Coccinelle release-1.0.0-rc11
[bpt/coccinelle.git] / parsing_c / unparse_c.ml
1 (* Yoann Padioleau, Julia Lawall
2 *
3 * Copyright (C) 2012, INRIA.
4 * Copyright (C) 2010, 2011, University of Copenhagen DIKU and INRIA.
5 * Copyright (C) 2006, 2007, 2008, 2009 Ecole des Mines de Nantes and DIKU
6 *
7 * This program is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License (GPL)
9 * version 2 as published by the Free Software Foundation.
10 *
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * file license.txt for more details.
15 *
16 *
17 * Modifications by Julia Lawall for better newline handling.
18 *)
19 open Common
20
21 open Ast_c
22
23 module TH = Token_helpers
24
25
26 (* should keep comments and directives in between adjacent deleted terms,
27 but not comments and directives within deleted terms. should use the
28 labels found in the control-flow graph *)
29
30
31
32 (*****************************************************************************)
33 (* Wrappers *)
34 (*****************************************************************************)
35 let pr2, pr2_once = Common.mk_pr2_wrappers Flag_parsing_c.verbose_unparsing
36
37 (*****************************************************************************)
38 (* Types used during the intermediate phases of the unparsing *)
39 (*****************************************************************************)
40
41 type token1 =
42 | Fake1 of info
43 | T1 of Parser_c.token
44
45 (* The cocci_tag of the token should always be a NOTHING. The mark of
46 * the token can only be OriginTok or ExpandedTok. Why not get rid of
47 * token and get something simpler ? because we need to know if the
48 * info is a TCommentCpp or TCommentSpace, etc for some of the further
49 * analysis so easier to keep with the token.
50 *
51 * This type contains the whole information. Have all the tokens with this
52 * type.
53 *)
54 type min =
55 Min of (int list (* match numbers from witness trees *) *
56 Ast_cocci.adjacency (* adjacency information *))
57 | Ctx
58
59 type token2 =
60 | T2 of Parser_c.token * min *
61 int option (* orig index, abstracting away comments and space *)
62 | Fake2 of min
63 | Cocci2 of string * int (* line *) * int (* lcol *) * int (* rcol *)
64 * Unparse_cocci.nlhint option
65 | C2 of string
66 | Comma of string
67 | Indent_cocci2
68 | Unindent_cocci2 of bool (* true for permanent, false for temporary *)
69 | EatSpace2
70
71 (* not used yet *)
72 type token3 =
73 | T3 of Parser_c.token
74 | Cocci3 of string
75 | C3 of string
76
77
78 (* similar to the tech in parsing_hack *)
79 type token_extended = {
80 tok2 : token2;
81 str : string;
82 idx: int option; (* to know if 2 tokens were consecutive in orig file *)
83 mutable new_tokens_before : token2 list;
84 mutable remove : bool;
85 }
86
87
88 (*****************************************************************************)
89 (* Helpers *)
90 (*****************************************************************************)
91
92 let info_of_token1 t =
93 match t with
94 | Fake1 info -> info
95 | T1 tok -> TH.info_of_tok tok
96
97 let print_token1 = function
98 T1 tok -> TH.str_of_tok tok
99 | Fake1 info -> "fake"
100
101 let str_of_token2 = function
102 | T2 (t,_,_) -> TH.str_of_tok t
103 | Fake2 _ -> ""
104 | Cocci2 (s,_,_,_,_) -> s
105 | C2 s -> s
106 | Comma s -> s
107 | Indent_cocci2 -> ""
108 | Unindent_cocci2 _ -> ""
109 | EatSpace2 -> ""
110
111 let print_token2 = function
112 | T2 (t,b,_) ->
113 let t_str =
114 match t with
115 | Parser_c.TCommentSpace _ -> " sp "
116 | Parser_c.TCommentNewline _ -> " nl "
117 | Parser_c.TCommentCpp _ -> " cp "
118 | Parser_c.TCommentMisc _ -> " misc "
119 | Parser_c.TComment _ -> " comment "
120 | _ -> "" in
121 let b_str =
122 match b with
123 Min (index,adj) ->
124 Printf.sprintf "-%d[%s]"
125 (match adj with Ast_cocci.ADJ n -> n | _ -> -1)
126 (String.concat " " (List.map string_of_int index))
127 | Ctx -> "" in
128 "T2:"^b_str^t_str^TH.str_of_tok t
129 | Fake2 b ->
130 let b_str =
131 match b with
132 Min (index,adj) ->
133 Printf.sprintf "-%d[%s]"
134 (match adj with Ast_cocci.ADJ n -> n | _ -> -1)
135 (String.concat " " (List.map string_of_int index))
136 | Ctx -> "" in
137 b_str^"fake"
138 | Cocci2 (s,_,lc,rc,_) -> Printf.sprintf "Cocci2:%d:%d%s" lc rc s
139 | C2 s -> "C2:"^s
140 | Comma s -> "Comma:"^s
141 | Indent_cocci2 -> "Indent"
142 | Unindent_cocci2 _ -> "Unindent"
143 | EatSpace2 -> "EatSpace"
144
145 let simple_print_all_tokens1 l =
146 List.iter (function x -> Printf.printf "|%s| " (print_token1 x)) l;
147 Printf.printf "\n"
148
149 let simple_print_all_tokens2 l =
150 List.iter (function x -> Printf.printf "|%s| " (print_token2 x)) l;
151 Printf.printf "\n"
152
153 let str_of_token3 = function
154 | T3 t -> TH.str_of_tok t
155 | Cocci3 s | C3 s -> s
156
157
158
159 let mk_token_extended x =
160 let origidx =
161 match x with
162 | T2 (_,_, idx) -> idx
163 | _ -> None
164 in
165 { tok2 = x;
166 str = str_of_token2 x;
167 idx = origidx;
168 new_tokens_before = [];
169 remove = false;
170 }
171
172 let rebuild_tokens_extented toks_ext =
173 let _tokens = ref [] in
174 toks_ext +> List.iter (fun tok ->
175 tok.new_tokens_before +> List.iter (fun x -> push2 x _tokens);
176 if not tok.remove then push2 tok.tok2 _tokens;
177 );
178 let tokens = List.rev !_tokens in
179 (tokens +> List.map mk_token_extended)
180
181
182 let mcode_contain_plus = function
183 | Ast_cocci.CONTEXT (_,Ast_cocci.NOTHING) -> false
184 | Ast_cocci.CONTEXT _ -> true
185 (* patch: when need full coccinelle transformation *)
186 | Ast_cocci.MINUS (_,_,_,Ast_cocci.NOREPLACEMENT) -> false
187 | Ast_cocci.MINUS (_,_,_,Ast_cocci.REPLACEMENT _) -> true(*REPL is not empty*)
188 | Ast_cocci.PLUS _ -> raise Impossible
189
190 let contain_plus info =
191 let mck = Ast_c.mcode_of_info info in
192 mcode_contain_plus mck
193
194 (*****************************************************************************)
195 (* Last fix on the ast *)
196 (*****************************************************************************)
197
198 (* Because of the ugly trick to handle initialiser, I generate fake ','
199 * for the last initializer element, but if there is nothing around it,
200 * I don't want in the end to print it.
