# Lexer current state
var _state: Int = 0
- # Name of the stream (as given to tokens)
- readable var _filename: String
+ # The source file
+ readable var _file: SourceFile
- # Input stream where character are read
- var _stream: IStream
-
- # Pushback buffer to store unread character
- var _stream_buf: Buffer
-
- # Number of character stored in the pushback buffer
- var _stream_pos: Int
+ # Current character in the stream
+ var _stream_pos: Int = 0
# Current line number in the input stream
var _line: Int = 0
# Was the last character a cariage-return?
var _cr: Bool = false
- # If the end of stream?
- var _eof: Bool = false
-
- # Current working text read from the input stream
- var _text: Buffer
-
$ foreach {lexer_data/state}
# Constante state values
private fun state_${translate(@name,"ABCDEFGHIJKLMNOPQRSTUVWXYZ","abcdefghijklmnopqrstuvwxyz")}: Int do return @id end
$ end foreach
# Create a new lexer for a stream (and a name)
- init(stream: IStream, fname: String)
+ init(file: SourceFile)
do
- _filename = fname
- _text = new Buffer
- _stream = stream
- _stream_pos = -1
- _stream_buf = new Buffer
+ _file = file
end
+ # The last peeked token to chain them
+ private var last_token: nullable Token = null
+
# Give the next token (but do not consume it)
fun peek: Token
do
- while _token == null do
- _token = get_token
+ var t = _token
+ if t != null then return t
+
+ t = get_token
+ while t == null do t = get_token
+
+ var l = last_token
+ if l != null then
+ l.next_token = t
+ t.prev_token = l
end
- return _token.as(not null)
+
+ last_token = t
+ _token = t
+ return t
end
# Give and consume the next token
fun next: Token
do
- var result = _token
- while result == null do
- result = get_token
- end
+ var result = peek
_token = null
- return result.as(not null)
+ return result
end
- # Get a token, or null if it is discarded
- private fun get_token: nullable Token
+ # Primitive method to return a token, or return null if it is discarded
+ # Is used to implement `peek` and `next`
+ protected fun get_token: nullable Token
do
var dfa_state = 0
+ var sp = _stream_pos
+ var start_stream_pos = sp
var start_pos = _pos
var start_line = _line
+ var string = _file.string
+ var string_len = string.length
var accept_state = -1
var accept_token = -1
var accept_pos = -1
var accept_line = -1
- var text = _text
- text.clear
-
loop
- var c = get_char
+ if sp >= string_len then
+ dfa_state = -1
+ else
+ var c = string[sp].ascii
+ sp += 1
- if c != -1 then
var cr = _cr
var line = _line
var pos = _pos
if c == 10 then
if cr then
cr = false
+ _file.line_starts[line] = sp
else
line = line + 1
pos = 0
+ _file.line_starts[line] = sp
end
else if c == 13 then
line = line + 1
pos = 0
cr = true
+ _file.line_starts[line] = sp
else
pos = pos + 1
cr = false
end
- text.add(c.ascii)
-
loop
var old_state = dfa_state
if dfa_state < -1 then
_cr = cr
_line = line
_pos = pos
- else
- dfa_state = -1
end
if dfa_state >= 0 then
if tok != -1 then
accept_state = dfa_state
accept_token = tok
- accept_length = text.length
+ accept_length = sp - start_stream_pos
accept_pos = _pos
accept_line = _line
end
else
if accept_state != -1 then
- var location = new Location(_filename, start_line + 1, accept_line + 1, start_pos + 1, accept_pos)
+ var location = new Location(_file, start_line + 1, accept_line + 1, start_pos + 1, accept_pos)
_pos = accept_pos
_line = accept_line
- push_back(accept_length)
+ _stream_pos = start_stream_pos + accept_length
$ foreach {//token}
if accept_token == ${position()-1} then
$ if {count(transition[@from!=@to])!=0}
$ end
$ end if
$ if {@parser_index}
-$ if {not(@text)}
- var token_text = text.substring(0, accept_length)
- return new @ename.init_tk(token_text, location)
-$ else
return new @ename.init_tk(location)
-$ end
$ else
return null
$ end
end
$ end foreach
else
- var location = new Location(_filename, start_line + 1, start_line + 1, start_pos + 1, start_pos + 1)
- if text.length > 0 then
- var token = new PError.init_error("Syntax error: unknown token {text}.", location)
+ _stream_pos = sp
+ var location = new Location(_file, start_line + 1, start_line + 1, start_pos + 1, start_pos + 1)
+ if sp > start_stream_pos then
+ var text = string.substring(start_stream_pos, sp-start_stream_pos)
+ var token = new PLexerError.init_lexer_error("Syntax error: unknown token {text}.", location, text)
return token
else
- var token = new EOF(location)
+ var token = new EOF.init_tk(location)
return token
end
end
end
- if false then break # FIXME remove once unreach loop exits are in c_src
- end
- return null # FIXME remove once unreach loop exits are in c_src
- end
-
- # Read the next character.
- # The character is read from the stream of from the pushback buffer.
- private fun get_char: Int
- do
- if _eof then
- return -1
- end
-
- var result: Int
-
- var sp = _stream_pos
- if sp >= 0 then
- var res = _stream_buf[_stream_pos]
- _stream_pos = sp - 1
- result = res.ascii
- else
- result = _stream.read_char
- end
-
- if result == -1 then
- _eof = true
- end
-
- return result
- end
-
- # Unread some characters.
- # Unread characters are stored in the pushback buffer.
- private fun push_back(accept_length: Int)
- do
- var length = _text.length
- var i = length - 1
- while i >= accept_length do
- _eof = false
- _stream_pos = _stream_pos + 1
- _stream_buf[_stream_pos] = _text[i]
- i = i - 1
end
end
end