?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
2009-07-20 Jb Evain <jbevain@novell.com>
[mcs.git] / mcs / cs-tokenizer.cs
blobfafb277aea9c64225ebd41392618f816ff1fb157
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 bool hidden = false;
47 int ref_line = 1;
48 int line = 1;
49 int col = 0;
50 int previous_col;
51 int current_token;
52 bool handle_get_set = false;
53 bool handle_remove_add = false;
54 bool handle_where = false;
55 bool handle_typeof = false;
56 bool lambda_arguments_parsing;
57 Location current_comment_location = Location.Null;
58 ArrayList escaped_identifiers;
59 int parsing_generic_less_than;
60
61 //
62 // Used mainly for parser optimizations. Some expressions for instance
63 // can appear only in block (including initializer, base initializer)
64 // scope only
65 //
66 public int parsing_block;
67 internal bool query_parsing;
68
69 //
70 // When parsing type only, useful for ambiguous nullable types
71 //
72 public int parsing_type;
73
74 //
75 // Set when parsing generic declaration (type or method header)
76 //
77 public bool parsing_generic_declaration;
78
79 //
80 // The value indicates that we have not reach any declaration or
81 // namespace yet
82 //
83 public int parsing_declaration;
84
85 //
86 // The special character to inject on streams to trigger the EXPRESSION_PARSE
87 // token to be returned. It just happens to be a Unicode character that
88 // would never be part of a program (can not be an identifier).
89 //
90 // This character is only tested just before the tokenizer is about to report
91 // an error; So on the regular operation mode, this addition will have no
92 // impact on the tokenizer's performance.
93 //
94
95 public const int EvalStatementParserCharacter = 0x2190; // Unicode Left Arrow
96 public const int EvalCompilationUnitParserCharacter = 0x2191; // Unicode Arrow
97 public const int EvalUsingDeclarationsParserCharacter = 0x2192; // Unicode Arrow
98
99 //
100 // XML documentation buffer. The save point is used to divide
101 // comments on types and comments on members.
102 //
103 StringBuilder xml_comment_buffer;
104
105 //
106 // See comment on XmlCommentState enumeration.
107 //
108 XmlCommentState xml_doc_state = XmlCommentState.Allowed;
109
110 //
111 // Whether tokens have been seen on this line
112 //
113 bool tokens_seen = false;
114
115 //
116 // Set to true once the GENERATE_COMPLETION token has bee
117 // returned. This helps produce one GENERATE_COMPLETION,
118 // as many COMPLETE_COMPLETION as necessary to complete the
119 // AST tree and one final EOF.
120 //
121 bool generated;
122
123 //
124 // Whether a token has been seen on the file
125 // This is needed because `define' is not allowed to be used
126 // after a token has been seen.
127 //
128 bool any_token_seen = false;
129
130 static readonly char[] simple_whitespaces = new char[] { ' ', '\t' };
131
132 public bool PropertyParsing {
133 get { return handle_get_set; }
134 set { handle_get_set = value; }
135 }
136
137 public bool EventParsing {
138 get { return handle_remove_add; }
139 set { handle_remove_add = value; }
140 }
141
142 public bool ConstraintsParsing {
143 get { return handle_where; }
144 set { handle_where = value; }
145 }
146
147 public bool TypeOfParsing {
148 get { return handle_typeof; }
149 set { handle_typeof = value; }
150 }
151
152 public XmlCommentState doc_state {
153 get { return xml_doc_state; }
154 set {
155 if (value == XmlCommentState.Allowed) {
156 check_incorrect_doc_comment ();
157 reset_doc_comment ();
158 }
159 xml_doc_state = value;
160 }
161 }
162
163 //
164 // This is used to trigger completion generation on the parser
165 public bool CompleteOnEOF;
166
167 void AddEscapedIdentifier (LocatedToken lt)
168 {
169 if (escaped_identifiers == null)
170 escaped_identifiers = new ArrayList ();
171
172 escaped_identifiers.Add (lt);
173 }
174
175 public bool IsEscapedIdentifier (Location loc)
176 {
177 if (escaped_identifiers != null) {
178 foreach (LocatedToken lt in escaped_identifiers)
179 if (lt.Location.Equals (loc))
180 return true;
181 }
182
183 return false;
184 }
185
186 //
187 // Class variables
188 //
189 static KeywordEntry[][] keywords;
190 static Hashtable keyword_strings;
191 static NumberStyles styles;
192 static NumberFormatInfo csharp_format_info;
193
194 //
195 // Values for the associated token returned
196 //
197 internal int putback_char; // Used by repl only
198 Object val;
199
200 //
201 // Pre-processor
202 //
203 const int TAKING = 1;
204 const int ELSE_SEEN = 4;
205 const int PARENT_TAKING = 8;
206 const int REGION = 16;
207
208 //
209 // pre-processor if stack state:
210 //
211 Stack ifstack;
212
213 static System.Text.StringBuilder string_builder;
214
215 const int max_id_size = 512;
216 static char [] id_builder = new char [max_id_size];
217
218 static CharArrayHashtable [] identifiers = new CharArrayHashtable [max_id_size + 1];
219
220 const int max_number_size = 512;
221 static char [] number_builder = new char [max_number_size];
222 static int number_pos;
223
224 //
225 // Details about the error encoutered by the tokenizer
226 //
227 string error_details;
228
229 public string error {
230 get {
231 return error_details;
232 }
233 }
234
235 public int Line {
236 get {
237 return ref_line;
238 }
239 }
240
241 //
242 // This is used when the tokenizer needs to save
243 // the current position as it needs to do some parsing
244 // on its own to deamiguate a token in behalf of the
245 // parser.
246 //
247 Stack position_stack = new Stack (2);
248 class Position {
249 public int position;
250 public int line;
251 public int ref_line;
252 public int col;
253 public bool hidden;
254 public int putback_char;
255 public int previous_col;
256 public Stack ifstack;
257 public int parsing_generic_less_than;
258 public int current_token;
259
260 public Position (Tokenizer t)
261 {
262 position = t.reader.Position;
263 line = t.line;
264 ref_line = t.ref_line;
265 col = t.col;
266 hidden = t.hidden;
267 putback_char = t.putback_char;
268 previous_col = t.previous_col;
269 if (t.ifstack != null && t.ifstack.Count != 0)
270 ifstack = (Stack)t.ifstack.Clone ();
271 parsing_generic_less_than = t.parsing_generic_less_than;
272 current_token = t.current_token;
273 }
274 }
275
276 public void PushPosition ()
277 {
278 position_stack.Push (new Position (this));
279 }
280
281 public void PopPosition ()
282 {
283 Position p = (Position) position_stack.Pop ();
284
285 reader.Position = p.position;
286 ref_line = p.ref_line;
287 line = p.line;
288 col = p.col;
289 hidden = p.hidden;
290 putback_char = p.putback_char;
291 previous_col = p.previous_col;
292 ifstack = p.ifstack;
293 parsing_generic_less_than = p.parsing_generic_less_than;
294 current_token = p.current_token;
295 }
296
297 // Do not reset the position, ignore it.
298 public void DiscardPosition ()
299 {
300 position_stack.Pop ();
301 }
302
303 static void AddKeyword (string kw, int token)
304 {
305 keyword_strings.Add (kw, kw);
306
307 int length = kw.Length;
308 if (keywords [length] == null) {
309 keywords [length] = new KeywordEntry ['z' - '_' + 1];
310 }
311
312 int char_index = kw [0] - '_';
313 KeywordEntry kwe = keywords [length] [char_index];
314 if (kwe == null) {
315 keywords [length] [char_index] = new KeywordEntry (kw, token);
316 return;
317 }
318
319 while (kwe.Next != null) {
320 kwe = kwe.Next;
321 }
322
323 kwe.Next = new KeywordEntry (kw, token);
324 }
325
326 static void InitTokens ()
327 {
328 keyword_strings = new Hashtable ();
329
330 // 11 is the length of the longest keyword for now
331 keywords = new KeywordEntry [11] [];
332
333 AddKeyword ("__arglist", Token.ARGLIST);
334 AddKeyword ("abstract", Token.ABSTRACT);
335 AddKeyword ("as", Token.AS);
336 AddKeyword ("add", Token.ADD);
337 AddKeyword ("base", Token.BASE);
338 AddKeyword ("bool", Token.BOOL);
339 AddKeyword ("break", Token.BREAK);
340 AddKeyword ("byte", Token.BYTE);
341 AddKeyword ("case", Token.CASE);
342 AddKeyword ("catch", Token.CATCH);
343 AddKeyword ("char", Token.CHAR);
344 AddKeyword ("checked", Token.CHECKED);
345 AddKeyword ("class", Token.CLASS);
346 AddKeyword ("const", Token.CONST);
347 AddKeyword ("continue", Token.CONTINUE);
348 AddKeyword ("decimal", Token.DECIMAL);
349 AddKeyword ("default", Token.DEFAULT);
350 AddKeyword ("delegate", Token.DELEGATE);
351 AddKeyword ("do", Token.DO);
352 AddKeyword ("double", Token.DOUBLE);
353 AddKeyword ("else", Token.ELSE);
354 AddKeyword ("enum", Token.ENUM);
355 AddKeyword ("event", Token.EVENT);
356 AddKeyword ("explicit", Token.EXPLICIT);
