1 (* Copyright (C) 2005, HELM Team.
3 * This file is part of HELM, an Hypertextual, Electronic
4 * Library of Mathematics, developed at the Computer Science
5 * Department, University of Bologna, Italy.
7 * HELM is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation; either version 2
10 * of the License, or (at your option) any later version.
12 * HELM is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 * GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with HELM; if not, write to the Free Software
19 * Foundation, Inc., 59 Temple Place - Suite 330, Boston,
22 * For details, see the HELM World-Wide-Web page,
23 * http://helm.cs.unibo.it/
31 exception Parse_error of Token.flocation * string
32 exception Level_not_found of int
34 let grammar = Grammar.gcreate CicNotationLexer.notation_lexer
36 let min_precedence = 0
37 let max_precedence = 100
38 let default_precedence = 50
40 let level1_pattern = Grammar.Entry.create grammar "level1_pattern"
41 let level2_pattern = Grammar.Entry.create grammar "level2_pattern"
42 let level3_term = Grammar.Entry.create grammar "level3_term"
43 let l2_pattern = Grammar.Entry.create grammar "l2_pattern"
44 let notation = Grammar.Entry.create grammar "notation" (* level1 <-> level 2 *)
46 Grammar.Entry.create grammar "interpretation" (* level2 <-> level 3 *)
47 let phrase = Grammar.Entry.create grammar "phrase"
49 let return_term loc term = ()
52 let (x, y) = loc_of_floc floc in
53 failwith (sprintf "Error at characters %d - %d: %s" x y msg)
57 Pervasives.int_of_string s
59 failwith (sprintf "Lexer failure: string_of_int \"%s\" failed" s)
61 (** {2 Grammar extension} *)
63 let symbol s = Gramext.Stoken ("SYMBOL", s)
64 let ident s = Gramext.Stoken ("IDENT", s)
65 let number s = Gramext.Stoken ("NUMBER", s)
66 let term = Gramext.Sself
68 let g_symbol_of_literal =
70 | `Symbol s -> symbol s
71 | `Keyword s -> ident s
72 | `Number s -> number s
76 | Binding of string * value_type
77 | Env of (string * value_type) list
79 let make_action action bindings =
80 let rec aux (vl : env_type) =
83 prerr_endline "aux: make_action";
84 Gramext.action (fun (loc: location) -> action vl loc)
86 prerr_endline "aux: none";
87 Gramext.action (fun _ -> aux vl tl)
88 (* LUCA: DEFCON 2 BEGIN *)
89 | Binding (name, TermType) :: tl ->
90 prerr_endline "aux: term";
92 (fun (v:term) -> aux ((name, (TermType, (TermValue v)))::vl) tl)
93 | Binding (name, StringType) :: tl ->
94 prerr_endline "aux: string";
97 aux ((name, (StringType, (StringValue v))) :: vl) tl)
98 | Binding (name, NumType) :: tl ->
99 prerr_endline "aux: num";
101 (fun (v:string) -> aux ((name, (NumType, (NumValue v))) :: vl) tl)
102 | Binding (name, OptType t) :: tl ->
103 prerr_endline "aux: opt";
105 (fun (v:'a option) ->
106 aux ((name, (OptType t, (OptValue v))) :: vl) tl)
107 | Binding (name, ListType t) :: tl ->
108 prerr_endline "aux: list";
111 aux ((name, (ListType t, (ListValue v))) :: vl) tl)
113 prerr_endline "aux: env";
114 Gramext.action (fun (v:env_type) -> aux (v @ vl) tl)
115 (* LUCA: DEFCON 2 END *)
117 aux [] (List.rev bindings)
123 | NoBinding :: tl -> aux acc tl
124 | Env names :: tl -> aux (List.rev names @ acc) tl
125 | Binding (name, ty) :: tl -> aux ((name, ty) :: acc) tl
129 (* given a level 1 pattern computes the new RHS of "term" grammar entry *)
130 let extract_term_production pattern =
