parser.ml 20.2 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 14 15 16 17
let tloc (i,j) = (i.Lexing.pos_cnum,j.Lexing.pos_cnum)
let nopos = (Lexing.dummy_pos, Lexing.dummy_pos)


let mk loc x = Location.mk (tloc loc) x

18 19
exception Error of string
let error (i,j) s = Location.raise_loc i j (Error s)
20
let error loc s = error (tloc loc) s
21

22
let gram    = Grammar.gcreate Ulexer.lex
23

24
let parse_ident = U.mk
25

26
let id_dummy = ident (U.mk "$$$")
27

28
let label = parse_ident
29
let ident s = ident (parse_ident s)
30

31
let prog    = Grammar.Entry.create gram "prog"
32
let top_phrases   = Grammar.Entry.create gram "toplevel phrases"
33 34 35
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"
36
let keyword = Grammar.Entry.create gram "keyword"
37
		
38
let exp pos e = LocatedExpr (loc_of_pos (tloc pos),e)
39

40 41 42 43 44
let rec multi_prod loc = function
  | [ x ] -> x
  | x :: l -> mk loc (Prod (x, multi_prod loc l))
  | [] -> assert false
      
45
let rec tuple = function
46
  | [ x ] -> x
47
  | x :: l -> Pair (x, tuple l)
48
  | [] -> assert false
49 50

let tuple_queue = 
51
  List.fold_right (fun x q -> Pair (x, q))
52 53


54
let char = mknoloc (Internal (Types.char Chars.any))
55
let string_regexp = Star (Elem char)
56
	       
57
let seq_of_string s =
58 59
  let s = Encodings.Utf8.mk s in
  let rec aux i j =
60 61
    if Encodings.Utf8.equal_index i j then []
    else let (c,i) = Encodings.Utf8.next s i in c :: (aux i j)
62
  in
63
  aux (Encodings.Utf8.start_index s) (Encodings.Utf8.end_index s)
64

65

66
let parse_char loc s =
67 68
  match seq_of_string s with
    | [ c ] -> c
69
    | _ -> error loc "Character litteral must have length 1"
70

71 72
let include_stack = ref []

73 74 75 76
let protect_exn f g =
  try let x = f () in g (); x
  with e -> g (); raise e

77 78 79 80
let localize_exn f = 
  try f ()
  with
  | Stdpp.Exc_located (_, (Location _ as e)) -> raise e
81 82
(*  | Stdpp.Exc_located ((i,j), e) -> raise_loc i j e *)
  | Stdpp.Exc_located ((i,j), e) -> raise_loc i.Lexing.pos_cnum j.Lexing.pos_cnum e
83

84 85 86 87
let is_fun_decl =
   Grammar.Entry.of_parser gram "[is_fun_decl]"
     (fun strm ->
       match Stream.npeek 3 strm with
88 89
	 | [ ("", "fun"); ("IDENT", _); ("", "(") ]
	 | [ ("IDENT", _) ; ("", "(") ; _ ] -> ()
90 91 92
	 | _ -> raise Stream.Failure
     )

93 94 95 96 97
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
98 99 100

let apply_op2_noloc op e1 e2 = Apply (Apply (Var (parse_ident op), e1), e2)
let apply_op2 loc op e1 e2 = exp loc (apply_op2_noloc op e1 e2)
101
 
102

103
EXTEND
104
  GLOBAL: top_phrases prog expr pat regexp keyword;
105 106

  top_phrases: [
107
    [ l = LIST0 phrase; ";;" -> List.flatten l  ]
108
  ];
109 110

  prog: [
111
    [ l = LIST0 [ p = phrase ; OPT ";;" -> p ]; EOI -> List.flatten l ]
112 113 114
  ];

