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 level1_pattern_grammar =
35 Grammar.gcreate CicNotationLexer.level1_pattern_lexer
36 let level2_ast_grammar = Grammar.gcreate CicNotationLexer.level2_ast_lexer
37 let level2_meta_grammar = Grammar.gcreate CicNotationLexer.level2_meta_lexer
39 let min_precedence = 0
40 let max_precedence = 100
43 Grammar.Entry.create level1_pattern_grammar "level1_pattern"
44 let level2_ast = Grammar.Entry.create level2_ast_grammar "level2_ast"
45 let term = Grammar.Entry.create level2_ast_grammar "term"
46 let let_defs = Grammar.Entry.create level2_ast_grammar "let_defs"
47 let level2_meta = Grammar.Entry.create level2_meta_grammar "level2_meta"
49 let return_term loc term = ()
53 Pervasives.int_of_string s
55 failwith (sprintf "Lexer failure: string_of_int \"%s\" failed" s)
57 (** {2 Grammar extension} *)
59 let gram_symbol s = Gramext.Stoken ("SYMBOL", s)
60 let gram_ident s = Gramext.Stoken ("IDENT", s)
61 let gram_number s = Gramext.Stoken ("NUMBER", s)
62 let gram_keyword s = Gramext.Stoken ("", s)
63 let gram_term = Gramext.Sself
67 | `Symbol s -> gram_symbol s
68 | `Keyword s -> gram_keyword s
69 | `Number s -> gram_number s
73 | Binding of string * value_type
74 | Env of (string * value_type) list
76 let make_action action bindings =
77 let rec aux (vl : CicNotationEnv.t) =
79 [] -> Gramext.action (fun (loc: location) -> action vl loc)
80 | NoBinding :: tl -> Gramext.action (fun _ -> aux vl tl)
81 (* LUCA: DEFCON 5 BEGIN *)
82 | Binding (name, TermType) :: tl ->
84 (fun (v:term) -> aux ((name, (TermType, TermValue v))::vl) tl)
85 | Binding (name, StringType) :: tl ->
88 aux ((name, (StringType, StringValue v)) :: vl) tl)
89 | Binding (name, NumType) :: tl ->
91 (fun (v:string) -> aux ((name, (NumType, NumValue v)) :: vl) tl)
92 | Binding (name, OptType t) :: tl ->
95 aux ((name, (OptType t, OptValue v)) :: vl) tl)
96 | Binding (name, ListType t) :: tl ->
99 aux ((name, (ListType t, ListValue v)) :: vl) tl)
101 Gramext.action (fun (v:CicNotationEnv.t) -> aux (v @ vl) tl)
102 (* LUCA: DEFCON 5 END *)
104 aux [] (List.rev bindings)
110 | NoBinding :: tl -> aux acc tl
111 | Env names :: tl -> aux (List.rev names @ acc) tl
112 | Binding (name, ty) :: tl -> aux ((name, ty) :: acc) tl
116 (* given a level 1 pattern computes the new RHS of "term" grammar entry *)
117 let extract_term_production pattern =
118 let rec aux = function
119 | AttributedTerm (_, t) -> aux t
120 | Literal l -> aux_literal l
121 | Layout l -> aux_layout l
122 | Magic m -> aux_magic m
123 | Variable v -> aux_variable v
125 prerr_endline (CicNotationPp.pp_term t);
129 | `Symbol s -> [NoBinding, gram_symbol s]
131 (* assumption: s will be registered as a keyword with the lexer *)
132 [NoBinding, gram_keyword s]
133 | `Number s -> [NoBinding, gram_number s]
134 and aux_layout = function
135 | Sub (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\sub"] @ aux p2
136 | Sup (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\sup"] @ aux p2
137 | Below (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\below"] @ aux p2
138 | Above (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\above"] @ aux p2