201 *)
202
203 let remove_useless_fakeInfo_struct program =
204 let bigf = { Visitor_c.default_visitor_c_s with
205 Visitor_c.kini_s = (fun (k,bigf) ini ->
206 match k ini with
207 | InitList args, ii ->
208 (match ii with
209 | [_i1;_i2] -> ini
210 | [i1;i2;iicommaopt] ->
211 if (not (contain_plus iicommaopt)) && (not (contain_plus i2))
212 && (Ast_c.is_fake iicommaopt)
213 (* sometimes the guy put a normal iicommaopt *)
214 then InitList args, [i1;i2]
215 else InitList args, [i1;i2;iicommaopt]
216 | [i1;i2;iicommaopt;end_comma_opt] ->
217 (* only in #define. end_comma_opt canot be fake *)
218 (* not sure if this will be considered ambiguous with a previous
219 case? *)
220 if (not (contain_plus iicommaopt)) && (not (contain_plus i2))
221 && (Ast_c.is_fake iicommaopt)
222 (* sometimes the guy put a normal iicommaopt *)
223 then InitList args, [i1;i2;end_comma_opt]
224 else InitList args, [i1;i2;iicommaopt;end_comma_opt]
225 | _ -> raise Impossible
226 )
227 | x -> x
228 )
229 } in
230 Visitor_c.vk_toplevel_s bigf program
231
232
233 (*****************************************************************************)
234 (* Tokens1 generation *)
235 (*****************************************************************************)
236
237 let get_fakeInfo_and_tokens celem toks =
238 let toks_in = ref toks in
239 let toks_out = ref [] in
240
241 (* todo? verify good order of position ? *)
242 let pr_elem info =
243 match Ast_c.pinfo_of_info info with
244 | FakeTok _ ->
245 Common.push2 (Fake1 info) toks_out
246 | OriginTok _ | ExpandedTok _ ->
247 (* get the associated comments/space/cppcomment tokens *)
248 let (before, x, after) =
249 !toks_in +> Common.split_when (fun tok ->
250 info =*= TH.info_of_tok tok)
251 in
252 assert(info =*= TH.info_of_tok x);
253 (*old: assert(before +> List.for_all (TH.is_comment)); *)
254 before +> List.iter (fun x ->
255 if not (TH.is_comment x)
256 then pr2 ("WEIRD: not a comment:" ^ TH.str_of_tok x)
257 (* case such as int asm d3("x"); not yet in ast *)
258 );
259 before +> List.iter (fun x -> Common.push2 (T1 x) toks_out);
260 push2 (T1 x) toks_out;
261 toks_in := after;
262 | AbstractLineTok _ ->
263 (* can be called on type info when for instance use -type_c *)
264 if !Flag_parsing_c.pretty_print_type_info
265 then Common.push2 (Fake1 info) toks_out
266 else raise Impossible (* at this stage *)
267 in
268
269 let pr_space _ = () in (* use the spacing that is there already *)
270
271 Pretty_print_c.pp_program_gen pr_elem pr_space celem;
272
273 if not (null !toks_in)
274 then failwith "WEIRD: unparsing not finished";
275
276 List.rev !toks_out
277
278 (* Fake nodes that have BEFORE code or are - should be moved over any subsequent
279 whitespace and newlines, but not any comments, to get as close to the affected
280 code as possible. Similarly, fake nodes that have AFTER code should be moved
281 backwards. No fake nodes should have both before and after code. *)
282
283 let displace_fake_nodes toks =
284 let is_fake = function Fake1 _ -> true | _ -> false in
285 let is_whitespace = function
286 T1(Parser_c.TCommentSpace _)
287 (* patch: cocci *)
288 | T1(Parser_c.TCommentNewline _) -> true
289 | _ -> false in
290 let rec loop toks =
291 let fake_info =
292 try Some (Common.split_when is_fake toks)
293 with Not_found -> None in
294 match fake_info with
295 Some(bef,((Fake1 info) as fake),aft) ->
296 (match !(info.cocci_tag) with
297 | Some x ->
298 (match x with
299 (Ast_cocci.CONTEXT(_,Ast_cocci.BEFORE _),_)
300 | (Ast_cocci.MINUS(_,_,_,Ast_cocci.REPLACEMENT _),_) ->
301 (* move the fake node forwards *)
302 let (whitespace,rest) = Common.span is_whitespace aft in
303 bef @ whitespace @ fake :: (loop rest)
304 | (Ast_cocci.CONTEXT(_,Ast_cocci.AFTER _),_) ->
305 (* move the fake node backwards *)
306 let revbef = List.rev bef in
307 let (revwhitespace,revprev) = Common.span is_whitespace revbef in
308 let whitespace = List.rev revwhitespace in
309 let prev = List.rev revprev in
310 prev @ fake :: (loop (whitespace @ aft))
311 | (Ast_cocci.CONTEXT(_,Ast_cocci.NOTHING),_) ->
312 bef @ fake :: (loop aft)
313 | (Ast_cocci.CONTEXT(_,Ast_cocci.BEFOREAFTER _),_) ->
314 failwith "fake node should not be before-after"
315 | _ -> bef @ fake :: (loop aft) (* old: was removed when have simpler yacfe *)
316 )
317 | None ->
318 bef @ fake :: (loop aft)
319 )
320 | None -> toks
321 | _ -> raise Impossible in
322 loop toks
323
324 (*****************************************************************************)
325 (* Tokens2 generation *)
326 (*****************************************************************************)
327
328 let comment2t2 = function
329 (Token_c.TCommentCpp
330 (* not sure iif the following list is exhaustive or complete *)
331 (Token_c.CppAttr|Token_c.CppMacro|Token_c.CppPassingCosWouldGetError),
332 (info : Token_c.info)) ->
333 C2(info.Common.str)
334 | (Token_c.TCommentCpp x,(info : Token_c.info)) ->
335 C2("\n"^info.Common.str^"\n")
336 | x -> failwith (Printf.sprintf "unexpected comment %s" (Common.dump x))
337
338 let expand_mcode toks =
339 let toks_out = ref [] in
340
341 let index = ref 0 in
342
343 let add_elem t minus =
344 match t with
345 | Fake1 info ->
346 let str = Ast_c.str_of_info info in
347 if str =$= ""
348 then push2 (Fake2 minus) toks_out
349 (* fx the fake "," at the end of a structure or enum.