  phrase: [
115 116 117 118
    [ (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"->
119
        [ mk loc (EvalStatement (exp loc (Match (e1,[p,e2])))) ]
120
    | "type"; x = IDENT; "="; t = pat -> [ mk loc (TypeDecl (ident x,t)) ]
121 122
    | "using"; name = IDENT; "="; cu = [ IDENT | STRING2 ] ->
	[ mk loc (Using (U.mk name, U.mk cu)) ]
123 124 125
    | "schema"; name = IDENT; 
      p = OPT [ "("; name = [ IDENT | keyword ]; ")" -> parse_ident name ];
      "="; uri = STRING2 ->
126
	protect_op "schema";
127
        [ mk loc (SchemaDecl (U.mk name, uri, p)) ]
128 129 130 131 132
    | (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)) ]
133
    | "debug"; d = debug_directive -> [ mk loc (Directive (`Debug d)) ]
134 135
    | DIRECTIVE "#verbose" -> [ mk loc (Directive `Verbose) ]
    | DIRECTIVE "#silent" -> [ mk loc (Directive `Silent) ]
136 137 138
    | DIRECTIVE "#utf8" -> Ulexer.enc := Ulexing.Utf8; [ ]
    | DIRECTIVE "#latin1" -> Ulexer.enc := Ulexing.Latin1; [ ]
    | DIRECTIVE "#ascii" -> Ulexer.enc := Ulexing.Ascii; [ ]
139 140
    | DIRECTIVE "#quit" -> [ mk loc (Directive `Quit) ]
    | DIRECTIVE "#env" -> [ mk loc (Directive `Env) ]
141
    | DIRECTIVE "#print_schema"; name = IDENT ->
142
        [ mk loc (Directive (`Print_schema (U.mk name))) ]
143 144
    | DIRECTIVE "#print_type"; t = pat ->
        [ mk loc (Directive (`Print_type t)) ]
145
    | DIRECTIVE "#dump_value"; e = expr -> [ mk loc (Directive (`Dump e)) ]
146
    | DIRECTIVE "#reinit_ns" -> [ mk loc (Directive `Reinit_ns) ]
147
    | DIRECTIVE "#help" -> [ mk loc (Directive `Help) ]
148
    | "include"; s = STRING2 ->
149 150 151 152
	let s = 
	  if Filename.is_relative s 
	  then Filename.concat (Location.current_dir ()) s
	  else s in
153 154
	protect_op "File inclusion";
	(* avoid looping; should issue an error ? *)
155 156
	(* it is possible to have looping with x/../x/../x/.. ....
	   Need to canonicalize filename *)
157 158 159 160
	if List.mem s !include_stack then [] 
	else (
	  include_stack := s :: !include_stack;
	  Location.push_source (`File s);
161 162
	  let saved_enc = !Ulexer.enc in
	  Ulexer.enc := Ulexing.Latin1;
163 164 165
	  protect_exn
	    (fun () ->
	       let chan = open_in s in
166 167 168
	       protect_exn
		 (fun () ->
		    let input = Stream.of_channel chan in
169
		    localize_exn (fun () -> Grammar.Entry.parse prog input))
170
		 (fun () -> close_in chan))
171
	    (fun () ->
172
	       Ulexer.enc := saved_enc;
173 174
	       Location.pop_source ();
	       include_stack := List.tl !include_stack)
175
	)
176
    ] | 
177
    [ e = expr -> [ mk loc (EvalStatement e) ]
178 179 180 181
    ]
  ];

  debug_directive: [
182 183 184 185 186
    [ 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)
187
    | IDENT "explain"; t0 = pat; t = pat; e = expr -> `Explain (t0,t,e)
188 189
    | IDENT "single"; t = pat -> `Single t
    | IDENT "approx"; p = pat; t = pat -> `Approx (p,t)
190
    ]
191
  ];
192

193 194 195 196 197
  keyword: [
    [ a = 
	[ "map" | "match" | "with" | "try" | "xtransform"
	| "if" | "then"	| "else"
	| "transform" | "fun" | "in"
198
	| "let" | "type" | "debug" | "include"
199
        | "and" | "validate" | "schema" | "namespace" | "ref" | "alias"
200
	| "not" | "as" | "where" | "external"
201 202 203 204 205
	]
	-> a
    ]
  ];