139 | Frac (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\frac"] @ aux p2
140 | Atop (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\atop"] @ aux p2
141 | Over (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\over"] @ aux p2
143 [NoBinding, gram_symbol "\\root"] @ aux p2
144 @ [NoBinding, gram_symbol "\\of"] @ aux p1
145 | Sqrt p -> [NoBinding, gram_symbol "\\sqrt"] @ aux p
147 | Box (_, pl) -> List.flatten (List.map aux pl)
148 | Group pl -> List.flatten (List.map aux pl)
149 and aux_magic magic =
152 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
153 let action (env_opt : CicNotationEnv.t option) (loc : location) =
155 | Some env -> List.map opt_binding_some env
156 | None -> List.map opt_binding_of_name p_names
158 [ Env (List.map opt_declaration p_names),
160 [ [ Gramext.Sopt (Gramext.srules [ p_atoms, p_action ]) ],
161 Gramext.action action ] ]
164 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
165 (* let env0 = List.map list_binding_of_name p_names in
166 let grow_env_entry env n v =
169 | (n', (ty, ListValue vl)) as entry ->
170 if n' = n then n', (ty, ListValue (v :: vl)) else entry
174 let grow_env env_i env =
176 (fun env (n, (_, v)) -> grow_env_entry env n v)
179 let action (env_list : CicNotationEnv.t list) (loc : location) =
180 CicNotationEnv.coalesce_env p_names env_list
184 | List0 (_, None) -> Gramext.Slist0 s
185 | List1 (_, None) -> Gramext.Slist1 s
186 | List0 (_, Some l) -> Gramext.Slist0sep (s, gram_of_literal l)
187 | List1 (_, Some l) -> Gramext.Slist1sep (s, gram_of_literal l)
190 [ Env (List.map list_declaration p_names),
192 [ [ gram_of_list (Gramext.srules [ p_atoms, p_action ]) ],
193 Gramext.action action ] ]
197 | NumVar s -> [Binding (s, NumType), gram_number ""]
198 | TermVar s -> [Binding (s, TermType), gram_term]
199 | IdentVar s -> [Binding (s, StringType), gram_ident ""]
200 | Ascription (p, s) -> assert false (* TODO *)
201 | FreshVar _ -> assert false
202 and inner_pattern p =
203 let p_bindings, p_atoms = List.split (aux p) in
204 let p_names = flatten_opt p_bindings in
206 make_action (fun (env : CicNotationEnv.t) (loc : location) -> env)
209 p_bindings, p_atoms, p_names, action
213 let level_of precedence associativity =
214 if precedence < min_precedence || precedence > max_precedence then
215 raise (Level_not_found precedence);
217 match associativity with
218 | Gramext.NonA -> "N"
219 | Gramext.LeftA -> "L"
220 | Gramext.RightA -> "R"
222 string_of_int precedence ^ assoc_string
224 type rule_id = Token.t Gramext.g_symbol list
226 (* mapping: rule_id -> owned keywords. (rule_id, string list) Hashtbl.t *)
227 let owned_keywords = Hashtbl.create 23
229 let extend level1_pattern ~precedence ~associativity action =
230 let p_bindings, p_atoms =
231 List.split (extract_term_production level1_pattern)
233 let level = level_of precedence associativity in
234 let p_names = flatten_opt p_bindings in
237 [ Grammar.Entry.obj (term: 'a Grammar.Entry.e),
238 Some (Gramext.Level level),
243 (fun (env: CicNotationEnv.t) (loc: location) -> (action env loc))
246 let keywords = CicNotationUtil.keywords_of_term level1_pattern in
