?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
revert 144379
[mcs.git] / mcs / cs-tokenizer.cs
blobfeff8405e1833bd4d48f2906238a6ebd026e1eb0
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 // Dual licensed under the terms of the MIT X11 or GNU GPL
9 //
10 // Copyright 2001, 2002 Ximian, Inc (http://www.ximian.com)
11 // Copyright 2004-2008 Novell, Inc
12 //
13 //
14
15 using System;
16 using System.Text;
17 using System.Collections;
18 using System.IO;
19 using System.Globalization;
20 using System.Reflection;
21
22 namespace Mono.CSharp
23 {
24 /// <summary>
25 /// Tokenizer for C# source code.
26 /// </summary>
27
28 public class Tokenizer : yyParser.yyInput
29 {
30 class KeywordEntry
31 {
32 public readonly int Token;
33 public KeywordEntry Next;
34 public readonly char[] Value;
35
36 public KeywordEntry (string value, int token)
37 {
38 this.Value = value.ToCharArray ();
39 this.Token = token;
40 }
41 }
42
43 SeekableStreamReader reader;
44 SourceFile ref_name;
45 CompilationUnit file_name;
46 CompilerContext context;
47 bool hidden = false;
48 int ref_line = 1;
49 int line = 1;
50 int col = 0;
51 int previous_col;
52 int current_token;
53 bool handle_get_set = false;
54 bool handle_remove_add = false;
55 bool handle_where = false;
56 bool handle_typeof = false;
57 bool lambda_arguments_parsing;
58 Location current_comment_location = Location.Null;
59 ArrayList escaped_identifiers;
60 int parsing_generic_less_than;
61
62 //
63 // Used mainly for parser optimizations. Some expressions for instance
64 // can appear only in block (including initializer, base initializer)
65 // scope only
66 //
67 public int parsing_block;
68 internal bool query_parsing;
69
70 //
71 // When parsing type only, useful for ambiguous nullable types
72 //
73 public int parsing_type;
74
75 //
76 // Set when parsing generic declaration (type or method header)
77 //
78 public bool parsing_generic_declaration;
79
80 //
81 // The value indicates that we have not reach any declaration or
82 // namespace yet
83 //
84 public int parsing_declaration;
85
86 //
87 // The special character to inject on streams to trigger the EXPRESSION_PARSE
88 // token to be returned. It just happens to be a Unicode character that
89 // would never be part of a program (can not be an identifier).
90 //
91 // This character is only tested just before the tokenizer is about to report
92 // an error; So on the regular operation mode, this addition will have no
93 // impact on the tokenizer's performance.
94 //
95
96 public const int EvalStatementParserCharacter = 0x2190; // Unicode Left Arrow
97 public const int EvalCompilationUnitParserCharacter = 0x2191; // Unicode Arrow
98 public const int EvalUsingDeclarationsParserCharacter = 0x2192; // Unicode Arrow
99
100 //
101 // XML documentation buffer. The save point is used to divide
102 // comments on types and comments on members.
103 //
104 StringBuilder xml_comment_buffer;
105
106 //
107 // See comment on XmlCommentState enumeration.
108 //
109 XmlCommentState xml_doc_state = XmlCommentState.Allowed;
110
111 //
112 // Whether tokens have been seen on this line
113 //
114 bool tokens_seen = false;
115
116 //
117 // Set to true once the GENERATE_COMPLETION token has bee
118 // returned. This helps produce one GENERATE_COMPLETION,
119 // as many COMPLETE_COMPLETION as necessary to complete the
120 // AST tree and one final EOF.
121 //
122 bool generated;
123
124 //
125 // Whether a token has been seen on the file
126 // This is needed because `define' is not allowed to be used
127 // after a token has been seen.
128 //
129 bool any_token_seen = false;
130
131 static readonly char[] simple_whitespaces = new char[] { ' ', '\t' };
132
133 public bool PropertyParsing {
134 get { return handle_get_set; }
135 set { handle_get_set = value; }
136 }
137
138 public bool EventParsing {
139 get { return handle_remove_add; }
140 set { handle_remove_add = value; }
141 }
142
143 public bool ConstraintsParsing {
144 get { return handle_where; }
145 set { handle_where = value; }
146 }
147
148 public bool TypeOfParsing {
149 get { return handle_typeof; }
150 set { handle_typeof = value; }
151 }
152
153 public XmlCommentState doc_state {
154 get { return xml_doc_state; }
155 set {
156 if (value == XmlCommentState.Allowed) {
157 check_incorrect_doc_comment ();
158 reset_doc_comment ();
159 }
160 xml_doc_state = value;
161 }
162 }
163
164 //
165 // This is used to trigger completion generation on the parser
166 public bool CompleteOnEOF;
167
168 void AddEscapedIdentifier (LocatedToken lt)
169 {
170 if (escaped_identifiers == null)
171 escaped_identifiers = new ArrayList ();
172
173 escaped_identifiers.Add (lt);
174 }
175
176 public bool IsEscapedIdentifier (Location loc)
177 {
178 if (escaped_identifiers != null) {
179 foreach (LocatedToken lt in escaped_identifiers)
180 if (lt.Location.Equals (loc))
181 return true;
182 }
183
184 return false;
185 }
186
187 //
188 // Class variables
189 //
190 static KeywordEntry[][] keywords;
191 static Hashtable keyword_strings;
192 static NumberStyles styles;
193 static NumberFormatInfo csharp_format_info;
194
195 //
196 // Values for the associated token returned
197 //
198 internal int putback_char; // Used by repl only
199 Object val;
200
201 //
202 // Pre-processor
203 //
204 const int TAKING = 1;
205 const int ELSE_SEEN = 4;
206 const int PARENT_TAKING = 8;
207 const int REGION = 16;
208
209 //
210 // pre-processor if stack state:
211 //
212 Stack ifstack;
213
214 static System.Text.StringBuilder string_builder;
215
216 const int max_id_size = 512;
217 static char [] id_builder = new char [max_id_size];
218
219 static CharArrayHashtable [] identifiers = new CharArrayHashtable [max_id_size + 1];
220
221 const int max_number_size = 512;
222 static char [] number_builder = new char [max_number_size];
223 static int number_pos;
224
225 //
226 // Details about the error encoutered by the tokenizer
227 //
228 string error_details;
229
230 public string error {
231 get {
232 return error_details;
233 }
234 }
235
236 public int Line {
237 get {
238 return ref_line;
239 }
240 }
241
242 //
243 // This is used when the tokenizer needs to save
244 // the current position as it needs to do some parsing
245 // on its own to deamiguate a token in behalf of the
246 // parser.
247 //
248 Stack position_stack = new Stack (2);
249 class Position {
250 public int position;
251 public int line;
252 public int ref_line;
253 public int col;
254 public bool hidden;
255 public int putback_char;
256 public int previous_col;
257 public Stack ifstack;
258 public int parsing_generic_less_than;
259 public int current_token;
260
261 public Position (Tokenizer t)
262 {
263 position = t.reader.Position;
264 line = t.line;
265 ref_line = t.ref_line;
266 col = t.col;
267 hidden = t.hidden;
268 putback_char = t.putback_char;
269 previous_col = t.previous_col;
270 if (t.ifstack != null && t.ifstack.Count != 0)
271 ifstack = (Stack)t.ifstack.Clone ();
272 parsing_generic_less_than = t.parsing_generic_less_than;
273 current_token = t.current_token;
274 }
275 }
276
277 public void PushPosition ()
278 {
279 position_stack.Push (new Position (this));
280 }
281
282 public void PopPosition ()
283 {
284 Position p = (Position) position_stack.Pop ();
285
286 reader.Position = p.position;
287 ref_line = p.ref_line;
288 line = p.line;
289 col = p.col;
290 hidden = p.hidden;
291 putback_char = p.putback_char;
292 previous_col = p.previous_col;
293 ifstack = p.ifstack;
294 parsing_generic_less_than = p.parsing_generic_less_than;
295 current_token = p.current_token;
296 }
297
298 // Do not reset the position, ignore it.
299 public void DiscardPosition ()
300 {
301 position_stack.Pop ();
302 }
303
304 static void AddKeyword (string kw, int token)
305 {
306 keyword_strings.Add (kw, kw);
307
308 int length = kw.Length;
309 if (keywords [length] == null) {
310 keywords [length] = new KeywordEntry ['z' - '_' + 1];
311 }
312
313 int char_index = kw [0] - '_';
314 KeywordEntry kwe = keywords [length] [char_index];
315 if (kwe == null) {
316 keywords [length] [char_index] = new KeywordEntry (kw, token);
317 return;
318 }
319
320 while (kwe.Next != null) {
321 kwe = kwe.Next;
322 }
323
324 kwe.Next = new KeywordEntry (kw, token);
325 }
326
327 static void InitTokens ()
328 {
329 keyword_strings = new Hashtable ();
330
331 // 11 is the length of the longest keyword for now
332 keywords = new KeywordEntry [11] [];
333
334 AddKeyword ("__arglist", Token.ARGLIST);
335 AddKeyword ("abstract", Token.ABSTRACT);
336 AddKeyword ("as", Token.AS);
337 AddKeyword ("add", Token.ADD);
338 AddKeyword ("base", Token.BASE);
339 AddKeyword ("bool", Token.BOOL);
340 AddKeyword ("break", Token.BREAK);
341 AddKeyword ("byte", Token.BYTE);
342 AddKeyword ("case", Token.CASE);
343 AddKeyword ("catch", Token.CATCH);
344 AddKeyword ("char", Token.CHAR);
345 AddKeyword ("checked", Token.CHECKED);
346 AddKeyword ("class", Token.CLASS);
347 AddKeyword ("const", Token.CONST);