357 AddKeyword ("extern", Token.EXTERN);
358 AddKeyword ("false", Token.FALSE);
359 AddKeyword ("finally", Token.FINALLY);
360 AddKeyword ("fixed", Token.FIXED);
361 AddKeyword ("float", Token.FLOAT);
362 AddKeyword ("for", Token.FOR);
363 AddKeyword ("foreach", Token.FOREACH);
364 AddKeyword ("goto", Token.GOTO);
365 AddKeyword ("get", Token.GET);
366 AddKeyword ("if", Token.IF);
367 AddKeyword ("implicit", Token.IMPLICIT);
368 AddKeyword ("in", Token.IN);
369 AddKeyword ("int", Token.INT);
370 AddKeyword ("interface", Token.INTERFACE);
371 AddKeyword ("internal", Token.INTERNAL);
372 AddKeyword ("is", Token.IS);
373 AddKeyword ("lock", Token.LOCK);
374 AddKeyword ("long", Token.LONG);
375 AddKeyword ("namespace", Token.NAMESPACE);
376 AddKeyword ("new", Token.NEW);
377 AddKeyword ("null", Token.NULL);
378 AddKeyword ("object", Token.OBJECT);
379 AddKeyword ("operator", Token.OPERATOR);
380 AddKeyword ("out", Token.OUT);
381 AddKeyword ("override", Token.OVERRIDE);
382 AddKeyword ("params", Token.PARAMS);
383 AddKeyword ("private", Token.PRIVATE);
384 AddKeyword ("protected", Token.PROTECTED);
385 AddKeyword ("public", Token.PUBLIC);
386 AddKeyword ("readonly", Token.READONLY);
387 AddKeyword ("ref", Token.REF);
388 AddKeyword ("remove", Token.REMOVE);
389 AddKeyword ("return", Token.RETURN);
390 AddKeyword ("sbyte", Token.SBYTE);
391 AddKeyword ("sealed", Token.SEALED);
392 AddKeyword ("set", Token.SET);
393 AddKeyword ("short", Token.SHORT);
394 AddKeyword ("sizeof", Token.SIZEOF);
395 AddKeyword ("stackalloc", Token.STACKALLOC);
396 AddKeyword ("static", Token.STATIC);
397 AddKeyword ("string", Token.STRING);
398 AddKeyword ("struct", Token.STRUCT);
399 AddKeyword ("switch", Token.SWITCH);
400 AddKeyword ("this", Token.THIS);
401 AddKeyword ("throw", Token.THROW);
402 AddKeyword ("true", Token.TRUE);
403 AddKeyword ("try", Token.TRY);
404 AddKeyword ("typeof", Token.TYPEOF);
405 AddKeyword ("uint", Token.UINT);
406 AddKeyword ("ulong", Token.ULONG);
407 AddKeyword ("unchecked", Token.UNCHECKED);
408 AddKeyword ("unsafe", Token.UNSAFE);
409 AddKeyword ("ushort", Token.USHORT);
410 AddKeyword ("using", Token.USING);
411 AddKeyword ("virtual", Token.VIRTUAL);
412 AddKeyword ("void", Token.VOID);
413 AddKeyword ("volatile", Token.VOLATILE);
414 AddKeyword ("while", Token.WHILE);
415 AddKeyword ("partial", Token.PARTIAL);
416 AddKeyword ("where", Token.WHERE);
417
418 // LINQ keywords
419 AddKeyword ("from", Token.FROM);
420 AddKeyword ("join", Token.JOIN);
421 AddKeyword ("on", Token.ON);
422 AddKeyword ("equals", Token.EQUALS);
423 AddKeyword ("select", Token.SELECT);
424 AddKeyword ("group", Token.GROUP);
425 AddKeyword ("by", Token.BY);
426 AddKeyword ("let", Token.LET);
427 AddKeyword ("orderby", Token.ORDERBY);
428 AddKeyword ("ascending", Token.ASCENDING);
429 AddKeyword ("descending", Token.DESCENDING);
430 AddKeyword ("into", Token.INTO);
431 }
432
433 //
434 // Class initializer
435 //
436 static Tokenizer ()
437 {
438 InitTokens ();
439 csharp_format_info = NumberFormatInfo.InvariantInfo;
440 styles = NumberStyles.Float;
441
442 string_builder = new System.Text.StringBuilder ();
443 }
444
445 int GetKeyword (char[] id, int id_len)
446 {
447 //
448 // Keywords are stored in an array of arrays grouped by their
449 // length and then by the first character
450 //
451 if (id_len >= keywords.Length || keywords [id_len] == null)
452 return -1;
453
454 int first_index = id [0] - '_';
455 if (first_index > 'z')
456 return -1;
457
458 KeywordEntry kwe = keywords [id_len] [first_index];
459 if (kwe == null)
460 return -1;
461
462 int res;
463 do {
464 res = kwe.Token;
465 for (int i = 1; i < id_len; ++i) {
466 if (id [i] != kwe.Value [i]) {
467 res = 0;
468 break;
469 }
470 }
471 kwe = kwe.Next;
472 } while (kwe != null && res == 0);
473
474 if (res == 0)
475 return -1;
476
477 int next_token;
478 switch (res) {
479 case Token.GET:
480 case Token.SET:
481 if (!handle_get_set)
482 res = -1;
483 break;
484 case Token.REMOVE:
485 case Token.ADD:
486 if (!handle_remove_add)
487 res = -1;
488 break;
489 case Token.EXTERN:
490 if (parsing_declaration == 0)
491 res = Token.EXTERN_ALIAS;
492 break;
493 case Token.DEFAULT:
494 if (peek_token () == Token.COLON) {
495 token ();
496 res = Token.DEFAULT_COLON;
497 }
498 break;
499 case Token.WHERE:
500 if (!handle_where && !query_parsing)
501 res = -1;
502 break;
503 case Token.FROM:
504 //
505 // A query expression is any expression that starts with `from identifier'
506 // followed by any token except ; , =
507 //
508 if (!query_parsing) {
509 if (lambda_arguments_parsing) {
510 res = -1;
511 break;
512 }
513
514 PushPosition ();
515 // HACK: to disable generics micro-parser, because PushPosition does not
516 // store identifiers array
517 parsing_generic_less_than = 1;
518 switch (xtoken ()) {
519 case Token.IDENTIFIER:
520 case Token.INT:
521 case Token.BOOL:
522 case Token.BYTE:
523 case Token.CHAR:
524 case Token.DECIMAL:
525 case Token.FLOAT:
526 case Token.LONG:
527 case Token.OBJECT:
528 case Token.STRING:
529 case Token.UINT:
530 case Token.ULONG:
531 next_token = xtoken ();
532 if (next_token == Token.SEMICOLON || next_token == Token.COMMA || next_token == Token.EQUALS)
533 goto default;
534
535 res = Token.FROM_FIRST;
536 query_parsing = true;
537 if (RootContext.Version <= LanguageVersion.ISO_2)
538 Report.FeatureIsNotAvailable (Location, "query expressions");
539 break;
540 case Token.VOID:
541 Expression.Error_VoidInvalidInTheContext (Location);
542 break;
543 default:
544 PopPosition ();
545 // HACK: A token is not a keyword so we need to restore identifiers buffer
546 // which has been overwritten before we grabbed the identifier
547 id_builder [0] = 'f'; id_builder [1] = 'r'; id_builder [2] = 'o'; id_builder [3] = 'm';
548 return -1;
549 }
550 PopPosition ();
551 }
552 break;
553 case Token.JOIN:
554 case Token.ON:
555 case Token.EQUALS:
556 case Token.SELECT:
557 case Token.GROUP:
558 case Token.BY:
559 case Token.LET:
560 case Token.ORDERBY:
561 case Token.ASCENDING:
562 case Token.DESCENDING:
563 case Token.INTO:
564 if (!query_parsing)
565 res = -1;
566 break;
567
568 case Token.USING:
569 case Token.NAMESPACE:
570 // TODO: some explanation needed
571 check_incorrect_doc_comment ();
572 break;
573
574 case Token.PARTIAL:
575 if (parsing_block > 0) {
576 res = -1;
577 break;
578 }
579
580 // Save current position and parse next token.
581 PushPosition ();
582
583 next_token = token ();
584 bool ok = (next_token == Token.CLASS) ||
585 (next_token == Token.STRUCT) ||
586 (next_token == Token.INTERFACE) ||
587 (next_token == Token.VOID);
588
589 PopPosition ();
590
591 if (ok) {
592 if (next_token == Token.VOID) {
593 if (RootContext.Version == LanguageVersion.ISO_1 ||
594 RootContext.Version == LanguageVersion.ISO_2)
595 Report.FeatureIsNotAvailable (Location, "partial methods");
596 } else if (RootContext.Version == LanguageVersion.ISO_1)
597 Report.FeatureIsNotAvailable (Location, "partial types");
598
599 return res;
600 }
601
602 if (next_token < Token.LAST_KEYWORD) {
603 Report.Error (267, Location,
604 "The `partial' modifier can be used only immediately before `class', `struct', `interface', or `void' keyword");
605 return token ();
606 }
607
608 res = -1;
609 break;
610 }
611
612 return res;
613 }
614
615 public Location Location {
616 get {
617 return new Location (ref_line, hidden ? -1 : col);
618 }
619 }
620
621 public Tokenizer (SeekableStreamReader input, CompilationUnit file)
622 {
623 this.ref_name = file;
624 this.file_name = file;
625 reader = input;
626
627 putback_char = -1;
628
629 xml_comment_buffer = new StringBuilder ();
630
631 //
632 // FIXME: This could be `Location.Push' but we have to
633 // find out why the MS compiler allows this
634 //
635 Mono.CSharp.Location.Push (file, file);
636 }
637
638 static bool is_identifier_start_character (int c)
639 {
640 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || Char.IsLetter ((char)c);
641 }
642
643 static bool is_identifier_part_character (char c)
644 {
645 if (c >= 'a' && c <= 'z')
646 return true;
647
648 if (c >= 'A' && c <= 'Z')
649 return true;
650
651 if (c == '_' || (c >= '0' && c <= '9'))
652 return true;
653
654 if (c < 0x80)
655 return false;
656
657 return Char.IsLetter (c) || Char.GetUnicodeCategory (c) == UnicodeCategory.ConnectorPunctuation;
658 }
659
660 public static bool IsKeyword (string s)
661 {
662 return keyword_strings [s] != null;
663 }
664
665 //
666 // Open parens micro parser. Detects both lambda and cast ambiguity.