206 207
  expr: [
    "top" RIGHTA
208 209
    [ "match"; e = SELF; "with"; b = branches -> 
	exp loc (Match (e,b))
210
    | "try"; e = SELF; "with"; b = branches -> 
211
	exp loc (Try (e,b))
212 213 214 215
    | "map"; e = SELF; "with"; b = branches -> 
	exp loc (Map (e,b))
    | "xtransform"; e = SELF; "with"; b = branches -> 
	exp loc (Xtrans (e,b))
216
    | "if"; e = SELF; "then"; e1 = SELF; "else"; e2 = SELF ->
217
	exp loc (if_then_else e e1 e2)
218
    | "transform"; e = SELF; "with"; b = branches -> 
219
	exp loc (Transform (e,b))
220 221
    | "validate"; e = SELF; "with"; (kind, schema, typ) = schema_ref ->
        exp loc (Validate (e, kind, schema, typ))
222
    | "fun"; (f,a,b) = fun_decl ->
223
	exp loc (Abstraction { fun_name = f; fun_iface = a; fun_body = b })
224
    | "external"; s = STRING2 ->
225
	exp loc (External (s,[]))
226
    | "external"; "{"; s = STRING2; pl = LIST0 pat; "}" ->
227
	exp loc (External (s,pl))
228
    | (_,p,e1) = let_binding; "in"; e2 = expr LEVEL "top"->
229
        exp loc (Match (e1,[p,e2]))
230 231
    | (name,ns) = namespace_binding; "in"; e2 = expr LEVEL "top" ->
	exp loc (NamespaceIn (name, ns, e2))
232
    | e = expr; ":"; p = pat ->
233
	exp loc (Forget (e,p))
234 235
    | e = expr; ":"; "?"; p = pat ->
	exp loc (Check (e,p))
236 237 238 239
    | e1 = expr; ";"; e2 = expr ->
	exp loc (Match (e1, [pat_nil,e2]))
    | "ref"; p = pat; e = expr ->
	exp loc (Ref (e,p))
240
    | "not"; e = expr -> exp loc (logical_not e)
241 242 243 244
    ]
    |
    [ e1 = expr; ":="; e2 = expr ->
	exp loc (Apply (Dot (e1, U.mk "set"), e2))
245
    ]
246 247 248 249 250 251
    | 
    [ e1 = expr; op = ["=" | "<=" | "<<" | ">>" | ">=" ]; e2 = expr -> 
	let op = match op with
	  | "<<" -> "<"
	  | ">>" -> ">"
	  | s -> s in
252
	apply_op2 loc op e1 e2
253 254
    ]

255
    | 
256
    [ e1 = expr; op = ["+" | "-" | "@" ]; e2 = expr -> apply_op2 loc op e1 e2
257
    | e1 = expr; "||"; e2 = expr -> exp loc (logical_or e1 e2)
258
    | e = expr; "\\"; l = [IDENT | keyword ] -> 
259
	exp loc (RemoveField (e, label l)) 
260 261
    ]
    |
262
    [ e1 = expr; op = ["*"]; e2 = expr -> apply_op2 loc op e1 e2
263
    | e1 = expr; "&&"; e2 = expr -> exp loc (logical_and e1 e2)
264
    | e = expr; op = "/"; p = pat LEVEL "simple" ->
265 266 267 268
	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
269
	let ct = mk loc (Regexp re) in
270 271 272
        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]))
273
    ]
274
    |
275
    [ e = expr;  "."; l = [IDENT | keyword ] -> 
276
	exp loc (Dot (e, label l)) 
277
    ]  
278 279 280
    | [ 
      e1 = SELF; IDENT "div"; e2 = expr -> apply_op2 loc "/" e1 e2
    | e1 = SELF; IDENT "mod"; e2 = expr -> apply_op2 loc "mod" e1 e2
281
    | e1 = SELF; e2 = expr -> exp loc (Apply (e1,e2))
282 283
    ]

