path: root/core/text/scanner/scanner.odin

// package text/scanner provides a scanner and tokenizer for UTF-8-encoded text.
// It takes a string providing the source, which then can be tokenized through
// repeated calls to the scan procedure.
// For compatibility with existing tooling and languages, the NUL character is not allowed.
// If an UTF-8 encoded byte order mark (BOM) is the first character in the source, it will be discarded.
//
// By default, a Scanner skips white space and Odin comments and recognizes all literals defined by the Odin programming language specification.
// A Scanner may be customized to recognize only a subset of those literals and to recognize different identifiers and white space characters.
package text_scanner

import "base:runtime"
import "core:fmt"
import "core:strings"
import "core:unicode"
import "core:unicode/utf8"

// Position represents a source position
// A position is valid if line > 0
Position :: struct {
	filename: string, // filename, if present
	offset:   int,    // byte offset, starting @ 0
	line:     int,    // line number, starting @ 1
	column:   int,    // column number, starting @ 1 (character count per line)
}

// position_is_valid reports where the position is valid
@(require_results)
position_is_valid :: proc(pos: Position) -> bool {
	return pos.line > 0
}

@(require_results)
position_to_string :: proc(pos: Position, allocator := context.temp_allocator) -> string {
	s := pos.filename
	if s == "" {
		s = "<input>"
	}

	context.allocator = allocator
	if position_is_valid(pos) {
		return fmt.aprintf("%s(%d:%d)", s, pos.line, pos.column)
	} else {
		return strings.clone(s)
	}
}

EOF        :: -1
Ident      :: -2
Int        :: -3
Float      :: -4
Char       :: -5
String     :: -6
Raw_String :: -7
Comment    :: -8

Scan_Flag :: enum u32 {
	Scan_Idents,
	Scan_Ints,
	Scan_C_Int_Prefixes,
	Scan_Floats, // Includes integers and hexadecimal floats
	Scan_Chars,
	Scan_Strings,
	Scan_Raw_Strings,
	Scan_Comments,
	Skip_Comments, // if set with .Scan_Comments, comments become white space
}
Scan_Flags :: distinct bit_set[Scan_Flag; u32]

Odin_Like_Tokens :: Scan_Flags{.Scan_Idents, .Scan_Ints, .Scan_Floats, .Scan_Chars, .Scan_Strings, .Scan_Raw_Strings, .Scan_Comments, .Skip_Comments}
C_Like_Tokens    :: Scan_Flags{.Scan_Idents, .Scan_Ints, .Scan_C_Int_Prefixes, .Scan_Floats, .Scan_Chars, .Scan_Strings, .Scan_Raw_Strings, .Scan_Comments, .Skip_Comments}

// Only allows for ASCII whitespace
Whitespace :: distinct bit_set['\x00'..<utf8.RUNE_SELF; u128]

// Odin_Whitespace is the default value for the Scanner's whitespace field
Odin_Whitespace :: Whitespace{'\t', '\n', '\r', ' '}
C_Whitespace    :: Whitespace{'\t', '\n', '\r', '\v', '\f', ' '}


// Scanner allows for the reading of Unicode characters and tokens from a string
Scanner :: struct {
	src: string,

	src_pos: int,
	src_end: int,

	tok_pos: int,
	tok_end: int,

	ch: rune,

	line:   int,
	column: int,
	prev_line_len: int,
	prev_char_len: int,

	// error is called for each error encountered
	// If no error procedure is set, the error is reported to os.stderr
	error: proc(s: ^Scanner, msg: string),

	// error_count is incremented by one for each error encountered
	error_count: int,

	// flags controls which tokens are recognized
	// e.g. to recognize integers, set the .Scan_Ints flag
	// This field may be changed by the user at any time during scanning
	flags: Scan_Flags,

	// The whitespace field controls which characters are recognized as white space
	// This field may be changed by the user at any time during scanning
	whitespace: Whitespace,

	// is_ident_rune is a predicate controlling the characters accepted as the ith rune in an identifier
	// The valid characters must not conflict with the set of white space characters
	// If is_ident_rune is not set, regular Odin-like identifiers are accepted
	// This field may be changed by the user at any time during scanning
	is_ident_rune: proc(ch: rune, i: int) -> bool,

	// Start position of most recently scanned token (set by scan(s))
	// Call init or next invalidates the position
	pos: Position,
}