667 //
668
669 int TokenizeOpenParens ()
670 {
671 int ptoken;
672 current_token = -1;
673
674 int bracket_level = 0;
675 bool is_type = false;
676 bool can_be_type = false;
677
678 while (true) {
679 ptoken = current_token;
680 token ();
681
682 switch (current_token) {
683 case Token.CLOSE_PARENS:
684 token ();
685
686 //
687 // Expression inside parens is lambda, (int i) =>
688 //
689 if (current_token == Token.ARROW) {
690 if (RootContext.Version <= LanguageVersion.ISO_2)
691 Report.FeatureIsNotAvailable (Location, "lambda expressions");
692
693 return Token.OPEN_PARENS_LAMBDA;
694 }
695
696 //
697 // Expression inside parens is single type, (int[])
698 //
699 if (is_type)
700 return Token.OPEN_PARENS_CAST;
701
702 //
703 // Expression is possible cast, look at next token, (T)null
704 //
705 if (can_be_type) {
706 switch (current_token) {
707 case Token.OPEN_PARENS:
708 case Token.BANG:
709 case Token.TILDE:
710 case Token.IDENTIFIER:
711 case Token.LITERAL_INTEGER:
712 case Token.LITERAL_FLOAT:
713 case Token.LITERAL_DOUBLE:
714 case Token.LITERAL_DECIMAL:
715 case Token.LITERAL_CHARACTER:
716 case Token.LITERAL_STRING:
717 case Token.BASE:
718 case Token.CHECKED:
719 case Token.DELEGATE:
720 case Token.FALSE:
721 case Token.FIXED:
722 case Token.NEW:
723 case Token.NULL:
724 case Token.SIZEOF:
725 case Token.THIS:
726 case Token.THROW:
727 case Token.TRUE:
728 case Token.TYPEOF:
729 case Token.UNCHECKED:
730 case Token.UNSAFE:
731 case Token.DEFAULT:
732
733 //
734 // These can be part of a member access
735 //
736 case Token.INT:
737 case Token.UINT:
738 case Token.SHORT:
739 case Token.USHORT:
740 case Token.LONG:
741 case Token.ULONG:
742 case Token.DOUBLE:
743 case Token.FLOAT:
744 case Token.CHAR:
745 case Token.BYTE:
746 case Token.DECIMAL:
747 case Token.BOOL:
748 return Token.OPEN_PARENS_CAST;
749 }
750 }
751 return Token.OPEN_PARENS;
752
753 case Token.DOT:
754 case Token.DOUBLE_COLON:
755 if (ptoken != Token.IDENTIFIER && ptoken != Token.OP_GENERICS_GT)
756 goto default;
757
758 continue;
759
760 case Token.IDENTIFIER:
761 switch (ptoken) {
762 case Token.DOT:
763 case Token.OP_GENERICS_LT:
764 case Token.COMMA:
765 case Token.DOUBLE_COLON:
766 case -1:
767 if (bracket_level == 0)
768 can_be_type = true;
769 continue;
770 default:
771 can_be_type = is_type = false;
772 continue;
773 }
774
775 case Token.OBJECT:
776 case Token.STRING:
777 case Token.BOOL:
778 case Token.DECIMAL:
779 case Token.FLOAT:
780 case Token.DOUBLE:
781 case Token.SBYTE:
782 case Token.BYTE:
783 case Token.SHORT:
784 case Token.USHORT:
785 case Token.INT:
786 case Token.UINT:
787 case Token.LONG:
788 case Token.ULONG:
789 case Token.CHAR:
790 case Token.VOID:
791 if (bracket_level == 0)
792 is_type = true;
793 continue;
794
795 case Token.COMMA:
796 if (bracket_level == 0) {
797 bracket_level = 100;
798 can_be_type = is_type = false;
799 }
800 continue;
801
802 case Token.OP_GENERICS_LT:
803 case Token.OPEN_BRACKET:
804 if (bracket_level++ == 0)
805 is_type = true;
806 continue;
807
808 case Token.OP_GENERICS_GT:
809 case Token.CLOSE_BRACKET:
810 --bracket_level;
811 continue;
812
813 case Token.INTERR_NULLABLE:
814 case Token.STAR:
815 if (bracket_level == 0)
816 is_type = true;
817 continue;
818
819 case Token.REF:
820 case Token.OUT:
821 can_be_type = is_type = false;
822 continue;
823
824 default:
825 return Token.OPEN_PARENS;
826 }
827 }
828 }
829
830 public static bool IsValidIdentifier (string s)
831 {
832 if (s == null || s.Length == 0)
833 return false;
834
835 if (!is_identifier_start_character (s [0]))
836 return false;
837
838 for (int i = 1; i < s.Length; i ++)
839 if (! is_identifier_part_character (s [i]))
840 return false;
841
842 return true;
843 }
844
845 bool parse_less_than ()
846 {
847 start:
848 int the_token = token ();
849 if (the_token == Token.OPEN_BRACKET) {
850 do {
851 the_token = token ();
852 } while (the_token != Token.CLOSE_BRACKET);
853 the_token = token ();
854 } else if (the_token == Token.IN || the_token == Token.OUT) {
855 the_token = token ();
856 }
857 switch (the_token) {
858 case Token.IDENTIFIER:
859 case Token.OBJECT:
860 case Token.STRING:
861 case Token.BOOL:
862 case Token.DECIMAL:
863 case Token.FLOAT:
864 case Token.DOUBLE:
865 case Token.SBYTE:
866 case Token.BYTE:
867 case Token.SHORT:
868 case Token.USHORT:
869 case Token.INT:
870 case Token.UINT:
871 case Token.LONG:
872 case Token.ULONG:
873 case Token.CHAR:
874 case Token.VOID:
875 break;
876 case Token.OP_GENERICS_GT:
877 return true;
878
879 default:
880 return false;
881 }
882 again:
883 the_token = token ();
884
885 if (the_token == Token.OP_GENERICS_GT)
886 return true;
887 else if (the_token == Token.COMMA || the_token == Token.DOT || the_token == Token.DOUBLE_COLON)
888 goto start;
889 else if (the_token == Token.INTERR_NULLABLE || the_token == Token.STAR)
890 goto again;
891 else if (the_token == Token.OP_GENERICS_LT) {
892 if (!parse_less_than ())
893 return false;
894 goto again;
895 } else if (the_token == Token.OPEN_BRACKET) {
896 rank_specifiers:
897 the_token = token ();
898 if (the_token == Token.CLOSE_BRACKET)
899 goto again;
900 else if (the_token == Token.COMMA)
901 goto rank_specifiers;
902 return false;
903 }
904
905 return false;
906 }
907
908 bool parse_generic_dimension (out int dimension)
909 {
910 dimension = 1;
911
912 again:
913 int the_token = token ();
914 if (the_token == Token.OP_GENERICS_GT)
915 return true;
916 else if (the_token == Token.COMMA) {
917 dimension++;
918 goto again;
919 }
920
921 return false;
922 }
923
924 public int peek_token ()
925 {
926 int the_token;
927
928 PushPosition ();
929 the_token = token ();
930 PopPosition ();
931
932 return the_token;
933 }
934
935 //
936 // Tonizes `?' using custom disambiguous rules to return one
937 // of following tokens: INTERR_NULLABLE, OP_COALESCING, INTERR
938 //
939 // Tricky expression look like:
940 //
941 // Foo ? a = x ? b : c;
942 //
943 int TokenizePossibleNullableType ()
944 {
945 if (parsing_block == 0 || parsing_type > 0)
946 return Token.INTERR_NULLABLE;
947
948 int d = peek_char ();
949 if (d == '?') {
950 get_char ();
951 return Token.OP_COALESCING;
952 }
953
954 switch (current_token) {
955 case Token.CLOSE_PARENS:
956 case Token.TRUE:
957 case Token.FALSE:
958 case Token.NULL:
959 case Token.LITERAL_INTEGER:
960 case Token.LITERAL_STRING:
961 return Token.INTERR;
962 }
963
964 if (d != ' ') {
965 if (d == ',' || d == ';' || d == '>')
966 return Token.INTERR_NULLABLE;
967 if (d == '*' || (d >= '0' && d <= '9'))
968 return Token.INTERR;
969 }
970
971 PushPosition ();
972 current_token = Token.NONE;
973 int next_token;
974 switch (xtoken ()) {
975 case Token.LITERAL_INTEGER:
976 case Token.LITERAL_STRING:
977 case Token.LITERAL_CHARACTER:
978 case Token.LITERAL_DECIMAL:
979 case Token.LITERAL_DOUBLE:
980 case Token.LITERAL_FLOAT:
981 case Token.TRUE:
982 case Token.FALSE:
983 case Token.NULL:
984 case Token.THIS:
985 case Token.NEW:
986 next_token = Token.INTERR;
987 break;
988
989 case Token.SEMICOLON:
990 case Token.COMMA:
991 case Token.CLOSE_PARENS:
992 case Token.OPEN_BRACKET:
993 case Token.OP_GENERICS_GT:
994 next_token = Token.INTERR_NULLABLE;
995 break;
996
997 default:
998 next_token = -1;
999 break;
1000 }
1001
1002 if (next_token == -1) {
1003 switch (xtoken ()) {
1004 case Token.COMMA:
1005 case Token.SEMICOLON:
1006 case Token.OPEN_BRACE:
1007 case Token.CLOSE_PARENS:
1008 case Token.IN:
1009 next_token = Token.INTERR_NULLABLE;
1010 break;
1011
1012 case Token.COLON:
1013 next_token = Token.INTERR;
1014 break;
1015
1016 default:
1017 int ntoken;
1018 int interrs = 1;
1019 int colons = 0;
1020 //
1021 // All shorcuts failed, do it hard way
1022 //
1023 while ((ntoken = xtoken ()) != Token.EOF) {
1024 if (ntoken == Token.SEMICOLON)
1025 break;
1026
1027 if (ntoken == Token.COLON) {
1028 if (++colons == interrs)
1029 break;
1030 continue;
1031 }
1032
1033 if (ntoken == Token.INTERR) {
1034 ++interrs;
1035 continue;
1036 }
1037 }
1038
1039 next_token = colons != interrs ? Token.INTERR_NULLABLE : Token.INTERR;
1040 break;
1041 }
1042 }
1043
1044 PopPosition ();
1045 return next_token;
1046 }
1047
1048 bool decimal_digits (int c)
1049 {
1050 int d;
1051 bool seen_digits = false;
1052
1053 if (c != -1){
1054 if (number_pos == max_number_size)
1055 Error_NumericConstantTooLong ();
1056 number_builder [number_pos++] = (char) c;
1057 }
1058
1059 //
1060 // We use peek_char2, because decimal_digits needs to do a
1061 // 2-character look-ahead (5.ToString for example).
1062 //
1063 while ((d = peek_char2 ()) != -1){
1064 if (d >= '0' && d <= '9'){
1065 if (number_pos == max_number_size)
1066 Error_NumericConstantTooLong ();
1067 number_builder [number_pos++] = (char) d;
1068 get_char ();
1069 seen_digits = true;
1070 } else
1071 break;
1072 }
1073
1074 return seen_digits;
1075 }
1076
1077 static bool is_hex (int e)
1078 {
1079 return (e >= '0' && e <= '9') || (e >= 'A' && e <= 'F') || (e >= 'a' && e <= 'f');
1080 }
1081
1082 static int real_type_suffix (int c)
1083 {
1084 int t;
1085
1086 switch (c){
1087 case 'F': case 'f':
1088 t = Token.LITERAL_FLOAT;
1089 break;
1090 case 'D': case 'd':
1091 t = Token.LITERAL_DOUBLE;
1092 break;
1093 case 'M': case 'm':
1094 t= Token.LITERAL_DECIMAL;
1095 break;
1096 default:
1097 return Token.NONE;
1098 }
1099 return t;
1100 }
1101
1102 int integer_type_suffix (ulong ul, int c)
1103 {
1104 bool is_unsigned = false;
1105 bool is_long = false;
1106
1107 if (c != -1){
1108 bool scanning = true;
1109 do {
1110 switch (c){
1111 case 'U': case 'u':
1112 if (is_unsigned)
1113 scanning = false;
1114 is_unsigned = true;
1115 get_char ();
1116 break;
1117
1118 case 'l':
1119 if (!is_unsigned){
1120 //
1121 // if we have not seen anything in between
1122 // report this error
1123 //
1124 Report.Warning (78, 4, Location, "The 'l' suffix is easily confused with the digit '1' (use 'L' for clarity)");
1125 }
1126 //
1127 // This goto statement causes the MS CLR 2.0 beta 1 csc to report an error, so
1128 // work around that.
1129 //
1130 //goto case 'L';
1131 if (is_long)
1132 scanning = false;
1133 is_long = true;
1134 get_char ();
1135 break;
1136
1137 case 'L':
1138 if (is_long)
1139 scanning = false;
1140 is_long = true;
1141 get_char ();
1142 break;
1143
1144 default:
1145 scanning = false;
1146 break;
1147 }
1148 c = peek_char ();
1149 } while (scanning);
1150 }
1151
1152 if (is_long && is_unsigned){
1153 val = ul;
1154 return Token.LITERAL_INTEGER;
1155 } else if (is_unsigned){
1156 // uint if possible, or ulong else.