348 AddKeyword ("continue", Token.CONTINUE);
349 AddKeyword ("decimal", Token.DECIMAL);
350 AddKeyword ("default", Token.DEFAULT);
351 AddKeyword ("delegate", Token.DELEGATE);
352 AddKeyword ("do", Token.DO);
353 AddKeyword ("double", Token.DOUBLE);
354 AddKeyword ("else", Token.ELSE);
355 AddKeyword ("enum", Token.ENUM);
356 AddKeyword ("event", Token.EVENT);
357 AddKeyword ("explicit", Token.EXPLICIT);
358 AddKeyword ("extern", Token.EXTERN);
359 AddKeyword ("false", Token.FALSE);
360 AddKeyword ("finally", Token.FINALLY);
361 AddKeyword ("fixed", Token.FIXED);
362 AddKeyword ("float", Token.FLOAT);
363 AddKeyword ("for", Token.FOR);
364 AddKeyword ("foreach", Token.FOREACH);
365 AddKeyword ("goto", Token.GOTO);
366 AddKeyword ("get", Token.GET);
367 AddKeyword ("if", Token.IF);
368 AddKeyword ("implicit", Token.IMPLICIT);
369 AddKeyword ("in", Token.IN);
370 AddKeyword ("int", Token.INT);
371 AddKeyword ("interface", Token.INTERFACE);
372 AddKeyword ("internal", Token.INTERNAL);
373 AddKeyword ("is", Token.IS);
374 AddKeyword ("lock", Token.LOCK);
375 AddKeyword ("long", Token.LONG);
376 AddKeyword ("namespace", Token.NAMESPACE);
377 AddKeyword ("new", Token.NEW);
378 AddKeyword ("null", Token.NULL);
379 AddKeyword ("object", Token.OBJECT);
380 AddKeyword ("operator", Token.OPERATOR);
381 AddKeyword ("out", Token.OUT);
382 AddKeyword ("override", Token.OVERRIDE);
383 AddKeyword ("params", Token.PARAMS);
384 AddKeyword ("private", Token.PRIVATE);
385 AddKeyword ("protected", Token.PROTECTED);
386 AddKeyword ("public", Token.PUBLIC);
387 AddKeyword ("readonly", Token.READONLY);
388 AddKeyword ("ref", Token.REF);
389 AddKeyword ("remove", Token.REMOVE);
390 AddKeyword ("return", Token.RETURN);
391 AddKeyword ("sbyte", Token.SBYTE);
392 AddKeyword ("sealed", Token.SEALED);
393 AddKeyword ("set", Token.SET);
394 AddKeyword ("short", Token.SHORT);
395 AddKeyword ("sizeof", Token.SIZEOF);
396 AddKeyword ("stackalloc", Token.STACKALLOC);
397 AddKeyword ("static", Token.STATIC);
398 AddKeyword ("string", Token.STRING);
399 AddKeyword ("struct", Token.STRUCT);
400 AddKeyword ("switch", Token.SWITCH);
401 AddKeyword ("this", Token.THIS);
402 AddKeyword ("throw", Token.THROW);
403 AddKeyword ("true", Token.TRUE);
404 AddKeyword ("try", Token.TRY);
405 AddKeyword ("typeof", Token.TYPEOF);
406 AddKeyword ("uint", Token.UINT);
407 AddKeyword ("ulong", Token.ULONG);
408 AddKeyword ("unchecked", Token.UNCHECKED);
409 AddKeyword ("unsafe", Token.UNSAFE);
410 AddKeyword ("ushort", Token.USHORT);
411 AddKeyword ("using", Token.USING);
412 AddKeyword ("virtual", Token.VIRTUAL);
413 AddKeyword ("void", Token.VOID);
414 AddKeyword ("volatile", Token.VOLATILE);
415 AddKeyword ("while", Token.WHILE);
416 AddKeyword ("partial", Token.PARTIAL);
417 AddKeyword ("where", Token.WHERE);
418
419 // LINQ keywords
420 AddKeyword ("from", Token.FROM);
421 AddKeyword ("join", Token.JOIN);
422 AddKeyword ("on", Token.ON);
423 AddKeyword ("equals", Token.EQUALS);
424 AddKeyword ("select", Token.SELECT);
425 AddKeyword ("group", Token.GROUP);
426 AddKeyword ("by", Token.BY);
427 AddKeyword ("let", Token.LET);
428 AddKeyword ("orderby", Token.ORDERBY);
429 AddKeyword ("ascending", Token.ASCENDING);
430 AddKeyword ("descending", Token.DESCENDING);
431 AddKeyword ("into", Token.INTO);
432 }
433
434 //
435 // Class initializer
436 //
437 static Tokenizer ()
438 {
439 InitTokens ();
440 csharp_format_info = NumberFormatInfo.InvariantInfo;
441 styles = NumberStyles.Float;
442
443 string_builder = new System.Text.StringBuilder ();
444 }
445
446 int GetKeyword (char[] id, int id_len)
447 {
448 //
449 // Keywords are stored in an array of arrays grouped by their
450 // length and then by the first character
451 //
452 if (id_len >= keywords.Length || keywords [id_len] == null)
453 return -1;
454
455 int first_index = id [0] - '_';
456 if (first_index > 'z')
457 return -1;
458
459 KeywordEntry kwe = keywords [id_len] [first_index];
460 if (kwe == null)
461 return -1;
462
463 int res;
464 do {
465 res = kwe.Token;
466 for (int i = 1; i < id_len; ++i) {
467 if (id [i] != kwe.Value [i]) {
468 res = 0;
469 break;
470 }
471 }
472 kwe = kwe.Next;
473 } while (kwe != null && res == 0);
474
475 if (res == 0)
476 return -1;
477
478 int next_token;
479 switch (res) {
480 case Token.GET:
481 case Token.SET:
482 if (!handle_get_set)
483 res = -1;
484 break;
485 case Token.REMOVE:
486 case Token.ADD:
487 if (!handle_remove_add)
488 res = -1;
489 break;
490 case Token.EXTERN:
491 if (parsing_declaration == 0)
492 res = Token.EXTERN_ALIAS;
493 break;
494 case Token.DEFAULT:
495 if (peek_token () == Token.COLON) {
496 token ();
497 res = Token.DEFAULT_COLON;
498 }
499 break;
500 case Token.WHERE:
501 if (!handle_where && !query_parsing)
502 res = -1;
503 break;
504 case Token.FROM:
505 //
506 // A query expression is any expression that starts with `from identifier'
507 // followed by any token except ; , =
508 //
509 if (!query_parsing) {
510 if (lambda_arguments_parsing) {
511 res = -1;
512 break;
513 }
514
515 PushPosition ();
516 // HACK: to disable generics micro-parser, because PushPosition does not
517 // store identifiers array
518 parsing_generic_less_than = 1;
519 switch (xtoken ()) {
520 case Token.IDENTIFIER:
521 case Token.INT:
522 case Token.BOOL:
523 case Token.BYTE:
524 case Token.CHAR:
525 case Token.DECIMAL:
526 case Token.FLOAT:
527 case Token.LONG:
528 case Token.OBJECT:
529 case Token.STRING:
530 case Token.UINT:
531 case Token.ULONG:
532 next_token = xtoken ();
533 if (next_token == Token.SEMICOLON || next_token == Token.COMMA || next_token == Token.EQUALS)
534 goto default;
535
536 res = Token.FROM_FIRST;
537 query_parsing = true;
538 if (RootContext.Version <= LanguageVersion.ISO_2)
539 Report.FeatureIsNotAvailable (Location, "query expressions");
540 break;
541 case Token.VOID:
542 Expression.Error_VoidInvalidInTheContext (Location, Report);
543 break;
544 default:
545 PopPosition ();
546 // HACK: A token is not a keyword so we need to restore identifiers buffer
547 // which has been overwritten before we grabbed the identifier
548 id_builder [0] = 'f'; id_builder [1] = 'r'; id_builder [2] = 'o'; id_builder [3] = 'm';
549 return -1;
550 }
551 PopPosition ();
552 }
553 break;
554 case Token.JOIN:
555 case Token.ON:
556 case Token.EQUALS:
557 case Token.SELECT:
558 case Token.GROUP:
559 case Token.BY:
560 case Token.LET:
561 case Token.ORDERBY:
562 case Token.ASCENDING:
563 case Token.DESCENDING:
564 case Token.INTO:
565 if (!query_parsing)
566 res = -1;
567 break;
568
569 case Token.USING:
570 case Token.NAMESPACE:
571 // TODO: some explanation needed
572 check_incorrect_doc_comment ();
573 break;
574
575 case Token.PARTIAL:
576 if (parsing_block > 0) {
577 res = -1;
578 break;
579 }
580
581 // Save current position and parse next token.
582 PushPosition ();
583
584 next_token = token ();
585 bool ok = (next_token == Token.CLASS) ||
586 (next_token == Token.STRUCT) ||
587 (next_token == Token.INTERFACE) ||
588 (next_token == Token.VOID);
589
590 PopPosition ();
591
592 if (ok) {
593 if (next_token == Token.VOID) {
594 if (RootContext.Version == LanguageVersion.ISO_1 ||
595 RootContext.Version == LanguageVersion.ISO_2)
596 Report.FeatureIsNotAvailable (Location, "partial methods");
597 } else if (RootContext.Version == LanguageVersion.ISO_1)
598 Report.FeatureIsNotAvailable (Location, "partial types");
599
600 return res;
601 }
602
603 if (next_token < Token.LAST_KEYWORD) {
604 Report.Error (267, Location,
605 "The `partial' modifier can be used only immediately before `class', `struct', `interface', or `void' keyword");
606 return token ();
607 }
608
609 res = -1;
610 break;
611 }
612
613 return res;
614 }
615
616 public Location Location {
617 get {
618 return new Location (ref_line, hidden ? -1 : col);
619 }
620 }
621
622 public Tokenizer (SeekableStreamReader input, CompilationUnit file, CompilerContext ctx)
623 {
624 this.ref_name = file;
625 this.file_name = file;
626 this.context = ctx;
627 reader = input;
628
629 putback_char = -1;
630
631 xml_comment_buffer = new StringBuilder ();
632
633 //
634 // FIXME: This could be `Location.Push' but we have to
635 // find out why the MS compiler allows this
636 //
637 Mono.CSharp.Location.Push (file, file);
638 }
639
640 static bool is_identifier_start_character (int c)
641 {
642 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || Char.IsLetter ((char)c);
643 }
644
645 static bool is_identifier_part_character (char c)
646 {
647 if (c >= 'a' && c <= 'z')
648 return true;
649
650 if (c >= 'A' && c <= 'Z')
651 return true;
652
653 if (c == '_' || (c >= '0' && c <= '9'))
654 return true;
655
656 if (c < 0x80)
657 return false;
658
659 return Char.IsLetter (c) || Char.GetUnicodeCategory (c) == UnicodeCategory.ConnectorPunctuation;
660 }
661
662 public static bool IsKeyword (string s)
663 {
664 return keyword_strings [s] != null;
665 }
666
667 //
668 // Open parens micro parser. Detects both lambda and cast ambiguity.
