eq2go/internal/packets/parser/lexer.go

package parser

import (
	"fmt"
	"sync"
	"unicode"
)

// Object pools for heavy reuse
var tokenPool = sync.Pool{
	New: func() any {
		return &Token{
			Attributes: make(map[string]string, 8),
			TagStart:   -1,
			TagEnd:     -1,
			TextStart:  -1,
			TextEnd:    -1,
		}
	},
}

// More efficient lexer using byte operations and minimal allocations
type Lexer struct {
	input []byte // Use byte slice for faster operations
	pos   int
	line  int
	col   int
}

// Creates a new lexer
func NewLexer(input string) *Lexer {
	return &Lexer{
		input: []byte(input),
		line:  1,
		col:   1,
	}
}

// Returns next byte without advancing
func (l *Lexer) peek() byte {
	if l.pos >= len(l.input) {
		return 0
	}
	return l.input[l.pos]
}

// Advances and returns next byte
func (l *Lexer) next() byte {
	if l.pos >= len(l.input) {
		return 0
	}
	ch := l.input[l.pos]
	l.pos++
	if ch == '\n' {
		l.line++
		l.col = 1
	} else {
		l.col++
	}
	return ch
}

// Checks if a tag should be treated as self-closing (using byte comparison)
func (l *Lexer) isSelfClosingTag(start, end int) bool {
	length := end - start
	if length < 2 || length > 6 {
		return false
	}

	// Fast byte-based comparison
	switch length {
	case 2:
		return (l.input[start] == 'i' && l.input[start+1] == '8') ||
			(l.input[start] == 'f' && l.input[start+1] == '2')
	case 3:
		return (l.input[start] == 'i' && l.input[start+1] == '1' && l.input[start+2] == '6') ||
			(l.input[start] == 'i' && l.input[start+1] == '3' && l.input[start+2] == '2') ||
			(l.input[start] == 'i' && l.input[start+1] == '6' && l.input[start+2] == '4') ||
			(l.input[start] == 's' && l.input[start+1] == 'i' && l.input[start+2] == '8') ||
			(l.input[start] == 'f' && l.input[start+1] == '3' && l.input[start+2] == '2') ||
			(l.input[start] == 'f' && l.input[start+1] == '6' && l.input[start+2] == '4')
	case 4:
		return (l.input[start] == 's' && l.input[start+1] == 'i' &&
			l.input[start+2] == '1' && l.input[start+3] == '6') ||
			(l.input[start] == 's' && l.input[start+1] == 'i' &&
				l.input[start+2] == '3' && l.input[start+3] == '2') ||
			(l.input[start] == 's' && l.input[start+1] == 'i' &&
				l.input[start+2] == '6' && l.input[start+3] == '4') ||
			(l.input[start] == 'c' && l.input[start+1] == 'h' &&
				l.input[start+2] == 'a' && l.input[start+3] == 'r') ||
			(l.input[start] == 's' && l.input[start+1] == 't' &&
				l.input[start+2] == 'r' && l.input[start+3] == '8')
	case 5:
		return (l.input[start] == 'c' && l.input[start+1] == 'o' &&
			l.input[start+2] == 'l' && l.input[start+3] == 'o' &&
			l.input[start+4] == 'r') ||
			(l.input[start] == 'e' && l.input[start+1] == 'q' &&
				l.input[start+2] == 'u' && l.input[start+3] == 'i' &&
				l.input[start+4] == 'p') ||
			(l.input[start] == 's' && l.input[start+1] == 't' &&
				l.input[start+2] == 'r' && l.input[start+3] == '1' &&
				l.input[start+4] == '6') ||
			(l.input[start] == 's' && l.input[start+1] == 't' &&
				l.input[start+2] == 'r' && l.input[start+3] == '3' &&
				l.input[start+4] == '2')
	case 6:
		return (l.input[start] == 'd' && l.input[start+1] == 'o' &&
			l.input[start+2] == 'u' && l.input[start+3] == 'b' &&
			l.input[start+4] == 'l' && l.input[start+5] == 'e')
	}
	return false
}

// Skips whitespace using byte operations
func (l *Lexer) skipWhitespace() {
	for l.pos < len(l.input) {
		ch := l.input[l.pos]
		if ch == ' ' || ch == '\t' || ch == '\n' || ch == '\r' {
			if ch == '\n' {
				l.line++
				l.col = 1
			} else {
				l.col++
			}
			l.pos++
		} else {
			break
		}
	}
}

// Optimized attribute parsing with minimal allocations - FIXED BUG
func (l *Lexer) parseAttributes(attrs map[string]string) error {
	// Clear existing attributes without deallocating
	for k := range attrs {
		delete(attrs, k)
	}

	for {
		l.skipWhitespace()
		if l.pos >= len(l.input) || l.peek() == '>' ||
			(l.peek() == '/' && l.pos+1 < len(l.input) && l.input[l.pos+1] == '>') {
			break
		}

