2 // cs-tokenizer.cs: The Tokenizer for the C# compiler
3 // This also implements the preprocessor
5 // Author: Miguel de Icaza (miguel@gnu.org)
6 // Marek Safar (marek.safar@gmail.com)
8 // Dual licensed under the terms of the MIT X11 or GNU GPL
10 // Copyright 2001, 2002 Ximian, Inc (http://www.ximian.com)
11 // Copyright 2004-2008 Novell, Inc
12 // Copyright 2011 Xamarin, Inc (http://www.xamarin.com)
17 using System.Collections.Generic;
18 using System.Globalization;
19 using System.Diagnostics;
20 using System.Collections;
25 // This class has to be used by parser only, it reuses token
26 // details once a file is parsed
28 public class LocatedToken
30 public int row, column;
32 public SourceFile file;
34 public LocatedToken ()
38 public LocatedToken (string value, Location loc)
41 file = loc.SourceFile;
46 public override string ToString ()
48 return string.Format ("Token '{0}' at {1},{2}", Value, row, column);
51 public Location Location
53 get { return new Location (file, row, column); }
63 /// Tokenizer for C# source code.
65 public class Tokenizer : yyParser.yyInput
69 public readonly T Token;
70 public KeywordEntry<T> Next;
71 public readonly char[] Value;
73 public KeywordEntry (string value, T token)
75 this.Value = value.ToCharArray ();
80 sealed class IdentifiersComparer : IEqualityComparer<char[]>
84 public IdentifiersComparer (int length)
89 public bool Equals (char[] x, char[] y)
91 for (int i = 0; i < length; ++i)
98 public int GetHashCode (char[] obj)
101 for (int i = 0; i < length; ++i)
102 h = (h << 5) - h + obj [i];
108 public class LocatedTokenBuffer
110 readonly LocatedToken[] buffer;
113 public LocatedTokenBuffer ()
115 buffer = new LocatedToken[0];
118 public LocatedTokenBuffer (LocatedToken[] buffer)
120 this.buffer = buffer ?? new LocatedToken[0];
123 public LocatedToken Create (SourceFile file, int row, int column)
125 return Create (null, file, row, column);
128 public LocatedToken Create (string value, SourceFile file, int row, int column)
131 // TODO: I am not very happy about the logic but it's the best
132 // what I could come up with for now.
133 // Ideally we should be using just tiny buffer (256 elements) which
134 // is enough to hold all details for currect stack and recycle elements
135 // poped from the stack but there is a trick needed to recycle
139 if (pos >= buffer.Length) {
140 entry = new LocatedToken ();
144 entry = new LocatedToken ();
153 entry.column = column;
158 // Used for token not required by expression evaluator
160 [Conditional ("FULL_AST")]
161 public void CreateOptional (SourceFile file, int row, int col, ref object token)
163 token = Create (file, row, col);
167 public enum PreprocessorDirective
173 If = 3 | RequiresArgument,
175 Elif = 5 | RequiresArgument,
177 Define = 7 | RequiresArgument,
178 Undef = 8 | RequiresArgument,
181 Pragma = 11 | CustomArgumentsParsing,
182 Line = 12 | CustomArgumentsParsing,
184 CustomArgumentsParsing = 1 << 10,
185 RequiresArgument = 1 << 11
188 readonly SeekableStreamReader reader;
189 readonly CompilationSourceFile source_file;
190 readonly CompilerContext context;
191 readonly Report Report;
194 SourceFile current_source;
195 Location hidden_block_start;
201 readonly int tab_size;
202 bool handle_get_set = false;
203 bool handle_remove_add = false;
205 bool lambda_arguments_parsing;
206 List<Location> escaped_identifiers;
207 int parsing_generic_less_than;
208 readonly bool doc_processing;
209 readonly LocatedTokenBuffer ltb;
212 // Used mainly for parser optimizations. Some expressions for instance
213 // can appear only in block (including initializer, base initializer)
216 public int parsing_block;
217 internal bool query_parsing;
220 // When parsing type only, useful for ambiguous nullable types
222 public int parsing_type;
225 // Set when parsing generic declaration (type or method header)
227 public bool parsing_generic_declaration;
228 public bool parsing_generic_declaration_doc;
231 // The value indicates that we have not reach any declaration or
234 public int parsing_declaration;
236 public bool parsing_attribute_section;
238 public bool parsing_modifiers;
240 public bool parsing_catch_when;
242 int parsing_string_interpolation;
244 public bool parsing_interpolation_format;
247 // The special characters to inject on streams to run the unit parser
248 // in the special expression mode. Using private characters from
249 // Plane Sixteen (U+100000 to U+10FFFD)
251 // This character is only tested just before the tokenizer is about to report
252 // an error; So on the regular operation mode, this addition will have no
253 // impact on the tokenizer's performance.