131 let rec aux = function
132 | AttributedTerm (_, t) -> aux t
133 | Literal l -> aux_literal l
134 | Layout l -> aux_layout l
135 | Magic m -> aux_magic m
136 | Variable v -> aux_variable v
138 prerr_endline (CicNotationPp.pp_term t);
142 | `Symbol s -> [NoBinding, symbol s]
143 | `Keyword s -> [NoBinding, ident s]
144 | `Number s -> [NoBinding, number s]
145 and aux_layout = function
146 | Sub (p1, p2) -> aux p1 @ [NoBinding, symbol "\\SUB"] @ aux p2
147 | Sup (p1, p2) -> aux p1 @ [NoBinding, symbol "\\SUP"] @ aux p2
148 | Below (p1, p2) -> aux p1 @ [NoBinding, symbol "\\BELOW"] @ aux p2
149 | Above (p1, p2) -> aux p1 @ [NoBinding, symbol "\\ABOVE"] @ aux p2
150 | Frac (p1, p2) -> aux p1 @ [NoBinding, symbol "\\FRAC"] @ aux p2
151 | Atop (p1, p2) -> aux p1 @ [NoBinding, symbol "\\ATOP"] @ aux p2
152 | Over (p1, p2) -> aux p1 @ [NoBinding, symbol "\\OVER"] @ aux p2
154 [NoBinding, symbol "\\ROOT"] @ aux p2 @ [NoBinding, symbol "\\OF"]
156 | Sqrt p -> [NoBinding, symbol "\\SQRT"] @ aux p
158 | Box (_, pl) -> List.flatten (List.map aux pl)
159 and aux_magic magic =
162 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
163 let action (env_opt : env_type option) (loc : location) =
165 | Some env -> List.map opt_binding_some env
166 | None -> List.map opt_binding_of_name p_names
168 [ Env (List.map opt_declaration p_names),
170 [ [ Gramext.Sopt (Gramext.srules [ p_atoms, p_action ]) ],
171 Gramext.action action ] ]
174 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
175 let env0 = List.map list_binding_of_name p_names in
176 let grow_env_entry env n v =
177 prerr_endline "grow_env_entry";
180 | (n', (ty, ListValue vl)) as entry ->
181 if n' = n then n', (ty, ListValue (v :: vl)) else entry
185 let grow_env env_i env =
186 prerr_endline "grow_env";
188 (fun env (n, (_, v)) -> grow_env_entry env n v)
191 let action (env_list : env_type list) (loc : location) =
192 prerr_endline "list action";
193 List.fold_right grow_env env_list env0
197 | List0 (_, None) -> Gramext.Slist0 s
198 | List1 (_, None) -> Gramext.Slist1 s
199 | List0 (_, Some l) -> Gramext.Slist0sep (s, g_symbol_of_literal l)
200 | List1 (_, Some l) -> Gramext.Slist1sep (s, g_symbol_of_literal l)
203 [ Env (List.map list_declaration p_names),
205 [ [ g_symbol (Gramext.srules [ p_atoms, p_action ]) ],
206 Gramext.action action ] ]
210 | NumVar s -> [Binding (s, NumType), number ""]
211 | TermVar s -> [Binding (s, TermType), term]
212 | IdentVar s -> [Binding (s, StringType), ident ""]
213 | Ascription (p, s) -> assert false (* TODO *)
214 | FreshVar _ -> assert false
215 and inner_pattern p =
216 let p_bindings, p_atoms = List.split (aux p) in
217 let p_names = flatten_opt p_bindings in
218 let _ = prerr_endline ("inner names: " ^ String.concat " " (List.map fst p_names)) in
220 make_action (fun (env : env_type) (loc : location) -> prerr_endline "inner action"; env) p_bindings
222 p_bindings, p_atoms, p_names, action
226 let level_of_int precedence =
227 if precedence < min_precedence || precedence > max_precedence then
228 raise (Level_not_found precedence);
229 string_of_int precedence
231 type rule_id = Token.t Gramext.g_symbol list
233 let extend level1_pattern ?(precedence = default_precedence)
234 ?associativity action
236 let p_bindings, p_atoms =
237 List.split (extract_term_production level1_pattern)