// init initializes a scanner with a new source and returns itself.
// error_count is set to 0, flags is set to Odin_Like_Tokens, whitespace is set to Odin_Whitespace
init :: proc(s: ^Scanner, src: string, filename := "") -> ^Scanner {
	s^ = {}

	s.error_count = 0
	s.src = src
	s.pos.filename = filename

	s.tok_pos = -1

	s.ch = -2 // no char read yet, not an EOF

	s.line = 1

	s.flags = Odin_Like_Tokens
	s.whitespace = Odin_Whitespace

	return s
}


@(private, require_results)
advance :: proc(s: ^Scanner) -> rune {
	if s.src_pos >= len(s.src) {
		s.prev_char_len = 0
		return EOF
	}
	ch, width := rune(s.src[s.src_pos]), 1

	if ch >= utf8.RUNE_SELF {
		ch, width = utf8.decode_rune_in_string(s.src[s.src_pos:])
		if ch == utf8.RUNE_ERROR && width == 1 {
			s.src_pos += width
			s.prev_char_len = width
			s.column += 1
			error(s, "invalid UTF-8 encoding")
			return ch
		}
	}

	s.src_pos += width
	s.prev_char_len = width
	s.column += 1

	switch ch {
	case 0:
		error(s, "invalid character NUL")
	case '\n':
		s.line += 1
		s.prev_line_len = s.column
		s.column = 0
	}

	return ch
}

// next reads and returns the next Unicode character. It returns EOF at the end of the source.
// next does not update the Scanner's pos field. Use 'position(s)' to get the current position
next :: proc(s: ^Scanner) -> rune {
	s.tok_pos = -1
	s.pos.line = 0
	ch := peek(s)
	if ch != EOF {
		s.ch = advance(s)
	}
	return ch
}

// peek returns the next Unicode character in the source without advancing the scanner
// It returns EOF if the scanner's position is at least the last character of the source
// if n > 0, it call next n times and return the nth Unicode character and then restore the Scanner's state
@(require_results)
peek :: proc(s: ^Scanner, n := 0) -> (ch: rune) {
	if s.ch == -2 {
		s.ch = advance(s)
		if s.ch == '\ufeff' { // Ignore BOM
			s.ch = advance(s)
		}
	}
	ch = s.ch
	if n > 0 {
		prev_s := s^
		for _ in 0..<n {
			next(s)
		}
		ch = s.ch
		s^ = prev_s
	}
	return ch
}
// peek returns the next token in the source
// It returns EOF if the scanner's position is at least the last character of the source
// if n > 0, it call next n times and return the nth token and then restore the Scanner's state
@(require_results)
peek_token :: proc(s: ^Scanner, n := 0) -> (tok: rune) {
	assert(n >= 0)
	prev_s := s^
	for _ in 0..<n {
		tok = scan(s)
	}
	tok = scan(s)
	s^ = prev_s
	return
}

error :: proc(s: ^Scanner, msg: string) {
	s.error_count += 1
	if s.error != nil {
		s.error(s, msg)
		return
	}
	p := s.pos
	if !position_is_valid(p) {
		p = position(s)
	}

	s := p.filename
	if s == "" {
		s = "<input>"
	}

	if position_is_valid(p) {
		fmt.eprintf("%s(%d:%d): %s\n", s, p.line, p.column, msg)
	} else {
		fmt.eprintf("%s: %s\n", s, msg)
	}
}

errorf :: proc(s: ^Scanner, format: string, args: ..any) {
	error(s, fmt.tprintf(format, ..args))
}

@(private, require_results)
is_ident_rune :: proc(s: ^Scanner, ch: rune, i: int) -> bool {
	if s.is_ident_rune != nil {
		return s.is_ident_rune(ch, i)
	}
	return ch == '_' || unicode.is_letter(ch) || unicode.is_digit(ch) && i > 0
}

@(private, require_results)
scan_identifier :: proc(s: ^Scanner) -> rune {
	ch := advance(s)
	for i := 1; is_ident_rune(s, ch, i); i += 1 {
		ch = advance(s)
	}
	return ch
}

@(private, require_results) lower      :: proc(ch: rune) -> rune { return ('a' - 'A') | ch }
@(private, require_results) is_decimal :: proc(ch: rune) -> bool { return '0' <= ch && ch <= '9' }
@(private, require_results) is_hex     :: proc(ch: rune) -> bool { return '0' <= ch && ch <= '9' || 'a' <= lower(ch) && lower(ch) <= 'f' }



@(private, require_results)
scan_number :: proc(s: ^Scanner, ch: rune, seen_dot: bool) -> (rune, rune) {
	lit_name :: proc(prefix: rune) -> string {
		switch prefix {
		case 'b': return "binary literal"
		case 'o': return "octal literal"
		case 'z': return "dozenal literal"
		case 'x': return "hexadecimal literal"
		case 'h': return "hexadecimal literal"
		}
		return "decimal literal"
	}

