?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
* TextControl.cs: Make this operation undoable.
[mono-project.git] / mcs / mcs / cs-tokenizer.cs
blob4c9ab645672cab54ef1d443df9df9e927bbd666c
1 //
2 // cs-tokenizer.cs: The Tokenizer for the C# compiler
3 // This also implements the preprocessor
4 //
5 // Author: Miguel de Icaza (miguel@gnu.org)
6 // Marek Safar (marek.safar@seznam.cz)
7 //
8 // Licensed under the terms of the GNU GPL
9 //
10 // (C) 2001, 2002 Ximian, Inc (http://www.ximian.com)
11 // (C) 2004 Novell, Inc
12 //
13
14 /*
15 * TODO:
16 * Make sure we accept the proper Unicode ranges, per the spec.
17 * Report error 1032
18 */
19
20 using System;
21 using System.Text;
22 using System.Collections;
23 using System.IO;
24 using System.Globalization;
25 using System.Reflection;
26
27 namespace Mono.CSharp
28 {
29 /// <summary>
30 /// Tokenizer for C# source code.
31 /// </summary>
32
33 public class Tokenizer : yyParser.yyInput
34 {
35 SeekableStreamReader reader;
36 SourceFile ref_name;
37 SourceFile file_name;
38 int ref_line = 1;
39 int line = 1;
40 int col = 0;
41 int previous_col;
42 int current_token;
43 bool handle_get_set = false;
44 bool handle_remove_add = false;
45 bool handle_assembly = false;
46 bool handle_constraints = false;
47 bool handle_typeof = false;
48 Location current_location;
49 Location current_comment_location = Location.Null;
50 ArrayList escapedIdentifiers = new ArrayList ();
51
52 //
53 // XML documentation buffer. The save point is used to divide
54 // comments on types and comments on members.
55 //
56 StringBuilder xml_comment_buffer;
57
58 //
59 // See comment on XmlCommentState enumeration.
60 //
61 XmlCommentState xmlDocState = XmlCommentState.Allowed;
62
63 //
64 // Whether tokens have been seen on this line
65 //
66 bool tokens_seen = false;
67
68 //
69 // Whether a token has been seen on the file
70 // This is needed because `define' is not allowed to be used
71 // after a token has been seen.
72 //
73 bool any_token_seen = false;
74
75 static Hashtable tokenValues;
76
77 private static Hashtable TokenValueName
78 {
79 get {
80 if (tokenValues == null)
81 tokenValues = GetTokenValueNameHash ();
82
83 return tokenValues;
84 }
85 }
86
87 private static Hashtable GetTokenValueNameHash ()
88 {
89 Type t = typeof (Token);
90 FieldInfo [] fields = t.GetFields ();
91 Hashtable hash = new Hashtable ();
92 foreach (FieldInfo field in fields) {
93 if (field.IsLiteral && field.IsStatic && field.FieldType == typeof (int))
94 hash.Add (field.GetValue (null), field.Name);
95 }
96 return hash;
97 }
98
99 //
100 // Returns a verbose representation of the current location
101 //
102 public string location {
103 get {
104 string det;
105
106 if (current_token == Token.ERROR)
107 det = "detail: " + error_details;
108 else
109 det = "";
110
111 // return "Line: "+line+" Col: "+col + "\n" +
112 // "VirtLine: "+ref_line +
113 // " Token: "+current_token + " " + det;
114 string current_token_name = TokenValueName [current_token] as string;
115 if (current_token_name == null)
116 current_token_name = current_token.ToString ();
117
118 return String.Format ("{0} ({1},{2}), Token: {3} {4}", ref_name.Name,
119 ref_line,
120 col,
121 current_token_name,
122 det);
123 }
124 }
125
126 public bool PropertyParsing {
127 get {
128 return handle_get_set;
129 }
130
131 set {
132 handle_get_set = value;
133 }
134 }
135
136 public bool AssemblyTargetParsing {
137 get {
138 return handle_assembly;
139 }
140
141 set {
142 handle_assembly = value;
143 }
144 }
145
146 public bool EventParsing {
147 get {
148 return handle_remove_add;
149 }
150
151 set {
152 handle_remove_add = value;
153 }
154 }
155
156 public bool ConstraintsParsing {
157 get {
158 return handle_constraints;
159 }
160
161 set {
162 handle_constraints = value;
163 }
164 }
165
166 public bool TypeOfParsing {
167 get {
168 return handle_typeof;
169 }
170
171 set {
172 handle_typeof = value;
173 }
174 }
175
176 public XmlCommentState doc_state {
177 get { return xmlDocState; }
178 set {
179 if (value == XmlCommentState.Allowed) {
180 check_incorrect_doc_comment ();
181 reset_doc_comment ();
182 }
183 xmlDocState = value;
184 }
185 }
186
187 public bool IsEscapedIdentifier (Location loc)
188 {
189 foreach (LocatedToken lt in escapedIdentifiers)
190 if (lt.Location.Equals (loc))
191 return true;
192 return false;
193 }
194
195 //
196 // Class variables
197 //
198 static CharArrayHashtable[] keywords;
199 static Hashtable keywordStrings = new Hashtable ();
200 static NumberStyles styles;
201 static NumberFormatInfo csharp_format_info;
202
203 //
204 // Values for the associated token returned
205 //
206 int putback_char;
207 Object val;
208
209 //
210 // Pre-processor
211 //
212 Hashtable defines;
213
214 const int TAKING = 1;
215 const int TAKEN_BEFORE = 2;
216 const int ELSE_SEEN = 4;
217 const int PARENT_TAKING = 8;
218 const int REGION = 16;
219
220 //
221 // pre-processor if stack state:
222 //
223 Stack ifstack;
224
225 static System.Text.StringBuilder string_builder;
226
227 const int max_id_size = 512;
228 static char [] id_builder = new char [max_id_size];
229
230 static CharArrayHashtable [] identifiers = new CharArrayHashtable [max_id_size + 1];
231
232 const int max_number_size = 512;
233 static char [] number_builder = new char [max_number_size];
234 static int number_pos;
235
236 //
237 // Details about the error encoutered by the tokenizer
238 //
239 string error_details;
240
241 public string error {
242 get {
243 return error_details;
244 }
245 }
246
247 public int Line {
248 get {
249 return ref_line;
250 }
251 }
252
253 public int Col {
254 get {
255 return col;
256 }
257 }
258
259 //
260 // This is used when the tokenizer needs to save
261 // the current position as it needs to do some parsing
262 // on its own to deamiguate a token in behalf of the
263 // parser.
264 //
265 Stack position_stack = new Stack (2);
266 class Position {
267 public int position;
268 public int ref_line;
269 public int col;
270 public int putback_char;
271 public int previous_col;
272 #if GMCS_SOURCES
273 public int parsing_generic_less_than;
274 #endif
275 public Position (Tokenizer t)
276 {
277 position = t.reader.Position;
278 ref_line = t.ref_line;
279 col = t.col;
280 putback_char = t.putback_char;
281 previous_col = t.previous_col;
282 #if GMCS_SOURCES
283 parsing_generic_less_than = t.parsing_generic_less_than;
284 #endif
285 }
286 }
287
288 public void PushPosition ()
289 {
290 position_stack.Push (new Position (this));
291 }
292
293 public void PopPosition ()
294 {
295 Position p = (Position) position_stack.Pop ();
296
297 reader.Position = p.position;
298 ref_line = p.ref_line;
299 col = p.col;
300 putback_char = p.putback_char;
301 previous_col = p.previous_col;
302
303 }
304
305 // Do not reset the position, ignore it.
