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 | Group pl -> List.flatten (List.map aux pl)
170 and aux_magic magic =
173 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
174 let action (env_opt : CicNotationEnv.t option) (loc : location) =
176 | Some env -> List.map opt_binding_some env
177 | None -> List.map opt_binding_of_name p_names
179 [ Env (List.map opt_declaration p_names),
181 [ [ Gramext.Sopt (Gramext.srules [ p_atoms, p_action ]) ],
182 Gramext.action action ] ]
185 let p_bindings, p_atoms, p_names, p_action = inner_pattern p in
186 (* let env0 = List.map list_binding_of_name p_names in
187 let grow_env_entry env n v =
190 | (n', (ty, ListValue vl)) as entry ->
191 if n' = n then n', (ty, ListValue (v :: vl)) else entry
195 let grow_env env_i env =
197 (fun env (n, (_, v)) -> grow_env_entry env n v)
200 let action (env_list : CicNotationEnv.t list) (loc : location) =
201 CicNotationEnv.coalesce_env p_names env_list
205 | List0 (_, None) -> Gramext.Slist0 s
206 | List1 (_, None) -> Gramext.Slist1 s
207 | List0 (_, Some l) -> Gramext.Slist0sep (s, gram_of_literal l)
208 | List1 (_, Some l) -> Gramext.Slist1sep (s, gram_of_literal l)
211 [ Env (List.map list_declaration p_names),
213 [ [ gram_of_list (Gramext.srules [ p_atoms, p_action ]) ],
214 Gramext.action action ] ]
218 | NumVar s -> [Binding (s, NumType), gram_number ""]
219 | TermVar s -> [Binding (s, TermType), gram_term]
220 | IdentVar s -> [Binding (s, StringType), gram_ident ""]
221 | Ascription (p, s) -> assert false (* TODO *)
222 | FreshVar _ -> assert false
223 and inner_pattern p =
224 let p_bindings, p_atoms = List.split (aux p) in
225 let p_names = flatten_opt p_bindings in
227 make_action (fun (env : CicNotationEnv.t) (loc : location) -> env)
230 p_bindings, p_atoms, p_names, action
234 let level_of_int precedence =
235 if precedence < min_precedence || precedence > max_precedence then
236 raise (Level_not_found precedence);
237 string_of_int precedence
239 type rule_id = Token.t Gramext.g_symbol list
241 (* mapping: rule_id -> owned keywords. (rule_id, string list) Hashtbl.t *)
242 let owned_keywords = Hashtbl.create 23
244 let extend level1_pattern ?(precedence = default_precedence)
245 ?associativity action
247 let p_bindings, p_atoms =
248 List.split (extract_term_production level1_pattern)
250 let level = level_of_int precedence in
251 let p_names = flatten_opt p_bindings in
254 [ Grammar.Entry.obj (term: 'a Grammar.Entry.e),
255 Some (Gramext.Level level),
260 (fun (env: CicNotationEnv.t) (loc: location) -> (action env loc))
263 let keywords = CicNotationUtil.keywords_of_term level1_pattern in
264 let rule_id = p_atoms in
265 List.iter CicNotationLexer.add_level2_ast_keyword keywords;
266 Hashtbl.add owned_keywords rule_id keywords; (* keywords may be [] *)
270 let atoms = rule_id in
272 let keywords = Hashtbl.find owned_keywords rule_id in
273 List.iter CicNotationLexer.remove_level2_ast_keyword keywords
274 with Not_found -> assert false);
275 Grammar.delete_rule term atoms
279 let parse_level1_pattern_ref = ref (fun _ -> assert false)
280 let parse_level2_ast_ref = ref (fun _ -> assert false)
281 let parse_level2_meta_ref = ref (fun _ -> assert false)
283 let fold_binder binder pt_names body =
284 let fold_cluster binder terms ty body =
286 (fun term body -> Binder (binder, (term, ty), body))