350 no idea what other fake info there can be... *)
351 else push2 (Comma str) toks_out
352
353
354 | T1 tok ->
355 (*let (a,b) = !((TH.info_of_tok tok).cocci_tag) in*)
356 (* no tag on expandedTok ! *)
357 let modified = function
358 None -> false
359 | Some (Ast_cocci.CONTEXT(pos,Ast_cocci.NOTHING),l) -> false
360 | _ -> true in
361 (if (TH.is_expanded tok &&
362 modified !((TH.info_of_tok tok).cocci_tag)
363 (*!((TH.info_of_tok tok).cocci_tag) <> Ast_c.emptyAnnot*))
364 then
365 failwith
366 (Printf.sprintf
367 "expanded token %s on line %d is either modified or stored in a metavariable"
368 (TH.str_of_tok tok) (TH.line_of_tok tok)));
369
370 let tok' = tok +> TH.visitor_info_of_tok (fun i ->
371 { i with cocci_tag = ref Ast_c.emptyAnnot; }
372 ) in
373
374 let optindex =
375 if TH.is_origin tok && not (TH.is_real_comment tok)
376 then begin
377 incr index;
378 Some !index
379 end
380 else None
381 in
382
383 push2 (T2 (tok', minus, optindex)) toks_out
384 in
385
386 let expand_info t =
387 let (mcode,env) =
388 Ast_c.mcode_and_env_of_cocciref ((info_of_token1 t).cocci_tag) in
389
390 let pr_cocci s ln col rcol hint =
391 push2 (Cocci2(s,ln,col,rcol,hint)) toks_out in
392 let pr_c info =
393 (match Ast_c.pinfo_of_info info with
394 Ast_c.AbstractLineTok _ ->
395 push2 (C2 (Ast_c.str_of_info info)) toks_out
396 | Ast_c.FakeTok (s,_) ->
397 push2 (C2 s) toks_out
398 | _ ->
399 Printf.printf "line: %s\n" (Common.dump info);
400 failwith "not an abstract line");
401 (!(info.Ast_c.comments_tag)).Ast_c.mafter +>
402 List.iter (fun x -> Common.push2 (comment2t2 x) toks_out) in
403
404 let pr_barrier ln col = (* marks a position, used around C code *)
405 push2 (Cocci2("",ln,col,col,None)) toks_out in
406 let pr_nobarrier ln col = () in (* not needed for linux spacing *)
407
408 let pr_cspace _ = push2 (C2 " ") toks_out in
409
410 let pr_space _ = () (* rely on add_space in cocci code *) in
411 let pr_arity _ = () (* not interested *) in
412
413 let indent _ = push2 Indent_cocci2 toks_out in
414 let unindent x = push2 (Unindent_cocci2 x) toks_out in
415 let eat_space _ = push2 EatSpace2 toks_out in
416
417 let args_pp =
418 (env, pr_cocci, pr_c, pr_cspace,
419 (match !Flag_parsing_c.spacing with
420 Flag_parsing_c.SMPL -> pr_space | _ -> pr_cspace),
421 pr_arity,
422 (match !Flag_parsing_c.spacing with
423 Flag_parsing_c.SMPL -> pr_barrier | _ -> pr_nobarrier),
424 indent, unindent, eat_space) in
425
426 (* old: when for yacfe with partial cocci:
427 * add_elem t false;
428 *)
429
430 (* patch: when need full coccinelle transformation *)
431 let unparser = Unparse_cocci.pp_list_list_any args_pp false in
432 match mcode with
433 | Ast_cocci.MINUS (_,inst,adj,any_xxs) ->
434 (* Why adding ? because I want to have all the information, the whole
435 * set of tokens, so I can then process and remove the
436 * is_between_two_minus for instance *)
437 add_elem t (Min (inst,adj));
438 (match any_xxs with
439 Ast_cocci.NOREPLACEMENT -> ()
440 | Ast_cocci.REPLACEMENT(any_xxs,_) ->
441 unparser any_xxs Unparse_cocci.InPlace)
442 | Ast_cocci.CONTEXT (_,any_befaft) ->
443 (match any_befaft with
444 | Ast_cocci.NOTHING ->
445 add_elem t Ctx
446 | Ast_cocci.BEFORE (xxs,_) ->
447 unparser xxs Unparse_cocci.Before;
448 add_elem t Ctx
449 | Ast_cocci.AFTER (xxs,_) ->
450 add_elem t Ctx;
451 unparser xxs Unparse_cocci.After;
452 | Ast_cocci.BEFOREAFTER (xxs, yys, _) ->
453 unparser xxs Unparse_cocci.Before;
454 add_elem t Ctx;
455 unparser yys Unparse_cocci.After;
456 )
457 | Ast_cocci.PLUS _ -> raise Impossible
458
459 in
460
461 toks +> List.iter expand_info;
462 List.rev !toks_out
463
464
465 (*****************************************************************************)
466 (* Tokens2 processing, filtering, adjusting *)
467 (*****************************************************************************)
468
469 let is_space = function
470 | T2(Parser_c.TCommentSpace _,_b,_i) -> true (* only whitespace *)
471 | _ -> false
472
473 let is_newline = function
474 | T2(Parser_c.TCommentNewline _,_b,_i) -> true
475 | _ -> false
476
477 let is_whitespace = function
478 | (T2 (t,_b,_i)) ->
479 (match t with
480 | Parser_c.TCommentSpace _ -> true (* only whitespace *)
481 | Parser_c.TCommentNewline _ (* newline plus whitespace *) -> true
482 | _ -> false
483 )
484 | _ -> false
485
486 let is_minusable_comment = function
487 | (T2 (t,_b,_i)) ->
488 (match t with
489 | Parser_c.TCommentSpace _ (* only whitespace *)
490 (* patch: coccinelle *)
491 | Parser_c.TCommentNewline _ (* newline plus whitespace *) -> true
492 | Parser_c.TComment _ when !Flag_parsing_c.keep_comments -> false
493 | Parser_c.TComment _
494 | Parser_c.TCommentCpp (Token_c.CppAttr, _)
495 | Parser_c.TCommentCpp (Token_c.CppMacro, _)
496 | Parser_c.TCommentCpp (Token_c.CppDirective, _) (* result was false *)
497 -> true
498
499 | Parser_c.TCommentMisc _
500 | Parser_c.TCommentCpp (Token_c.CppPassingCosWouldGetError, _)
501 -> false
502
503 | _ -> false
504 )
505 | _ -> false
506
507 let is_minusable_comment_nocpp = function
508 | (T2 (t,_b,_i)) ->
509 (match t with
510 | Parser_c.TCommentSpace _ (* only whitespace *)
511 (* patch: coccinelle *)
512 | Parser_c.TCommentNewline _ (* newline plus whitespace *) -> true
513 | Parser_c.TComment _ when !Flag_parsing_c.keep_comments -> false
514 | Parser_c.TComment _ -> true
515 | Parser_c.TCommentCpp (Token_c.CppAttr, _)
516 | Parser_c.TCommentCpp (Token_c.CppMacro, _)
517 | Parser_c.TCommentCpp (Token_c.CppDirective, _)
518 -> false
519
520 | Parser_c.TCommentMisc _
521 | Parser_c.TCommentCpp (Token_c.CppPassingCosWouldGetError, _)
522 -> false
523
524 | _ -> false
525 )
526 | _ -> false
527
528 let all_coccis = function
529 Cocci2 _ | C2 _ | Comma _ | Indent_cocci2 | Unindent_cocci2 _
530 | EatSpace2 -> true
531 | _ -> false
532
533 (*previously gave up if the first character was a newline, but not clear why*)
534 let is_minusable_comment_or_plus x = is_minusable_comment x or all_coccis x