256 public const int EvalStatementParserCharacter = 0x100000;
257 public const int EvalCompilationUnitParserCharacter = 0x100001;
258 public const int EvalUsingDeclarationsParserCharacter = 0x100002;
259 public const int DocumentationXref = 0x100003;
261 const int UnicodeLS = 0x2028;
262 const int UnicodePS = 0x2029;
265 // XML documentation buffer. The save point is used to divide
266 // comments on types and comments on members.
268 StringBuilder xml_comment_buffer;
271 // See comment on XmlCommentState enumeration.
273 XmlCommentState xml_doc_state = XmlCommentState.Allowed;
276 // Whether tokens have been seen on this line
278 bool tokens_seen = false;
281 // Set to true once the GENERATE_COMPLETION token has bee
282 // returned. This helps produce one GENERATE_COMPLETION,
283 // as many COMPLETE_COMPLETION as necessary to complete the
284 // AST tree and one final EOF.
289 // Whether a token has been seen on the file
290 // This is needed because `define' is not allowed to be used
291 // after a token has been seen.
298 static readonly KeywordEntry<int>[][] keywords;
299 static readonly KeywordEntry<PreprocessorDirective>[][] keywords_preprocessor;
300 static readonly HashSet<string> keyword_strings;
301 static readonly NumberStyles styles;
302 static readonly NumberFormatInfo csharp_format_info;
305 static readonly char[] pragma_warning = "warning".ToCharArray ();
306 static readonly char[] pragma_warning_disable = "disable".ToCharArray ();
307 static readonly char[] pragma_warning_restore = "restore".ToCharArray ();
308 static readonly char[] pragma_checksum = "checksum".ToCharArray ();
309 static readonly char[] line_hidden = "hidden".ToCharArray ();
310 static readonly char[] line_default = "default".ToCharArray ();
312 static readonly char[] simple_whitespaces = new char[] { ' ', '\t' };
314 public bool PropertyParsing {
315 get { return handle_get_set; }
316 set { handle_get_set = value; }
319 public bool EventParsing {
320 get { return handle_remove_add; }
321 set { handle_remove_add = value; }
324 public bool ConstraintsParsing {
325 get { return handle_where; }
326 set { handle_where = value; }
329 public XmlCommentState doc_state {
330 get { return xml_doc_state; }
332 if (value == XmlCommentState.Allowed) {
333 check_incorrect_doc_comment ();
334 reset_doc_comment ();
336 xml_doc_state = value;
341 // This is used to trigger completion generation on the parser
342 public bool CompleteOnEOF;
344 void AddEscapedIdentifier (Location loc)
346 if (escaped_identifiers == null)
347 escaped_identifiers = new List<Location> ();
349 escaped_identifiers.Add (loc);
352 public bool IsEscapedIdentifier (ATypeNameExpression name)
354 return escaped_identifiers != null && escaped_identifiers.Contains (name.Location);
358 // Values for the associated token returned
360 internal int putback_char; // Used by repl only
366 const int TAKING = 1;
367 const int ELSE_SEEN = 4;
368 const int PARENT_TAKING = 8;
369 const int REGION = 16;
372 // pre-processor if stack state:
376 public const int MaxIdentifierLength = 512;
377 public const int MaxNumberLength = 512;
379 readonly char[] id_builder;
380 readonly Dictionary<char[], string>[] identifiers;
381 readonly char[] number_builder;
384 char[] value_builder = new char[64];
393 // This is used when the tokenizer needs to save
394 // the current position as it needs to do some parsing
395 // on its own to deamiguate a token in behalf of the
398 Stack<Position> position_stack = new Stack<Position> (2);
405 public Location hidden;
406 public int putback_char;
407 public int previous_col;
408 public Stack<int> ifstack;
409 public int parsing_generic_less_than;
410 public int current_token;
412 public int parsing_string_interpolation;
414 public Position (Tokenizer t)
416 position = t.reader.Position;
418 ref_line = t.ref_line;
420 hidden = t.hidden_block_start;
421 putback_char = t.putback_char;
422 previous_col = t.previous_col;
423 if (t.ifstack != null && t.ifstack.Count != 0) {
424 // There is no simple way to clone Stack<T> all
425 // methods reverse the order
426 var clone = t.ifstack.ToArray ();
427 Array.Reverse (clone);
428 ifstack = new Stack<int> (clone);
430 parsing_generic_less_than = t.parsing_generic_less_than;
431 parsing_string_interpolation = t.parsing_string_interpolation;
432 current_token = t.current_token;
437 public Tokenizer (SeekableStreamReader input, CompilationSourceFile file, ParserSession session, Report report)
439 this.source_file = file;
440 this.context = file.Compiler;
441 this.current_source = file.SourceFile;
442 this.identifiers = session.Identifiers;
443 this.id_builder = session.IDBuilder;
444 this.number_builder = session.NumberBuilder;
445 this.ltb = new LocatedTokenBuffer (session.LocatedTokens);
446 this.Report = report;
452 xml_comment_buffer = new StringBuilder ();
453 doc_processing = context.Settings.DocumentationFile != null;
455 tab_size = context.Settings.TabSize;
458 public void PushPosition ()
460 position_stack.Push (new Position (this));
463 public void PopPosition ()
465 Position p = position_stack.Pop ();
467 reader.Position = p.position;
468 ref_line = p.ref_line;
471 hidden_block_start = p.hidden;
472 putback_char = p.putback_char;
473 previous_col = p.previous_col;
475 parsing_generic_less_than = p.parsing_generic_less_than;
476 parsing_string_interpolation = p.parsing_string_interpolation;
477 current_token = p.current_token;
481 // Do not reset the position, ignore it.
