?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Fix make dist
[mcs.git] / mcs / cs-tokenizer.cs
blob023649905cec5923c55db59ec6430a5527a89974
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.Generic;
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 sealed class IdentifiersComparer : IEqualityComparer<char[]>
44 {
45 readonly int length;
46
47 public IdentifiersComparer (int length)
48 {
49 this.length = length;
50 }
51
52 public bool Equals (char[] x, char[] y)
53 {
54 for (int i = 0; i < length; ++i)
55 if (x [i] != y [i])
56 return false;
57
58 return true;
59 }
60
61 public int GetHashCode (char[] obj)
62 {
63 int h = 0;
64 for (int i = 0; i < length; ++i)
65 h = (h << 5) - h + obj [i];
66
67 return h;
68 }
69 }
70
71 //
72 // This class has to be used in the parser only, it reuses token
73 // details after each parse
74 //
75 public class LocatedToken
76 {
77 int row, column;
78 string value;
79
80 static LocatedToken[] buffer;
81 static int pos;
82
83 private LocatedToken ()
84 {
85 }
86
87 public static LocatedToken Create (int row, int column)
88 {
89 return Create (null, row, column);
90 }
91
92 public static LocatedToken Create (string value, int row, int column)
93 {
94 //
95 // TODO: I am not very happy about the logic but it's the best
96 // what I could come up with for now.
97 // Ideally we should be using just tiny buffer (256 elements) which
98 // is enough to hold all details for currect stack and recycle elements
99 // poped from the stack but there is a trick needed to recycle
100 // them properly.
101 //
102 LocatedToken entry;
103 if (pos >= buffer.Length) {
104 entry = new LocatedToken ();
105 } else {
106 entry = buffer [pos];
107 if (entry == null) {
108 entry = new LocatedToken ();
109 buffer [pos] = entry;
110 }
111
112 ++pos;
113 }
114 entry.value = value;
115 entry.row = row;
116 entry.column = column;
117 return entry;
118 }
119
120 //
121 // Used for token not required by expression evaluator
122 //
123 public static LocatedToken CreateOptional (int row, int col)
124 {
125 #if false
126 return Create (row, col);
127 #endif
128 return null;
129 }
130
131 public static void Initialize ()
132 {
133 if (buffer == null)
134 buffer = new LocatedToken [10000];
135 pos = 0;
136 }
137
138 public Location Location {
139 get { return new Location (row, column); }
140 }
141
142 public string Value {
143 get { return value; }
144 }
145 }
146
147 SeekableStreamReader reader;
148 SourceFile ref_name;
149 CompilationUnit file_name;
150 CompilerContext context;
151 bool hidden = false;
152 int ref_line = 1;
153 int line = 1;
154 int col = 0;
155 int previous_col;
156 int current_token;
157 bool handle_get_set = false;
158 bool handle_remove_add = false;
159 bool handle_where = false;
160 bool handle_typeof = false;
161 bool lambda_arguments_parsing;
162 Location current_comment_location = Location.Null;
163 List<Location> escaped_identifiers;
164 int parsing_generic_less_than;
165
166 //
167 // Used mainly for parser optimizations. Some expressions for instance
168 // can appear only in block (including initializer, base initializer)
169 // scope only
170 //
171 public int parsing_block;
172 internal bool query_parsing;
173
174 //
175 // When parsing type only, useful for ambiguous nullable types
176 //
177 public int parsing_type;
178
179 //
180 // Set when parsing generic declaration (type or method header)
181 //
182 public bool parsing_generic_declaration;
183
184 //
185 // The value indicates that we have not reach any declaration or
186 // namespace yet
187 //
188 public int parsing_declaration;
189
190 //
191 // The special character to inject on streams to trigger the EXPRESSION_PARSE
192 // token to be returned. It just happens to be a Unicode character that
193 // would never be part of a program (can not be an identifier).
194 //
195 // This character is only tested just before the tokenizer is about to report
196 // an error; So on the regular operation mode, this addition will have no
197 // impact on the tokenizer's performance.
198 //
199
200 public const int EvalStatementParserCharacter = 0x2190; // Unicode Left Arrow
201 public const int EvalCompilationUnitParserCharacter = 0x2191; // Unicode Arrow
202 public const int EvalUsingDeclarationsParserCharacter = 0x2192; // Unicode Arrow
203
204 //
205 // XML documentation buffer. The save point is used to divide
206 // comments on types and comments on members.
207 //
208 StringBuilder xml_comment_buffer;
209
210 //
211 // See comment on XmlCommentState enumeration.
212 //
213 XmlCommentState xml_doc_state = XmlCommentState.Allowed;
214
215 //
216 // Whether tokens have been seen on this line
217 //
218 bool tokens_seen = false;
219
220 //
221 // Set to true once the GENERATE_COMPLETION token has bee
222 // returned. This helps produce one GENERATE_COMPLETION,
223 // as many COMPLETE_COMPLETION as necessary to complete the
224 // AST tree and one final EOF.
225 //
226 bool generated;
227
228 //
229 // Whether a token has been seen on the file
230 // This is needed because `define' is not allowed to be used
231 // after a token has been seen.
232 //
233 bool any_token_seen = false;
234
235 static readonly char[] simple_whitespaces = new char[] { ' ', '\t' };
236
237 public bool PropertyParsing {
238 get { return handle_get_set; }
239 set { handle_get_set = value; }
240 }
241
242 public bool EventParsing {
243 get { return handle_remove_add; }
244 set { handle_remove_add = value; }
245 }
246
247 public bool ConstraintsParsing {
248 get { return handle_where; }
249 set { handle_where = value; }
250 }
251
252 public bool TypeOfParsing {
253 get { return handle_typeof; }
254 set { handle_typeof = value; }
255 }
256
257 public XmlCommentState doc_state {
258 get { return xml_doc_state; }
259 set {
260 if (value == XmlCommentState.Allowed) {
261 check_incorrect_doc_comment ();
262 reset_doc_comment ();
263 }
264 xml_doc_state = value;
265 }
266 }
267
268 //
269 // This is used to trigger completion generation on the parser
270 public bool CompleteOnEOF;
271
272 void AddEscapedIdentifier (Location loc)
273 {
274 if (escaped_identifiers == null)
275 escaped_identifiers = new List<Location> ();
276
277 escaped_identifiers.Add (loc);
278 }
279
280 public bool IsEscapedIdentifier (Location loc)
281 {
282 if (escaped_identifiers != null) {
283 foreach (Location lt in escaped_identifiers)
284 if (lt.Equals (loc))
285 return true;
286 }
287
288 return false;
289 }
290
291 //
292 // Class variables
293 //
294 static KeywordEntry[][] keywords;
295 static Dictionary<string, object> keyword_strings; // TODO: HashSet
296 static NumberStyles styles;
297 static NumberFormatInfo csharp_format_info;
298
299 //
300 // Values for the associated token returned
301 //
302 internal int putback_char; // Used by repl only
303 object val;
304
305 //
306 // Pre-processor
307 //
308 const int TAKING = 1;
309 const int ELSE_SEEN = 4;
310 const int PARENT_TAKING = 8;
311 const int REGION = 16;
312
313 //
314 // pre-processor if stack state:
315 //
316 Stack<int> ifstack;
317
318 static System.Text.StringBuilder string_builder;
319
320 const int max_id_size = 512;
321 static char [] id_builder = new char [max_id_size];
322
323 public static Dictionary<char[], string>[] identifiers = new Dictionary<char[], string>[max_id_size + 1];
324
325 const int max_number_size = 512;
326 static char [] number_builder = new char [max_number_size];
327 static int number_pos;
328
329 static StringBuilder static_cmd_arg = new System.Text.StringBuilder ();
330
331 //
332 // Details about the error encoutered by the tokenizer
333 //
334 string error_details;
335
336 public string error {
337 get {
338 return error_details;
339 }
340 }
341
342 public int Line {
343 get {
344 return ref_line;
345 }
346 }
347
348 //
349 // This is used when the tokenizer needs to save
350 // the current position as it needs to do some parsing
351 // on its own to deamiguate a token in behalf of the
352 // parser.
353 //
354 Stack<Position> position_stack = new Stack<Position> (2);
355
356 class Position {
357 public int position;
358 public int line;
359 public int ref_line;
360 public int col;
361 public bool hidden;
362 public int putback_char;
363 public int previous_col;
364 public Stack<int> ifstack;
365 public int parsing_generic_less_than;
366 public int current_token;
367 public object val;
368
369 public Position (Tokenizer t)
370 {
371 position = t.reader.Position;
372 line = t.line;
373 ref_line = t.ref_line;
374 col = t.col;
375 hidden = t.hidden;
376 putback_char = t.putback_char;
377 previous_col = t.previous_col;
378 if (t.ifstack != null && t.ifstack.Count != 0) {
379 // There is no simple way to clone Stack<T> all
380 // methods reverse the order
381 var clone = t.ifstack.ToArray ();
382 Array.Reverse (clone);
383 ifstack = new Stack<int> (clone);
384 }
385 parsing_generic_less_than = t.parsing_generic_less_than;
386 current_token = t.current_token;
387 val = t.val;
388 }
389 }
390
391 public void PushPosition ()
392 {
393 position_stack.Push (new Position (this));
394 }
395
396 public void PopPosition ()
397 {
398 Position p = position_stack.Pop ();
399
400 reader.Position = p.position;
401 ref_line = p.ref_line;
402 line = p.line;
403 col = p.col;
404 hidden = p.hidden;
405 putback_char = p.putback_char;
406 previous_col = p.previous_col;
407 ifstack = p.ifstack;
408 parsing_generic_less_than = p.parsing_generic_less_than;
409 current_token = p.current_token;
410 val = p.val;
411 }
412
413 // Do not reset the position, ignore it.
