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
41 let default_precedence = 50
44 Grammar.Entry.create level1_pattern_grammar "level1_pattern"
45 let level2_ast = Grammar.Entry.create level2_ast_grammar "level2_ast"
46 let term = Grammar.Entry.create level2_ast_grammar "term"
47 let let_defs = Grammar.Entry.create level2_ast_grammar "let_defs"
48 let level2_meta = Grammar.Entry.create level2_meta_grammar "level2_meta"
50 let return_term loc term = ()
54 Pervasives.int_of_string s
56 failwith (sprintf "Lexer failure: string_of_int \"%s\" failed" s)
58 (** {2 Grammar extension} *)
60 let gram_symbol s = Gramext.Stoken ("SYMBOL", s)
61 let gram_ident s = Gramext.Stoken ("IDENT", s)
62 let gram_number s = Gramext.Stoken ("NUMBER", s)
63 let gram_keyword s = Gramext.Stoken ("", s)
64 let gram_term = Gramext.Sself
68 | `Symbol s -> gram_symbol s
69 | `Keyword s -> gram_keyword s
70 | `Number s -> gram_number s
74 | Binding of string * value_type
75 | Env of (string * value_type) list
77 let make_action action bindings =
78 let rec aux (vl : CicNotationEnv.t) =
80 [] -> Gramext.action (fun (loc: location) -> action vl loc)
81 | NoBinding :: tl -> Gramext.action (fun _ -> aux vl tl)
82 (* LUCA: DEFCON 5 BEGIN *)
83 | Binding (name, TermType) :: tl ->
85 (fun (v:term) -> aux ((name, (TermType, TermValue v))::vl) tl)
86 | Binding (name, StringType) :: tl ->
89 aux ((name, (StringType, StringValue v)) :: vl) tl)
90 | Binding (name, NumType) :: tl ->
92 (fun (v:string) -> aux ((name, (NumType, NumValue v)) :: vl) tl)
93 | Binding (name, OptType t) :: tl ->
96 aux ((name, (OptType t, OptValue v)) :: vl) tl)
97 | Binding (name, ListType t) :: tl ->
100 aux ((name, (ListType t, ListValue v)) :: vl) tl)
102 Gramext.action (fun (v:CicNotationEnv.t) -> aux (v @ vl) tl)
103 (* LUCA: DEFCON 5 END *)
105 aux [] (List.rev bindings)
111 | NoBinding :: tl -> aux acc tl
112 | Env names :: tl -> aux (List.rev names @ acc) tl
113 | Binding (name, ty) :: tl -> aux ((name, ty) :: acc) tl
117 (* given a level 1 pattern computes the new RHS of "term" grammar entry *)
118 let extract_term_production pattern =
119 let rec aux = function
120 | AttributedTerm (_, t) -> aux t
121 | Literal l -> aux_literal l
122 | Layout l -> aux_layout l
123 | Magic m -> aux_magic m
124 | Variable v -> aux_variable v
126 prerr_endline (CicNotationPp.pp_term t);
130 | `Symbol s -> [NoBinding, gram_symbol s]
132 (* assumption: s will be registered as a keyword with the lexer *)
133 [NoBinding, gram_keyword s]
134 | `Number s -> [NoBinding, gram_number s]
135 and aux_layout = function
136 | Sub (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\sub"] @ aux p2
137 | Sup (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\sup"] @ aux p2
138 | Below (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\below"] @ aux p2
139 | Above (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\above"] @ aux p2
140 | Frac (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\frac"] @ aux p2
