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
45 let let_in_assoc = Gramext.NonA
46 let binder_assoc = Gramext.RightA
47 let apply_assoc = Gramext.LeftA
48 let simple_assoc = Gramext.NonA
50 let level1_pattern = Grammar.Entry.create grammar "level1_pattern"
51 let level2_pattern = Grammar.Entry.create grammar "level2_pattern"
52 let level3_term = Grammar.Entry.create grammar "level3_term"
53 let l2_pattern = Grammar.Entry.create grammar "l2_pattern"
54 let notation = Grammar.Entry.create grammar "notation" (* level1 <-> level 2 *)
56 Grammar.Entry.create grammar "interpretation" (* level2 <-> level 3 *)
57 let phrase = Grammar.Entry.create grammar "phrase"
59 let return_term loc term = ()
62 let (x, y) = loc_of_floc floc in
63 failwith (sprintf "Error at characters %d - %d: %s" x y msg)
67 Pervasives.int_of_string s
69 failwith (sprintf "Lexer failure: string_of_int \"%s\" failed" s)
71 (** {2 Grammar extension} *)
73 let symbol s = Gramext.Stoken ("SYMBOL", s)
74 let ident s = Gramext.Stoken ("IDENT", s)
75 let number s = Gramext.Stoken ("NUMBER", s)
76 let term = Gramext.Sself
78 let g_symbol_of_literal =
80 | `Symbol s -> symbol s
81 | `Keyword s -> ident s
82 | `Number s -> number s
86 | Binding of string * value_type
87 | Env of (string * value_type) list
89 let make_action action bindings =
90 let rec aux (vl : CicNotationEnv.t) =
92 [] -> Gramext.action (fun (loc: location) -> action vl loc)
93 | NoBinding :: tl -> Gramext.action (fun _ -> aux vl tl)
94 (* LUCA: DEFCON 5 BEGIN *)
95 | Binding (name, TermType) :: tl ->
97 (fun (v:term) -> aux ((name, (TermType, TermValue v))::vl) tl)
98 | Binding (name, StringType) :: tl ->
101 aux ((name, (StringType, StringValue v)) :: vl) tl)
102 | Binding (name, NumType) :: tl ->
104 (fun (v:string) -> aux ((name, (NumType, NumValue v)) :: vl) tl)
105 | Binding (name, OptType t) :: tl ->
107 (fun (v:'a option) ->
108 aux ((name, (OptType t, OptValue v)) :: vl) tl)
109 | Binding (name, ListType t) :: tl ->
112 aux ((name, (ListType t, ListValue v)) :: vl) tl)
114 Gramext.action (fun (v:CicNotationEnv.t) -> aux (v @ vl) tl)
115 (* LUCA: DEFCON 5 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 : CicNotationEnv.t 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 =
179 | (n', (ty, ListValue vl)) as entry ->
180 if n' = n then n', (ty, ListValue (v :: vl)) else entry
184 let grow_env env_i env =
186 (fun env (n, (_, v)) -> grow_env_entry env n v)
189 let action (env_list : CicNotationEnv.t list) (loc : location) =
190 CicNotationEnv.coalesce_env p_names env_list
194 | List0 (_, None) -> Gramext.Slist0 s
195 | List1 (_, None) -> Gramext.Slist1 s
196 | List0 (_, Some l) -> Gramext.Slist0sep (s, g_symbol_of_literal l)
197 | List1 (_, Some l) -> Gramext.Slist1sep (s, g_symbol_of_literal l)
200 [ Env (List.map list_declaration p_names),
202 [ [ g_symbol (Gramext.srules [ p_atoms, p_action ]) ],
203 Gramext.action action ] ]
207 | NumVar s -> [Binding (s, NumType), number ""]
208 | TermVar s -> [Binding (s, TermType), term]
209 | IdentVar s -> [Binding (s, StringType), ident ""]
210 | Ascription (p, s) -> assert false (* TODO *)
211 | FreshVar _ -> assert false
212 and inner_pattern p =
213 let p_bindings, p_atoms = List.split (aux p) in
214 let p_names = flatten_opt p_bindings in