1157
1158 if ((ul & 0xffffffff00000000) == 0)
1159 val = (uint) ul;
1160 else
1161 val = ul;
1162 } else if (is_long){
1163 // long if possible, ulong otherwise
1164 if ((ul & 0x8000000000000000) != 0)
1165 val = ul;
1166 else
1167 val = (long) ul;
1168 } else {
1169 // int, uint, long or ulong in that order
1170 if ((ul & 0xffffffff00000000) == 0){
1171 uint ui = (uint) ul;
1172
1173 if ((ui & 0x80000000) != 0)
1174 val = ui;
1175 else
1176 val = (int) ui;
1177 } else {
1178 if ((ul & 0x8000000000000000) != 0)
1179 val = ul;
1180 else
1181 val = (long) ul;
1182 }
1183 }
1184 return Token.LITERAL_INTEGER;
1185 }
1186
1187 //
1188 // given `c' as the next char in the input decide whether
1189 // we need to convert to a special type, and then choose
1190 // the best representation for the integer
1191 //
1192 int adjust_int (int c)
1193 {
1194 try {
1195 if (number_pos > 9){
1196 ulong ul = (uint) (number_builder [0] - '0');
1197
1198 for (int i = 1; i < number_pos; i++){
1199 ul = checked ((ul * 10) + ((uint)(number_builder [i] - '0')));
1200 }
1201 return integer_type_suffix (ul, c);
1202 } else {
1203 uint ui = (uint) (number_builder [0] - '0');
1204
1205 for (int i = 1; i < number_pos; i++){
1206 ui = checked ((ui * 10) + ((uint)(number_builder [i] - '0')));
1207 }
1208 return integer_type_suffix (ui, c);
1209 }
1210 } catch (OverflowException) {
1211 error_details = "Integral constant is too large";
1212 Report.Error (1021, Location, error_details);
1213 val = 0ul;
1214 return Token.LITERAL_INTEGER;
1215 }
1216 catch (FormatException) {
1217 Report.Error (1013, Location, "Invalid number");
1218 val = 0ul;
1219 return Token.LITERAL_INTEGER;
1220 }
1221 }
1222
1223 int adjust_real (int t)
1224 {
1225 string s = new String (number_builder, 0, number_pos);
1226 const string error_details = "Floating-point constant is outside the range of type `{0}'";
1227
1228 switch (t){
1229 case Token.LITERAL_DECIMAL:
1230 try {
1231 val = System.Decimal.Parse (s, styles, csharp_format_info);
1232 } catch (OverflowException) {
1233 val = 0m;
1234 Report.Error (594, Location, error_details, "decimal");
1235 }
1236 break;
1237 case Token.LITERAL_FLOAT:
1238 try {
1239 val = float.Parse (s, styles, csharp_format_info);
1240 } catch (OverflowException) {
1241 val = 0.0f;
1242 Report.Error (594, Location, error_details, "float");
1243 }
1244 break;
1245
1246 case Token.LITERAL_DOUBLE:
1247 case Token.NONE:
1248 t = Token.LITERAL_DOUBLE;
1249 try {
1250 val = System.Double.Parse (s, styles, csharp_format_info);
1251 } catch (OverflowException) {
1252 val = 0.0;
1253 Report.Error (594, Location, error_details, "double");
1254 }
1255 break;
1256 }
1257 return t;
1258 }
1259
1260 int handle_hex ()
1261 {
1262 int d;
1263 ulong ul;
1264
1265 get_char ();
1266 while ((d = peek_char ()) != -1){
1267 if (is_hex (d)){
1268 number_builder [number_pos++] = (char) d;
1269 get_char ();
1270 } else
1271 break;
1272 }
1273
1274 string s = new String (number_builder, 0, number_pos);
1275 try {
1276 if (number_pos <= 8)
1277 ul = System.UInt32.Parse (s, NumberStyles.HexNumber);
1278 else
1279 ul = System.UInt64.Parse (s, NumberStyles.HexNumber);
1280 } catch (OverflowException){
1281 error_details = "Integral constant is too large";
1282 Report.Error (1021, Location, error_details);
1283 val = 0ul;
1284 return Token.LITERAL_INTEGER;
1285 }
1286 catch (FormatException) {
1287 Report.Error (1013, Location, "Invalid number");
1288 val = 0ul;
1289 return Token.LITERAL_INTEGER;
1290 }
1291
1292 return integer_type_suffix (ul, peek_char ());
1293 }
1294
1295 //
1296 // Invoked if we know we have .digits or digits
1297 //
1298 int is_number (int c)
1299 {
1300 bool is_real = false;
1301 int type;
1302
1303 number_pos = 0;
1304
1305 if (c >= '0' && c <= '9'){
1306 if (c == '0'){
1307 int peek = peek_char ();
1308
1309 if (peek == 'x' || peek == 'X')
1310 return handle_hex ();
1311 }
1312 decimal_digits (c);
1313 c = get_char ();
1314 }
1315
1316 //
1317 // We need to handle the case of
1318 // "1.1" vs "1.string" (LITERAL_FLOAT vs NUMBER DOT IDENTIFIER)
1319 //
1320 if (c == '.'){
1321 if (decimal_digits ('.')){
1322 is_real = true;
1323 c = get_char ();
1324 } else {
1325 putback ('.');
1326 number_pos--;
1327 return adjust_int (-1);
1328 }
1329 }
1330
1331 if (c == 'e' || c == 'E'){
1332 is_real = true;
1333 if (number_pos == max_number_size)
1334 Error_NumericConstantTooLong ();
1335 number_builder [number_pos++] = 'e';
1336 c = get_char ();
1337
1338 if (c == '+'){
1339 if (number_pos == max_number_size)
1340 Error_NumericConstantTooLong ();
1341 number_builder [number_pos++] = '+';
1342 c = -1;
1343 } else if (c == '-') {
1344 if (number_pos == max_number_size)
1345 Error_NumericConstantTooLong ();
1346 number_builder [number_pos++] = '-';
1347 c = -1;
1348 } else {
1349 if (number_pos == max_number_size)
1350 Error_NumericConstantTooLong ();
1351 number_builder [number_pos++] = '+';
1352 }
1353
1354 decimal_digits (c);
1355 c = get_char ();
1356 }
1357
1358 type = real_type_suffix (c);
1359 if (type == Token.NONE && !is_real){
1360 putback (c);
1361 return adjust_int (c);
1362 } else
1363 is_real = true;
1364
1365 if (type == Token.NONE){
1366 putback (c);
1367 }
1368
1369 if (is_real)
1370 return adjust_real (type);
1371
1372 Console.WriteLine ("This should not be reached");
1373 throw new Exception ("Is Number should never reach this point");
1374 }
1375
1376 //
1377 // Accepts exactly count (4 or 8) hex, no more no less
1378 //
1379 int getHex (int count, out int surrogate, out bool error)
1380 {
1381 int i;
1382 int total = 0;
1383 int c;
1384 int top = count != -1 ? count : 4;
1385
1386 get_char ();
1387 error = false;
1388 surrogate = 0;
1389 for (i = 0; i < top; i++){
1390 c = get_char ();
1391
1392 if (c >= '0' && c <= '9')
1393 c = (int) c - (int) '0';
1394 else if (c >= 'A' && c <= 'F')
1395 c = (int) c - (int) 'A' + 10;
1396 else if (c >= 'a' && c <= 'f')
1397 c = (int) c - (int) 'a' + 10;
1398 else {
1399 error = true;
1400 return 0;
1401 }
1402
1403 total = (total * 16) + c;
1404 if (count == -1){
1405 int p = peek_char ();
1406 if (p == -1)
1407 break;
1408 if (!is_hex ((char)p))
1409 break;
1410 }
1411 }
1412
1413 if (top == 8) {
1414 if (total > 0x0010FFFF) {
1415 error = true;
1416 return 0;
1417 }
1418
1419 if (total >= 0x00010000) {
1420 surrogate = ((total - 0x00010000) % 0x0400 + 0xDC00);
1421 total = ((total - 0x00010000) / 0x0400 + 0xD800);
1422 }
1423 }
1424
1425 return total;
1426 }
1427
1428 int escape (int c, out int surrogate)
1429 {
1430 bool error;
1431 int d;
1432 int v;
1433
1434 d = peek_char ();
1435 if (c != '\\') {
1436 surrogate = 0;
1437 return c;
1438 }
1439
1440 switch (d){
1441 case 'a':
1442 v = '\a'; break;
1443 case 'b':
1444 v = '\b'; break;
1445 case 'n':
1446 v = '\n'; break;
1447 case 't':
1448 v = '\t'; break;
1449 case 'v':
1450 v = '\v'; break;
1451 case 'r':
1452 v = '\r'; break;
1453 case '\\':
1454 v = '\\'; break;
1455 case 'f':
1456 v = '\f'; break;
1457 case '0':
1458 v = 0; break;
1459 case '"':
1460 v = '"'; break;
1461 case '\'':
1462 v = '\''; break;
1463 case 'x':
1464 v = getHex (-1, out surrogate, out error);
1465 if (error)
1466 goto default;
1467 return v;
1468 case 'u':
1469 case 'U':
1470 return EscapeUnicode (d, out surrogate);
1471 default:
1472 surrogate = 0;
1473 Report.Error (1009, Location, "Unrecognized escape sequence `\\{0}'", ((char)d).ToString ());
1474 return d;
1475 }
1476
1477 get_char ();
1478 surrogate = 0;
1479 return v;
1480 }
1481
1482 int EscapeUnicode (int ch, out int surrogate)
1483 {
1484 bool error;
1485 if (ch == 'U') {
1486 ch = getHex (8, out surrogate, out error);
1487 } else {
1488 ch = getHex (4, out surrogate, out error);
1489 }
1490
1491 if (error)
1492 Report.Error (1009, Location, "Unrecognized escape sequence");
1493
1494 return ch;
1495 }
1496
1497 int get_char ()
1498 {
1499 int x;
1500 if (putback_char != -1) {
1501 x = putback_char;
1502 putback_char = -1;
1503 } else
1504 x = reader.Read ();
1505 if (x == '\n') {
1506 advance_line ();
1507 } else {
1508 col++;
1509 }
1510 return x;
1511 }
1512
1513 void advance_line ()
1514 {
1515 line++;
1516 ref_line++;
1517 previous_col = col;
1518 col = 0;
1519 }
1520
1521 int peek_char ()
1522 {
1523 if (putback_char == -1)
1524 putback_char = reader.Read ();
1525 return putback_char;
1526 }
1527
1528 int peek_char2 ()
1529 {
1530 if (putback_char != -1)
1531 return putback_char;
1532 return reader.Peek ();
1533 }
1534
1535 void putback (int c)
1536 {
1537 if (putback_char != -1){
1538 Console.WriteLine ("Col: " + col);
1539 Console.WriteLine ("Row: " + line);
1540 Console.WriteLine ("Name: " + ref_name.Name);
1541 Console.WriteLine ("Current [{0}] putting back [{1}] ", putback_char, c);
1542 throw new Exception ("This should not happen putback on putback");
1543 }
1544 if (c == '\n' || col == 0) {
1545 // It won't happen though.