482 public void DiscardPosition ()
484 position_stack.Pop ();
487 static void AddKeyword (string kw, int token)
489 keyword_strings.Add (kw);
491 AddKeyword (keywords, kw, token);
494 static void AddPreprocessorKeyword (string kw, PreprocessorDirective directive)
496 AddKeyword (keywords_preprocessor, kw, directive);
499 static void AddKeyword<T> (KeywordEntry<T>[][] keywords, string kw, T token)
501 int length = kw.Length;
502 if (keywords[length] == null) {
503 keywords[length] = new KeywordEntry<T>['z' - '_' + 1];
506 int char_index = kw[0] - '_';
507 var kwe = keywords[length][char_index];
509 keywords[length][char_index] = new KeywordEntry<T> (kw, token);
513 while (kwe.Next != null) {
517 kwe.Next = new KeywordEntry<T> (kw, token);
525 keyword_strings = new HashSet<string> ();
527 // 11 is the length of the longest keyword for now
528 keywords = new KeywordEntry<int>[11][];
530 AddKeyword ("__arglist", Token.ARGLIST);
531 AddKeyword ("__makeref", Token.MAKEREF);
532 AddKeyword ("__reftype", Token.REFTYPE);
533 AddKeyword ("__refvalue", Token.REFVALUE);
534 AddKeyword ("abstract", Token.ABSTRACT);
535 AddKeyword ("as", Token.AS);
536 AddKeyword ("add", Token.ADD);
537 AddKeyword ("base", Token.BASE);
538 AddKeyword ("bool", Token.BOOL);
539 AddKeyword ("break", Token.BREAK);
540 AddKeyword ("byte", Token.BYTE);
541 AddKeyword ("case", Token.CASE);
542 AddKeyword ("catch", Token.CATCH);
543 AddKeyword ("char", Token.CHAR);
544 AddKeyword ("checked", Token.CHECKED);
545 AddKeyword ("class", Token.CLASS);
546 AddKeyword ("const", Token.CONST);
547 AddKeyword ("continue", Token.CONTINUE);
548 AddKeyword ("decimal", Token.DECIMAL);
549 AddKeyword ("default", Token.DEFAULT);
550 AddKeyword ("delegate", Token.DELEGATE);
551 AddKeyword ("do", Token.DO);
552 AddKeyword ("double", Token.DOUBLE);
553 AddKeyword ("else", Token.ELSE);
554 AddKeyword ("enum", Token.ENUM);
555 AddKeyword ("event", Token.EVENT);
556 AddKeyword ("explicit", Token.EXPLICIT);
557 AddKeyword ("extern", Token.EXTERN);
558 AddKeyword ("false", Token.FALSE);
559 AddKeyword ("finally", Token.FINALLY);
560 AddKeyword ("fixed", Token.FIXED);
561 AddKeyword ("float", Token.FLOAT);
562 AddKeyword ("for", Token.FOR);
563 AddKeyword ("foreach", Token.FOREACH);
564 AddKeyword ("goto", Token.GOTO);
565 AddKeyword ("get", Token.GET);
566 AddKeyword ("if", Token.IF);
567 AddKeyword ("implicit", Token.IMPLICIT);
568 AddKeyword ("in", Token.IN);
569 AddKeyword ("int", Token.INT);
570 AddKeyword ("interface", Token.INTERFACE);
571 AddKeyword ("internal", Token.INTERNAL);
572 AddKeyword ("is", Token.IS);
573 AddKeyword ("lock", Token.LOCK);
574 AddKeyword ("long", Token.LONG);
575 AddKeyword ("namespace", Token.NAMESPACE);
576 AddKeyword ("new", Token.NEW);
577 AddKeyword ("null", Token.NULL);
578 AddKeyword ("object", Token.OBJECT);
579 AddKeyword ("operator", Token.OPERATOR);
580 AddKeyword ("out", Token.OUT);
581 AddKeyword ("override", Token.OVERRIDE);
582 AddKeyword ("params", Token.PARAMS);
583 AddKeyword ("private", Token.PRIVATE);
584 AddKeyword ("protected", Token.PROTECTED);
585 AddKeyword ("public", Token.PUBLIC);
586 AddKeyword ("readonly", Token.READONLY);
587 AddKeyword ("ref", Token.REF);
588 AddKeyword ("remove", Token.REMOVE);
589 AddKeyword ("return", Token.RETURN);
590 AddKeyword ("sbyte", Token.SBYTE);
591 AddKeyword ("sealed", Token.SEALED);
592 AddKeyword ("set", Token.SET);
593 AddKeyword ("short", Token.SHORT);
594 AddKeyword ("sizeof", Token.SIZEOF);
595 AddKeyword ("stackalloc", Token.STACKALLOC);
596 AddKeyword ("static", Token.STATIC);
597 AddKeyword ("string", Token.STRING);
598 AddKeyword ("struct", Token.STRUCT);
599 AddKeyword ("switch", Token.SWITCH);
600 AddKeyword ("this", Token.THIS);
601 AddKeyword ("throw", Token.THROW);