287 terms body (* terms are names: either Ident or FreshVar *)
290 (fun (names, ty) body -> fold_cluster binder names ty body)
293 let return_term loc term = AttributedTerm (`Loc loc, term)
295 let _ = (* create empty precedence level for "term" *)
296 let mk_level_list first last =
297 let rec aux acc = function
298 | i when i < first -> acc
299 | i -> aux ((Some (string_of_int i), None, []) :: acc) (i - 1)
304 [ Grammar.Entry.obj (term: 'a Grammar.Entry.e),
306 mk_level_list min_precedence max_precedence ]
308 (* {{{ Grammar for concrete syntax patterns, notation level 1 *)
310 GLOBAL: level1_pattern;
312 level1_pattern: [ [ p = l1_pattern; EOI -> CicNotationUtil.boxify p ] ];
313 l1_pattern: [ [ p = LIST1 l1_simple_pattern -> p ] ];
315 [ s = SYMBOL -> `Symbol s
316 | k = QKEYWORD -> `Keyword k
317 | n = NUMBER -> `Number n
320 sep: [ [ "sep"; sep = literal -> sep ] ];
321 (* row_sep: [ [ "rowsep"; sep = literal -> sep ] ];
322 field_sep: [ [ "fieldsep"; sep = literal -> sep ] ]; *)
324 [ "list0"; p = l1_simple_pattern; sep = OPT sep -> List0 (p, sep)
325 | "list1"; p = l1_simple_pattern; sep = OPT sep -> List1 (p, sep)
326 | "opt"; p = l1_simple_pattern -> Opt p
329 l1_pattern_variable: [
330 [ "term"; id = IDENT -> TermVar id
331 | "number"; id = IDENT -> NumVar id
332 | "ident"; id = IDENT -> IdentVar id
337 [ p1 = SELF; SYMBOL "\\sub"; p2 = SELF ->
338 return_term loc (Layout (Sub (p1, p2)))
339 | p1 = SELF; SYMBOL "\\sup"; p2 = SELF ->
340 return_term loc (Layout (Sup (p1, p2)))
341 | p1 = SELF; SYMBOL "\\below"; p2 = SELF ->
342 return_term loc (Layout (Below (p1, p2)))
343 | p1 = SELF; SYMBOL "\\above"; p2 = SELF ->
344 return_term loc (Layout (Above (p1, p2)))
345 | p1 = SELF; SYMBOL "\\over"; p2 = SELF ->
346 return_term loc (Layout (Over (p1, p2)))
347 | p1 = SELF; SYMBOL "\\atop"; p2 = SELF ->
348 return_term loc (Layout (Atop (p1, p2)))
349 (* | "array"; p = SELF; csep = OPT field_sep; rsep = OPT row_sep ->
350 return_term loc (Array (p, csep, rsep)) *)
351 | SYMBOL "\\frac"; p1 = SELF; p2 = SELF ->
352 return_term loc (Layout (Frac (p1, p2)))
353 | SYMBOL "\\sqrt"; p = SELF -> return_term loc (Layout (Sqrt p))
354 | SYMBOL "\\root"; index = SELF; SYMBOL "\\OF"; arg = SELF ->
355 return_term loc (Layout (Root (arg, index)))
356 | "hbox"; LPAREN; p = l1_pattern; RPAREN ->
357 return_term loc (Layout (Box ((H, false, false), p)))
358 | "vbox"; LPAREN; p = l1_pattern; RPAREN ->
359 return_term loc (Layout (Box ((V, false, false), p)))
360 | "hvbox"; LPAREN; p = l1_pattern; RPAREN ->
361 return_term loc (Layout (Box ((HV, false, false), p)))
362 | "hovbox"; LPAREN; p = l1_pattern; RPAREN ->
363 return_term loc (Layout (Box ((HOV, false, false), p)))
364 | "break" -> return_term loc (Layout Break)
365 (* | SYMBOL "\\SPACE" -> return_term loc (Layout Space) *)
366 | LPAREN; p = l1_pattern; RPAREN ->
367 return_term loc (CicNotationUtil.group p)
370 [ i = IDENT -> return_term loc (Variable (TermVar i))
371 | m = l1_magic_pattern -> return_term loc (Magic m)
372 | v = l1_pattern_variable -> return_term loc (Variable v)
373 | l = literal -> return_term loc (Literal l)
383 [ "term"; id = IDENT -> TermVar id
384 | "number"; id = IDENT -> NumVar id
385 | "ident"; id = IDENT -> IdentVar id
386 | "fresh"; id = IDENT -> FreshVar id
387 | "anonymous" -> TermVar "_"
388 | id = IDENT -> TermVar id
