Пример #1
0
        /*
        * Method:  FindNextToken
        *
        * Find the next token. Return 'true' if one was found. False, otherwise.
        */
        internal override bool FindNextToken()
        {
            int startPosition = _reader.Position;

            // VB docs claim whitespace is Unicode category Zs. However,
            // this category does not contain tabs. Assuming a less restrictive
            // definition for whitespace...
            if (_reader.SinkWhiteSpace())
            {
                while (_reader.SinkWhiteSpace())
                {
                }

                // Now, we need to check for the line continuation character.
                if (_reader.SinkLineContinuationCharacter())    // Line continuation is '_'
                {
                    // Save the current position because we may need to come back here.
                    int savePosition = _reader.Position - 1;

                    // Skip all whitespace after the '_'
                    while (_reader.SinkWhiteSpace())
                    {
                    }

                    // Now, skip all the newlines.
                    // Need at least one newline for this to count as line continuation.
                    int count = 0;
                    while (_reader.SinkNewLine())
                    {
                        ++count;
                    }

                    if (count > 0)
                    {
                        current = new VisualBasicTokenizer.LineContinuationToken();
                        return true;
                    }

                    // Otherwise, fall back to plain old whitespace.
                    _reader.Position = savePosition;
                }

                current = new WhitespaceToken();
                return true;
            }
            // Line terminators are separate from whitespace and are significant.
            else if (_reader.SinkNewLine())
            {
                // We want one token per line terminator.
                current = new VisualBasicTokenizer.LineTerminatorToken();
                return true;
            }
            // Check for a comment--either those that start with ' or rem.
            else if (_reader.SinkLineCommentStart())
            {
                // Skip to the first EOL.
                _reader.SinkToEndOfLine();

                current = new CommentToken();
                return true;
            }
            // Identifier or keyword?
            else if
            (
                // VB allows escaping of identifiers by surrounding them with []
                // In other words,
                //      Date is a keyword but,
                //      [Date] is an identifier.
                _reader.CurrentCharacter == '[' ||
                _reader.MatchNextIdentifierStart()
            )
            {
                bool escapedIdentifier = false;
                if (_reader.CurrentCharacter == '[')
                {
                    escapedIdentifier = true;
                    _reader.SinkCharacter();

                    // Now, the next character must be an identifier start.
                    if (!_reader.SinkIdentifierStart())
                    {
                        current = new ExpectedIdentifierToken();
                        return true;
                    }
                }

                // Sink the rest of the identifier.
                while (_reader.SinkIdentifierPart())
                {
                }

                // If this was an escaped identifier the we need to get the terminating ']'.
                if (escapedIdentifier)
                {
                    if (!_reader.Sink("]"))
                    {
                        current = new ExpectedIdentifierToken();
                        return true;
                    }
                }
                else
                {
                    // Escaped identifiers are not allowed to have trailing type character.
                    _reader.SinkTypeCharacter(); // Type character is optional.
                }

                // An identifier that is only a '_' is illegal because it is
                // ambiguous with line continuation
                string identifierOrKeyword = _reader.GetCurrentMatchedString(startPosition);
                if (identifierOrKeyword == "_" || identifierOrKeyword == "[_]" || identifierOrKeyword == "[]")
                {
                    current = new ExpectedIdentifierToken();
                    return true;
                }

                // Make an upper-case version in order to check whether this may be a keyword.
                string upper = identifierOrKeyword.ToUpper(CultureInfo.InvariantCulture);

                switch (upper)
                {
                    default:

                        if (Array.IndexOf(s_keywordList, upper) >= 0)
                        {
                            current = new KeywordToken();
                            return true;
                        }

                        // Create the token.
                        current = new IdentifierToken();

                        // Trim off the [] if this is an escaped identifier.
                        if (escapedIdentifier)
                        {
                            current.InnerText = identifierOrKeyword.Substring(1, identifierOrKeyword.Length - 2);
                        }
                        return true;
                    case "FALSE":
                    case "TRUE":
                        current = new BooleanLiteralToken();
                        return true;
                }
            }
            // Is it a hex integer?
            else if (_reader.SinkHexIntegerPrefix())
            {
                if (!_reader.SinkMultipleHexDigits())
                {
                    current = new ExpectedValidHexDigitToken();
                    return true;
                }

                // Sink a suffix if there is one.
                _reader.SinkIntegerSuffix();

                current = new HexIntegerLiteralToken();
                return true;
            }
            // Is it an octal integer?
            else if (_reader.SinkOctalIntegerPrefix())
            {
                if (!_reader.SinkMultipleOctalDigits())
                {
                    current = new VisualBasicTokenizer.ExpectedValidOctalDigitToken();
                    return true;
                }