215 let _ = prerr_endline ("inner names: " ^ String.concat " " (List.map fst p_names)) in
217 make_action (fun (env : CicNotationEnv.t) (loc : location) -> env)
220 p_bindings, p_atoms, p_names, action
224 let level_of_int precedence =
225 if precedence < min_precedence || precedence > max_precedence then
226 raise (Level_not_found precedence);
227 string_of_int precedence
229 type rule_id = Token.t Gramext.g_symbol list
231 let extend level1_pattern ?(precedence = default_precedence)
232 ?associativity action
234 let p_bindings, p_atoms =
235 List.split (extract_term_production level1_pattern)
237 let level = level_of_int precedence in
238 let p_names = flatten_opt p_bindings in
240 prerr_endline (string_of_int (List.length p_bindings));
242 [ Grammar.Entry.obj (l2_pattern: 'a Grammar.Entry.e),
243 Some (Gramext.Level level),
248 (fun (env: CicNotationEnv.t) (loc: location) -> (action env loc))
253 let delete atoms = Grammar.delete_rule l2_pattern atoms
257 let fold_binder binder pt_names body =
258 let fold_cluster binder terms ty body =
260 (fun term body -> Binder (binder, (term, ty), body))
261 terms body (* terms are names: either Ident or FreshVar *)
264 (fun (names, ty) body -> fold_cluster binder names ty body)
267 let return_term loc term = AttributedTerm (`Loc loc, term)
269 let _ = (* create empty precedence level for "l2_pattern" *)
270 let mk_level_list first last =
271 let rec aux acc = function
272 | i when i < first -> acc
273 | i -> aux ((Some (string_of_int i), None, []) :: acc) (i - 1)
278 [ Grammar.Entry.obj (l2_pattern: 'a Grammar.Entry.e),
280 mk_level_list min_precedence max_precedence ]
283 GLOBAL: level1_pattern level2_pattern level3_term
285 notation interpretation
287 (* {{{ Grammar for concrete syntax patterns, notation level 1 *)
288 level1_pattern: [ [ p = l1_simple_pattern -> p ] ];
289 l1_pattern: [ [ p = LIST1 l1_simple_pattern -> p ] ];
291 [ s = SYMBOL -> `Symbol s
292 | k = KEYWORD -> `Keyword k
293 | n = NUMBER -> `Number n
296 sep: [ [ SYMBOL "\\SEP"; sep = literal -> sep ] ];
297 (* row_sep: [ [ SYMBOL "\\ROWSEP"; sep = literal -> sep ] ];
298 field_sep: [ [ SYMBOL "\\FIELDSEP"; sep = literal -> sep ] ]; *)
300 [ SYMBOL "\\LIST0"; p = l1_simple_pattern; sep = OPT sep -> List0 (p, sep)
301 | SYMBOL "\\LIST1"; p = l1_simple_pattern; sep = OPT sep -> List1 (p, sep)
302 | SYMBOL "\\OPT"; p = l1_simple_pattern -> Opt p
305 l1_pattern_variable: [
306 [ SYMBOL "\\TERM"; id = IDENT -> TermVar id
307 | SYMBOL "\\NUM"; id = IDENT -> NumVar id
308 | SYMBOL "\\IDENT"; id = IDENT -> IdentVar id
313 [ p1 = SELF; SYMBOL "\\SUB"; p2 = SELF ->
314 return_term loc (Layout (Sub (p1, p2)))
315 | p1 = SELF; SYMBOL "\\SUP"; p2 = SELF ->
316 return_term loc (Layout (Sup (p1, p2)))
317 | p1 = SELF; SYMBOL "\\BELOW"; p2 = SELF ->
318 return_term loc (Layout (Below (p1, p2)))
319 | p1 = SELF; SYMBOL "\\ABOVE"; p2 = SELF ->
320 return_term loc (Layout (Above (p1, p2)))
321 | p1 = SELF; SYMBOL "\\OVER"; p2 = SELF ->
322 return_term loc (Layout (Over (p1, p2)))
323 | p1 = SELF; SYMBOL "\\ATOP"; p2 = SELF ->
324 return_term loc (Layout (Atop (p1, p2)))
325 (* | SYMBOL "\\ARRAY"; p = SELF; csep = OPT field_sep; rsep = OPT row_sep ->
326 return_term loc (Array (p, csep, rsep)) *)
327 | SYMBOL "\\FRAC"; p1 = SELF; p2 = SELF ->
328 return_term loc (Layout (Frac (p1, p2)))