669 //
670
671 int TokenizeOpenParens ()
672 {
673 int ptoken;
674 current_token = -1;
675
676 int bracket_level = 0;
677 bool is_type = false;
678 bool can_be_type = false;
679
680 while (true) {
681 ptoken = current_token;
682 token ();
683
684 switch (current_token) {
685 case Token.CLOSE_PARENS:
686 token ();
687
688 //
689 // Expression inside parens is lambda, (int i) =>
690 //
691 if (current_token == Token.ARROW) {
692 if (RootContext.Version <= LanguageVersion.ISO_2)
693 Report.FeatureIsNotAvailable (Location, "lambda expressions");
694
695 return Token.OPEN_PARENS_LAMBDA;
696 }
697
698 //
699 // Expression inside parens is single type, (int[])
700 //
701 if (is_type)
702 return Token.OPEN_PARENS_CAST;
703
704 //
705 // Expression is possible cast, look at next token, (T)null
706 //
707 if (can_be_type) {
708 switch (current_token) {
709 case Token.OPEN_PARENS:
710 case Token.BANG:
711 case Token.TILDE:
712 case Token.IDENTIFIER:
713 case Token.LITERAL_INTEGER:
714 case Token.LITERAL_FLOAT:
715 case Token.LITERAL_DOUBLE:
716 case Token.LITERAL_DECIMAL:
717 case Token.LITERAL_CHARACTER:
718 case Token.LITERAL_STRING:
719 case Token.BASE:
720 case Token.CHECKED:
721 case Token.DELEGATE:
722 case Token.FALSE:
723 case Token.FIXED:
724 case Token.NEW:
725 case Token.NULL:
726 case Token.SIZEOF:
727 case Token.THIS:
728 case Token.THROW:
729 case Token.TRUE:
730 case Token.TYPEOF:
731 case Token.UNCHECKED:
732 case Token.UNSAFE:
733 case Token.DEFAULT:
734
735 //
736 // These can be part of a member access
737 //
738 case Token.INT:
739 case Token.UINT:
740 case Token.SHORT:
741 case Token.USHORT:
742 case Token.LONG:
743 case Token.ULONG:
744 case Token.DOUBLE:
745 case Token.FLOAT:
746 case Token.CHAR:
747 case Token.BYTE:
748 case Token.DECIMAL:
749 case Token.BOOL:
750 return Token.OPEN_PARENS_CAST;
751 }
752 }
753 return Token.OPEN_PARENS;
754
755 case Token.DOT:
756 case Token.DOUBLE_COLON:
757 if (ptoken != Token.IDENTIFIER && ptoken != Token.OP_GENERICS_GT)
758 goto default;
759
760 continue;
761
762 case Token.IDENTIFIER:
763 switch (ptoken) {
764 case Token.DOT:
765 case Token.OP_GENERICS_LT:
766 case Token.COMMA:
767 case Token.DOUBLE_COLON:
768 case -1:
769 if (bracket_level == 0)
770 can_be_type = true;
771 continue;
772 default:
773 can_be_type = is_type = false;
774 continue;
775 }
776
777 case Token.OBJECT:
778 case Token.STRING:
779 case Token.BOOL:
780 case Token.DECIMAL:
781 case Token.FLOAT:
782 case Token.DOUBLE:
783 case Token.SBYTE:
784 case Token.BYTE:
785 case Token.SHORT:
786 case Token.USHORT:
787 case Token.INT:
788 case Token.UINT:
789 case Token.LONG:
790 case Token.ULONG:
791 case Token.CHAR:
792 case Token.VOID:
793 if (bracket_level == 0)
794 is_type = true;
795 continue;
796
797 case Token.COMMA:
798 if (bracket_level == 0) {
799 bracket_level = 100;
800 can_be_type = is_type = false;
801 }
802 continue;
803
804 case Token.OP_GENERICS_LT:
805 case Token.OPEN_BRACKET:
806 if (bracket_level++ == 0)
807 is_type = true;
808 continue;
809
810 case Token.OP_GENERICS_GT:
811 case Token.CLOSE_BRACKET:
812 --bracket_level;
813 continue;
814
815 case Token.INTERR_NULLABLE:
816 case Token.STAR:
817 if (bracket_level == 0)
818 is_type = true;
819 continue;
820
821 case Token.REF:
822 case Token.OUT:
823 can_be_type = is_type = false;
824 continue;
825
826 default:
827 return Token.OPEN_PARENS;
828 }
829 }
830 }
831
832 public static bool IsValidIdentifier (string s)
833 {
834 if (s == null || s.Length == 0)
835 return false;
836
837 if (!is_identifier_start_character (s [0]))
838 return false;
839
840 for (int i = 1; i < s.Length; i ++)
841 if (! is_identifier_part_character (s [i]))
842 return false;
843
844 return true;
845 }
846
847 bool parse_less_than ()
848 {
849 start:
850 int the_token = token ();
851 if (the_token == Token.OPEN_BRACKET) {
852 do {
853 the_token = token ();
854 } while (the_token != Token.CLOSE_BRACKET);
855 the_token = token ();
856 } else if (the_token == Token.IN || the_token == Token.OUT) {
857 the_token = token ();
858 }
859 switch (the_token) {
860 case Token.IDENTIFIER:
861 case Token.OBJECT:
862 case Token.STRING:
863 case Token.BOOL:
864 case Token.DECIMAL:
865 case Token.FLOAT:
866 case Token.DOUBLE:
867 case Token.SBYTE:
868 case Token.BYTE:
869 case Token.SHORT:
870 case Token.USHORT:
871 case Token.INT:
872 case Token.UINT:
873 case Token.LONG:
874 case Token.ULONG:
875 case Token.CHAR:
876 case Token.VOID:
877 break;
878 case Token.OP_GENERICS_GT:
879 return true;
880
881 default:
882 return false;
883 }
884 again:
885 the_token = token ();
886
887 if (the_token == Token.OP_GENERICS_GT)
888 return true;
889 else if (the_token == Token.COMMA || the_token == Token.DOT || the_token == Token.DOUBLE_COLON)
890 goto start;
891 else if (the_token == Token.INTERR_NULLABLE || the_token == Token.STAR)
892 goto again;
893 else if (the_token == Token.OP_GENERICS_LT) {
894 if (!parse_less_than ())
895 return false;
896 goto again;
897 } else if (the_token == Token.OPEN_BRACKET) {
898 rank_specifiers:
899 the_token = token ();
900 if (the_token == Token.CLOSE_BRACKET)
901 goto again;
902 else if (the_token == Token.COMMA)
903 goto rank_specifiers;
904 return false;
905 }
906
907 return false;
908 }
909
910 bool parse_generic_dimension (out int dimension)
911 {
912 dimension = 1;
913
914 again:
915 int the_token = token ();
916 if (the_token == Token.OP_GENERICS_GT)
917 return true;
918 else if (the_token == Token.COMMA) {
919 dimension++;
920 goto again;
921 }
922
923 return false;
924 }
925
926 public int peek_token ()
927 {
928 int the_token;
929
930 PushPosition ();
931 the_token = token ();
932 PopPosition ();
933
934 return the_token;
935 }
936
937 //
938 // Tonizes `?' using custom disambiguous rules to return one
939 // of following tokens: INTERR_NULLABLE, OP_COALESCING, INTERR
940 //
941 // Tricky expression look like:
942 //
943 // Foo ? a = x ? b : c;
944 //
945 int TokenizePossibleNullableType ()
946 {
947 if (parsing_block == 0 || parsing_type > 0)
948 return Token.INTERR_NULLABLE;
949
950 int d = peek_char ();
951 if (d == '?') {
952 get_char ();
953 return Token.OP_COALESCING;
954 }
955
956 switch (current_token) {
957 case Token.CLOSE_PARENS:
958 case Token.TRUE:
959 case Token.FALSE:
960 case Token.NULL:
961 case Token.LITERAL_INTEGER:
962 case Token.LITERAL_STRING:
963 return Token.INTERR;
964 }
965
966 if (d != ' ') {
967 if (d == ',' || d == ';' || d == '>')
968 return Token.INTERR_NULLABLE;
969 if (d == '*' || (d >= '0' && d <= '9'))
970 return Token.INTERR;
971 }
972
973 PushPosition ();
974 current_token = Token.NONE;
975 int next_token;
976 switch (xtoken ()) {
977 case Token.LITERAL_INTEGER:
978 case Token.LITERAL_STRING:
979 case Token.LITERAL_CHARACTER:
980 case Token.LITERAL_DECIMAL:
981 case Token.LITERAL_DOUBLE:
982 case Token.LITERAL_FLOAT:
983 case Token.TRUE:
984 case Token.FALSE:
985 case Token.NULL:
986 case Token.THIS:
987 case Token.NEW:
988 next_token = Token.INTERR;
989 break;
990
991 case Token.SEMICOLON:
992 case Token.COMMA:
993 case Token.CLOSE_PARENS:
994 case Token.OPEN_BRACKET:
995 case Token.OP_GENERICS_GT:
996 next_token = Token.INTERR_NULLABLE;
997 break;
998
999 default:
1000 next_token = -1;
1001 break;
1002 }
1003
1004 if (next_token == -1) {
1005 switch (xtoken ()) {
1006 case Token.COMMA:
1007 case Token.SEMICOLON:
1008 case Token.OPEN_BRACE:
1009 case Token.CLOSE_PARENS:
1010 case Token.IN:
1011 next_token = Token.INTERR_NULLABLE;
1012 break;
1013
1014 case Token.COLON:
1015 next_token = Token.INTERR;
1016 break;
1017
1018 default:
1019 int ntoken;
1020 int interrs = 1;
1021 int colons = 0;
1022 //
1023 // All shorcuts failed, do it hard way
1024 //
1025 while ((ntoken = xtoken ()) != Token.EOF) {
1026 if (ntoken == Token.SEMICOLON)
1027 break;
1028
1029 if (ntoken == Token.COLON) {
1030 if (++colons == interrs)
1031 break;
1032 continue;
1033 }
1034
1035 if (ntoken == Token.INTERR) {
1036 ++interrs;
1037 continue;
1038 }
1039 }
1040
1041 next_token = colons != interrs ? Token.INTERR_NULLABLE : Token.INTERR;
1042 break;
1043 }
1044 }
1045
1046 PopPosition ();
1047 return next_token;
1048 }
1049
1050 bool decimal_digits (int c)
1051 {
1052 int d;
1053 bool seen_digits = false;
1054
1055 if (c != -1){
1056 if (number_pos == max_number_size)
1057 Error_NumericConstantTooLong ();
1058 number_builder [number_pos++] = (char) c;
1059 }
1060
1061 //
1062 // We use peek_char2, because decimal_digits needs to do a
1063 // 2-character look-ahead (5.ToString for example).
1064 //
1065 while ((d = peek_char2 ()) != -1){
1066 if (d >= '0' && d <= '9'){
1067 if (number_pos == max_number_size)
1068 Error_NumericConstantTooLong ();
1069 number_builder [number_pos++] = (char) d;
1070 get_char ();
1071 seen_digits = true;
1072 } else
1073 break;
1074 }
1075
1076 return seen_digits;
1077 }
1078
1079 static bool is_hex (int e)
1080 {
1081 return (e >= '0' && e <= '9') || (e >= 'A' && e <= 'F') || (e >= 'a' && e <= 'f');
1082 }
1083
1084 static int real_type_suffix (int c)
1085 {
1086 int t;
1087
1088 switch (c){
1089 case 'F': case 'f':
1090 t = Token.LITERAL_FLOAT;
1091 break;
1092 case 'D': case 'd':
1093 t = Token.LITERAL_DOUBLE;
1094 break;
1095 case 'M': case 'm':
1096 t= Token.LITERAL_DECIMAL;
1097 break;
1098 default:
1099 return Token.NONE;
1100 }
1101 return t;
1102 }
1103
1104 int integer_type_suffix (ulong ul, int c)
1105 {
1106 bool is_unsigned = false;
1107 bool is_long = false;
1108
1109 if (c != -1){
1110 bool scanning = true;
1111 do {
1112 switch (c){
1113 case 'U': case 'u':
1114 if (is_unsigned)
1115 scanning = false;
1116 is_unsigned = true;
1117 get_char ();
1118 break;
1119
1120 case 'l':
1121 if (!is_unsigned){
1122 //
1123 // if we have not seen anything in between
1124 // report this error
1125 //
1126 Report.Warning (78, 4, Location, "The 'l' suffix is easily confused with the digit '1' (use 'L' for clarity)");
1127 }
1128 //
1129 // This goto statement causes the MS CLR 2.0 beta 1 csc to report an error, so
1130 // work around that.
1131 //
1132 //goto case 'L';
1133 if (is_long)
1134 scanning = false;
1135 is_long = true;
1136 get_char ();
1137 break;
1138
1139 case 'L':
1140 if (is_long)
1141 scanning = false;
1142 is_long = true;
1143 get_char ();
1144 break;
1145
1146 default:
1147 scanning = false;
1148 break;
1149 }
1150 c = peek_char ();
1151 } while (scanning);
1152 }
1153
1154 if (is_long && is_unsigned){
1155 val = ul;
1156 return Token.LITERAL_INTEGER;
1157 } else if (is_unsigned){
1158 // uint if possible, or ulong else.