                // Sink a suffix if there is one.
                _reader.SinkIntegerSuffix();

                current = new VisualBasicTokenizer.OctalIntegerLiteralToken();
                return true;
            }
            // Is it a decimal integer?
            else if (_reader.SinkMultipleDecimalDigits())
            {
                // Sink a suffix if there is one.
                _reader.SinkDecimalIntegerSuffix();

                current = new DecimalIntegerLiteralToken();
                return true;
            }
            // Preprocessor line
            else if (_reader.CurrentCharacter == '#')
            {
                if (_reader.SinkIgnoreCase("#if"))
                {
                    current = new OpenConditionalDirectiveToken();
                }
                else if (_reader.SinkIgnoreCase("#end if"))
                {
                    current = new CloseConditionalDirectiveToken();
                }
                else
                {
                    current = new PreprocessorToken();
                }

                _reader.SinkToEndOfLine();

                return true;
            }
            // Is it a separator?
            else if (_reader.SinkSeparatorCharacter())
            {
                current = new VisualBasicTokenizer.SeparatorToken();
                return true;
            }
            // Is it an operator?
            else if (_reader.SinkOperator())
            {
                current = new OperatorToken();
                return true;
            }
            // A string?
            else if (_reader.Sink("\""))
            {
                do
                {
                    // Inside a verbatim string "" is treated as a special character
                    while (_reader.Sink("\"\""))
                    {
                    }
                }
                while (!_reader.EndOfLines && _reader.SinkCharacter() != '\"');

                // Can't end a file inside a string
                if (_reader.EndOfLines)
                {
                    current = new EndOfFileInsideStringToken();
                    return true;
                }

                current = new StringLiteralToken();
                return true;
            }

            // We didn't recognize the token, so this is a syntax error.
            _reader.SinkCharacter();
            current = new UnrecognizedToken();
            return true;
        }
Пример #2
0
        /*
         * Method:  FindNextToken
         *
         * Find the next token. Return 'true' if one was found. False, otherwise.
         */
        override internal bool FindNextToken()
        {
            int startPosition = _reader.Position;

            // VB docs claim whitespace is Unicode category Zs. However,
            // this category does not contain tabs. Assuming a less restrictive
            // definition for whitespace...
            if (_reader.SinkWhiteSpace())
            {
                while (_reader.SinkWhiteSpace())
                {
                }

                // Now, we need to check for the line continuation character.
                if (_reader.SinkLineContinuationCharacter())    // Line continuation is '_'
                {
                    // Save the current position because we may need to come back here.
                    int savePosition = _reader.Position - 1;

                    // Skip all whitespace after the '_'
                    while (_reader.SinkWhiteSpace())
                    {
                    }

                    // Now, skip all the newlines.
                    // Need at least one newline for this to count as line continuation.
                    int count = 0;
                    while (_reader.SinkNewLine())
                    {
                        ++count;
                    }

                    if (count > 0)
                    {
                        current = new VisualBasicTokenizer.LineContinuationToken();
                        return(true);
                    }

                    // Otherwise, fall back to plain old whitespace.
                    _reader.Position = savePosition;
                }

                current = new WhitespaceToken();
                return(true);
            }
            // Line terminators are separate from whitespace and are significant.
            else if (_reader.SinkNewLine())
            {
                // We want one token per line terminator.
                current = new VisualBasicTokenizer.LineTerminatorToken();
                return(true);
            }
            // Check for a comment--either those that start with ' or rem.
            else if (_reader.SinkLineCommentStart())
            {
                // Skip to the first EOL.
                _reader.SinkToEndOfLine();

                current = new CommentToken();
                return(true);
            }
            // Identifier or keyword?
            else if
            (
                // VB allows escaping of identifiers by surrounding them with []
                // In other words,
                //      Date is a keyword but,
                //      [Date] is an identifier.
                _reader.CurrentCharacter == '[' ||
                _reader.MatchNextIdentifierStart()
            )
            {
                bool escapedIdentifier = false;
                if (_reader.CurrentCharacter == '[')
                {
                    escapedIdentifier = true;
                    _reader.SinkCharacter();

                    // Now, the next character must be an identifier start.
                    if (!_reader.SinkIdentifierStart())
                    {
                        current = new ExpectedIdentifierToken();
                        return(true);
                    }
                }

                // Sink the rest of the identifier.
                while (_reader.SinkIdentifierPart())
                {
                }

                // If this was an escaped identifier the we need to get the terminating ']'.
                if (escapedIdentifier)
                {
                    if (!_reader.Sink("]"))
                    {
                        current = new ExpectedIdentifierToken();
                        return(true);
                    }
                }
                else
                {
                    // Escaped identifiers are not allowed to have trailing type character.
                    _reader.SinkTypeCharacter(); // Type character is optional.
                }

                // An identifier that is only a '_' is illegal because it is
                // ambiguous with line continuation
                string identifierOrKeyword = _reader.GetCurrentMatchedString(startPosition);
                if (identifierOrKeyword == "_" || identifierOrKeyword == "[_]" || identifierOrKeyword == "[]")
                {
                    current = new ExpectedIdentifierToken();
                    return(true);
                }