329 | SYMBOL "\\SQRT"; p = SELF -> return_term loc (Layout (Sqrt p))
330 | SYMBOL "\\ROOT"; index = SELF; SYMBOL "\\OF"; arg = SELF ->
331 return_term loc (Layout (Root (arg, index)));
332 | SYMBOL "\\HBOX"; DELIM "\\["; p = l1_pattern; DELIM "\\]" ->
333 return_term loc (Layout (Box ((H, false, false), p)))
334 | SYMBOL "\\VBOX"; DELIM "\\["; p = l1_pattern; DELIM "\\]" ->
335 return_term loc (Layout (Box ((V, false, false), p)))
336 | SYMBOL "\\HVBOX"; DELIM "\\["; p = l1_pattern; DELIM "\\]" ->
337 return_term loc (Layout (Box ((HV, false, false), p)))
338 | SYMBOL "\\HOVBOX"; DELIM "\\["; p = l1_pattern; DELIM "\\]" ->
339 return_term loc (Layout (Box ((HOV, false, false), p)))
340 (* | SYMBOL "\\BREAK" -> return_term loc (Layout Break) *)
341 (* | SYMBOL "\\SPACE" -> return_term loc (Layout Space) *)
342 | DELIM "\\["; p = l1_pattern; DELIM "\\]" ->
343 return_term loc (CicNotationUtil.boxify p)
344 | p = SELF; SYMBOL "\\AS"; id = IDENT ->
345 return_term loc (Variable (Ascription (p, id)))
348 [ i = IDENT -> return_term loc (Ident (i, None))
349 | m = l1_magic_pattern -> return_term loc (Magic m)
350 | v = l1_pattern_variable -> return_term loc (Variable v)
351 | l = literal -> return_term loc (Literal l)
355 (* {{{ Grammar for ast patterns, notation level 2 *)
356 level2_pattern: [ [ p = l2_pattern -> p ] ];
358 [ SYMBOL "\\PROP" -> `Prop
359 | SYMBOL "\\SET" -> `Set
360 | SYMBOL "\\TYPE" -> `Type
364 [ SYMBOL "\\subst"; (* to avoid catching frequent "a [1]" cases *)
367 i = IDENT; SYMBOL <:unicode<Assign>> (* ≔ *); t = l2_pattern -> (i, t)
374 [ s = SYMBOL "_" -> None
375 | p = l2_pattern -> Some p ]
378 [ SYMBOL "["; substs = LIST0 meta_subst; SYMBOL "]" -> substs ]
380 possibly_typed_name: [
381 [ SYMBOL "("; id = bound_name; SYMBOL ":"; typ = l2_pattern; SYMBOL ")" ->
383 | id = bound_name -> id, None
387 [ id = IDENT -> id, []
388 | SYMBOL "("; id = IDENT; vars = LIST1 possibly_typed_name; SYMBOL ")" ->
393 [ SYMBOL <:unicode<Pi>> (* Π *) -> `Pi
394 | SYMBOL <:unicode<exists>> (* ∃ *) -> `Exists
395 | SYMBOL <:unicode<forall>> (* ∀ *) -> `Forall
396 | SYMBOL <:unicode<lambda>> (* λ *) -> `Lambda
400 [ i = IDENT -> Ident (i, None)
401 | SYMBOL "\\FRESH"; i = IDENT -> Variable (FreshVar i)
405 [ vars = LIST1 bound_name SEP SYMBOL ",";
406 ty = OPT [ SYMBOL ":"; p = l2_pattern -> p ] ->
410 vars = LIST1 bound_name SEP SYMBOL ",";
411 ty = OPT [ SYMBOL ":"; p = l2_pattern -> p ];
419 [ IDENT "rec" -> `Inductive
420 | IDENT "corec" -> `CoInductive
425 name = bound_name; args = bound_names;
426 index_name = OPT [ IDENT "on"; id = bound_name -> id ];
427 ty = OPT [ SYMBOL ":" ; p = l2_pattern -> p ];
428 SYMBOL <:unicode<def>> (* ≝ *); body = l2_pattern ->
429 let body = fold_binder `Lambda args body in
433 | Some ty -> Some (fold_binder `Pi args ty)
435 let rec position_of name p = function
437 | n :: _ when n = name -> Some p, p
438 | _ :: tl -> position_of name (p + 1) tl
440 let rec find_arg name n = function
442 fail loc (sprintf "Argument %s not found"
443 (CicNotationPp.pp_term name))
445 (match position_of name 0 l with
446 | None, len -> find_arg name (n + len) tl
447 | Some where, len -> n + where)
450 match index_name with
452 | Some name -> find_arg name 0 args
454 (name, ty), body, index
459 l2_pattern_variable: [
460 [ SYMBOL "\\TERM"; id = IDENT -> TermVar id