1159
1160 if ((ul & 0xffffffff00000000) == 0)
1161 val = (uint) ul;
1162 else
1163 val = ul;
1164 } else if (is_long){
1165 // long if possible, ulong otherwise
1166 if ((ul & 0x8000000000000000) != 0)
1167 val = ul;
1168 else
1169 val = (long) ul;
1170 } else {
1171 // int, uint, long or ulong in that order
1172 if ((ul & 0xffffffff00000000) == 0){
1173 uint ui = (uint) ul;
1174
1175 if ((ui & 0x80000000) != 0)
1176 val = ui;
1177 else
1178 val = (int) ui;
1179 } else {
1180 if ((ul & 0x8000000000000000) != 0)
1181 val = ul;
1182 else
1183 val = (long) ul;
1184 }
1185 }
1186 return Token.LITERAL_INTEGER;
1187 }
1188
1189 //
1190 // given `c' as the next char in the input decide whether
1191 // we need to convert to a special type, and then choose
1192 // the best representation for the integer
1193 //
1194 int adjust_int (int c)
1195 {
1196 try {
1197 if (number_pos > 9){
1198 ulong ul = (uint) (number_builder [0] - '0');
1199
1200 for (int i = 1; i < number_pos; i++){
1201 ul = checked ((ul * 10) + ((uint)(number_builder [i] - '0')));
1202 }
1203 return integer_type_suffix (ul, c);
1204 } else {
1205 uint ui = (uint) (number_builder [0] - '0');
1206
1207 for (int i = 1; i < number_pos; i++){
1208 ui = checked ((ui * 10) + ((uint)(number_builder [i] - '0')));
1209 }
1210 return integer_type_suffix (ui, c);
1211 }
1212 } catch (OverflowException) {
1213 error_details = "Integral constant is too large";
1214 Report.Error (1021, Location, error_details);
1215 val = 0ul;
1216 return Token.LITERAL_INTEGER;
1217 }
1218 catch (FormatException) {
1219 Report.Error (1013, Location, "Invalid number");
1220 val = 0ul;
1221 return Token.LITERAL_INTEGER;
1222 }
1223 }
1224
1225 int adjust_real (int t)
1226 {
1227 string s = new String (number_builder, 0, number_pos);
1228 const string error_details = "Floating-point constant is outside the range of type `{0}'";
1229
1230 switch (t){
1231 case Token.LITERAL_DECIMAL:
1232 try {
1233 val = System.Decimal.Parse (s, styles, csharp_format_info);
1234 } catch (OverflowException) {
1235 val = 0m;
1236 Report.Error (594, Location, error_details, "decimal");
1237 }
1238 break;
1239 case Token.LITERAL_FLOAT:
1240 try {
1241 val = float.Parse (s, styles, csharp_format_info);
1242 } catch (OverflowException) {
1243 val = 0.0f;
1244 Report.Error (594, Location, error_details, "float");
1245 }
1246 break;
1247
1248 case Token.LITERAL_DOUBLE:
1249 case Token.NONE:
1250 t = Token.LITERAL_DOUBLE;
1251 try {
1252 val = System.Double.Parse (s, styles, csharp_format_info);
1253 } catch (OverflowException) {
1254 val = 0.0;
1255 Report.Error (594, Location, error_details, "double");
1256 }
1257 break;
1258 }
1259 return t;
1260 }
1261
1262 int handle_hex ()
1263 {
1264 int d;
1265 ulong ul;
1266
1267 get_char ();
1268 while ((d = peek_char ()) != -1){
1269 if (is_hex (d)){
1270 number_builder [number_pos++] = (char) d;
1271 get_char ();
1272 } else
1273 break;
1274 }
1275
1276 string s = new String (number_builder, 0, number_pos);
1277 try {
1278 if (number_pos <= 8)
1279 ul = System.UInt32.Parse (s, NumberStyles.HexNumber);
1280 else
1281 ul = System.UInt64.Parse (s, NumberStyles.HexNumber);
1282 } catch (OverflowException){
1283 error_details = "Integral constant is too large";
1284 Report.Error (1021, Location, error_details);
1285 val = 0ul;
1286 return Token.LITERAL_INTEGER;
1287 }
1288 catch (FormatException) {
1289 Report.Error (1013, Location, "Invalid number");
1290 val = 0ul;
1291 return Token.LITERAL_INTEGER;
1292 }
1293
1294 return integer_type_suffix (ul, peek_char ());
1295 }
1296
1297 //
1298 // Invoked if we know we have .digits or digits
1299 //
1300 int is_number (int c)
1301 {
1302 bool is_real = false;
1303 int type;
1304
1305 number_pos = 0;
1306
1307 if (c >= '0' && c <= '9'){
1308 if (c == '0'){
1309 int peek = peek_char ();
1310
1311 if (peek == 'x' || peek == 'X')
1312 return handle_hex ();
1313 }
1314 decimal_digits (c);
1315 c = get_char ();
1316 }
1317
1318 //
1319 // We need to handle the case of
1320 // "1.1" vs "1.string" (LITERAL_FLOAT vs NUMBER DOT IDENTIFIER)
1321 //
1322 if (c == '.'){
1323 if (decimal_digits ('.')){
1324 is_real = true;
1325 c = get_char ();
1326 } else {
1327 putback ('.');
1328 number_pos--;
1329 return adjust_int (-1);
1330 }
1331 }
1332
1333 if (c == 'e' || c == 'E'){
1334 is_real = true;
1335 if (number_pos == max_number_size)
1336 Error_NumericConstantTooLong ();
1337 number_builder [number_pos++] = 'e';
1338 c = get_char ();
1339
1340 if (c == '+'){
1341 if (number_pos == max_number_size)
1342 Error_NumericConstantTooLong ();
1343 number_builder [number_pos++] = '+';
1344 c = -1;
1345 } else if (c == '-') {
1346 if (number_pos == max_number_size)
1347 Error_NumericConstantTooLong ();
1348 number_builder [number_pos++] = '-';
1349 c = -1;
1350 } else {
1351 if (number_pos == max_number_size)
1352 Error_NumericConstantTooLong ();
1353 number_builder [number_pos++] = '+';
1354 }
1355
1356 decimal_digits (c);
1357 c = get_char ();
1358 }
1359
1360 type = real_type_suffix (c);
1361 if (type == Token.NONE && !is_real){
1362 putback (c);
1363 return adjust_int (c);
1364 } else
1365 is_real = true;
1366
1367 if (type == Token.NONE){
1368 putback (c);
1369 }
1370
1371 if (is_real)
1372 return adjust_real (type);
1373
1374 Console.WriteLine ("This should not be reached");
1375 throw new Exception ("Is Number should never reach this point");
1376 }
1377
1378 //
1379 // Accepts exactly count (4 or 8) hex, no more no less
1380 //
1381 int getHex (int count, out int surrogate, out bool error)
1382 {
1383 int i;
1384 int total = 0;
1385 int c;
1386 int top = count != -1 ? count : 4;
1387
1388 get_char ();
1389 error = false;
1390 surrogate = 0;
1391 for (i = 0; i < top; i++){
1392 c = get_char ();
1393
1394 if (c >= '0' && c <= '9')
1395 c = (int) c - (int) '0';
1396 else if (c >= 'A' && c <= 'F')
1397 c = (int) c - (int) 'A' + 10;
1398 else if (c >= 'a' && c <= 'f')
1399 c = (int) c - (int) 'a' + 10;
1400 else {
1401 error = true;
1402 return 0;
1403 }
1404
1405 total = (total * 16) + c;
1406 if (count == -1){
1407 int p = peek_char ();
1408 if (p == -1)
1409 break;
1410 if (!is_hex ((char)p))
1411 break;
1412 }
1413 }
1414
1415 if (top == 8) {
1416 if (total > 0x0010FFFF) {
1417 error = true;
1418 return 0;
1419 }
1420
1421 if (total >= 0x00010000) {
1422 surrogate = ((total - 0x00010000) % 0x0400 + 0xDC00);
1423 total = ((total - 0x00010000) / 0x0400 + 0xD800);
1424 }
1425 }
1426
1427 return total;
1428 }
1429
1430 int escape (int c, out int surrogate)
1431 {
1432 bool error;
1433 int d;
1434 int v;
1435
1436 d = peek_char ();
1437 if (c != '\\') {
1438 surrogate = 0;
1439 return c;
1440 }
1441
1442 switch (d){
1443 case 'a':
1444 v = '\a'; break;
1445 case 'b':
1446 v = '\b'; break;
1447 case 'n':
1448 v = '\n'; break;
1449 case 't':
1450 v = '\t'; break;
1451 case 'v':
1452 v = '\v'; break;
1453 case 'r':
1454 v = '\r'; break;
1455 case '\\':
1456 v = '\\'; break;
1457 case 'f':
1458 v = '\f'; break;
1459 case '0':
1460 v = 0; break;
1461 case '"':
1462 v = '"'; break;
1463 case '\'':
1464 v = '\''; break;
1465 case 'x':
1466 v = getHex (-1, out surrogate, out error);
1467 if (error)
1468 goto default;
1469 return v;
1470 case 'u':
1471 case 'U':
1472 return EscapeUnicode (d, out surrogate);
1473 default:
1474 surrogate = 0;
1475 Report.Error (1009, Location, "Unrecognized escape sequence `\\{0}'", ((char)d).ToString ());
1476 return d;
1477 }
1478
1479 get_char ();
1480 surrogate = 0;
1481 return v;
1482 }
1483
1484 int EscapeUnicode (int ch, out int surrogate)
1485 {
1486 bool error;
1487 if (ch == 'U') {
1488 ch = getHex (8, out surrogate, out error);
1489 } else {
1490 ch = getHex (4, out surrogate, out error);
1491 }
1492
1493 if (error)
1494 Report.Error (1009, Location, "Unrecognized escape sequence");
1495
1496 return ch;
1497 }
1498
1499 int get_char ()
1500 {
1501 int x;
1502 if (putback_char != -1) {
1503 x = putback_char;
1504 putback_char = -1;
1505 } else
1506 x = reader.Read ();
1507 if (x == '\n') {
1508 advance_line ();
1509 } else {
1510 col++;
1511 }
1512 return x;
1513 }
1514
1515 void advance_line ()
1516 {
1517 line++;
1518 ref_line++;
1519 previous_col = col;
1520 col = 0;
1521 }
1522
1523 int peek_char ()
1524 {
1525 if (putback_char == -1)
1526 putback_char = reader.Read ();
1527 return putback_char;
1528 }
1529
1530 int peek_char2 ()
1531 {
1532 if (putback_char != -1)
1533 return putback_char;
1534 return reader.Peek ();
1535 }
1536
1537 void putback (int c)
1538 {
1539 if (putback_char != -1){
1540 Console.WriteLine ("Col: " + col);
1541 Console.WriteLine ("Row: " + line);
1542 Console.WriteLine ("Name: " + ref_name.Name);
1543 Console.WriteLine ("Current [{0}] putting back [{1}] ", putback_char, c);
1544 throw new Exception ("This should not happen putback on putback");
1545 }
1546 if (c == '\n' || col == 0) {
1547 // It won't happen though.