602 AddKeyword ("true", Token.TRUE);
603 AddKeyword ("try", Token.TRY);
604 AddKeyword ("typeof", Token.TYPEOF);
605 AddKeyword ("uint", Token.UINT);
606 AddKeyword ("ulong", Token.ULONG);
607 AddKeyword ("unchecked", Token.UNCHECKED);
608 AddKeyword ("unsafe", Token.UNSAFE);
609 AddKeyword ("ushort", Token.USHORT);
610 AddKeyword ("using", Token.USING);
611 AddKeyword ("virtual", Token.VIRTUAL);
612 AddKeyword ("void", Token.VOID);
613 AddKeyword ("volatile", Token.VOLATILE);
614 AddKeyword ("while", Token.WHILE);
615 AddKeyword ("partial", Token.PARTIAL);
616 AddKeyword ("where", Token.WHERE);
619 AddKeyword ("from", Token.FROM);
620 AddKeyword ("join", Token.JOIN);
621 AddKeyword ("on", Token.ON);
622 AddKeyword ("equals", Token.EQUALS);
623 AddKeyword ("select", Token.SELECT);
624 AddKeyword ("group", Token.GROUP);
625 AddKeyword ("by", Token.BY);
626 AddKeyword ("let", Token.LET);
627 AddKeyword ("orderby", Token.ORDERBY);
628 AddKeyword ("ascending", Token.ASCENDING);
629 AddKeyword ("descending", Token.DESCENDING);
630 AddKeyword ("into", Token.INTO);
632 // Contextual async keywords
633 AddKeyword ("async", Token.ASYNC);
634 AddKeyword ("await", Token.AWAIT);
636 // Contextual filter catch keyword
637 AddKeyword ("when", Token.WHEN);
639 keywords_preprocessor = new KeywordEntry<PreprocessorDirective>[10][];
641 AddPreprocessorKeyword ("region", PreprocessorDirective.Region);
642 AddPreprocessorKeyword ("endregion", PreprocessorDirective.Endregion);
643 AddPreprocessorKeyword ("if", PreprocessorDirective.If);
644 AddPreprocessorKeyword ("endif", PreprocessorDirective.Endif);
645 AddPreprocessorKeyword ("elif", PreprocessorDirective.Elif);
646 AddPreprocessorKeyword ("else", PreprocessorDirective.Else);
647 AddPreprocessorKeyword ("define", PreprocessorDirective.Define);
648 AddPreprocessorKeyword ("undef", PreprocessorDirective.Undef);
649 AddPreprocessorKeyword ("error", PreprocessorDirective.Error);
650 AddPreprocessorKeyword ("warning", PreprocessorDirective.Warning);
651 AddPreprocessorKeyword ("pragma", PreprocessorDirective.Pragma);
652 AddPreprocessorKeyword ("line", PreprocessorDirective.Line);
654 csharp_format_info = NumberFormatInfo.InvariantInfo;
655 styles = NumberStyles.Float;
658 int GetKeyword (char[] id, int id_len)
661 // Keywords are stored in an array of arrays grouped by their
662 // length and then by the first character
664 if (id_len >= keywords.Length || keywords [id_len] == null)
667 int first_index = id [0] - '_';
668 if (first_index > 'z' - '_')
671 var kwe = keywords [id_len] [first_index];
678 for (int i = 1; i < id_len; ++i) {
679 if (id [i] != kwe.Value [i]) {
685 } while (res == 0 && kwe != null);
699 if (!handle_remove_add)
703 if (parsing_declaration == 0)
704 res = Token.EXTERN_ALIAS;
707 if (peek_token () == Token.COLON) {
709 res = Token.DEFAULT_COLON;
713 if (current_token != Token.CATCH && !parsing_catch_when)
717 if (!(handle_where && current_token != Token.COLON) && !query_parsing)
722 // A query expression is any expression that starts with `from identifier'
723 // followed by any token except ; , =
725 if (!query_parsing) {
726 if (lambda_arguments_parsing || parsing_block == 0) {
732 // HACK: to disable generics micro-parser, because PushPosition does not
733 // store identifiers array
734 parsing_generic_less_than = 1;
736 case Token.IDENTIFIER:
749 next_token = xtoken ();
750 if (next_token == Token.SEMICOLON || next_token == Token.COMMA || next_token == Token.EQUALS || next_token == Token.ASSIGN)
753 res = Token.FROM_FIRST;
754 query_parsing = true;
755 if (context.Settings.Version <= LanguageVersion.ISO_2)
756 Report.FeatureIsNotAvailable (context, Location, "query expressions");
759 Expression.Error_VoidInvalidInTheContext (Location, Report);
