parser.ml 18.8 KB
Newer Older
1 2
#load "pa_extend.cmo";;

3 4
open Location
open Ast
5
open Ident
6
open Printf
7

8
(*
9
let ()  = Grammar.error_verbose := true
10
*)
11

12 13
exception Error of string
let error (i,j) s = Location.raise_loc i j (Error s)
14

15
let gram    = Grammar.gcreate Ulexer.lex
16

17
let parse_ident = U.mk
18

19
let id_dummy = ident (U.mk "$$$")
20

21
let label = parse_ident
22
let ident s = ident (parse_ident s)
23

24
let prog    = Grammar.Entry.create gram "prog"
25
let top_phrases   = Grammar.Entry.create gram "toplevel phrases"
26 27 28
let expr    = Grammar.Entry.create gram "expression"
let pat     = Grammar.Entry.create gram "type/pattern expression"
let regexp  = Grammar.Entry.create gram "type/pattern regexp"
29
let keyword = Grammar.Entry.create gram "keyword"
30
		
31 32
let exp pos e = LocatedExpr (loc_of_pos pos,e)

33 34 35 36 37
let rec multi_prod loc = function
  | [ x ] -> x
  | x :: l -> mk loc (Prod (x, multi_prod loc l))
  | [] -> assert false
      
38
let rec tuple = function
39
  | [ x ] -> x
40
  | x :: l -> Pair (x, tuple l)
41
  | [] -> assert false
42 43

let tuple_queue = 
44
  List.fold_right (fun x q -> Pair (x, q))
45 46


47
let char = mknoloc (Internal (Types.char Chars.any))
48
let string_regexp = Star (Elem char)
49
	       
50
let seq_of_string s =
51 52
  let s = Encodings.Utf8.mk s in
  let rec aux i j =
53 54
    if Encodings.Utf8.equal_index i j then []
    else let (c,i) = Encodings.Utf8.next s i in c :: (aux i j)
55
  in
56
  aux (Encodings.Utf8.start_index s) (Encodings.Utf8.end_index s)
57

58

59
let parse_char loc s =
60 61
  match seq_of_string s with
    | [ c ] -> c
62
    | _ -> error loc "Character litteral must have length 1"
63

64 65
let include_stack = ref []

66 67 68 69
let protect_exn f g =
  try let x = f () in g (); x
  with e -> g (); raise e

70 71 72 73 74 75 76
let localize_exn f = 
  try f ()
  with
  | Stdpp.Exc_located (_, (Location _ as e)) -> raise e
  | Stdpp.Exc_located ((i,j), e) -> raise_loc i j e
  

77 78 79 80
let is_fun_decl =
   Grammar.Entry.of_parser gram "[is_fun_decl]"
     (fun strm ->
       match Stream.npeek 3 strm with
81 82
	 | [ ("", "fun"); ("IDENT", _); ("", "(") ]
	 | [ ("IDENT", _) ; ("", "(") ; _ ] -> ()
83 84 85
	 | _ -> raise Stream.Failure
     )

86 87 88 89 90 91
let if_then_else cond e1 e2 = Match (cond, [pat_true,e1; pat_false,e2])
 
let logical_and e1 e2 = if_then_else e1 e2 cst_false
let logical_or e1 e2 = if_then_else e1 cst_true e2
let logical_not e = if_then_else e cst_false cst_true
 
92

93
EXTEND
94
  GLOBAL: top_phrases prog expr pat regexp keyword;
95 96

  top_phrases: [
97
    [ l = LIST0 phrase; ";;" -> List.flatten l  ]
98
  ];
99 100

  prog: [
101
    [ l = LIST0 [ p = phrase ; OPT ";;" -> p ]; EOI -> List.flatten l ]
102 103 104
  ];