414 public void DiscardPosition ()
415 {
416 position_stack.Pop ();
417 }
418
419 static void AddKeyword (string kw, int token)
420 {
421 keyword_strings.Add (kw, null);
422
423 int length = kw.Length;
424 if (keywords [length] == null) {
425 keywords [length] = new KeywordEntry ['z' - '_' + 1];
426 }
427
428 int char_index = kw [0] - '_';
429 KeywordEntry kwe = keywords [length] [char_index];
430 if (kwe == null) {
431 keywords [length] [char_index] = new KeywordEntry (kw, token);
432 return;
433 }
434
435 while (kwe.Next != null) {
436 kwe = kwe.Next;
437 }
438
439 kwe.Next = new KeywordEntry (kw, token);
440 }
441
442 static void InitTokens ()
443 {
444 keyword_strings = new Dictionary<string, object> ();
445
446 // 11 is the length of the longest keyword for now
447 keywords = new KeywordEntry [11] [];
448
449 AddKeyword ("__arglist", Token.ARGLIST);
450 AddKeyword ("abstract", Token.ABSTRACT);
451 AddKeyword ("as", Token.AS);
452 AddKeyword ("add", Token.ADD);
453 AddKeyword ("base", Token.BASE);
454 AddKeyword ("bool", Token.BOOL);
455 AddKeyword ("break", Token.BREAK);
456 AddKeyword ("byte", Token.BYTE);
457 AddKeyword ("case", Token.CASE);
458 AddKeyword ("catch", Token.CATCH);
459 AddKeyword ("char", Token.CHAR);
460 AddKeyword ("checked", Token.CHECKED);
461 AddKeyword ("class", Token.CLASS);
462 AddKeyword ("const", Token.CONST);
463 AddKeyword ("continue", Token.CONTINUE);
464 AddKeyword ("decimal", Token.DECIMAL);
465 AddKeyword ("default", Token.DEFAULT);
466 AddKeyword ("delegate", Token.DELEGATE);
467 AddKeyword ("do", Token.DO);
468 AddKeyword ("double", Token.DOUBLE);
469 AddKeyword ("else", Token.ELSE);
470 AddKeyword ("enum", Token.ENUM);
471 AddKeyword ("event", Token.EVENT);
472 AddKeyword ("explicit", Token.EXPLICIT);
473 AddKeyword ("extern", Token.EXTERN);
474 AddKeyword ("false", Token.FALSE);
475 AddKeyword ("finally", Token.FINALLY);
476 AddKeyword ("fixed", Token.FIXED);
477 AddKeyword ("float", Token.FLOAT);
478 AddKeyword ("for", Token.FOR);
479 AddKeyword ("foreach", Token.FOREACH);
480 AddKeyword ("goto", Token.GOTO);
481 AddKeyword ("get", Token.GET);
482 AddKeyword ("if", Token.IF);
483 AddKeyword ("implicit", Token.IMPLICIT);
484 AddKeyword ("in", Token.IN);
485 AddKeyword ("int", Token.INT);
486 AddKeyword ("interface", Token.INTERFACE);
487 AddKeyword ("internal", Token.INTERNAL);
488 AddKeyword ("is", Token.IS);
489 AddKeyword ("lock", Token.LOCK);
490 AddKeyword ("long", Token.LONG);
491 AddKeyword ("namespace", Token.NAMESPACE);
492 AddKeyword ("new", Token.NEW);
493 AddKeyword ("null", Token.NULL);
494 AddKeyword ("object", Token.OBJECT);
495 AddKeyword ("operator", Token.OPERATOR);
496 AddKeyword ("out", Token.OUT);
497 AddKeyword ("override", Token.OVERRIDE);
498 AddKeyword ("params", Token.PARAMS);
499 AddKeyword ("private", Token.PRIVATE);
500 AddKeyword ("protected", Token.PROTECTED);
501 AddKeyword ("public", Token.PUBLIC);
502 AddKeyword ("readonly", Token.READONLY);
503 AddKeyword ("ref", Token.REF);
504 AddKeyword ("remove", Token.REMOVE);
505 AddKeyword ("return", Token.RETURN);
506 AddKeyword ("sbyte", Token.SBYTE);
507 AddKeyword ("sealed", Token.SEALED);
508 AddKeyword ("set", Token.SET);
509 AddKeyword ("short", Token.SHORT);
510 AddKeyword ("sizeof", Token.SIZEOF);
511 AddKeyword ("stackalloc", Token.STACKALLOC);
512 AddKeyword ("static", Token.STATIC);
513 AddKeyword ("string", Token.STRING);
514 AddKeyword ("struct", Token.STRUCT);
515 AddKeyword ("switch", Token.SWITCH);
516 AddKeyword ("this", Token.THIS);
517 AddKeyword ("throw", Token.THROW);
518 AddKeyword ("true", Token.TRUE);
519 AddKeyword ("try", Token.TRY);
520 AddKeyword ("typeof", Token.TYPEOF);
521 AddKeyword ("uint", Token.UINT);
522 AddKeyword ("ulong", Token.ULONG);
523 AddKeyword ("unchecked", Token.UNCHECKED);
524 AddKeyword ("unsafe", Token.UNSAFE);
525 AddKeyword ("ushort", Token.USHORT);
526 AddKeyword ("using", Token.USING);
527 AddKeyword ("virtual", Token.VIRTUAL);
528 AddKeyword ("void", Token.VOID);
529 AddKeyword ("volatile", Token.VOLATILE);
530 AddKeyword ("while", Token.WHILE);
531 AddKeyword ("partial", Token.PARTIAL);
532 AddKeyword ("where", Token.WHERE);
533
534 // LINQ keywords
535 AddKeyword ("from", Token.FROM);
536 AddKeyword ("join", Token.JOIN);
537 AddKeyword ("on", Token.ON);
538 AddKeyword ("equals", Token.EQUALS);
539 AddKeyword ("select", Token.SELECT);
540 AddKeyword ("group", Token.GROUP);
541 AddKeyword ("by", Token.BY);
542 AddKeyword ("let", Token.LET);
543 AddKeyword ("orderby", Token.ORDERBY);
544 AddKeyword ("ascending", Token.ASCENDING);
545 AddKeyword ("descending", Token.DESCENDING);
546 AddKeyword ("into", Token.INTO);
547 }
548
549 //
550 // Class initializer
551 //
552 static Tokenizer ()
553 {
554 InitTokens ();
555 csharp_format_info = NumberFormatInfo.InvariantInfo;
556 styles = NumberStyles.Float;
557
558 string_builder = new System.Text.StringBuilder ();
559 }
560
561 int GetKeyword (char[] id, int id_len)
562 {
563 //
564 // Keywords are stored in an array of arrays grouped by their
565 // length and then by the first character
566 //
567 if (id_len >= keywords.Length || keywords [id_len] == null)
568 return -1;
569
570 int first_index = id [0] - '_';
571 if (first_index > 'z')
572 return -1;
573
574 KeywordEntry kwe = keywords [id_len] [first_index];
575 if (kwe == null)
576 return -1;
577
578 int res;
579 do {
580 res = kwe.Token;
581 for (int i = 1; i < id_len; ++i) {
582 if (id [i] != kwe.Value [i]) {
583 res = 0;
584 kwe = kwe.Next;
585 break;
586 }
587 }
588 } while (res == 0 && kwe != null);
589
590 if (res == 0)
591 return -1;
592
593 int next_token;
594 switch (res) {
595 case Token.GET:
596 case Token.SET:
597 if (!handle_get_set)
598 res = -1;
599 break;
600 case Token.REMOVE:
601 case Token.ADD:
602 if (!handle_remove_add)
603 res = -1;
604 break;
605 case Token.EXTERN:
606 if (parsing_declaration == 0)
607 res = Token.EXTERN_ALIAS;
608 break;
609 case Token.DEFAULT:
610 if (peek_token () == Token.COLON) {
611 token ();
612 res = Token.DEFAULT_COLON;
613 }
614 break;
615 case Token.WHERE:
616 if (!handle_where && !query_parsing)
617 res = -1;
618 break;
619 case Token.FROM:
620 //
621 // A query expression is any expression that starts with `from identifier'
622 // followed by any token except ; , =
623 //
624 if (!query_parsing) {
625 if (lambda_arguments_parsing) {
626 res = -1;
627 break;
628 }
629
630 PushPosition ();
631 // HACK: to disable generics micro-parser, because PushPosition does not
632 // store identifiers array
633 parsing_generic_less_than = 1;
634 switch (xtoken ()) {
635 case Token.IDENTIFIER:
636 case Token.INT:
637 case Token.BOOL:
638 case Token.BYTE:
639 case Token.CHAR:
640 case Token.DECIMAL:
641 case Token.FLOAT:
642 case Token.LONG:
643 case Token.OBJECT:
644 case Token.STRING:
645 case Token.UINT:
646 case Token.ULONG:
647 next_token = xtoken ();
648 if (next_token == Token.SEMICOLON || next_token == Token.COMMA || next_token == Token.EQUALS)
649 goto default;
650
651 res = Token.FROM_FIRST;
652 query_parsing = true;
653 if (RootContext.Version <= LanguageVersion.ISO_2)
654 Report.FeatureIsNotAvailable (Location, "query expressions");
655 break;
656 case Token.VOID:
657 Expression.Error_VoidInvalidInTheContext (Location, Report);
658 break;
659 default:
660 PopPosition ();
661 // HACK: A token is not a keyword so we need to restore identifiers buffer
662 // which has been overwritten before we grabbed the identifier
663 id_builder [0] = 'f'; id_builder [1] = 'r'; id_builder [2] = 'o'; id_builder [3] = 'm';
664 return -1;
665 }
666 PopPosition ();
667 }
668 break;
669 case Token.JOIN:
670 case Token.ON:
671 case Token.EQUALS:
672 case Token.SELECT:
673 case Token.GROUP:
674 case Token.BY:
675 case Token.LET:
676 case Token.ORDERBY:
677 case Token.ASCENDING:
678 case Token.DESCENDING:
679 case Token.INTO:
680 if (!query_parsing)
681 res = -1;
682 break;
683
684 case Token.USING:
685 case Token.NAMESPACE:
686 // TODO: some explanation needed
687 check_incorrect_doc_comment ();
688 break;
689
690 case Token.PARTIAL:
691 if (parsing_block > 0) {
692 res = -1;
693 break;
694 }
695
696 // Save current position and parse next token.
697 PushPosition ();
698
699 next_token = token ();
700 bool ok = (next_token == Token.CLASS) ||
701 (next_token == Token.STRUCT) ||
702 (next_token == Token.INTERFACE) ||
703 (next_token == Token.VOID);
704
705 PopPosition ();
706
707 if (ok) {
708 if (next_token == Token.VOID) {
709 if (RootContext.Version == LanguageVersion.ISO_1 ||
710 RootContext.Version == LanguageVersion.ISO_2)
711 Report.FeatureIsNotAvailable (Location, "partial methods");
712 } else if (RootContext.Version == LanguageVersion.ISO_1)
713 Report.FeatureIsNotAvailable (Location, "partial types");
714
715 return res;
716 }
717
718 if (next_token < Token.LAST_KEYWORD) {
719 Report.Error (267, Location,
720 "The `partial' modifier can be used only immediately before `class', `struct', `interface', or `void' keyword");
721 return token ();
722 }
723
724 res = -1;
725 break;
726 }
727
728 return res;
729 }
730
731 public Location Location {
732 get {
733 return new Location (ref_line, hidden ? -1 : col);
734 }
735 }
736
737 public Tokenizer (SeekableStreamReader input, CompilationUnit file, CompilerContext ctx)
738 {
739 this.ref_name = file;
740 this.file_name = file;
741 this.context = ctx;
742 reader = input;
743
744 putback_char = -1;
745
746 xml_comment_buffer = new StringBuilder ();
747
748 //
749 // FIXME: This could be `Location.Push' but we have to
750 // find out why the MS compiler allows this
751 //
752 Mono.CSharp.Location.Push (file, file);
753 }
754
755 static bool is_identifier_start_character (int c)
756 {
757 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || Char.IsLetter ((char)c);
758 }
759
760 static bool is_identifier_part_character (char c)
761 {
762 if (c >= 'a' && c <= 'z')
763 return true;
764
765 if (c >= 'A' && c <= 'Z')
766 return true;
767
768 if (c == '_' || (c >= '0' && c <= '9'))
769 return true;
770
771 if (c < 0x80)
772 return false;
773
774 return Char.IsLetter (c) || Char.GetUnicodeCategory (c) == UnicodeCategory.ConnectorPunctuation;
775 }
776
777 public static bool IsKeyword (string s)
778 {
779 return keyword_strings.ContainsKey (s);
780 }
781
782 //
783 // Open parens micro parser. Detects both lambda and cast ambiguity.