763 // HACK: A token is not a keyword so we need to restore identifiers buffer
764 // which has been overwritten before we grabbed the identifier
765 id_builder [0] = 'f'; id_builder [1] = 'r'; id_builder [2] = 'o'; id_builder [3] = 'm';
779 case Token.ASCENDING:
780 case Token.DESCENDING:
787 case Token.NAMESPACE:
788 // TODO: some explanation needed
789 check_incorrect_doc_comment ();
790 parsing_modifiers = false;
794 if (parsing_block > 0) {
799 // Save current position and parse next token.
802 next_token = token ();
803 bool ok = (next_token == Token.CLASS) ||
804 (next_token == Token.STRUCT) ||
805 (next_token == Token.INTERFACE) ||
806 (next_token == Token.VOID);
811 if (next_token == Token.VOID) {
812 if (context.Settings.Version <= LanguageVersion.ISO_2)
813 Report.FeatureIsNotAvailable (context, Location, "partial methods");
814 } else if (context.Settings.Version == LanguageVersion.ISO_1)
815 Report.FeatureIsNotAvailable (context, Location, "partial types");
820 if (next_token < Token.LAST_KEYWORD) {
821 Report.Error (267, Location,
822 "The `partial' modifier can be used only immediately before `class', `struct', `interface', or `void' keyword");
826 // HACK: A token is not a keyword so we need to restore identifiers buffer
827 // which has been overwritten before we grabbed the identifier
839 if (parsing_modifiers) {
841 // Skip attributes section or constructor called async
843 if (parsing_attribute_section || peek_token () == Token.OPEN_PARENS) {
848 } else if (parsing_block > 0) {
849 switch (peek_token ()) {
851 case Token.OPEN_PARENS_LAMBDA:
854 case Token.IDENTIFIER:
857 if (xtoken () != Token.ARROW) {
865 // peek_token could overwrite id_buffer
866 id_builder [0] = 'a'; id_builder [1] = 's'; id_builder [2] = 'y'; id_builder [3] = 'n'; id_builder [4] = 'c';
874 if (res == Token.ASYNC && context.Settings.Version <= LanguageVersion.V_4) {
875 Report.FeatureIsNotAvailable (context, Location, "asynchronous functions");
881 if (parsing_block == 0)
891 static PreprocessorDirective GetPreprocessorDirective (char[] id, int id_len)
894 // Keywords are stored in an array of arrays grouped by their
895 // length and then by the first character
897 if (id_len >= keywords_preprocessor.Length || keywords_preprocessor[id_len] == null)
898 return PreprocessorDirective.Invalid;
900 int first_index = id[0] - '_';
901 if (first_index > 'z' - '_')
902 return PreprocessorDirective.Invalid;
904 var kwe = keywords_preprocessor[id_len][first_index];
906 return PreprocessorDirective.Invalid;
908 PreprocessorDirective res = PreprocessorDirective.Invalid;
911 for (int i = 1; i < id_len; ++i) {
912 if (id[i] != kwe.Value[i]) {
918 } while (res == PreprocessorDirective.Invalid && kwe != null);
923 public Location Location {
925 return new Location (current_source, ref_line, col);
929 static bool is_identifier_start_character (int c)
931 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || Char.IsLetter ((char)c);
934 static bool is_identifier_part_character (char c)
936 if (c >= 'a' && c <= 'z')
939 if (c >= 'A' && c <= 'Z')
942 if (c == '_' || (c >= '0' && c <= '9'))
948 return is_identifier_part_character_slow_part (c);
951 static bool is_identifier_part_character_slow_part (char c)
953 if (Char.IsLetter (c))
956 switch (Char.GetUnicodeCategory (c)) {
957 case UnicodeCategory.ConnectorPunctuation:
959 // combining-character: A Unicode character of classes Mn or Mc
960 case UnicodeCategory.NonSpacingMark:
961 case UnicodeCategory.SpacingCombiningMark:
963 // decimal-digit-character: A Unicode character of the class Nd
964 case UnicodeCategory.DecimalDigitNumber:
971 public static bool IsKeyword (string s)
973 return keyword_strings.Contains (s);
977 // Open parens micro parser. Detects both lambda and cast ambiguity.