  phrase: [
105 106 107 108
    [ (f,p,e) = let_binding -> 
	if f then [ mk loc (FunDecl e) ] else
	  [ mk loc (LetDecl (p,e)) ]
    | (_,p,e1) = let_binding; "in"; e2 = expr LEVEL "top"->
109
        [ mk loc (EvalStatement (exp loc (Match (e1,[p,e2])))) ]
110
    | "type"; x = IDENT; "="; t = pat -> [ mk loc (TypeDecl (ident x,t)) ]
111 112
    | "using"; name = IDENT; "="; cu = [ IDENT | STRING2 ] ->
	[ mk loc (Using (U.mk name, U.mk cu)) ]
113
    | "schema"; name = IDENT; "="; uri = STRING2 ->
114
	protect_op "schema";
115
        [ mk loc (SchemaDecl (U.mk name, uri)) ]
116 117 118 119 120
    | (name,ns) = namespace_binding ->
	[ mk loc (Namespace (name, ns)) ]
    | (name,ns) = namespace_binding; "in"; e2 = expr LEVEL "top" ->
	let e = exp loc (NamespaceIn (name, ns, e2)) in
        [ mk loc (EvalStatement (exp loc e)) ]
121
    | "debug"; d = debug_directive -> [ mk loc (Directive (`Debug d)) ]
122 123 124
    | DIRECTIVE "#utf8" -> Ulexer.enc := Ulexing.Utf8; [ ]
    | DIRECTIVE "#latin1" -> Ulexer.enc := Ulexing.Latin1; [ ]
    | DIRECTIVE "#ascii" -> Ulexer.enc := Ulexing.Ascii; [ ]
125 126
    | DIRECTIVE "#quit" -> [ mk loc (Directive `Quit) ]
    | DIRECTIVE "#env" -> [ mk loc (Directive `Env) ]
127
    | DIRECTIVE "#print_schema"; name = IDENT ->
128
        [ mk loc (Directive (`Print_schema (U.mk name))) ]
129 130 131 132 133 134 135
    | DIRECTIVE "#print_type"; name = IDENT;
      schema_part = OPT [
        "#"; typ = [ IDENT | keyword ];
        kind = OPT [ "as"; k = schema_kind -> k] ->
          (kind, typ)
      ] ->
        (match schema_part with
136
        | None -> [ mk loc (Directive (`Print_type (U.mk name))) ]
137
        | Some (kind, typ) ->
138 139
            [ mk loc
              (Directive (`Print_schema_type (kind, U.mk name, U.mk typ))) ])
140
    | DIRECTIVE "#dump_value"; e = expr -> [ mk loc (Directive (`Dump e)) ]
141
    | DIRECTIVE "#reinit_ns" -> [ mk loc (Directive `Reinit_ns) ]
142
    | DIRECTIVE "#help" -> [ mk loc (Directive `Help) ]
143
    | "include"; s = STRING2 ->
144 145 146 147
	let s = 
	  if Filename.is_relative s 
	  then Filename.concat (Location.current_dir ()) s
	  else s in
148 149
	protect_op "File inclusion";
	(* avoid looping; should issue an error ? *)
150 151
	(* it is possible to have looping with x/../x/../x/.. ....
	   Need to canonicalize filename *)
152 153 154 155
	if List.mem s !include_stack then [] 
	else (
	  include_stack := s :: !include_stack;
	  Location.push_source (`File s);
156 157
	  let saved_enc = !Ulexer.enc in
	  Ulexer.enc := Ulexing.Latin1;
158 159 160
	  protect_exn
	    (fun () ->
	       let chan = open_in s in
161 162 163
	       protect_exn
		 (fun () ->
		    let input = Stream.of_channel chan in
164
		    localize_exn (fun () -> Grammar.Entry.parse prog input))
165
		 (fun () -> close_in chan))
166
	    (fun () ->
167
	       Ulexer.enc := saved_enc;
168 169
	       Location.pop_source ();
	       include_stack := List.tl !include_stack)
170
	)
171
    ] | 
172
    [ e = expr -> [ mk loc (EvalStatement e) ]
173 174 175 176
    ]
  ];

  debug_directive: [
177 178 179 180 181
    [ IDENT "filter"; t = pat; p = pat -> `Filter(t,p)
    | IDENT "accept"; p = pat -> `Accept p
    | IDENT "compile"; t = pat; p = LIST1 pat -> `Compile (t,p)
    | IDENT "sample"; t = pat -> `Sample t
    | IDENT "subtype"; t1 = pat; t2 = pat -> `Subtype (t1,t2)
182
    | IDENT "explain"; t = pat; e = expr -> `Explain (t,e)
183
    ]
184
  ];
185