784 //
785 int TokenizeOpenParens ()
786 {
787 int ptoken;
788 current_token = -1;
789
790 int bracket_level = 0;
791 bool is_type = false;
792 bool can_be_type = false;
793
794 while (true) {
795 ptoken = current_token;
796 token ();
797
798 switch (current_token) {
799 case Token.CLOSE_PARENS:
800 token ();
801
802 //
803 // Expression inside parens is lambda, (int i) =>
804 //
805 if (current_token == Token.ARROW)
806 return Token.OPEN_PARENS_LAMBDA;
807
808 //
809 // Expression inside parens is single type, (int[])
810 //
811 if (is_type)
812 return Token.OPEN_PARENS_CAST;
813
814 //
815 // Expression is possible cast, look at next token, (T)null
816 //
817 if (can_be_type) {
818 switch (current_token) {
819 case Token.OPEN_PARENS:
820 case Token.BANG:
821 case Token.TILDE:
822 case Token.IDENTIFIER:
823 case Token.LITERAL:
824 case Token.BASE:
825 case Token.CHECKED:
826 case Token.DELEGATE:
827 case Token.FALSE:
828 case Token.FIXED:
829 case Token.NEW:
830 case Token.NULL:
831 case Token.SIZEOF:
832 case Token.THIS:
833 case Token.THROW:
834 case Token.TRUE:
835 case Token.TYPEOF:
836 case Token.UNCHECKED:
837 case Token.UNSAFE:
838 case Token.DEFAULT:
839
840 //
841 // These can be part of a member access
842 //
843 case Token.INT:
844 case Token.UINT:
845 case Token.SHORT:
846 case Token.USHORT:
847 case Token.LONG:
848 case Token.ULONG:
849 case Token.DOUBLE:
850 case Token.FLOAT:
851 case Token.CHAR:
852 case Token.BYTE:
853 case Token.DECIMAL:
854 case Token.BOOL:
855 return Token.OPEN_PARENS_CAST;
856 }
857 }
858 return Token.OPEN_PARENS;
859
860 case Token.DOT:
861 case Token.DOUBLE_COLON:
862 if (ptoken != Token.IDENTIFIER && ptoken != Token.OP_GENERICS_GT)
863 goto default;
864
865 continue;
866
867 case Token.IDENTIFIER:
868 switch (ptoken) {
869 case Token.DOT:
870 case Token.OP_GENERICS_LT:
871 case Token.COMMA:
872 case Token.DOUBLE_COLON:
873 case -1:
874 if (bracket_level == 0)
875 can_be_type = true;
876 continue;
877 default:
878 can_be_type = is_type = false;
879 continue;
880 }
881
882 case Token.OBJECT:
883 case Token.STRING:
884 case Token.BOOL:
885 case Token.DECIMAL:
886 case Token.FLOAT:
887 case Token.DOUBLE:
888 case Token.SBYTE:
889 case Token.BYTE:
890 case Token.SHORT:
891 case Token.USHORT:
892 case Token.INT:
893 case Token.UINT:
894 case Token.LONG:
895 case Token.ULONG:
896 case Token.CHAR:
897 case Token.VOID:
898 if (bracket_level == 0)
899 is_type = true;
900 continue;
901
902 case Token.COMMA:
903 if (bracket_level == 0) {
904 bracket_level = 100;
905 can_be_type = is_type = false;
906 }
907 continue;
908
909 case Token.OP_GENERICS_LT:
910 case Token.OPEN_BRACKET:
911 if (bracket_level++ == 0)
912 is_type = true;
913 continue;
914
915 case Token.OP_GENERICS_GT:
916 case Token.CLOSE_BRACKET:
917 --bracket_level;
918 continue;
919
920 case Token.INTERR_NULLABLE:
921 case Token.STAR:
922 if (bracket_level == 0)
923 is_type = true;
924 continue;
925
926 case Token.REF:
927 case Token.OUT:
928 can_be_type = is_type = false;
929 continue;
930
931 default:
932 return Token.OPEN_PARENS;
933 }
934 }
935 }
936
937 public static bool IsValidIdentifier (string s)
938 {
939 if (s == null || s.Length == 0)
940 return false;
941
942 if (!is_identifier_start_character (s [0]))
943 return false;
944
945 for (int i = 1; i < s.Length; i ++)
946 if (! is_identifier_part_character (s [i]))
947 return false;
948
949 return true;
950 }
951
952 bool parse_less_than ()
953 {
954 start:
955 int the_token = token ();
956 if (the_token == Token.OPEN_BRACKET) {
957 do {
958 the_token = token ();
959 } while (the_token != Token.CLOSE_BRACKET);
960 the_token = token ();
961 } else if (the_token == Token.IN || the_token == Token.OUT) {
962 the_token = token ();
963 }
964 switch (the_token) {
965 case Token.IDENTIFIER:
966 case Token.OBJECT:
967 case Token.STRING:
968 case Token.BOOL:
969 case Token.DECIMAL:
970 case Token.FLOAT:
971 case Token.DOUBLE:
972 case Token.SBYTE:
973 case Token.BYTE:
974 case Token.SHORT:
975 case Token.USHORT:
976 case Token.INT:
977 case Token.UINT:
978 case Token.LONG:
979 case Token.ULONG:
980 case Token.CHAR:
981 case Token.VOID:
982 break;
983 case Token.OP_GENERICS_GT:
984 return true;
985
986 default:
987 return false;
988 }
989 again:
990 the_token = token ();
991
992 if (the_token == Token.OP_GENERICS_GT)
993 return true;
994 else if (the_token == Token.COMMA || the_token == Token.DOT || the_token == Token.DOUBLE_COLON)
995 goto start;
996 else if (the_token == Token.INTERR_NULLABLE || the_token == Token.STAR)
997 goto again;
998 else if (the_token == Token.OP_GENERICS_LT) {
999 if (!parse_less_than ())
1000 return false;
1001 goto again;
1002 } else if (the_token == Token.OPEN_BRACKET) {
1003 rank_specifiers:
1004 the_token = token ();
1005 if (the_token == Token.CLOSE_BRACKET)
1006 goto again;
1007 else if (the_token == Token.COMMA)
1008 goto rank_specifiers;
1009 return false;
1010 }
1011
1012 return false;
1013 }
1014
1015 bool parse_generic_dimension (out int dimension)
1016 {
1017 dimension = 1;
1018
1019 again:
1020 int the_token = token ();
1021 if (the_token == Token.OP_GENERICS_GT)
1022 return true;
1023 else if (the_token == Token.COMMA) {
1024 dimension++;
1025 goto again;
1026 }
1027
1028 return false;
1029 }
1030
1031 public int peek_token ()
1032 {
1033 int the_token;
1034
1035 PushPosition ();
1036 the_token = token ();
1037 PopPosition ();
1038
1039 return the_token;
1040 }
1041
1042 //
1043 // Tonizes `?' using custom disambiguous rules to return one
1044 // of following tokens: INTERR_NULLABLE, OP_COALESCING, INTERR
1045 //
1046 // Tricky expression look like:
1047 //
1048 // Foo ? a = x ? b : c;
1049 //
1050 int TokenizePossibleNullableType ()
1051 {
1052 if (parsing_block == 0 || parsing_type > 0)
1053 return Token.INTERR_NULLABLE;
1054
1055 int d = peek_char ();
1056 if (d == '?') {
1057 get_char ();
1058 return Token.OP_COALESCING;
1059 }
1060
1061 switch (current_token) {
1062 case Token.CLOSE_PARENS:
1063 case Token.TRUE:
1064 case Token.FALSE:
1065 case Token.NULL:
1066 case Token.LITERAL:
1067 return Token.INTERR;
1068 }
1069
1070 if (d != ' ') {
1071 if (d == ',' || d == ';' || d == '>')
1072 return Token.INTERR_NULLABLE;
1073 if (d == '*' || (d >= '0' && d <= '9'))
1074 return Token.INTERR;
1075 }
1076
1077 PushPosition ();
1078 current_token = Token.NONE;
1079 int next_token;
1080 switch (xtoken ()) {
1081 case Token.LITERAL:
1082 case Token.TRUE:
1083 case Token.FALSE:
1084 case Token.NULL:
1085 case Token.THIS:
1086 case Token.NEW:
1087 next_token = Token.INTERR;
1088 break;
1089
1090 case Token.SEMICOLON:
1091 case Token.COMMA:
1092 case Token.CLOSE_PARENS:
1093 case Token.OPEN_BRACKET:
1094 case Token.OP_GENERICS_GT:
1095 next_token = Token.INTERR_NULLABLE;
1096 break;
1097
1098 default:
1099 next_token = -1;
1100 break;
1101 }
1102
1103 if (next_token == -1) {
1104 switch (xtoken ()) {
1105 case Token.COMMA:
1106 case Token.SEMICOLON:
1107 case Token.OPEN_BRACE:
1108 case Token.CLOSE_PARENS:
1109 case Token.IN:
1110 next_token = Token.INTERR_NULLABLE;
1111 break;
1112
1113 case Token.COLON:
1114 next_token = Token.INTERR;
1115 break;
1116
1117 default:
1118 int ntoken;
1119 int interrs = 1;
1120 int colons = 0;
1121 //
1122 // All shorcuts failed, do it hard way
1123 //
1124 while ((ntoken = xtoken ()) != Token.EOF) {
1125 if (ntoken == Token.SEMICOLON)
1126 break;
1127
1128 if (ntoken == Token.COLON) {
1129 if (++colons == interrs)
1130 break;
1131 continue;
1132 }
1133
1134 if (ntoken == Token.INTERR) {
1135 ++interrs;
1136 continue;
1137 }
1138 }
1139
1140 next_token = colons != interrs ? Token.INTERR_NULLABLE : Token.INTERR;
1141 break;
1142 }
1143 }
1144
1145 PopPosition ();
1146 return next_token;
1147 }
1148
1149 bool decimal_digits (int c)
1150 {
1151 int d;
1152 bool seen_digits = false;
1153
1154 if (c != -1){
1155 if (number_pos == max_number_size)
1156 Error_NumericConstantTooLong ();
1157 number_builder [number_pos++] = (char) c;
1158 }
1159
1160 //
1161 // We use peek_char2, because decimal_digits needs to do a
1162 // 2-character look-ahead (5.ToString for example).
1163 //
1164 while ((d = peek_char2 ()) != -1){
1165 if (d >= '0' && d <= '9'){
1166 if (number_pos == max_number_size)
1167 Error_NumericConstantTooLong ();
1168 number_builder [number_pos++] = (char) d;
1169 get_char ();
1170 seen_digits = true;
1171 } else
1172 break;
1173 }
1174
1175 return seen_digits;
1176 }
1177
1178 static bool is_hex (int e)
1179 {
1180 return (e >= '0' && e <= '9') || (e >= 'A' && e <= 'F') || (e >= 'a' && e <= 'f');
1181 }
1182
1183 static TypeCode real_type_suffix (int c)
1184 {
1185 switch (c){
1186 case 'F': case 'f':
1187 return TypeCode.Single;
1188 case 'D': case 'd':
1189 return TypeCode.Double;
1190 case 'M': case 'm':
1191 return TypeCode.Decimal;
1192 default:
1193 return TypeCode.Empty;
1194 }
1195 }
1196
1197 int integer_type_suffix (ulong ul, int c)
1198 {
1199 bool is_unsigned = false;
1200 bool is_long = false;
1201
1202 if (c != -1){
1203 bool scanning = true;
1204 do {
1205 switch (c){
1206 case 'U': case 'u':
1207 if (is_unsigned)
1208 scanning = false;
1209 is_unsigned = true;
1210 get_char ();
1211 break;
1212
1213 case 'l':
1214 if (!is_unsigned){
1215 //
1216 // if we have not seen anything in between
1217 // report this error
1218 //
1219 Report.Warning (78, 4, Location, "The 'l' suffix is easily confused with the digit '1' (use 'L' for clarity)");
1220 }
1221
1222 goto case 'L';
1223
1224 case 'L':
1225 if (is_long)
1226 scanning = false;
1227 is_long = true;
1228 get_char ();
1229 break;
1230
1231 default:
1232 scanning = false;
1233 break;
1234 }
1235 c = peek_char ();
1236 } while (scanning);
1237 }
1238
1239 if (is_long && is_unsigned){
1240 val = new ULongLiteral (ul, Location);
1241 return Token.LITERAL;
1242 }
1243
1244 if (is_unsigned){
1245 // uint if possible, or ulong else.