535
536 let set_minus_comment adj = function
537 | T2 (t,Ctx,idx) ->
538 let str = TH.str_of_tok t in
539 (match t with
540 | Parser_c.TCommentSpace _
541 (* patch: coccinelle *)
542 | Parser_c.TCommentNewline _ -> ()
543
544 | Parser_c.TComment _
545 | Parser_c.TCommentCpp (Token_c.CppAttr, _)
546 | Parser_c.TCommentCpp (Token_c.CppMacro, _)
547 | Parser_c.TCommentCpp (Token_c.CppDirective, _)
548 ->
549 pr2 (Printf.sprintf "%d: ERASING_COMMENTS: %s"
550 (TH.line_of_tok t) str)
551 | _ -> raise Impossible
552 );
553 T2 (t, Min adj, idx)
554 (* patch: coccinelle *)
555 | T2 (t,Min adj,idx) as x -> x
556 | Fake2 _ as x -> x
557 | _ -> raise Impossible
558
559 let set_minus_comment_or_plus adj = function
560 Cocci2 _ | C2 _ | Comma _ | Indent_cocci2 | Unindent_cocci2 _
561 | EatSpace2 as x -> x
562 | x -> set_minus_comment adj x
563
564 let drop_minus xs =
565 xs +> Common.exclude (function
566 | T2 (t,Min adj,_) -> true
567 | _ -> false
568 )
569
570 let drop_expanded xs =
571 xs +> Common.exclude (function
572 | T2 (t,_,_) when TH.is_expanded t -> true
573 | _ -> false
574 )
575
576 let drop_fake xs =
577 xs +> Common.exclude (function
578 | Fake2 _ -> true
579 | _ -> false
580 )
581
582 let remove_minus_and_between_and_expanded_and_fake xs =
583
584 (* get rid of expanded tok *)
585 let xs = drop_expanded xs in
586
587 let minus_or_comment = function
588 T2(_,Min adj,_) -> true
589 | x -> is_minusable_comment x in
590
591 let minus_or_comment_nocpp = function
592 T2(_,Min adj,_) -> true
593 | x -> is_minusable_comment_nocpp x in
594
595 let common_adj (index1,adj1) (index2,adj2) =
596 let same_adj = (* same adjacency info *)
597 match (adj1,adj2) with
598 (Ast_cocci.ADJ adj1,Ast_cocci.ADJ adj2) -> adj1 = adj2
599 | (Ast_cocci.ALLMINUS,_) | (_,Ast_cocci.ALLMINUS) -> true in
600 same_adj &&
601 (* non-empty intersection of witness trees *)
602 not ((Common.inter_set index1 index2) = []) in
603
604 (* new idea: collects regions not containing non-space context code
605 if two adjacent adjacent minus tokens satisfy common_adj then delete
606 all spaces, comments etc between them
607 if two adjacent minus tokens do not satisfy common_adj only delete
608 the spaces between them if there are no comments, etc.
609 if the region contain no plus code and is both preceded and followed
610 by a newline, delete the initial newline. *)
611
612 let rec adjust_around_minus = function
613 [] -> []
614 | (T2(Parser_c.TCommentNewline c,_b,_i) as x)::
615 ((Fake2(Min adj1) | T2(_,Min adj1,_)) as t1)::xs ->
616 let (minus_list,rest) = Common.span not_context (t1::xs) in
617 let contains_plus = List.exists is_plus minus_list in
618 let x =
619 match List.rev minus_list with
620 (T2(Parser_c.TCommentNewline c,_b,_i))::rest
621 when List.for_all minus_or_comment minus_list ->
622 set_minus_comment_or_plus adj1 x
623 | _ -> x in
624 x :: adjust_within_minus contains_plus minus_list @
625 adjust_around_minus rest
626 | ((Fake2(Min adj1) | T2(_,Min adj1,_)) as t1)::xs ->
627 let (minus_list,rest) = Common.span not_context (t1::xs) in
628 let contains_plus = List.exists is_plus minus_list in
629 adjust_within_minus contains_plus minus_list @ adjust_around_minus rest
630 | x::xs ->
631 x :: adjust_around_minus xs
632 and adjust_within_minus cp (* contains plus *) = function
633 ((Fake2(Min adj1) | T2(_,Min adj1,_)) as t1)::xs ->
634 let not_minus = function T2(_,Min _,_) -> false | _ -> true in
635 let (not_minus_list,rest) = Common.span not_minus xs in
636 t1 ::
637 (match rest with
638 ((Fake2(Min adj2) | T2(_,Min adj2,_)) as t2)::xs
639 when common_adj adj1 adj2 ->
640 (List.map (set_minus_comment_or_plus adj1) not_minus_list)
641 @ (adjust_within_minus cp (t2::xs))
642 | ((Fake2(Min adj2) | T2(_,Min adj2,_)) as t2)::xs ->
643 if not cp && List.for_all is_whitespace not_minus_list
644 then
645 (List.map (set_minus_comment_or_plus adj1) not_minus_list)
646 @ (adjust_within_minus cp (t2::xs))
647 else
648 not_minus_list @ (adjust_within_minus cp (t2::xs))
649 | _ ->
650 if cp
651 then xs
652 else
653 (* remove spaces after removed stuff, eg a comma after a
654 function argument *)
655 let (spaces,rest) = Common.span is_space xs in
656 (List.map (set_minus_comment_or_plus adj1) spaces)
657 @ rest)
658 | xs -> failwith "should always start with minus"
659 and not_context = function
660 (T2(_,Ctx,_) as x) when not (is_minusable_comment x) -> false
661 | _ -> true
662 and is_plus = function
663 C2 _ | Comma _ | Cocci2 _ -> true
664 | _ -> false in
665
666 let xs = adjust_around_minus xs in
667
668 (* get rid of fake tok *)
669 let xs = drop_fake xs in
670
671 (* this drops blank lines after a brace introduced by removing code *)
672 let minus_or_comment_nonl = function
673 T2(_,Min adj,_) -> true
674 | T2(Parser_c.TCommentNewline _,_b,_i) -> false
675 | x -> is_minusable_comment x in
676
677 let rec adjust_after_brace = function
678 [] -> []
679 | ((T2(_,Ctx,_)) as x)::((T2(_,Min adj,_)::_) as xs)
680 when str_of_token2 x =$= "{" ->
681 let (between_minus,rest) = Common.span minus_or_comment_nonl xs in
682 let is_whitespace = function
683 T2(Parser_c.TCommentSpace _,_b,_i)
684 (* patch: cocci *)
685 | T2(Parser_c.TCommentNewline _,_b,_i) -> true
686 | _ -> false in
687 let (newlines,rest) = Common.span is_whitespace rest in
688 let (drop_newlines,last_newline) =
689 let rec loop = function
690 [] -> ([],[])
691 | ((T2(Parser_c.TCommentNewline _,_b,_i)) as x) :: rest ->
692 (List.rev rest,[x])
693 | x::xs ->
694 let (drop_newlines,last_newline) = loop xs in
695 (drop_newlines,x::last_newline) in
696 loop (List.rev newlines) in
697 x::between_minus@(List.map (set_minus_comment adj) drop_newlines)@
698 last_newline@
699 adjust_after_brace rest
700 | x::xs -> x::adjust_after_brace xs in
701
702 let xs = adjust_after_brace xs in
703
704 (* search backwards from context } over spaces until reaching a newline.
705 then go back over all minus code until reaching some context or + code.