306 public void DiscardPosition ()
307 {
308 position_stack.Pop ();
309 }
310
311 static void AddKeyword (string kw, int token) {
312 keywordStrings.Add (kw, kw);
313 if (keywords [kw.Length] == null) {
314 keywords [kw.Length] = new CharArrayHashtable (kw.Length);
315 }
316 keywords [kw.Length] [kw.ToCharArray ()] = token;
317 }
318
319 static void InitTokens ()
320 {
321 keywords = new CharArrayHashtable [64];
322
323 AddKeyword ("__arglist", Token.ARGLIST);
324 AddKeyword ("abstract", Token.ABSTRACT);
325 AddKeyword ("as", Token.AS);
326 AddKeyword ("add", Token.ADD);
327 AddKeyword ("assembly", Token.ASSEMBLY);
328 AddKeyword ("base", Token.BASE);
329 AddKeyword ("bool", Token.BOOL);
330 AddKeyword ("break", Token.BREAK);
331 AddKeyword ("byte", Token.BYTE);
332 AddKeyword ("case", Token.CASE);
333 AddKeyword ("catch", Token.CATCH);
334 AddKeyword ("char", Token.CHAR);
335 AddKeyword ("checked", Token.CHECKED);
336 AddKeyword ("class", Token.CLASS);
337 AddKeyword ("const", Token.CONST);
338 AddKeyword ("continue", Token.CONTINUE);
339 AddKeyword ("decimal", Token.DECIMAL);
340 AddKeyword ("default", Token.DEFAULT);
341 AddKeyword ("delegate", Token.DELEGATE);
342 AddKeyword ("do", Token.DO);
343 AddKeyword ("double", Token.DOUBLE);
344 AddKeyword ("else", Token.ELSE);
345 AddKeyword ("enum", Token.ENUM);
346 AddKeyword ("event", Token.EVENT);
347 AddKeyword ("explicit", Token.EXPLICIT);
348 AddKeyword ("extern", Token.EXTERN);
349 AddKeyword ("false", Token.FALSE);
350 AddKeyword ("finally", Token.FINALLY);
351 AddKeyword ("fixed", Token.FIXED);
352 AddKeyword ("float", Token.FLOAT);
353 AddKeyword ("for", Token.FOR);
354 AddKeyword ("foreach", Token.FOREACH);
355 AddKeyword ("goto", Token.GOTO);
356 AddKeyword ("get", Token.GET);
357 AddKeyword ("if", Token.IF);
358 AddKeyword ("implicit", Token.IMPLICIT);
359 AddKeyword ("in", Token.IN);
360 AddKeyword ("int", Token.INT);
361 AddKeyword ("interface", Token.INTERFACE);
362 AddKeyword ("internal", Token.INTERNAL);
363 AddKeyword ("is", Token.IS);
364 AddKeyword ("lock", Token.LOCK);
365 AddKeyword ("long", Token.LONG);
366 AddKeyword ("namespace", Token.NAMESPACE);
367 AddKeyword ("new", Token.NEW);
368 AddKeyword ("null", Token.NULL);
369 AddKeyword ("object", Token.OBJECT);
370 AddKeyword ("operator", Token.OPERATOR);
371 AddKeyword ("out", Token.OUT);
372 AddKeyword ("override", Token.OVERRIDE);
373 AddKeyword ("params", Token.PARAMS);
374 AddKeyword ("private", Token.PRIVATE);
375 AddKeyword ("protected", Token.PROTECTED);
376 AddKeyword ("public", Token.PUBLIC);
377 AddKeyword ("readonly", Token.READONLY);
378 AddKeyword ("ref", Token.REF);
379 AddKeyword ("remove", Token.REMOVE);
380 AddKeyword ("return", Token.RETURN);
381 AddKeyword ("sbyte", Token.SBYTE);
382 AddKeyword ("sealed", Token.SEALED);
383 AddKeyword ("set", Token.SET);
384 AddKeyword ("short", Token.SHORT);
385 AddKeyword ("sizeof", Token.SIZEOF);
386 AddKeyword ("stackalloc", Token.STACKALLOC);
387 AddKeyword ("static", Token.STATIC);
388 AddKeyword ("string", Token.STRING);
389 AddKeyword ("struct", Token.STRUCT);
390 AddKeyword ("switch", Token.SWITCH);
391 AddKeyword ("this", Token.THIS);
392 AddKeyword ("throw", Token.THROW);
393 AddKeyword ("true", Token.TRUE);
394 AddKeyword ("try", Token.TRY);
395 AddKeyword ("typeof", Token.TYPEOF);
396 AddKeyword ("uint", Token.UINT);
397 AddKeyword ("ulong", Token.ULONG);
398 AddKeyword ("unchecked", Token.UNCHECKED);
399 AddKeyword ("unsafe", Token.UNSAFE);
400 AddKeyword ("ushort", Token.USHORT);
401 AddKeyword ("using", Token.USING);
402 AddKeyword ("virtual", Token.VIRTUAL);
403 AddKeyword ("void", Token.VOID);
404 AddKeyword ("volatile", Token.VOLATILE);
405 AddKeyword ("while", Token.WHILE);
406 AddKeyword ("partial", Token.PARTIAL);
407 #if GMCS_SOURCE
408 AddKeyword ("where", Token.WHERE);
409 #endif
410 }
411
412 //
413 // Class initializer
414 //
415 static Tokenizer ()
416 {
417 InitTokens ();
418 csharp_format_info = NumberFormatInfo.InvariantInfo;
419 styles = NumberStyles.Float;
420
421 string_builder = new System.Text.StringBuilder ();
422 }
423
424 int GetKeyword (char[] id, int id_len)
425 {
426 /*
427 * Keywords are stored in an array of hashtables grouped by their
428 * length.
429 */
430
431 if ((id_len >= keywords.Length) || (keywords [id_len] == null))
432 return -1;
433 object o = keywords [id_len] [id];
434
435 if (o == null)
436 return -1;
437
438 int res = (int) o;
439
440 if (handle_get_set == false && (res == Token.GET || res == Token.SET))
441 return -1;
442 if (handle_remove_add == false && (res == Token.REMOVE || res == Token.ADD))
443 return -1;
444 if (handle_assembly == false && res == Token.ASSEMBLY)
445 return -1;
446 #if GMCS_SOURCE
447 if (handle_constraints == false && res == Token.WHERE)
448 return -1;
449 #endif
450 return res;
451
452 }
453
454 public Location Location {
455 get { return current_location; }
456 }
457
458 void define (string def)
459 {
460 if (!RootContext.AllDefines.Contains (def)){
461 RootContext.AllDefines [def] = true;
462 }
463 if (defines.Contains (def))
464 return;
465 defines [def] = true;
466 }
467
468 public Tokenizer (SeekableStreamReader input, SourceFile file, ArrayList defs)
469 {
470 this.ref_name = file;
471 this.file_name = file;
472 reader = input;
473
474 putback_char = -1;
475
476 if (defs != null){
477 defines = new Hashtable ();
478 foreach (string def in defs)
479 define (def);
480 }
481
482 xml_comment_buffer = new StringBuilder ();
483
484 //
485 // FIXME: This could be `Location.Push' but we have to
486 // find out why the MS compiler allows this
487 //
488 Mono.CSharp.Location.Push (file);
489 }
490
491 static bool is_identifier_start_character (char c)
492 {
493 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || Char.IsLetter (c);
494 }
495
496 static bool is_identifier_part_character (char c)
497 {
498 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || (c >= '0' && c <= '9') || Char.IsLetter (c);
499 }
500
501 public static bool IsKeyword (string s)
502 {
503 return keywordStrings [s] != null;
504 }
505
506 public static bool IsValidIdentifier (string s)
507 {
508 if (s == null || s.Length == 0)
509 return false;
510
511 if (!is_identifier_start_character (s [0]))
512 return false;
513
514 for (int i = 1; i < s.Length; i ++)
515 if (! is_identifier_part_character (s [i]))
516 return false;
517
518 return true;
519 }
520
521 #if GMCS_SOURCE
522 bool parse_generic_dimension (out int dimension)
523 {
524 dimension = 1;
525
526 again:
527 int the_token = token ();
528 if (the_token == Token.OP_GENERICS_GT)
529 return true;
530 else if (the_token == Token.COMMA) {
531 dimension++;
532 goto again;
533 }
534
535 return false;
536 }
537
538 bool parse_less_than ()
539 {
540 start:
541 int the_token = token ();
542 if (the_token == Token.OPEN_BRACKET) {
543 do {
544 the_token = token ();
545 } while (the_token != Token.CLOSE_BRACKET);
546 the_token = token ();
547 }
548 switch (the_token) {
549 case Token.IDENTIFIER:
550 case Token.OBJECT:
551 case Token.STRING:
552 case Token.BOOL:
553 case Token.DECIMAL:
554 case Token.FLOAT:
555 case Token.DOUBLE:
556 case Token.SBYTE:
557 case Token.BYTE:
558 case Token.SHORT:
559 case Token.USHORT:
560 case Token.INT:
561 case Token.UINT:
562 case Token.LONG:
563 case Token.ULONG:
564 case Token.CHAR:
565 case Token.VOID:
566 break;
567
568 default:
569 return false;
570 }
571 again:
572 the_token = token ();
573
574 if (the_token == Token.OP_GENERICS_GT)
575 return true;
576 else if ((the_token == Token.COMMA) || (the_token == Token.DOT))
577 goto start;
578 else if (the_token == Token.INTERR || the_token == Token.STAR)
579 goto again;
580 else if (the_token == Token.OP_GENERICS_LT) {
581 if (!parse_less_than ())
582 return false;
583 goto again;
584 } else if (the_token == Token.OPEN_BRACKET) {
585 rank_specifiers:
586 the_token = token ();
587 if (the_token == Token.CLOSE_BRACKET)
588 goto again;
589 else if (the_token == Token.COMMA)
590 goto rank_specifiers;
591 return false;
592 }
593
594 return false;
595 }
596
597 int parsing_generic_less_than = 0;
598
599 public void PutbackNullable ()
600 {
601 if (nullable_pos < 0)
602 throw new Exception ();
603
604 current_token = -1;
605 val = null;
606 reader.Position = nullable_pos;
607
608 putback_char = '?';
609 }
610
611 public void PutbackCloseParens ()
612 {
613 putback_char = ')';
614 }
615
616
617 int nullable_pos = -1;
618
619 public void CheckNullable (bool is_nullable)
620 {
621 if (is_nullable)
622 nullable_pos = reader.Position;
623 else
624 nullable_pos = -1;
625 }
626 #endif
627 int is_punct (char c, ref bool doread)
628 {
629 int d;
630 int t;
631
632 doread = false;
633
634 switch (c){
635 case '{':
636 val = Location;
637 return Token.OPEN_BRACE;
638 case '}':
639 val = Location;
640 return Token.CLOSE_BRACE;
641 case '[':
642 // To block doccomment inside attribute declaration.