392 [ "fold"; kind = [ "left" -> `Left | "right" -> `Right ];
393 base = level2_meta; "rec"; id = IDENT; recursive = level2_meta ->
394 Fold (kind, base, [id], recursive)
395 | "default"; some = level2_meta; none = level2_meta -> Default (some, none)
396 | "if"; p_test = level2_meta;
397 "then"; p_true = level2_meta;
398 "else"; p_false = level2_meta ->
399 If (p_test, p_true, p_false)
404 [ magic = l2_magic -> Magic magic
405 | var = l2_variable -> Variable var
406 | blob = UNPARSED_AST -> !parse_level2_ast_ref (Stream.of_string blob)
412 GLOBAL: level2_ast term
414 notation interpretation
416 (* {{{ Grammar for ast patterns, notation level 2 *)
417 level2_ast: [ [ p = term -> p ] ];
425 [ SYMBOL "\\subst"; (* to avoid catching frequent "a [1]" cases *)
428 i = IDENT; SYMBOL <:unicode<Assign>> (* ≔ *); t = term -> (i, t)
435 [ s = SYMBOL "_" -> None
436 | p = term -> Some p ]
439 [ SYMBOL "["; substs = LIST0 meta_subst; SYMBOL "]" -> substs ]
441 possibly_typed_name: [
442 [ LPAREN; id = bound_name; SYMBOL ":"; typ = term; RPAREN ->
444 | id = bound_name -> id, None
448 [ id = IDENT -> id, []
449 | LPAREN; id = IDENT; vars = LIST1 possibly_typed_name; RPAREN ->
454 [ SYMBOL <:unicode<Pi>> (* Π *) -> `Pi
455 | SYMBOL <:unicode<exists>> (* ∃ *) -> `Exists
456 | SYMBOL <:unicode<forall>> (* ∀ *) -> `Forall
457 | SYMBOL <:unicode<lambda>> (* λ *) -> `Lambda
461 [ i = IDENT -> Ident (i, None)
462 | SYMBOL "\\FRESH"; i = IDENT -> Variable (FreshVar i)
466 [ vars = LIST1 bound_name SEP SYMBOL ",";
467 ty = OPT [ SYMBOL ":"; p = term -> p ] ->
471 vars = LIST1 bound_name SEP SYMBOL ",";
472 ty = OPT [ SYMBOL ":"; p = term -> p ];
480 [ "rec" -> `Inductive
481 | "corec" -> `CoInductive
486 name = bound_name; args = bound_names;
487 index_name = OPT [ "on"; id = bound_name -> id ];
488 ty = OPT [ SYMBOL ":" ; p = term -> p ];
489 SYMBOL <:unicode<def>> (* ≝ *); body = term ->
490 let body = fold_binder `Lambda args body in
494 | Some ty -> Some (fold_binder `Pi args ty)
496 let rec position_of name p = function
498 | n :: _ when n = name -> Some p, p
499 | _ :: tl -> position_of name (p + 1) tl
501 let rec find_arg name n = function
503 fail loc (sprintf "Argument %s not found"
504 (CicNotationPp.pp_term name))
506 (match position_of name 0 l with
507 | None, len -> find_arg name (n + len) tl
508 | Some where, len -> n + where)
511 match index_name with
513 | Some name -> find_arg name 0 args
515 (name, ty), body, index
520 term: LEVEL "10" (* let in *)
522 [ "let"; var = possibly_typed_name; SYMBOL <:unicode<def>> (* ≝ *);
523 p1 = term; "in"; p2 = term ->
524 return_term loc (LetIn (var, p1, p2))
525 | "let"; k = induction_kind; defs = let_defs; "in";
527 return_term loc (LetRec (k, defs, body))
530 term: LEVEL "20" (* binder *)
532 [ b = binder; names = bound_names; SYMBOL "."; body = term ->
533 return_term loc (fold_binder b names body)
536 term: LEVEL "70" (* apply *)
538 [ p1 = term; p2 = term ->
539 let rec aux = function
541 | AttributedTerm (_, Appl (hd :: tl)) ->
545 return_term loc (Appl (aux p1 @ [p2]))
548 term: LEVEL "90" (* simple *)
550 [ id = IDENT -> return_term loc (Ident (id, None))
551 | id = IDENT; s = explicit_subst -> return_term loc (Ident (id, Some s))
552 | s = CSYMBOL -> return_term loc (Symbol (s, 0))
553 | u = URI -> return_term loc (Uri (u, None))