141 | Atop (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\atop"] @ aux p2
142 | Over (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\over"] @ aux p2
144 [NoBinding, gram_symbol "\\root"] @ aux p2
145 @ [NoBinding, gram_symbol "\\of"] @ aux p1
146 | Sqrt p -> [NoBinding, gram_symbol "\\sqrt"] @ aux p
148 | Box (_, pl) -> List.flatten (List.map aux pl)
149 | Group pl -> List.flatten (List.map aux pl)
150 and aux_magic magic =
153 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
154 let action (env_opt : CicNotationEnv.t option) (loc : location) =
156 | Some env -> List.map opt_binding_some env
157 | None -> List.map opt_binding_of_name p_names
159 [ Env (List.map opt_declaration p_names),
161 [ [ Gramext.Sopt (Gramext.srules [ p_atoms, p_action ]) ],
162 Gramext.action action ] ]
165 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
166 (* let env0 = List.map list_binding_of_name p_names in
167 let grow_env_entry env n v =
170 | (n', (ty, ListValue vl)) as entry ->
171 if n' = n then n', (ty, ListValue (v :: vl)) else entry
175 let grow_env env_i env =
177 (fun env (n, (_, v)) -> grow_env_entry env n v)
180 let action (env_list : CicNotationEnv.t list) (loc : location) =
181 CicNotationEnv.coalesce_env p_names env_list
185 | List0 (_, None) -> Gramext.Slist0 s
186 | List1 (_, None) -> Gramext.Slist1 s
187 | List0 (_, Some l) -> Gramext.Slist0sep (s, gram_of_literal l)
188 | List1 (_, Some l) -> Gramext.Slist1sep (s, gram_of_literal l)
191 [ Env (List.map list_declaration p_names),
193 [ [ gram_of_list (Gramext.srules [ p_atoms, p_action ]) ],
194 Gramext.action action ] ]
198 | NumVar s -> [Binding (s, NumType), gram_number ""]
199 | TermVar s -> [Binding (s, TermType), gram_term]
200 | IdentVar s -> [Binding (s, StringType), gram_ident ""]
201 | Ascription (p, s) -> assert false (* TODO *)
202 | FreshVar _ -> assert false
203 and inner_pattern p =
204 let p_bindings, p_atoms = List.split (aux p) in
205 let p_names = flatten_opt p_bindings in
207 make_action (fun (env : CicNotationEnv.t) (loc : location) -> env)
210 p_bindings, p_atoms, p_names, action
214 let level_of_int precedence =
215 if precedence < min_precedence || precedence > max_precedence then
216 raise (Level_not_found precedence);
217 string_of_int precedence
219 type rule_id = Token.t Gramext.g_symbol list
221 (* mapping: rule_id -> owned keywords. (rule_id, string list) Hashtbl.t *)
222 let owned_keywords = Hashtbl.create 23
224 let extend level1_pattern ?(precedence = default_precedence)
225 ?associativity action
227 let p_bindings, p_atoms =
228 List.split (extract_term_production level1_pattern)
230 let level = level_of_int precedence in
231 let p_names = flatten_opt p_bindings in
234 [ Grammar.Entry.obj (term: 'a Grammar.Entry.e),
235 Some (Gramext.Level level),
240 (fun (env: CicNotationEnv.t) (loc: location) -> (action env loc))
243 let keywords = CicNotationUtil.keywords_of_term level1_pattern in