643 if (doc_state == XmlCommentState.Allowed)
644 doc_state = XmlCommentState.NotAllowed;
645 return Token.OPEN_BRACKET;
646 case ']':
647 return Token.CLOSE_BRACKET;
648 case '(':
649 return Token.OPEN_PARENS;
650 case ')': {
651 if (deambiguate_close_parens == 0)
652 return Token.CLOSE_PARENS;
653
654 --deambiguate_close_parens;
655
656 PushPosition ();
657
658 // disable preprocessing directives when peeking
659 process_directives = false;
660 int new_token = xtoken ();
661 process_directives = true;
662
663 PopPosition ();
664
665 if (new_token == Token.OPEN_PARENS)
666 return Token.CLOSE_PARENS_OPEN_PARENS;
667 else if (new_token == Token.MINUS)
668 return Token.CLOSE_PARENS_MINUS;
669 else if (IsCastToken (new_token))
670 return Token.CLOSE_PARENS_CAST;
671 else
672 return Token.CLOSE_PARENS_NO_CAST;
673 }
674
675 case ',':
676 return Token.COMMA;
677 case ';':
678 val = Location;
679 return Token.SEMICOLON;
680 case '~':
681 val = Location;
682 return Token.TILDE;
683 case '?':
684 return Token.INTERR;
685 }
686 #if GMCS_SOURCE
687 if (c == '<') {
688 if (parsing_generic_less_than++ > 0)
689 return Token.OP_GENERICS_LT;
690
691 if (handle_typeof) {
692 int dimension;
693 PushPosition ();
694 if (parse_generic_dimension (out dimension)) {
695 val = dimension;
696 DiscardPosition ();
697 return Token.GENERIC_DIMENSION;
698 }
699 PopPosition ();
700 }
701
702 // Save current position and parse next token.
703 PushPosition ();
704 bool is_generic_lt = parse_less_than ();
705 PopPosition ();
706
707 if (is_generic_lt) {
708 parsing_generic_less_than++;
709 return Token.OP_GENERICS_LT;
710 } else
711 parsing_generic_less_than = 0;
712
713 d = peekChar ();
714 if (d == '<'){
715 getChar ();
716 d = peekChar ();
717
718 if (d == '='){
719 doread = true;
720 return Token.OP_SHIFT_LEFT_ASSIGN;
721 }
722 return Token.OP_SHIFT_LEFT;
723 } else if (d == '='){
724 doread = true;
725 return Token.OP_LE;
726 }
727 return Token.OP_LT;
728 } else if (c == '>') {
729 if (parsing_generic_less_than > 0) {
730 parsing_generic_less_than--;
731 return Token.OP_GENERICS_GT;
732 }
733
734 d = peekChar ();
735 if (d == '>'){
736 getChar ();
737 d = peekChar ();
738
739 if (d == '='){
740 doread = true;
741 return Token.OP_SHIFT_RIGHT_ASSIGN;
742 }
743 return Token.OP_SHIFT_RIGHT;
744 } else if (d == '='){
745 doread = true;
746 return Token.OP_GE;
747 }
748 return Token.OP_GT;
749 }
750 #endif
751 d = peekChar ();
752 if (c == '+'){
753
754 if (d == '+') {
755 val = Location;
756 t = Token.OP_INC;
757 }
758 else if (d == '=')
759 t = Token.OP_ADD_ASSIGN;
760 else {
761 val = Location;
762 return Token.PLUS;
763 }
764 doread = true;
765 return t;
766 }
767 if (c == '-'){
768 if (d == '-') {
769 val = Location;
770 t = Token.OP_DEC;
771 }
772 else if (d == '=')
773 t = Token.OP_SUB_ASSIGN;
774 else if (d == '>')
775 t = Token.OP_PTR;
776 else {
777 val = Location;
778 return Token.MINUS;
779 }
780 doread = true;
781 return t;
782 }
783
784 if (c == '!'){
785 if (d == '='){
786 doread = true;
787 return Token.OP_NE;
788 }
789 val = Location;
790 return Token.BANG;
791 }
792
793 if (c == '='){
794 if (d == '='){
795 doread = true;
796 return Token.OP_EQ;
797 }
798 return Token.ASSIGN;
799 }
800
801 if (c == '&'){
802 if (d == '&'){
803 doread = true;
804 return Token.OP_AND;
805 } else if (d == '='){
806 doread = true;
807 return Token.OP_AND_ASSIGN;
808 }
809 val = Location;
810 return Token.BITWISE_AND;
811 }
812
813 if (c == '|'){
814 if (d == '|'){
815 doread = true;
816 return Token.OP_OR;
817 } else if (d == '='){
818 doread = true;
819 return Token.OP_OR_ASSIGN;
820 }
821 return Token.BITWISE_OR;
822 }
823
824 if (c == '*'){
825 if (d == '='){
826 doread = true;
827 return Token.OP_MULT_ASSIGN;
828 }
829 val = Location;
830 return Token.STAR;
831 }
832
833 if (c == '/'){
834 if (d == '='){
835 doread = true;
836 return Token.OP_DIV_ASSIGN;
837 }
838 return Token.DIV;
839 }
840
841 if (c == '%'){
842 if (d == '='){
843 doread = true;
844 return Token.OP_MOD_ASSIGN;
845 }
846 return Token.PERCENT;
847 }
848
849 if (c == '^'){
850 if (d == '='){
851 doread = true;
852 return Token.OP_XOR_ASSIGN;
853 }
854 return Token.CARRET;
855 }
856
857 #if !GMCS_SOURCE
858 if (c == '<'){
859 if (d == '<'){
860 getChar ();
861 d = peekChar ();
862
863 if (d == '='){
864 doread = true;
865 return Token.OP_SHIFT_LEFT_ASSIGN;
866 }
867 return Token.OP_SHIFT_LEFT;
868 } else if (d == '='){
869 doread = true;
870 return Token.OP_LE;
871 }
872 return Token.OP_LT;
873 }
874
875 if (c == '>'){
876 if (d == '>'){
877 getChar ();
878 d = peekChar ();
879
880 if (d == '='){
881 doread = true;
882 return Token.OP_SHIFT_RIGHT_ASSIGN;
883 }
884 return Token.OP_SHIFT_RIGHT;
885 } else if (d == '='){
886 doread = true;
887 return Token.OP_GE;
888 }
889 return Token.OP_GT;
890 }
891 #endif
892 if (c == ':'){
893 if (d == ':'){
894 doread = true;
895 return Token.DOUBLE_COLON;
896 }
897 val = Location;
898 return Token.COLON;
899 }
900
901 return Token.ERROR;
902 }
903
904 int deambiguate_close_parens = 0;
905
906 public void Deambiguate_CloseParens (object expression)
907 {
908 putback (')');
909
910 // When any binary operation is used we are sure it is not a cast
911 if (expression is Binary)
912 return;
913
914 deambiguate_close_parens++;
915 }
916
917 bool decimal_digits (int c)
918 {
919 int d;
920 bool seen_digits = false;
921
922 if (c != -1){
923 if (number_pos == max_number_size)
924 Error_NumericConstantTooLong ();
925 number_builder [number_pos++] = (char) c;
926 }
927
928 //
929 // We use peekChar2, because decimal_digits needs to do a
930 // 2-character look-ahead (5.ToString for example).
931 //
932 while ((d = peekChar2 ()) != -1){
933 if (d >= '0' && d <= '9'){
934 if (number_pos == max_number_size)
935 Error_NumericConstantTooLong ();
936 number_builder [number_pos++] = (char) d;
937 getChar ();
938 seen_digits = true;
939 } else
940 break;
941 }
942
943 return seen_digits;
944 }
945
946 static bool is_hex (int e)
947 {
948 return (e >= '0' && e <= '9') || (e >= 'A' && e <= 'F') || (e >= 'a' && e <= 'f');
949 }
950
951 static int real_type_suffix (int c)
952 {
953 int t;
954
955 switch (c){
956 case 'F': case 'f':
957 t = Token.LITERAL_FLOAT;
958 break;
959 case 'D': case 'd':
960 t = Token.LITERAL_DOUBLE;
961 break;
962 case 'M': case 'm':
963 t= Token.LITERAL_DECIMAL;
964 break;
965 default:
966 return Token.NONE;
967 }
968 return t;
969 }
970
971 int integer_type_suffix (ulong ul, int c)
972 {
973 bool is_unsigned = false;
974 bool is_long = false;
975
976 if (c != -1){
977 bool scanning = true;
978 do {
979 switch (c){
980 case 'U': case 'u':
981 if (is_unsigned)
982 scanning = false;
983 is_unsigned = true;
984 getChar ();
985 break;
986
987 case 'l':
988 if (!is_unsigned && (RootContext.WarningLevel >= 4)){
989 //
990 // if we have not seen anything in between
991 // report this error
992 //
993 Report.Warning (78, 4, Location, "The 'l' suffix is easily confused with the digit '1' (use 'L' for clarity)");
994 }
995 //
996 // This goto statement causes the MS CLR 2.0 beta 1 csc to report an error, so
997 // work around that.