284
    | "no_appl" 
285 286
    [ 
      "("; l = LIST1 expr SEP ","; ")" -> exp loc (tuple l)
287
    | "[";  l = LIST0 seq_elem; e = OPT [ ";"; e = expr -> e ]; 
288
      loc_end = ["]" -> loc] ->
289
	let e = match e with Some e -> e | None -> cst_nil in
290 291
        let e = exp loc_end e in
        let (_,loc_end) = loc_end in
292 293 294
	let l = List.fold_right 
		  (fun x q ->
		     match x with
295
		       | `String (loc,i,j,s) -> exp loc (String (i,j,s,q))
296
		       | `Elems ((loc,_),x) -> exp (loc,loc_end) (Pair(x,q))
297
		       | `Explode x -> apply_op2_noloc "@" x q
298 299 300
		  ) l e
	in
	exp loc l
301
    | "<"; t = [ "("; e = expr; ")" -> e
302
	       | a = tag -> exp loc a
303
	       ];
304
	a = expr_attrib_spec; ">"; c = expr ->
305
	  exp loc (Xml (t, Pair (a,c)))
306
    | "{"; r = [ expr_record_spec | -> exp loc (RecordLitt []) ]; "}" -> r
307
    | s = STRING2 ->
308
	let s = U.mk s in
309
	exp loc (String (U.start_index s, U.end_index s, s, cst_nil))
310
    | a = IDENT -> exp loc (Var (U.mk a))
311 312
    | "!"; e = expr ->
	exp loc (Apply (Dot (e, U.mk "get"), cst_nil))
313 314 315
    | i = INT -> exp loc (Integer (Intervals.V.mk i))
    | "`"; a = tag -> a
    | c = char -> exp loc (Char c)
316 317 318
    ]

  ];
319

320
  tag: [ [ a = [ IDENT | keyword ] -> exp loc (Atom (parse_ident a)) ] ];
321 322

  tag_type: [
323
    [ IDENT "_" ->  mk loc (Internal (Types.atom Atoms.any))
324
    | a = [ IDENT | keyword ] -> mk loc (Cst (Atom (parse_ident a)))
325
    | t = ANY_IN_NS -> mk loc (NsT (parse_ident t)) 
326 327 328
    ]
  ];

329
  seq_elem: [
330
    [ x = STRING1 -> 
331
	let s = U.mk x in
332
	`String (loc, U.start_index s, U.end_index s, s)
333
    | e = expr LEVEL "no_appl" -> `Elems (loc,e)
334
    | "!"; e = expr LEVEL "no_appl" -> `Explode e
335 336
    ]
  ];
337 338 339
	
  namespace_binding: [
    [ "namespace"; 
340
	name = [ name = [ IDENT | keyword ]; "=" ->
341 342 343
		   parse_ident name
	       | -> U.mk "" ];
	uri = STRING2 ->
344
	  let ns = Ns.mk (parse_ident uri) in
345 346 347 348 349
	  (name,ns)
    ]
  ];

  
350
  let_binding: [
351 352
    [ "let"; is_fun_decl; OPT "fun"; (f,a,b) = fun_decl ->
	let f = match f with Some x -> x | None -> assert false in
353
	let p = mk loc (PatVar (Id.value f)) in
354
	let abst = { fun_name = Some f; fun_iface = a; fun_body = b } in
355
        let e = exp loc (Abstraction abst) in
356
        (true,p,e)
357 358
    | "let"; p = pat; "="; e = expr -> (false,p,e)
    | "let"; p = pat; ":"; t = pat; "="; e = expr -> (false,p, Forget (e,t))
359 360
    | "let"; p = pat; ":"; "?"; t = pat; "="; e = expr -> 
	(false,p, Check (e,t))
361 362 363
    ] 
  ];