1548 line--;
1549 ref_line--;
1550 col = previous_col;
1551 }
1552 else
1553 col--;
1554 putback_char = c;
1555 }
1556
1557 public bool advance ()
1558 {
1559 return peek_char () != -1 || CompleteOnEOF;
1560 }
1561
1562 public Object Value {
1563 get {
1564 return val;
1565 }
1566 }
1567
1568 public Object value ()
1569 {
1570 return val;
1571 }
1572
1573 public int token ()
1574 {
1575 current_token = xtoken ();
1576 return current_token;
1577 }
1578
1579 static StringBuilder static_cmd_arg = new System.Text.StringBuilder ();
1580
1581 void get_cmd_arg (out string cmd, out string arg)
1582 {
1583 int c;
1584
1585 tokens_seen = false;
1586 arg = "";
1587
1588 // skip over white space
1589 do {
1590 c = get_char ();
1591 } while (c == '\r' || c == ' ' || c == '\t');
1592
1593 static_cmd_arg.Length = 0;
1594 while (c != -1 && is_identifier_part_character ((char)c)) {
1595 static_cmd_arg.Append ((char)c);
1596 c = get_char ();
1597 if (c == '\\') {
1598 int peek = peek_char ();
1599 if (peek == 'U' || peek == 'u') {
1600 int surrogate;
1601 c = EscapeUnicode (c, out surrogate);
1602 if (surrogate != 0) {
1603 if (is_identifier_part_character ((char) c))
1604 static_cmd_arg.Append ((char) c);
1605 c = surrogate;
1606 }
1607 }
1608 }
1609 }
1610
1611 cmd = static_cmd_arg.ToString ();
1612
1613 // skip over white space
1614 while (c == '\r' || c == ' ' || c == '\t')
1615 c = get_char ();
1616
1617 static_cmd_arg.Length = 0;
1618 int has_identifier_argument = 0;
1619
1620 while (c != -1 && c != '\n' && c != '\r') {
1621 if (c == '\\' && has_identifier_argument >= 0) {
1622 if (has_identifier_argument != 0 || (cmd == "define" || cmd == "if" || cmd == "elif" || cmd == "undef")) {
1623 has_identifier_argument = 1;
1624
1625 int peek = peek_char ();
1626 if (peek == 'U' || peek == 'u') {
1627 int surrogate;
1628 c = EscapeUnicode (c, out surrogate);
1629 if (surrogate != 0) {
1630 if (is_identifier_part_character ((char) c))
1631 static_cmd_arg.Append ((char) c);
1632 c = surrogate;
1633 }
1634 }
1635 } else {
1636 has_identifier_argument = -1;
1637 }
1638 }
1639 static_cmd_arg.Append ((char) c);
1640 c = get_char ();
1641 }
1642
1643 if (static_cmd_arg.Length != 0)
1644 arg = static_cmd_arg.ToString ();
1645 }
1646
1647 //
1648 // Handles the #line directive
1649 //
1650 bool PreProcessLine (string arg)
1651 {
1652 if (arg.Length == 0)
1653 return false;
1654
1655 if (arg == "default"){
1656 ref_line = line;
1657 ref_name = file_name;
1658 hidden = false;
1659 Location.Push (file_name, ref_name);
1660 return true;
1661 } else if (arg == "hidden"){
1662 hidden = true;
1663 return true;
1664 }
1665
1666 try {
1667 int pos;
1668
1669 if ((pos = arg.IndexOf (' ')) != -1 && pos != 0){
1670 ref_line = System.Int32.Parse (arg.Substring (0, pos));
1671 pos++;
1672
1673 char [] quotes = { '\"' };
1674
1675 string name = arg.Substring (pos). Trim (quotes);
1676 ref_name = Location.LookupFile (file_name, name);
1677 file_name.AddFile (ref_name);
1678 hidden = false;
1679 Location.Push (file_name, ref_name);
1680 } else {
1681 ref_line = System.Int32.Parse (arg);
1682 hidden = false;
1683 }
1684 } catch {
1685 return false;
1686 }
1687
1688 return true;
1689 }
1690
1691 //
1692 // Handles #define and #undef
1693 //
1694 void PreProcessDefinition (bool is_define, string ident, bool caller_is_taking)
1695 {
1696 if (ident.Length == 0 || ident == "true" || ident == "false"){
1697 Report.Error (1001, Location, "Missing identifier to pre-processor directive");
1698 return;
1699 }
1700
1701 if (ident.IndexOfAny (simple_whitespaces) != -1){
1702 Error_EndLineExpected ();
1703 return;
1704 }
1705
1706 if (!is_identifier_start_character (ident [0]))
1707 Report.Error (1001, Location, "Identifier expected: {0}", ident);
1708
1709 foreach (char c in ident.Substring (1)){
1710 if (!is_identifier_part_character (c)){
1711 Report.Error (1001, Location, "Identifier expected: {0}", ident);
1712 return;
1713 }
1714 }
1715
1716 if (!caller_is_taking)
1717 return;
1718
1719 if (is_define) {
1720 //
1721 // #define ident
1722 //
1723 if (RootContext.IsConditionalDefined (ident))
1724 return;
1725
1726 file_name.AddDefine (ident);
1727 } else {
1728 //
1729 // #undef ident
1730 //
1731 file_name.AddUndefine (ident);
1732 }
1733 }
1734
1735 static byte read_hex (string arg, int pos, out bool error)
1736 {
1737 error = false;
1738
1739 int total;
1740 char c = arg [pos];
1741
1742 if ((c >= '0') && (c <= '9'))
1743 total = (int) c - (int) '0';
1744 else if ((c >= 'A') && (c <= 'F'))
1745 total = (int) c - (int) 'A' + 10;
1746 else if ((c >= 'a') && (c <= 'f'))
1747 total = (int) c - (int) 'a' + 10;
1748 else {
1749 error = true;
1750 return 0;
1751 }
1752
1753 total *= 16;
1754 c = arg [pos+1];
1755
1756 if ((c >= '0') && (c <= '9'))
1757 total += (int) c - (int) '0';
1758 else if ((c >= 'A') && (c <= 'F'))
1759 total += (int) c - (int) 'A' + 10;
1760 else if ((c >= 'a') && (c <= 'f'))
1761 total += (int) c - (int) 'a' + 10;
1762 else {
1763 error = true;
1764 return 0;
1765 }
1766
1767 return (byte) total;
1768 }
1769
1770 /// <summary>
1771 /// Handles #pragma checksum
1772 /// </summary>
1773 bool PreProcessPragmaChecksum (string arg)
1774 {
1775 if ((arg [0] != ' ') && (arg [0] != '\t'))
1776 return false;
1777
1778 arg = arg.Trim (simple_whitespaces);
1779 if ((arg.Length < 2) || (arg [0] != '"'))
1780 return false;
1781
1782 StringBuilder file_sb = new StringBuilder ();
1783
1784 int pos = 1;
1785 char ch;
1786 while ((ch = arg [pos++]) != '"') {
1787 if (pos >= arg.Length)
1788 return false;
1789
1790 if (ch == '\\') {
1791 if (pos+1 >= arg.Length)
1792 return false;
1793 ch = arg [pos++];
1794 }
1795
1796 file_sb.Append (ch);
1797 }
1798
1799 if ((pos+2 >= arg.Length) || ((arg [pos] != ' ') && (arg [pos] != '\t')))
1800 return false;
1801
1802 arg = arg.Substring (pos).Trim (simple_whitespaces);
1803 if ((arg.Length < 42) || (arg [0] != '"') || (arg [1] != '{') ||
1804 (arg [10] != '-') || (arg [15] != '-') || (arg [20] != '-') ||
1805 (arg [25] != '-') || (arg [38] != '}') || (arg [39] != '"'))
1806 return false;
1807
1808 bool error;
1809 byte[] guid_bytes = new byte [16];
1810
1811 for (int i = 0; i < 4; i++) {
1812 guid_bytes [i] = read_hex (arg, 2+2*i, out error);
1813 if (error)
1814 return false;
1815 }
1816 for (int i = 0; i < 2; i++) {
1817 guid_bytes [i+4] = read_hex (arg, 11+2*i, out error);
1818 if (error)
1819 return false;
1820 guid_bytes [i+6] = read_hex (arg, 16+2*i, out error);
1821 if (error)
1822 return false;
1823 guid_bytes [i+8] = read_hex (arg, 21+2*i, out error);
1824 if (error)
1825 return false;
1826 }
1827
1828 for (int i = 0; i < 6; i++) {
1829 guid_bytes [i+10] = read_hex (arg, 26+2*i, out error);
1830 if (error)
1831 return false;
1832 }
1833
1834 arg = arg.Substring (40).Trim (simple_whitespaces);
1835 if ((arg.Length < 34) || (arg [0] != '"') || (arg [33] != '"'))
1836 return false;
1837
1838 byte[] checksum_bytes = new byte [16];
1839 for (int i = 0; i < 16; i++) {
1840 checksum_bytes [i] = read_hex (arg, 1+2*i, out error);
1841 if (error)
1842 return false;
1843 }
1844
1845 arg = arg.Substring (34).Trim (simple_whitespaces);
1846 if (arg.Length > 0)
1847 return false;
1848
1849 SourceFile file = Location.LookupFile (file_name, file_sb.ToString ());
1850 file.SetChecksum (guid_bytes, checksum_bytes);
1851 ref_name.AutoGenerated = true;
1852 return true;
1853 }
1854
1855 /// <summary>
1856 /// Handles #pragma directive
1857 /// </summary>
1858 void PreProcessPragma (string arg)
1859 {
1860 const string warning = "warning";
1861 const string w_disable = "warning disable";
1862 const string w_restore = "warning restore";
1863 const string checksum = "checksum";
1864
1865 if (arg == w_disable) {
1866 Report.RegisterWarningRegion (Location).WarningDisable (Location.Row);
1867 return;
1868 }
1869
1870 if (arg == w_restore) {
1871 Report.RegisterWarningRegion (Location).WarningEnable (Location.Row);
1872 return;
1873 }
1874
1875 if (arg.StartsWith (w_disable)) {
1876 int[] codes = ParseNumbers (arg.Substring (w_disable.Length));
1877 foreach (int code in codes) {
1878 if (code != 0)
1879 Report.RegisterWarningRegion (Location).WarningDisable (Location, code, Report);
1880 }
1881 return;
1882 }
1883
1884 if (arg.StartsWith (w_restore)) {
1885 int[] codes = ParseNumbers (arg.Substring (w_restore.Length));
1886 Hashtable w_table = Report.warning_ignore_table;
1887 foreach (int code in codes) {
1888 if (w_table != null && w_table.Contains (code))
1889 Report.Warning (1635, 1, Location, "Cannot restore warning `CS{0:0000}' because it was disabled globally", code);
1890 Report.RegisterWarningRegion (Location).WarningEnable (Location, code, Report);
1891 }
1892 return;
1893 }
1894
1895 if (arg.StartsWith (warning)) {
1896 Report.Warning (1634, 1, Location, "Expected disable or restore");
1897 return;
1898 }
1899
1900 if (arg.StartsWith (checksum)) {
1901 if (!PreProcessPragmaChecksum (arg.Substring (checksum.Length)))
1902 Warning_InvalidPragmaChecksum ();
1903 return;
1904 }
1905
1906 Report.Warning (1633, 1, Location, "Unrecognized #pragma directive");
1907 }
1908
1909 int[] ParseNumbers (string text)
1910 {
1911 string[] string_array = text.Split (',');
1912 int[] values = new int [string_array.Length];
1913 int index = 0;
1914 foreach (string string_code in string_array) {
1915 try {