998 //
999 //goto case 'L';
1000 if (is_long)
1001 scanning = false;
1002 is_long = true;
1003 getChar ();
1004 break;
1005
1006 case 'L':
1007 if (is_long)
1008 scanning = false;
1009 is_long = true;
1010 getChar ();
1011 break;
1012
1013 default:
1014 scanning = false;
1015 break;
1016 }
1017 c = peekChar ();
1018 } while (scanning);
1019 }
1020
1021 if (is_long && is_unsigned){
1022 val = ul;
1023 return Token.LITERAL_INTEGER;
1024 } else if (is_unsigned){
1025 // uint if possible, or ulong else.
1026
1027 if ((ul & 0xffffffff00000000) == 0)
1028 val = (uint) ul;
1029 else
1030 val = ul;
1031 } else if (is_long){
1032 // long if possible, ulong otherwise
1033 if ((ul & 0x8000000000000000) != 0)
1034 val = ul;
1035 else
1036 val = (long) ul;
1037 } else {
1038 // int, uint, long or ulong in that order
1039 if ((ul & 0xffffffff00000000) == 0){
1040 uint ui = (uint) ul;
1041
1042 if ((ui & 0x80000000) != 0)
1043 val = ui;
1044 else
1045 val = (int) ui;
1046 } else {
1047 if ((ul & 0x8000000000000000) != 0)
1048 val = ul;
1049 else
1050 val = (long) ul;
1051 }
1052 }
1053 return Token.LITERAL_INTEGER;
1054 }
1055
1056 //
1057 // given `c' as the next char in the input decide whether
1058 // we need to convert to a special type, and then choose
1059 // the best representation for the integer
1060 //
1061 int adjust_int (int c)
1062 {
1063 try {
1064 if (number_pos > 9){
1065 ulong ul = (uint) (number_builder [0] - '0');
1066
1067 for (int i = 1; i < number_pos; i++){
1068 ul = checked ((ul * 10) + ((uint)(number_builder [i] - '0')));
1069 }
1070 return integer_type_suffix (ul, c);
1071 } else {
1072 uint ui = (uint) (number_builder [0] - '0');
1073
1074 for (int i = 1; i < number_pos; i++){
1075 ui = checked ((ui * 10) + ((uint)(number_builder [i] - '0')));
1076 }
1077 return integer_type_suffix (ui, c);
1078 }
1079 } catch (OverflowException) {
1080 error_details = "Integral constant is too large";
1081 Report.Error (1021, Location, error_details);
1082 val = 0ul;
1083 return Token.LITERAL_INTEGER;
1084 }
1085 catch (FormatException) {
1086 Report.Error (1013, Location, "Invalid number");
1087 val = 0ul;
1088 return Token.LITERAL_INTEGER;
1089 }
1090 }
1091
1092 int adjust_real (int t)
1093 {
1094 string s = new String (number_builder, 0, number_pos);
1095 const string error_details = "Floating-point constant is outside the range of type `{0}'";
1096
1097 switch (t){
1098 case Token.LITERAL_DECIMAL:
1099 try {
1100 val = System.Decimal.Parse (s, styles, csharp_format_info);
1101 } catch (OverflowException) {
1102 val = 0m;
1103 Report.Error (594, Location, error_details, "decimal");
1104 }
1105 break;
1106 case Token.LITERAL_FLOAT:
1107 try {
1108 val = float.Parse (s, styles, csharp_format_info);
1109 } catch (OverflowException) {
1110 val = 0.0f;
1111 Report.Error (594, Location, error_details, "float");
1112 }
1113 break;
1114
1115 case Token.LITERAL_DOUBLE:
1116 case Token.NONE:
1117 t = Token.LITERAL_DOUBLE;
1118 try {
1119 val = System.Double.Parse (s, styles, csharp_format_info);
1120 } catch (OverflowException) {
1121 val = 0.0;
1122 Report.Error (594, Location, error_details, "double");
1123 }
1124 break;
1125 }
1126 return t;
1127 }
1128
1129 int handle_hex ()
1130 {
1131 int d;
1132 ulong ul;
1133
1134 getChar ();
1135 while ((d = peekChar ()) != -1){
1136 if (is_hex (d)){
1137 number_builder [number_pos++] = (char) d;
1138 getChar ();
1139 } else
1140 break;
1141 }
1142
1143 string s = new String (number_builder, 0, number_pos);
1144 try {
1145 if (number_pos <= 8)
1146 ul = System.UInt32.Parse (s, NumberStyles.HexNumber);
1147 else
1148 ul = System.UInt64.Parse (s, NumberStyles.HexNumber);
1149 } catch (OverflowException){
1150 error_details = "Integral constant is too large";
1151 Report.Error (1021, Location, error_details);
1152 val = 0ul;
1153 return Token.LITERAL_INTEGER;
1154 }
1155 catch (FormatException) {
1156 Report.Error (1013, Location, "Invalid number");
1157 val = 0ul;
1158 return Token.LITERAL_INTEGER;
1159 }
1160
1161 return integer_type_suffix (ul, peekChar ());
1162 }
1163
1164 //
1165 // Invoked if we know we have .digits or digits
1166 //
1167 int is_number (int c)
1168 {
1169 bool is_real = false;
1170 int type;
1171
1172 number_pos = 0;
1173
1174 if (c >= '0' && c <= '9'){
1175 if (c == '0'){
1176 int peek = peekChar ();
1177
1178 if (peek == 'x' || peek == 'X')
1179 return handle_hex ();
1180 }
1181 decimal_digits (c);
1182 c = getChar ();
1183 }
1184
1185 //
1186 // We need to handle the case of
1187 // "1.1" vs "1.string" (LITERAL_FLOAT vs NUMBER DOT IDENTIFIER)
1188 //
1189 if (c == '.'){
1190 if (decimal_digits ('.')){
1191 is_real = true;
1192 c = getChar ();
1193 } else {
1194 putback ('.');
1195 number_pos--;
1196 return adjust_int (-1);
1197 }
1198 }
1199
1200 if (c == 'e' || c == 'E'){
1201 is_real = true;
1202 if (number_pos == max_number_size)
1203 Error_NumericConstantTooLong ();
1204 number_builder [number_pos++] = 'e';
1205 c = getChar ();
1206
1207 if (c == '+'){
1208 if (number_pos == max_number_size)
1209 Error_NumericConstantTooLong ();
1210 number_builder [number_pos++] = '+';
1211 c = -1;
1212 } else if (c == '-') {
1213 if (number_pos == max_number_size)
1214 Error_NumericConstantTooLong ();
1215 number_builder [number_pos++] = '-';
1216 c = -1;
1217 } else {
1218 if (number_pos == max_number_size)
1219 Error_NumericConstantTooLong ();
1220 number_builder [number_pos++] = '+';
1221 }
1222
1223 decimal_digits (c);
1224 c = getChar ();
1225 }
1226
1227 type = real_type_suffix (c);
1228 if (type == Token.NONE && !is_real){
1229 putback (c);
1230 return adjust_int (c);
1231 } else
1232 is_real = true;
1233
1234 if (type == Token.NONE){
1235 putback (c);
1236 }
1237
1238 if (is_real)
1239 return adjust_real (type);
1240
1241 Console.WriteLine ("This should not be reached");
1242 throw new Exception ("Is Number should never reach this point");
1243 }
1244
1245 //
1246 // Accepts exactly count (4 or 8) hex, no more no less
1247 //
1248 int getHex (int count, out bool error)
1249 {
1250 int i;
1251 int total = 0;
1252 int c;
1253 int top = count != -1 ? count : 4;
1254
1255 getChar ();
1256 error = false;
1257 for (i = 0; i < top; i++){
1258 c = getChar ();
1259
1260 if (c >= '0' && c <= '9')
1261 c = (int) c - (int) '0';
1262 else if (c >= 'A' && c <= 'F')
1263 c = (int) c - (int) 'A' + 10;
1264 else if (c >= 'a' && c <= 'f')
1265 c = (int) c - (int) 'a' + 10;
1266 else {
1267 error = true;
1268 return 0;
1269 }
1270
1271 total = (total * 16) + c;
1272 if (count == -1){
1273 int p = peekChar ();
1274 if (p == -1)
1275 break;
1276 if (!is_hex ((char)p))
1277 break;
1278 }
1279 }
1280 return total;
1281 }
1282
1283 int escape (int c)
1284 {
1285 bool error;
1286 int d;
1287 int v;
1288
1289 d = peekChar ();
1290 if (c != '\\')
1291 return c;
1292
1293 switch (d){
1294 case 'a':
1295 v = '\a'; break;
1296 case 'b':
1297 v = '\b'; break;
1298 case 'n':
1299 v = '\n'; break;
1300 case 't':
1301 v = '\t'; break;
1302 case 'v':
1303 v = '\v'; break;
1304 case 'r':
1305 v = '\r'; break;
1306 case '\\':
1307 v = '\\'; break;
1308 case 'f':
1309 v = '\f'; break;
1310 case '0':
1311 v = 0; break;
1312 case '"':
1313 v = '"'; break;
1314 case '\'':
1315 v = '\''; break;
1316 case 'x':
1317 v = getHex (-1, out error);
1318 if (error)
1319 goto default;
1320 return v;
1321 case 'u':
1322 v = getHex (4, out error);
1323 if (error)
1324 goto default;
1325 return v;
1326 case 'U':
1327 v = getHex (8, out error);
1328 if (error)
1329 goto default;
1330 return v;
1331 default:
1332 Report.Error (1009, Location, "Unrecognized escape sequence `\\{0}'", ((char)d).ToString ());
1333 return d;
1334 }
1335 getChar ();
1336 return v;
1337 }
1338
1339 int getChar ()
1340 {
1341 int x;
1342 if (putback_char != -1) {
1343 x = putback_char;
1344 putback_char = -1;
1345 } else
1346 x = reader.Read ();
1347 if (x == '\n') {
1348 line++;
1349 ref_line++;
1350 previous_col = col;
1351 col = 0;
1352 }
1353 else
1354 col++;
1355 return x;
1356 }
1357
1358 int peekChar ()
1359 {
1360 if (putback_char != -1)
1361 return putback_char;
1362 putback_char = reader.Read ();
1363 return putback_char;
1364 }
1365
1366 int peekChar2 ()
1367 {
1368 if (putback_char != -1)
1369 return putback_char;
1370 return reader.Peek ();
1371 }
1372
1373 void putback (int c)
1374 {
1375 if (putback_char != -1){
1376 Console.WriteLine ("Col: " + col);
1377 Console.WriteLine ("Row: " + line);
1378 Console.WriteLine ("Name: " + ref_name.Name);
1379 Console.WriteLine ("Current [{0}] putting back [{1}] ", putback_char, c);
1380 throw new Exception ("This should not happen putback on putback");
1381 }
1382 if (c == '\n' || col == 0) {
1383 // It won't happen though.