                // Make an upper-case version in order to check whether this may be a keyword.
                string upper = identifierOrKeyword.ToUpperInvariant();

                switch (upper)
                {
                default:

                    if (Array.IndexOf(s_keywordList, upper) >= 0)
                    {
                        current = new KeywordToken();
                        return(true);
                    }

                    // Create the token.
                    current = new IdentifierToken();

                    // Trim off the [] if this is an escaped identifier.
                    if (escapedIdentifier)
                    {
                        current.InnerText = identifierOrKeyword.Substring(1, identifierOrKeyword.Length - 2);
                    }
                    return(true);

                case "FALSE":
                case "TRUE":
                    current = new BooleanLiteralToken();
                    return(true);
                }
            }
            // Is it a hex integer?
            else if (_reader.SinkHexIntegerPrefix())
            {
                if (!_reader.SinkMultipleHexDigits())
                {
                    current = new ExpectedValidHexDigitToken();
                    return(true);
                }

                // Sink a suffix if there is one.
                _reader.SinkIntegerSuffix();

                current = new HexIntegerLiteralToken();
                return(true);
            }
            // Is it an octal integer?
            else if (_reader.SinkOctalIntegerPrefix())
            {
                if (!_reader.SinkMultipleOctalDigits())
                {
                    current = new VisualBasicTokenizer.ExpectedValidOctalDigitToken();
                    return(true);
                }

                // Sink a suffix if there is one.
                _reader.SinkIntegerSuffix();

                current = new VisualBasicTokenizer.OctalIntegerLiteralToken();
                return(true);
            }
            // Is it a decimal integer?
            else if (_reader.SinkMultipleDecimalDigits())
            {
                // Sink a suffix if there is one.
                _reader.SinkDecimalIntegerSuffix();

                current = new DecimalIntegerLiteralToken();
                return(true);
            }
            // Preprocessor line
            else if (_reader.CurrentCharacter == '#')
            {
                if (_reader.SinkIgnoreCase("#if"))
                {
                    current = new OpenConditionalDirectiveToken();
                }
                else if (_reader.SinkIgnoreCase("#end if"))
                {
                    current = new CloseConditionalDirectiveToken();
                }
                else
                {
                    current = new PreprocessorToken();
                }

                _reader.SinkToEndOfLine();

                return(true);
            }
            // Is it a separator?
            else if (_reader.SinkSeparatorCharacter())
            {
                current = new VisualBasicTokenizer.SeparatorToken();
                return(true);
            }
            // Is it an operator?
            else if (_reader.SinkOperator())
            {
                current = new OperatorToken();
                return(true);
            }
            // A string?
            else if (_reader.Sink("\""))
            {
                do
                {
                    // Inside a verbatim string "" is treated as a special character
                    while (_reader.Sink("\"\""))
                    {
                    }
                }while (!_reader.EndOfLines && _reader.SinkCharacter() != '\"');

                // Can't end a file inside a string
                if (_reader.EndOfLines)
                {
                    current = new EndOfFileInsideStringToken();
                    return(true);
                }

                current = new StringLiteralToken();
                return(true);
            }


            // We didn't recognize the token, so this is a syntax error.
            _reader.SinkCharacter();
            current = new UnrecognizedToken();
            return(true);
        }
Пример #3
0
        public IEnumerable <IToken> Tokenize(SafeStreamReader source)
        {
            uint lineNumber = 1;
            uint lexemeStartPositionInLine = 1;
            uint absolutePosition          = 1;
            var  maybeCurrentChar          = Option <int> .None;

            var currentLexemeBuffer = new StringBuilder();

            var maybeToken = Option <IToken> .None;;

            while ((maybeCurrentChar = source.Read()).IsSome)
            {
                var currentChar = maybeCurrentChar.Value();

                maybeToken = Option <IToken> .None;

                switch (currentChar)
                {
                case var c when string.IsNullOrWhiteSpace(char.ConvertFromUtf32(c)):
                    // if a whitespace was encountered - strip it
                    // and yield whatever in the buffer to the output

                    maybeToken = FlushBuffer(
                        currentLexemeBuffer,
                        ref absolutePosition,
                        lineNumber,
                        ref lexemeStartPositionInLine
                        );

                    if (maybeToken.IsSome)
                    {
                        yield return(maybeToken.ValueUnsafe());
                    }

                    switch (c)
                    {
                    case '\r':
                        yield return(source.Read()
                                     .Some <IToken>(cn =>
                                                    cn == '\n' ?
                                                    (IToken) new NewLineSymbolToken(
                                                        absolutePosition,
                                                        lineNumber,
                                                        lexemeStartPositionInLine
                                                        ) :
                                                    (IToken) new UnrecognizedToken(
                                                        $"\r{cn}",
                                                        absolutePosition,
                                                        lineNumber,
                                                        lexemeStartPositionInLine
                                                        )
                                                    )
                                     .None(new UnrecognizedToken(
                                               $"\r",
                                               absolutePosition,
                                               lineNumber,
                                               lexemeStartPositionInLine
                                               ))
                                     );