364
 fun_decl_after_lparen: [
365
(* need an hack to do this, because both productions would
366
   match   [ OPT IDENT; "("; pat ] .... *)
367 368 369 370 371 372
   [ 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) ]; 
373 374 375 376 377 378 379 380 381
	     ")";
	     others = LIST0 
			[ "(";
			  args = 
			    LIST1 
			      [ arg = pat; ":"; targ = pat -> (arg,targ) ]
			      SEP ",";
			  ")" -> args ];
	     ":"; tres = pat ; 
382
	     "="; body = expr ->
383
	       `Compact (targ1,args,others,tres,body)
384 385 386
	   ] ->
       match res with
	 | `Classic (p2,a,b) -> (p1,p2)::a,b
387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
	 | `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)]
	      ] ];
406 407 408


 fun_decl: [
409 410 411
   [ 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)
412
   ]
413 414
 ];

415
 arrow: [
416
    [ t1 = pat LEVEL "no_arrow"; "->"; t2 = pat -> (t1,t2)]
417 418 419
  ];

  branches: [
420
    [ OPT "|"; l = LIST1 branch SEP "|" -> l ]
421 422 423
  ];

  branch: [
424
    [ p = pat LEVEL "no_arrow"; "->"; e = expr -> (p,e) ]
425 426 427 428
  ];

	  
  regexp: [ 
429 430 431 432 433
    [ x = regexp; "|"; y = regexp -> 
	match (x,y) with
	  | Elem x, Elem y -> Elem (mk loc (Or (x,y)))
	  | _ -> Alt (x,y) 
    ]
434
  | [ x = regexp; y = regexp -> Seq (x,y) ]
435 436 437 438 439
  | [ 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"
    ]
440
  | [ a = IDENT; "::"; x = regexp -> SeqCapture (ident a,x) ]
441 442 443 444
  | [ x = regexp; "*" -> Star x
    | x = regexp; "*?" -> WeakStar x
    | x = regexp; "+" -> Seq (x, Star x)
    | x = regexp; "+?" -> Seq (x, WeakStar x)
445
    | x = regexp; "?" ->  Alt (x, Epsilon)
446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462
    | x = regexp; "??" -> Alt (Epsilon, x) 
    | x = regexp; "**"; i = INT ->
	let rec aux i accu =
	  if (i = 0) then accu
	  else aux (pred i) (Seq (x, accu))
	in
	let i = 
	  try 
	    let i = int_of_string i in
	    if (i > 1024) then raise Exit else i
	      (* We cannot handle type that huge... *)
	  with Failure _ | Exit -> error loc "Repetition number too large"
	in
	if (i <= 0) then 
	  error loc "Repetition number must be a positive integer";
	aux i Epsilon
    ]
463
  | [ "("; x = LIST1 regexp SEP ","; ")" ->
464 465 466 467 468 469 470 471 472 473 474
	(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))
475 476
    | "("; a = IDENT; ":="; c = expr; ")" -> 
	Elem (mk loc (Constant ((ident a,c))))
477
    | "/"; p = pat LEVEL "simple" -> Guard p
478
    | IDENT "PCDATA" -> string_regexp
479
    | i = STRING1; "--"; j = STRING1 ->
480 481
	let i = Chars.V.mk_int (parse_char loc i)
	and j = Chars.V.mk_int (parse_char loc j) in
482
        Elem (mk loc (Internal (Types.char (Chars.char_class i j))))
483
    |  s = STRING1 ->
484
	List.fold_right
485
	  (fun c accu -> 
486
	     let c = Chars.V.mk_int c in
487
	     let c = Chars.atom c in
488 489
	     Seq (Elem (mknoloc (Internal (Types.char c))), accu))
	  (seq_of_string s)
490 491
	  Epsilon ]
    | [ e = pat LEVEL "simple" -> Elem e
492 493 494
    ]
  ];

495
  schema_kind : [
496
    [ IDENT "element" -> `Element
497
    | "type" -> `Type
498 499 500
    | IDENT "attribute" -> `Attribute
    | IDENT "attribute_group" -> `Attribute_group
    | IDENT "model_group" -> `Model_group
501 502 503
    ]
  ];