1384 line--;
1385 ref_line--;
1386 col = previous_col;
1387 }
1388 else
1389 col--;
1390 putback_char = c;
1391 }
1392
1393 public bool advance ()
1394 {
1395 return peekChar () != -1;
1396 }
1397
1398 public Object Value {
1399 get {
1400 return val;
1401 }
1402 }
1403
1404 public Object value ()
1405 {
1406 return val;
1407 }
1408
1409 static bool IsCastToken (int token)
1410 {
1411 switch (token) {
1412 case Token.BANG:
1413 case Token.TILDE:
1414 case Token.IDENTIFIER:
1415 case Token.LITERAL_INTEGER:
1416 case Token.LITERAL_FLOAT:
1417 case Token.LITERAL_DOUBLE:
1418 case Token.LITERAL_DECIMAL:
1419 case Token.LITERAL_CHARACTER:
1420 case Token.LITERAL_STRING:
1421 case Token.BASE:
1422 case Token.CHECKED:
1423 case Token.DELEGATE:
1424 case Token.FALSE:
1425 case Token.FIXED:
1426 case Token.NEW:
1427 case Token.NULL:
1428 case Token.SIZEOF:
1429 case Token.THIS:
1430 case Token.THROW:
1431 case Token.TRUE:
1432 case Token.TYPEOF:
1433 case Token.UNCHECKED:
1434 case Token.UNSAFE:
1435 #if GMCS_SOURCE
1436 case Token.DEFAULT:
1437 #endif
1438
1439 //
1440 // These can be part of a member access
1441 //
1442 case Token.INT:
1443 case Token.UINT:
1444 case Token.SHORT:
1445 case Token.USHORT:
1446 case Token.LONG:
1447 case Token.ULONG:
1448 case Token.DOUBLE:
1449 case Token.FLOAT:
1450 case Token.CHAR:
1451 return true;
1452
1453 default:
1454 return false;
1455 }
1456 }
1457
1458 public int token ()
1459 {
1460 current_token = xtoken ();
1461
1462 #if GMCS_SOURCE
1463 if (current_token != Token.DEFAULT)
1464 return current_token;
1465
1466 PushPosition();
1467 int c = xtoken();
1468 if (c == -1)
1469 current_token = Token.ERROR;
1470 else if (c == Token.OPEN_PARENS)
1471 current_token = Token.DEFAULT_OPEN_PARENS;
1472 else if (c == Token.COLON)
1473 current_token = Token.DEFAULT_COLON;
1474 else
1475 PopPosition();
1476 #endif
1477 return current_token;
1478 }
1479
1480 static StringBuilder static_cmd_arg = new System.Text.StringBuilder ();
1481
1482 void get_cmd_arg (out string cmd, out string arg)
1483 {
1484 int c;
1485
1486 tokens_seen = false;
1487 arg = "";
1488 static_cmd_arg.Length = 0;
1489
1490 // skip over white space
1491 while ((c = getChar ()) != -1 && (c != '\n') && ((c == '\r') || (c == ' ') || (c == '\t')))
1492 ;
1493
1494 while ((c != -1) && (c != '\n') && (c != ' ') && (c != '\t') && (c != '\r')){
1495 if (is_identifier_part_character ((char) c)){
1496 static_cmd_arg.Append ((char) c);
1497 c = getChar ();
1498 } else {
1499 putback (c);
1500 break;
1501 }
1502 }
1503
1504 cmd = static_cmd_arg.ToString ();
1505
1506 if (c == '\n'){
1507 return;
1508 }
1509
1510 // skip over white space
1511 while ((c = getChar ()) != -1 && (c != '\n') && ((c == '\r') || (c == ' ') || (c == '\t')))
1512 ;
1513
1514 if (c == '\n'){
1515 return;
1516 } else if (c == '\r'){
1517 return;
1518 } else if (c == -1){
1519 arg = "";
1520 return;
1521 }
1522
1523 static_cmd_arg.Length = 0;
1524 static_cmd_arg.Append ((char) c);
1525
1526 while ((c = getChar ()) != -1 && (c != '\n') && (c != '\r')){
1527 static_cmd_arg.Append ((char) c);
1528 }
1529
1530 arg = static_cmd_arg.ToString ().Trim ();
1531 }
1532
1533 //
1534 // Handles the #line directive
1535 //
1536 bool PreProcessLine (string arg)
1537 {
1538 if (arg.Length == 0)
1539 return false;
1540
1541 if (arg == "default"){
1542 ref_line = line;
1543 ref_name = file_name;
1544 Location.Push (ref_name);
1545 return true;
1546 } else if (arg == "hidden"){
1547 //
1548 // We ignore #line hidden
1549 //
1550 return true;
1551 }
1552
1553 try {
1554 int pos;
1555
1556 if ((pos = arg.IndexOf (' ')) != -1 && pos != 0){
1557 ref_line = System.Int32.Parse (arg.Substring (0, pos));
1558 pos++;
1559
1560 char [] quotes = { '\"' };
1561
1562 string name = arg.Substring (pos). Trim (quotes);
1563 ref_name = Location.LookupFile (name);
1564 file_name.HasLineDirective = true;
1565 ref_name.HasLineDirective = true;
1566 Location.Push (ref_name);
1567 } else {
1568 ref_line = System.Int32.Parse (arg);
1569 }
1570 } catch {
1571 return false;
1572 }
1573
1574 return true;
1575 }
1576
1577 //
1578 // Handles #define and #undef
1579 //
1580 void PreProcessDefinition (bool is_define, string arg)
1581 {
1582 if (arg.Length == 0 || arg == "true" || arg == "false"){
1583 Report.Error (1001, Location, "Missing identifer to pre-processor directive");
1584 return;
1585 }
1586
1587 char[] whitespace = { ' ', '\t' };
1588 if (arg.IndexOfAny (whitespace) != -1){
1589 Report.Error (1025, Location, "Single-line comment or end-of-line expected");
1590 return;
1591 }
1592
1593 if (!is_identifier_start_character (arg [0]))
1594 Report.Error (1001, Location, "Identifier expected: " + arg);
1595
1596 foreach (char c in arg.Substring (1)){
1597 if (!is_identifier_part_character (c)){
1598 Report.Error (1001, Location, "Identifier expected: " + arg);
1599 return;
1600 }
1601 }
1602
1603 if (is_define){
1604 if (defines == null)
1605 defines = new Hashtable ();
1606 define (arg);
1607 } else {
1608 if (defines == null)
1609 return;
1610 if (defines.Contains (arg))
1611 defines.Remove (arg);
1612 }
1613 }
1614
1615 /// <summary>
1616 /// Handles #pragma directive
1617 /// </summary>
1618 void PreProcessPragma (string arg)
1619 {
1620 const string warning = "warning";
1621 const string w_disable = "warning disable";
1622 const string w_restore = "warning restore";
1623
1624 if (arg == w_disable) {
1625 Report.RegisterWarningRegion (Location).WarningDisable (line);
1626 return;
1627 }
1628
1629 if (arg == w_restore) {
1630 Report.RegisterWarningRegion (Location).WarningEnable (line);
1631 return;
1632 }
1633
1634 if (arg.StartsWith (w_disable)) {
1635 int[] codes = ParseNumbers (arg.Substring (w_disable.Length));
1636 foreach (int code in codes) {
1637 if (code != 0)
1638 Report.RegisterWarningRegion (Location).WarningDisable (Location, code);
1639 }
1640 return;
1641 }
1642
1643 if (arg.StartsWith (w_restore)) {
1644 int[] codes = ParseNumbers (arg.Substring (w_restore.Length));
1645 Hashtable w_table = Report.warning_ignore_table;
1646 foreach (int code in codes) {
1647 if (w_table != null && w_table.Contains (code))
1648 Report.Warning (1635, 1, Location, String.Format ("Cannot restore warning `CS{0:0000}' because it was disabled globally", code));
1649 Report.RegisterWarningRegion (Location).WarningEnable (Location, code);
1650 }
1651 return;
1652 }
1653
1654 if (arg.StartsWith (warning)) {
1655 Report.Warning (1634, 1, Location, "Expected disable or restore");
1656 return;
1657 }
1658
1659 Report.Warning (1633, 1, Location, "Unrecognized #pragma directive");
1660 }
1661
1662 int[] ParseNumbers (string text)
1663 {
1664 string[] string_array = text.Split (',');
1665 int[] values = new int [string_array.Length];
1666 int index = 0;
1667 foreach (string string_code in string_array) {
1668 try {
1669 values[index++] = int.Parse (string_code, System.Globalization.CultureInfo.InvariantCulture);
1670 }
1671 catch (FormatException) {
1672 Report.Warning (1692, 1, Location, "Invalid number");
1673 }
1674 }
1675 return values;
