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
48 let let_in_assoc = Gramext.NonA
49 let binder_assoc = Gramext.RightA
50 let apply_assoc = Gramext.LeftA
51 let simple_assoc = Gramext.NonA
54 Grammar.Entry.create level1_pattern_grammar "level1_pattern"
55 let level2_ast = Grammar.Entry.create level2_ast_grammar "level2_ast"
56 let term = Grammar.Entry.create level2_ast_grammar "term"
57 let level2_meta = Grammar.Entry.create level2_meta_grammar "level2_meta"
59 let level3_term = Grammar.Entry.create level2_ast_grammar "level3_term"
60 let notation = (* level1 <-> level 2 *)
61 Grammar.Entry.create level2_ast_grammar "notation"
62 let interpretation = (* level2 <-> level 3 *)
63 Grammar.Entry.create level2_ast_grammar "interpretation"
64 let phrase = Grammar.Entry.create level2_ast_grammar "phrase"
66 let return_term loc term = ()
69 let (x, y) = loc_of_floc floc in
70 failwith (sprintf "Error at characters %d - %d: %s" x y msg)
74 Pervasives.int_of_string s
76 failwith (sprintf "Lexer failure: string_of_int \"%s\" failed" s)
78 (** {2 Grammar extension} *)
80 let gram_symbol s = Gramext.Stoken ("SYMBOL", s)
81 let gram_ident s = Gramext.Stoken ("IDENT", s)
82 let gram_number s = Gramext.Stoken ("NUMBER", s)
83 let gram_keyword s = Gramext.Stoken ("", s)
84 let gram_term = Gramext.Sself
88 | `Symbol s -> gram_symbol s
89 | `Keyword s -> gram_keyword s
90 | `Number s -> gram_number s
94 | Binding of string * value_type
95 | Env of (string * value_type) list
97 let make_action action bindings =
98 let rec aux (vl : CicNotationEnv.t) =
100 [] -> Gramext.action (fun (loc: location) -> action vl loc)
101 | NoBinding :: tl -> Gramext.action (fun _ -> aux vl tl)
102 (* LUCA: DEFCON 5 BEGIN *)
103 | Binding (name, TermType) :: tl ->
105 (fun (v:term) -> aux ((name, (TermType, TermValue v))::vl) tl)
106 | Binding (name, StringType) :: tl ->
109 aux ((name, (StringType, StringValue v)) :: vl) tl)
110 | Binding (name, NumType) :: tl ->
112 (fun (v:string) -> aux ((name, (NumType, NumValue v)) :: vl) tl)
113 | Binding (name, OptType t) :: tl ->
115 (fun (v:'a option) ->
116 aux ((name, (OptType t, OptValue v)) :: vl) tl)
117 | Binding (name, ListType t) :: tl ->
120 aux ((name, (ListType t, ListValue v)) :: vl) tl)
122 Gramext.action (fun (v:CicNotationEnv.t) -> aux (v @ vl) tl)
123 (* LUCA: DEFCON 5 END *)
125 aux [] (List.rev bindings)
131 | NoBinding :: tl -> aux acc tl
132 | Env names :: tl -> aux (List.rev names @ acc) tl
133 | Binding (name, ty) :: tl -> aux ((name, ty) :: acc) tl
137 (* given a level 1 pattern computes the new RHS of "term" grammar entry *)
138 let extract_term_production pattern =
139 let rec aux = function
140 | AttributedTerm (_, t) -> aux t
141 | Literal l -> aux_literal l
142 | Layout l -> aux_layout l
143 | Magic m -> aux_magic m
144 | Variable v -> aux_variable v
146 prerr_endline (CicNotationPp.pp_term t);
150 | `Symbol s -> [NoBinding, gram_symbol s]
152 (* assumption: s will be registered as a keyword with the lexer *)
153 [NoBinding, gram_keyword s]
154 | `Number s -> [NoBinding, gram_number s]
155 and aux_layout = function
156 | Sub (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\sub"] @ aux p2
157 | Sup (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\sup"] @ aux p2
158 | Below (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\below"] @ aux p2
159 | Above (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\above"] @ aux p2