1246
1247 if ((ul & 0xffffffff00000000) == 0)
1248 val = new UIntLiteral ((uint) ul, Location);
1249 else
1250 val = new ULongLiteral (ul, Location);
1251 } else if (is_long){
1252 // long if possible, ulong otherwise
1253 if ((ul & 0x8000000000000000) != 0)
1254 val = new ULongLiteral (ul, Location);
1255 else
1256 val = new LongLiteral ((long) ul, Location);
1257 } else {
1258 // int, uint, long or ulong in that order
1259 if ((ul & 0xffffffff00000000) == 0){
1260 uint ui = (uint) ul;
1261
1262 if ((ui & 0x80000000) != 0)
1263 val = new UIntLiteral (ui, Location);
1264 else
1265 val = new IntLiteral ((int) ui, Location);
1266 } else {
1267 if ((ul & 0x8000000000000000) != 0)
1268 val = new ULongLiteral (ul, Location);
1269 else
1270 val = new LongLiteral ((long) ul, Location);
1271 }
1272 }
1273 return Token.LITERAL;
1274 }
1275
1276 //
1277 // given `c' as the next char in the input decide whether
1278 // we need to convert to a special type, and then choose
1279 // the best representation for the integer
1280 //
1281 int adjust_int (int c)
1282 {
1283 try {
1284 if (number_pos > 9){
1285 ulong ul = (uint) (number_builder [0] - '0');
1286
1287 for (int i = 1; i < number_pos; i++){
1288 ul = checked ((ul * 10) + ((uint)(number_builder [i] - '0')));
1289 }
1290 return integer_type_suffix (ul, c);
1291 } else {
1292 uint ui = (uint) (number_builder [0] - '0');
1293
1294 for (int i = 1; i < number_pos; i++){
1295 ui = checked ((ui * 10) + ((uint)(number_builder [i] - '0')));
1296 }
1297 return integer_type_suffix (ui, c);
1298 }
1299 } catch (OverflowException) {
1300 error_details = "Integral constant is too large";
1301 Report.Error (1021, Location, error_details);
1302 val = new IntLiteral (0, Location);
1303 return Token.LITERAL;
1304 }
1305 catch (FormatException) {
1306 Report.Error (1013, Location, "Invalid number");
1307 val = new IntLiteral (0, Location);
1308 return Token.LITERAL;
1309 }
1310 }
1311
1312 int adjust_real (TypeCode t)
1313 {
1314 string s = new String (number_builder, 0, number_pos);
1315 const string error_details = "Floating-point constant is outside the range of type `{0}'";
1316
1317 switch (t){
1318 case TypeCode.Decimal:
1319 try {
1320 val = new DecimalLiteral (decimal.Parse (s, styles, csharp_format_info), Location);
1321 } catch (OverflowException) {
1322 val = new DecimalLiteral (0, Location);
1323 Report.Error (594, Location, error_details, "decimal");
1324 }
1325 break;
1326 case TypeCode.Single:
1327 try {
1328 val = new FloatLiteral (float.Parse (s, styles, csharp_format_info), Location);
1329 } catch (OverflowException) {
1330 val = new FloatLiteral (0, Location);
1331 Report.Error (594, Location, error_details, "float");
1332 }
1333 break;
1334 default:
1335 try {
1336 val = new DoubleLiteral (double.Parse (s, styles, csharp_format_info), Location);
1337 } catch (OverflowException) {
1338 val = new DoubleLiteral (0, Location);
1339 Report.Error (594, Location, error_details, "double");
1340 }
1341 break;
1342 }
1343
1344 return Token.LITERAL;
1345 }
1346
1347 int handle_hex ()
1348 {
1349 int d;
1350 ulong ul;
1351
1352 get_char ();
1353 while ((d = peek_char ()) != -1){
1354 if (is_hex (d)){
1355 number_builder [number_pos++] = (char) d;
1356 get_char ();
1357 } else
1358 break;
1359 }
1360
1361 string s = new String (number_builder, 0, number_pos);
1362 try {
1363 if (number_pos <= 8)
1364 ul = System.UInt32.Parse (s, NumberStyles.HexNumber);
1365 else
1366 ul = System.UInt64.Parse (s, NumberStyles.HexNumber);
1367 } catch (OverflowException){
1368 error_details = "Integral constant is too large";
1369 Report.Error (1021, Location, error_details);
1370 val = new IntLiteral (0, Location);
1371 return Token.LITERAL;
1372 }
1373 catch (FormatException) {
1374 Report.Error (1013, Location, "Invalid number");
1375 val = new IntLiteral (0, Location);
1376 return Token.LITERAL;
1377 }
1378
1379 return integer_type_suffix (ul, peek_char ());
1380 }
1381
1382 //
1383 // Invoked if we know we have .digits or digits
1384 //
1385 int is_number (int c)
1386 {
1387 bool is_real = false;
1388
1389 number_pos = 0;
1390
1391 if (c >= '0' && c <= '9'){
1392 if (c == '0'){
1393 int peek = peek_char ();
1394
1395 if (peek == 'x' || peek == 'X')
1396 return handle_hex ();
1397 }
1398 decimal_digits (c);
1399 c = get_char ();
1400 }
1401
1402 //
1403 // We need to handle the case of
1404 // "1.1" vs "1.string" (LITERAL_FLOAT vs NUMBER DOT IDENTIFIER)
1405 //
1406 if (c == '.'){
1407 if (decimal_digits ('.')){
1408 is_real = true;
1409 c = get_char ();
1410 } else {
1411 putback ('.');
1412 number_pos--;
1413 return adjust_int (-1);
1414 }
1415 }
1416
1417 if (c == 'e' || c == 'E'){
1418 is_real = true;
1419 if (number_pos == max_number_size)
1420 Error_NumericConstantTooLong ();
1421 number_builder [number_pos++] = 'e';
1422 c = get_char ();
1423
1424 if (c == '+'){
1425 if (number_pos == max_number_size)
1426 Error_NumericConstantTooLong ();
1427 number_builder [number_pos++] = '+';
1428 c = -1;
1429 } else if (c == '-') {
1430 if (number_pos == max_number_size)
1431 Error_NumericConstantTooLong ();
1432 number_builder [number_pos++] = '-';
1433 c = -1;
1434 } else {
1435 if (number_pos == max_number_size)
1436 Error_NumericConstantTooLong ();
1437 number_builder [number_pos++] = '+';
1438 }
1439
1440 decimal_digits (c);
1441 c = get_char ();
1442 }
1443
1444 var type = real_type_suffix (c);
1445 if (type == TypeCode.Empty && !is_real){
1446 putback (c);
1447 return adjust_int (c);
1448 }
1449
1450 is_real = true;
1451
1452 if (type == TypeCode.Empty){
1453 putback (c);
1454 }
1455
1456 if (is_real)
1457 return adjust_real (type);
1458
1459 throw new Exception ("Is Number should never reach this point");
1460 }
1461
1462 //
1463 // Accepts exactly count (4 or 8) hex, no more no less
1464 //
1465 int getHex (int count, out int surrogate, out bool error)
1466 {
1467 int i;
1468 int total = 0;
1469 int c;
1470 int top = count != -1 ? count : 4;
1471
1472 get_char ();
1473 error = false;
1474 surrogate = 0;
1475 for (i = 0; i < top; i++){
1476 c = get_char ();
1477
1478 if (c >= '0' && c <= '9')
1479 c = (int) c - (int) '0';
1480 else if (c >= 'A' && c <= 'F')
1481 c = (int) c - (int) 'A' + 10;
1482 else if (c >= 'a' && c <= 'f')
1483 c = (int) c - (int) 'a' + 10;
1484 else {
1485 error = true;
1486 return 0;
1487 }
1488
1489 total = (total * 16) + c;
1490 if (count == -1){
1491 int p = peek_char ();
1492 if (p == -1)
1493 break;
1494 if (!is_hex ((char)p))
1495 break;
1496 }
1497 }
1498
1499 if (top == 8) {
1500 if (total > 0x0010FFFF) {
1501 error = true;
1502 return 0;
1503 }
1504
1505 if (total >= 0x00010000) {
1506 surrogate = ((total - 0x00010000) % 0x0400 + 0xDC00);
1507 total = ((total - 0x00010000) / 0x0400 + 0xD800);
1508 }
1509 }
1510
1511 return total;
1512 }
1513
1514 int escape (int c, out int surrogate)
1515 {
1516 bool error;
1517 int d;
1518 int v;
1519
1520 d = peek_char ();
1521 if (c != '\\') {
1522 surrogate = 0;
1523 return c;
1524 }
1525
1526 switch (d){
1527 case 'a':
1528 v = '\a'; break;
1529 case 'b':
1530 v = '\b'; break;
1531 case 'n':
1532 v = '\n'; break;
1533 case 't':
1534 v = '\t'; break;
1535 case 'v':
1536 v = '\v'; break;
1537 case 'r':
1538 v = '\r'; break;
1539 case '\\':
1540 v = '\\'; break;
1541 case 'f':
1542 v = '\f'; break;
1543 case '0':
1544 v = 0; break;
1545 case '"':
1546 v = '"'; break;
1547 case '\'':
1548 v = '\''; break;
1549 case 'x':
1550 v = getHex (-1, out surrogate, out error);
1551 if (error)
1552 goto default;
1553 return v;
1554 case 'u':
1555 case 'U':
1556 return EscapeUnicode (d, out surrogate);
1557 default:
1558 surrogate = 0;
1559 Report.Error (1009, Location, "Unrecognized escape sequence `\\{0}'", ((char)d).ToString ());
1560 return d;
1561 }
1562
1563 get_char ();
1564 surrogate = 0;
1565 return v;
1566 }
1567
1568 int EscapeUnicode (int ch, out int surrogate)
1569 {
1570 bool error;
1571 if (ch == 'U') {
1572 ch = getHex (8, out surrogate, out error);
1573 } else {
1574 ch = getHex (4, out surrogate, out error);
1575 }
1576
1577 if (error)
1578 Report.Error (1009, Location, "Unrecognized escape sequence");
1579
1580 return ch;
1581 }
1582
1583 int get_char ()
1584 {
1585 int x;
1586 if (putback_char != -1) {
1587 x = putback_char;
1588 putback_char = -1;
1589 } else
1590 x = reader.Read ();
1591 if (x == '\n') {
1592 advance_line ();
1593 } else {
1594 col++;
1595 }
1596 return x;
1597 }
1598
1599 void advance_line ()
1600 {
1601 line++;
1602 ref_line++;
1603 previous_col = col;
1604 col = 0;
1605 }
1606
1607 int peek_char ()
1608 {
1609 if (putback_char == -1)
1610 putback_char = reader.Read ();
1611 return putback_char;
1612 }
1613
1614 int peek_char2 ()
1615 {
1616 if (putback_char != -1)
1617 return putback_char;
1618 return reader.Peek ();
1619 }
1620
1621 void putback (int c)
1622 {
1623 if (putback_char != -1){
1624 Console.WriteLine ("Col: " + col);
1625 Console.WriteLine ("Row: " + line);
1626 Console.WriteLine ("Name: " + ref_name.Name);
1627 Console.WriteLine ("Current [{0}] putting back [{1}] ", putback_char, c);
1628 throw new Exception ("This should not happen putback on putback");
1629 }
1630 if (c == '\n' || col == 0) {
1631 // It won't happen though.