1676 }
1677
1678 bool eval_val (string s)
1679 {
1680 if (s == "true")
1681 return true;
1682 if (s == "false")
1683 return false;
1684
1685 if (defines == null)
1686 return false;
1687 if (defines.Contains (s))
1688 return true;
1689
1690 return false;
1691 }
1692
1693 bool pp_primary (ref string s)
1694 {
1695 s = s.Trim ();
1696 int len = s.Length;
1697
1698 if (len > 0){
1699 char c = s [0];
1700
1701 if (c == '('){
1702 s = s.Substring (1);
1703 bool val = pp_expr (ref s);
1704 if (s.Length > 0 && s [0] == ')'){
1705 s = s.Substring (1);
1706 return val;
1707 }
1708 Error_InvalidDirective ();
1709 return false;
1710 }
1711
1712 if (is_identifier_start_character (c)){
1713 int j = 1;
1714
1715 while (j < len){
1716 c = s [j];
1717
1718 if (is_identifier_part_character (c)){
1719 j++;
1720 continue;
1721 }
1722 bool v = eval_val (s.Substring (0, j));
1723 s = s.Substring (j);
1724 return v;
1725 }
1726 bool vv = eval_val (s);
1727 s = "";
1728 return vv;
1729 }
1730 }
1731 Error_InvalidDirective ();
1732 return false;
1733 }
1734
1735 bool pp_unary (ref string s)
1736 {
1737 s = s.Trim ();
1738 int len = s.Length;
1739
1740 if (len > 0){
1741 if (s [0] == '!'){
1742 if (len > 1 && s [1] == '='){
1743 Error_InvalidDirective ();
1744 return false;
1745 }
1746 s = s.Substring (1);
1747 return ! pp_primary (ref s);
1748 } else
1749 return pp_primary (ref s);
1750 } else {
1751 Error_InvalidDirective ();
1752 return false;
1753 }
1754 }
1755
1756 bool pp_eq (ref string s)
1757 {
1758 bool va = pp_unary (ref s);
1759
1760 s = s.Trim ();
1761 int len = s.Length;
1762 if (len > 0){
1763 if (s [0] == '='){
1764 if (len > 2 && s [1] == '='){
1765 s = s.Substring (2);
1766 return va == pp_unary (ref s);
1767 } else {
1768 Error_InvalidDirective ();
1769 return false;
1770 }
1771 } else if (s [0] == '!' && len > 1 && s [1] == '='){
1772 s = s.Substring (2);
1773
1774 return va != pp_unary (ref s);
1775
1776 }
1777 }
1778
1779 return va;
1780
1781 }
1782
1783 bool pp_and (ref string s)
1784 {
1785 bool va = pp_eq (ref s);
1786
1787 s = s.Trim ();
1788 int len = s.Length;
1789 if (len > 0){
1790 if (s [0] == '&'){
1791 if (len > 2 && s [1] == '&'){
1792 s = s.Substring (2);
1793 return (va & pp_and (ref s));
1794 } else {
1795 Error_InvalidDirective ();
1796 return false;
1797 }
1798 }
1799 }
1800 return va;
1801 }
1802
1803 //
1804 // Evaluates an expression for `#if' or `#elif'
1805 //
1806 bool pp_expr (ref string s)
1807 {
1808 bool va = pp_and (ref s);
1809 s = s.Trim ();
1810 int len = s.Length;
1811 if (len > 0){
1812 char c = s [0];
1813
1814 if (c == '|'){
1815 if (len > 2 && s [1] == '|'){
1816 s = s.Substring (2);
1817 return va | pp_expr (ref s);
1818 } else {
1819 Error_InvalidDirective ();
1820 return false;
1821 }
1822 }
1823 }
1824
1825 return va;
1826 }
1827
1828 bool eval (string s)
1829 {
1830 bool v = pp_expr (ref s);
1831 s = s.Trim ();
1832 if (s.Length != 0){
1833 return false;
1834 }
1835
1836 return v;
1837 }
1838
1839 void Error_NumericConstantTooLong ()
1840 {
1841 Report.Error (1021, Location, "Numeric constant too long");
1842 }
1843
1844 void Error_InvalidDirective ()
1845 {
1846 Report.Error (1517, Location, "Invalid preprocessor directive");
1847 }
1848
1849 void Error_UnexpectedDirective (string extra)
1850 {
1851 Report.Error (
1852 1028, Location,
1853 "Unexpected processor directive (" + extra + ")");
1854 }
1855
1856 void Error_TokensSeen ()
1857 {
1858 Report.Error (1032, Location,
1859 "Cannot define or undefine preprocessor symbols after first token in file");
1860 }
1861
1862 //
1863 // Set to false to stop handling preprocesser directives
1864 //
1865 bool process_directives = true;
1866
1867 //
1868 // if true, then the code continues processing the code
1869 // if false, the code stays in a loop until another directive is
1870 // reached.
1871 //
1872 bool handle_preprocessing_directive (bool caller_is_taking)
1873 {
1874 string cmd, arg;
1875 bool region_directive = false;
1876
1877 get_cmd_arg (out cmd, out arg);
1878
1879 // Eat any trailing whitespaces and single-line comments
1880 if (arg.IndexOf ("//") != -1)
1881 arg = arg.Substring (0, arg.IndexOf ("//"));
1882 arg = arg.TrimEnd (' ', '\t');
1883
1884 //
1885 // The first group of pre-processing instructions is always processed
1886 //
1887 switch (cmd){
1888 case "pragma":
1889 if (RootContext.Version == LanguageVersion.ISO_1) {
1890 Report.FeatureIsNotStandardized (Location, "#pragma");
1891 return caller_is_taking;
1892 }
1893
1894 PreProcessPragma (arg);
1895 return caller_is_taking;
1896
1897 case "line":
1898 if (!PreProcessLine (arg))
1899 Report.Error (
1900 1576, Location,
1901 "The line number specified for #line directive is missing or invalid");
1902 return caller_is_taking;
1903
1904 case "region":
1905 region_directive = true;
1906 arg = "true";
1907 goto case "if";
1908
1909 case "endregion":
1910 region_directive = true;
1911 goto case "endif";
1912
1913 case "if":
1914 if (arg.Length == 0){
1915 Error_InvalidDirective ();
1916 return true;
1917 }
1918 bool taking = false;
1919 if (ifstack == null)
1920 ifstack = new Stack (2);
1921
1922 if (ifstack.Count == 0){
1923 taking = true;
1924 } else {
1925 int state = (int) ifstack.Peek ();
1926 if ((state & TAKING) != 0)
1927 taking = true;
1928 }
1929
1930 if (eval (arg) && taking){
1931 int push = TAKING | TAKEN_BEFORE | PARENT_TAKING;
1932 if (region_directive)
1933 push |= REGION;
1934 ifstack.Push (push);
1935 return true;
1936 } else {
1937 int push = (taking ? PARENT_TAKING : 0);
1938 if (region_directive)
1939 push |= REGION;
1940 ifstack.Push (push);
1941 return false;
1942 }
1943
1944 case "endif":
1945 if (ifstack == null || ifstack.Count == 0){
1946 Error_UnexpectedDirective ("no #if for this #endif");
1947 return true;
1948 } else {
1949 int pop = (int) ifstack.Pop ();
1950
1951 if (region_directive && ((pop & REGION) == 0))
1952 Report.Error (1027, Location, "Expected `#endif' directive");
1953 else if (!region_directive && ((pop & REGION) != 0))
1954 Report.Error (1038, Location, "#endregion directive expected");
1955
1956 if (!region_directive && arg.Length != 0) {
1957 Report.Error (1025, Location, "Single-line comment or end-of-line expected");
1958 }
1959
1960 if (ifstack.Count == 0)
1961 return true;
1962 else {
1963 int state = (int) ifstack.Peek ();
1964
1965 if ((state & TAKING) != 0)
1966 return true;
1967 else
1968 return false;
1969 }
1970 }
1971
1972 case "elif":
1973 if (ifstack == null || ifstack.Count == 0){
1974 Error_UnexpectedDirective ("no #if for this #elif");
1975 return true;
1976 } else {
1977 int state = (int) ifstack.Peek ();
1978
1979 if ((state & REGION) != 0) {
1980 Report.Error (1038, Location, "#endregion directive expected");
1981 return true;
1982 }
1983
1984 if ((state & ELSE_SEEN) != 0){
1985 Error_UnexpectedDirective ("#elif not valid after #else");
1986 return true;
1987 }
1988
1989 if ((state & (TAKEN_BEFORE | TAKING)) != 0)
1990 return false;
1991
1992 if (eval (arg) && ((state & PARENT_TAKING) != 0)){