239 let level = level_of_int precedence in
240 let p_names = flatten_opt p_bindings in
242 prerr_endline (string_of_int (List.length p_bindings));
244 [ Grammar.Entry.obj (l2_pattern: 'a Grammar.Entry.e),
245 Some (Gramext.Level level),
250 (fun (env: env_type) (loc: location) -> (action env loc))
255 let delete atoms = Grammar.delete_rule l2_pattern atoms
259 let boxify = function
261 | l -> Layout (Box (H, l))
263 let fold_binder binder pt_names body =
264 let fold_cluster binder terms ty body =
266 (fun term body -> Binder (binder, (term, ty), body))
267 terms body (* terms are names: either Ident or FreshVar *)
270 (fun (names, ty) body -> fold_cluster binder names ty body)
273 let return_term loc term = AttributedTerm (`Loc loc, term)
275 let _ = (* create empty precedence level for "l2_pattern" *)
276 let mk_level_list first last =
277 let rec aux acc = function
278 | i when i < first -> acc
279 | i -> aux ((Some (string_of_int i), None, []) :: acc) (i - 1)
284 [ Grammar.Entry.obj (l2_pattern: 'a Grammar.Entry.e),
286 mk_level_list min_precedence max_precedence ]
289 GLOBAL: level1_pattern level2_pattern level3_term
291 notation interpretation
293 (* {{{ Grammar for concrete syntax patterns, notation level 1 *)
294 level1_pattern: [ [ p = l1_simple_pattern -> p ] ];
295 l1_pattern: [ [ p = LIST1 l1_simple_pattern -> p ] ];
297 [ s = SYMBOL -> `Symbol s
298 | k = KEYWORD -> `Keyword k
299 | n = NUMBER -> `Number n
302 sep: [ [ SYMBOL "\\SEP"; sep = literal -> sep ] ];
303 (* row_sep: [ [ SYMBOL "\\ROWSEP"; sep = literal -> sep ] ];
304 field_sep: [ [ SYMBOL "\\FIELDSEP"; sep = literal -> sep ] ]; *)
306 [ SYMBOL "\\LIST0"; p = l1_simple_pattern; sep = OPT sep -> List0 (p, sep)
307 | SYMBOL "\\LIST1"; p = l1_simple_pattern; sep = OPT sep -> List1 (p, sep)
308 | SYMBOL "\\OPT"; p = l1_simple_pattern -> Opt p
311 l1_pattern_variable: [
312 [ SYMBOL "\\TERM"; id = IDENT -> TermVar id
313 | SYMBOL "\\NUM"; id = IDENT -> NumVar id
314 | SYMBOL "\\IDENT"; id = IDENT -> IdentVar id
319 [ p1 = SELF; SYMBOL "\\SUB"; p2 = SELF ->
320 return_term loc (Layout (Sub (p1, p2)))
321 | p1 = SELF; SYMBOL "\\SUP"; p2 = SELF ->
322 return_term loc (Layout (Sup (p1, p2)))
323 | p1 = SELF; SYMBOL "\\BELOW"; p2 = SELF ->
324 return_term loc (Layout (Below (p1, p2)))
325 | p1 = SELF; SYMBOL "\\ABOVE"; p2 = SELF ->
326 return_term loc (Layout (Above (p1, p2)))
327 | p1 = SELF; SYMBOL "\\OVER"; p2 = SELF ->
328 return_term loc (Layout (Over (p1, p2)))
329 | p1 = SELF; SYMBOL "\\ATOP"; p2 = SELF ->
330 return_term loc (Layout (Atop (p1, p2)))
331 (* | SYMBOL "\\ARRAY"; p = SELF; csep = OPT field_sep; rsep = OPT row_sep ->
332 return_term loc (Array (p, csep, rsep)) *)
333 | SYMBOL "\\FRAC"; p1 = SELF; p2 = SELF ->
334 return_term loc (Layout (Frac (p1, p2)))
335 | SYMBOL "\\SQRT"; p = SELF -> return_term loc (Layout (Sqrt p))
336 | SYMBOL "\\ROOT"; index = SELF; SYMBOL "\\OF"; arg = SELF ->
337 return_term loc (Layout (Root (arg, index)));
338 | SYMBOL "\\HBOX"; DELIM "\\["; p = l1_pattern; DELIM "\\]" ->
339 return_term loc (Layout (Box (H, p)))
340 | SYMBOL "\\VBOX"; DELIM "\\["; p = l1_pattern; DELIM "\\]" ->
341 return_term loc (Layout (Box (V, p)))
342 | SYMBOL "\\BREAK" -> return_term loc (Layout Break)
343 | DELIM "\\["; p = l1_pattern; DELIM "\\]" ->
344 return_term loc (boxify p)