1546 line--;
1547 ref_line--;
1548 col = previous_col;
1549 }
1550 else
1551 col--;
1552 putback_char = c;
1553 }
1554
1555 public bool advance ()
1556 {
1557 return peek_char () != -1 || CompleteOnEOF;
1558 }
1559
1560 public Object Value {
1561 get {
1562 return val;
1563 }
1564 }
1565
1566 public Object value ()
1567 {
1568 return val;
1569 }
1570
1571 public int token ()
1572 {
1573 current_token = xtoken ();
1574 return current_token;
1575 }
1576
1577 static StringBuilder static_cmd_arg = new System.Text.StringBuilder ();
1578
1579 void get_cmd_arg (out string cmd, out string arg)
1580 {
1581 int c;
1582
1583 tokens_seen = false;
1584 arg = "";
1585
1586 // skip over white space
1587 do {
1588 c = get_char ();
1589 } while (c == '\r' || c == ' ' || c == '\t');
1590
1591 static_cmd_arg.Length = 0;
1592 while (c != -1 && is_identifier_part_character ((char)c)) {
1593 static_cmd_arg.Append ((char)c);
1594 c = get_char ();
1595 if (c == '\\') {
1596 int peek = peek_char ();
1597 if (peek == 'U' || peek == 'u') {
1598 int surrogate;
1599 c = EscapeUnicode (c, out surrogate);
1600 if (surrogate != 0) {
1601 if (is_identifier_part_character ((char) c))
1602 static_cmd_arg.Append ((char) c);
1603 c = surrogate;
1604 }
1605 }
1606 }
1607 }
1608
1609 cmd = static_cmd_arg.ToString ();
1610
1611 // skip over white space
1612 while (c == '\r' || c == ' ' || c == '\t')
1613 c = get_char ();
1614
1615 static_cmd_arg.Length = 0;
1616 int has_identifier_argument = 0;
1617
1618 while (c != -1 && c != '\n' && c != '\r') {
1619 if (c == '\\' && has_identifier_argument >= 0) {
1620 if (has_identifier_argument != 0 || (cmd == "define" || cmd == "if" || cmd == "elif" || cmd == "undef")) {
1621 has_identifier_argument = 1;
1622
1623 int peek = peek_char ();
1624 if (peek == 'U' || peek == 'u') {
1625 int surrogate;
1626 c = EscapeUnicode (c, out surrogate);
1627 if (surrogate != 0) {
1628 if (is_identifier_part_character ((char) c))
1629 static_cmd_arg.Append ((char) c);
1630 c = surrogate;
1631 }
1632 }
1633 } else {
1634 has_identifier_argument = -1;
1635 }
1636 }
1637 static_cmd_arg.Append ((char) c);
1638 c = get_char ();
1639 }
1640
1641 if (static_cmd_arg.Length != 0)
1642 arg = static_cmd_arg.ToString ();
1643 }
1644
1645 //
1646 // Handles the #line directive
1647 //
1648 bool PreProcessLine (string arg)
1649 {
1650 if (arg.Length == 0)
1651 return false;
1652
1653 if (arg == "default"){
1654 ref_line = line;
1655 ref_name = file_name;
1656 hidden = false;
1657 Location.Push (file_name, ref_name);
1658 return true;
1659 } else if (arg == "hidden"){
1660 hidden = true;
1661 return true;
1662 }
1663
1664 try {
1665 int pos;
1666
1667 if ((pos = arg.IndexOf (' ')) != -1 && pos != 0){
1668 ref_line = System.Int32.Parse (arg.Substring (0, pos));
1669 pos++;
1670
1671 char [] quotes = { '\"' };
1672
1673 string name = arg.Substring (pos). Trim (quotes);
1674 ref_name = Location.LookupFile (file_name, name);
1675 file_name.AddFile (ref_name);
1676 hidden = false;
1677 Location.Push (file_name, ref_name);
1678 } else {
1679 ref_line = System.Int32.Parse (arg);
1680 hidden = false;
1681 }
1682 } catch {
1683 return false;
1684 }
1685
1686 return true;
1687 }
1688
1689 //
1690 // Handles #define and #undef
1691 //
1692 void PreProcessDefinition (bool is_define, string ident, bool caller_is_taking)
1693 {
1694 if (ident.Length == 0 || ident == "true" || ident == "false"){
1695 Report.Error (1001, Location, "Missing identifier to pre-processor directive");
1696 return;
1697 }
1698
1699 if (ident.IndexOfAny (simple_whitespaces) != -1){
1700 Error_EndLineExpected ();
1701 return;
1702 }
1703
1704 if (!is_identifier_start_character (ident [0]))
1705 Report.Error (1001, Location, "Identifier expected: {0}", ident);
1706
1707 foreach (char c in ident.Substring (1)){
1708 if (!is_identifier_part_character (c)){
1709 Report.Error (1001, Location, "Identifier expected: {0}", ident);
1710 return;
1711 }
1712 }
1713
1714 if (!caller_is_taking)
1715 return;
1716
1717 if (is_define) {
1718 //
1719 // #define ident
1720 //
1721 if (RootContext.IsConditionalDefined (ident))
1722 return;
1723
1724 file_name.AddDefine (ident);
1725 } else {
1726 //
1727 // #undef ident
1728 //
1729 file_name.AddUndefine (ident);
1730 }
1731 }
1732
1733 static byte read_hex (string arg, int pos, out bool error)
1734 {
1735 error = false;
1736
1737 int total;
1738 char c = arg [pos];
1739
1740 if ((c >= '0') && (c <= '9'))
1741 total = (int) c - (int) '0';
1742 else if ((c >= 'A') && (c <= 'F'))
1743 total = (int) c - (int) 'A' + 10;
1744 else if ((c >= 'a') && (c <= 'f'))
1745 total = (int) c - (int) 'a' + 10;
1746 else {
1747 error = true;
1748 return 0;
1749 }
1750
1751 total *= 16;
1752 c = arg [pos+1];
1753
1754 if ((c >= '0') && (c <= '9'))
1755 total += (int) c - (int) '0';
1756 else if ((c >= 'A') && (c <= 'F'))
1757 total += (int) c - (int) 'A' + 10;
1758 else if ((c >= 'a') && (c <= 'f'))
1759 total += (int) c - (int) 'a' + 10;
1760 else {
1761 error = true;
1762 return 0;
1763 }
1764
1765 return (byte) total;
1766 }
1767
1768 /// <summary>
1769 /// Handles #pragma checksum
1770 /// </summary>
1771 bool PreProcessPragmaChecksum (string arg)
1772 {
1773 if ((arg [0] != ' ') && (arg [0] != '\t'))
1774 return false;
1775
1776 arg = arg.Trim (simple_whitespaces);
1777 if ((arg.Length < 2) || (arg [0] != '"'))
1778 return false;
1779
1780 StringBuilder file_sb = new StringBuilder ();
1781
1782 int pos = 1;
1783 char ch;
1784 while ((ch = arg [pos++]) != '"') {
1785 if (pos >= arg.Length)
1786 return false;
1787
1788 if (ch == '\\') {
1789 if (pos+1 >= arg.Length)
1790 return false;
1791 ch = arg [pos++];
1792 }
1793
1794 file_sb.Append (ch);
1795 }
1796
1797 if ((pos+2 >= arg.Length) || ((arg [pos] != ' ') && (arg [pos] != '\t')))
1798 return false;
1799
1800 arg = arg.Substring (pos).Trim (simple_whitespaces);
1801 if ((arg.Length < 42) || (arg [0] != '"') || (arg [1] != '{') ||
1802 (arg [10] != '-') || (arg [15] != '-') || (arg [20] != '-') ||
1803 (arg [25] != '-') || (arg [38] != '}') || (arg [39] != '"'))
1804 return false;
1805
1806 bool error;
1807 byte[] guid_bytes = new byte [16];
1808
1809 for (int i = 0; i < 4; i++) {
1810 guid_bytes [i] = read_hex (arg, 2+2*i, out error);
1811 if (error)
1812 return false;
1813 }
1814 for (int i = 0; i < 2; i++) {
1815 guid_bytes [i+4] = read_hex (arg, 11+2*i, out error);
1816 if (error)
1817 return false;
1818 guid_bytes [i+6] = read_hex (arg, 16+2*i, out error);
1819 if (error)
1820 return false;
1821 guid_bytes [i+8] = read_hex (arg, 21+2*i, out error);
1822 if (error)
1823 return false;
1824 }
1825
1826 for (int i = 0; i < 6; i++) {
1827 guid_bytes [i+10] = read_hex (arg, 26+2*i, out error);
1828 if (error)
1829 return false;
1830 }
1831
1832 arg = arg.Substring (40).Trim (simple_whitespaces);
1833 if ((arg.Length < 34) || (arg [0] != '"') || (arg [33] != '"'))
1834 return false;
1835
1836 byte[] checksum_bytes = new byte [16];
1837 for (int i = 0; i < 16; i++) {
1838 checksum_bytes [i] = read_hex (arg, 1+2*i, out error);
1839 if (error)
1840 return false;
1841 }
1842
1843 arg = arg.Substring (34).Trim (simple_whitespaces);
1844 if (arg.Length > 0)
1845 return false;
1846
1847 SourceFile file = Location.LookupFile (file_name, file_sb.ToString ());
1848 file.SetChecksum (guid_bytes, checksum_bytes);
1849 ref_name.AutoGenerated = true;
1850 return true;
1851 }
1852
1853 /// <summary>
1854 /// Handles #pragma directive
1855 /// </summary>
1856 void PreProcessPragma (string arg)
1857 {
1858 const string warning = "warning";
1859 const string w_disable = "warning disable";
1860 const string w_restore = "warning restore";
1861 const string checksum = "checksum";
1862
1863 if (arg == w_disable) {
1864 Report.RegisterWarningRegion (Location).WarningDisable (Location.Row);
1865 return;
1866 }
1867
1868 if (arg == w_restore) {
1869 Report.RegisterWarningRegion (Location).WarningEnable (Location.Row);
1870 return;
1871 }
1872
1873 if (arg.StartsWith (w_disable)) {
1874 int[] codes = ParseNumbers (arg.Substring (w_disable.Length));
1875 foreach (int code in codes) {
1876 if (code != 0)
1877 Report.RegisterWarningRegion (Location).WarningDisable (Location, code);
1878 }
1879 return;
1880 }
1881
1882 if (arg.StartsWith (w_restore)) {
1883 int[] codes = ParseNumbers (arg.Substring (w_restore.Length));
1884 Hashtable w_table = Report.warning_ignore_table;
1885 foreach (int code in codes) {
1886 if (w_table != null && w_table.Contains (code))
1887 Report.Warning (1635, 1, Location, String.Format ("Cannot restore warning `CS{0:0000}' because it was disabled globally", code));
1888 Report.RegisterWarningRegion (Location).WarningEnable (Location, code);
1889 }
1890 return;
1891 }
1892
1893 if (arg.StartsWith (warning)) {
1894 Report.Warning (1634, 1, Location, "Expected disable or restore");
1895 return;
1896 }
1897
1898 if (arg.StartsWith (checksum)) {
1899 if (!PreProcessPragmaChecksum (arg.Substring (checksum.Length)))
1900 Warning_InvalidPragmaChecksum ();
1901 return;
1902 }
1903
1904 Report.Warning (1633, 1, Location, "Unrecognized #pragma directive");
1905 }
1906
1907 int[] ParseNumbers (string text)
1908 {
1909 string[] string_array = text.Split (',');
1910 int[] values = new int [string_array.Length];
1911 int index = 0;
1912 foreach (string string_code in string_array) {
1913 try {