1916 values[index++] = int.Parse (string_code, System.Globalization.CultureInfo.InvariantCulture);
1917 }
1918 catch (FormatException) {
1919 Report.Warning (1692, 1, Location, "Invalid number");
1920 }
1921 }
1922 return values;
1923 }
1924
1925 bool eval_val (string s)
1926 {
1927 if (s == "true")
1928 return true;
1929 if (s == "false")
1930 return false;
1931
1932 return file_name.IsConditionalDefined (s);
1933 }
1934
1935 bool pp_primary (ref string s)
1936 {
1937 s = s.Trim ();
1938 int len = s.Length;
1939
1940 if (len > 0){
1941 char c = s [0];
1942
1943 if (c == '('){
1944 s = s.Substring (1);
1945 bool val = pp_expr (ref s, false);
1946 if (s.Length > 0 && s [0] == ')'){
1947 s = s.Substring (1);
1948 return val;
1949 }
1950 Error_InvalidDirective ();
1951 return false;
1952 }
1953
1954 if (is_identifier_start_character (c)){
1955 int j = 1;
1956
1957 while (j < len){
1958 c = s [j];
1959
1960 if (is_identifier_part_character (c)){
1961 j++;
1962 continue;
1963 }
1964 bool v = eval_val (s.Substring (0, j));
1965 s = s.Substring (j);
1966 return v;
1967 }
1968 bool vv = eval_val (s);
1969 s = "";
1970 return vv;
1971 }
1972 }
1973 Error_InvalidDirective ();
1974 return false;
1975 }
1976
1977 bool pp_unary (ref string s)
1978 {
1979 s = s.Trim ();
1980 int len = s.Length;
1981
1982 if (len > 0){
1983 if (s [0] == '!'){
1984 if (len > 1 && s [1] == '='){
1985 Error_InvalidDirective ();
1986 return false;
1987 }
1988 s = s.Substring (1);
1989 return ! pp_primary (ref s);
1990 } else
1991 return pp_primary (ref s);
1992 } else {
1993 Error_InvalidDirective ();
1994 return false;
1995 }
1996 }
1997
1998 bool pp_eq (ref string s)
1999 {
2000 bool va = pp_unary (ref s);
2001
2002 s = s.Trim ();
2003 int len = s.Length;
2004 if (len > 0){
2005 if (s [0] == '='){
2006 if (len > 2 && s [1] == '='){
2007 s = s.Substring (2);
2008 return va == pp_unary (ref s);
2009 } else {
2010 Error_InvalidDirective ();
2011 return false;
2012 }
2013 } else if (s [0] == '!' && len > 1 && s [1] == '='){
2014 s = s.Substring (2);
2015
2016 return va != pp_unary (ref s);
2017
2018 }
2019 }
2020
2021 return va;
2022
2023 }
2024
2025 bool pp_and (ref string s)
2026 {
2027 bool va = pp_eq (ref s);
2028
2029 s = s.Trim ();
2030 int len = s.Length;
2031 if (len > 0){
2032 if (s [0] == '&'){
2033 if (len > 2 && s [1] == '&'){
2034 s = s.Substring (2);
2035 return (va & pp_and (ref s));
2036 } else {
2037 Error_InvalidDirective ();
2038 return false;
2039 }
2040 }
2041 }
2042 return va;
2043 }
2044
2045 //
2046 // Evaluates an expression for `#if' or `#elif'
2047 //
2048 bool pp_expr (ref string s, bool isTerm)
2049 {
2050 bool va = pp_and (ref s);
2051 s = s.Trim ();
2052 int len = s.Length;
2053 if (len > 0){
2054 char c = s [0];
2055
2056 if (c == '|'){
2057 if (len > 2 && s [1] == '|'){
2058 s = s.Substring (2);
2059 return va | pp_expr (ref s, isTerm);
2060 } else {
2061 Error_InvalidDirective ();
2062 return false;
2063 }
2064 }
2065 if (isTerm) {
2066 Error_EndLineExpected ();
2067 return false;
2068 }
2069 }
2070
2071 return va;
2072 }
2073
2074 bool eval (string s)
2075 {
2076 bool v = pp_expr (ref s, true);
2077 s = s.Trim ();
2078 if (s.Length != 0){
2079 return false;
2080 }
2081
2082 return v;
2083 }
2084
2085 void Error_NumericConstantTooLong ()
2086 {
2087 Report.Error (1021, Location, "Numeric constant too long");
2088 }
2089
2090 void Error_InvalidDirective ()
2091 {
2092 Report.Error (1517, Location, "Invalid preprocessor directive");
2093 }
2094
2095 void Error_UnexpectedDirective (string extra)
2096 {
2097 Report.Error (
2098 1028, Location,
2099 "Unexpected processor directive ({0})", extra);
2100 }
2101
2102 void Error_TokensSeen ()
2103 {
2104 Report.Error (1032, Location,
2105 "Cannot define or undefine preprocessor symbols after first token in file");
2106 }
2107
2108 void Eror_WrongPreprocessorLocation ()
2109 {
2110 Report.Error (1040, Location,
2111 "Preprocessor directives must appear as the first non-whitespace character on a line");
2112 }
2113
2114 void Error_EndLineExpected ()
2115 {
2116 Report.Error (1025, Location, "Single-line comment or end-of-line expected");
2117 }
2118
2119 void Warning_InvalidPragmaChecksum ()
2120 {
2121 Report.Warning (1695, 1, Location,
2122 "Invalid #pragma checksum syntax; should be " +
2123 "#pragma checksum \"filename\" " +
2124 "\"{XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX}\" \"XXXX...\"");
2125 }
2126 //
2127 // if true, then the code continues processing the code
2128 // if false, the code stays in a loop until another directive is
2129 // reached.
2130 // When caller_is_taking is false we ignore all directives except the ones
2131 // which can help us to identify where the #if block ends
2132 bool handle_preprocessing_directive (bool caller_is_taking)
2133 {
2134 string cmd, arg;
2135 bool region_directive = false;
2136
2137 get_cmd_arg (out cmd, out arg);
2138
2139 // Eat any trailing whitespaces and single-line comments
2140 if (arg.IndexOf ("//") != -1)
2141 arg = arg.Substring (0, arg.IndexOf ("//"));
2142 arg = arg.Trim (simple_whitespaces);
2143
2144 //
2145 // The first group of pre-processing instructions is always processed
2146 //
2147 switch (cmd){
2148 case "region":
2149 region_directive = true;
2150 arg = "true";
2151 goto case "if";
2152
2153 case "endregion":
2154 if (ifstack == null || ifstack.Count == 0){
2155 Error_UnexpectedDirective ("no #region for this #endregion");
2156 return true;
2157 }
2158 int pop = (int) ifstack.Pop ();
2159
2160 if ((pop & REGION) == 0)
2161 Report.Error (1027, Location, "Expected `#endif' directive");
2162
2163 return caller_is_taking;
2164
2165 case "if":
2166 if (ifstack == null)
2167 ifstack = new Stack (2);
2168
2169 int flags = region_directive ? REGION : 0;
2170 if (ifstack.Count == 0){
2171 flags |= PARENT_TAKING;
2172 } else {
2173 int state = (int) ifstack.Peek ();
2174 if ((state & TAKING) != 0) {
2175 flags |= PARENT_TAKING;
2176 }
2177 }
2178
2179 if (caller_is_taking && eval (arg)) {
2180 ifstack.Push (flags | TAKING);
2181 return true;
2182 }
2183 ifstack.Push (flags);
2184 return false;
2185
2186 case "endif":
2187 if (ifstack == null || ifstack.Count == 0){
2188 Error_UnexpectedDirective ("no #if for this #endif");
2189 return true;
2190 } else {
2191 pop = (int) ifstack.Pop ();
2192
2193 if ((pop & REGION) != 0)
2194 Report.Error (1038, Location, "#endregion directive expected");
2195
2196 if (arg.Length != 0) {
2197 Error_EndLineExpected ();
2198 }
2199
2200 if (ifstack.Count == 0)
2201 return true;
2202
2203 int state = (int) ifstack.Peek ();
2204 return (state & TAKING) != 0;
2205 }
2206
2207 case "elif":
2208 if (ifstack == null || ifstack.Count == 0){
2209 Error_UnexpectedDirective ("no #if for this #elif");
2210 return true;
2211 } else {
2212 int state = (int) ifstack.Pop ();
2213
2214 if ((state & REGION) != 0) {
2215 Report.Error (1038, Location, "#endregion directive expected");
2216 return true;
2217 }
2218
2219 if ((state & ELSE_SEEN) != 0){
2220 Error_UnexpectedDirective ("#elif not valid after #else");
2221 return true;
2222 }
2223
2224 if ((state & TAKING) != 0) {
2225 ifstack.Push (0);
2226 return false;
2227 }
2228
2229 if (eval (arg) && ((state & PARENT_TAKING) != 0)){
2230 ifstack.Push (state | TAKING);
2231 return true;
2232 }
2233
2234 ifstack.Push (state);
2235 return false;
2236 }
2237
2238 case "else":
2239 if (ifstack == null || ifstack.Count == 0){
2240 Error_UnexpectedDirective ("no #if for this #else");
2241 return true;
2242 } else {
2243 int state = (int) ifstack.Peek ();
2244
2245 if ((state & REGION) != 0) {
2246 Report.Error (1038, Location, "#endregion directive expected");
2247 return true;
2248 }
2249
2250 if ((state & ELSE_SEEN) != 0){
2251 Error_UnexpectedDirective ("#else within #else");
2252 return true;
2253 }
2254
2255 ifstack.Pop ();
2256
2257 if (arg.Length != 0) {
2258 Error_EndLineExpected ();
2259 return true;
2260 }
2261
2262 bool ret = false;
2263 if ((state & PARENT_TAKING) != 0) {
2264 ret = (state & TAKING) == 0;
2265
2266 if (ret)
2267 state |= TAKING;
2268 else
2269 state &= ~TAKING;
2270 }
2271
2272 ifstack.Push (state | ELSE_SEEN);
2273
2274 return ret;
2275 }
2276 case "define":
2277 if (any_token_seen){
2278 Error_TokensSeen ();
2279 return caller_is_taking;
2280 }
2281 PreProcessDefinition (true, arg, caller_is_taking);
2282 return caller_is_taking;
2283
2284 case "undef":
2285 if (any_token_seen){
2286 Error_TokensSeen ();
2287 return caller_is_taking;
2288 }
2289 PreProcessDefinition (false, arg, caller_is_taking);
2290 return caller_is_taking;
2291 }
2292
2293 //
2294 // These are only processed if we are in a `taking' block
2295 //
2296 if (!caller_is_taking)
2297 return false;
2298
2299 switch (cmd){
2300 case "error":
2301 Report.Error (1029, Location, "#error: '{0}'", arg);
2302 return true;
2303
2304 case "warning":
2305 Report.Warning (1030, 1, Location, "#warning: `{0}'", arg);
2306 return true;
2307
2308 case "pragma":
2309 if (RootContext.Version == LanguageVersion.ISO_1) {
2310 Report.FeatureIsNotAvailable (Location, "#pragma");
2311 return true;
2312 }
2313
2314 PreProcessPragma (arg);
2315 return true;
2316
2317 case "line":
2318 if (!PreProcessLine (arg))
2319 Report.Error (
2320 1576, Location,
2321 "The line number specified for #line directive is missing or invalid");
2322 return caller_is_taking;