                        absolutePosition         += 2;
                        lineNumber               += 1;
                        lexemeStartPositionInLine = 1;

                        break;

                    case '\n':
                        yield return(new NewLineSymbolToken(
                                         absolutePosition,
                                         lineNumber,
                                         lexemeStartPositionInLine
                                         ));

                        absolutePosition         += 1;
                        lineNumber               += 1;
                        lexemeStartPositionInLine = 1;

                        break;

                    default:
                        absolutePosition          += 1;
                        lexemeStartPositionInLine += 1;
                        break;
                    }

                    break;

                case '.':
                    var currentLexeme = currentLexemeBuffer.ToString();

                    var maybeBeforeToken =
                        IntegerLiteralToken.FromString(
                            currentLexeme,
                            absolutePosition,
                            lineNumber,
                            lexemeStartPositionInLine
                            ) ||
                        IdentifierToken.FromString(
                            currentLexeme,
                            absolutePosition,
                            lineNumber,
                            lexemeStartPositionInLine
                            ) ||
                        UnrecognizedToken.FromString(
                            currentLexeme,
                            absolutePosition,
                            lineNumber,
                            lexemeStartPositionInLine
                            )
                    ;

                    var tokes =
                        source.Peek()
                        .Some <ImmutableList <IToken> >(c =>
                    {
                        var result        = ImmutableList <IToken> .Empty;
                        IToken tokenToAdd = null;

                        switch (c)
                        {
                        case var _ when IsDigit(char.ConvertFromUtf32(c)):
                            currentLexemeBuffer.Append('.');
                            return(ImmutableList <IToken> .Empty);

                        case '.':
                            absolutePosition += maybeBeforeToken
                                                .Map(t => (uint)t.Lexeme.Length)
                                                .IfNone(0);
                            lexemeStartPositionInLine += maybeBeforeToken
                                                         .Some(t => (uint)t.Lexeme.Length)
                                                         .None(0u);

                            tokenToAdd = new RangeSymbolToken(
                                absolutePosition,
                                lineNumber,
                                lexemeStartPositionInLine
                                );

                            result = maybeBeforeToken
                                     .ToImmutableList()
                                     .Add(tokenToAdd);
                            source.Read();
                            currentLexemeBuffer.Clear();
                            lexemeStartPositionInLine += (uint)(tokenToAdd?.Lexeme.Length ?? 0);
                            absolutePosition          += (uint)(tokenToAdd?.Lexeme.Length ?? 0);

                            return(result);

                        default:
                            absolutePosition += maybeBeforeToken
                                                .Map(t => (uint)t.Lexeme.Length)
                                                .IfNone(0);
                            lexemeStartPositionInLine += maybeBeforeToken
                                                         .Some(t => (uint)t.Lexeme.Length)
                                                         .None(0u);

                            tokenToAdd = new DotSymbolToken(
                                absolutePosition,
                                lineNumber,
                                lexemeStartPositionInLine
                                );

                            result = maybeBeforeToken
                                     .ToImmutableList()
                                     .Add(tokenToAdd);
                            currentLexemeBuffer.Clear();
                            lexemeStartPositionInLine += (uint)(tokenToAdd?.Lexeme.Length ?? 0);
                            absolutePosition          += (uint)(tokenToAdd?.Lexeme.Length ?? 0);

                            return(result);
                        }
                    })
                        .None(() =>
                    {
                        absolutePosition += maybeBeforeToken
                                            .Map(t => (uint)t.Lexeme.Length)
                                            .IfNone(0);
                        lexemeStartPositionInLine += maybeBeforeToken
                                                     .Some(t => (uint)t.Lexeme.Length)
                                                     .None(0u);

                        var tokenToAdd = new DotSymbolToken(
                            absolutePosition,
                            lineNumber,
                            lexemeStartPositionInLine
                            );

                        var result = maybeBeforeToken
                                     .ToImmutableList()
                                     .Add(tokenToAdd);
                        currentLexemeBuffer.Clear();
                        lexemeStartPositionInLine += (uint)(tokenToAdd?.Lexeme.Length ?? 0);
                        absolutePosition          += (uint)(tokenToAdd?.Lexeme.Length ?? 0);

                        return(result);
                    })
                    ;

                    foreach (var token in tokes)
                    {
                        yield return(token);
                    }

                    break;

                case '/':
                    maybeToken = FlushBuffer(
                        currentLexemeBuffer,
                        ref absolutePosition,
                        lineNumber,
                        ref lexemeStartPositionInLine
                        );
                    if (maybeToken.IsSome)
                    {
                        yield return(maybeToken.ValueUnsafe());
                    }

                    yield return(source.Peek()
                                 .Some <IToken>(c =>
                    {
                        switch (c)
                        {
                        case '/':
                            var commentContent = source.ReadLine();