979 int TokenizeOpenParens ()
984 int bracket_level = 0;
985 bool is_type = false;
986 bool can_be_type = false;
989 ptoken = current_token;
992 switch (current_token) {
993 case Token.CLOSE_PARENS:
997 // Expression inside parens is lambda, (int i) =>
999 if (current_token == Token.ARROW)
1000 return Token.OPEN_PARENS_LAMBDA;
1003 // Expression inside parens is single type, (int[])
1006 if (current_token == Token.SEMICOLON)
1007 return Token.OPEN_PARENS;
1009 return Token.OPEN_PARENS_CAST;
1013 // Expression is possible cast, look at next token, (T)null
1016 switch (current_token) {
1017 case Token.OPEN_PARENS:
1020 case Token.IDENTIFIER:
1024 case Token.DELEGATE:
1034 case Token.UNCHECKED:
1040 // These can be part of a member access
1055 return Token.OPEN_PARENS_CAST;
1058 return Token.OPEN_PARENS;
1061 case Token.DOUBLE_COLON:
1062 if (ptoken != Token.IDENTIFIER && ptoken != Token.OP_GENERICS_GT)
1067 case Token.IDENTIFIER:
1071 if (bracket_level == 0) {
1077 case Token.OP_GENERICS_LT:
1079 case Token.DOUBLE_COLON:
1081 if (bracket_level == 0)
1085 can_be_type = is_type = false;
1105 if (bracket_level == 0)
1110 if (bracket_level == 0) {
1111 bracket_level = 100;
1112 can_be_type = is_type = false;
1116 case Token.OP_GENERICS_LT:
1117 case Token.OPEN_BRACKET:
1118 if (bracket_level++ == 0)
1122 case Token.OP_GENERICS_GT:
1123 case Token.CLOSE_BRACKET:
1127 case Token.INTERR_NULLABLE:
1129 if (bracket_level == 0)
1135 can_be_type = is_type = false;
1139 return Token.OPEN_PARENS;
1144 public static bool IsValidIdentifier (string s)
1146 if (s == null || s.Length == 0)
1149 if (!is_identifier_start_character (s [0]))
1152 for (int i = 1; i < s.Length; i ++)
1153 if (! is_identifier_part_character (s [i]))
1159 bool parse_less_than (ref int genericDimension)
1162 int the_token = token ();
1163 if (the_token == Token.OPEN_BRACKET) {
1165 the_token = token ();
1166 if (the_token == Token.EOF)
1169 if (the_token == Token.CLOSE_BRACKET)
1172 the_token = token ();
1173 } else if (the_token == Token.IN || the_token == Token.OUT) {
1174 the_token = token ();
1176 switch (the_token) {
1177 case Token.IDENTIFIER:
1195 case Token.OP_GENERICS_GT:
1196 genericDimension = 1;
1204 the_token = token ();
1205 } while (the_token == Token.COMMA);
1207 if (the_token == Token.OP_GENERICS_GT) {
1217 the_token = token ();
1219 if (the_token == Token.OP_GENERICS_GT)
1221 else if (the_token == Token.COMMA || the_token == Token.DOT || the_token == Token.DOUBLE_COLON)
1223 else if (the_token == Token.INTERR_NULLABLE || the_token == Token.STAR)
1225 else if (the_token == Token.OP_GENERICS_LT) {
1226 if (!parse_less_than (ref genericDimension))
1229 } else if (the_token == Token.OPEN_BRACKET) {
1231 the_token = token ();
1232 if (the_token == Token.CLOSE_BRACKET)
1234 else if (the_token == Token.COMMA)
1235 goto rank_specifiers;
1242 public int peek_token ()
1247 the_token = token ();
1254 // Tonizes `?' using custom disambiguous rules to return one
1255 // of following tokens: INTERR_NULLABLE, OP_COALESCING, INTERR
1257 // Tricky expression looks like:
1259 // Foo ? a = x ? b : c;
1261 int TokenizePossibleNullableType ()
1263 if (parsing_block == 0 || parsing_type > 0)
1264 return Token.INTERR_NULLABLE;
1266 int d = peek_char ();
1269 return Token.OP_COALESCING;
1273 return Token.INTERR_OPERATOR;
1277 if (d == ',' || d == ';' || d == '>')