160 | Frac (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\frac"] @ aux p2
161 | Atop (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\atop"] @ aux p2
162 | Over (p1, p2) -> aux p1 @ [NoBinding, gram_symbol "\\over"] @ aux p2
164 [NoBinding, gram_symbol "\\root"] @ aux p2
165 @ [NoBinding, gram_symbol "\\of"] @ aux p1
166 | Sqrt p -> [NoBinding, gram_symbol "\\sqrt"] @ aux p
168 | Box (_, pl) -> List.flatten (List.map aux pl)
169 and aux_magic magic =
172 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
173 let action (env_opt : CicNotationEnv.t option) (loc : location) =
175 | Some env -> List.map opt_binding_some env
176 | None -> List.map opt_binding_of_name p_names
178 [ Env (List.map opt_declaration p_names),
180 [ [ Gramext.Sopt (Gramext.srules [ p_atoms, p_action ]) ],
181 Gramext.action action ] ]
184 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
185 (* let env0 = List.map list_binding_of_name p_names in
186 let grow_env_entry env n v =
189 | (n', (ty, ListValue vl)) as entry ->
190 if n' = n then n', (ty, ListValue (v :: vl)) else entry
194 let grow_env env_i env =
196 (fun env (n, (_, v)) -> grow_env_entry env n v)
199 let action (env_list : CicNotationEnv.t list) (loc : location) =
200 CicNotationEnv.coalesce_env p_names env_list
204 | List0 (_, None) -> Gramext.Slist0 s
205 | List1 (_, None) -> Gramext.Slist1 s
206 | List0 (_, Some l) -> Gramext.Slist0sep (s, gram_of_literal l)
207 | List1 (_, Some l) -> Gramext.Slist1sep (s, gram_of_literal l)
210 [ Env (List.map list_declaration p_names),
212 [ [ gram_of_list (Gramext.srules [ p_atoms, p_action ]) ],
213 Gramext.action action ] ]
217 | NumVar s -> [Binding (s, NumType), gram_number ""]
218 | TermVar s -> [Binding (s, TermType), gram_term]
219 | IdentVar s -> [Binding (s, StringType), gram_ident ""]
220 | Ascription (p, s) -> assert false (* TODO *)
221 | FreshVar _ -> assert false
222 and inner_pattern p =
223 let p_bindings, p_atoms = List.split (aux p) in
224 let p_names = flatten_opt p_bindings in
226 make_action (fun (env : CicNotationEnv.t) (loc : location) -> env)
229 p_bindings, p_atoms, p_names, action
233 let level_of_int precedence =
234 if precedence < min_precedence || precedence > max_precedence then
235 raise (Level_not_found precedence);
236 string_of_int precedence
238 type rule_id = Token.t Gramext.g_symbol list
240 (* mapping: rule_id -> owned keywords. (rule_id, string list) Hashtbl.t *)
241 let owned_keywords = Hashtbl.create 23
243 let extend level1_pattern ?(precedence = default_precedence)
244 ?associativity action
246 let p_bindings, p_atoms =
247 List.split (extract_term_production level1_pattern)
249 let level = level_of_int precedence in
250 let p_names = flatten_opt p_bindings in
253 [ Grammar.Entry.obj (term: 'a Grammar.Entry.e),
254 Some (Gramext.Level level),
259 (fun (env: CicNotationEnv.t) (loc: location) -> (action env loc))
262 let keywords = CicNotationUtil.keywords_of_term level1_pattern in
263 let rule_id = p_atoms in
264 List.iter CicNotationLexer.add_level2_ast_keyword keywords;
265 Hashtbl.add owned_keywords rule_id keywords; (* keywords may be [] *)
269 let atoms = rule_id in
271 let keywords = Hashtbl.find owned_keywords rule_id in
272 List.iter CicNotationLexer.remove_level2_ast_keyword keywords
273 with Not_found -> assert false);
274 Grammar.delete_rule term atoms
278 let parse_level1_pattern_ref = ref (fun _ -> assert false)