461 | SYMBOL "\\NUM"; id = IDENT -> NumVar id
462 | SYMBOL "\\IDENT"; id = IDENT -> IdentVar id
463 | SYMBOL "\\FRESH"; id = IDENT -> FreshVar id
468 kind = [ IDENT "left" -> `Left | IDENT "right" -> `Right ];
469 DELIM "\\["; base = l2_pattern; DELIM "\\]";
470 SYMBOL "\\LAMBDA"; id = IDENT;
471 DELIM "\\["; recursive = l2_pattern; DELIM "\\]" ->
472 Fold (kind, base, [id], recursive)
473 | SYMBOL "\\DEFAULT";
474 DELIM "\\["; some = l2_pattern; DELIM "\\]";
475 DELIM "\\["; none = l2_pattern; DELIM "\\]" ->
479 l2_pattern: LEVEL "10" (* let in *)
481 [ IDENT "let"; var = possibly_typed_name; SYMBOL <:unicode<def>> (* ≝ *);
482 p1 = l2_pattern; "in"; p2 = l2_pattern ->
483 return_term loc (LetIn (var, p1, p2))
484 | IDENT "let"; k = induction_kind; defs = let_defs; IDENT "in";
486 return_term loc (LetRec (k, defs, body))
489 l2_pattern: LEVEL "20" (* binder *)
491 [ b = binder; names = bound_names; SYMBOL "."; body = l2_pattern ->
492 return_term loc (fold_binder b names body)
495 l2_pattern: LEVEL "70" (* apply *)
497 [ p1 = l2_pattern; p2 = l2_pattern ->
498 let rec aux = function
500 | AttributedTerm (_, Appl (hd :: tl)) ->
504 return_term loc (Appl (aux p1 @ [p2]))
507 l2_pattern: LEVEL "90" (* simple *)
509 [ id = IDENT -> return_term loc (Ident (id, None))
510 | id = IDENT; s = explicit_subst -> return_term loc (Ident (id, Some s))
511 | s = CSYMBOL -> return_term loc (Symbol (s, 0))
512 | u = URI -> return_term loc (Uri (u, None))
513 | n = 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 (0, id)
540 | l = LIST1 [ SYMBOL <:unicode<eta>> (* η *) -> () ] SEP SYMBOL ".";
541 SYMBOL "."; id = IDENT ->
542 IdentArg (List.length l, id)
546 [ u = URI -> UriPattern (UriManager.uri_of_string u)
547 | id = IDENT -> VarPattern id
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 ]
567 [ p1 = level1_pattern;
568 assoc = OPT associativity; prec = OPT precedence;
569 IDENT "for"; p2 = level2_pattern ->
570 (p1, assoc, prec, p2)
574 [ s = CSYMBOL; args = LIST1 argument; SYMBOL "="; t = level3_term ->
579 (* {{{ Top-level phrases *)
581 [ IDENT "print"; p2 = level2_pattern; SYMBOL "." -> Print p2
582 | IDENT "notation"; (l1, assoc, prec, l2) = notation; SYMBOL "." ->
583 Notation (l1, assoc, prec, l2)
584 | IDENT "interpretation"; (symbol, args, l3) = interpretation; SYMBOL "." ->
585 Interpretation ((symbol, args), l3)
586 | IDENT "render"; u = URI; SYMBOL "." -> Render (UriManager.uri_of_string u)
592 (** {2 API implementation} *)
594 let exc_located_wrapper f =
598 | Stdpp.Exc_located (floc, Stream.Error msg) ->
599 raise (Parse_error (floc, msg))
600 | Stdpp.Exc_located (floc, exn) ->
601 raise (Parse_error (floc, (Printexc.to_string exn)))
603 let parse_syntax_pattern stream =
604 exc_located_wrapper (fun () -> Grammar.Entry.parse level1_pattern stream)
605 let parse_ast_pattern stream =
606 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_pattern stream)
607 let parse_interpretation stream =
608 exc_located_wrapper (fun () -> Grammar.Entry.parse level3_term stream)
609 let parse_phrase stream =
610 exc_located_wrapper (fun () -> Grammar.Entry.parse phrase stream)
614 let print_l2_pattern () =
615 Grammar.print_entry Format.std_formatter (Grammar.Entry.obj l2_pattern);
616 Format.pp_print_flush Format.std_formatter ();
619 (* vim:set encoding=utf8 foldmethod=marker: *)