186 187 188 189 190
  keyword: [
    [ a = 
	[ "map" | "match" | "with" | "try" | "xtransform"
	| "if" | "then"	| "else"
	| "transform" | "fun" | "in"
191
	| "let" | "type" | "debug" | "include"
192
        | "and" | "validate" | "schema" | "namespace" | "ref" | "alias"
193
	| "not" | "as" | "where" | "external"
194 195 196 197 198
	]
	-> a
    ]
  ];

199 200
  expr: [
    "top" RIGHTA
201 202
    [ "match"; e = SELF; "with"; b = branches -> 
	exp loc (Match (e,b))
203
    | "try"; e = SELF; "with"; b = branches -> 
204
	exp loc (Try (e,b))
205 206 207 208
    | "map"; e = SELF; "with"; b = branches -> 
	exp loc (Map (e,b))
    | "xtransform"; e = SELF; "with"; b = branches -> 
	exp loc (Xtrans (e,b))
209
    | "if"; e = SELF; "then"; e1 = SELF; "else"; e2 = SELF ->
210
	exp loc (if_then_else e e1 e2)
211
    | "transform"; e = SELF; "with"; b = branches -> 
212
	exp loc (Transform (e,b))
213 214
    | "validate"; e = SELF; "with"; (kind, schema, typ) = schema_ref ->
        exp loc (Validate (e, kind, schema, typ))
215
    | "fun"; (f,a,b) = fun_decl ->
216
	exp loc (Abstraction { fun_name = f; fun_iface = a; fun_body = b })
217
    | "external"; s = STRING2 ->
218
	exp loc (External (s,[]))
219
    | "external"; "{"; s = STRING2; pl = LIST0 pat; "}" ->
220
	exp loc (External (s,pl))
221
    | (_,p,e1) = let_binding; "in"; e2 = expr LEVEL "top"->
222
        exp loc (Match (e1,[p,e2]))
223 224
    | (name,ns) = namespace_binding; "in"; e2 = expr LEVEL "top" ->
	exp loc (NamespaceIn (name, ns, e2))
225
    | e = expr; ":"; p = pat ->
226
	exp loc (Forget (e,p))
227 228 229 230
    | e1 = expr; ";"; e2 = expr ->
	exp loc (Match (e1, [pat_nil,e2]))
    | "ref"; p = pat; e = expr ->
	exp loc (Ref (e,p))
231
    | "not"; e = expr -> exp loc (logical_not e)
232 233 234 235
    ]
    |
    [ e1 = expr; ":="; e2 = expr ->
	exp loc (Apply (Dot (e1, U.mk "set"), e2))
236
    ]
237 238 239 240 241 242
    | 
    [ e1 = expr; op = ["=" | "<=" | "<<" | ">>" | ">=" ]; e2 = expr -> 
	let op = match op with
	  | "<<" -> "<"
	  | ">>" -> ">"
	  | s -> s in
243
	exp loc (Op (op,[e1;e2]))
244 245
    ]

246
    | 
247
    [ e1 = expr; op = ["+" | "-" | "@" ]; e2 = expr -> 
248
	exp loc (Op (op,[e1;e2]))
249
    | e1 = expr; "||"; e2 = expr -> exp loc (logical_or e1 e2)
250
    | e = expr; "\\"; l = [IDENT | keyword ] -> 
251
	exp loc (RemoveField (e, label l)) 
252 253
    ]
    |
254
    [ e1 = expr; op = ["*"]; e2 = expr -> exp loc (Op (op,[e1;e2]))  
255
    | e1 = expr; "&&"; e2 = expr -> exp loc (logical_and e1 e2)
256
    | e = expr; op = "/"; p = pat LEVEL "simple" ->
257 258 259 260 261 262 263 264
	let tag = mk loc (Internal (Types.atom (Atoms.any))) in
	let att = mk loc (Internal Types.Record.any) in
	let any = mk loc (Internal (Types.any)) in
	let re = Star(Alt(SeqCapture(id_dummy,Elem p), Elem any)) in
	let ct = mk loc (Regexp (re,any)) in
        let p = mk loc (XmlT (tag, multi_prod loc [att;ct])) in
	let b = (p,  Var (Id.value id_dummy)) in
	exp loc (Transform (e,[b]))
265
    ]
266
    |
267
    [ e = expr;  "."; l = [IDENT | keyword ] -> 
268
	exp loc (Dot (e, label l)) 
269
    ]  
270
    |
271
    [ op = [ IDENT "print_xml" | IDENT "print_xml_utf8"
272
	   ]; 
273
      e = expr -> exp loc (Op (op,[e]))
274
    | op = [ IDENT "dump_to_file" | IDENT "dump_to_file_utf8" ];
275
      e1 = expr LEVEL "no_appl"; e2 = expr -> exp loc (Op (op, [e1;e2])) 
276 277
    | e1 = SELF; IDENT "div"; e2 = expr -> exp loc (Op ("/", [e1;e2]))
    | e1 = SELF; IDENT "mod"; e2 = expr -> exp loc (Op ("mod", [e1;e2]))
278
    | e1 = SELF; e2 = expr -> exp loc (Apply (e1,e2))
279 280
    ]