247 let rule_id = p_atoms in
248 List.iter CicNotationLexer.add_level2_ast_keyword keywords;
249 Hashtbl.add owned_keywords rule_id keywords; (* keywords may be [] *)
253 let atoms = rule_id in
255 let keywords = Hashtbl.find owned_keywords rule_id in
256 List.iter CicNotationLexer.remove_level2_ast_keyword keywords
257 with Not_found -> assert false);
258 Grammar.delete_rule term atoms
262 let parse_level1_pattern_ref = ref (fun _ -> assert false)
263 let parse_level2_ast_ref = ref (fun _ -> assert false)
264 let parse_level2_meta_ref = ref (fun _ -> assert false)
266 let fold_cluster binder terms ty body =
268 (fun term body -> Binder (binder, (term, ty), body))
269 terms body (* terms are names: either Ident or FreshVar *)
271 let fold_binder binder pt_names body =
273 (fun (names, ty) body -> fold_cluster binder names ty body)
276 let return_term loc term = AttributedTerm (`Loc loc, term)
278 let _ = (* create empty precedence level for "term" *)
279 let mk_level_list first last =
280 let rec aux acc = function
281 | i when i < first -> acc
284 ((Some (string_of_int i ^ "N"), Some Gramext.NonA, [])
285 :: (Some (string_of_int i ^ "L"), Some Gramext.LeftA, [])
286 :: (Some (string_of_int i ^ "R"), Some Gramext.RightA, [])
293 [ Grammar.Entry.obj (term: 'a Grammar.Entry.e),
295 mk_level_list min_precedence max_precedence ]
297 (* {{{ Grammar for concrete syntax patterns, notation level 1 *)
299 GLOBAL: level1_pattern;
301 level1_pattern: [ [ p = l1_pattern; EOI -> CicNotationUtil.boxify p ] ];
302 l1_pattern: [ [ p = LIST1 l1_simple_pattern -> p ] ];
304 [ s = SYMBOL -> `Symbol s
305 | k = QKEYWORD -> `Keyword k
306 | n = NUMBER -> `Number n
309 sep: [ [ "sep"; sep = literal -> sep ] ];
310 (* row_sep: [ [ "rowsep"; sep = literal -> sep ] ];
311 field_sep: [ [ "fieldsep"; sep = literal -> sep ] ]; *)
313 [ "list0"; p = l1_simple_pattern; sep = OPT sep -> List0 (p, sep)
314 | "list1"; p = l1_simple_pattern; sep = OPT sep -> List1 (p, sep)
315 | "opt"; p = l1_simple_pattern -> Opt p
318 l1_pattern_variable: [
319 [ "term"; id = IDENT -> TermVar id
320 | "number"; id = IDENT -> NumVar id
321 | "ident"; id = IDENT -> IdentVar id
326 [ p1 = SELF; SYMBOL "\\sub"; p2 = SELF ->
327 return_term loc (Layout (Sub (p1, p2)))
328 | p1 = SELF; SYMBOL "\\sup"; p2 = SELF ->
329 return_term loc (Layout (Sup (p1, p2)))
330 | p1 = SELF; SYMBOL "\\below"; p2 = SELF ->
331 return_term loc (Layout (Below (p1, p2)))
332 | p1 = SELF; SYMBOL "\\above"; p2 = SELF ->
333 return_term loc (Layout (Above (p1, p2)))
334 | p1 = SELF; SYMBOL "\\over"; p2 = SELF ->
335 return_term loc (Layout (Over (p1, p2)))
336 | p1 = SELF; SYMBOL "\\atop"; p2 = SELF ->
337 return_term loc (Layout (Atop (p1, p2)))
338 (* | "array"; p = SELF; csep = OPT field_sep; rsep = OPT row_sep ->
339 return_term loc (Array (p, csep, rsep)) *)
340 | SYMBOL "\\frac"; p1 = SELF; p2 = SELF ->
341 return_term loc (Layout (Frac (p1, p2)))
342 | SYMBOL "\\sqrt"; p = SELF -> return_term loc (Layout (Sqrt p))
343 | SYMBOL "\\root"; index = SELF; SYMBOL "\\of"; arg = SELF ->
344 return_term loc (Layout (Root (arg, index)))
345 | "hbox"; LPAREN; p = l1_pattern; RPAREN ->