2323 }
2324
2325 Report.Error (1024, Location, "Wrong preprocessor directive");
2326 return true;
2327
2328 }
2329
2330 private int consume_string (bool quoted)
2331 {
2332 int c;
2333 string_builder.Length = 0;
2334
2335 while ((c = get_char ()) != -1){
2336 if (c == '"'){
2337 if (quoted && peek_char () == '"'){
2338 string_builder.Append ((char) c);
2339 get_char ();
2340 continue;
2341 } else {
2342 val = string_builder.ToString ();
2343 return Token.LITERAL_STRING;
2344 }
2345 }
2346
2347 if (c == '\n'){
2348 if (!quoted)
2349 Report.Error (1010, Location, "Newline in constant");
2350 }
2351
2352 if (!quoted){
2353 int surrogate;
2354 c = escape (c, out surrogate);
2355 if (c == -1)
2356 return Token.ERROR;
2357 if (surrogate != 0) {
2358 string_builder.Append ((char) c);
2359 c = surrogate;
2360 }
2361 }
2362 string_builder.Append ((char) c);
2363 }
2364
2365 Report.Error (1039, Location, "Unterminated string literal");
2366 return Token.EOF;
2367 }
2368
2369 private int consume_identifier (int s)
2370 {
2371 int res = consume_identifier (s, false);
2372
2373 if (doc_state == XmlCommentState.Allowed)
2374 doc_state = XmlCommentState.NotAllowed;
2375
2376 return res;
2377 }
2378
2379 private int consume_identifier (int c, bool quoted)
2380 {
2381 int pos = 0;
2382
2383 if (c == '\\') {
2384 int surrogate;
2385 c = escape (c, out surrogate);
2386 if (surrogate != 0) {
2387 id_builder [pos++] = (char) c;
2388 c = surrogate;
2389 }
2390 }
2391
2392 id_builder [pos++] = (char) c;
2393 Location loc = Location;
2394
2395 while ((c = get_char ()) != -1) {
2396 loop:
2397 if (is_identifier_part_character ((char) c)){
2398 if (pos == max_id_size){
2399 Report.Error (645, loc, "Identifier too long (limit is 512 chars)");
2400 return Token.ERROR;
2401 }
2402
2403 id_builder [pos++] = (char) c;
2404 } else if (c == '\\') {
2405 int surrogate;
2406 c = escape (c, out surrogate);
2407 if (surrogate != 0) {
2408 if (is_identifier_part_character ((char) c))
2409 id_builder [pos++] = (char) c;
2410 c = surrogate;
2411 }
2412 goto loop;
2413 } else {
2414 putback (c);
2415 break;
2416 }
2417 }
2418
2419 //
2420 // Optimization: avoids doing the keyword lookup
2421 // on uppercase letters
2422 //
2423 if (id_builder [0] >= '_' && !quoted) {
2424 int keyword = GetKeyword (id_builder, pos);
2425 if (keyword != -1) {
2426 // TODO: No need to store location for keyword, required location cleanup
2427 val = loc;
2428 return keyword;
2429 }
2430 }
2431
2432 //
2433 // Keep identifiers in an array of hashtables to avoid needless
2434 // allocations
2435 //
2436 CharArrayHashtable identifiers_group = identifiers [pos];
2437 if (identifiers_group != null) {
2438 val = identifiers_group [id_builder];
2439 if (val != null) {
2440 val = new LocatedToken (loc, (string) val);
2441 if (quoted)
2442 AddEscapedIdentifier ((LocatedToken) val);
2443 return Token.IDENTIFIER;
2444 }
2445 } else {
2446 identifiers_group = new CharArrayHashtable (pos);
2447 identifiers [pos] = identifiers_group;
2448 }
2449
2450 char [] chars = new char [pos];
2451 Array.Copy (id_builder, chars, pos);
2452
2453 val = new String (id_builder, 0, pos);
2454 identifiers_group.Add (chars, val);
2455
2456 if (RootContext.Version == LanguageVersion.ISO_1) {
2457 for (int i = 1; i < chars.Length; i += 3) {
2458 if (chars [i] == '_' && (chars [i - 1] == '_' || chars [i + 1] == '_')) {
2459 Report.Error (1638, loc,
2460 "`{0}': Any identifier with double underscores cannot be used when ISO language version mode is specified", val.ToString ());
2461 }
2462 }
2463 }
2464
2465 val = new LocatedToken (loc, (string) val);
2466 if (quoted)
2467 AddEscapedIdentifier ((LocatedToken) val);
2468 return Token.IDENTIFIER;
2469 }
2470
2471 public int xtoken ()
2472 {
2473 int d, c;
2474
2475 // Whether we have seen comments on the current line
2476 bool comments_seen = false;
2477 while ((c = get_char ()) != -1) {
2478 switch (c) {
2479 case '\t':
2480 col = ((col + 8) / 8) * 8;
2481 continue;
2482
2483 case ' ':
2484 case '\f':
2485 case '\v':
2486 case 0xa0:
2487 case 0:
2488 case 0xFEFF: // Ignore BOM anywhere in the file
2489 continue;
2490
2491 /* This is required for compatibility with .NET
2492 case 0xEF:
2493 if (peek_char () == 0xBB) {
2494 PushPosition ();
2495 get_char ();
2496 if (get_char () == 0xBF)
2497 continue;
2498 PopPosition ();
2499 }
2500 break;
2501 */
2502 case '\r':
2503 if (peek_char () != '\n')
2504 advance_line ();
2505 else
2506 get_char ();
2507
2508 any_token_seen |= tokens_seen;
2509 tokens_seen = false;
2510 comments_seen = false;
2511 continue;
2512
2513 case '\\':
2514 tokens_seen = true;
2515 return consume_identifier (c);
2516
2517 case '{':
2518 val = Location;
2519 return Token.OPEN_BRACE;
2520 case '}':
2521 val = Location;
2522 return Token.CLOSE_BRACE;
2523 case '[':
2524 // To block doccomment inside attribute declaration.
2525 if (doc_state == XmlCommentState.Allowed)
2526 doc_state = XmlCommentState.NotAllowed;
2527 return Token.OPEN_BRACKET;
2528 case ']':
2529 return Token.CLOSE_BRACKET;
2530 case '(':
2531 val = Location;
2532 //
2533 // An expression versions of parens can appear in block context only
2534 //
2535 if (parsing_block != 0 && !lambda_arguments_parsing) {
2536
2537 //
2538 // Optmize most common case where we know that parens
2539 // is not special
2540 //
2541 switch (current_token) {
2542 case Token.IDENTIFIER:
2543 case Token.IF:
2544 case Token.FOR:
2545 case Token.FOREACH:
2546 case Token.TYPEOF:
2547 case Token.WHILE:
2548 case Token.USING:
2549 case Token.DEFAULT:
2550 case Token.DELEGATE:
2551 case Token.OP_GENERICS_GT:
2552 return Token.OPEN_PARENS;
2553 }
2554
2555 // Optimize using peek
2556 int xx = peek_char ();
2557 switch (xx) {
2558 case '(':
2559 case '\'':
2560 case '"':
2561 case '0':
2562 case '1':
2563 return Token.OPEN_PARENS;
2564 }
2565
2566 lambda_arguments_parsing = true;
2567 PushPosition ();
2568 d = TokenizeOpenParens ();
2569 PopPosition ();
2570 lambda_arguments_parsing = false;
2571 return d;
2572 }
2573
2574 return Token.OPEN_PARENS;
2575 case ')':
2576 return Token.CLOSE_PARENS;
2577 case ',':
2578 return Token.COMMA;
2579 case ';':
2580 return Token.SEMICOLON;
2581 case '~':
2582 return Token.TILDE;
2583 case '?':
2584 return TokenizePossibleNullableType ();
2585 case '<':
2586 if (parsing_generic_less_than++ > 0)
2587 return Token.OP_GENERICS_LT;
2588
2589 return TokenizeLessThan ();
2590
2591 case '>':
2592 d = peek_char ();
2593
2594 if (d == '='){
2595 get_char ();
2596 return Token.OP_GE;
2597 }
2598
2599 if (parsing_generic_less_than > 1 || (parsing_generic_less_than == 1 && d != '>')) {
2600 parsing_generic_less_than--;
2601 return Token.OP_GENERICS_GT;
2602 }
2603
2604 if (d == '>') {
2605 get_char ();
2606 d = peek_char ();
2607
2608 if (d == '=') {
2609 get_char ();
2610 return Token.OP_SHIFT_RIGHT_ASSIGN;
2611 }
2612 return Token.OP_SHIFT_RIGHT;
2613 }
2614
2615 return Token.OP_GT;
2616
2617 case '+':
2618 d = peek_char ();
2619 if (d == '+') {
2620 d = Token.OP_INC;
2621 } else if (d == '=') {
2622 d = Token.OP_ADD_ASSIGN;
2623 } else {
2624 return Token.PLUS;
2625 }
2626 get_char ();
2627 return d;
2628
2629 case '-':
2630 d = peek_char ();
2631 if (d == '-') {
2632 d = Token.OP_DEC;
2633 } else if (d == '=')
2634 d = Token.OP_SUB_ASSIGN;
2635 else if (d == '>')
2636 d = Token.OP_PTR;
2637 else {
2638 return Token.MINUS;
2639 }
2640 get_char ();
2641 return d;
2642
2643 case '!':
2644 if (peek_char () == '='){
2645 get_char ();
2646 return Token.OP_NE;
2647 }
2648 return Token.BANG;
2649
2650 case '=':
2651 d = peek_char ();
2652 if (d == '='){
2653 get_char ();
2654 return Token.OP_EQ;
2655 }
2656 if (d == '>'){
2657 get_char ();
2658 return Token.ARROW;
2659 }
2660
2661 return Token.ASSIGN;
2662
2663 case '&':
2664 d = peek_char ();
2665 if (d == '&'){
2666 get_char ();
2667 return Token.OP_AND;
2668 }
2669 if (d == '='){
2670 get_char ();
2671 return Token.OP_AND_ASSIGN;
2672 }
2673 return Token.BITWISE_AND;
2674
2675 case '|':
2676 d = peek_char ();
2677 if (d == '|'){
2678 get_char ();
2679 return Token.OP_OR;
2680 }
2681 if (d == '='){
2682 get_char ();
2683 return Token.OP_OR_ASSIGN;
2684 }
2685 return Token.BITWISE_OR;
2686
2687 case '*':
2688 if (peek_char () == '='){
2689 get_char ();
2690 return Token.OP_MULT_ASSIGN;
2691 }
2692 val = Location;
2693 return Token.STAR;
2694
2695 case '/':
2696 d = peek_char ();
2697 if (d == '='){
2698 get_char ();
2699 return Token.OP_DIV_ASSIGN;
2700 }
2701
2702 // Handle double-slash comments.
2703 if (d == '/'){
2704 get_char ();
2705 if (RootContext.Documentation != null && peek_char () == '/') {
2706 get_char ();
2707 // Don't allow ////.
2708 if ((d = peek_char ()) != '/') {
2709 update_comment_location ();
2710 if (doc_state == XmlCommentState.Allowed)
2711 handle_one_line_xml_comment ();
2712 else if (doc_state == XmlCommentState.NotAllowed)
2713 warn_incorrect_doc_comment ();
2714 }
2715 }
2716 while ((d = get_char ()) != -1 && (d != '\n') && d != '\r');
2717
2718 any_token_seen |= tokens_seen;
2719 tokens_seen = false;
2720 comments_seen = false;
2721 continue;
2722 } else if (d == '*'){
2723 get_char ();
2724 bool docAppend = false;
2725 if (RootContext.Documentation != null && peek_char () == '*') {
2726 get_char ();
2727 update_comment_location ();
2728 // But when it is /**/, just do nothing.