1278 return Token.INTERR_NULLABLE;
1279 if (d == '*' || (d >= '0' && d <= '9'))
1280 return Token.INTERR;
1284 current_token = Token.NONE;
1293 case Token.OPEN_BRACKET_EXPR:
1294 next_token = Token.INTERR_OPERATOR;
1302 case Token.INTERPOLATED_STRING:
1303 next_token = Token.INTERR;
1306 case Token.SEMICOLON:
1308 case Token.CLOSE_PARENS:
1309 case Token.OPEN_BRACKET:
1310 case Token.OP_GENERICS_GT:
1312 case Token.OP_COALESCING:
1314 next_token = Token.INTERR_NULLABLE;
1317 case Token.OPEN_PARENS:
1318 case Token.OPEN_PARENS_CAST:
1319 case Token.OPEN_PARENS_LAMBDA:
1324 case Token.OP_GENERICS_LT:
1325 case Token.OP_GENERICS_LT_DECL:
1326 case Token.GENERIC_DIMENSION:
1336 if (next_token == -1) {
1337 switch (xtoken ()) {
1339 case Token.SEMICOLON:
1340 case Token.OPEN_BRACE:
1342 next_token = Token.INTERR_NULLABLE;
1346 next_token = Token.INTERR;
1349 case Token.OPEN_PARENS:
1350 case Token.OPEN_PARENS_CAST:
1351 case Token.OPEN_PARENS_LAMBDA:
1355 case Token.OPEN_BRACKET:
1356 case Token.OPEN_BRACKET_EXPR:
1360 case Token.CLOSE_PARENS:
1364 case Token.OP_GENERICS_LT:
1365 case Token.OP_GENERICS_LT_DECL:
1366 case Token.GENERIC_DIMENSION:
1376 // All shorcuts failed, do it hard way
1378 while ((ntoken = xtoken ()) != Token.EOF) {
1380 case Token.OPEN_BRACE:
1383 case Token.OPEN_PARENS:
1384 case Token.OPEN_PARENS_CAST:
1385 case Token.OPEN_PARENS_LAMBDA:
1388 case Token.CLOSE_BRACE:
1391 case Token.OP_GENERICS_LT:
1392 case Token.OP_GENERICS_LT_DECL:
1393 case Token.GENERIC_DIMENSION:
1396 case Token.OPEN_BRACKET:
1397 case Token.OPEN_BRACKET_EXPR:
1400 case Token.CLOSE_BRACKET:
1403 case Token.CLOSE_PARENS:
1410 return Token.INTERR_NULLABLE;
1412 case Token.OP_GENERICS_GT:
1419 return Token.INTERR_NULLABLE;
1425 if (ntoken == Token.SEMICOLON)
1431 if (ntoken == Token.COMMA) {
1432 if (generics != 0 || brackets != 0)
1436 return Token.INTERR_NULLABLE;
1439 if (ntoken == Token.COLON) {
1440 if (++colons == interrs)
1445 if (ntoken == Token.INTERR) {
1451 next_token = colons != interrs && braces == 0 ? Token.INTERR_NULLABLE : Token.INTERR;
1460 bool decimal_digits (int c)
1463 bool seen_digits = false;
1466 if (number_pos == MaxNumberLength)
1467 Error_NumericConstantTooLong ();
1468 number_builder [number_pos++] = (char) c;
1472 // We use peek_char2, because decimal_digits needs to do a
1473 // 2-character look-ahead (5.ToString for example).
1475 while ((d = peek_char2 ()) != -1){
1476 if (d >= '0' && d <= '9'){
1477 if (number_pos == MaxNumberLength)
1478 Error_NumericConstantTooLong ();
1479 number_builder [number_pos++] = (char) d;
1489 static bool is_hex (int e)
1491 return (e >= '0' && e <= '9') || (e >= 'A' && e <= 'F') || (e >= 'a' && e <= 'f');
1494 static TypeCode real_type_suffix (int c)
1498 return TypeCode.Single;
1500 return TypeCode.Double;
1502 return TypeCode.Decimal;
1504 return TypeCode.Empty;
1508 ILiteralConstant integer_type_suffix (ulong ul, int c, Location loc)
1510 bool is_unsigned = false;
1511 bool is_long = false;
1514 bool scanning = true;
1527 // if we have not seen anything in between
1528 // report this error
1530 Report.Warning (78, 4, Location, "The `l' suffix is easily confused with the digit `1' (use `L' for clarity)");
1550 if (is_long && is_unsigned){
1551 return new ULongLiteral (context.BuiltinTypes, ul, loc);
1555 // uint if possible, or ulong else.