281
    | "no_appl" 
282 283
    [ 
      "("; l = LIST1 expr SEP ","; ")" -> exp loc (tuple l)
284
    | "[";  l = LIST0 seq_elem; e = OPT [ ";"; e = expr -> e ]; 
285
      loc_end = ["]" -> loc] ->
286
	let e = match e with Some e -> e | None -> cst_nil in
287 288
        let e = exp loc_end e in
        let (_,loc_end) = loc_end in
289 290 291
	let l = List.fold_right 
		  (fun x q ->
		     match x with
292
		       | `String (loc,i,j,s) -> exp loc (String (i,j,s,q))
293
		       | `Elems ((loc,_),x) -> exp (loc,loc_end) (Pair(x,q))
294 295 296 297
		       | `Explode x -> Op ("@",[x;q])
		  ) l e
	in
	exp loc l
298
    | "<"; t = [ "("; e = expr; ")" -> e
299
	       | a = tag -> exp loc a
300
	       ];
301
	a = expr_attrib_spec; ">"; c = expr ->
302
	  exp loc (Xml (t, Pair (a,c)))
303
    | "{"; r = [ expr_record_spec | -> exp loc (RecordLitt []) ]; "}" -> r
304
    | s = STRING2 ->
305
	let s = U.mk s in
306
	exp loc (String (U.start_index s, U.end_index s, s, cst_nil))
307
    | a = IDENT -> exp loc (Var (U.mk a))
308 309
    | "!"; e = expr ->
	exp loc (Apply (Dot (e, U.mk "get"), cst_nil))
310 311 312
    | i = INT -> exp loc (Integer (Intervals.V.mk i))
    | "`"; a = tag -> a
    | c = char -> exp loc (Char c)
313 314 315
    ]

  ];
316

317
  tag: [ [ a = [ IDENT | keyword ] -> exp loc (Atom (parse_ident a)) ] ];
318 319

  tag_type: [
320
    [ IDENT "_" ->  mk loc (Internal (Types.atom Atoms.any))
321
    | a = [ IDENT | keyword ] -> mk loc (Cst (Atom (parse_ident a)))
322
    | t = ANY_IN_NS -> mk loc (NsT (parse_ident t)) 
323 324 325
    ]
  ];

326
  seq_elem: [
327
    [ x = STRING1 -> 
328
	let s = U.mk x in
329
	`String (loc, U.start_index s, U.end_index s, s)
330
    | e = expr LEVEL "no_appl" -> `Elems (loc,e)
331
    | "!"; e = expr LEVEL "no_appl" -> `Explode e
332 333
    ]
  ];
334 335 336
	
  namespace_binding: [
    [ "namespace"; 
337
	name = [ name = [ IDENT | keyword ]; "=" ->
338 339 340
		   parse_ident name
	       | -> U.mk "" ];
	uri = STRING2 ->
341
	  let ns = Ns.mk (parse_ident uri) in
342 343 344 345 346
	  (name,ns)
    ]
  ];