706 get rid of all intervening spaces, newlines, and comments
707 input is reversed *)
708 let rec adjust_before_brace = function
709 [] -> []
710 | ((T2(t,Ctx,_)) as x)::xs when str_of_token2 x =$= "}" or is_newline x ->
711 let (outer_spaces,rest) = Common.span is_space xs in
712 x :: outer_spaces @
713 (match rest with
714 ((T2 (Parser_c.TCommentNewline _,Ctx,_i)) as h) ::
715 (* the rest of this code is the same as from_newline below
716 but merging them seems to be error prone... *)
717 ((T2 (t, Min adj, idx)) as m) :: rest ->
718 let (spaces,rest) = Common.span minus_or_comment_nocpp rest in
719 h :: m ::
720 (List.map (set_minus_comment adj) spaces) @
721 (adjust_before_brace rest)
722 | _ -> adjust_before_brace rest)
723 | x::xs -> x :: (adjust_before_brace xs) in
724
725 let from_newline = function
726 ((T2 (t, Min adj, idx)) as m) :: rest ->
727 let (spaces,rest) = Common.span minus_or_comment_nocpp rest in
728 m ::
729 (List.map (set_minus_comment adj) spaces) @
730 (adjust_before_brace rest)
731 | ((T2 (t0, Ctx, idx0)) as m0) :: ((T2 (t, Min adj, idx)) as m) :: rest
732 when TH.str_of_tok t0 = "" ->
733 (* This is for the case of a #define that is completely deleted,
734 because a #define has a strange EOL token at the end.
735 We hope there i no other kind of token that is represented by
736 "", but it seems like changing the kind of token might break
737 the end of entity recognition in the C parser.
738 See parsing_hacks.ml *)
739 let (spaces,rest) = Common.span minus_or_comment_nocpp rest in
740 m0 :: m ::
741 (List.map (set_minus_comment adj) spaces) @
742 (adjust_before_brace rest)
743 | rest -> adjust_before_brace rest in
744
745 let xs = List.rev (from_newline (List.rev xs)) in
746 let xs = drop_minus xs in
747 xs
748
749 (* things that should not be followed by space - boundary between SmPL
750 code and C code *)
751 let adjust_eat_space toks =
752 let rec loop = function
753 [] -> []
754 | EatSpace2 :: x :: rest when is_space x -> loop rest
755 | EatSpace2 :: rest -> loop rest
756 | x :: xs -> x :: loop xs in
757 loop toks
758
759 (* normally, in C code, a semicolon is not preceded by a space or newline *)
760 let adjust_before_semicolon toks =
761 let toks = List.rev toks in
762 let rec search_semic = function
763 [] -> []
764 | ((T2(_,Ctx,_)) as x)::xs | ((Cocci2 _) as x)::xs ->
765 if List.mem (str_of_token2 x) [";";")";","]
766 then x :: search_semic (search_minus false xs)
767 else x :: search_semic xs
768 | x::xs -> x :: search_semic xs
769 and search_minus seen_minus xs =
770 let (spaces, rest) = Common.span is_space xs in
771 (* only delete spaces if something is actually deleted *)
772 match rest with
773 ((T2(_,Min _,_)) as a)::rerest -> a :: search_minus true rerest
774 | _ -> if seen_minus then rest else xs in
775 List.rev (search_semic toks)
776
777 (* normally, in C code, a ( is not followed by a space or newline *)
778 let adjust_after_paren toks =
779 let rec search_paren = function
780 [] -> []
781 | ((T2(_,Ctx,_)) as x)::xs | ((Cocci2 _) as x)::xs ->
782 if List.mem (str_of_token2 x) ["("] (* other things? *)
783 then x :: search_paren(search_minus false xs)
784 else x :: search_paren xs
785 | x::xs -> x :: search_paren xs
786 and search_minus seen_minus xs =
787 let (spaces, rest) = Common.span is_whitespace xs in
788 (* only delete spaces if something is actually deleted *)
789 match rest with
790 ((T2(_,Min _,_)) as a)::rerest -> (* minus *)
791 a :: search_minus true rerest
792 | ((T2(_,Ctx,_)) as a)::rerest when seen_minus && str_of_token2 a = "," ->
793 (* comma after ( will be deleted, so consider it as minus code
794 already *)
795 a :: search_minus true rerest
796 | _ -> if seen_minus then rest else xs in (* drop trailing space *)
797 search_paren toks
798
799 (* this is for the case where braces are added around an if branch *)
800 let paren_then_brace toks =
801 let rec search_paren = function
802 [] -> []
803 | ((T2(_,Ctx,_)) as x)::xs ->
804 if List.mem (str_of_token2 x) [")"]
805 then x :: search_paren(search_plus xs)
806 else x :: search_paren xs
807 | x::xs -> x :: search_paren xs
808 and search_plus xs =
809 let (spaces, rest) = Common.span is_whitespace xs in
810 match rest with
811 (* move the brace up to the previous line *)
812 ((Cocci2("{",_,_,_,_)) as x) :: (((Cocci2 _) :: _) as rest) ->
813 (C2 " ") :: x :: spaces @ rest
814 | _ -> xs in
815 search_paren toks
816
817 let is_ident_like s = s ==~ Common.regexp_alpha
818
819 let rec drop_space_at_endline = function
820 [] -> []
821 | [x] -> [x]
822 | (C2 " ") ::
823 ((((T2(Parser_c.TCommentSpace _,Ctx,_)) | Cocci2("\n",_,_,_,_) |
824 (T2(Parser_c.TCommentNewline _,Ctx,_))) :: _) as rest) ->
825 (* when unparse_cocci doesn't know whether space is needed *)
826 drop_space_at_endline rest
827 | ((T2(Parser_c.TCommentSpace _,Ctx,_i)) as a)::rest ->
828 let (outer_spaces,rest) = Common.span is_space rest in
829 let minus_or_comment_or_space_nocpp = function
830 T2(_,Min adj,_) -> true
831 | (T2(Parser_c.TCommentSpace _,Ctx,_i)) -> true
832 | (T2(Parser_c.TCommentNewline _,Ctx,_i)) -> false
833 | x -> false in
834 let (minus,rest) = Common.span minus_or_comment_or_space_nocpp rest in
835 let fail _ = a :: outer_spaces @ minus @ (drop_space_at_endline rest) in
836 if List.exists (function T2(_,Min adj,_) -> true | _ -> false) minus
837 then
838 match rest with
839 ((T2(Parser_c.TCommentNewline _,Ctx,_i)) as a)::rest ->
840 (* drop trailing spaces *)
841 minus@a::(drop_space_at_endline rest)
842 | _ -> fail()
843 else fail()
844 | a :: rest ->
845 a :: drop_space_at_endline rest
846