                            var commentToken = new CommentToken(
                                $"/{commentContent}",
                                absolutePosition,
                                lineNumber,
                                lexemeStartPositionInLine
                                );

                            absolutePosition += (uint)commentContent.Length;
                            lineNumber += 1;
                            lexemeStartPositionInLine = 0;

                            return commentToken;

                        case '=':
                            var notEqualsToken = new NotEqualsOperatorToken(
                                absolutePosition,
                                lineNumber,
                                lexemeStartPositionInLine
                                );

                            source.Read();
                            absolutePosition += 1;
                            lexemeStartPositionInLine = 1;

                            return notEqualsToken;

                        default:
                            return new DivideOperatorToken(
                                (uint)source.BaseStream.Position,
                                lineNumber,
                                lexemeStartPositionInLine
                                );
                        }
                    })
                                 .None(() => new DivideOperatorToken(
                                           (uint)source.BaseStream.Position,
                                           lineNumber,
                                           lexemeStartPositionInLine
                                           )));

                    absolutePosition          += 1;
                    lexemeStartPositionInLine += 1;

                    break;

                case ':':
                    maybeToken = FlushBuffer(
                        currentLexemeBuffer,
                        ref absolutePosition,
                        lineNumber,
                        ref lexemeStartPositionInLine
                        );
                    if (maybeToken.IsSome)
                    {
                        yield return(maybeToken.ValueUnsafe());
                    }

                    yield return(source.Peek()
                                 .Filter(c => c == '=')
                                 .Some <IToken>(c =>
                    {
                        var result = new AssignmentOperatorToken(
                            absolutePosition,
                            lineNumber,
                            lexemeStartPositionInLine
                            );

                        source.Read();
                        absolutePosition += 1;
                        lexemeStartPositionInLine += 1;

                        return result;
                    })
                                 .None(new ColonSymbolToken(
                                           absolutePosition,
                                           lineNumber,
                                           lexemeStartPositionInLine
                                           )));

                    absolutePosition          += 1;
                    lexemeStartPositionInLine += 1;

                    break;

                case '>':
                    maybeToken = FlushBuffer(
                        currentLexemeBuffer,
                        ref absolutePosition,
                        lineNumber,
                        ref lexemeStartPositionInLine
                        );
                    if (maybeToken.IsSome)
                    {
                        yield return(maybeToken.ValueUnsafe());
                    }

                    yield return(source.Peek()
                                 .Filter(c => c == '=')
                                 .Some <IToken>(_ =>
                    {
                        var result = new GeOperatorToken(
                            absolutePosition,
                            lineNumber,
                            lexemeStartPositionInLine
                            );

                        source.Read();
                        absolutePosition += 1;
                        lexemeStartPositionInLine += 1;

                        return result;
                    })
                                 .None(new GtOperatorToken(
                                           (uint)absolutePosition,
                                           lineNumber,
                                           lexemeStartPositionInLine
                                           )));

                    absolutePosition          += 1;
                    lexemeStartPositionInLine += 1;

                    break;

                case '<':
                    maybeToken = FlushBuffer(
                        currentLexemeBuffer,
                        ref absolutePosition,
                        lineNumber,
                        ref lexemeStartPositionInLine
                        );
                    if (maybeToken.IsSome)
                    {
                        yield return(maybeToken.ValueUnsafe());
                    }

                    yield return(source.Peek()
                                 .Filter(c => c == '=')
                                 .Some <IToken>(_ =>
                    {
                        var result = new LeOperatorToken(
                            absolutePosition,
                            lineNumber,
                            lexemeStartPositionInLine
                            );

                        source.Read();
                        absolutePosition += 1;
                        lexemeStartPositionInLine += 1;

                        return result;
                    })
                                 .None(new LtOperatorToken(
                                           absolutePosition,
                                           lineNumber,
                                           lexemeStartPositionInLine
                                           )));

                    absolutePosition          += 1;
                    lexemeStartPositionInLine += 1;

                    break;

                case '*':
                case '%':
                case '+':
                case '-':
                case '=':
                case ',':
                case '[':
                case ']':
                case '(':
                case ')':
                case ';':
                    maybeToken = FlushBuffer(
                        currentLexemeBuffer,
                        ref absolutePosition,
                        lineNumber,
                        ref lexemeStartPositionInLine
                        );
                    if (maybeToken.IsSome)
                    {
                        yield return(maybeToken.ValueUnsafe());
                    }

                    yield return(SymbolLexemes
                                 .TryGetValue(((char)currentChar).ToString())
                                 .Some(cons => cons(
                                           absolutePosition,
                                           lineNumber,
                                           lexemeStartPositionInLine
                                           ))
                                 .None(() => new UnrecognizedToken(
                                           currentChar.ToString(),
                                           absolutePosition,
                                           lineNumber,
                                           lexemeStartPositionInLine
                                           )
                                       ));

                    absolutePosition          += 1;
                    lexemeStartPositionInLine += 1;

                    break;

                default:
                    currentLexemeBuffer.Append(char.ConvertFromUtf32(currentChar));
                    break;
                }
            }

            maybeToken = FlushBuffer(
                currentLexemeBuffer,
                ref absolutePosition,
                lineNumber,
                ref lexemeStartPositionInLine
                );
            if (maybeToken.IsSome)
            {
                yield return(maybeToken.ValueUnsafe());
            }
        }
Пример #4
0
        /*
         * Method:  FindNextToken
         *
         * Find the next token. Return 'true' if one was found. False, otherwise.
         */
        override internal bool FindNextToken()
        {
            int startPosition = _reader.Position;

