAtNewLines; break;
                case '\\':  out << "\\\\"; lastWasHexEscapeCode = false; charsOnLine += 2; break;
                case '\"':  out << "\\\""; lastWasHexEscapeCode = false; charsOnLine += 2; break;

                case 0:
                    if (numBytesToRead < 0)
                        return;

                    out << "\\0";
                    lastWasHexEscapeCode = true;
                    charsOnLine += 2;
                    break;

                case '\'':
                    if (replaceSingleQuotes)
                    {
                        out << "\\\'";
                        lastWasHexEscapeCode = false;
                        charsOnLine += 2;
                        break;
                    }

                    // deliberate fall-through...

                default:
                    if (c >= 32 && c < 127 && ! (lastWasHexEscapeCode  // (have to avoid following a hex escape sequence with a valid hex digit)
                                                   && CharacterFunctions::getHexDigitValue (c) >= 0))
                    {
                        out << (char) c;
                        lastWasHexEscapeCode = false;
                        ++charsOnLine;
                    }
                    else if (allowStringBreaks && lastWasHexEscapeCode && c >= 32 && c < 127)
                    {
                        out << "\"\"" << (char) c;
                        lastWasHexEscapeCode = false;
                        charsOnLine += 3;
                    }
                    else
                    {
                        out << (c < 16 ? "\\x0" : "\\x") << String::toHexString ((int) c);
                        lastWasHexEscapeCode = true;
                        charsOnLine += 4;
                    }

                    break;
            }

            if ((startNewLine || (maxCharsOnLine > 0 && charsOnLine >= maxCharsOnLine))
                 && (numBytesToRead < 0 || i < numBytesToRead - 1))
            {
                charsOnLine = 0;
                out << "\"" << newLine << "\"";
                lastWasHexEscapeCode = false;
            }
        }
    }

    /** Takes a string and returns a version of it where standard C++ escape sequences have been
        used to replace any non-ascii bytes.

        Although not strictly a tokenising function, this is still a function that often comes in
        handy when working with C++ code!

        @see writeEscapeChars
    */
    static String addEscapeChars (const String& s)
    {
        MemoryOutputStream mo;
        writeEscapeChars (mo, s.toRawUTF8(), -1, -1, false, true, true);
        return mo.toString();
    }
};


#endif   // JUCE_CPLUSPLUSCODETOKENISERFUNCTIONS_H_INCLUDED