1914 values[index++] = int.Parse (string_code, System.Globalization.CultureInfo.InvariantCulture);
1915 }
1916 catch (FormatException) {
1917 Report.Warning (1692, 1, Location, "Invalid number");
1918 }
1919 }
1920 return values;
1921 }
1922
1923 bool eval_val (string s)
1924 {
1925 if (s == "true")
1926 return true;
1927 if (s == "false")
1928 return false;
1929
1930 return file_name.IsConditionalDefined (s);
1931 }
1932
1933 bool pp_primary (ref string s)
1934 {
1935 s = s.Trim ();
1936 int len = s.Length;
1937
1938 if (len > 0){
1939 char c = s [0];
1940
1941 if (c == '('){
1942 s = s.Substring (1);
1943 bool val = pp_expr (ref s, false);
1944 if (s.Length > 0 && s [0] == ')'){
1945 s = s.Substring (1);
1946 return val;
1947 }
1948 Error_InvalidDirective ();
1949 return false;
1950 }
1951
1952 if (is_identifier_start_character (c)){
1953 int j = 1;
1954
1955 while (j < len){
1956 c = s [j];
1957
1958 if (is_identifier_part_character (c)){
1959 j++;
1960 continue;
1961 }
1962 bool v = eval_val (s.Substring (0, j));
1963 s = s.Substring (j);
1964 return v;
1965 }
1966 bool vv = eval_val (s);
1967 s = "";
1968 return vv;
1969 }
1970 }
1971 Error_InvalidDirective ();
1972 return false;
1973 }
1974
1975 bool pp_unary (ref string s)
1976 {
1977 s = s.Trim ();
1978 int len = s.Length;
1979
1980 if (len > 0){
1981 if (s [0] == '!'){
1982 if (len > 1 && s [1] == '='){
1983 Error_InvalidDirective ();
1984 return false;
1985 }
1986 s = s.Substring (1);
1987 return ! pp_primary (ref s);
1988 } else
1989 return pp_primary (ref s);
1990 } else {
1991 Error_InvalidDirective ();
1992 return false;
1993 }
1994 }
1995
1996 bool pp_eq (ref string s)
1997 {
1998 bool va = pp_unary (ref s);
1999
2000 s = s.Trim ();
2001 int len = s.Length;
2002 if (len > 0){
2003 if (s [0] == '='){
2004 if (len > 2 && s [1] == '='){
2005 s = s.Substring (2);
2006 return va == pp_unary (ref s);
2007 } else {
2008 Error_InvalidDirective ();
2009 return false;
2010 }
2011 } else if (s [0] == '!' && len > 1 && s [1] == '='){
2012 s = s.Substring (2);
2013
2014 return va != pp_unary (ref s);
2015
2016 }
2017 }
2018
2019 return va;
2020
2021 }
2022
2023 bool pp_and (ref string s)
2024 {
2025 bool va = pp_eq (ref s);
2026
2027 s = s.Trim ();
2028 int len = s.Length;
2029 if (len > 0){
2030 if (s [0] == '&'){
2031 if (len > 2 && s [1] == '&'){
2032 s = s.Substring (2);
2033 return (va & pp_and (ref s));
2034 } else {
2035 Error_InvalidDirective ();
2036 return false;
2037 }
2038 }
2039 }
2040 return va;
2041 }
2042
2043 //
2044 // Evaluates an expression for `#if' or `#elif'
2045 //
2046 bool pp_expr (ref string s, bool isTerm)
2047 {
2048 bool va = pp_and (ref s);
2049 s = s.Trim ();
2050 int len = s.Length;
2051 if (len > 0){
2052 char c = s [0];
2053
2054 if (c == '|'){
2055 if (len > 2 && s [1] == '|'){
2056 s = s.Substring (2);
2057 return va | pp_expr (ref s, isTerm);
2058 } else {
2059 Error_InvalidDirective ();
2060 return false;
2061 }
2062 }
2063 if (isTerm) {
2064 Error_EndLineExpected ();
2065 return false;
2066 }
2067 }
2068
2069 return va;
2070 }
2071
2072 bool eval (string s)
2073 {
2074 bool v = pp_expr (ref s, true);
2075 s = s.Trim ();
2076 if (s.Length != 0){
2077 return false;
2078 }
2079
2080 return v;
2081 }
2082
2083 void Error_NumericConstantTooLong ()
2084 {
2085 Report.Error (1021, Location, "Numeric constant too long");
2086 }
2087
2088 void Error_InvalidDirective ()
2089 {
2090 Report.Error (1517, Location, "Invalid preprocessor directive");
2091 }
2092
2093 void Error_UnexpectedDirective (string extra)
2094 {
2095 Report.Error (
2096 1028, Location,
2097 "Unexpected processor directive ({0})", extra);
2098 }
2099
2100 void Error_TokensSeen ()
2101 {
2102 Report.Error (1032, Location,
2103 "Cannot define or undefine preprocessor symbols after first token in file");
2104 }
2105
2106 void Eror_WrongPreprocessorLocation ()
2107 {
2108 Report.Error (1040, Location,
2109 "Preprocessor directives must appear as the first non-whitespace character on a line");
2110 }
2111
2112 void Error_EndLineExpected ()
2113 {
2114 Report.Error (1025, Location, "Single-line comment or end-of-line expected");
2115 }
2116
2117 void Warning_InvalidPragmaChecksum ()
2118 {
2119 Report.Warning (1695, 1, Location,
2120 "Invalid #pragma checksum syntax; should be " +
2121 "#pragma checksum \"filename\" " +
2122 "\"{XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX}\" \"XXXX...\"");
2123 }
2124 //
2125 // if true, then the code continues processing the code
2126 // if false, the code stays in a loop until another directive is
2127 // reached.
2128 // When caller_is_taking is false we ignore all directives except the ones
2129 // which can help us to identify where the #if block ends
2130 bool handle_preprocessing_directive (bool caller_is_taking)
2131 {
2132 string cmd, arg;
2133 bool region_directive = false;
2134
2135 get_cmd_arg (out cmd, out arg);
2136
2137 // Eat any trailing whitespaces and single-line comments
2138 if (arg.IndexOf ("//") != -1)
2139 arg = arg.Substring (0, arg.IndexOf ("//"));
2140 arg = arg.Trim (simple_whitespaces);
2141
2142 //
2143 // The first group of pre-processing instructions is always processed
2144 //
2145 switch (cmd){
2146 case "region":
2147 region_directive = true;
2148 arg = "true";
2149 goto case "if";
2150
2151 case "endregion":
2152 if (ifstack == null || ifstack.Count == 0){
2153 Error_UnexpectedDirective ("no #region for this #endregion");
2154 return true;
2155 }
2156 int pop = (int) ifstack.Pop ();
2157
2158 if ((pop & REGION) == 0)
2159 Report.Error (1027, Location, "Expected `#endif' directive");
2160
2161 return caller_is_taking;
2162
2163 case "if":
2164 if (ifstack == null)
2165 ifstack = new Stack (2);
2166
2167 int flags = region_directive ? REGION : 0;
2168 if (ifstack.Count == 0){
2169 flags |= PARENT_TAKING;
2170 } else {
2171 int state = (int) ifstack.Peek ();
2172 if ((state & TAKING) != 0) {
2173 flags |= PARENT_TAKING;
2174 }
2175 }
2176
2177 if (caller_is_taking && eval (arg)) {
2178 ifstack.Push (flags | TAKING);
2179 return true;
2180 }
2181 ifstack.Push (flags);
2182 return false;
2183
2184 case "endif":
2185 if (ifstack == null || ifstack.Count == 0){
2186 Error_UnexpectedDirective ("no #if for this #endif");
2187 return true;
2188 } else {
2189 pop = (int) ifstack.Pop ();
2190
2191 if ((pop & REGION) != 0)
2192 Report.Error (1038, Location, "#endregion directive expected");
2193
2194 if (arg.Length != 0) {
2195 Error_EndLineExpected ();
2196 }
2197
2198 if (ifstack.Count == 0)
2199 return true;
2200
2201 int state = (int) ifstack.Peek ();
2202 return (state & TAKING) != 0;
2203 }
2204
2205 case "elif":
2206 if (ifstack == null || ifstack.Count == 0){
2207 Error_UnexpectedDirective ("no #if for this #elif");
2208 return true;
2209 } else {
2210 int state = (int) ifstack.Pop ();
2211
2212 if ((state & REGION) != 0) {
2213 Report.Error (1038, Location, "#endregion directive expected");
2214 return true;
2215 }
2216
2217 if ((state & ELSE_SEEN) != 0){
2218 Error_UnexpectedDirective ("#elif not valid after #else");
2219 return true;
2220 }
2221
2222 if ((state & TAKING) != 0) {
2223 ifstack.Push (0);
2224 return false;
2225 }
2226
2227 if (eval (arg) && ((state & PARENT_TAKING) != 0)){
2228 ifstack.Push (state | TAKING);
2229 return true;
2230 }
2231
2232 ifstack.Push (state);
2233 return false;
2234 }
2235
2236 case "else":
2237 if (ifstack == null || ifstack.Count == 0){
2238 Error_UnexpectedDirective ("no #if for this #else");
2239 return true;
2240 } else {
2241 int state = (int) ifstack.Peek ();
2242
2243 if ((state & REGION) != 0) {
2244 Report.Error (1038, Location, "#endregion directive expected");
2245 return true;
2246 }
2247
2248 if ((state & ELSE_SEEN) != 0){
2249 Error_UnexpectedDirective ("#else within #else");
2250 return true;
2251 }
2252
2253 ifstack.Pop ();
2254
2255 if (arg.Length != 0) {
2256 Error_EndLineExpected ();
2257 return true;
2258 }
2259
2260 bool ret = false;
2261 if ((state & PARENT_TAKING) != 0) {
2262 ret = (state & TAKING) == 0;
2263
2264 if (ret)
2265 state |= TAKING;
2266 else
2267 state &= ~TAKING;
2268 }
2269
2270 ifstack.Push (state | ELSE_SEEN);
2271
2272 return ret;
2273 }
2274 case "define":
2275 if (any_token_seen){
2276 Error_TokensSeen ();
2277 return caller_is_taking;
2278 }
2279 PreProcessDefinition (true, arg, caller_is_taking);
2280 return caller_is_taking;
2281
2282 case "undef":
2283 if (any_token_seen){
2284 Error_TokensSeen ();
2285 return caller_is_taking;
2286 }
2287 PreProcessDefinition (false, arg, caller_is_taking);
2288 return caller_is_taking;
2289 }
2290
2291 //
2292 // These are only processed if we are in a `taking' block
2293 //
2294 if (!caller_is_taking)
2295 return false;
2296
2297 switch (cmd){
2298 case "error":
2299 Report.Error (1029, Location, "#error: '{0}'", arg);
2300 return true;
2301
2302 case "warning":
2303 Report.Warning (1030, 1, Location, "#warning: `{0}'", arg);
2304 return true;
2305
2306 case "pragma":
2307 if (RootContext.Version == LanguageVersion.ISO_1) {
2308 Report.FeatureIsNotAvailable (Location, "#pragma");
2309 return true;
2310 }
2311
2312 PreProcessPragma (arg);
2313 return true;
2314
2315 case "line":
2316 if (!PreProcessLine (arg))
2317 Report.Error (
2318 1576, Location,
2319 "The line number specified for #line directive is missing or invalid");
2320 return caller_is_taking;