244 let rule_id = p_atoms in
245 List.iter CicNotationLexer.add_level2_ast_keyword keywords;
246 Hashtbl.add owned_keywords rule_id keywords; (* keywords may be [] *)
250 let atoms = rule_id in
252 let keywords = Hashtbl.find owned_keywords rule_id in
253 List.iter CicNotationLexer.remove_level2_ast_keyword keywords
254 with Not_found -> assert false);
255 Grammar.delete_rule term atoms
259 let parse_level1_pattern_ref = ref (fun _ -> assert false)
260 let parse_level2_ast_ref = ref (fun _ -> assert false)
261 let parse_level2_meta_ref = ref (fun _ -> assert false)
263 let fold_cluster binder terms ty body =
265 (fun term body -> Binder (binder, (term, ty), body))
266 terms body (* terms are names: either Ident or FreshVar *)
268 let fold_binder binder pt_names body =
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 "term" *)
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 (term: 'a Grammar.Entry.e),
286 mk_level_list min_precedence max_precedence ]
288 (* {{{ Grammar for concrete syntax patterns, notation level 1 *)
290 GLOBAL: level1_pattern;
292 level1_pattern: [ [ p = l1_pattern; EOI -> CicNotationUtil.boxify p ] ];
293 l1_pattern: [ [ p = LIST1 l1_simple_pattern -> p ] ];
295 [ s = SYMBOL -> `Symbol s
296 | k = QKEYWORD -> `Keyword k
297 | n = NUMBER -> `Number n
300 sep: [ [ "sep"; sep = literal -> sep ] ];
301 (* row_sep: [ [ "rowsep"; sep = literal -> sep ] ];
302 field_sep: [ [ "fieldsep"; sep = literal -> sep ] ]; *)
304 [ "list0"; p = l1_simple_pattern; sep = OPT sep -> List0 (p, sep)
305 | "list1"; p = l1_simple_pattern; sep = OPT sep -> List1 (p, sep)
306 | "opt"; p = l1_simple_pattern -> Opt p
309 l1_pattern_variable: [
310 [ "term"; id = IDENT -> TermVar id
311 | "number"; id = IDENT -> NumVar id
312 | "ident"; id = IDENT -> IdentVar id
317 [ p1 = SELF; SYMBOL "\\sub"; p2 = SELF ->
318 return_term loc (Layout (Sub (p1, p2)))
319 | p1 = SELF; SYMBOL "\\sup"; p2 = SELF ->
320 return_term loc (Layout (Sup (p1, p2)))
321 | p1 = SELF; SYMBOL "\\below"; p2 = SELF ->
322 return_term loc (Layout (Below (p1, p2)))
323 | p1 = SELF; SYMBOL "\\above"; p2 = SELF ->
324 return_term loc (Layout (Above (p1, p2)))
325 | p1 = SELF; SYMBOL "\\over"; p2 = SELF ->
326 return_term loc (Layout (Over (p1, p2)))
327 | p1 = SELF; SYMBOL "\\atop"; p2 = SELF ->
328 return_term loc (Layout (Atop (p1, p2)))
329 (* | "array"; p = SELF; csep = OPT field_sep; rsep = OPT row_sep ->
330 return_term loc (Array (p, csep, rsep)) *)
331 | SYMBOL "\\frac"; p1 = SELF; p2 = SELF ->
332 return_term loc (Layout (Frac (p1, p2)))
333 | SYMBOL "\\sqrt"; p = SELF -> return_term loc (Layout (Sqrt p))
334 | SYMBOL "\\root"; index = SELF; SYMBOL "\\OF"; arg = SELF ->
335 return_term loc (Layout (Root (arg, index)))
336 | "hbox"; LPAREN; p = l1_pattern; RPAREN ->
337 return_term loc (Layout (Box ((H, false, false), p)))
338 | "vbox"; LPAREN; p = l1_pattern; RPAREN ->