1993 state = (int) ifstack.Pop ();
1994 ifstack.Push (state | TAKING | TAKEN_BEFORE);
1995 return true;
1996 } else
1997 return false;
1998 }
1999
2000 case "else":
2001 if (ifstack == null || ifstack.Count == 0){
2002 Error_UnexpectedDirective ("no #if for this #else");
2003 return true;
2004 } else {
2005 int state = (int) ifstack.Peek ();
2006
2007 if ((state & REGION) != 0) {
2008 Report.Error (1038, Location, "#endregion directive expected");
2009 return true;
2010 }
2011
2012 if ((state & ELSE_SEEN) != 0){
2013 Error_UnexpectedDirective ("#else within #else");
2014 return true;
2015 }
2016
2017 ifstack.Pop ();
2018
2019 bool ret;
2020 if ((state & TAKEN_BEFORE) == 0){
2021 ret = ((state & PARENT_TAKING) != 0);
2022 } else
2023 ret = false;
2024
2025 if (ret)
2026 state |= TAKING;
2027 else
2028 state &= ~TAKING;
2029
2030 ifstack.Push (state | ELSE_SEEN);
2031
2032 return ret;
2033 }
2034 }
2035
2036 //
2037 // These are only processed if we are in a `taking' block
2038 //
2039 if (!caller_is_taking)
2040 return false;
2041
2042 switch (cmd){
2043 case "define":
2044 if (any_token_seen){
2045 Error_TokensSeen ();
2046 return true;
2047 }
2048 PreProcessDefinition (true, arg);
2049 return true;
2050
2051 case "undef":
2052 if (any_token_seen){
2053 Error_TokensSeen ();
2054 return true;
2055 }
2056 PreProcessDefinition (false, arg);
2057 return true;
2058
2059 case "error":
2060 Report.Error (1029, Location, "#error: '" + arg + "'");
2061 return true;
2062
2063 case "warning":
2064 Report.Warning (1030, 1, Location, "#warning: `{0}'", arg);
2065 return true;
2066 }
2067
2068 Report.Error (1024, Location, "Wrong preprocessor directive");
2069 return true;
2070
2071 }
2072
2073 private int consume_string (bool quoted)
2074 {
2075 int c;
2076 string_builder.Length = 0;
2077
2078 while ((c = getChar ()) != -1){
2079 if (c == '"'){
2080 if (quoted && peekChar () == '"'){
2081 string_builder.Append ((char) c);
2082 getChar ();
2083 continue;
2084 } else {
2085 val = string_builder.ToString ();
2086 return Token.LITERAL_STRING;
2087 }
2088 }
2089
2090 if (c == '\n'){
2091 if (!quoted)
2092 Report.Error (1010, Location, "Newline in constant");
2093 }
2094
2095 if (!quoted){
2096 c = escape (c);
2097 if (c == -1)
2098 return Token.ERROR;
2099 }
2100 string_builder.Append ((char) c);
2101 }
2102
2103 Report.Error (1039, Location, "Unterminated string literal");
2104 return Token.EOF;
2105 }
2106
2107 private int consume_identifier (int s)
2108 {
2109 int res = consume_identifier (s, false);
2110
2111 if (doc_state == XmlCommentState.Allowed)
2112 doc_state = XmlCommentState.NotAllowed;
2113 switch (res) {
2114 case Token.USING:
2115 case Token.NAMESPACE:
2116 check_incorrect_doc_comment ();
2117 break;
2118 }
2119
2120 if (res == Token.PARTIAL) {
2121 // Save current position and parse next token.
2122 PushPosition ();
2123
2124 int next_token = token ();
2125 bool ok = (next_token == Token.CLASS) ||
2126 (next_token == Token.STRUCT) ||
2127 (next_token == Token.INTERFACE);
2128
2129 PopPosition ();
2130
2131 if (ok)
2132 return res;
2133
2134 // Report.Error (267, Location, "The `partial' modifier can be used only immediately before keyword `class', `struct', or `interface'");
2135 val = new LocatedToken (Location, "partial");
2136 return Token.IDENTIFIER;
2137 }
2138
2139 return res;
2140 }
2141
2142 private int consume_identifier (int s, bool quoted)
2143 {
2144 int pos = 1;
2145 int c = -1;
2146
2147 id_builder [0] = (char) s;
2148
2149 current_location = new Location (ref_line, Col);
2150
2151 while ((c = getChar ()) != -1) {
2152 loop:
2153 if (is_identifier_part_character ((char) c)){
2154 if (pos == max_id_size){
2155 Report.Error (645, Location, "Identifier too long (limit is 512 chars)");
2156 return Token.ERROR;
2157 }
2158
2159 id_builder [pos++] = (char) c;
2160 // putback_char = -1;
2161 } else if (c == '\\') {
2162 c = escape (c);
2163 goto loop;
2164 } else {
2165 // putback_char = c;
2166 putback (c);
2167 break;
2168 }
2169 }
2170
2171 //
2172 // Optimization: avoids doing the keyword lookup
2173 // on uppercase letters and _
2174 //
2175 if (!quoted && (s >= 'a' || s == '_')){
2176 int keyword = GetKeyword (id_builder, pos);
2177 if (keyword != -1) {
2178 val = Location;
2179 return keyword;
2180 }
2181 }
2182
2183 //
2184 // Keep identifiers in an array of hashtables to avoid needless
2185 // allocations
2186 //
2187
2188 if (identifiers [pos] != null) {
2189 val = identifiers [pos][id_builder];
2190 if (val != null) {
2191 val = new LocatedToken (Location, (string) val);
2192 if (quoted)
2193 escapedIdentifiers.Add (val);
2194 return Token.IDENTIFIER;
2195 }
2196 }
2197 else
2198 identifiers [pos] = new CharArrayHashtable (pos);
2199
2200 val = new String (id_builder, 0, pos);
2201 if (RootContext.Version == LanguageVersion.ISO_1) {
2202 for (int i = 1; i < id_builder.Length; i += 3) {
2203 if (id_builder [i] == '_' && (id_builder [i - 1] == '_' || id_builder [i + 1] == '_')) {
2204 Report.Error (1638, Location,
2205 "`{0}': Any identifier with double underscores cannot be used when ISO language version mode is specified", val.ToString ());
2206 break;
2207 }
2208 }
2209 }
2210
2211 char [] chars = new char [pos];
2212 Array.Copy (id_builder, chars, pos);
2213
2214 identifiers [pos] [chars] = val;
2215
2216 val = new LocatedToken (Location, (string) val);
2217 if (quoted)
2218 escapedIdentifiers.Add (val);
2219 return Token.IDENTIFIER;
2220 }
2221
2222 public int xtoken ()
2223 {
2224 int t;
2225 bool doread = false;
2226 int c;
2227
2228 // Whether we have seen comments on the current line
2229 bool comments_seen = false;
2230
2231 val = null;
2232 // optimization: eliminate col and implement #directive semantic correctly.
2233 for (;(c = getChar ()) != -1;) {
2234 if (c == ' ')
2235 continue;
2236
2237 if (c == '\t') {
2238 continue;
2239 }
2240
2241 if (c == ' ' || c == '\f' || c == '\v' || c == 0xa0)
2242 continue;
2243
2244 if (c == '\r') {
2245 if (peekChar () == '\n')
2246 getChar ();
2247
2248 any_token_seen |= tokens_seen;
2249 tokens_seen = false;
2250 comments_seen = false;
2251 continue;
2252 }
2253
2254 // Handle double-slash comments.
2255 if (c == '/'){
2256 int d = peekChar ();
2257
2258 if (d == '/'){
2259 getChar ();
2260 if (RootContext.Documentation != null && peekChar () == '/') {
2261 getChar ();
2262 // Don't allow ////.
2263 if ((d = peekChar ()) != '/') {
2264 update_comment_location ();
2265 if (doc_state == XmlCommentState.Allowed)
2266 handle_one_line_xml_comment ();
2267 else if (doc_state == XmlCommentState.NotAllowed)
2268 warn_incorrect_doc_comment ();
2269 }
2270 }
2271 while ((d = getChar ()) != -1 && (d != '\n') && d != '\r')
2272 if (d == '\n'){
2273 }
2274 any_token_seen |= tokens_seen;
2275 tokens_seen = false;
2276 comments_seen = false;
2277 continue;
2278 } else if (d == '*'){
2279 getChar ();
2280 bool docAppend = false;
2281 if (RootContext.Documentation != null && peekChar () == '*') {
2282 getChar ();
2283 update_comment_location ();
2284 // But when it is /**/, just do nothing.