  
347
  let_binding: [
348 349
    [ "let"; is_fun_decl; OPT "fun"; (f,a,b) = fun_decl ->
	let f = match f with Some x -> x | None -> assert false in
350
	let p = mk loc (PatVar (Id.value f)) in
351
	let abst = { fun_name = Some f; fun_iface = a; fun_body = b } in
352
        let e = exp loc (Abstraction abst) in
353
        (true,p,e)
354 355
    | "let"; p = pat; "="; e = expr -> (false,p,e)
    | "let"; p = pat; ":"; t = pat; "="; e = expr -> (false,p, Forget (e,t))
356 357 358
    ] 
  ];

359
 fun_decl_after_lparen: [
360
(* need an hack to do this, because both productions would
361
   match   [ OPT IDENT; "("; pat ] .... *)
362 363 364 365 366 367
   [ p1 = pat LEVEL "no_arrow";
     res = [ "->"; p2 = pat;
	     a = [ ";"; a = LIST0 arrow SEP ";" -> a | -> [] ];
	     ")"; b = branches -> `Classic (p2,a,b)
	   | ":"; targ1 = pat;
	     args = LIST0 [ ","; arg = pat; ":"; targ = pat -> (arg,targ) ]; 
368 369 370 371 372 373 374 375 376
	     ")";
	     others = LIST0 
			[ "(";
			  args = 
			    LIST1 
			      [ arg = pat; ":"; targ = pat -> (arg,targ) ]
			      SEP ",";
			  ")" -> args ];
	     ":"; tres = pat ; 
377
	     "="; body = expr ->
378
	       `Compact (targ1,args,others,tres,body)
379 380 381
	   ] ->
       match res with
	 | `Classic (p2,a,b) -> (p1,p2)::a,b
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400
	 | `Compact (targ1,args,others,tres,body) ->
	     let mkfun args =
	       multi_prod nopos (List.map snd args),
	       multi_prod nopos (List.map fst args)
	     in
	     let (tres,body) =
	       List.fold_right
		 (fun args (tres,body) ->
		    let (targ,arg) = mkfun args in
		    let e = Abstraction 
			      { fun_name = None; fun_iface = [targ,tres]; 
				fun_body = [arg,body] } in
		    let t = mknoloc (Arrow (targ,tres)) in
		    (t,e)
		 )
		 others (tres,body) in
	     let (targ,arg) = mkfun ((p1,targ1) :: args) in
	     [(targ,tres)],[(arg,body)]
	      ] ];
401 402 403


 fun_decl: [
404 405 406
   [ f = OPT IDENT; "("; (a,b) = fun_decl_after_lparen ->
       let f = match f with Some x -> Some (ident x) | None -> None in
       (f,a,b)
407
   ]
408 409
 ];

410
 arrow: [
411
    [ t1 = pat LEVEL "no_arrow"; "->"; t2 = pat -> (t1,t2)]
412 413 414
  ];

  branches: [
415
    [ OPT "|"; l = LIST1 branch SEP "|" -> l ]
416 417 418
  ];

  branch: [
419
    [ p = pat LEVEL "no_arrow"; "->"; e = expr -> (p,e) ]
420 421 422 423
  ];