2321 }
2322
2323 Report.Error (1024, Location, "Wrong preprocessor directive");
2324 return true;
2325
2326 }
2327
2328 private int consume_string (bool quoted)
2329 {
2330 int c;
2331 string_builder.Length = 0;
2332
2333 while ((c = get_char ()) != -1){
2334 if (c == '"'){
2335 if (quoted && peek_char () == '"'){
2336 string_builder.Append ((char) c);
2337 get_char ();
2338 continue;
2339 } else {
2340 val = string_builder.ToString ();
2341 return Token.LITERAL_STRING;
2342 }
2343 }
2344
2345 if (c == '\n'){
2346 if (!quoted)
2347 Report.Error (1010, Location, "Newline in constant");
2348 }
2349
2350 if (!quoted){
2351 int surrogate;
2352 c = escape (c, out surrogate);
2353 if (c == -1)
2354 return Token.ERROR;
2355 if (surrogate != 0) {
2356 string_builder.Append ((char) c);
2357 c = surrogate;
2358 }
2359 }
2360 string_builder.Append ((char) c);
2361 }
2362
2363 Report.Error (1039, Location, "Unterminated string literal");
2364 return Token.EOF;
2365 }
2366
2367 private int consume_identifier (int s)
2368 {
2369 int res = consume_identifier (s, false);
2370
2371 if (doc_state == XmlCommentState.Allowed)
2372 doc_state = XmlCommentState.NotAllowed;
2373
2374 return res;
2375 }
2376
2377 private int consume_identifier (int c, bool quoted)
2378 {
2379 int pos = 0;
2380
2381 if (c == '\\') {
2382 int surrogate;
2383 c = escape (c, out surrogate);
2384 if (surrogate != 0) {
2385 id_builder [pos++] = (char) c;
2386 c = surrogate;
2387 }
2388 }
2389
2390 id_builder [pos++] = (char) c;
2391 Location loc = Location;
2392
2393 while ((c = get_char ()) != -1) {
2394 loop:
2395 if (is_identifier_part_character ((char) c)){
2396 if (pos == max_id_size){
2397 Report.Error (645, loc, "Identifier too long (limit is 512 chars)");
2398 return Token.ERROR;
2399 }
2400
2401 id_builder [pos++] = (char) c;
2402 } else if (c == '\\') {
2403 int surrogate;
2404 c = escape (c, out surrogate);
2405 if (surrogate != 0) {
2406 if (is_identifier_part_character ((char) c))
2407 id_builder [pos++] = (char) c;
2408 c = surrogate;
2409 }
2410 goto loop;
2411 } else {
2412 putback (c);
2413 break;
2414 }
2415 }
2416
2417 //
2418 // Optimization: avoids doing the keyword lookup
2419 // on uppercase letters
2420 //
2421 if (id_builder [0] >= '_' && !quoted) {
2422 int keyword = GetKeyword (id_builder, pos);
2423 if (keyword != -1) {
2424 // TODO: No need to store location for keyword, required location cleanup
2425 val = loc;
2426 return keyword;
2427 }
2428 }
2429
2430 //
2431 // Keep identifiers in an array of hashtables to avoid needless
2432 // allocations
2433 //
2434 CharArrayHashtable identifiers_group = identifiers [pos];
2435 if (identifiers_group != null) {
2436 val = identifiers_group [id_builder];
2437 if (val != null) {
2438 val = new LocatedToken (loc, (string) val);
2439 if (quoted)
2440 AddEscapedIdentifier ((LocatedToken) val);
2441 return Token.IDENTIFIER;
2442 }
2443 } else {
2444 identifiers_group = new CharArrayHashtable (pos);
2445 identifiers [pos] = identifiers_group;
2446 }
2447
2448 char [] chars = new char [pos];
2449 Array.Copy (id_builder, chars, pos);
2450
2451 val = new String (id_builder, 0, pos);
2452 identifiers_group.Add (chars, val);
2453
2454 if (RootContext.Version == LanguageVersion.ISO_1) {
2455 for (int i = 1; i < chars.Length; i += 3) {
2456 if (chars [i] == '_' && (chars [i - 1] == '_' || chars [i + 1] == '_')) {
2457 Report.Error (1638, loc,
2458 "`{0}': Any identifier with double underscores cannot be used when ISO language version mode is specified", val.ToString ());
2459 }
2460 }
2461 }
2462
2463 val = new LocatedToken (loc, (string) val);
2464 if (quoted)
2465 AddEscapedIdentifier ((LocatedToken) val);
2466 return Token.IDENTIFIER;
2467 }
2468
2469 public int xtoken ()
2470 {
2471 int d, c;
2472
2473 // Whether we have seen comments on the current line
2474 bool comments_seen = false;
2475 while ((c = get_char ()) != -1) {
2476 switch (c) {
2477 case '\t':
2478 col = ((col + 8) / 8) * 8;
2479 continue;
2480
2481 case ' ':
2482 case '\f':
2483 case '\v':
2484 case 0xa0:
2485 case 0:
2486 case 0xFEFF: // Ignore BOM anywhere in the file
2487 continue;
2488
2489 /* This is required for compatibility with .NET
2490 case 0xEF:
2491 if (peek_char () == 0xBB) {
2492 PushPosition ();
2493 get_char ();
2494 if (get_char () == 0xBF)
2495 continue;
2496 PopPosition ();
2497 }
2498 break;
2499 */
2500 case '\r':
2501 if (peek_char () != '\n')
2502 advance_line ();
2503 else
2504 get_char ();
2505
2506 any_token_seen |= tokens_seen;
2507 tokens_seen = false;
2508 comments_seen = false;
2509 continue;
2510
2511 case '\\':
2512 tokens_seen = true;
2513 return consume_identifier (c);
2514
2515 case '{':
2516 val = Location;
2517 return Token.OPEN_BRACE;
2518 case '}':
2519 val = Location;
2520 return Token.CLOSE_BRACE;
2521 case '[':
2522 // To block doccomment inside attribute declaration.
2523 if (doc_state == XmlCommentState.Allowed)
2524 doc_state = XmlCommentState.NotAllowed;
2525 return Token.OPEN_BRACKET;
2526 case ']':
2527 return Token.CLOSE_BRACKET;
2528 case '(':
2529 val = Location;
2530 //
2531 // An expression versions of parens can appear in block context only
2532 //
2533 if (parsing_block != 0 && !lambda_arguments_parsing) {
2534
2535 //
2536 // Optmize most common case where we know that parens
2537 // is not special
2538 //
2539 switch (current_token) {
2540 case Token.IDENTIFIER:
2541 case Token.IF:
2542 case Token.FOR:
2543 case Token.FOREACH:
2544 case Token.TYPEOF:
2545 case Token.WHILE:
2546 case Token.USING:
2547 case Token.DEFAULT:
2548 case Token.DELEGATE:
2549 case Token.OP_GENERICS_GT:
2550 return Token.OPEN_PARENS;
2551 }
2552
2553 // Optimize using peek
2554 int xx = peek_char ();
2555 switch (xx) {
2556 case '(':
2557 case '\'':
2558 case '"':
2559 case '0':
2560 case '1':
2561 return Token.OPEN_PARENS;
2562 }
2563
2564 lambda_arguments_parsing = true;
2565 PushPosition ();
2566 d = TokenizeOpenParens ();
2567 PopPosition ();
2568 lambda_arguments_parsing = false;
2569 return d;
2570 }
2571
2572 return Token.OPEN_PARENS;
2573 case ')':
2574 return Token.CLOSE_PARENS;
2575 case ',':
2576 return Token.COMMA;
2577 case ';':
2578 return Token.SEMICOLON;
2579 case '~':
2580 return Token.TILDE;
2581 case '?':
2582 return TokenizePossibleNullableType ();
2583 case '<':
2584 if (parsing_generic_less_than++ > 0)
2585 return Token.OP_GENERICS_LT;
2586
2587 return TokenizeLessThan ();
2588
2589 case '>':
2590 d = peek_char ();
2591
2592 if (d == '='){
2593 get_char ();
2594 return Token.OP_GE;
2595 }
2596
2597 if (parsing_generic_less_than > 1 || (parsing_generic_less_than == 1 && d != '>')) {
2598 parsing_generic_less_than--;
2599 return Token.OP_GENERICS_GT;
2600 }
2601
2602 if (d == '>') {
2603 get_char ();
2604 d = peek_char ();
2605
2606 if (d == '=') {
2607 get_char ();
2608 return Token.OP_SHIFT_RIGHT_ASSIGN;
2609 }
2610 return Token.OP_SHIFT_RIGHT;
2611 }
2612
2613 return Token.OP_GT;
2614
2615 case '+':
2616 d = peek_char ();
2617 if (d == '+') {
2618 d = Token.OP_INC;
2619 } else if (d == '=') {
2620 d = Token.OP_ADD_ASSIGN;
2621 } else {
2622 return Token.PLUS;
2623 }
2624 get_char ();
2625 return d;
2626
2627 case '-':
2628 d = peek_char ();
2629 if (d == '-') {
2630 d = Token.OP_DEC;
2631 } else if (d == '=')
2632 d = Token.OP_SUB_ASSIGN;
2633 else if (d == '>')
2634 d = Token.OP_PTR;
2635 else {
2636 return Token.MINUS;
2637 }
2638 get_char ();
2639 return d;
2640
2641 case '!':
2642 if (peek_char () == '='){
2643 get_char ();
2644 return Token.OP_NE;
2645 }
2646 return Token.BANG;
2647
2648 case '=':
2649 d = peek_char ();
2650 if (d == '='){
2651 get_char ();
2652 return Token.OP_EQ;
2653 }
2654 if (d == '>'){
2655 get_char ();
2656 return Token.ARROW;
2657 }
2658
2659 return Token.ASSIGN;
2660
2661 case '&':
2662 d = peek_char ();
2663 if (d == '&'){
2664 get_char ();
2665 return Token.OP_AND;
2666 }
2667 if (d == '='){
2668 get_char ();
2669 return Token.OP_AND_ASSIGN;
2670 }
2671 return Token.BITWISE_AND;
2672
2673 case '|':
2674 d = peek_char ();
2675 if (d == '|'){
2676 get_char ();
2677 return Token.OP_OR;
2678 }
2679 if (d == '='){
2680 get_char ();
2681 return Token.OP_OR_ASSIGN;
2682 }
2683 return Token.BITWISE_OR;
2684
2685 case '*':
2686 if (peek_char () == '='){
2687 get_char ();
2688 return Token.OP_MULT_ASSIGN;
2689 }
2690 val = Location;
2691 return Token.STAR;
2692
2693 case '/':
2694 d = peek_char ();
2695 if (d == '='){
2696 get_char ();
2697 return Token.OP_DIV_ASSIGN;
2698 }
2699
2700 // Handle double-slash comments.
2701 if (d == '/'){
2702 get_char ();
2703 if (RootContext.Documentation != null && peek_char () == '/') {
2704 get_char ();
2705 // Don't allow ////.
2706 if ((d = peek_char ()) != '/') {
2707 update_comment_location ();
2708 if (doc_state == XmlCommentState.Allowed)
2709 handle_one_line_xml_comment ();
2710 else if (doc_state == XmlCommentState.NotAllowed)
2711 warn_incorrect_doc_comment ();
2712 }
2713 }
2714 while ((d = get_char ()) != -1 && (d != '\n') && d != '\r');
2715
2716 any_token_seen |= tokens_seen;
2717 tokens_seen = false;
2718 comments_seen = false;
2719 continue;
2720 } else if (d == '*'){
2721 get_char ();
2722 bool docAppend = false;
2723 if (RootContext.Documentation != null && peek_char () == '*') {
2724 get_char ();
2725 update_comment_location ();
2726 // But when it is /**/, just do nothing.