		// Read attribute name using byte operations
		nameStart := l.pos
		for l.pos < len(l.input) {
			ch := l.input[l.pos]
			if ch == '=' || ch == ' ' || ch == '\t' || ch == '\n' || ch == '>' {
				break
			}
			l.pos++
			if ch != '\n' {
				l.col++
			}
		}

		nameEnd := l.pos // FIXED: Store end of name here

		if nameStart == nameEnd {
			break
		}

		l.skipWhitespace()
		if l.peek() != '=' {
			return fmt.Errorf("expected '=' after attribute name")
		}
		l.next() // skip '='
		l.skipWhitespace()

		// Read attribute value
		quote := l.peek()
		if quote != '"' && quote != '\'' {
			return fmt.Errorf("attribute value must be quoted")
		}
		l.next() // skip opening quote

		valueStart := l.pos
		for l.pos < len(l.input) && l.input[l.pos] != quote {
			if l.input[l.pos] == '\n' {
				l.line++
				l.col = 1
			} else {
				l.col++
			}
			l.pos++
		}

		if l.pos >= len(l.input) {
			return fmt.Errorf("unclosed attribute value")
		}

		// FIXED: Correct name and value extraction
		name := string(l.input[nameStart:nameEnd])
		value := string(l.input[valueStart:l.pos])
		attrs[name] = value

		l.next() // skip closing quote
	}

	return nil
}

// Optimized token generation with pooling
func (l *Lexer) NextToken() *Token {
	token := tokenPool.Get().(*Token)
	token.Type = TokenError
	token.TagStart = -1
	token.TagEnd = -1
	token.TextStart = -1
	token.TextEnd = -1
	token.Line = l.line
	token.Col = l.col

	l.skipWhitespace()
	if l.pos >= len(l.input) {
		token.Type = TokenEOF
		return token
	}

	if l.peek() == '<' {
		l.next() // skip '<'

		// Check for comment using byte comparison
		if l.pos+2 < len(l.input) &&
			l.input[l.pos] == '!' && l.input[l.pos+1] == '-' && l.input[l.pos+2] == '-' {
			l.pos += 3
			start := l.pos
			// Find end of comment efficiently
			for l.pos+2 < len(l.input) {
				if l.input[l.pos] == '-' && l.input[l.pos+1] == '-' && l.input[l.pos+2] == '>' {
					token.Type = TokenComment
					token.TextStart = start
					token.TextEnd = l.pos
					l.pos += 3
					return token
				}
				if l.input[l.pos] == '\n' {
					l.line++
					l.col = 1
				} else {
					l.col++
				}
				l.pos++
			}
			token.Type = TokenError
			return token
		}

		// Check for closing tag
		if l.peek() == '/' {
			l.next() // skip '/'
			start := l.pos
			for l.pos < len(l.input) && l.input[l.pos] != '>' {
				l.pos++
				l.col++
			}
			if l.pos >= len(l.input) {
				token.Type = TokenError
				return token
			}
			token.Type = TokenCloseTag
			token.TagStart = start
			token.TagEnd = l.pos
			l.next() // skip '>'
			return token
		}

		// Opening or self-closing tag
		start := l.pos
		for l.pos < len(l.input) {
			ch := l.input[l.pos]
			if ch == '>' || ch == '/' || ch == ' ' || ch == '\t' || ch == '\n' {
				break
			}
			l.pos++
			l.col++
		}

		if start == l.pos {
			token.Type = TokenError
			return token
		}

		token.TagStart = start
		token.TagEnd = l.pos

		if err := l.parseAttributes(token.Attributes); err != nil {
			token.Type = TokenError
			return token
		}

		l.skipWhitespace()
		if l.pos >= len(l.input) {
			token.Type = TokenError
			return token
		}

		if l.peek() == '/' && l.pos+1 < len(l.input) && l.input[l.pos+1] == '>' {
			token.Type = TokenSelfCloseTag
			l.pos += 2
		} else {
			// Check if this is a self-closing field type
			if l.isSelfClosingTag(token.TagStart, token.TagEnd) {
				token.Type = TokenSelfCloseTag
			} else {
				token.Type = TokenOpenTag
			}
			if l.peek() == '>' {
				l.next()
			} else {
				token.Type = TokenError
				return token
			}
		}

		return token
	}

	// Text content - find range without copying
	start := l.pos
	for l.pos < len(l.input) && l.input[l.pos] != '<' {
		if l.input[l.pos] == '\n' {
			l.line++
			l.col = 1
		} else {
			l.col++
		}
		l.pos++
	}

	// Trim whitespace from range
	for start < l.pos && unicode.IsSpace(rune(l.input[start])) {
		start++
	}
	end := l.pos
	for end > start && unicode.IsSpace(rune(l.input[end-1])) {
		end--
	}

	if start < end {
		token.Type = TokenText
		token.TextStart = start
		token.TextEnd = end
		return token
	}

	// Skip empty text, get next token
	return l.NextToken()
}

// Returns token to pool
func (l *Lexer) ReleaseToken(token *Token) {
	if token != nil {
		tokenPool.Put(token)
	}
}