504 505 506
  schema_ref: [
    [ schema = IDENT; "#"; typ = [ IDENT | keyword ];
      kind = OPT [ "as"; k = schema_kind -> k] ->
507
        (kind, U.mk schema, U.mk typ)
508 509 510
    ]
  ];

511
  pat: [ 
512
      [ x = pat; "where"; 
513
        b = LIST1 [ a = IDENT; "="; y = pat -> (ident a,y) ] SEP "and"
514 515
            -> mk loc (Recurs (x,b)) ]
    | RIGHTA [ x = pat; "->"; y = pat -> mk loc (Arrow (x,y)) ]
516
    | "no_arrow" [ x = pat; "|"; y = pat -> mk loc (Or (x,y)) ] 
517
    | "simple" [ x = pat; "&"; y = pat -> mk loc (And (x,y)) 
518
	       | x = pat; "\\"; y = pat -> mk loc (Diff (x,y)) ]
519
    | 
520 521
      [ "{"; r = record_spec; "}" -> mk loc (Record (true,r))
      | "{|"; r = record_spec; "|}" -> mk loc (Record (false,r))
522 523
      | "ref"; p = pat ->
	  let get_fun = mk loc (Arrow (pat_nil, p)) 
524 525 526
	  and set_fun = mk loc (Arrow (p, pat_nil))in
	  let fields = 
	    [ label "get", (get_fun, None); label "set", (set_fun, None) ] in
527
	  mk loc (Record (false, fields))
528
      | IDENT "_" -> mk loc (Internal Types.any)
529
      | "("; a = IDENT; ":="; c = expr; ")" -> 
530
	  mk loc (Constant (ident a,c))
531
      | schema = IDENT; "#"; typ = [ IDENT | keyword ];
532
        kind = OPT [ "as"; k = schema_kind -> k] ->
533
          mk loc (SchemaVar (kind, U.mk schema, U.mk typ))
534 535
      | "!"; a = IDENT ->
	  mk loc (Internal (Types.abstract (Types.Abstract.atom a)))
536
      | a = IDENT -> 
537
	  mk loc (PatVar (U.mk a))
538
      | i = INT ; "--"; j = INT -> 
539 540
          let i =  Intervals.V.mk i 
	  and j =  Intervals.V.mk j in
541 542
          mk loc (Internal (Types.interval (Intervals.bounded i j)))
      | i = INT -> 
543
          let i =  Intervals.V.mk i  in
544
          mk loc (Internal (Types.interval (Intervals.atom i)))
545
      | "*"; "--"; j = INT ->
546
	  let j =  Intervals.V.mk j in
547
          mk loc (Internal (Types.interval (Intervals.left j)))
548
      | i = INT; "--"; "*" ->
549
	  let i = Intervals.V.mk i in
550
          mk loc (Internal (Types.interval (Intervals.right i)))
551 552
      | i = char ->
          mk loc (Internal (Types.char (Chars.char_class i i)))
553
      | i = char ; "--"; j = char ->
554
          mk loc (Internal (Types.char (Chars.char_class i j)))
555
      | "`"; c = tag_type -> c
556 557
      | "("; l = LIST1 pat SEP ","; ")" -> multi_prod loc l
      | "["; r = [ r = regexp -> r | -> Epsilon ];
558 559 560 561 562 563 564 565 566 567
             q = [ ";"; q = pat -> Some q
                 | -> None ]; 
             "]" -> 
	       let r = match q with
		 | Some q -> 	
		     let any = mk loc (Internal (Types.any)) in
		     Seq(r,Seq(Guard q, Star (Elem any)))
		 | None -> r
	       in
	       mk loc (Regexp r)
568
      | "<"; t =
569 570
            [ x = tag_type -> x
            | "("; t = pat; ")" -> t ];
571
	a = attrib_spec; ">"; c = pat ->
572
          mk loc (XmlT (t, multi_prod loc [a;c]))
573
      | s = STRING2 ->
574 575 576 577 578 579
	  let s = 
	    List.map 
	      (fun c -> 
		 mknoloc (Internal
			     (Types.char
				(Chars.atom
580
				   (Chars.V.mk_int c))))) 
581 582
	      (seq_of_string s) in
	  let s = s @ [mknoloc (Internal (Sequence.nil_type))] in
583
	  multi_prod loc s
584 585 586 587
      ]
    