1632 line--;
1633 ref_line--;
1634 col = previous_col;
1635 }
1636 else
1637 col--;
1638 putback_char = c;
1639 }
1640
1641 public bool advance ()
1642 {
1643 return peek_char () != -1 || CompleteOnEOF;
1644 }
1645
1646 public Object Value {
1647 get {
1648 return val;
1649 }
1650 }
1651
1652 public Object value ()
1653 {
1654 return val;
1655 }
1656
1657 public int token ()
1658 {
1659 current_token = xtoken ();
1660 return current_token;
1661 }
1662
1663 void get_cmd_arg (out string cmd, out string arg)
1664 {
1665 int c;
1666
1667 tokens_seen = false;
1668 arg = "";
1669
1670 // skip over white space
1671 do {
1672 c = get_char ();
1673 } while (c == '\r' || c == ' ' || c == '\t');
1674
1675 static_cmd_arg.Length = 0;
1676 while (c != -1 && is_identifier_part_character ((char)c)) {
1677 static_cmd_arg.Append ((char)c);
1678 c = get_char ();
1679 if (c == '\\') {
1680 int peek = peek_char ();
1681 if (peek == 'U' || peek == 'u') {
1682 int surrogate;
1683 c = EscapeUnicode (c, out surrogate);
1684 if (surrogate != 0) {
1685 if (is_identifier_part_character ((char) c))
1686 static_cmd_arg.Append ((char) c);
1687 c = surrogate;
1688 }
1689 }
1690 }
1691 }
1692
1693 cmd = static_cmd_arg.ToString ();
1694
1695 // skip over white space
1696 while (c == '\r' || c == ' ' || c == '\t')
1697 c = get_char ();
1698
1699 static_cmd_arg.Length = 0;
1700 int has_identifier_argument = 0;
1701
1702 while (c != -1 && c != '\n' && c != '\r') {
1703 if (c == '\\' && has_identifier_argument >= 0) {
1704 if (has_identifier_argument != 0 || (cmd == "define" || cmd == "if" || cmd == "elif" || cmd == "undef")) {
1705 has_identifier_argument = 1;
1706
1707 int peek = peek_char ();
1708 if (peek == 'U' || peek == 'u') {
1709 int surrogate;
1710 c = EscapeUnicode (c, out surrogate);
1711 if (surrogate != 0) {
1712 if (is_identifier_part_character ((char) c))
1713 static_cmd_arg.Append ((char) c);
1714 c = surrogate;
1715 }
1716 }
1717 } else {
1718 has_identifier_argument = -1;
1719 }
1720 }
1721 static_cmd_arg.Append ((char) c);
1722 c = get_char ();
1723 }
1724
1725 if (static_cmd_arg.Length != 0)
1726 arg = static_cmd_arg.ToString ();
1727 }
1728
1729 //
1730 // Handles the #line directive
1731 //
1732 bool PreProcessLine (string arg)
1733 {
1734 if (arg.Length == 0)
1735 return false;
1736
1737 if (arg == "default"){
1738 ref_line = line;
1739 ref_name = file_name;
1740 hidden = false;
1741 Location.Push (file_name, ref_name);
1742 return true;
1743 } else if (arg == "hidden"){
1744 hidden = true;
1745 return true;
1746 }
1747
1748 try {
1749 int pos;
1750
1751 if ((pos = arg.IndexOf (' ')) != -1 && pos != 0){
1752 ref_line = System.Int32.Parse (arg.Substring (0, pos));
1753 pos++;
1754
1755 char [] quotes = { '\"' };
1756
1757 string name = arg.Substring (pos). Trim (quotes);
1758 ref_name = Location.LookupFile (file_name, name);
1759 file_name.AddFile (ref_name);
1760 hidden = false;
1761 Location.Push (file_name, ref_name);
1762 } else {
1763 ref_line = System.Int32.Parse (arg);
1764 hidden = false;
1765 }
1766 } catch {
1767 return false;
1768 }
1769
1770 return true;
1771 }
1772
1773 //
1774 // Handles #define and #undef
1775 //
1776 void PreProcessDefinition (bool is_define, string ident, bool caller_is_taking)
1777 {
1778 if (ident.Length == 0 || ident == "true" || ident == "false"){
1779 Report.Error (1001, Location, "Missing identifier to pre-processor directive");
1780 return;
1781 }
1782
1783 if (ident.IndexOfAny (simple_whitespaces) != -1){
1784 Error_EndLineExpected ();
1785 return;
1786 }
1787
1788 if (!is_identifier_start_character (ident [0]))
1789 Report.Error (1001, Location, "Identifier expected: {0}", ident);
1790
1791 foreach (char c in ident.Substring (1)){
1792 if (!is_identifier_part_character (c)){
1793 Report.Error (1001, Location, "Identifier expected: {0}", ident);
1794 return;
1795 }
1796 }
1797
1798 if (!caller_is_taking)
1799 return;
1800
1801 if (is_define) {
1802 //
1803 // #define ident
1804 //
1805 if (RootContext.IsConditionalDefined (ident))
1806 return;
1807
1808 file_name.AddDefine (ident);
1809 } else {
1810 //
1811 // #undef ident
1812 //
1813 file_name.AddUndefine (ident);
1814 }
1815 }
1816
1817 static byte read_hex (string arg, int pos, out bool error)
1818 {
1819 error = false;
1820
1821 int total;
1822 char c = arg [pos];
1823
1824 if ((c >= '0') && (c <= '9'))
1825 total = (int) c - (int) '0';
1826 else if ((c >= 'A') && (c <= 'F'))
1827 total = (int) c - (int) 'A' + 10;
1828 else if ((c >= 'a') && (c <= 'f'))
1829 total = (int) c - (int) 'a' + 10;
1830 else {
1831 error = true;
1832 return 0;
1833 }
1834
1835 total *= 16;
1836 c = arg [pos+1];
1837
1838 if ((c >= '0') && (c <= '9'))
1839 total += (int) c - (int) '0';
1840 else if ((c >= 'A') && (c <= 'F'))
1841 total += (int) c - (int) 'A' + 10;
1842 else if ((c >= 'a') && (c <= 'f'))
1843 total += (int) c - (int) 'a' + 10;
1844 else {
1845 error = true;
1846 return 0;
1847 }
1848
1849 return (byte) total;
1850 }
1851
1852 /// <summary>
1853 /// Handles #pragma checksum
1854 /// </summary>
1855 bool PreProcessPragmaChecksum (string arg)
1856 {
1857 if ((arg [0] != ' ') && (arg [0] != '\t'))
1858 return false;
1859
1860 arg = arg.Trim (simple_whitespaces);
1861 if ((arg.Length < 2) || (arg [0] != '"'))
1862 return false;
1863
1864 StringBuilder file_sb = new StringBuilder ();
1865
1866 int pos = 1;
1867 char ch;
1868 while ((ch = arg [pos++]) != '"') {
1869 if (pos >= arg.Length)
1870 return false;
1871
1872 if (ch == '\\') {
1873 if (pos+1 >= arg.Length)
1874 return false;
1875 ch = arg [pos++];
1876 }
1877
1878 file_sb.Append (ch);
1879 }
1880
1881 if ((pos+2 >= arg.Length) || ((arg [pos] != ' ') && (arg [pos] != '\t')))
1882 return false;
1883
1884 arg = arg.Substring (pos).Trim (simple_whitespaces);
1885 if ((arg.Length < 42) || (arg [0] != '"') || (arg [1] != '{') ||
1886 (arg [10] != '-') || (arg [15] != '-') || (arg [20] != '-') ||
1887 (arg [25] != '-') || (arg [38] != '}') || (arg [39] != '"'))
1888 return false;
1889
1890 bool error;
1891 byte[] guid_bytes = new byte [16];
1892
1893 for (int i = 0; i < 4; i++) {
1894 guid_bytes [i] = read_hex (arg, 2+2*i, out error);
1895 if (error)
1896 return false;
1897 }
1898 for (int i = 0; i < 2; i++) {
1899 guid_bytes [i+4] = read_hex (arg, 11+2*i, out error);
1900 if (error)
1901 return false;
1902 guid_bytes [i+6] = read_hex (arg, 16+2*i, out error);
1903 if (error)
1904 return false;
1905 guid_bytes [i+8] = read_hex (arg, 21+2*i, out error);
1906 if (error)
1907 return false;
1908 }
1909
1910 for (int i = 0; i < 6; i++) {
1911 guid_bytes [i+10] = read_hex (arg, 26+2*i, out error);
1912 if (error)
1913 return false;
1914 }
1915
1916 arg = arg.Substring (40).Trim (simple_whitespaces);
1917 if ((arg.Length < 34) || (arg [0] != '"') || (arg [33] != '"'))
1918 return false;
1919
1920 byte[] checksum_bytes = new byte [16];
1921 for (int i = 0; i < 16; i++) {
1922 checksum_bytes [i] = read_hex (arg, 1+2*i, out error);
1923 if (error)
1924 return false;
1925 }
1926
1927 arg = arg.Substring (34).Trim (simple_whitespaces);
1928 if (arg.Length > 0)
1929 return false;
1930
1931 SourceFile file = Location.LookupFile (file_name, file_sb.ToString ());
1932 file.SetChecksum (guid_bytes, checksum_bytes);
1933 ref_name.AutoGenerated = true;
1934 return true;
1935 }
1936
1937 /// <summary>
1938 /// Handles #pragma directive
1939 /// </summary>
1940 void PreProcessPragma (string arg)
1941 {
1942 const string warning = "warning";
1943 const string w_disable = "warning disable";
1944 const string w_restore = "warning restore";
1945 const string checksum = "checksum";
1946
1947 if (arg == w_disable) {
1948 Report.RegisterWarningRegion (Location).WarningDisable (Location.Row);
1949 return;
1950 }
1951
1952 if (arg == w_restore) {
1953 Report.RegisterWarningRegion (Location).WarningEnable (Location.Row);
1954 return;
1955 }
1956
1957 if (arg.StartsWith (w_disable)) {
1958 int[] codes = ParseNumbers (arg.Substring (w_disable.Length));
1959 foreach (int code in codes) {
1960 if (code != 0)
1961 Report.RegisterWarningRegion (Location).WarningDisable (Location, code, Report);
1962 }
1963 return;
1964 }
1965
1966 if (arg.StartsWith (w_restore)) {
1967 int[] codes = ParseNumbers (arg.Substring (w_restore.Length));
1968 var w_table = Report.warning_ignore_table;
1969 foreach (int code in codes) {
1970 if (w_table != null && w_table.ContainsKey (code))
1971 Report.Warning (1635, 1, Location, "Cannot restore warning `CS{0:0000}' because it was disabled globally", code);
1972 Report.RegisterWarningRegion (Location).WarningEnable (Location, code, Report);
1973 }
1974 return;
1975 }
1976
1977 if (arg.StartsWith (warning)) {
1978 Report.Warning (1634, 1, Location, "Expected disable or restore");
1979 return;
1980 }
1981
1982 if (arg.StartsWith (checksum)) {
1983 if (!PreProcessPragmaChecksum (arg.Substring (checksum.Length)))
1984 Warning_InvalidPragmaChecksum ();
1985 return;
1986 }
1987
1988 Report.Warning (1633, 1, Location, "Unrecognized #pragma directive");
1989 }
1990
1991 int[] ParseNumbers (string text)
1992 {
1993 string[] string_array = text.Split (',');
1994 int[] values = new int [string_array.Length];
1995 int index = 0;
1996 foreach (string string_code in string_array) {
1997 try {
1998 values[index++] = int.Parse (string_code, System.Globalization.CultureInfo.InvariantCulture);
1999 }
2000 catch (FormatException) {
2001 Report.Warning (1692, 1, Location, "Invalid number");
2002 }
2003 }
2004 return values;
2005 }
2006
2007 bool eval_val (string s)
2008 {
2009 if (s == "true")
2010 return true;
2011 if (s == "false")
2012 return false;
2013
2014 return file_name.IsConditionalDefined (s);
2015 }
2016
2017 bool pp_primary (ref string s)