279 let parse_level2_ast_ref = ref (fun _ -> assert false)
280 let parse_level2_meta_ref = ref (fun _ -> assert false)
282 let fold_binder binder pt_names body =
283 let fold_cluster binder terms ty body =
285 (fun term body -> Binder (binder, (term, ty), body))
286 terms body (* terms are names: either Ident or FreshVar *)
289 (fun (names, ty) body -> fold_cluster binder names ty body)
292 let return_term loc term = AttributedTerm (`Loc loc, term)
294 let _ = (* create empty precedence level for "term" *)
295 let mk_level_list first last =
296 let rec aux acc = function
297 | i when i < first -> acc
298 | i -> aux ((Some (string_of_int i), None, []) :: acc) (i - 1)
303 [ Grammar.Entry.obj (term: 'a Grammar.Entry.e),
305 mk_level_list min_precedence max_precedence ]
307 (* {{{ Grammar for concrete syntax patterns, notation level 1 *)
309 GLOBAL: level1_pattern;
311 level1_pattern: [ [ p = l1_pattern; EOI -> CicNotationUtil.boxify p ] ];
312 l1_pattern: [ [ p = LIST1 l1_simple_pattern -> p ] ];
314 [ s = SYMBOL -> `Symbol s
315 | k = QKEYWORD -> `Keyword k
316 | n = NUMBER -> `Number n
319 sep: [ [ "sep"; sep = literal -> sep ] ];
320 (* row_sep: [ [ "rowsep"; sep = literal -> sep ] ];
321 field_sep: [ [ "fieldsep"; sep = literal -> sep ] ]; *)
323 [ "list0"; p = l1_simple_pattern; sep = OPT sep -> List0 (p, sep)
324 | "list1"; p = l1_simple_pattern; sep = OPT sep -> List1 (p, sep)
325 | "opt"; p = l1_simple_pattern -> Opt p
328 l1_pattern_variable: [
329 [ "term"; id = IDENT -> TermVar id
330 | "number"; id = IDENT -> NumVar id
331 | "ident"; id = IDENT -> IdentVar id
336 [ p1 = SELF; SYMBOL "\\sub"; p2 = SELF ->
337 return_term loc (Layout (Sub (p1, p2)))
338 | p1 = SELF; SYMBOL "\\sup"; p2 = SELF ->
339 return_term loc (Layout (Sup (p1, p2)))
340 | p1 = SELF; SYMBOL "\\below"; p2 = SELF ->
341 return_term loc (Layout (Below (p1, p2)))
342 | p1 = SELF; SYMBOL "\\above"; p2 = SELF ->
343 return_term loc (Layout (Above (p1, p2)))
344 | p1 = SELF; SYMBOL "\\over"; p2 = SELF ->
345 return_term loc (Layout (Over (p1, p2)))
346 | p1 = SELF; SYMBOL "\\atop"; p2 = SELF ->
347 return_term loc (Layout (Atop (p1, p2)))
348 (* | "array"; p = SELF; csep = OPT field_sep; rsep = OPT row_sep ->
349 return_term loc (Array (p, csep, rsep)) *)
350 | SYMBOL "\\frac"; p1 = SELF; p2 = SELF ->
351 return_term loc (Layout (Frac (p1, p2)))
352 | SYMBOL "\\sqrt"; p = SELF -> return_term loc (Layout (Sqrt p))
353 | SYMBOL "\\root"; index = SELF; SYMBOL "\\OF"; arg = SELF ->
354 return_term loc (Layout (Root (arg, index)))
355 | "hbox"; LPAREN; p = l1_pattern; RPAREN ->
356 return_term loc (CicNotationUtil.boxify p)
357 | "vbox"; LPAREN; p = l1_pattern; RPAREN ->
358 return_term loc (CicNotationUtil.boxify p)
359 | "hvbox"; LPAREN; p = l1_pattern; RPAREN ->
360 return_term loc (CicNotationUtil.boxify p)
361 | "hovbox"; LPAREN; p = l1_pattern; RPAREN ->
362 return_term loc (CicNotationUtil.boxify p)
363 | "break" -> return_term loc (Layout Break)
364 (* | SYMBOL "\\SPACE" -> return_term loc (Layout Space) *)
365 | LPAREN; p = l1_pattern; RPAREN ->
366 return_term loc (CicNotationUtil.boxify p)
369 [ i = IDENT -> return_term loc (Variable (TermVar i))
370 | m = l1_magic_pattern -> return_term loc (Magic m)
371 | v = l1_pattern_variable -> return_term loc (Variable v)
372 | l = literal -> return_term loc (Literal l)
382 [ "term"; id = IDENT -> TermVar id