	digits :: proc(s: ^Scanner, ch0: rune, base: int, invalid: ^rune) -> (ch: rune, digsep: int) {
		ch = ch0
		if base <= 10 {
			max := rune('0' + base)
			for is_decimal(ch) || ch == '_' {
				ds := 1
				if ch == '_' {
					ds = 2
				} else if ch >= max && invalid^ == 0 {
					invalid^ = ch
				}
				digsep |= ds
				ch = advance(s)
			}
		} else {
			for is_hex(ch) || ch == '_' {
				ds := 1
				if ch == '_' {
					ds = 2
				}
				digsep |= ds
				ch = advance(s)
			}
		}
		return
	}

	ch, seen_dot := ch, seen_dot

	base := 10
	prefix := rune(0)
	digsep := 0
	invalid := rune(0)

	tok: rune
	ds: int

	if !seen_dot {
		tok = Int
		if ch == '0' {
			ch = advance(s)

			p := lower(ch)
			if .Scan_C_Int_Prefixes in s.flags {
				switch p {
				case 'b':
					ch = advance(s)
					base, prefix = 2, 'b'
				case 'x':
					ch = advance(s)
					base, prefix = 16, 'x'
				case:
					base, prefix = 8, 'o'
					digsep = 1 // Leading zero
				}
			} else {
				switch p {
				case 'b':
					ch = advance(s)
					base, prefix = 2, 'b'
				case 'o':
					ch = advance(s)
					base, prefix = 8, 'o'
				case 'd':
					ch = advance(s)
					base, prefix = 10, 'd'
				case 'z':
					ch = advance(s)
					base, prefix = 12, 'z'
				case 'h':
					tok = Float
					ch = advance(s)
					base, prefix = 16, 'h'
				case 'x':
					ch = advance(s)
					base, prefix = 16, 'x'
				case:
					digsep = 1 // Leading zero
				}
			}
		}

		ch, ds = digits(s, ch, base, &invalid)
		digsep |= ds
		if ch == '.' && .Scan_Floats in s.flags {
			ch = advance(s)
			seen_dot = true
		}
	}

	if seen_dot {
		tok = Float
		if prefix != 0 && prefix != 'x' {
			errorf(s, "invalid radix point in %s", lit_name(prefix))
		}
		ch, ds = digits(s, ch, base, &invalid)
		digsep |= ds
	}

	if digsep&1 == 0 {
		errorf(s, "%s has no digits", lit_name(prefix))
	}

	if e := lower(ch); (e == 'e' || e == 'p') && .Scan_Floats in s.flags {
		switch {
		case e == 'e' && prefix != 0:
			errorf(s, "%q exponent requires decimal mantissa", ch)
		case e == 'p' && prefix != 'x':
			errorf(s, "%q exponent requires hexadecimal mantissa", ch)
		}
		ch = advance(s)
		tok = Float
		if ch == '+' || ch == '-' {
			ch = advance(s)
		}
		ch, ds = digits(s, ch, 10, nil)
		digsep |= ds
		if ds&1 == 0 {
			error(s, "exponent has no digits")
		}
	} else if prefix == 'x' && tok == Float {
		error(s, "hexadecimal mantissa requires a 'p' exponent")
	}

	if tok == Int && invalid != 0 {
		errorf(s, "invalid digit %q in %s", invalid, lit_name(prefix))
	}

	if digsep&2 != 0 {
		s.tok_end = s.src_pos - s.prev_char_len
	}
	return tok, ch
}

@(private, require_results)
scan_string :: proc(s: ^Scanner, quote: rune) -> (n: int) {
	digit_val :: proc(ch: rune) -> int {
		switch v := lower(ch); v {
		case '0'..='9': return int(v - '0')
		case 'a'..='z': return int(v - 'a')
		}
		return 16
	}

	scan_digits :: proc(s: ^Scanner, ch: rune, base, n: int) -> rune {
		ch, n := ch, n
		for n > 0 && digit_val(ch) < base {
			ch = advance(s)
			n -= 1
		}
		if n > 0 {
			error(s, "invalid char escape")
		}
		return ch
	}

	ch := advance(s)
	for ch != quote {
		if ch == '\n' || ch < 0 {
			error(s, "literal not terminated")
			return
		}
		if ch == '\\' {
			ch = advance(s)
			switch ch {
			case quote, 'a', 'b', 'e', 'f', 'n', 'r', 't', 'v', '\\':
				ch = advance(s)
			case '0'..='7': ch = scan_digits(s, advance(s), 8, 3)
			case 'x':       ch = scan_digits(s, advance(s), 16, 2)
			case 'u':       ch = scan_digits(s, advance(s), 16, 4)
			case 'U':       ch = scan_digits(s, advance(s), 16, 8)
			case:
				error(s, "invalid char escape")
			}
		} else {
			ch = advance(s)
		}
		n += 1
	}
	return
}

@(private)
scan_raw_string :: proc(s: ^Scanner) {
	ch := advance(s)
	for ch != '`' {
		if ch < 0 {
			error(s, "literal not terminated")
			return
		}
		ch = advance(s)
	}
}