339 return_term loc (Layout (Box ((V, false, false), p)))
340 | "hvbox"; LPAREN; p = l1_pattern; RPAREN ->
341 return_term loc (Layout (Box ((HV, false, false), p)))
342 | "hovbox"; LPAREN; p = l1_pattern; RPAREN ->
343 return_term loc (Layout (Box ((HOV, false, false), p)))
344 | "break" -> return_term loc (Layout Break)
345 (* | SYMBOL "\\SPACE" -> return_term loc (Layout Space) *)
346 | LPAREN; p = l1_pattern; RPAREN ->
347 return_term loc (CicNotationUtil.group p)
350 [ i = IDENT -> return_term loc (Variable (TermVar i))
351 | m = l1_magic_pattern -> return_term loc (Magic m)
352 | v = l1_pattern_variable -> return_term loc (Variable v)
353 | l = literal -> return_term loc (Literal l)
363 [ "term"; id = IDENT -> TermVar id
364 | "number"; id = IDENT -> NumVar id
365 | "ident"; id = IDENT -> IdentVar id
366 | "fresh"; id = IDENT -> FreshVar id
367 | "anonymous" -> TermVar "_"
368 | id = IDENT -> TermVar id
372 [ "fold"; kind = [ "left" -> `Left | "right" -> `Right ];
373 base = level2_meta; "rec"; id = IDENT; recursive = level2_meta ->
374 Fold (kind, base, [id], recursive)
375 | "default"; some = level2_meta; none = level2_meta -> Default (some, none)
376 | "if"; p_test = level2_meta;
377 "then"; p_true = level2_meta;
378 "else"; p_false = level2_meta ->
379 If (p_test, p_true, p_false)
384 [ magic = l2_magic -> Magic magic
385 | var = l2_variable -> Variable var
386 | blob = UNPARSED_AST -> !parse_level2_ast_ref (Stream.of_string blob)
392 GLOBAL: level2_ast term let_defs;
393 (* {{{ Grammar for ast patterns, notation level 2 *)
394 level2_ast: [ [ p = term -> p ] ];
402 [ SYMBOL "\\subst"; (* to avoid catching frequent "a [1]" cases *)
405 i = IDENT; SYMBOL <:unicode<Assign>> (* ≔ *); t = term -> (i, t)
412 [ s = SYMBOL "_" -> None
413 | p = term -> Some p ]
416 [ SYMBOL "["; substs = LIST0 meta_subst; SYMBOL "]" -> substs ]
418 possibly_typed_name: [
419 [ LPAREN; id = single_arg; SYMBOL ":"; typ = term; RPAREN ->
421 | arg = single_arg -> arg, None
425 [ id = IDENT -> id, []
426 | LPAREN; id = IDENT; vars = LIST1 possibly_typed_name; RPAREN ->
431 [ SYMBOL <:unicode<Pi>> (* Π *) -> `Pi
432 | SYMBOL <:unicode<exists>> (* ∃ *) -> `Exists
433 | SYMBOL <:unicode<forall>> (* ∀ *) -> `Forall
434 | SYMBOL <:unicode<lambda>> (* λ *) -> `Lambda
438 [ LPAREN; names = LIST1 IDENT SEP SYMBOL ",";
439 SYMBOL ":"; ty = term; RPAREN ->
440 List.map (fun n -> Ident (n, None)) names, Some ty
441 | name = IDENT -> [Ident (name, None)], None
442 | blob = UNPARSED_META ->
443 let meta = !parse_level2_meta_ref (Stream.of_string blob) in
445 | Variable (FreshVar _) -> [meta], None
446 | Variable (TermVar "_") -> [Ident ("_", None)], None
447 | _ -> failwith "Invalid bound name."
451 [ name = IDENT -> Ident (name, None)
452 | blob = UNPARSED_META ->
453 let meta = !parse_level2_meta_ref (Stream.of_string blob) in
455 | Variable (FreshVar _) -> meta
456 | Variable (TermVar "_") -> Ident ("_", None)
457 | _ -> failwith "Invalid index name."
