lexer.mll 7.19 KB
Newer Older
1 2
(**************************************************************************)
(*                                                                        *)
Jean-Christophe Filliâtre's avatar
headers  
Jean-Christophe Filliâtre committed
3 4 5 6 7
(*  Copyright (C) 2010-                                                   *)
(*    Francois Bobot                                                      *)
(*    Jean-Christophe Filliatre                                           *)
(*    Johannes Kanig                                                      *)
(*    Andrei Paskevich                                                    *)
8 9 10 11 12 13 14 15 16 17 18
(*                                                                        *)
(*  This software is free software; you can redistribute it and/or        *)
(*  modify it under the terms of the GNU Library General Public           *)
(*  License version 2.1, with the special exception on linking            *)
(*  described in file LICENSE.                                            *)
(*                                                                        *)
(*  This software is distributed in the hope that it will be useful,      *)
(*  but WITHOUT ANY WARRANTY; without even the implied warranty of        *)
(*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.                  *)
(*                                                                        *)
(**************************************************************************)
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
19 20

{
21
  open Format
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
22
  open Lexing
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
23
  open Term
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
24 25 26
  open Ptree
  open Parser

27 28 29 30 31 32 33
  (* lexical errors *)

  type error = 
    | IllegalCharacter of char
    | UnterminatedComment
    | UnterminatedString

34
  exception Error of error
35 36 37 38 39 40

  let report fmt = function
    | IllegalCharacter c -> fprintf fmt "illegal character %c" c
    | UnterminatedComment -> fprintf fmt "unterminated comment"
    | UnterminatedString -> fprintf fmt "unterminated string"

41 42 43 44 45 46 47

  let () = Exn_printer.register
    (fun fmt exn -> match exn with
      | Error error -> report fmt error
      | _ -> raise exn)


Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
48 49 50 51
  let keywords = Hashtbl.create 97
  let () = 
    List.iter 
      (fun (x,y) -> Hashtbl.add keywords x y)
52
      [ 
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
53 54 55
	"and", AND;
	"as", AS;
	"axiom", AXIOM;
56
	"clone", CLONE;
57
	"else", ELSE;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
58
	"end", END;
Andrei Paskevich's avatar
Andrei Paskevich committed
59
	"epsilon", EPSILON;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
60
	"exists", EXISTS;
61
	"export", EXPORT;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
62 63 64 65
        "false", FALSE;
	"forall", FORALL;
	"goal", GOAL;
	"if", IF;
66
	"import", IMPORT;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
67 68
	"in", IN;
	"inductive", INDUCTIVE;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
69
	"lemma", LEMMA;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
70 71 72
	"let", LET;
	"logic", LOGIC;
	"match", MATCH;
73
	"meta", META;
74
	"namespace", NAMESPACE;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
75 76
	"not", NOT;
	"or", OR;
77
        "prop", PROP;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
78
	"then", THEN;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
79
	"theory", THEORY;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
80 81
	"true", TRUE;
	"type", TYPE;
82
	"use", USE;
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
83
	"with", WITH;
84
      ]
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
85 86 87 88 89 90
	       
  let newline lexbuf =
    let pos = lexbuf.lex_curr_p in
    lexbuf.lex_curr_p <- 
      { pos with pos_lnum = pos.pos_lnum + 1; pos_bol = pos.pos_cnum }

91
  let string_start_loc = ref Loc.dummy_position
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
92 93
  let string_buf = Buffer.create 1024

94 95
  let comment_start_loc = ref Loc.dummy_position

Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114
  let char_for_backslash = function
    | 'n' -> '\n'
    | 't' -> '\t'
    | c -> c

  let update_loc lexbuf file line chars =
    let pos = lexbuf.lex_curr_p in
    let new_file = match file with None -> pos.pos_fname | Some s -> s in
    lexbuf.lex_curr_p <- 
      { pos with
	  pos_fname = new_file;
	  pos_lnum = int_of_string line;
	  pos_bol = pos.pos_cnum - int_of_string chars;
      }

  let remove_leading_plus s =
    let n = String.length s in 
    if n > 0 && s.[0] = '+' then String.sub s 1 (n-1) else s

115 116
  let loc lb = (lexeme_start_p lb, lexeme_end_p lb)

Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
117 118 119 120
}

let newline = '\n'
let space = [' ' '\t' '\r']
121 122 123
let lalpha = ['a'-'z' '_']
let ualpha = ['A'-'Z']
let alpha = lalpha | ualpha
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
124
let digit = ['0'-'9']
125 126
let lident = lalpha (alpha | digit | '\'')*
let uident = ualpha (alpha | digit | '\'')*
127 128 129 130 131 132 133 134 135 136
let decimal_literal =
  ['0'-'9'] ['0'-'9' '_']*
let hex_literal =
  '0' ['x' 'X'] ['0'-'9' 'A'-'F' 'a'-'f']['0'-'9' 'A'-'F' 'a'-'f' '_']*
let oct_literal =
  '0' ['o' 'O'] ['0'-'7'] ['0'-'7' '_']*
let bin_literal =
  '0' ['b' 'B'] ['0'-'1'] ['0'-'1' '_']*
let int_literal =
  decimal_literal | hex_literal | oct_literal | bin_literal
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
137 138
let hexadigit = ['0'-'9' 'a'-'f' 'A'-'F']

139
let op_char_1 = ['=' '<' '>' '~']
140 141
let op_char_2 = ['+' '-']
let op_char_3 = ['*' '/' '%']
142
let op_char_4 = ['!' '$' '&' '?' '@' '^' '.' ':' '|' '#']
143 144 145 146
let op_char_34 = op_char_3 | op_char_4
let op_char_234 = op_char_2 | op_char_34
let op_char_1234 = op_char_1 | op_char_234

147 148
let op_char_pref = ['!' '?']

Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
149 150 151 152 153 154 155 156
rule token = parse
  | "#" space* ("\"" ([^ '\010' '\013' '"' ]* as file) "\"")?
    space* (digit+ as line) space* (digit+ as char) space* "#"
      { update_loc lexbuf file line char; token lexbuf }
  | newline 
      { newline lexbuf; token lexbuf }
  | space+  
      { token lexbuf }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
157 158
  | '_'
      { UNDERSCORE }
159 160 161 162
  | lident as id  
      { try Hashtbl.find keywords id with Not_found -> LIDENT id }
  | uident as id  
      { UIDENT id }
163
  | int_literal as s
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
164 165 166 167
      { INTEGER s }
  | (digit+ as i) ("" as f) ['e' 'E'] (['-' '+']? digit+ as e)
  | (digit+ as i) '.' (digit* as f) (['e' 'E'] (['-' '+']? digit+ as e))?
  | (digit* as i) '.' (digit+ as f) (['e' 'E'] (['-' '+']? digit+ as e))?
168
      { FLOAT (RConstDecimal (i, f, Util.option_map remove_leading_plus e)) }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
169 170 171 172 173
  | '0' ['x' 'X'] ((hexadigit* as i) '.' (hexadigit+ as f) 
                  |(hexadigit+ as i) '.' (hexadigit* as f)
		  |(hexadigit+ as i) ("" as f))
    ['p' 'P'] (['-' '+']? digit+ as e)
      { FLOAT (RConstHexa (i, f, remove_leading_plus e)) }
174 175
  | "(*)"
      { LEFTPAR_STAR_RIGHTPAR }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
176
  | "(*"
177
      { comment_start_loc := loc lexbuf; comment lexbuf; token lexbuf }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193
  | "'"
      { QUOTE }
  | ","
      { COMMA }
  | "("
      { LEFTPAR }
  | ")"
      { RIGHTPAR }
  | ":"
      { COLON }
  | "->"
      { ARROW }
  | "<->"
      { LRARROW }
  | "."
      { DOT }
194 195 196 197
  | "|"
      { BAR }
  | "="
      { EQUAL }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
198 199 200 201
  | "["
      { LEFTSQ }
  | "]"
      { RIGHTSQ }
202 203
  | op_char_pref op_char_4* as s
      { OPPREF s }
204 205 206 207 208 209 210 211
  | op_char_1234* op_char_1 op_char_1234* as s
      { OP1 s }
  | op_char_234*  op_char_2 op_char_234*  as s
      { OP2 s }
  | op_char_34*   op_char_3 op_char_34*  as s
      { OP3 s }
  | op_char_4+ as s
      { OP4 s }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
212
  | "\""
213
      { string_start_loc := loc lexbuf; STRING (string lexbuf) }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
214 215 216
  | eof 
      { EOF }
  | _ as c
217
      { raise (Error (IllegalCharacter c)) }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
218 219 220 221 222 223 224 225 226

and comment = parse
  | "*)" 
      { () }
  | "(*" 
      { comment lexbuf; comment lexbuf }
  | newline 
      { newline lexbuf; comment lexbuf }
  | eof
227
      { raise (Loc.Located (!comment_start_loc, Error UnterminatedComment)) }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
228 229 230 231 232
  | _ 
      { comment lexbuf }

and string = parse
  | "\""
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
233 234 235
      { let s = Buffer.contents string_buf in
	Buffer.clear string_buf; 
	s }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
236 237 238 239 240
  | "\\" (_ as c)
      { Buffer.add_char string_buf (char_for_backslash c); string lexbuf }
  | newline 
      { newline lexbuf; Buffer.add_char string_buf '\n'; string lexbuf }
  | eof
241
      { raise (Loc.Located (!string_start_loc, Error UnterminatedString)) }
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
242 243 244 245 246 247 248 249
  | _ as c
      { Buffer.add_char string_buf c; string lexbuf }



{

  let with_location f lb =
250 251 252 253 254
    try 
      f lb 
    with 
      | Loc.Located _ as e -> raise e
      | e -> raise (Loc.Located (loc lb, e))
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
255

256 257
  let parse_logic_file = with_location (logic_file token)

Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
258
  let parse_lexpr = with_location (lexpr token)
259
  let parse_list0_decl = with_location (list0_decl token)
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
260 261 262 263 264

}

(*
Local Variables: 
265
compile-command: "unset LANG; make -C ../.. test"
Jean-Christophe Filliâtre's avatar
Jean-Christophe Filliâtre committed
266 267 268
End: 
*)