2727 if (peek_char () == '/') {
2728 get_char ();
2729 continue;
2730 }
2731 if (doc_state == XmlCommentState.Allowed)
2732 docAppend = true;
2733 else if (doc_state == XmlCommentState.NotAllowed)
2734 warn_incorrect_doc_comment ();
2735 }
2736
2737 int current_comment_start = 0;
2738 if (docAppend) {
2739 current_comment_start = xml_comment_buffer.Length;
2740 xml_comment_buffer.Append (Environment.NewLine);
2741 }
2742
2743 while ((d = get_char ()) != -1){
2744 if (d == '*' && peek_char () == '/'){
2745 get_char ();
2746 comments_seen = true;
2747 break;
2748 }
2749 if (docAppend)
2750 xml_comment_buffer.Append ((char) d);
2751
2752 if (d == '\n'){
2753 any_token_seen |= tokens_seen;
2754 tokens_seen = false;
2755 //
2756 // Reset 'comments_seen' just to be consistent.
2757 // It doesn't matter either way, here.
2758 //
2759 comments_seen = false;
2760 }
2761 }
2762 if (!comments_seen)
2763 Report.Error (1035, Location, "End-of-file found, '*/' expected");
2764
2765 if (docAppend)
2766 update_formatted_doc_comment (current_comment_start);
2767 continue;
2768 }
2769 return Token.DIV;
2770
2771 case '%':
2772 if (peek_char () == '='){
2773 get_char ();
2774 return Token.OP_MOD_ASSIGN;
2775 }
2776 return Token.PERCENT;
2777
2778 case '^':
2779 if (peek_char () == '='){
2780 get_char ();
2781 return Token.OP_XOR_ASSIGN;
2782 }
2783 return Token.CARRET;
2784
2785 case ':':
2786 if (peek_char () == ':') {
2787 get_char ();
2788 return Token.DOUBLE_COLON;
2789 }
2790 return Token.COLON;
2791
2792 case '0': case '1': case '2': case '3': case '4':
2793 case '5': case '6': case '7': case '8': case '9':
2794 tokens_seen = true;
2795 return is_number (c);
2796
2797 case '\n': // white space
2798 any_token_seen |= tokens_seen;
2799 tokens_seen = false;
2800 comments_seen = false;
2801 continue;
2802
2803 case '.':
2804 tokens_seen = true;
2805 d = peek_char ();
2806 if (d >= '0' && d <= '9')
2807 return is_number (c);
2808 return Token.DOT;
2809
2810 case '#':
2811 if (tokens_seen || comments_seen) {
2812 Eror_WrongPreprocessorLocation ();
2813 return Token.ERROR;
2814 }
2815
2816 if (handle_preprocessing_directive (true))
2817 continue;
2818
2819 bool directive_expected = false;
2820 while ((c = get_char ()) != -1) {
2821 if (col == 1) {
2822 directive_expected = true;
2823 } else if (!directive_expected) {
2824 // TODO: Implement comment support for disabled code and uncomment this code
2825 // if (c == '#') {
2826 // Eror_WrongPreprocessorLocation ();
2827 // return Token.ERROR;
2828 // }
2829 continue;
2830 }
2831
2832 if (c == ' ' || c == '\t' || c == '\r' || c == '\n' || c == '\f' || c == '\v' )
2833 continue;
2834
2835 if (c == '#') {
2836 if (handle_preprocessing_directive (false))
2837 break;
2838 }
2839 directive_expected = false;
2840 }
2841
2842 if (c != -1) {
2843 tokens_seen = false;
2844 continue;
2845 }
2846
2847 return Token.EOF;
2848
2849 case '"':
2850 return consume_string (false);
2851
2852 case '\'':
2853 return TokenizeBackslash ();
2854
2855 case '@':
2856 c = get_char ();
2857 if (c == '"') {
2858 tokens_seen = true;
2859 return consume_string (true);
2860 }
2861
2862 if (is_identifier_start_character (c)){
2863 return consume_identifier (c, true);
2864 }
2865
2866 Report.Error (1646, Location, "Keyword, identifier, or string expected after verbatim specifier: @");
2867 return Token.ERROR;
2868
2869 case EvalStatementParserCharacter:
2870 return Token.EVAL_STATEMENT_PARSER;
2871 case EvalCompilationUnitParserCharacter:
2872 return Token.EVAL_COMPILATION_UNIT_PARSER;
2873 case EvalUsingDeclarationsParserCharacter:
2874 return Token.EVAL_USING_DECLARATIONS_UNIT_PARSER;
2875 }
2876
2877 if (is_identifier_start_character (c)) {
2878 tokens_seen = true;
2879 return consume_identifier (c);
2880 }
2881
2882 error_details = ((char)c).ToString ();
2883 return Token.ERROR;
2884 }
2885
2886 if (CompleteOnEOF){
2887 if (generated)
2888 return Token.COMPLETE_COMPLETION;
2889
2890 generated = true;
2891 return Token.GENERATE_COMPLETION;
2892 }
2893
2894
2895 return Token.EOF;
2896 }
2897
2898 int TokenizeBackslash ()
2899 {
2900 int c = get_char ();
2901 tokens_seen = true;
2902 if (c == '\'') {
2903 error_details = "Empty character literal";
2904 Report.Error (1011, Location, error_details);
2905 return Token.ERROR;
2906 }
2907 if (c == '\r' || c == '\n') {
2908 Report.Error (1010, Location, "Newline in constant");
2909 return Token.ERROR;
2910 }
2911
2912 int d;
2913 c = escape (c, out d);
2914 if (c == -1)
2915 return Token.ERROR;
2916 if (d != 0)
2917 throw new NotImplementedException ();
2918
2919 val = (char) c;
2920 c = get_char ();
2921
2922 if (c != '\'') {
2923 Report.Error (1012, Location, "Too many characters in character literal");
2924
2925 // Try to recover, read until newline or next "'"
2926 while ((c = get_char ()) != -1) {
2927 if (c == '\n' || c == '\'')
2928 break;
2929 }
2930 return Token.ERROR;
2931 }
2932
2933 return Token.LITERAL_CHARACTER;
2934 }
2935
2936 int TokenizeLessThan ()
2937 {
2938 int d;
2939 if (handle_typeof) {
2940 PushPosition ();
2941 if (parse_generic_dimension (out d)) {
2942 val = d;
2943 DiscardPosition ();
2944 return Token.GENERIC_DIMENSION;
2945 }
2946 PopPosition ();
2947 }
2948
2949 // Save current position and parse next token.
2950 PushPosition ();
2951 if (parse_less_than ()) {
2952 if (parsing_generic_declaration && token () != Token.DOT) {
2953 d = Token.OP_GENERICS_LT_DECL;
2954 } else {
2955 d = Token.OP_GENERICS_LT;
2956 }
2957 PopPosition ();
2958 return d;
2959 }
2960
2961 PopPosition ();
2962 parsing_generic_less_than = 0;
2963
2964 d = peek_char ();
2965 if (d == '<') {
2966 get_char ();
2967 d = peek_char ();
2968
2969 if (d == '=') {
2970 get_char ();
2971 return Token.OP_SHIFT_LEFT_ASSIGN;
2972 }
2973 return Token.OP_SHIFT_LEFT;
2974 }
2975
2976 if (d == '=') {
2977 get_char ();
2978 return Token.OP_LE;
2979 }
2980 return Token.OP_LT;
2981 }
2982
2983 //
2984 // Handles one line xml comment
2985 //
2986 private void handle_one_line_xml_comment ()
2987 {
2988 int c;
2989 while ((c = peek_char ()) == ' ')
2990 get_char (); // skip heading whitespaces.
2991 while ((c = peek_char ()) != -1 && c != '\n' && c != '\r') {
2992 xml_comment_buffer.Append ((char) get_char ());
2993 }
2994 if (c == '\r' || c == '\n')
2995 xml_comment_buffer.Append (Environment.NewLine);
2996 }
2997
2998 //
2999 // Remove heading "*" in Javadoc-like xml documentation.
3000 //
3001 private void update_formatted_doc_comment (int current_comment_start)
3002 {
3003 int length = xml_comment_buffer.Length - current_comment_start;
3004 string [] lines = xml_comment_buffer.ToString (
3005 current_comment_start,
3006 length).Replace ("\r", "").Split ('\n');
3007
3008 // The first line starts with /**, thus it is not target
3009 // for the format check.
3010 for (int i = 1; i < lines.Length; i++) {
3011 string s = lines [i];
3012 int idx = s.IndexOf ('*');
3013 string head = null;
3014 if (idx < 0) {
3015 if (i < lines.Length - 1)
3016 return;
3017 head = s;
3018 } else
3019 head = s.Substring (0, idx);
3020 foreach (char c in head)
3021 if (c != ' ')
3022 return;
3023 lines [i] = s.Substring (idx + 1);
3024 }
3025 xml_comment_buffer.Remove (current_comment_start, length);
3026 xml_comment_buffer.Insert (current_comment_start, String.Join (Environment.NewLine, lines));
3027 }
3028
3029 //
3030 // Updates current comment location.
3031 //
3032 private void update_comment_location ()
3033 {
3034 if (current_comment_location.IsNull) {
3035 // "-2" is for heading "//" or "/*"
3036 current_comment_location =
3037 new Location (ref_line, hidden ? -1 : col - 2);
3038 }
3039 }
3040
3041 //
3042 // Checks if there was incorrect doc comments and raise
3043 // warnings.
3044 //
3045 public void check_incorrect_doc_comment ()
3046 {
3047 if (xml_comment_buffer.Length > 0)
3048 warn_incorrect_doc_comment ();
3049 }
3050
3051 //
3052 // Raises a warning when tokenizer found incorrect doccomment
3053 // markup.
3054 //
3055 private void warn_incorrect_doc_comment ()
3056 {
3057 if (doc_state != XmlCommentState.Error) {
3058 doc_state = XmlCommentState.Error;
3059 // in csc, it is 'XML comment is not placed on
3060 // a valid language element'. But that does not
3061 // make sense.
3062 Report.Warning (1587, 2, Location, "XML comment is not placed on a valid language element");
3063 }
3064 }
3065
3066 //
3067 // Consumes the saved xml comment lines (if any)
3068 // as for current target member or type.
3069 //
3070 public string consume_doc_comment ()
3071 {
3072 if (xml_comment_buffer.Length > 0) {
3073 string ret = xml_comment_buffer.ToString ();
3074 reset_doc_comment ();
3075 return ret;
3076 }
3077 return null;
3078 }
3079
3080 void reset_doc_comment ()
3081 {
3082 xml_comment_buffer.Length = 0;
3083 current_comment_location = Location.Null;
3084 }
3085
3086 public void cleanup ()
3087 {
3088 if (ifstack != null && ifstack.Count >= 1) {
3089 int state = (int) ifstack.Pop ();
3090 if ((state & REGION) != 0)
3091 Report.Error (1038, Location, "#endregion directive expected");
3092 else
3093 Report.Error (1027, Location, "Expected `#endif' directive");
3094 }
3095 }
3096 }
3097
3098 //
3099 // Indicates whether it accepts XML documentation or not.
3100 //
3101 public enum XmlCommentState {
3102 // comment is allowed in this state.
3103 Allowed,
3104 // comment is not allowed in this state.
3105 NotAllowed,
3106 // once comments appeared when it is NotAllowed, then the
3107 // state is changed to it, until the state is changed to
3108 // .Allowed.
3109 Error
3110 }
3111 }
3112
git-svn mirror of mcs