461 [ "rec" -> `Inductive
462 | "corec" -> `CoInductive
469 index_name = OPT [ "on"; id = single_arg -> id ];
470 ty = OPT [ SYMBOL ":" ; p = term -> p ];
471 SYMBOL <:unicode<def>> (* ≝ *); body = term ->
472 let body = fold_binder `Lambda args body in
476 | Some ty -> Some (fold_binder `Pi args ty)
478 let rec position_of name p = function
480 | n :: _ when n = name -> Some p, p
481 | _ :: tl -> position_of name (p + 1) tl
483 let rec find_arg name n = function
485 fail loc (sprintf "Argument %s not found"
486 (CicNotationPp.pp_term name))
488 (match position_of name 0 l with
489 | None, len -> find_arg name (n + len) tl
490 | Some where, len -> n + where)
493 match index_name with
495 | Some index_name -> find_arg index_name 0 args
497 (name, ty), body, index
504 l = LIST1 single_arg SEP SYMBOL "," -> l
505 | SYMBOL "_" -> [Ident ("_", None)] ];
506 typ = OPT [ SYMBOL ":"; t = term -> t ] -> (vars, typ)
509 l = LIST1 single_arg SEP SYMBOL "," -> l
510 | SYMBOL "_" -> [Ident ("_", None)] ];
511 typ = OPT [ SYMBOL ":"; t = term -> t ];
512 RPAREN -> (vars, typ)
515 term: LEVEL "10" (* let in *)
517 [ "let"; var = possibly_typed_name; SYMBOL <:unicode<def>> (* ≝ *);
518 p1 = term; "in"; p2 = term ->
519 return_term loc (LetIn (var, p1, p2))
520 | "let"; k = induction_kind; defs = let_defs; "in";
522 return_term loc (LetRec (k, defs, body))
525 term: LEVEL "20" (* binder *)
527 [ b = binder; (vars, typ) = binder_vars; SYMBOL "."; body = term ->
528 return_term loc (fold_cluster b vars typ body)
531 term: LEVEL "70" (* apply *)
533 [ p1 = term; p2 = term ->
534 let rec aux = function
536 | AttributedTerm (_, Appl (hd :: tl)) ->
540 return_term loc (Appl (aux p1 @ [p2]))
543 term: LEVEL "90" (* simple *)
545 [ id = IDENT -> return_term loc (Ident (id, None))
546 | id = IDENT; s = explicit_subst -> return_term loc (Ident (id, Some s))
547 | s = CSYMBOL -> return_term loc (Symbol (s, 0))
548 | u = URI -> return_term loc (Uri (u, None))
549 | n = NUMBER -> return_term loc (Num (n, 0))
550 | IMPLICIT -> return_term loc (Implicit)
551 | PLACEHOLDER -> return_term loc UserInput
552 | m = META -> return_term loc (Meta (int_of_string m, []))
553 | m = META; s = meta_substs -> return_term loc (Meta (int_of_string m, s))
554 | s = sort -> return_term loc (Sort s)
555 | outtyp = OPT [ SYMBOL "["; ty = term; SYMBOL "]" -> ty ];
557 indty_ident = OPT [ "in"; id = IDENT -> id ];
560 lhs = match_pattern; SYMBOL <:unicode<Rightarrow>> (* ⇒ *);
565 return_term loc (Case (t, indty_ident, outtyp, patterns))
566 | LPAREN; p1 = term; SYMBOL ":"; p2 = term; RPAREN ->
567 return_term loc (Cast (p1, p2))
568 | LPAREN; p = term; RPAREN -> p
569 | blob = UNPARSED_META -> !parse_level2_meta_ref (Stream.of_string blob)
575 (** {2 API implementation} *)
577 let exc_located_wrapper f =
581 | Stdpp.Exc_located (floc, Stream.Error msg) ->
582 raise (Parse_error (floc, msg))
583 | Stdpp.Exc_located (floc, exn) ->
584 raise (Parse_error (floc, (Printexc.to_string exn)))
586 let parse_level1_pattern stream =
587 exc_located_wrapper (fun () -> Grammar.Entry.parse level1_pattern stream)
588 let parse_level2_ast stream =
589 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_ast stream)
590 let parse_level2_meta stream =
591 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_meta stream)
594 parse_level1_pattern_ref := parse_level1_pattern;
595 parse_level2_ast_ref := parse_level2_ast;
596 parse_level2_meta_ref := parse_level2_meta
600 let print_l2_pattern () =
601 Grammar.print_entry Format.std_formatter (Grammar.Entry.obj term);
602 Format.pp_print_flush Format.std_formatter ();
605 (* vim:set encoding=utf8 foldmethod=marker: *)