346 return_term loc (Layout (Box ((H, false, false), p)))
347 | "vbox"; LPAREN; p = l1_pattern; RPAREN ->
348 return_term loc (Layout (Box ((V, false, false), p)))
349 | "hvbox"; LPAREN; p = l1_pattern; RPAREN ->
350 return_term loc (Layout (Box ((HV, false, false), p)))
351 | "hovbox"; LPAREN; p = l1_pattern; RPAREN ->
352 return_term loc (Layout (Box ((HOV, false, false), p)))
353 | "break" -> return_term loc (Layout Break)
354 (* | SYMBOL "\\SPACE" -> return_term loc (Layout Space) *)
355 | LPAREN; p = l1_pattern; RPAREN ->
356 return_term loc (CicNotationUtil.group p)
359 [ i = IDENT -> return_term loc (Variable (TermVar i))
360 | m = l1_magic_pattern -> return_term loc (Magic m)
361 | v = l1_pattern_variable -> return_term loc (Variable v)
362 | l = literal -> return_term loc (Literal l)
368 (* {{{ Grammar for ast magics, notation level 2 *)
372 [ "term"; id = IDENT -> TermVar id
373 | "number"; id = IDENT -> NumVar id
374 | "ident"; id = IDENT -> IdentVar id
375 | "fresh"; id = IDENT -> FreshVar id
376 | "anonymous" -> TermVar "_"
377 | id = IDENT -> TermVar id
381 [ "fold"; kind = [ "left" -> `Left | "right" -> `Right ];
382 base = level2_meta; "rec"; id = IDENT; recursive = level2_meta ->
383 Fold (kind, base, [id], recursive)
384 | "default"; some = level2_meta; none = level2_meta -> Default (some, none)
385 | "if"; p_test = level2_meta;
386 "then"; p_true = level2_meta;
387 "else"; p_false = level2_meta ->
388 If (p_test, p_true, p_false)
393 [ magic = l2_magic -> Magic magic
394 | var = l2_variable -> Variable var
395 | blob = UNPARSED_AST -> !parse_level2_ast_ref (Stream.of_string blob)
401 (* {{{ Grammar for ast patterns, notation level 2 *)
403 GLOBAL: level2_ast term let_defs;
404 level2_ast: [ [ p = term -> p ] ];
413 [ SYMBOL "\\subst"; (* to avoid catching frequent "a [1]" cases *)
416 i = IDENT; SYMBOL <:unicode<Assign>> (* ≔ *); t = term -> (i, t)
423 [ s = SYMBOL "_" -> None
424 | p = term -> Some p ]
427 [ SYMBOL "["; substs = LIST0 meta_subst; SYMBOL "]" -> substs ]
429 possibly_typed_name: [
430 [ LPAREN; id = single_arg; SYMBOL ":"; typ = term; RPAREN ->
432 | arg = single_arg -> arg, None
436 [ id = IDENT -> id, []
437 | LPAREN; id = IDENT; vars = LIST1 possibly_typed_name; RPAREN ->
442 [ SYMBOL <:unicode<Pi>> (* Π *) -> `Pi
443 | SYMBOL <:unicode<exists>> (* ∃ *) -> `Exists
444 | SYMBOL <:unicode<forall>> (* ∀ *) -> `Forall
445 | SYMBOL <:unicode<lambda>> (* λ *) -> `Lambda
449 [ LPAREN; names = LIST1 IDENT SEP SYMBOL ",";
450 SYMBOL ":"; ty = term; RPAREN ->
451 List.map (fun n -> Ident (n, None)) names, Some ty
452 | name = IDENT -> [Ident (name, None)], None
453 | blob = UNPARSED_META ->
454 let meta = !parse_level2_meta_ref (Stream.of_string blob) in
456 | Variable (FreshVar _) -> [meta], None
457 | Variable (TermVar "_") -> [Ident ("_", None)], None
458 | _ -> failwith "Invalid bound name."
462 [ name = IDENT -> Ident (name, None)
463 | blob = UNPARSED_META ->
464 let meta = !parse_level2_meta_ref (Stream.of_string blob) in
466 | Variable (FreshVar _) -> meta
467 | Variable (TermVar "_") -> Ident ("_", None)
468 | _ -> failwith "Invalid index name."