	  
  regexp: [ 
424 425 426 427 428
    [ x = regexp; "|"; y = regexp -> 
	match (x,y) with
	  | Elem x, Elem y -> Elem (mk loc (Or (x,y)))
	  | _ -> Alt (x,y) 
    ]
429
  | [ x = regexp; y = regexp -> Seq (x,y) ]
430 431 432 433 434
  | [ x = regexp; "&"; y = regexp ->
	match (x,y) with
	  | Elem x, Elem y -> Elem (mk loc (And (x,y)))
	  | _ -> error loc "Conjunction not allowed in regular expression"
    ]
435
  | [ a = IDENT; "::"; x = regexp -> SeqCapture (ident a,x) ]
436 437 438 439
  | [ x = regexp; "*" -> Star x
    | x = regexp; "*?" -> WeakStar x
    | x = regexp; "+" -> Seq (x, Star x)
    | x = regexp; "+?" -> Seq (x, WeakStar x)
440
    | x = regexp; "?" ->  Alt (x, Epsilon)
441
    | x = regexp; "??" -> Alt (Epsilon, x) ]
442
  | [ "("; x = LIST1 regexp SEP ","; ")" ->
443 444 445 446 447 448 449 450 451 452 453
	(match x with
	  | [ x ] -> x
	  | _ ->
	      let x = 
		List.map 
		  (function
		     | Elem x -> x
		     | _ -> error loc 
			 "Mixing regular expressions and products")
		  x in
	      Elem (multi_prod loc x))
454 455
    | "("; a = IDENT; ":="; c = expr; ")" -> 
	Elem (mk loc (Constant ((ident a,c))))
456
    | IDENT "PCDATA" -> string_regexp
457
    | i = STRING1; "--"; j = STRING1 ->
458 459
	let i = Chars.V.mk_int (parse_char loc i)
	and j = Chars.V.mk_int (parse_char loc j) in
460
        Elem (mk loc (Internal (Types.char (Chars.char_class i j))))
461
    |  s = STRING1 ->
462
	List.fold_right
463
	  (fun c accu -> 
464
	     let c = Chars.V.mk_int c in
465
	     let c = Chars.atom c in
466 467
	     Seq (Elem (mknoloc (Internal (Types.char c))), accu))
	  (seq_of_string s)
468 469
	  Epsilon ]
    | [ e = pat LEVEL "simple" -> Elem e
470 471 472
    ]
  ];

473
  schema_kind : [
474
    [ IDENT "element" -> `Element
475
    | "type" -> `Type
476 477 478
    | IDENT "attribute" -> `Attribute
    | IDENT "attribute_group" -> `Attribute_group
    | IDENT "model_group" -> `Model_group
479 480 481
    ]
  ];

482 483 484
  schema_ref: [
    [ schema = IDENT; "#"; typ = [ IDENT | keyword ];
      kind = OPT [ "as"; k = schema_kind -> k] ->
485
        (kind, U.mk schema, U.mk typ)
486 487 488
    ]
  ];

489
  pat: [ 
490
      [ x = pat; "where"; 
491
        b = LIST1 [ a = IDENT; "="; y = pat -> (ident a,y) ] SEP "and"
492 493
            -> mk loc (Recurs (x,b)) ]
    | RIGHTA [ x = pat; "->"; y = pat -> mk loc (Arrow (x,y)) ]
494
    | "no_arrow" [ x = pat; "|"; y = pat -> mk loc (Or (x,y)) ] 
495
    | "simple" [ x = pat; "&"; y = pat -> mk loc (And (x,y)) 
496
	       | x = pat; "\\"; y = pat -> mk loc (Diff (x,y)) ]
497
    | 
498 499
      [ "{"; r = record_spec; "}" -> mk loc (Record (true,r))
      | "{|"; r = record_spec; "|}" -> mk loc (Record (false,r))
500 501 502 503 504
      | "ref"; p = pat ->
	  let get_fun = mk loc (Arrow (pat_nil, p)) 
	  and set_fun = mk loc (Arrow (p, pat_nil)) in
	  let fields = [ label "get", get_fun; label "set", set_fun ] in
	  mk loc (Record (false, fields))
505
      | IDENT "_" -> mk loc (Internal Types.any)
506
      | "("; a = IDENT; ":="; c = expr; ")" -> 
507
	  mk loc (Constant (ident a,c))
508
      | schema = IDENT; "#"; typ = [ IDENT | keyword ];
509
        kind = OPT [ "as"; k = schema_kind -> k] ->
510
          mk loc (SchemaVar (kind, U.mk schema, U.mk typ))
511 512
      | "!"; a = IDENT ->
	  mk loc (Internal (Types.abstract (Types.Abstract.atom a)))
513
      | a = IDENT -> 
514
	  mk loc (PatVar (U.mk a))
515
      | i = INT ; "--"; j = INT -> 
516 517
          let i =  Intervals.V.mk i 
	  and j =  Intervals.V.mk j in
518 519
          mk loc (Internal (Types.interval (Intervals.bounded i j)))
      | i = INT -> 
520
          let i =  Intervals.V.mk i  in
521
          mk loc (Internal (Types.interval (Intervals.atom i)))
522
      | "*"; "--"; j = INT ->
523
	  let j =  Intervals.V.mk j in
524
          mk loc (Internal (Types.interval (Intervals.left j)))
525
      | i = INT; "--"; "*" ->
526
	  let i = Intervals.V.mk i in
527
          mk loc (Internal (Types.interval (Intervals.right i)))
528 529
      | i = char ->
          mk loc (Internal (Types.char (Chars.char_class i i)))
530
      | i = char ; "--"; j = char ->
531
          mk loc (Internal (Types.char (Chars.char_class i j)))
532
      | "`"; c = tag_type -> c
533 534 535
      | "("; l = LIST1 pat SEP ","; ")" -> multi_prod loc l
      | "["; r = [ r = regexp -> r | -> Epsilon ];
             q = [ ";"; q = pat -> q 
536
                 | -> pat_nil ]; 
537
             "]" -> mk loc (Regexp (r,q))
538
      | "<"; t =
539 540
            [ x = tag_type -> x
            | "("; t = pat; ")" -> t ];
541
	a = attrib_spec; ">"; c = pat ->
542
          mk loc (XmlT (t, multi_prod loc [a;c]))
543
      | s = STRING2 ->
544 545 546 547 548 549
	  let s = 
	    List.map 
	      (fun c -> 
		 mknoloc (Internal
			     (Types.char
				(Chars.atom
550
				   (Chars.V.mk_int c))))) 
551 552
	      (seq_of_string s) in
	  let s = s @ [mknoloc (Internal (Sequence.nil_type))] in
553
	  multi_prod loc s
554 555 556 557 558
      ]
    