554 | n = NUMBER -> return_term loc (Num (n, 0))
555 | IMPLICIT -> return_term loc (Implicit)
556 | m = META -> return_term loc (Meta (int_of_string m, []))
557 | m = META; s = meta_substs -> return_term loc (Meta (int_of_string m, s))
558 | s = sort -> return_term loc (Sort s)
559 | outtyp = OPT [ SYMBOL "["; ty = term; SYMBOL "]" -> ty ];
561 indty_ident = OPT [ SYMBOL ":"; id = IDENT -> id ];
564 lhs = match_pattern; SYMBOL <:unicode<Rightarrow>> (* ⇒ *);
569 return_term loc (Case (t, indty_ident, outtyp, patterns))
570 | LPAREN; p1 = term; SYMBOL ":"; p2 = term; RPAREN ->
571 return_term loc (Appl [ Symbol ("cast", 0); p1; p2 ])
572 | LPAREN; p = term; RPAREN -> p
573 | blob = UNPARSED_META -> !parse_level2_meta_ref (Stream.of_string blob)
577 (* {{{ Grammar for interpretation, notation level 3 *)
579 [ id = IDENT -> IdentArg (0, id)
580 | l = LIST1 [ SYMBOL <:unicode<eta>> (* η *) -> () ] SEP SYMBOL ".";
581 SYMBOL "."; id = IDENT ->
582 IdentArg (List.length l, id)
586 [ u = URI -> UriPattern (UriManager.uri_of_string u)
587 | id = IDENT -> VarPattern id
588 | LPAREN; terms = LIST1 SELF; RPAREN ->
592 | terms -> ApplPattern terms)
596 (* {{{ Notation glues *)
598 [ IDENT "left"; IDENT "associative" -> Gramext.LeftA
599 | IDENT "right"; IDENT "associative" -> Gramext.RightA
600 | IDENT "non"; IDENT "associative" -> Gramext.NonA
604 [ "with"; IDENT "precedence"; n = NUMBER -> int_of_string n ]
608 assoc = OPT associativity; prec = OPT precedence;
611 [ blob = UNPARSED_AST -> !parse_level2_ast_ref (Stream.of_string blob)
612 | blob = UNPARSED_META ->
613 !parse_level2_meta_ref (Stream.of_string blob) ]
615 (!parse_level1_pattern_ref (Stream.of_string s), assoc, prec, p2)
619 [ s = CSYMBOL; args = LIST1 argument; SYMBOL "="; t = level3_term ->
624 (* {{{ Top-level phrases *)
626 [ IDENT "print"; t = term; SYMBOL "." -> Print t
627 | IDENT "notation"; (l1, assoc, prec, l2) = notation; SYMBOL "." ->
628 Notation (l1, assoc, prec, l2)
629 | IDENT "interpretation"; (symbol, args, l3) = interpretation; SYMBOL "." ->
630 Interpretation ((symbol, args), l3)
631 | IDENT "render"; u = URI; SYMBOL "." -> Render (UriManager.uri_of_string u)
632 | IDENT "dump"; SYMBOL "." -> Dump
638 (** {2 API implementation} *)
640 let exc_located_wrapper f =
644 | Stdpp.Exc_located (floc, Stream.Error msg) ->
645 raise (Parse_error (floc, msg))
646 | Stdpp.Exc_located (floc, exn) ->
647 raise (Parse_error (floc, (Printexc.to_string exn)))
649 let parse_level1_pattern stream =
650 exc_located_wrapper (fun () -> Grammar.Entry.parse level1_pattern stream)
651 let parse_level2_ast stream =
652 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_ast stream)
653 let parse_level2_meta stream =
654 exc_located_wrapper (fun () -> Grammar.Entry.parse level2_meta stream)
655 let parse_interpretation stream =
656 exc_located_wrapper (fun () -> Grammar.Entry.parse level3_term stream)
657 let parse_phrase stream =
658 exc_located_wrapper (fun () -> Grammar.Entry.parse phrase stream)
661 parse_level1_pattern_ref := parse_level1_pattern;
662 parse_level2_ast_ref := parse_level2_ast;
663 parse_level2_meta_ref := parse_level2_meta
667 let print_l2_pattern () =
668 Grammar.print_entry Format.std_formatter (Grammar.Entry.obj term);
669 Format.pp_print_flush Format.std_formatter ();
672 (* vim:set encoding=utf8 foldmethod=marker: *)