847 (* if a removed ( is between two tokens, then add a space *)
848 let rec paren_to_space = function
849 [] -> []
850 | [x] -> [x]
851 | [x;y] -> [x;y]
852 | ((T2(_,Ctx,_)) as a)::((T2(t,Min _,_)) as b)::((T2(_,Ctx,_)) as c)::rest
853 when not (is_whitespace a) && TH.str_of_tok t = "(" ->
854 a :: b :: (C2 " ") :: (paren_to_space (c :: rest))
855 | a :: rest -> a :: (paren_to_space rest)
856
857 let rec add_space xs =
858 match xs with
859 | [] -> []
860 | [x] -> [x]
861 | (Cocci2(sx,lnx,_,rcolx,_) as x)::((Cocci2(sy,lny,lcoly,_,_)) as y)::xs
862 when !Flag_parsing_c.spacing = Flag_parsing_c.SMPL &&
863 not (lnx = -1) && lnx = lny && not (rcolx = -1) && rcolx < lcoly ->
864 (* this only works within a line. could consider whether
865 something should be done to add newlines too, rather than
866 printing them explicitly in unparse_cocci. *)
867 x::C2 (String.make (lcoly-rcolx) ' ')::add_space (y::xs)
868 | (Cocci2(sx,lnx,_,rcolx,_) as x)::((Cocci2(sy,lny,lcoly,_,_)) as y)::xs
869 when !Flag_parsing_c.spacing = Flag_parsing_c.SMPL &&
870 not (lnx = -1) && lnx < lny && not (rcolx = -1) ->
871 (* this only works within a line. could consider whether
872 something should be done to add newlines too, rather than
873 printing them explicitly in unparse_cocci. *)
874 x::C2 (String.make (lny-lnx) '\n')::
875 C2 (String.make (lcoly-1) ' '):: (* -1 is for the + *)
876 add_space (y::xs)
877 | ((T2(_,Ctx,_)) as x)::((Cocci2 _) as y)::xs -> (* add space on boundary *)
878 let sx = str_of_token2 x in
879 let sy = str_of_token2 y in
880 if is_ident_like sx && (is_ident_like sy or List.mem sy ["="])
881 then x::C2 " "::(add_space (y::xs))
882 else x::(add_space (y::xs))
883 | x::y::xs -> (* not boundary, not sure if it is possible *)
884 let sx = str_of_token2 x in
885 let sy = str_of_token2 y in
886 if is_ident_like sx && is_ident_like sy
887 then x::C2 " "::(add_space (y::xs))
888 else x::(add_space (y::xs))
889
890 (* A fake comma is added at the end of an unordered initlist or a enum
891 decl, if the initlist or enum doesn't already end in a comma. This is only
892 needed if there is + code, ie if we see Cocci after it in the code sequence *)
893
894 let rec drop_end_comma = function
895 [] -> []
896 | [x] -> [x]
897 | ((Comma ",") as x) :: rest ->
898 let (newlines,rest2) = Common.span is_whitespace rest in
899 (match rest2 with
900 (Cocci2 _) :: _ -> x :: drop_end_comma rest
901 | _ -> drop_end_comma rest)
902 | x :: xs -> x :: drop_end_comma xs
903
904 (* The following only works for the outermost function call. Stack records
905 the column of all open parentheses. Space_cell contains the most recent
906 comma in the outermost function call. The goal is to decide whether this
907 should be followed by a space or a newline and indent. *)
908 let add_newlines toks tabbing_unit =
909 let create_indent n =
910 let (tu,tlen) =
911 match tabbing_unit with
912 Some ("\t",_) -> ("\t",8)
913 | Some ("",_) -> ("\t",8) (* not sure why... *)
914 | Some (s,_) -> (s,String.length s) (* assuming only spaces *)
915 | None -> ("\t",8) in
916 let rec loop seen =
917 if seen + tlen <= n
918 then tu ^ loop (seen + tlen)
919 else String.make (n-seen) ' ' in
920 loop 0 in
921 let check_for_newline count x = function
922 Some (start,space_cell) when count > Flag_parsing_c.max_width ->
923 space_cell := "\n"^(create_indent x);
924 Some (x + (count - start))
925 | _ -> None in
926 (* the following is for strings that may contain newline *)
927 let string_length s count =
928 let l = list_of_string s in
929 List.fold_left
930 (function count ->
931 function
932 '\t' -> count + 8
933 | '\n' -> 0
934 | c -> count + 1)
935 count l in
936 let rec loop info count = function
937 [] -> []
938 | ((T2(tok,_,_)) as a)::xs ->
939 a :: loop info (string_length (TH.str_of_tok tok) count) xs
940 | ((Cocci2(s,line,lcol,rcol,hint)) as a)::xs ->
941 let (stack,space_cell) = info in
942 let rest =
943 match hint with
944 None -> loop info (string_length s count) xs
945 | Some Unparse_cocci.StartBox ->
946 let count = string_length s count in
947 loop (count::stack,space_cell) count xs
948 | Some Unparse_cocci.EndBox ->
949 let count = string_length s count in
950 (match stack with
951 [x] ->
952 (match check_for_newline count x space_cell with
953 Some count -> loop ([],None) count xs
954 | None -> loop ([],None) count xs)
955 | _ -> loop (List.tl stack,space_cell) count xs)
956 | Some (Unparse_cocci.SpaceOrNewline sp) ->
957 let count = string_length s (count + 1 (*space*)) in
958 (match stack with
959 [x] ->
960 (match check_for_newline count x space_cell with
961 Some count -> loop (stack,Some (x,sp)) count xs
962 | None -> loop (stack,Some (count,sp)) count xs)
963 | _ -> loop info count xs) in
964 a :: rest
965 | ((C2(s)) as a)::xs -> a :: loop info (string_length s count) xs
966 | ((Comma(s)) as a)::xs -> a :: loop info (string_length s count) xs
967 | Fake2 _ :: _ | Indent_cocci2 :: _
968 | Unindent_cocci2 _::_ | EatSpace2::_ ->
969 failwith "unexpected fake, indent, unindent, or eatspace" in
970 let redo_spaces prev = function
971 Cocci2(s,line,lcol,rcol,Some (Unparse_cocci.SpaceOrNewline sp)) ->
972 C2 !sp :: Cocci2(s,line,lcol,rcol,None) :: prev
973 | t -> t::prev in
974 (match !Flag_parsing_c.spacing with
975 Flag_parsing_c.SMPL -> toks
976 | _ -> List.rev (List.fold_left redo_spaces [] (loop ([],None) 0 toks)))
977
978 (* When insert some new code, because of a + in a SP, we must add this
979 * code at the right place, with the good indentation. So each time we
980 * encounter some spacing info, with some newline, we maintain the
981 * current indentation level used.
