pekoe/internal/protocol/protocol.go

// @tag protocol
// @tag core
package protocol

import (
	"bufio"
	"bytes"
	"errors"
	"fmt"
	"io"
	"strconv"
)

const (
	CRLF        = "\r\n"
	MaxBulkLen  = 512 * 1024 * 1024 // 512 MB limit for bulk strings - prevents memory exhaustion
	MaxArrayLen = 1024 * 1024       // 1M elements max - prevents array parsing attacks
)

// Protocol errors - separated for clear error handling
var (
	ErrProtocolError = errors.New("protocol error")
	ErrNotEnoughData = errors.New("not enough data")
)

// Value represents any RESP value - using interface for simplicity with type switches.
// This mirrors Redis protocol flexibility while keeping implementation simple.
type Value any

// Error represents an error response from Redis (-ERR...)
type Error struct {
	Message string
}

func (e Error) Error() string {
	return e.Message
}

// Array represents an array of values (*...) - used for commands and responses.
type Array []Value

// Bulk represents a bulk string ($...) - binary safe, used for most arguments.
type Bulk []byte

func (b Bulk) String() string {
	return string(b)
}

// Int represents an integer response (:12345) - used for counts and numeric responses.
type Int int64

// Status represents a simple string response (+OK) - used for simple confirmations.
type Status string

func (s Status) String() string {
	return string(s)
}

// Parse reads and parses a single RESP value from the reader.
// This is the main entry point for parsing incoming Redis commands.
// Uses bufio.Reader for efficient I/O with built-in buffering.
func Parse(reader *bufio.Reader) (Value, error) {
	// Read first line to determine message type
	line, err := readLine(reader)
	if err != nil {
		return nil, err
	}
	if len(line) == 0 {
		return nil, ErrProtocolError
	}

	// Route based on first character - this is the RESP type indicator
	switch line[0] {
	case '+': // Simple string: +OK\r\n
		return Status(line[1:]), nil
	case '-': // Error: -ERR message\r\n
		return Error{Message: string(line[1:])}, nil
	case ':': // Integer: :12345\r\n
		n, err := strconv.ParseInt(string(line[1:]), 10, 64)
		if err != nil {
			return nil, ErrProtocolError
		}
		return Int(n), nil
	case '$': // Bulk string: $6\r\nfoobar\r\n or $-1 for null
		return parseBulk(reader, line)
	case '*': // Array: *2\r\n$3\r\nfoo\r\n$3\r\nbar\r\n
		return parseArray(reader, line)
	default:
		return nil, fmt.Errorf("unknown type: %c", line[0])
	}
}

// readLine reads a single line from the reader, handling CRLF.
// Uses ReadLine for efficiency - it handles buffering internally.
func readLine(reader *bufio.Reader) ([]byte, error) {
	line, isPrefix, err := reader.ReadLine()
	if err != nil {
		return nil, err
	}
	if isPrefix {
		return nil, fmt.Errorf("line too long")
	}
	return line, nil
}

// parseBulk handles bulk string parsing ($<length>\r\n<data>\r\n).
// Bulk strings are used for binary-safe string arguments and responses.
func parseBulk(reader *bufio.Reader, line []byte) (Value, error) {
	// Parse the length field after the $ prefix
	size, err := strconv.Atoi(string(line[1:]))
	if err != nil {
		return nil, ErrProtocolError
	}

	// $-1 means null bulk string - legitimate empty value
	if size == -1 {
		return nil, nil
	}

	// Validate size to prevent memory exhaustion attacks
	if size < 0 || size > MaxBulkLen {
		return nil, fmt.Errorf("bulk string too large: %d bytes", size)
	}

	// Read exactly size bytes plus CRLF delimiter
	// Using ReadFull ensures we get all data or an error
	buf := make([]byte, size+len(CRLF))
	_, err = io.ReadFull(reader, buf)
	if err != nil {
		return nil, err
	}

	// Verify proper CRLF termination - prevents protocol injection
	if !bytes.HasSuffix(buf, []byte(CRLF)) {
		return nil, ErrProtocolError
	}

	return Bulk(buf[:size]), nil
}

// parseArray handles array parsing (*<count>\r\n<elements>...).
// Arrays are the fundamental structure for Redis commands.
func parseArray(reader *bufio.Reader, line []byte) (Value, error) {
	count, err := strconv.Atoi(string(line[1:]))
	if err != nil {
		return nil, ErrProtocolError
	}

	// *-1 means null array - rare but valid
	if count == -1 {
		return nil, nil
	}

	// Limit array size to prevent DoS via large arrays
	if count < 0 || count > MaxArrayLen {
		return nil, fmt.Errorf("array too large: %d elements", count)
	}

	// Pre-allocate with expected size to reduce reallocations
	result := make([]Value, count)
	for i := range count {
		val, err := Parse(reader)
		if err != nil {
			return nil, err
		}
		result[i] = val
	}

	return Array(result), nil
}

// Response builders - these create RESP-encoded responses to send to clients

// MakeError creates an error response (-ERR message).
func MakeError(msg string) []byte {
	return []byte("-" + msg + CRLF)
}

// MakeStatus creates a simple status response (+OK).
func MakeStatus(status string) []byte {
	return []byte("+" + status + CRLF)
}

// MakeInt creates an integer response (:12345).
func MakeInt(n int64) []byte {
	return []byte(":" + strconv.FormatInt(n, 10) + CRLF)
}

// MakeBulk creates a bulk string response ($<len>\r\n<data>\r\n).
// Handles nil gracefully by returning null bulk string.
func MakeBulk(data []byte) []byte {
	if data == nil {
		return []byte("$-1" + CRLF)
	}
	// Build directly to avoid string conversion
	b := make([]byte, 0, len(data)+len("$1000000000\r\n\r\n"))
	b = append(b, '$')
	b = strconv.AppendInt(b, int64(len(data)), 10)
	b = append(b, CRLF...)
	b = append(b, data...)
	b = append(b, CRLF...)
	return b
}

// MakeMultiBulk creates a RESP array from Value slices - used for pub/sub messages.
func MakeMultiBulk(values []Value) []byte {
	var buf bytes.Buffer
	buf.WriteString("*" + strconv.Itoa(len(values)) + CRLF)
	for _, v := range values {
		switch val := v.(type) {
		case nil:
			buf.WriteString("$-1\r\n")
		case string:
			buf.Write(MakeBulk([]byte(val)))
		case []byte:
			buf.Write(MakeBulk(val))
		case int64:
			buf.Write(MakeInt(val))
		case Int:
			buf.Write(MakeInt(int64(val)))
		case Error:
			buf.Write(MakeError(val.Message))
		case Status:
			buf.Write(MakeStatus(string(val)))
		default:
			buf.Write(MakeBulk([]byte(fmt.Sprintf("%v", val))))
		}
	}
	return buf.Bytes()
}

// ToBytes converts a parsed Value back to its RESP wire format.
// This is useful for converting parsed commands back to bytes for relay or logging.
// nolint: cyclop
func ToBytes(v Value) []byte {
	if v == nil {
		return MakeBulk(nil)
	}
	switch val := v.(type) {
	case string:
		return MakeBulk([]byte(val))
	case []byte:
		return MakeBulk(val)
	case Bulk:
		return MakeBulk([]byte(val))
	case int64:
		return MakeInt(val)
	case Int:
		return MakeInt(int64(val))
	case Error:
		return MakeError(val.Message)
	case Status:
		return MakeStatus(string(val))
	case Array:
		return MakeMultiBulk(val)
	default:
		return MakeBulk([]byte(fmt.Sprintf("%v", val)))
	}
}