Merge: doc: fixed some typos and other misc. corrections

[nit.git] / src / parser / lexer_work.nit

diff --git a/src/parser/lexer_work.nit b/src/parser/lexer_work.nit
index e7102b6..e061613 100644 (file)
--- a/src/parser/lexer_work.nit
+++ b/src/parser/lexer_work.nit
@@ -39,11 +39,6 @@ redef class Token
 end
 
 redef class EOF
-    redef fun parser_index: Int
-    do
-       return 97
-    end
-
     init init_tk(loc: Location)
     do
         _cached_text = ""
@@ -85,8 +80,9 @@ end
 # It is better user with the Parser
 class Lexer
        super TablesCapable
+
        # Last peeked token
-       var token: nullable Token
+       var token: nullable Token = null
 
        # Lexer current state
        private var state: Int = 0
@@ -103,18 +99,12 @@ class Lexer
        # Current column in the input stream
        var pos: Int = 0
 
-       # Was the last character a cariage-return?
+       # Was the last character a carriage-return?
        var cr: Bool = false
 
        # Constante state values
        private fun state_initial: Int do return 0 end
 
-       # Create a new lexer for a stream (and a name)
-       init(file: SourceFile)
-       do
-               self.file = file
-       end
-
        # The last peeked token to chain them
        private var last_token: nullable Token = null
 
@@ -174,7 +164,17 @@ class Lexer
                        if sp >= string_len then
                                dfa_state = -1
                        else
-                               var c = string[sp].ascii
+                               # Very ugly hack, this is because of the way SableCC generates its tables.
+                               # Due to the 0xFFFF limit of a Java char, when a big Nit char is read (i.e.
+                               # code point > 65535), it crashes.
+                               #
+                               # Hence, if a char has a code point <= 255 (ISO8859 range), it is left as is.
+                               # Else, it is replaced by 255.
+                               # This does not corrupt the lexer and works perfectly on any character.
+                               #
+                               # TL;DR: Java fucked up, need retarded solution to cope for retarded decision
+                               var c = string[sp].code_point
+                               if c >= 256 then c = 255
                                sp += 1
 
                                var cr = _cr
@@ -244,20 +244,21 @@ class Lexer
                                end
                        else
                                if accept_state != -1 then
-                                       var location = new Location(file, start_line + 1, accept_line + 1, start_pos + 1, accept_pos)
                                        _pos = accept_pos
                                        _line = accept_line
                                        _stream_pos = start_stream_pos + accept_length
                                        if accept_token == 0 then
+                                               # Ignored token (whitespaces)
                                                return null
                                        end
+                                       var location = new Location(file, start_line + 1, accept_line + 1, start_pos + 1, accept_pos)
                                        return make_token(accept_token, location)
                                else
                                        _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 ALexerError.init_lexer_error("Syntax error: unknown token {text}.", location, text)
+                                               var token = new ALexerError.init_lexer_error("Syntax Error: unknown token `{text}`.", location, text)
                                                file.last_token = token
                                                return token
                                        else