982 *
983 * TODO problems: not accurate. ex: TODO
984 *
985 * TODO: if in #define region, should add a \ \n
986 *)
987 let new_tabbing2 space =
988 (list_of_string space)
989 +> List.rev
990 +> Common.take_until (fun c -> c =<= '\n')
991 +> List.rev
992 +> List.map string_of_char
993 +> String.concat ""
994
995 let new_tabbing a =
996 Common.profile_code "C unparsing.new_tabbing" (fun () -> new_tabbing2 a)
997
998
999 let rec adjust_indentation xs =
1000
1001 let _current_tabbing = ref ([] : string list) in
1002 let tabbing_unit = ref None in
1003
1004 let string_of_list l = String.concat "" (List.map string_of_char l) in
1005
1006 (* try to pick a tabbing unit for the plus code *)
1007 let adjust_tabbing_unit old_tab new_tab =
1008 if !tabbing_unit =*= None && String.length new_tab > String.length old_tab
1009 then
1010 let old_tab = list_of_string old_tab in
1011 let new_tab = list_of_string new_tab in
1012 let rec loop = function
1013 ([],new_tab) ->
1014 tabbing_unit := Some(string_of_list new_tab,List.rev new_tab)
1015 | (_,[]) -> failwith "not possible"
1016 | (o::os,n::ns) -> loop (os,ns) in (* could check for equality *)
1017 loop (old_tab,new_tab) in
1018
1019 (*
1020 let remtab tu current_tab =
1021 let current_tab = List.rev(list_of_string current_tab) in
1022 let rec loop = function
1023 ([],new_tab) -> string_of_list (List.rev new_tab)
1024 | (_,[]) -> (-*weird; tabbing unit used up more than the current tab*-)
1025 ""
1026 | (t::ts,n::ns) when t =<= n -> loop (ts,ns)
1027 | (_,ns) -> (-* mismatch; remove what we can *-)
1028 string_of_list (List.rev ns) in
1029 loop (tu,current_tab) in
1030 *)
1031
1032 let rec find_first_tab started = function
1033 [] -> ()
1034 | ((T2 (tok,_,_)) as x)::xs when str_of_token2 x =$= "{" ->
1035 find_first_tab true xs
1036 (* patch: coccinelle *)
1037 | ((T2 (Parser_c.TCommentNewline s, _, _)) as x)::_
1038 when started ->
1039 let s = str_of_token2 x +> new_tabbing in
1040 tabbing_unit := Some (s,List.rev (list_of_string s))
1041 | x::xs -> find_first_tab started xs in
1042 find_first_tab false xs;
1043
1044 let rec balanced ct = function
1045 [] -> ct >= 0
1046 | ((T2(tok,_,_)) as x)::xs ->
1047 (match str_of_token2 x with
1048 "(" -> balanced (ct+1) xs
1049 | ")" -> balanced (ct-1) xs
1050 | _ -> balanced ct xs)
1051 | x::xs -> balanced ct xs in
1052
1053 let update_tabbing started s x =
1054 let old_tabbing = !_current_tabbing in
1055 str_of_token2 x +> new_tabbing +> (fun s -> _current_tabbing := [s]);
1056 (* only trust the indentation after the first { *)
1057 if started
1058 then
1059 adjust_tabbing_unit
1060 (String.concat "" old_tabbing)
1061 (String.concat "" !_current_tabbing) in
1062
1063 let rec aux started xs =
1064 match xs with
1065 | [] -> []
1066 (* patch: coccinelle *)
1067 | ((T2 (Parser_c.TCommentNewline s,_,_)) as x)::
1068 Unindent_cocci2(false)::xs ->
1069 update_tabbing started s x;
1070 (C2 "\n")::aux started xs
1071 | (Cocci2("\n",_,_,_,_))::Unindent_cocci2(false)::xs ->
1072 (C2 "\n")::aux started xs
1073 | ((T2 (tok,_,_)) as x)::(T2 (Parser_c.TCommentNewline s, _, _))::
1074 ((Cocci2 ("{",_,_,_,_)) as a)::xs
1075 when started && str_of_token2 x =$= ")" ->
1076 (* to be done for if, etc, but not for a function header *)
1077 x::(C2 " ")::a::(aux started xs)
1078 | ((T2 (Parser_c.TCommentNewline s, _, _)) as x)::xs
1079 when
1080 balanced 0 (fst(Common.span (function x -> not(is_newline x)) xs)) ->
1081 update_tabbing started s x;
1082 let coccis_rest = Common.span all_coccis xs in
1083 (match coccis_rest with
1084 (_::_,((T2 (tok,_,_)) as y)::_) when str_of_token2 y =$= "}" ->
1085 (* the case where cocci code has been added before a close } *)
1086 x::aux started (Indent_cocci2::xs)
1087 | _ -> x::aux started xs)
1088 | Indent_cocci2::((Cocci2(sy,lny,lcoly,_,_)) as y)::xs
1089 when !Flag_parsing_c.spacing = Flag_parsing_c.SMPL ->
1090 let tu = String.make (lcoly-1) ' ' in
1091 _current_tabbing := tu::(!_current_tabbing);
1092 C2 (tu)::aux started (y::xs)
1093 | Indent_cocci2::xs ->
1094 (match !tabbing_unit with
1095 None -> aux started xs
1096 | Some (tu,_) ->
1097 _current_tabbing := tu::(!_current_tabbing);
1098 (* can't be C2, for later phases *)
1099 Cocci2 (tu,-1,-1,-1,None)::aux started xs)
1100 | Unindent_cocci2(permanent)::xs ->
1101 (match !_current_tabbing with
1102 [] -> aux started xs
1103 | _::new_tabbing ->
1104 let s = String.concat "" new_tabbing in
1105 _current_tabbing := new_tabbing;
1106 Cocci2 (s,-1,-1,-1,None)::aux started xs)
1107 (* border between existing code and cocci code *)
1108 | ((T2 (tok,_,_)) as x)::((Cocci2("\n",_,_,_,_)) as y)::xs
1109 when str_of_token2 x =$= "{" ->
1110 x::aux true (y::Indent_cocci2::xs)
1111 | ((Cocci2 _) as x)::((T2 (tok,_,_)) as y)::xs
1112 when str_of_token2 y =$= "}" ->
1113 x::aux started (Unindent_cocci2 true::y::xs)
1114 (* starting the body of the function *)
1115 | ((T2 (tok,_,_)) as x)::xs when str_of_token2 x =$= "{" -> x::aux true xs
1116 | ((Cocci2("{",_,_,_,_)) as a)::xs -> a::aux true xs
1117 | ((Cocci2("\n",_,_,_,_)) as x)::xs ->
1118 (* dont inline in expr because of weird eval order of ocaml *)
1119 let s = String.concat "" !_current_tabbing in
1120 (* can't be C2, for later phases *)
1121 x::Cocci2 (s,-1,-1,-1,None)::aux started xs
1122 | x::xs -> x::aux started xs in
1123 (aux false xs,!tabbing_unit)
1124
1125
1126 let rec find_paren_comma = function
1127 | [] -> ()
1128
1129 (* do nothing if was like this in original file *)
1130 | ({ str = "("; idx = Some p1 } as _x1)::({ str = ","; idx = Some p2} as x2)
1131 ::xs when p2 =|= p1 + 1 ->
1132 find_paren_comma (x2::xs)
1133
1134 | ({ str = ","; idx = Some p1 } as _x1)::({ str = ","; idx = Some p2} as x2)
1135 ::xs when p2 =|= p1 + 1 ->
1136 find_paren_comma (x2::xs)
1137
1138 | ({ str = ","; idx = Some p1 } as _x1)::({ str = ")"; idx = Some p2} as x2)
1139 ::xs when p2 =|= p1 + 1 ->
1140 find_paren_comma (x2::xs)
1141
1142 (* otherwise yes can adjust *)
1143 | ({ str = "(" } as _x1)::({ str = ","} as x2)::xs ->
1144 x2.remove <- true;
1145 find_paren_comma (x2::xs)
1146 | ({ str = "," } as x1)::({ str = ","} as x2)::xs ->
1147 x1.remove <- true;
1148 find_paren_comma (x2::xs)
1149
1150 | ({ str = "," } as x1)::({ str = ")"} as x2)::xs ->
1151 x1.remove <- true;
1152 find_paren_comma (x2::xs)
1153
1154 | x::xs ->
1155 find_paren_comma xs
1156
1157
1158 let fix_tokens toks =
1159 let toks = toks +> List.map mk_token_extended in
1160
1161 let cleaner = toks +> Common.exclude (function
1162 | {tok2 = T2 (t,_,_)} -> TH.is_real_comment t (* I want the ifdef *)
1163 | _ -> false
1164 ) in
1165 find_paren_comma cleaner;
1166
1167 let toks = rebuild_tokens_extented toks in
1168 toks +> List.map (fun x -> x.tok2)
1169
1170
1171
1172 (*****************************************************************************)
1173 (* Final unparsing (and debugging support) *)
1174 (*****************************************************************************)
1175
1176 (* for debugging *)
1177 type kind_token2 = KFake | KCocci | KC | KExpanded | KOrigin
1178
1179 let kind_of_token2 = function
1180 | Fake2 _ -> KFake
1181 | Cocci2 _ -> KCocci
1182 | C2 _ -> KC
1183 | Comma _ -> KC
1184 | T2 (t,_,_) ->
1185 (match TH.pinfo_of_tok t with
1186 | ExpandedTok _ -> KExpanded
1187 | OriginTok _ -> KOrigin
1188 | FakeTok _ -> raise Impossible (* now a Fake2 *)
1189 | AbstractLineTok _ -> raise Impossible (* now a KC *)
1190 )
1191 | Unindent_cocci2 _ | Indent_cocci2 | EatSpace2 -> raise Impossible
1192
1193 let end_mark = "!"
