Refactor: Lexer string reading code duplication

[clank-hayen]

Problem

In `lexer/lexer.cpp`, there are four nearly identical string reading functions:

1. read_string() (lines 106-138): Double-quoted with escapes
2. read_single_quoted() (lines 145-177): Single-quoted with escapes
3. read_raw_string() (lines 183-199): Raw double-quoted
4. read_raw_single_quoted() (lines 205-221): Raw single-quoted

Functions 1 and 2 differ only in the quote character ('"' vs '\'').
Functions 3 and 4 differ only in the quote character.

Suggested Solution

Consolidate into two functions with a quote parameter:

Token read_string_literal(char quote_char, bool handle_escapes) {
    int start_line = line;
    advance();  // skip opening quote
    string value;

    while (current() != quote_char && current() != '\0') {
        if (handle_escapes && current() == '\\\\') {
            advance();
            switch (current()) {
                case 'n': value += '\\n'; break;
                case 't': value += '\\t'; break;
                case 'r': value += '\\r'; break;
                case '\\\\': value += '\\\\'; break;
                case '\"': value += '\"'; break;
                case '\\'': value += '\\''; break;
                default: value += current(); break;
            }
            advance();
        } else {
            value += current();
            advance();
        }
    }

    if (current() == '\0') {
        throw runtime_error("Unterminated string literal at line " + to_string(start_line));
    }

    advance();  // skip closing quote
    return {TokenType::STRING, value, start_line};
}

Usage:

// In tokenize():
if (current() == '\"') {
    tokens.push_back(read_string_literal('\"', true));
}
if (current() == '\\'') {
    tokens.push_back(read_string_literal('\\'', true));
}
// For raw strings: read_string_literal(quote, false)

Impact

• Reduces ~80 lines to ~30 lines
• Single place to add new escape sequences
• Consistent behavior across string types