2018 {
2019 s = s.Trim ();
2020 int len = s.Length;
2021
2022 if (len > 0){
2023 char c = s [0];
2024
2025 if (c == '('){
2026 s = s.Substring (1);
2027 bool val = pp_expr (ref s, false);
2028 if (s.Length > 0 && s [0] == ')'){
2029 s = s.Substring (1);
2030 return val;
2031 }
2032 Error_InvalidDirective ();
2033 return false;
2034 }
2035
2036 if (is_identifier_start_character (c)){
2037 int j = 1;
2038
2039 while (j < len){
2040 c = s [j];
2041
2042 if (is_identifier_part_character (c)){
2043 j++;
2044 continue;
2045 }
2046 bool v = eval_val (s.Substring (0, j));
2047 s = s.Substring (j);
2048 return v;
2049 }
2050 bool vv = eval_val (s);
2051 s = "";
2052 return vv;
2053 }
2054 }
2055 Error_InvalidDirective ();
2056 return false;
2057 }
2058
2059 bool pp_unary (ref string s)
2060 {
2061 s = s.Trim ();
2062 int len = s.Length;
2063
2064 if (len > 0){
2065 if (s [0] == '!'){
2066 if (len > 1 && s [1] == '='){
2067 Error_InvalidDirective ();
2068 return false;
2069 }
2070 s = s.Substring (1);
2071 return ! pp_primary (ref s);
2072 } else
2073 return pp_primary (ref s);
2074 } else {
2075 Error_InvalidDirective ();
2076 return false;
2077 }
2078 }
2079
2080 bool pp_eq (ref string s)
2081 {
2082 bool va = pp_unary (ref s);
2083
2084 s = s.Trim ();
2085 int len = s.Length;
2086 if (len > 0){
2087 if (s [0] == '='){
2088 if (len > 2 && s [1] == '='){
2089 s = s.Substring (2);
2090 return va == pp_unary (ref s);
2091 } else {
2092 Error_InvalidDirective ();
2093 return false;
2094 }
2095 } else if (s [0] == '!' && len > 1 && s [1] == '='){
2096 s = s.Substring (2);
2097
2098 return va != pp_unary (ref s);
2099
2100 }
2101 }
2102
2103 return va;
2104
2105 }
2106
2107 bool pp_and (ref string s)
2108 {
2109 bool va = pp_eq (ref s);
2110
2111 s = s.Trim ();
2112 int len = s.Length;
2113 if (len > 0){
2114 if (s [0] == '&'){
2115 if (len > 2 && s [1] == '&'){
2116 s = s.Substring (2);
2117 return (va & pp_and (ref s));
2118 } else {
2119 Error_InvalidDirective ();
2120 return false;
2121 }
2122 }
2123 }
2124 return va;
2125 }
2126
2127 //
2128 // Evaluates an expression for `#if' or `#elif'
2129 //
2130 bool pp_expr (ref string s, bool isTerm)
2131 {
2132 bool va = pp_and (ref s);
2133 s = s.Trim ();
2134 int len = s.Length;
2135 if (len > 0){
2136 char c = s [0];
2137
2138 if (c == '|'){
2139 if (len > 2 && s [1] == '|'){
2140 s = s.Substring (2);
2141 return va | pp_expr (ref s, isTerm);
2142 } else {
2143 Error_InvalidDirective ();
2144 return false;
2145 }
2146 }
2147 if (isTerm) {
2148 Error_EndLineExpected ();
2149 return false;
2150 }
2151 }
2152
2153 return va;
2154 }
2155
2156 bool eval (string s)
2157 {
2158 bool v = pp_expr (ref s, true);
2159 s = s.Trim ();
2160 if (s.Length != 0){
2161 return false;
2162 }
2163
2164 return v;
2165 }
2166
2167 void Error_NumericConstantTooLong ()
2168 {
2169 Report.Error (1021, Location, "Numeric constant too long");
2170 }
2171
2172 void Error_InvalidDirective ()
2173 {
2174 Report.Error (1517, Location, "Invalid preprocessor directive");
2175 }
2176
2177 void Error_UnexpectedDirective (string extra)
2178 {
2179 Report.Error (
2180 1028, Location,
2181 "Unexpected processor directive ({0})", extra);
2182 }
2183
2184 void Error_TokensSeen ()
2185 {
2186 Report.Error (1032, Location,
2187 "Cannot define or undefine preprocessor symbols after first token in file");
2188 }
2189
2190 void Eror_WrongPreprocessorLocation ()
2191 {
2192 Report.Error (1040, Location,
2193 "Preprocessor directives must appear as the first non-whitespace character on a line");
2194 }
2195
2196 void Error_EndLineExpected ()
2197 {
2198 Report.Error (1025, Location, "Single-line comment or end-of-line expected");
2199 }
2200
2201 void Warning_InvalidPragmaChecksum ()
2202 {
2203 Report.Warning (1695, 1, Location,
2204 "Invalid #pragma checksum syntax; should be " +
2205 "#pragma checksum \"filename\" " +
2206 "\"{XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX}\" \"XXXX...\"");
2207 }
2208 //
2209 // if true, then the code continues processing the code
2210 // if false, the code stays in a loop until another directive is
2211 // reached.
2212 // When caller_is_taking is false we ignore all directives except the ones
2213 // which can help us to identify where the #if block ends
2214 bool handle_preprocessing_directive (bool caller_is_taking)
2215 {
2216 string cmd, arg;
2217 bool region_directive = false;
2218
2219 get_cmd_arg (out cmd, out arg);
2220
2221 // Eat any trailing whitespaces and single-line comments
2222 if (arg.IndexOf ("//") != -1)
2223 arg = arg.Substring (0, arg.IndexOf ("//"));
2224 arg = arg.Trim (simple_whitespaces);
2225
2226 //
2227 // The first group of pre-processing instructions is always processed
2228 //
2229 switch (cmd){
2230 case "region":
2231 region_directive = true;
2232 arg = "true";
2233 goto case "if";
2234
2235 case "endregion":
2236 if (ifstack == null || ifstack.Count == 0){
2237 Error_UnexpectedDirective ("no #region for this #endregion");
2238 return true;
2239 }
2240 int pop = ifstack.Pop ();
2241
2242 if ((pop & REGION) == 0)
2243 Report.Error (1027, Location, "Expected `#endif' directive");
2244
2245 return caller_is_taking;
2246
2247 case "if":
2248 if (ifstack == null)
2249 ifstack = new Stack<int> (2);
2250
2251 int flags = region_directive ? REGION : 0;
2252 if (ifstack.Count == 0){
2253 flags |= PARENT_TAKING;
2254 } else {
2255 int state = ifstack.Peek ();
2256 if ((state & TAKING) != 0) {
2257 flags |= PARENT_TAKING;
2258 }
2259 }
2260
2261 if (caller_is_taking && eval (arg)) {
2262 ifstack.Push (flags | TAKING);
2263 return true;
2264 }
2265 ifstack.Push (flags);
2266 return false;
2267
2268 case "endif":
2269 if (ifstack == null || ifstack.Count == 0){
2270 Error_UnexpectedDirective ("no #if for this #endif");
2271 return true;
2272 } else {
2273 pop = ifstack.Pop ();
2274
2275 if ((pop & REGION) != 0)
2276 Report.Error (1038, Location, "#endregion directive expected");
2277
2278 if (arg.Length != 0) {
2279 Error_EndLineExpected ();
2280 }
2281
2282 if (ifstack.Count == 0)
2283 return true;
2284
2285 int state = ifstack.Peek ();
2286 return (state & TAKING) != 0;
2287 }
2288
2289 case "elif":
2290 if (ifstack == null || ifstack.Count == 0){
2291 Error_UnexpectedDirective ("no #if for this #elif");
2292 return true;
2293 } else {
2294 int state = ifstack.Pop ();
2295
2296 if ((state & REGION) != 0) {
2297 Report.Error (1038, Location, "#endregion directive expected");
2298 return true;
2299 }
2300
2301 if ((state & ELSE_SEEN) != 0){
2302 Error_UnexpectedDirective ("#elif not valid after #else");
2303 return true;
2304 }
2305
2306 if ((state & TAKING) != 0) {
2307 ifstack.Push (0);
2308 return false;
2309 }
2310
2311 if (eval (arg) && ((state & PARENT_TAKING) != 0)){
2312 ifstack.Push (state | TAKING);
2313 return true;
2314 }
2315
2316 ifstack.Push (state);
2317 return false;
2318 }
2319
2320 case "else":
2321 if (ifstack == null || ifstack.Count == 0){
2322 Error_UnexpectedDirective ("no #if for this #else");
2323 return true;
2324 } else {
2325 int state = ifstack.Peek ();
2326
2327 if ((state & REGION) != 0) {
2328 Report.Error (1038, Location, "#endregion directive expected");
2329 return true;
2330 }
2331
2332 if ((state & ELSE_SEEN) != 0){
2333 Error_UnexpectedDirective ("#else within #else");
2334 return true;
2335 }
2336
2337 ifstack.Pop ();
2338
2339 if (arg.Length != 0) {
2340 Error_EndLineExpected ();
2341 return true;
2342 }
2343
2344 bool ret = false;
2345 if ((state & PARENT_TAKING) != 0) {
2346 ret = (state & TAKING) == 0;
2347
2348 if (ret)
2349 state |= TAKING;
2350 else
2351 state &= ~TAKING;
2352 }
2353
2354 ifstack.Push (state | ELSE_SEEN);
2355
2356 return ret;
2357 }
2358 case "define":
2359 if (any_token_seen){
2360 Error_TokensSeen ();
2361 return caller_is_taking;
2362 }
2363 PreProcessDefinition (true, arg, caller_is_taking);
2364 return caller_is_taking;
2365
2366 case "undef":
2367 if (any_token_seen){
2368 Error_TokensSeen ();
2369 return caller_is_taking;
2370 }
2371 PreProcessDefinition (false, arg, caller_is_taking);
2372 return caller_is_taking;
2373 }
2374
2375 //
2376 // These are only processed if we are in a `taking' block
2377 //
2378 if (!caller_is_taking)
2379 return false;
2380
2381 switch (cmd){
2382 case "error":
2383 Report.Error (1029, Location, "#error: '{0}'", arg);
2384 return true;
2385
2386 case "warning":
2387 Report.Warning (1030, 1, Location, "#warning: `{0}'", arg);
2388 return true;
2389
2390 case "pragma":
2391 if (RootContext.Version == LanguageVersion.ISO_1) {
2392 Report.FeatureIsNotAvailable (Location, "#pragma");
2393 return true;
2394 }
2395
2396 PreProcessPragma (arg);
2397 return true;
2398
2399 case "line":
2400 if (!PreProcessLine (arg))
2401 Report.Error (
2402 1576, Location,
2403 "The line number specified for #line directive is missing or invalid");
2404 return caller_is_taking;
2405 }
2406
2407 Report.Error (1024, Location, "Wrong preprocessor directive");
2408 return true;
2409
2410 }
2411
2412 private int consume_string (bool quoted)
2413 {
2414 int c;
2415 string_builder.Length = 0;
2416
2417 while ((c = get_char ()) != -1){
2418 if (c == '"'){
2419 if (quoted && peek_char () == '"'){
2420 string_builder.Append ((char) c);
2421 get_char ();
2422 continue;
2423 } else {
2424 val = new StringLiteral (string_builder.ToString (), Location);
2425 return Token.LITERAL;
2426 }
2427 }
2428
2429 if (c == '\n'){
2430 if (!quoted)
2431 Report.Error (1010, Location, "Newline in constant");
2432 }
2433
2434 if (!quoted){
2435 int surrogate;
2436 c = escape (c, out surrogate);
2437 if (c == -1)
2438 return Token.ERROR;
2439 if (surrogate != 0) {
2440 string_builder.Append ((char) c);
2441 c = surrogate;
2442 }
2443 }