  ];

  record_spec:
559
    [ [ r = LIST0 [ l = [IDENT | keyword ]; "=";
560
                  o = [ "?" -> true | -> false]; 
561 562
                  x = pat -> 
		    let x = if o then mk loc (Optional x) else x in
563
		    (label l, x)
564
                ] SEP ";" ->
565
	  r
566 567
      ] ];
  
568 569
  char:
    [ 
570
      [ c = STRING1 -> Chars.V.mk_int (parse_char loc c) ]
571 572 573
    ];
     

574
  attrib_spec:
575
    [ [ r = LIST0 [ l = [IDENT | keyword ]; "=";
576 577 578 579 580 581
                  o = [ "?" -> true | -> false]; 
                  x = pat; OPT ";" -> 
		    let x = if o then mk loc (Optional x) else x in
		    (label l, x)
                ] ->
	  mk loc (Record (true,r)) 
582 583 584 585
      | "("; t = pat; ")" -> t 
      | "{"; r = record_spec; "}" -> mk loc (Record (true,r))
      | "{|"; r = record_spec; "|}" -> mk loc (Record (false,r))
      ] ];
586 587 588

  expr_record_spec:
    [ [ r = LIST1
589
	      [ l = [IDENT | keyword ]; "="; x = expr LEVEL "no_appl" -> 
590
		  (label l,x) ] 
591
	       SEP ";"  ->
592
	  exp loc (RecordLitt r)
593
      ] ];
594

595
  expr_attrib_spec:
596
    [ [ r = LIST1
597
	      [ l = [IDENT | keyword ]; "="; x = expr LEVEL "no_appl"; OPT ";" -> 
598 599 600
		  (label l,x) ] ->
	  exp loc (RecordLitt r)
      ]
601
    | [ e = expr LEVEL "no_appl" -> e 
602
      | -> exp loc (RecordLitt []) 
603 604 605 606
      ] 
    ];
END

607 608 609
module Hook = struct
  let expr = expr
  let pat = pat
610
  let keyword = keyword
611 612
end

613
let pat = Grammar.Entry.parse pat
614 615
and expr = Grammar.Entry.parse expr
and prog = Grammar.Entry.parse prog
616 617
and top_phrases = Grammar.Entry.parse top_phrases

618
let sync () = 
619
  match !Ulexer.lexbuf with
620 621 622
    | None -> ()
    | Some lb ->
	let rec aux () =
623
	  match !Ulexer.last_tok with
624 625
	    | ("",";;") | ("EOI","") -> ()
	    | _ -> 
626
		Ulexer.last_tok := fst (Ulexer.token lb); 
627 628 629
		aux ()
	in
	aux ()
630 631 632 633

let sync () =
  try sync ()
  with Ulexing.Error -> ()