383 | "number"; id = IDENT -> NumVar id
384 | "ident"; id = IDENT -> IdentVar id
385 | "fresh"; id = IDENT -> FreshVar id
386 | "anonymous" -> TermVar "_"
387 | id = IDENT -> TermVar id
391 [ "fold"; kind = [ "left" -> `Left | "right" -> `Right ];
392 base = level2_meta; "rec"; id = IDENT; recursive = level2_meta ->
393 Fold (kind, base, [id], recursive)
394 | "default"; some = level2_meta; none = level2_meta -> Default (some, none)
395 | "if"; p_test = level2_meta;
396 "then"; p_true = level2_meta;
397 "else"; p_false = level2_meta ->
398 If (p_test, p_true, p_false)
403 [ magic = l2_magic -> Magic magic
404 | var = l2_variable -> Variable var
405 | blob = UNPARSED_AST -> !parse_level2_ast_ref (Stream.of_string blob)
411 GLOBAL: level2_ast term
413 notation interpretation
415 (* {{{ Grammar for ast patterns, notation level 2 *)
416 level2_ast: [ [ p = term -> p ] ];
424 [ SYMBOL "\\subst"; (* to avoid catching frequent "a [1]" cases *)
427 i = IDENT; SYMBOL <:unicode<Assign>> (* ≔ *); t = term -> (i, t)
434 [ s = SYMBOL "_" -> None
435 | p = term -> Some p ]
438 [ SYMBOL "["; substs = LIST0 meta_subst; SYMBOL "]" -> substs ]
440 possibly_typed_name: [
441 [ LPAREN; id = bound_name; SYMBOL ":"; typ = term; RPAREN ->
443 | id = bound_name -> id, None
447 [ id = IDENT -> id, []
448 | LPAREN; id = IDENT; vars = LIST1 possibly_typed_name; RPAREN ->
453 [ SYMBOL <:unicode<Pi>> (* Π *) -> `Pi
454 | SYMBOL <:unicode<exists>> (* ∃ *) -> `Exists
455 | SYMBOL <:unicode<forall>> (* ∀ *) -> `Forall
456 | SYMBOL <:unicode<lambda>> (* λ *) -> `Lambda
460 [ i = IDENT -> Ident (i, None)
461 | SYMBOL "\\FRESH"; i = IDENT -> Variable (FreshVar i)
465 [ vars = LIST1 bound_name SEP SYMBOL ",";
466 ty = OPT [ SYMBOL ":"; p = term -> p ] ->
470 vars = LIST1 bound_name SEP SYMBOL ",";
471 ty = OPT [ SYMBOL ":"; p = term -> p ];
479 [ "rec" -> `Inductive
480 | "corec" -> `CoInductive
485 name = bound_name; args = bound_names;
486 index_name = OPT [ "on"; id = bound_name -> id ];
487 ty = OPT [ SYMBOL ":" ; p = term -> p ];
488 SYMBOL <:unicode<def>> (* ≝ *); body = term ->
489 let body = fold_binder `Lambda args body in
493 | Some ty -> Some (fold_binder `Pi args ty)
495 let rec position_of name p = function
497 | n :: _ when n = name -> Some p, p
498 | _ :: tl -> position_of name (p + 1) tl
500 let rec find_arg name n = function
502 fail loc (sprintf "Argument %s not found"
503 (CicNotationPp.pp_term name))
505 (match position_of name 0 l with
506 | None, len -> find_arg name (n + len) tl
507 | Some where, len -> n + where)
510 match index_name with
512 | Some name -> find_arg name 0 args
514 (name, ty), body, index
519 term: LEVEL "10" (* let in *)
521 [ "let"; var = possibly_typed_name; SYMBOL <:unicode<def>> (* ≝ *);
522 p1 = term; "in"; p2 = term ->
523 return_term loc (LetIn (var, p1, p2))
524 | "let"; k = induction_kind; defs = let_defs; "in";
526 return_term loc (LetRec (k, defs, body))
529 term: LEVEL "20" (* binder *)
531 [ b = binder; names = bound_names; SYMBOL "."; body = term ->
532 return_term loc (fold_binder b names body)
535 term: LEVEL "70" (* apply *)
537 [ p1 = term; p2 = term ->
538 let rec aux = function
540 | AttributedTerm (_, Appl (hd :: tl)) ->
544 return_term loc (Appl (aux p1 @ [p2]))
547 term: LEVEL "90" (* simple *)
549 [ id = IDENT -> return_term loc (Ident (id, None))
550 | id = IDENT; s = explicit_subst -> return_term loc (Ident (id, Some s))
551 | s = CSYMBOL -> return_term loc (Symbol (s, 0))