            // Dealing with whitespace?
            if (_reader.SinkMultipleWhiteSpace())
            {
                current = new WhitespaceToken();
                return(true);
            }
            // Check for one-line comment
            else if (_reader.Sink("//"))
            {
                // Looks like a one-line comment. Follow it to the End-of-line
                _reader.SinkToEndOfLine();

                current = new CommentToken();
                return(true);
            }
            // Check for multi-line comment
            else if (_reader.Sink("/*"))
            {
                _reader.SinkUntil("*/");

                // Was the ending */ found?
                if (_reader.EndOfLines)
                {
                    // No. There was a /* without a */. Return this a syntax error token.
                    current = new CSharpTokenizer.EndOfFileInsideCommentToken();
                    return(true);
                }

                current = new CommentToken();
                return(true);
            }
            // Handle chars
            else if (_reader.Sink("\'"))
            {
                while (_reader.CurrentCharacter != '\'')
                {
                    if (_reader.Sink("\\"))
                    {
                        /* reader.Skip the escape sequence.
                         *  This isn't exactly right. We should detect:
                         *
                         *  simple-escape-sequence: one of
                         \' \" \\ \0 \a \b \f \n \r \t \v
                         *
                         *  hexadecimal-escape-sequence:
                         *  \x   hex-digit   hex-digit[opt]   hex-digit[opt]  hex-digit[opt]
                         */
                    }

                    _reader.SinkCharacter();
                }

                if (_reader.SinkCharacter() != '\'')
                {
                    Debug.Assert(false, "Code defect in tokenizer: Should have yielded a closing tick.");
                }
                current = new CSharpTokenizer.CharLiteralToken();
                return(true);
            }
            // Check for verbatim string
            else if (_reader.Sink("@\""))
            {
                do
                {
                    // Inside a verbatim string "" is treated as a special character
                    while (_reader.Sink("\"\""))
                    {
                    }
                }while (!_reader.EndOfLines && _reader.SinkCharacter() != '\"');

                // Can't end a file inside a string
                if (_reader.EndOfLines)
                {
                    current = new EndOfFileInsideStringToken();
                    return(true);
                }

                // reader.Skip the ending quote.
                current           = new StringLiteralToken();
                current.InnerText = _reader.GetCurrentMatchedString(startPosition).Substring(1);
                return(true);
            }
            // Check for a quoted string.
            else if (_reader.Sink("\""))
            {
                while (_reader.CurrentCharacter == '\\' || _reader.MatchRegularStringLiteral())
                {
                    // See if we have an escape sequence.
                    if (_reader.SinkCharacter() == '\\')
                    {
                        // This is probably an escape character.
                        if (_reader.SinkStringEscape())
                        {
                            // This isn't nearly right. We just do barely enough to make a string
                            // with an embedded escape sequence return _some_ string whose start and
                            // end match the real bounds of the string.
                        }
                        else
                        {
                            // This is a compiler error.
                            _reader.SinkCharacter();
                            current = new CSharpTokenizer.UnrecognizedStringEscapeToken();
                            return(true);
                        }
                    }
                }

                // Is it a newline?
                if (TokenChar.IsNewLine(_reader.CurrentCharacter))
                {
                    current = new CSharpTokenizer.NewlineInsideStringToken();
                    return(true);
                }

                // Create the token.
                if (_reader.SinkCharacter() != '\"')
                {
                    Debug.Assert(false, "Defect in tokenizer: Should have yielded a terminating quote.");
                }
                current = new StringLiteralToken();
                return(true);
            }
            // Identifier or keyword?
            else if
            (
                // From 2.4.2 Identifiers: A '@' can be used to prefix an identifier so that a keyword can be used as an identifier.
                _reader.CurrentCharacter == '@' ||
                _reader.MatchNextIdentifierStart()
            )
            {
                if (_reader.CurrentCharacter == '@')
                {
                    _reader.SinkCharacter();
                }

                // Now, the next character must be an identifier start.
                if (!_reader.SinkIdentifierStart())
                {
                    current = new ExpectedIdentifierToken();
                    return(true);
                }

                // Sink the rest of the identifier.
                while (_reader.SinkIdentifierPart())
                {
                }
                string identifierOrKeyword = _reader.GetCurrentMatchedString(startPosition);

                switch (identifierOrKeyword)
                {
                default:

                    if (Array.IndexOf(s_keywordList, identifierOrKeyword) >= 0)
                    {
                        current = new KeywordToken();
                        return(true);
                    }