2444 string_builder.Append ((char) c);
2445 }
2446
2447 Report.Error (1039, Location, "Unterminated string literal");
2448 return Token.EOF;
2449 }
2450
2451 private int consume_identifier (int s)
2452 {
2453 int res = consume_identifier (s, false);
2454
2455 if (doc_state == XmlCommentState.Allowed)
2456 doc_state = XmlCommentState.NotAllowed;
2457
2458 return res;
2459 }
2460
2461 int consume_identifier (int c, bool quoted)
2462 {
2463 //
2464 // This method is very performance sensitive. It accounts
2465 // for approximately 25% of all parser time
2466 //
2467
2468 int pos = 0;
2469 int column = col;
2470
2471 if (c == '\\') {
2472 int surrogate;
2473 c = escape (c, out surrogate);
2474 if (surrogate != 0) {
2475 id_builder [pos++] = (char) c;
2476 c = surrogate;
2477 }
2478 }
2479
2480 id_builder [pos++] = (char) c;
2481
2482 try {
2483 while (true) {
2484 c = reader.Read ();
2485
2486 if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || (c >= '0' && c <= '9')) {
2487 id_builder [pos++] = (char) c;
2488 continue;
2489 }
2490
2491 if (c < 0x80) {
2492 if (c == '\\') {
2493 int surrogate;
2494 c = escape (c, out surrogate);
2495 if (surrogate != 0) {
2496 if (is_identifier_part_character ((char) c))
2497 id_builder[pos++] = (char) c;
2498 c = surrogate;
2499 }
2500
2501 continue;
2502 }
2503 } else if (Char.IsLetter ((char) c) || Char.GetUnicodeCategory ((char) c) == UnicodeCategory.ConnectorPunctuation) {
2504 id_builder [pos++] = (char) c;
2505 continue;
2506 }
2507
2508 putback_char = c;
2509 break;
2510 }
2511 } catch (IndexOutOfRangeException) {
2512 Report.Error (645, Location, "Identifier too long (limit is 512 chars)");
2513 col += pos - 1;
2514 return Token.ERROR;
2515 }
2516
2517 col += pos - 1;
2518
2519 //
2520 // Optimization: avoids doing the keyword lookup
2521 // on uppercase letters
2522 //
2523 if (id_builder [0] >= '_' && !quoted) {
2524 int keyword = GetKeyword (id_builder, pos);
2525 if (keyword != -1) {
2526 val = LocatedToken.Create (null, ref_line, column);
2527 return keyword;
2528 }
2529 }
2530
2531 //
2532 // Keep identifiers in an array of hashtables to avoid needless
2533 // allocations
2534 //
2535 var identifiers_group = identifiers [pos];
2536 string s;
2537 if (identifiers_group != null) {
2538 if (identifiers_group.TryGetValue (id_builder, out s)) {
2539 val = LocatedToken.Create (s, ref_line, column);
2540 if (quoted)
2541 AddEscapedIdentifier (((LocatedToken) val).Location);
2542 return Token.IDENTIFIER;
2543 }
2544 } else {
2545 // TODO: this should be number of files dependant
2546 // corlib compilation peaks at 1000 and System.Core at 150
2547 int capacity = pos > 20 ? 10 : 100;
2548 identifiers_group = new Dictionary<char[],string> (capacity, new IdentifiersComparer (pos));
2549 identifiers [pos] = identifiers_group;
2550 }
2551
2552 char [] chars = new char [pos];
2553 Array.Copy (id_builder, chars, pos);
2554
2555 s = new string (id_builder, 0, pos);
2556 identifiers_group.Add (chars, s);
2557
2558 val = LocatedToken.Create (s, ref_line, column);
2559 if (quoted)
2560 AddEscapedIdentifier (((LocatedToken) val).Location);
2561
2562 return Token.IDENTIFIER;
2563 }
2564
2565 public int xtoken ()
2566 {
2567 int d, c;
2568
2569 // Whether we have seen comments on the current line
2570 bool comments_seen = false;
2571 while ((c = get_char ()) != -1) {
2572 switch (c) {
2573 case '\t':
2574 col = ((col + 8) / 8) * 8;
2575 continue;
2576
2577 case ' ':
2578 case '\f':
2579 case '\v':
2580 case 0xa0:
2581 case 0:
2582 case 0xFEFF: // Ignore BOM anywhere in the file
2583 continue;
2584
2585 /* This is required for compatibility with .NET
2586 case 0xEF:
2587 if (peek_char () == 0xBB) {
2588 PushPosition ();
2589 get_char ();
2590 if (get_char () == 0xBF)
2591 continue;
2592 PopPosition ();
2593 }
2594 break;
2595 */
2596 case '\r':
2597 if (peek_char () != '\n')
2598 advance_line ();
2599 else
2600 get_char ();
2601
2602 any_token_seen |= tokens_seen;
2603 tokens_seen = false;
2604 comments_seen = false;
2605 continue;
2606
2607 case '\\':
2608 tokens_seen = true;
2609 return consume_identifier (c);
2610
2611 case '{':
2612 val = LocatedToken.Create (ref_line, col);
2613 return Token.OPEN_BRACE;
2614 case '}':
2615 val = LocatedToken.Create (ref_line, col);
2616 return Token.CLOSE_BRACE;
2617 case '[':
2618 // To block doccomment inside attribute declaration.
2619 if (doc_state == XmlCommentState.Allowed)
2620 doc_state = XmlCommentState.NotAllowed;
2621 val = LocatedToken.CreateOptional (ref_line, col);
2622 return Token.OPEN_BRACKET;
2623 case ']':
2624 val = LocatedToken.CreateOptional (ref_line, col);
2625 return Token.CLOSE_BRACKET;
2626 case '(':
2627 val = LocatedToken.Create (ref_line, col);
2628 //
2629 // An expression versions of parens can appear in block context only
2630 //
2631 if (parsing_block != 0 && !lambda_arguments_parsing) {
2632
2633 //
2634 // Optmize most common case where we know that parens
2635 // is not special
2636 //
2637 switch (current_token) {
2638 case Token.IDENTIFIER:
2639 case Token.IF:
2640 case Token.FOR:
2641 case Token.FOREACH:
2642 case Token.TYPEOF:
2643 case Token.WHILE:
2644 case Token.USING:
2645 case Token.DEFAULT:
2646 case Token.DELEGATE:
2647 case Token.OP_GENERICS_GT:
2648 return Token.OPEN_PARENS;
2649 }
2650
2651 // Optimize using peek
2652 int xx = peek_char ();
2653 switch (xx) {
2654 case '(':
2655 case '\'':
2656 case '"':
2657 case '0':
2658 case '1':
2659 return Token.OPEN_PARENS;
2660 }
2661
2662 lambda_arguments_parsing = true;
2663 PushPosition ();
2664 d = TokenizeOpenParens ();
2665 PopPosition ();
2666 lambda_arguments_parsing = false;
2667 return d;
2668 }
2669
2670 return Token.OPEN_PARENS;
2671 case ')':
2672 val = LocatedToken.CreateOptional (ref_line, col);
2673 return Token.CLOSE_PARENS;
2674 case ',':
2675 val = LocatedToken.CreateOptional (ref_line, col);
2676 return Token.COMMA;
2677 case ';':
2678 val = LocatedToken.CreateOptional (ref_line, col);
2679 return Token.SEMICOLON;
2680 case '~':
2681 val = LocatedToken.Create (ref_line, col);
2682 return Token.TILDE;
2683 case '?':
2684 val = LocatedToken.Create (ref_line, col);
2685 return TokenizePossibleNullableType ();
2686 case '<':
2687 val = LocatedToken.Create (ref_line, col);
2688 if (parsing_generic_less_than++ > 0)
2689 return Token.OP_GENERICS_LT;
2690
2691 return TokenizeLessThan ();
2692
2693 case '>':
2694 val = LocatedToken.Create (ref_line, col);
2695 d = peek_char ();
2696
2697 if (d == '='){
2698 get_char ();
2699 return Token.OP_GE;
2700 }
2701
2702 if (parsing_generic_less_than > 1 || (parsing_generic_less_than == 1 && d != '>')) {
2703 parsing_generic_less_than--;
2704 return Token.OP_GENERICS_GT;
2705 }
2706
2707 if (d == '>') {
2708 get_char ();
2709 d = peek_char ();
2710
2711 if (d == '=') {
2712 get_char ();
2713 return Token.OP_SHIFT_RIGHT_ASSIGN;
2714 }
2715 return Token.OP_SHIFT_RIGHT;
2716 }
2717
2718 return Token.OP_GT;
2719
2720 case '+':
2721 val = LocatedToken.Create (ref_line, col);
2722 d = peek_char ();
2723 if (d == '+') {
2724 d = Token.OP_INC;
2725 } else if (d == '=') {
2726 d = Token.OP_ADD_ASSIGN;
2727 } else {
2728 return Token.PLUS;
2729 }
2730 get_char ();
2731 return d;
2732
2733 case '-':
2734 val = LocatedToken.Create (ref_line, col);
2735 d = peek_char ();
2736 if (d == '-') {
2737 d = Token.OP_DEC;
2738 } else if (d == '=')
2739 d = Token.OP_SUB_ASSIGN;
2740 else if (d == '>')
2741 d = Token.OP_PTR;
2742 else {
2743 return Token.MINUS;
2744 }
2745 get_char ();
2746 return d;
2747
2748 case '!':
2749 val = LocatedToken.Create (ref_line, col);
2750 if (peek_char () == '='){
2751 get_char ();
2752 return Token.OP_NE;
2753 }
2754 return Token.BANG;
2755
2756 case '=':
2757 val = LocatedToken.Create (ref_line, col);
2758 d = peek_char ();
2759 if (d == '='){
2760 get_char ();
2761 return Token.OP_EQ;
2762 }
2763 if (d == '>'){
2764 get_char ();
2765 return Token.ARROW;
2766 }
2767
2768 return Token.ASSIGN;
2769
2770 case '&':
2771 val = LocatedToken.Create (ref_line, col);
2772 d = peek_char ();
2773 if (d == '&'){
2774 get_char ();
2775 return Token.OP_AND;
2776 }
2777 if (d == '='){
2778 get_char ();
2779 return Token.OP_AND_ASSIGN;
2780 }
2781 return Token.BITWISE_AND;
2782
2783 case '|':
2784 val = LocatedToken.Create (ref_line, col);
2785 d = peek_char ();
2786 if (d == '|'){
2787 get_char ();
2788 return Token.OP_OR;
2789 }
2790 if (d == '='){
2791 get_char ();
2792 return Token.OP_OR_ASSIGN;
2793 }
2794 return Token.BITWISE_OR;
2795
2796 case '*':
2797 val = LocatedToken.Create (ref_line, col);
2798 if (peek_char () == '='){
2799 get_char ();
2800 return Token.OP_MULT_ASSIGN;
2801 }
2802 return Token.STAR;
2803
2804 case '/':
2805 d = peek_char ();
2806 if (d == '='){
2807 val = LocatedToken.Create (ref_line, col);
2808 get_char ();
2809 return Token.OP_DIV_ASSIGN;
2810 }
2811
2812 // Handle double-slash comments.
2813 if (d == '/'){
2814 get_char ();
2815 if (RootContext.Documentation != null && peek_char () == '/') {
2816 get_char ();
2817 // Don't allow ////.
2818 if ((d = peek_char ()) != '/') {
2819 update_comment_location ();
2820 if (doc_state == XmlCommentState.Allowed)
2821 handle_one_line_xml_comment ();
2822 else if (doc_state == XmlCommentState.NotAllowed)
2823 warn_incorrect_doc_comment ();
2824 }
2825 }
2826 while ((d = get_char ()) != -1 && (d != '\n') && d != '\r');
2827
2828 any_token_seen |= tokens_seen;
2829 tokens_seen = false;
2830 comments_seen = false;
2831 continue;
2832 } else if (d == '*'){
2833 get_char ();
2834 bool docAppend = false;
2835 if (RootContext.Documentation != null && peek_char () == '*') {
2836 get_char ();
2837 update_comment_location ();
2838 // But when it is /**/, just do nothing.