  ];

588 589
  or_else : [ [ OPT [ "else"; y = pat -> y ]  ] ];

590
  opt_field_pat: [ [ OPT [ "=";
591
                  o = [ "?" -> true | -> false]; 
592 593 594 595 596 597 598 599 600 601 602 603
                  x = pat;  y = or_else -> (o,x,y) ] ] ];

  record_spec:
    [ [ r = LIST0 [ l = [IDENT | keyword ]; f = opt_field_pat ->
		      let (o,x,y) =
			match f with
			  | None -> (false, mknoloc (PatVar (U.mk l)), None)
			  | Some z -> z
		      in
		      let x = if o then mk loc (Optional x) else x in
		      (label l, (x,y))
                  ] SEP ";" ->
604
	  r
605 606
      ] ];
  
607 608
  char:
    [ 
609
      [ c = STRING1 -> Chars.V.mk_int (parse_char loc c) ]
610 611 612
    ];
     

613
  attrib_spec:
614 615 616 617 618 619 620 621 622
    [ [ r = LIST0 [ l = [IDENT | keyword ]; f = opt_field_pat; OPT ";" -> 
		      let (o,x,y) =
			match f with
			  | None -> (false, mknoloc (PatVar (U.mk l)), None)
			  | Some z -> z
		      in
		      let x = if o then mk loc (Optional x) else x in
		      (label l, (x, y))
                  ] ->
623
	  mk loc (Record (true,r)) 
624 625 626 627
      | "("; t = pat; ")" -> t 
      | "{"; r = record_spec; "}" -> mk loc (Record (true,r))
      | "{|"; r = record_spec; "|}" -> mk loc (Record (false,r))
      ] ];
628

629 630
  opt_field_expr: [ [ OPT [ "="; x = expr LEVEL "no_appl" -> x ] ] ];

631 632
  expr_record_spec:
    [ [ r = LIST1
633 634 635
	      [ l = [IDENT | keyword ]; 
		x = opt_field_expr ->
		  let x = match x with Some x -> x | None ->  Var (U.mk l) in
636
		  (label l,x) ] 
637
	       SEP ";"  ->
638
	  exp loc (RecordLitt r)
639
      ] ];
640

641
  expr_attrib_spec:
642
    [ [ r = LIST1
643 644 645
	      [ l = [IDENT | keyword ]; 
		x = opt_field_expr; OPT ";" ->
		  let x = match x with Some x -> x | None ->  Var (U.mk l) in
646 647 648
		  (label l,x) ] ->
	  exp loc (RecordLitt r)
      ]
649
    | [ e = expr LEVEL "no_appl" -> e 
650
      | -> exp loc (RecordLitt []) 
651 652 653 654
      ] 
    ];
END

655 656 657
module Hook = struct
  let expr = expr
  let pat = pat
658
  let keyword = keyword
659 660
end

661
let pat = Grammar.Entry.parse pat
662 663
and expr = Grammar.Entry.parse expr
and prog = Grammar.Entry.parse prog
664 665
and top_phrases = Grammar.Entry.parse top_phrases

666
let sync () = 
667
  match !Ulexer.lexbuf with
668 669 670
    | None -> ()
    | Some lb ->
	let rec aux () =
671
	  match !Ulexer.last_tok with
672 673
	    | ("",";;") | ("EOI","") -> ()
	    | _ -> 
674
		Ulexer.last_tok := fst (Ulexer.token lb); 
675 676 677
		aux ()
	in
	aux ()
678 679 680 681

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