                    // If the identifier starts with '@' then we need to strip it off.
                    // The '@' is for escaping so that we can have an identifier called
                    // the same thing as a reserved keyword (i.e. class, if, foreach, etc)
                    string identifier = _reader.GetCurrentMatchedString(startPosition);
                    if (identifier.StartsWith("@", StringComparison.Ordinal))
                    {
                        identifier = identifier.Substring(1);
                    }

                    // Create the token.
                    current           = new IdentifierToken();
                    current.InnerText = identifier;
                    return(true);

                case "false":
                case "true":
                    current = new BooleanLiteralToken();
                    return(true);

                case "null":
                    current = new CSharpTokenizer.NullLiteralToken();
                    return(true);
                }
            }
            // Open scope
            else if (_reader.Sink("{"))
            {
                current = new CSharpTokenizer.OpenScopeToken();
                return(true);
            }
            // Close scope
            else if (_reader.Sink("}"))
            {
                current = new CSharpTokenizer.CloseScopeToken();
                return(true);
            }
            // Hexidecimal integer literal
            else if (_reader.SinkIgnoreCase("0x"))
            {
                // Sink the hex digits.
                if (!_reader.SinkMultipleHexDigits())
                {
                    current = new ExpectedValidHexDigitToken();
                    return(true);
                }

                // Skip the L, U, l, u, ul, etc.
                _reader.SinkLongIntegerSuffix();

                current = new HexIntegerLiteralToken();
                return(true);
            }
            // Decimal integer literal
            else if (_reader.SinkMultipleDecimalDigits())
            {
                // reader.Skip the L, U, l, u, ul, etc.
                _reader.SinkLongIntegerSuffix();

                current = new DecimalIntegerLiteralToken();
                return(true);
            }
            // Check for single-digit operators and punctuators
            else if (_reader.SinkOperatorOrPunctuator())
            {
                current = new OperatorOrPunctuatorToken();
                return(true);
            }
            // Preprocessor line
            else if (_reader.CurrentCharacter == '#')
            {
                if (_reader.Sink("#if"))
                {
                    current = new OpenConditionalDirectiveToken();
                }
                else if (_reader.Sink("#endif"))
                {
                    current = new CloseConditionalDirectiveToken();
                }
                else
                {
                    current = new PreprocessorToken();
                }

                _reader.SinkToEndOfLine();

                return(true);
            }

            // We didn't recognize the token, so this is a syntax error.
            _reader.SinkCharacter();
            current = new UnrecognizedToken();
            return(true);
        }
Пример #5
0
        /*
        * Method:  FindNextToken
        * 
        * Find the next token. Return 'true' if one was found. False, otherwise.
        */
        override internal bool FindNextToken()
        {
            int startPosition = _reader.Position;

            // Dealing with whitespace?
            if (_reader.SinkMultipleWhiteSpace())
            {
                current = new WhitespaceToken();
                return true;
            }
            // Check for one-line comment
            else if (_reader.Sink("//"))
            {
                // Looks like a one-line comment. Follow it to the End-of-line
                _reader.SinkToEndOfLine();

                current = new CommentToken();
                return true;
            }
            // Check for multi-line comment
            else if (_reader.Sink("/*"))
            {
                _reader.SinkUntil("*/");

                // Was the ending */ found?
                if (_reader.EndOfLines)
                {
                    // No. There was a /* without a */. Return this a syntax error token.
                    current = new CSharpTokenizer.EndOfFileInsideCommentToken();
                    return true;
                }

                current = new CommentToken();
                return true;
            }
            // Handle chars
            else if (_reader.Sink("\'"))
            {
                while (_reader.CurrentCharacter != '\'')
                {
                    if (_reader.Sink("\\"))
                    {
                        /* reader.Skip the escape sequence. 
                            This isn't exactly right. We should detect:
                            
                            simple-escape-sequence: one of 
                            \' \" \\ \0 \a \b \f \n \r \t \v 
                            
                            hexadecimal-escape-sequence: 
                            \x   hex-digit   hex-digit[opt]   hex-digit[opt]  hex-digit[opt]                                
                        */
                    }

                    _reader.SinkCharacter();
                }

                if (_reader.SinkCharacter() != '\'')
                {
                    Debug.Assert(false, "Code defect in tokenizer: Should have yielded a closing tick.");
                }
                current = new CSharpTokenizer.CharLiteralToken();
                return true;
            }
            // Check for verbatim string
            else if (_reader.Sink("@\""))
            {
                do
                {
                    // Inside a verbatim string "" is treated as a special character
                    while (_reader.Sink("\"\""))
                    {
                    }
                }
                while (!_reader.EndOfLines && _reader.SinkCharacter() != '\"');