2729 if (peek_char () == '/') {
2730 get_char ();
2731 continue;
2732 }
2733 if (doc_state == XmlCommentState.Allowed)
2734 docAppend = true;
2735 else if (doc_state == XmlCommentState.NotAllowed)
2736 warn_incorrect_doc_comment ();
2737 }
2738
2739 int current_comment_start = 0;
2740 if (docAppend) {
2741 current_comment_start = xml_comment_buffer.Length;
2742 xml_comment_buffer.Append (Environment.NewLine);
2743 }
2744
2745 while ((d = get_char ()) != -1){
2746 if (d == '*' && peek_char () == '/'){
2747 get_char ();
2748 comments_seen = true;
2749 break;
2750 }
2751 if (docAppend)
2752 xml_comment_buffer.Append ((char) d);
2753
2754 if (d == '\n'){
2755 any_token_seen |= tokens_seen;
2756 tokens_seen = false;
2757 //
2758 // Reset 'comments_seen' just to be consistent.
2759 // It doesn't matter either way, here.
2760 //
2761 comments_seen = false;
2762 }
2763 }
2764 if (!comments_seen)
2765 Report.Error (1035, Location, "End-of-file found, '*/' expected");
2766
2767 if (docAppend)
2768 update_formatted_doc_comment (current_comment_start);
2769 continue;
2770 }
2771 return Token.DIV;
2772
2773 case '%':
2774 if (peek_char () == '='){
2775 get_char ();
2776 return Token.OP_MOD_ASSIGN;
2777 }
2778 return Token.PERCENT;
2779
2780 case '^':
2781 if (peek_char () == '='){
2782 get_char ();
2783 return Token.OP_XOR_ASSIGN;
2784 }
2785 return Token.CARRET;
2786
2787 case ':':
2788 if (peek_char () == ':') {
2789 get_char ();
2790 return Token.DOUBLE_COLON;
2791 }
2792 return Token.COLON;
2793
2794 case '0': case '1': case '2': case '3': case '4':
2795 case '5': case '6': case '7': case '8': case '9':
2796 tokens_seen = true;
2797 return is_number (c);
2798
2799 case '\n': // white space
2800 any_token_seen |= tokens_seen;
2801 tokens_seen = false;
2802 comments_seen = false;
2803 continue;
2804
2805 case '.':
2806 tokens_seen = true;
2807 d = peek_char ();
2808 if (d >= '0' && d <= '9')
2809 return is_number (c);
2810 return Token.DOT;
2811
2812 case '#':
2813 if (tokens_seen || comments_seen) {
2814 Eror_WrongPreprocessorLocation ();
2815 return Token.ERROR;
2816 }
2817
2818 if (handle_preprocessing_directive (true))
2819 continue;
2820
2821 bool directive_expected = false;
2822 while ((c = get_char ()) != -1) {
2823 if (col == 1) {
2824 directive_expected = true;
2825 } else if (!directive_expected) {
2826 // TODO: Implement comment support for disabled code and uncomment this code
2827 // if (c == '#') {
2828 // Eror_WrongPreprocessorLocation ();
2829 // return Token.ERROR;
2830 // }
2831 continue;
2832 }
2833
2834 if (c == ' ' || c == '\t' || c == '\r' || c == '\n' || c == '\f' || c == '\v' )
2835 continue;
2836
2837 if (c == '#') {
2838 if (handle_preprocessing_directive (false))
2839 break;
2840 }
2841 directive_expected = false;
2842 }
2843
2844 if (c != -1) {
2845 tokens_seen = false;
2846 continue;
2847 }
2848
2849 return Token.EOF;
2850
2851 case '"':
2852 return consume_string (false);
2853
2854 case '\'':
2855 return TokenizeBackslash ();
2856
2857 case '@':
2858 c = get_char ();
2859 if (c == '"') {
2860 tokens_seen = true;
2861 return consume_string (true);
2862 }
2863
2864 if (is_identifier_start_character (c)){
2865 return consume_identifier (c, true);
2866 }
2867
2868 Report.Error (1646, Location, "Keyword, identifier, or string expected after verbatim specifier: @");
2869 return Token.ERROR;
2870
2871 case EvalStatementParserCharacter:
2872 return Token.EVAL_STATEMENT_PARSER;
2873 case EvalCompilationUnitParserCharacter:
2874 return Token.EVAL_COMPILATION_UNIT_PARSER;
2875 case EvalUsingDeclarationsParserCharacter:
2876 return Token.EVAL_USING_DECLARATIONS_UNIT_PARSER;
2877 }
2878
2879 if (is_identifier_start_character (c)) {
2880 tokens_seen = true;
2881 return consume_identifier (c);
2882 }
2883
2884 error_details = ((char)c).ToString ();
2885 return Token.ERROR;
2886 }
2887
2888 if (CompleteOnEOF){
2889 if (generated)
2890 return Token.COMPLETE_COMPLETION;
2891
2892 generated = true;
2893 return Token.GENERATE_COMPLETION;
2894 }
2895
2896
2897 return Token.EOF;
2898 }
2899
2900 int TokenizeBackslash ()
2901 {
2902 int c = get_char ();
2903 tokens_seen = true;
2904 if (c == '\'') {
2905 error_details = "Empty character literal";
2906 Report.Error (1011, Location, error_details);
2907 return Token.ERROR;
2908 }
2909 if (c == '\r' || c == '\n') {
2910 Report.Error (1010, Location, "Newline in constant");
2911 return Token.ERROR;
2912 }
2913
2914 int d;
2915 c = escape (c, out d);
2916 if (c == -1)
2917 return Token.ERROR;
2918 if (d != 0)
2919 throw new NotImplementedException ();
2920
2921 val = (char) c;
2922 c = get_char ();
2923
2924 if (c != '\'') {
2925 Report.Error (1012, Location, "Too many characters in character literal");
2926
2927 // Try to recover, read until newline or next "'"
2928 while ((c = get_char ()) != -1) {
2929 if (c == '\n' || c == '\'')
2930 break;
2931 }
2932 return Token.ERROR;
2933 }
2934
2935 return Token.LITERAL_CHARACTER;
2936 }
2937
2938 int TokenizeLessThan ()
2939 {
2940 int d;
2941 if (handle_typeof) {
2942 PushPosition ();
2943 if (parse_generic_dimension (out d)) {
2944 val = d;
2945 DiscardPosition ();
2946 return Token.GENERIC_DIMENSION;
2947 }
2948 PopPosition ();
2949 }
2950
2951 // Save current position and parse next token.
2952 PushPosition ();
2953 if (parse_less_than ()) {
2954 if (parsing_generic_declaration && token () != Token.DOT) {
2955 d = Token.OP_GENERICS_LT_DECL;
2956 } else {
2957 d = Token.OP_GENERICS_LT;
2958 }
2959 PopPosition ();
2960 return d;
2961 }
2962
2963 PopPosition ();
2964 parsing_generic_less_than = 0;
2965
2966 d = peek_char ();
2967 if (d == '<') {
2968 get_char ();
2969 d = peek_char ();
2970
2971 if (d == '=') {
2972 get_char ();
2973 return Token.OP_SHIFT_LEFT_ASSIGN;
2974 }
2975 return Token.OP_SHIFT_LEFT;
2976 }
2977
2978 if (d == '=') {
2979 get_char ();
2980 return Token.OP_LE;
2981 }
2982 return Token.OP_LT;
2983 }
2984
2985 //
2986 // Handles one line xml comment
2987 //
2988 private void handle_one_line_xml_comment ()
2989 {
2990 int c;
2991 while ((c = peek_char ()) == ' ')
2992 get_char (); // skip heading whitespaces.
2993 while ((c = peek_char ()) != -1 && c != '\n' && c != '\r') {
2994 xml_comment_buffer.Append ((char) get_char ());
2995 }
2996 if (c == '\r' || c == '\n')
2997 xml_comment_buffer.Append (Environment.NewLine);
2998 }
2999
3000 //
3001 // Remove heading "*" in Javadoc-like xml documentation.
3002 //
3003 private void update_formatted_doc_comment (int current_comment_start)
3004 {
3005 int length = xml_comment_buffer.Length - current_comment_start;
3006 string [] lines = xml_comment_buffer.ToString (
3007 current_comment_start,
3008 length).Replace ("\r", "").Split ('\n');
3009
3010 // The first line starts with /**, thus it is not target
3011 // for the format check.
3012 for (int i = 1; i < lines.Length; i++) {
3013 string s = lines [i];
3014 int idx = s.IndexOf ('*');
3015 string head = null;
3016 if (idx < 0) {
3017 if (i < lines.Length - 1)
3018 return;
3019 head = s;
3020 } else
3021 head = s.Substring (0, idx);
3022 foreach (char c in head)
3023 if (c != ' ')
3024 return;
3025 lines [i] = s.Substring (idx + 1);
3026 }
3027 xml_comment_buffer.Remove (current_comment_start, length);
3028 xml_comment_buffer.Insert (current_comment_start, String.Join (Environment.NewLine, lines));
3029 }
3030
3031 //
3032 // Updates current comment location.
3033 //
3034 private void update_comment_location ()
3035 {
3036 if (current_comment_location.IsNull) {
3037 // "-2" is for heading "//" or "/*"
3038 current_comment_location =
3039 new Location (ref_line, hidden ? -1 : col - 2);
3040 }
3041 }
3042
3043 //
3044 // Checks if there was incorrect doc comments and raise
3045 // warnings.
3046 //
3047 public void check_incorrect_doc_comment ()
3048 {
3049 if (xml_comment_buffer.Length > 0)
3050 warn_incorrect_doc_comment ();
3051 }
3052
3053 //
3054 // Raises a warning when tokenizer found incorrect doccomment
3055 // markup.
3056 //
3057 private void warn_incorrect_doc_comment ()
3058 {
3059 if (doc_state != XmlCommentState.Error) {
3060 doc_state = XmlCommentState.Error;
3061 // in csc, it is 'XML comment is not placed on
3062 // a valid language element'. But that does not
3063 // make sense.
3064 Report.Warning (1587, 2, Location, "XML comment is not placed on a valid language element");
3065 }
3066 }
3067
3068 //
3069 // Consumes the saved xml comment lines (if any)
3070 // as for current target member or type.
3071 //
3072 public string consume_doc_comment ()
3073 {
3074 if (xml_comment_buffer.Length > 0) {
3075 string ret = xml_comment_buffer.ToString ();
3076 reset_doc_comment ();
3077 return ret;
3078 }
3079 return null;
3080 }
3081
3082 Report Report {
3083 get { return context.Report; }
3084 }
3085
3086 void reset_doc_comment ()
3087 {
3088 xml_comment_buffer.Length = 0;
3089 current_comment_location = Location.Null;
3090 }
3091
3092 public void cleanup ()
3093 {
3094 if (ifstack != null && ifstack.Count >= 1) {
3095 int state = (int) ifstack.Pop ();
3096 if ((state & REGION) != 0)
3097 Report.Error (1038, Location, "#endregion directive expected");
3098 else
3099 Report.Error (1027, Location, "Expected `#endif' directive");
3100 }
3101 }
3102 }
3103
3104 //
3105 // Indicates whether it accepts XML documentation or not.
3106 //
3107 public enum XmlCommentState {
3108 // comment is allowed in this state.
3109 Allowed,
3110 // comment is not allowed in this state.
3111 NotAllowed,
3112 // once comments appeared when it is NotAllowed, then the
3113 // state is changed to it, until the state is changed to
3114 // .Allowed.
3115 Error
3116 }
3117 }
3118
git-svn mirror of mcs