@(private)
scan_char :: proc(s: ^Scanner) {
	if scan_string(s, '\'') != 1 {
		error(s, "invalid char literal")
	}
}

@(private, require_results)
scan_comment :: proc(s: ^Scanner, ch: rune) -> rune {
	ch := ch
	if ch == '/' { // line comment
		ch = advance(s)
		for ch != '\n' && ch >= 0 {
			ch = advance(s)
		}
		return ch
	}

	// block /**/ comment
	ch = advance(s)
	for {
		if ch < 0 {
			error(s, "comment not terminated")
			break
		}
		ch0 := ch
		ch = advance(s)
		if ch0 == '*' && ch == '/' {
			return advance(s)
		}
	}
	return ch
}

// scan reads the next token or Unicode character from source and returns it
// It only recognizes tokens for which the respective flag that is set
// It returns EOF at the end of the source
// It reports Scanner errors by calling s.error, if not nil; otherwise it will print the error message to os.stderr
scan :: proc(s: ^Scanner) -> (tok: rune) {
	ch := peek(s)
	if ch == EOF {
		return ch
	}

	// reset position
	s.tok_pos = -1
	s.pos.line = 0

	redo: for {
		for ch < utf8.RUNE_SELF && (ch in s.whitespace) {
			ch = advance(s)
		}

		s.tok_pos = s.src_pos - s.prev_char_len
		s.pos.offset = s.tok_pos

		if s.column > 0 {
			s.pos.line = s.line
			s.pos.column = s.column
		} else {
			// previous character was newline
			s.pos.line = s.line - 1
			s.pos.column = s.prev_line_len
		}

		tok = ch
		if is_ident_rune(s, ch, 0) {
			if .Scan_Idents in s.flags {
				tok = Ident
				ch = scan_identifier(s)
			} else {
				ch = advance(s)
			}

		} else if is_decimal(ch) {
			if .Scan_Ints in s.flags || .Scan_Floats in s.flags {
				tok, ch = scan_number(s, ch, false)
			} else {
				ch = advance(s)
			}
		} else {
			switch ch {
			case EOF:
				break
			case '"':
				if .Scan_Strings in s.flags {
					_ = scan_string(s, '"')
					tok = String
				}
				ch = advance(s)
			case '\'':
				if .Scan_Chars in s.flags {
					_ = scan_string(s, '\'')
					tok = Char
				}
				ch = advance(s)
			case '`':
				if .Scan_Raw_Strings in s.flags {
					scan_raw_string(s)
					tok = Raw_String
				}
				ch = advance(s)
			case '.':
				ch = advance(s)
				if is_decimal(ch) && .Scan_Floats in s.flags {
					tok, ch = scan_number(s, ch, true)
				}
			case '/':
				ch = advance(s)
				if (ch == '/' || ch == '*') && .Scan_Comments in s.flags {
					if .Skip_Comments in s.flags {
						s.tok_pos = -1
						ch = scan_comment(s, ch)
						continue redo
					}
					ch = scan_comment(s, ch)
					tok = Comment
				}
			case:
				ch = advance(s)
			}
		}

		break redo
	}

	s.tok_end = s.src_pos - s.prev_char_len

	s.ch = ch
	return tok
}

// position returns the position of the character immediately after the character or token returns by the previous call to next or scan
// Use the Scanner's position field for the most recently scanned token position
@(require_results)
position :: proc(s: ^Scanner) -> Position {
	pos: Position
	pos.filename = s.pos.filename
	pos.offset = s.src_pos - s.prev_char_len
	switch {
	case s.column > 0:
		pos.line = s.line
		pos.column = s.column
	case s.prev_line_len > 0:
		pos.line = s.line-1
		pos.column = s.prev_line_len
	case:
		pos.line = 1
		pos.column = 1
	}
	return pos
}

// token_text returns the string of the most recently scanned token
@(require_results)
token_text :: proc(s: ^Scanner) -> string {
	if s.tok_pos < 0 {
		return ""
	}
	return string(s.src[s.tok_pos:s.tok_end])
}

// token_string returns a printable string for a token or Unicode character
// By default, it uses the context.temp_allocator to produce the string
@(require_results)
token_string :: proc(tok: rune, allocator: runtime.Allocator) -> string {
	context.allocator = allocator
	switch tok {
	case EOF:        return strings.clone("EOF")
	case Ident:      return strings.clone("Ident")
	case Int:        return strings.clone("Int")
	case Float:      return strings.clone("Float")
	case Char:       return strings.clone("Char")
	case String:     return strings.clone("String")
	case Raw_String: return strings.clone("Raw_String")
	case Comment:    return strings.clone("Comment")
	}
	return fmt.aprintf("%q", tok)
}