                // Can't end a file inside a string 
                if (_reader.EndOfLines)
                {
                    current = new EndOfFileInsideStringToken();
                    return true;
                }

                // reader.Skip the ending quote.
                current = new StringLiteralToken();
                current.InnerText = _reader.GetCurrentMatchedString(startPosition).Substring(1);
                return true;
            }
            // Check for a quoted string.
            else if (_reader.Sink("\""))
            {
                while (_reader.CurrentCharacter == '\\' || _reader.MatchRegularStringLiteral())
                {
                    // See if we have an escape sequence.
                    if (_reader.SinkCharacter() == '\\')
                    {
                        // This is probably an escape character.
                        if (_reader.SinkStringEscape())
                        {
                            // This isn't nearly right. We just do barely enough to make a string
                            // with an embedded escape sequence return _some_ string whose start and 
                            // end match the real bounds of the string.
                        }
                        else
                        {
                            // This is a compiler error. 
                            _reader.SinkCharacter();
                            current = new CSharpTokenizer.UnrecognizedStringEscapeToken();
                            return true;
                        }
                    }
                }

                // Is it a newline?
                if (TokenChar.IsNewLine(_reader.CurrentCharacter))
                {
                    current = new CSharpTokenizer.NewlineInsideStringToken();
                    return true;
                }

                // Create the token.
                if (_reader.SinkCharacter() != '\"')
                {
                    Debug.Assert(false, "Defect in tokenizer: Should have yielded a terminating quote.");
                }
                current = new StringLiteralToken();
                return true;
            }
            // Identifier or keyword?
            else if
            (
                // From 2.4.2 Identifiers: A '@' can be used to prefix an identifier so that a keyword can be used as an identifier.
                _reader.CurrentCharacter == '@' ||
                _reader.MatchNextIdentifierStart()
            )
            {
                if (_reader.CurrentCharacter == '@')
                {
                    _reader.SinkCharacter();
                }

                // Now, the next character must be an identifier start.
                if (!_reader.SinkIdentifierStart())
                {
                    current = new ExpectedIdentifierToken();
                    return true;
                }

                // Sink the rest of the identifier.                     
                while (_reader.SinkIdentifierPart())
                {
                }
                string identifierOrKeyword = _reader.GetCurrentMatchedString(startPosition);

                switch (identifierOrKeyword)
                {
                    default:

                        if (Array.IndexOf(s_keywordList, identifierOrKeyword) >= 0)
                        {
                            current = new KeywordToken();
                            return true;
                        }

                        // If the identifier starts with '@' then we need to strip it off.
                        // The '@' is for escaping so that we can have an identifier called
                        // the same thing as a reserved keyword (i.e. class, if, foreach, etc)
                        string identifier = _reader.GetCurrentMatchedString(startPosition);
                        if (identifier.StartsWith("@", StringComparison.Ordinal))
                        {
                            identifier = identifier.Substring(1);
                        }

                        // Create the token.
                        current = new IdentifierToken();
                        current.InnerText = identifier;
                        return true;
                    case "false":
                    case "true":
                        current = new BooleanLiteralToken();
                        return true;
                    case "null":
                        current = new CSharpTokenizer.NullLiteralToken();
                        return true;
                }
            }
            // Open scope
            else if (_reader.Sink("{"))
            {
                current = new CSharpTokenizer.OpenScopeToken();
                return true;
            }
            // Close scope
            else if (_reader.Sink("}"))
            {
                current = new CSharpTokenizer.CloseScopeToken();
                return true;
            }
            // Hexidecimal integer literal
            else if (_reader.SinkIgnoreCase("0x"))
            {
                // Sink the hex digits.
                if (!_reader.SinkMultipleHexDigits())
                {
                    current = new ExpectedValidHexDigitToken();
                    return true;
                }

                // Skip the L, U, l, u, ul, etc.                    
                _reader.SinkLongIntegerSuffix();

                current = new HexIntegerLiteralToken();
                return true;
            }
            // Decimal integer literal
            else if (_reader.SinkMultipleDecimalDigits())
            {
                // reader.Skip the L, U, l, u, ul, etc.                    
                _reader.SinkLongIntegerSuffix();

                current = new DecimalIntegerLiteralToken();
                return true;
            }
            // Check for single-digit operators and punctuators
            else if (_reader.SinkOperatorOrPunctuator())
            {
                current = new OperatorOrPunctuatorToken();
                return true;
            }
            // Preprocessor line
            else if (_reader.CurrentCharacter == '#')
            {
                if (_reader.Sink("#if"))
                {
                    current = new OpenConditionalDirectiveToken();
                }
                else if (_reader.Sink("#endif"))
                {
                    current = new CloseConditionalDirectiveToken();
                }
                else
                {
                    current = new PreprocessorToken();
                }

                _reader.SinkToEndOfLine();

                return true;
            }

            // We didn't recognize the token, so this is a syntax error. 
            _reader.SinkCharacter();
            current = new UnrecognizedToken();
            return true;
        }