345 | p = SELF; SYMBOL "\\AS"; id = IDENT ->
346 return_term loc (Variable (Ascription (p, id)))
349 [ i = IDENT -> return_term loc (Ident (i, None))
350 | m = l1_magic_pattern -> return_term loc (Magic m)
351 | v = l1_pattern_variable -> return_term loc (Variable v)
352 | l = literal -> return_term loc (Literal l)
356 (* {{{ Grammar for ast patterns, notation level 2 *)
357 level2_pattern: [ [ p = l2_pattern -> p ] ];
359 [ SYMBOL "\\PROP" -> `Prop
360 | SYMBOL "\\SET" -> `Set
361 | SYMBOL "\\TYPE" -> `Type
365 [ SYMBOL "\\subst"; (* to avoid catching frequent "a [1]" cases *)
368 i = IDENT; SYMBOL <:unicode<Assign>> (* ≔ *); t = l2_pattern -> (i, t)
375 [ s = SYMBOL "_" -> None
376 | p = l2_pattern -> Some p ]
379 [ SYMBOL "["; substs = LIST0 meta_subst; SYMBOL "]" -> substs ]
381 possibly_typed_name: [
382 [ SYMBOL "("; id = bound_name; SYMBOL ":"; typ = l2_pattern; SYMBOL ")" ->
384 | id = bound_name -> id, None
388 [ id = IDENT -> id, []
389 | SYMBOL "("; id = IDENT; vars = LIST1 possibly_typed_name; SYMBOL ")" ->
394 [ SYMBOL <:unicode<Pi>> (* Π *) -> `Pi
395 | SYMBOL <:unicode<exists>> (* ∃ *) -> `Exists
396 | SYMBOL <:unicode<forall>> (* ∀ *) -> `Forall
397 | SYMBOL <:unicode<lambda>> (* λ *) -> `Lambda
401 [ i = IDENT -> Ident (i, None)
402 | SYMBOL "\\FRESH"; i = IDENT -> Variable (FreshVar i)
406 [ vars = LIST1 bound_name SEP SYMBOL ",";
407 ty = OPT [ SYMBOL ":"; p = l2_pattern -> p ] ->
411 vars = LIST1 bound_name SEP SYMBOL ",";
412 ty = OPT [ SYMBOL ":"; p = l2_pattern -> p ];
420 [ IDENT "rec" -> `Inductive
421 | IDENT "corec" -> `CoInductive
426 name = bound_name; args = bound_names;
427 index_name = OPT [ IDENT "on"; id = bound_name -> id ];
428 ty = OPT [ SYMBOL ":" ; p = l2_pattern -> p ];
429 SYMBOL <:unicode<def>> (* ≝ *); body = l2_pattern ->
430 let body = fold_binder `Lambda args body in
434 | Some ty -> Some (fold_binder `Pi args ty)
436 let rec position_of name p = function
438 | n :: _ when n = name -> Some p, p
439 | _ :: tl -> position_of name (p + 1) tl
441 let rec find_arg name n = function
443 fail loc (sprintf "Argument %s not found"
444 (CicNotationPp.pp_term name))
446 (match position_of name 0 l with
447 | None, len -> find_arg name (n + len) tl
448 | Some where, len -> n + where)
451 match index_name with
453 | Some name -> find_arg name 0 args
455 (name, ty), body, index
460 l2_pattern_variable: [
461 [ SYMBOL "\\TERM"; id = IDENT -> TermVar id
462 | SYMBOL "\\NUM"; id = IDENT -> NumVar id
463 | SYMBOL "\\IDENT"; id = IDENT -> IdentVar id
464 | SYMBOL "\\FRESH"; id = IDENT -> FreshVar id
469 kind = [ IDENT "left" -> `Left | IDENT "right" -> `Right ];
470 DELIM "\\["; base = l2_pattern; DELIM "\\]";
471 SYMBOL "\\LAMBDA"; id = IDENT;
472 DELIM "\\["; recursive = l2_pattern; DELIM "\\]" ->
473 Fold (kind, base, [id], recursive)
474 | SYMBOL "\\DEFAULT";
475 DELIM "\\["; some = l2_pattern; DELIM "\\]";
476 DELIM "\\["; none = l2_pattern; DELIM "\\]" ->
480 l2_pattern: LEVEL "10" (* let in *)
482 [ IDENT "let"; var = possibly_typed_name; SYMBOL <:unicode<def>> (* ≝ *);
483 p1 = l2_pattern; "in"; p2 = l2_pattern ->
484 return_term loc (LetIn (var, p1, p2))
485 | IDENT "let"; k = induction_kind; defs = let_defs; IDENT "in";
487 return_term loc (LetRec (k, defs, body))
490 l2_pattern: LEVEL "20" (* binder *)
492 [ b = binder; names = bound_names; SYMBOL "."; body = l2_pattern ->