2839 if (peek_char () == '/') {
2840 get_char ();
2841 continue;
2842 }
2843 if (doc_state == XmlCommentState.Allowed)
2844 docAppend = true;
2845 else if (doc_state == XmlCommentState.NotAllowed)
2846 warn_incorrect_doc_comment ();
2847 }
2848
2849 int current_comment_start = 0;
2850 if (docAppend) {
2851 current_comment_start = xml_comment_buffer.Length;
2852 xml_comment_buffer.Append (Environment.NewLine);
2853 }
2854
2855 while ((d = get_char ()) != -1){
2856 if (d == '*' && peek_char () == '/'){
2857 get_char ();
2858 comments_seen = true;
2859 break;
2860 }
2861 if (docAppend)
2862 xml_comment_buffer.Append ((char) d);
2863
2864 if (d == '\n'){
2865 any_token_seen |= tokens_seen;
2866 tokens_seen = false;
2867 //
2868 // Reset 'comments_seen' just to be consistent.
2869 // It doesn't matter either way, here.
2870 //
2871 comments_seen = false;
2872 }
2873 }
2874 if (!comments_seen)
2875 Report.Error (1035, Location, "End-of-file found, '*/' expected");
2876
2877 if (docAppend)
2878 update_formatted_doc_comment (current_comment_start);
2879 continue;
2880 }
2881 return Token.DIV;
2882
2883 case '%':
2884 val = LocatedToken.Create (ref_line, col);
2885 if (peek_char () == '='){
2886 get_char ();
2887 return Token.OP_MOD_ASSIGN;
2888 }
2889 return Token.PERCENT;
2890
2891 case '^':
2892 val = LocatedToken.Create (ref_line, col);
2893 if (peek_char () == '='){
2894 get_char ();
2895 return Token.OP_XOR_ASSIGN;
2896 }
2897 return Token.CARRET;
2898
2899 case ':':
2900 val = LocatedToken.Create (ref_line, col);
2901 if (peek_char () == ':') {
2902 get_char ();
2903 return Token.DOUBLE_COLON;
2904 }
2905 return Token.COLON;
2906
2907 case '0': case '1': case '2': case '3': case '4':
2908 case '5': case '6': case '7': case '8': case '9':
2909 tokens_seen = true;
2910 return is_number (c);
2911
2912 case '\n': // white space
2913 any_token_seen |= tokens_seen;
2914 tokens_seen = false;
2915 comments_seen = false;
2916 continue;
2917
2918 case '.':
2919 tokens_seen = true;
2920 d = peek_char ();
2921 if (d >= '0' && d <= '9')
2922 return is_number (c);
2923 return Token.DOT;
2924
2925 case '#':
2926 if (tokens_seen || comments_seen) {
2927 Eror_WrongPreprocessorLocation ();
2928 return Token.ERROR;
2929 }
2930
2931 if (handle_preprocessing_directive (true))
2932 continue;
2933
2934 bool directive_expected = false;
2935 while ((c = get_char ()) != -1) {
2936 if (col == 1) {
2937 directive_expected = true;
2938 } else if (!directive_expected) {
2939 // TODO: Implement comment support for disabled code and uncomment this code
2940 // if (c == '#') {
2941 // Eror_WrongPreprocessorLocation ();
2942 // return Token.ERROR;
2943 // }
2944 continue;
2945 }
2946
2947 if (c == ' ' || c == '\t' || c == '\r' || c == '\n' || c == '\f' || c == '\v' )
2948 continue;
2949
2950 if (c == '#') {
2951 if (handle_preprocessing_directive (false))
2952 break;
2953 }
2954 directive_expected = false;
2955 }
2956
2957 if (c != -1) {
2958 tokens_seen = false;
2959 continue;
2960 }
2961
2962 return Token.EOF;
2963
2964 case '"':
2965 return consume_string (false);
2966
2967 case '\'':
2968 return TokenizeBackslash ();
2969
2970 case '@':
2971 c = get_char ();
2972 if (c == '"') {
2973 tokens_seen = true;
2974 return consume_string (true);
2975 }
2976
2977 if (is_identifier_start_character (c)){
2978 return consume_identifier (c, true);
2979 }
2980
2981 Report.Error (1646, Location, "Keyword, identifier, or string expected after verbatim specifier: @");
2982 return Token.ERROR;
2983
2984 case EvalStatementParserCharacter:
2985 return Token.EVAL_STATEMENT_PARSER;
2986 case EvalCompilationUnitParserCharacter:
2987 return Token.EVAL_COMPILATION_UNIT_PARSER;
2988 case EvalUsingDeclarationsParserCharacter:
2989 return Token.EVAL_USING_DECLARATIONS_UNIT_PARSER;
2990 }
2991
2992 if (is_identifier_start_character (c)) {
2993 tokens_seen = true;
2994 return consume_identifier (c);
2995 }
2996
2997 error_details = ((char)c).ToString ();
2998 return Token.ERROR;
2999 }
3000
3001 if (CompleteOnEOF){
3002 if (generated)
3003 return Token.COMPLETE_COMPLETION;
3004
3005 generated = true;
3006 return Token.GENERATE_COMPLETION;
3007 }
3008
3009
3010 return Token.EOF;
3011 }
3012
3013 int TokenizeBackslash ()
3014 {
3015 int c = get_char ();
3016 tokens_seen = true;
3017 if (c == '\'') {
3018 error_details = "Empty character literal";
3019 Report.Error (1011, Location, error_details);
3020 return Token.ERROR;
3021 }
3022 if (c == '\r' || c == '\n') {
3023 Report.Error (1010, Location, "Newline in constant");
3024 return Token.ERROR;
3025 }
3026
3027 int d;
3028 c = escape (c, out d);
3029 if (c == -1)
3030 return Token.ERROR;
3031 if (d != 0)
3032 throw new NotImplementedException ();
3033
3034 val = new CharLiteral ((char) c, Location);
3035 c = get_char ();
3036
3037 if (c != '\'') {
3038 Report.Error (1012, Location, "Too many characters in character literal");
3039
3040 // Try to recover, read until newline or next "'"
3041 while ((c = get_char ()) != -1) {
3042 if (c == '\n' || c == '\'')
3043 break;
3044 }
3045 return Token.ERROR;
3046 }
3047
3048 return Token.LITERAL;
3049 }
3050
3051 int TokenizeLessThan ()
3052 {
3053 int d;
3054 if (handle_typeof) {
3055 PushPosition ();
3056 if (parse_generic_dimension (out d)) {
3057 val = d;
3058 DiscardPosition ();
3059 return Token.GENERIC_DIMENSION;
3060 }
3061 PopPosition ();
3062 }
3063
3064 // Save current position and parse next token.
3065 PushPosition ();
3066 if (parse_less_than ()) {
3067 if (parsing_generic_declaration && token () != Token.DOT) {
3068 d = Token.OP_GENERICS_LT_DECL;
3069 } else {
3070 d = Token.OP_GENERICS_LT;
3071 }
3072 PopPosition ();
3073 return d;
3074 }
3075
3076 PopPosition ();
3077 parsing_generic_less_than = 0;
3078
3079 d = peek_char ();
3080 if (d == '<') {
3081 get_char ();
3082 d = peek_char ();
3083
3084 if (d == '=') {
3085 get_char ();
3086 return Token.OP_SHIFT_LEFT_ASSIGN;
3087 }
3088 return Token.OP_SHIFT_LEFT;
3089 }
3090
3091 if (d == '=') {
3092 get_char ();
3093 return Token.OP_LE;
3094 }
3095 return Token.OP_LT;
3096 }
3097
3098 //
3099 // Handles one line xml comment
3100 //
3101 private void handle_one_line_xml_comment ()
3102 {
3103 int c;
3104 while ((c = peek_char ()) == ' ')
3105 get_char (); // skip heading whitespaces.
3106 while ((c = peek_char ()) != -1 && c != '\n' && c != '\r') {
3107 xml_comment_buffer.Append ((char) get_char ());
3108 }
3109 if (c == '\r' || c == '\n')
3110 xml_comment_buffer.Append (Environment.NewLine);
3111 }
3112
3113 //
3114 // Remove heading "*" in Javadoc-like xml documentation.
3115 //
3116 private void update_formatted_doc_comment (int current_comment_start)
3117 {
3118 int length = xml_comment_buffer.Length - current_comment_start;
3119 string [] lines = xml_comment_buffer.ToString (
3120 current_comment_start,
3121 length).Replace ("\r", "").Split ('\n');
3122
3123 // The first line starts with /**, thus it is not target
3124 // for the format check.
3125 for (int i = 1; i < lines.Length; i++) {
3126 string s = lines [i];
3127 int idx = s.IndexOf ('*');
3128 string head = null;
3129 if (idx < 0) {
3130 if (i < lines.Length - 1)
3131 return;
3132 head = s;
3133 } else
3134 head = s.Substring (0, idx);
3135 foreach (char c in head)
3136 if (c != ' ')
3137 return;
3138 lines [i] = s.Substring (idx + 1);
3139 }
3140 xml_comment_buffer.Remove (current_comment_start, length);
3141 xml_comment_buffer.Insert (current_comment_start, String.Join (Environment.NewLine, lines));
3142 }
3143
3144 //
3145 // Updates current comment location.
3146 //
3147 private void update_comment_location ()
3148 {
3149 if (current_comment_location.IsNull) {
3150 // "-2" is for heading "//" or "/*"
3151 current_comment_location =
3152 new Location (ref_line, hidden ? -1 : col - 2);
3153 }
3154 }
3155
3156 //
3157 // Checks if there was incorrect doc comments and raise
3158 // warnings.
3159 //
3160 public void check_incorrect_doc_comment ()
3161 {
3162 if (xml_comment_buffer.Length > 0)
3163 warn_incorrect_doc_comment ();
3164 }
3165
3166 //
3167 // Raises a warning when tokenizer found incorrect doccomment
3168 // markup.
3169 //
3170 private void warn_incorrect_doc_comment ()
3171 {
3172 if (doc_state != XmlCommentState.Error) {
3173 doc_state = XmlCommentState.Error;
3174 // in csc, it is 'XML comment is not placed on
3175 // a valid language element'. But that does not
3176 // make sense.
3177 Report.Warning (1587, 2, Location, "XML comment is not placed on a valid language element");
3178 }
3179 }
3180
3181 //
3182 // Consumes the saved xml comment lines (if any)
3183 // as for current target member or type.
3184 //
3185 public string consume_doc_comment ()
3186 {
3187 if (xml_comment_buffer.Length > 0) {
3188 string ret = xml_comment_buffer.ToString ();
3189 reset_doc_comment ();
3190 return ret;
3191 }
3192 return null;
3193 }
3194
3195 Report Report {
3196 get { return context.Report; }
3197 }
3198
3199 void reset_doc_comment ()
3200 {
3201 xml_comment_buffer.Length = 0;
3202 current_comment_location = Location.Null;
3203 }
3204
3205 public void cleanup ()
3206 {
3207 if (ifstack != null && ifstack.Count >= 1) {
3208 int state = ifstack.Pop ();
3209 if ((state & REGION) != 0)
3210 Report.Error (1038, Location, "#endregion directive expected");
3211 else
3212 Report.Error (1027, Location, "Expected `#endif' directive");
3213 }
3214 }
3215 }
3216
3217 //
3218 // Indicates whether it accepts XML documentation or not.
3219 //
3220 public enum XmlCommentState {
3221 // comment is allowed in this state.
3222 Allowed,
3223 // comment is not allowed in this state.
3224 NotAllowed,
3225 // once comments appeared when it is NotAllowed, then the
3226 // state is changed to it, until the state is changed to
3227 // .Allowed.
3228 Error
3229 }
3230 }
3231
git-svn mirror of mcs