552 | u = URI -> return_term loc (Uri (u, None))
553 | n = NUMBER -> return_term loc (Num (n, 0))
554 | IMPLICIT -> return_term loc (Implicit)
555 | m = META -> return_term loc (Meta (int_of_string m, []))
556 | m = META; s = meta_substs -> return_term loc (Meta (int_of_string m, s))
557 | s = sort -> return_term loc (Sort s)
558 | outtyp = OPT [ SYMBOL "["; ty = term; SYMBOL "]" -> ty ];
560 indty_ident = OPT [ SYMBOL ":"; id = IDENT -> id ];
563 lhs = match_pattern; SYMBOL <:unicode<Rightarrow>> (* ⇒ *);
568 return_term loc (Case (t, indty_ident, outtyp, patterns))
569 | LPAREN; p1 = term; SYMBOL ":"; p2 = term; RPAREN ->
570 return_term loc (Appl [ Symbol ("cast", 0); p1; p2 ])
571 | LPAREN; p = term; RPAREN -> p
572 | blob = UNPARSED_META -> !parse_level2_meta_ref (Stream.of_string blob)
576 (* {{{ Grammar for interpretation, notation level 3 *)
578 [ id = IDENT -> IdentArg (0, id)
579 | l = LIST1 [ SYMBOL <:unicode<eta>> (* η *) -> () ] SEP SYMBOL ".";
580 SYMBOL "."; id = IDENT ->
581 IdentArg (List.length l, id)
585 [ u = URI -> UriPattern (UriManager.uri_of_string u)
586 | id = IDENT -> VarPattern id
587 | LPAREN; terms = LIST1 SELF; RPAREN ->
591 | terms -> ApplPattern terms)
595 (* {{{ Notation glues *)
597 [ IDENT "left"; IDENT "associative" -> Gramext.LeftA
598 | IDENT "right"; IDENT "associative" -> Gramext.RightA
599 | IDENT "non"; IDENT "associative" -> Gramext.NonA
603 [ "with"; IDENT "precedence"; n = NUMBER -> int_of_string n ]
607 assoc = OPT associativity; prec = OPT precedence;
610 [ blob = UNPARSED_AST -> !parse_level2_ast_ref (Stream.of_string blob)
611 | blob = UNPARSED_META ->
612 !parse_level2_meta_ref (Stream.of_string blob) ]
614 (!parse_level1_pattern_ref (Stream.of_string s), assoc, prec, p2)
618 [ s = CSYMBOL; args = LIST1 argument; SYMBOL "="; t = level3_term ->
623 (* {{{ Top-level phrases *)
625 [ IDENT "print"; t = term; SYMBOL "." -> Print t
626 | IDENT "notation"; (l1, assoc, prec, l2) = notation; SYMBOL "." ->
627 Notation (l1, assoc, prec, l2)
628 | IDENT "interpretation"; (symbol, args, l3) = interpretation; SYMBOL "." ->
629 Interpretation ((symbol, args), l3)
630 | IDENT "render"; u = URI; SYMBOL "." -> Render (UriManager.uri_of_string u)
631 | IDENT "dump"; SYMBOL "." -> Dump
637 (** {2 API implementation} *)
639 let exc_located_wrapper f =
643 | Stdpp.Exc_located (floc, Stream.Error msg) ->
644 raise (Parse_error (floc, msg))
645 | Stdpp.Exc_located (floc, exn) ->
646 raise (Parse_error (floc, (Printexc.to_string exn)))
648 let parse_level1_pattern stream =
649 exc_located_wrapper (fun () -> Grammar.Entry.parse level1_pattern stream)
650 let parse_level2_ast stream =
651 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_ast stream)
652 let parse_level2_meta stream =
653 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_meta stream)
654 let parse_interpretation stream =
655 exc_located_wrapper (fun () -> Grammar.Entry.parse level3_term stream)
656 let parse_phrase stream =
657 exc_located_wrapper (fun () -> Grammar.Entry.parse phrase stream)
660 parse_level1_pattern_ref := parse_level1_pattern;
661 parse_level2_ast_ref := parse_level2_ast;
662 parse_level2_meta_ref := parse_level2_meta
666 let print_l2_pattern () =
667 Grammar.print_entry Format.std_formatter (Grammar.Entry.obj term);
668 Format.pp_print_flush Format.std_formatter ();
671 (* vim:set encoding=utf8 foldmethod=marker: *)