1194
1195 let start_mark = function
1196 | KFake -> "!F!"
1197 | KCocci -> "!S!"
1198 | KC -> "!A!"
1199 | KExpanded -> "!E!"
1200 | KOrigin -> ""
1201
1202 let print_all_tokens2 pr xs =
1203 if !Flag_parsing_c.debug_unparsing
1204 then
1205 let current_kind = ref KOrigin in
1206 xs +> List.iter (fun t ->
1207 let newkind = kind_of_token2 t in
1208 if newkind =*= !current_kind
1209 then pr (str_of_token2 t)
1210 else begin
1211 pr (end_mark);
1212 pr (start_mark newkind);
1213 pr (str_of_token2 t);
1214 current_kind := newkind
1215 end
1216 );
1217 else
1218 xs +> List.iter (fun x -> pr (str_of_token2 x))
1219
1220
1221
1222
1223 (*****************************************************************************)
1224 (* Entry points *)
1225 (*****************************************************************************)
1226
1227 (* old: PPviatok was made in the beginning to allow to pretty print a
1228 * complete C file, including a modified C file by transformation.ml,
1229 * even if we don't handle yet in pretty_print_c.ml, ast_to_flow (and
1230 * maybe flow_to_ast) all the cases. Indeed we don't need to do some
1231 * fancy stuff when a function was not modified at all. Just need to
1232 * print the list of token as-is. But now pretty_print_c.ml handles
1233 * almost everything so maybe less useful. Maybe PPviatok allows to
1234 * optimize a little the pretty printing.
1235 *
1236 * update: now have PPviastr which goes even faster than PPviatok, so
1237 * PPviatok has disappeared.
1238 *)
1239
1240 type ppmethod = PPnormal | PPviastr
1241
1242
1243
1244
1245 (* The pp_program function will call pretty_print_c.ml with a special
1246 * function to print the leaf components, the tokens. When we want to
1247 * print a token, we need to print also maybe the space and comments that
1248 * were close to it in the original file (and that was omitted during the
1249 * parsing phase), and honor what the cocci-info attached to the token says.
1250 * Maybe we will not print the token if it's a MINUS-token, and maybe we will
1251 * print it and also print some cocci-code attached in a PLUS to it.
1252 * So we will also maybe call unparse_cocci. Because the cocci-code may
1253 * contain metavariables, unparse_cocci will in fact sometimes call back
1254 * pretty_print_c (which will this time don't call back again unparse_cocci)
1255 *)
1256
1257 let pp_program2 xs outfile =
1258 Common.with_open_outfile outfile (fun (pr,chan) ->
1259 let pr s =
1260 if !Flag_parsing_c.debug_unparsing
1261 then begin pr2_no_nl s; flush stderr end
1262 else pr s
1263 (* flush chan; *)
1264 (* Common.pr2 ("UNPARSING: >" ^ s ^ "<"); *)
1265 in
1266
1267 xs +> List.iter (fun ((e,(str, toks_e)), ppmethod) ->
1268 (* here can still work on ast *)
1269 let e = remove_useless_fakeInfo_struct e in
1270
1271 match ppmethod with
1272 | PPnormal ->
1273 (* now work on tokens *)
1274 (* phase1: just get all the tokens, all the information *)
1275 assert(toks_e +> List.for_all (fun t ->
1276 TH.is_origin t or TH.is_expanded t
1277 ));
1278 let toks = get_fakeInfo_and_tokens e toks_e in
1279 let toks = displace_fake_nodes toks in
1280 (* assert Origin;ExpandedTok;Faketok *)
1281 let toks = expand_mcode toks in
1282
1283 (* assert Origin;ExpandedTok; + Cocci + C (was AbstractLineTok)
1284 * and no tag information, just NOTHING. *)
1285
1286 let toks =
1287 if !Flag.sgrep_mode2
1288 then
1289 (* nothing else to do for sgrep *)
1290 drop_expanded(drop_fake(drop_minus toks))
1291 else
1292 (* phase2: can now start to filter and adjust *)
1293 (let (toks,tu) = adjust_indentation toks in
1294 let toks = adjust_eat_space toks in
1295 let toks = adjust_before_semicolon toks in(*before remove minus*)
1296 let toks = adjust_after_paren toks in(*also before remove minus*)
1297 let toks = drop_space_at_endline toks in
1298 let toks = paren_to_space toks in
1299 let toks = drop_end_comma toks in
1300 let toks = remove_minus_and_between_and_expanded_and_fake toks in
1301 (* assert Origin + Cocci + C and no minus *)
1302 let toks = add_space toks in
1303 let toks = add_newlines toks tu in
1304 let toks = paren_then_brace toks in
1305 let toks = fix_tokens toks in
1306 toks) in
1307
1308 (* in theory here could reparse and rework the ast! or
1309 * apply some SP. Not before cos julia may have generated
1310 * not parsable file. Need do unparsing_tricks call before being
1311 * ready to reparse. *)
1312 print_all_tokens2 pr toks;
1313
1314 | PPviastr -> pr str
1315 )
1316 )
1317
1318 let pp_program a b =
1319 Common.profile_code "C unparsing" (fun () -> pp_program2 a b)
1320
1321
1322 let pp_program_default xs outfile =
1323 let xs' = xs +> List.map (fun x -> x, PPnormal) in
1324 pp_program xs' outfile