2285 if (peekChar () == '/') {
2286 getChar ();
2287 continue;
2288 }
2289 if (doc_state == XmlCommentState.Allowed)
2290 docAppend = true;
2291 else if (doc_state == XmlCommentState.NotAllowed)
2292 warn_incorrect_doc_comment ();
2293 }
2294
2295 int current_comment_start = 0;
2296 if (docAppend) {
2297 current_comment_start = xml_comment_buffer.Length;
2298 xml_comment_buffer.Append (Environment.NewLine);
2299 }
2300
2301 Location start_location = Location;
2302
2303 while ((d = getChar ()) != -1){
2304 if (d == '*' && peekChar () == '/'){
2305 getChar ();
2306 comments_seen = true;
2307 break;
2308 }
2309 if (docAppend)
2310 xml_comment_buffer.Append ((char) d);
2311
2312 if (d == '\n'){
2313 any_token_seen |= tokens_seen;
2314 tokens_seen = false;
2315 //
2316 // Reset 'comments_seen' just to be consistent.
2317 // It doesn't matter either way, here.
2318 //
2319 comments_seen = false;
2320 }
2321 }
2322 if (!comments_seen)
2323 Report.Error (1035, start_location, "End-of-file found, '*/' expected");
2324
2325 if (docAppend)
2326 update_formatted_doc_comment (current_comment_start);
2327 continue;
2328 }
2329 goto is_punct_label;
2330 }
2331
2332
2333 if (c == '\\' || is_identifier_start_character ((char)c)){
2334 tokens_seen = true;
2335 return consume_identifier (c);
2336 }
2337
2338 is_punct_label:
2339 current_location = new Location (ref_line, Col);
2340 if ((t = is_punct ((char)c, ref doread)) != Token.ERROR){
2341 tokens_seen = true;
2342 if (doread){
2343 getChar ();
2344 }
2345 return t;
2346 }
2347
2348 // white space
2349 if (c == '\n'){
2350 any_token_seen |= tokens_seen;
2351 tokens_seen = false;
2352 comments_seen = false;
2353 continue;
2354 }
2355
2356 if (c >= '0' && c <= '9'){
2357 tokens_seen = true;
2358 return is_number (c);
2359 }
2360
2361 if (c == '.'){
2362 tokens_seen = true;
2363 int peek = peekChar ();
2364 if (peek >= '0' && peek <= '9')
2365 return is_number (c);
2366 return Token.DOT;
2367 }
2368
2369 /* For now, ignore pre-processor commands */
2370 // FIXME: In C# the '#' is not limited to appear
2371 // on the first column.
2372 if (c == '#') {
2373 // return NONE if we're not processing directives (during token peeks)
2374 if (!process_directives)
2375 return Token.NONE;
2376
2377 bool cont = true;
2378 if (tokens_seen || comments_seen) {
2379 error_details = "Preprocessor directives must appear as the first" +
2380 " non-whitespace character on a line.";
2381
2382 Report.Error (1040, Location, error_details);
2383
2384 return Token.ERROR;
2385 }
2386
2387 start_again:
2388
2389 cont = handle_preprocessing_directive (cont);
2390
2391 if (cont){
2392 continue;
2393 }
2394
2395 bool skipping = false;
2396 for (;(c = getChar ()) != -1;){
2397 if (c == '\n'){
2398 skipping = false;
2399 } else if (c == ' ' || c == '\t' || c == '\v' || c == '\r' || c == 0xa0)
2400 continue;
2401 else if (c != '#')
2402 skipping = true;
2403 if (c == '#' && !skipping)
2404 goto start_again;
2405 }
2406 any_token_seen |= tokens_seen;
2407 tokens_seen = false;
2408 if (c == -1)
2409 Report.Error (1027, Location, "Expected `#endif' directive");
2410 continue;
2411 }
2412
2413 if (c == '"')
2414 return consume_string (false);
2415
2416 if (c == '\''){
2417 c = getChar ();
2418 tokens_seen = true;
2419 if (c == '\''){
2420 error_details = "Empty character literal";
2421 Report.Error (1011, Location, error_details);
2422 return Token.ERROR;
2423 }
2424 if (c == '\r' || c == '\n') {
2425 Report.Error (1010, Location, "Newline in constant");
2426 return Token.ERROR;
2427 }
2428 c = escape (c);
2429 if (c == -1)
2430 return Token.ERROR;
2431 val = new System.Char ();
2432 val = (char) c;
2433 c = getChar ();
2434
2435 if (c != '\''){
2436 error_details = "Too many characters in character literal";
2437 Report.Error (1012, Location, error_details);
2438
2439 // Try to recover, read until newline or next "'"
2440 while ((c = getChar ()) != -1){
2441 if (c == '\n'){
2442 break;
2443 }
2444 else if (c == '\'')
2445 break;
2446 }
2447 return Token.ERROR;
2448 }
2449 return Token.LITERAL_CHARACTER;
2450 }
2451
2452 if (c == '@') {
2453 c = getChar ();
2454 if (c == '"') {
2455 tokens_seen = true;
2456 return consume_string (true);
2457 } else if (is_identifier_start_character ((char) c)){
2458 return consume_identifier (c, true);
2459 } else {
2460 Report.Error (1646, Location, "Keyword, identifier, or string expected after verbatim specifier: @");
2461 }
2462 }
2463
2464 error_details = ((char)c).ToString ();
2465
2466 return Token.ERROR;
2467 }
2468
2469 return Token.EOF;
2470 }
2471
2472 //
2473 // Handles one line xml comment
2474 //
2475 private void handle_one_line_xml_comment ()
2476 {
2477 int c;
2478 while ((c = peekChar ()) == ' ')
2479 getChar (); // skip heading whitespaces.
2480 while ((c = peekChar ()) != -1 && c != '\n' && c != '\r') {
2481 xml_comment_buffer.Append ((char) getChar ());
2482 }
2483 if (c == '\r' || c == '\n')
2484 xml_comment_buffer.Append (Environment.NewLine);
2485 }
2486
2487 //
2488 // Remove heading "*" in Javadoc-like xml documentation.
2489 //
2490 private void update_formatted_doc_comment (int current_comment_start)
2491 {
2492 int length = xml_comment_buffer.Length - current_comment_start;
2493 string [] lines = xml_comment_buffer.ToString (
2494 current_comment_start,
2495 length).Replace ("\r", "").Split ('\n');
2496
2497 // The first line starts with /**, thus it is not target
2498 // for the format check.
2499 for (int i = 1; i < lines.Length; i++) {
2500 string s = lines [i];
2501 int idx = s.IndexOf ('*');
2502 string head = null;
2503 if (idx < 0) {
2504 if (i < lines.Length - 1)
2505 return;
2506 head = s;
2507 } else
2508 head = s.Substring (0, idx);
2509 foreach (char c in head)
2510 if (c != ' ')
2511 return;
2512 lines [i] = s.Substring (idx + 1);
2513 }
2514 xml_comment_buffer.Remove (current_comment_start, length);
2515 xml_comment_buffer.Insert (current_comment_start, String.Join (Environment.NewLine, lines));
2516 }
2517
2518 //
2519 // Updates current comment location.
2520 //
2521 private void update_comment_location ()
2522 {
2523 if (current_comment_location.IsNull) {
2524 // "-2" is for heading "//" or "/*"
2525 current_comment_location =
2526 new Location (ref_line, col - 2);
2527 }
2528 }
2529
2530 //
2531 // Checks if there was incorrect doc comments and raise
2532 // warnings.
2533 //
2534 public void check_incorrect_doc_comment ()
2535 {
2536 if (xml_comment_buffer.Length > 0)
2537 warn_incorrect_doc_comment ();
2538 }
2539
2540 //
2541 // Raises a warning when tokenizer found incorrect doccomment
2542 // markup.
2543 //
2544 private void warn_incorrect_doc_comment ()
2545 {
2546 if (doc_state != XmlCommentState.Error) {
2547 doc_state = XmlCommentState.Error;
2548 // in csc, it is 'XML comment is not placed on
2549 // a valid language element'. But that does not
2550 // make sense.
2551 Report.Warning (1587, 2, Location, "XML comment is not placed on a valid language element");
2552 }
2553 }
2554
2555 //
2556 // Consumes the saved xml comment lines (if any)
2557 // as for current target member or type.
2558 //
2559 public string consume_doc_comment ()
2560 {
2561 if (xml_comment_buffer.Length > 0) {
2562 string ret = xml_comment_buffer.ToString ();
2563 reset_doc_comment ();
2564 return ret;
2565 }
2566 return null;
2567 }
2568
2569 void reset_doc_comment ()
2570 {
2571 xml_comment_buffer.Length = 0;
2572 current_comment_location = Location.Null;
2573 }
2574
2575 public void cleanup ()
2576 {
2577 if (ifstack != null && ifstack.Count >= 1) {
2578 int state = (int) ifstack.Pop ();
2579 if ((state & REGION) != 0)
2580 Report.Error (1038, Location, "#endregion directive expected");
2581 // else
2582 // Report.Error (1027, Location, "Expected `#endif' directive");
2583 }
2584 }
2585 }
2586
2587 //
2588 // Indicates whether it accepts XML documentation or not.
2589 //
2590 public enum XmlCommentState {
2591 // comment is allowed in this state.
2592 Allowed,
2593 // comment is not allowed in this state.
2594 NotAllowed,
2595 // once comments appeared when it is NotAllowed, then the
2596 // state is changed to it, until the state is changed to
2597 // .Allowed.
2598 Error
2599 }
2600 }
2601
Mirror of the official Mono project git repository