472 [ "rec" -> `Inductive
473 | "corec" -> `CoInductive
480 index_name = OPT [ "on"; id = single_arg -> id ];
481 ty = OPT [ SYMBOL ":" ; p = term -> p ];
482 SYMBOL <:unicode<def>> (* ≝ *); body = term ->
483 let body = fold_binder `Lambda args body in
487 | Some ty -> Some (fold_binder `Pi args ty)
489 let rec position_of name p = function
491 | n :: _ when n = name -> Some p, p
492 | _ :: tl -> position_of name (p + 1) tl
494 let rec find_arg name n = function
496 fail loc (sprintf "Argument %s not found"
497 (CicNotationPp.pp_term name))
499 (match position_of name 0 l with
500 | None, len -> find_arg name (n + len) tl
501 | Some where, len -> n + where)
504 match index_name with
506 | Some index_name -> find_arg index_name 0 args
508 (name, ty), body, index
515 l = LIST1 single_arg SEP SYMBOL "," -> l
516 | SYMBOL "_" -> [Ident ("_", None)] ];
517 typ = OPT [ SYMBOL ":"; t = term -> t ] -> (vars, typ)
520 l = LIST1 single_arg SEP SYMBOL "," -> l
521 | SYMBOL "_" -> [Ident ("_", None)] ];
522 typ = OPT [ SYMBOL ":"; t = term -> t ];
523 RPAREN -> (vars, typ)
526 term: LEVEL "10N" [ (* let in *)
527 [ "let"; var = possibly_typed_name; SYMBOL <:unicode<def>> (* ≝ *);
528 p1 = term; "in"; p2 = term ->
529 return_term loc (LetIn (var, p1, p2))
530 | "let"; k = induction_kind; defs = let_defs; "in";
532 return_term loc (LetRec (k, defs, body))
535 term: LEVEL "20R" (* binder *)
537 [ b = binder; (vars, typ) = binder_vars; SYMBOL "."; body = term ->
538 return_term loc (fold_cluster b vars typ body)
541 term: LEVEL "70L" (* apply *)
543 [ p1 = term; p2 = term ->
544 let rec aux = function
546 | AttributedTerm (_, Appl (hd :: tl)) ->
550 return_term loc (Appl (aux p1 @ [p2]))
553 term: LEVEL "90N" (* simple *)
555 [ id = IDENT -> return_term loc (Ident (id, None))
556 | id = IDENT; s = explicit_subst -> return_term loc (Ident (id, Some s))
557 | s = CSYMBOL -> return_term loc (Symbol (s, 0))
558 | u = URI -> return_term loc (Uri (u, None))
559 | n = NUMBER -> return_term loc (Num (n, 0))
560 | IMPLICIT -> return_term loc (Implicit)
561 | PLACEHOLDER -> return_term loc UserInput
562 | m = META -> return_term loc (Meta (int_of_string m, []))
563 | m = META; s = meta_substs -> return_term loc (Meta (int_of_string m, s))
564 | s = sort -> return_term loc (Sort s)
565 | outtyp = OPT [ SYMBOL "["; ty = term; SYMBOL "]" -> ty ];
567 indty_ident = OPT [ "in"; id = IDENT -> id ];
570 lhs = match_pattern; SYMBOL <:unicode<Rightarrow>> (* ⇒ *);
575 return_term loc (Case (t, indty_ident, outtyp, patterns))
576 | LPAREN; p1 = term; SYMBOL ":"; p2 = term; RPAREN ->
577 return_term loc (Cast (p1, p2))
578 | LPAREN; p = term; RPAREN -> p
579 | blob = UNPARSED_META -> !parse_level2_meta_ref (Stream.of_string blob)
585 (** {2 API implementation} *)
587 let exc_located_wrapper f =
591 | Stdpp.Exc_located (floc, Stream.Error msg) ->
592 raise (Parse_error (floc, msg))
593 | Stdpp.Exc_located (floc, exn) ->
594 raise (Parse_error (floc, (Printexc.to_string exn)))
596 let parse_level1_pattern stream =
597 exc_located_wrapper (fun () -> Grammar.Entry.parse level1_pattern stream)
598 let parse_level2_ast stream =
599 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_ast stream)
600 let parse_level2_meta stream =
601 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_meta stream)
604 parse_level1_pattern_ref := parse_level1_pattern;
605 parse_level2_ast_ref := parse_level2_ast;
606 parse_level2_meta_ref := parse_level2_meta
610 let print_l2_pattern () =
611 Grammar.print_entry Format.std_formatter (Grammar.Entry.obj term);
612 Format.pp_print_flush Format.std_formatter ();
615 (* vim:set encoding=utf8 foldmethod=marker: *)