493 return_term loc (fold_binder b names body)
496 l2_pattern: LEVEL "70" (* apply *)
498 [ p1 = l2_pattern; p2 = l2_pattern ->
499 let rec aux = function
501 | AttributedTerm (_, Appl (hd :: tl)) ->
505 return_term loc (Appl (aux p1 @ [p2]))
508 l2_pattern: LEVEL "90" (* simple *)
510 [ id = IDENT -> return_term loc (Ident (id, None))
511 | id = IDENT; s = explicit_subst -> return_term loc (Ident (id, Some s))
512 | u = URI -> return_term loc (Uri (u, None))
513 | n = NUMBER -> prerr_endline "number"; return_term loc (Num (n, 0))
514 | IMPLICIT -> return_term loc (Implicit)
515 | m = META -> return_term loc (Meta (int_of_string m, []))
516 | m = META; s = meta_substs -> return_term loc (Meta (int_of_string m, s))
517 | s = sort -> return_term loc (Sort s)
518 | outtyp = OPT [ SYMBOL "["; ty = l2_pattern; SYMBOL "]" -> ty ];
519 IDENT "match"; t = l2_pattern;
520 indty_ident = OPT [ SYMBOL ":"; id = IDENT -> id ];
521 IDENT "with"; SYMBOL "[";
523 lhs = match_pattern; SYMBOL <:unicode<Rightarrow>> (* ⇒ *);
528 return_term loc (Case (t, indty_ident, outtyp, patterns))
529 | SYMBOL "("; p1 = l2_pattern; SYMBOL ":"; p2 = l2_pattern; SYMBOL ")" ->
530 return_term loc (Appl [ Symbol ("cast", 0); p1; p2 ])
531 | SYMBOL "("; p = l2_pattern; SYMBOL ")" -> p
532 | v = l2_pattern_variable -> return_term loc (Variable v)
533 | m = l2_magic_pattern -> return_term loc (Magic m)
537 (* {{{ Grammar for interpretation, notation level 3 *)
539 [ id = IDENT -> IdentArg id
540 | SYMBOL <:unicode<eta>> (* η *); SYMBOL "."; a = SELF -> EtaArg (None, a)
541 | SYMBOL <:unicode<eta>> (* η *); id = IDENT; SYMBOL "."; a = SELF ->
546 [ u = URI -> UriPattern u
547 | a = argument -> ArgPattern a
548 | SYMBOL "("; terms = LIST1 SELF; SYMBOL ")" ->
552 | terms -> ApplPattern terms)
556 (* {{{ Notation glues *)
558 [ IDENT "left"; IDENT "associative" -> Gramext.LeftA
559 | IDENT "right"; IDENT "associative" -> Gramext.RightA
560 | IDENT "non"; IDENT "associative" -> Gramext.NonA
564 [ IDENT "with"; IDENT "precedence"; n = NUMBER -> int_of_string n ]
569 assoc = OPT associativity; prec = OPT precedence;
570 IDENT "for"; p2 = level2_pattern ->
571 (p1, assoc, prec, p2)
575 [ IDENT "interpretation"; s = SYMBOL; args = LIST1 argument; IDENT "as";
581 (* {{{ Top-level phrases *)
583 [ IDENT "print"; p2 = level2_pattern; SYMBOL "." -> Print p2
584 | (l1, assoc, prec, l2) = notation; SYMBOL "." ->
585 Notation (l1, assoc, prec, l2)
591 (** {2 API implementation} *)
593 let exc_located_wrapper f =
597 | Stdpp.Exc_located (floc, Stream.Error msg) ->
598 raise (Parse_error (floc, msg))
599 | Stdpp.Exc_located (floc, exn) ->
600 raise (Parse_error (floc, (Printexc.to_string exn)))
602 let parse_syntax_pattern stream =
603 exc_located_wrapper (fun () -> Grammar.Entry.parse level1_pattern stream)
604 let parse_ast_pattern stream =
605 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_pattern stream)
606 let parse_interpretation stream =
607 exc_located_wrapper (fun () -> Grammar.Entry.parse level3_term stream)
608 let parse_phrase stream =
609 exc_located_wrapper (fun () -> Grammar.Entry.parse phrase stream)
613 let print_l2_pattern () =
614 Grammar.print_entry Format.std_formatter (Grammar.Entry.obj l2_pattern);
615 Format.pp_print_flush Format.std_formatter ();
618 (* vim:set encoding=utf8 foldmethod=marker: *)