1557 if ((ul & 0xffffffff00000000) == 0)
1558 return new UIntLiteral (context.BuiltinTypes, (uint) ul, loc);
1560 return new ULongLiteral (context.BuiltinTypes, ul, loc);
1561 } else if (is_long){
1562 // long if possible, ulong otherwise
1563 if ((ul & 0x8000000000000000) != 0)
1564 return new ULongLiteral (context.BuiltinTypes, ul, loc);
1566 return new LongLiteral (context.BuiltinTypes, (long) ul, loc);
1568 // int, uint, long or ulong in that order
1569 if ((ul & 0xffffffff00000000) == 0){
1570 uint ui = (uint) ul;
1572 if ((ui & 0x80000000) != 0)
1573 return new UIntLiteral (context.BuiltinTypes, ui, loc);
1575 return new IntLiteral (context.BuiltinTypes, (int) ui, loc);
1577 if ((ul & 0x8000000000000000) != 0)
1578 return new ULongLiteral (context.BuiltinTypes, ul, loc);
1580 return new LongLiteral (context.BuiltinTypes, (long) ul, loc);
1586 // given `c' as the next char in the input decide whether
1587 // we need to convert to a special type, and then choose
1588 // the best representation for the integer
1590 ILiteralConstant adjust_int (int c, Location loc)
1593 if (number_pos > 9){
1594 ulong ul = (uint) (number_builder [0] - '0');
1596 for (int i = 1; i < number_pos; i++){
1597 ul = checked ((ul * 10) + ((uint)(number_builder [i] - '0')));
1600 return integer_type_suffix (ul, c, loc);
1602 uint ui = (uint) (number_builder [0] - '0');
1604 for (int i = 1; i < number_pos; i++){
1605 ui = checked ((ui * 10) + ((uint)(number_builder [i] - '0')));
1608 return integer_type_suffix (ui, c, loc);
1610 } catch (OverflowException) {
1611 Error_NumericConstantTooLong ();
1612 return new IntLiteral (context.BuiltinTypes, 0, loc);
1614 catch (FormatException) {
1615 Report.Error (1013, Location, "Invalid number");
1616 return new IntLiteral (context.BuiltinTypes, 0, loc);
1620 ILiteralConstant adjust_real (TypeCode t, Location loc)
1622 string s = new string (number_builder, 0, number_pos);
1623 const string error_details = "Floating-point constant is outside the range of type `{0}'";
1626 case TypeCode.Decimal:
1628 return new DecimalLiteral (context.BuiltinTypes, decimal.Parse (s, styles, csharp_format_info), loc);
1629 } catch (OverflowException) {
1630 Report.Error (594, Location, error_details, "decimal");
1631 return new DecimalLiteral (context.BuiltinTypes, 0, loc);
1633 case TypeCode.Single:
1635 return new FloatLiteral (context.BuiltinTypes, float.Parse (s, styles, csharp_format_info), loc);
1636 } catch (OverflowException) {
1637 Report.Error (594, Location, error_details, "float");
1638 return new FloatLiteral (context.BuiltinTypes, 0, loc);
1642 return new DoubleLiteral (context.BuiltinTypes, double.Parse (s, styles, csharp_format_info), loc);
1643 } catch (OverflowException) {
1644 Report.Error (594, loc, error_details, "double");
1645 return new DoubleLiteral (context.BuiltinTypes, 0, loc);
1650 ILiteralConstant handle_hex (Location loc)
1656 while ((d = peek_char ()) != -1){
1658 number_builder [number_pos++] = (char) d;
1664 string s = new String (number_builder, 0, number_pos);
1667 if (number_pos <= 8)
1668 ul = System.UInt32.Parse (s, NumberStyles.HexNumber);
1670 ul = System.UInt64.Parse (s, NumberStyles.HexNumber);
1672 return integer_type_suffix (ul, peek_char (), loc);
1673 } catch (OverflowException){
1674 Error_NumericConstantTooLong ();
1675 return new IntLiteral (context.BuiltinTypes, 0, loc);
1677 catch (FormatException) {
1678 Report.Error (1013, Location, "Invalid number");
1679 return new IntLiteral (context.BuiltinTypes, 0, loc);
1684 // Invoked if we know we have .digits or digits
1686 int is_number (int c, bool dotLead)
1688 ILiteralConstant res;
1691 int read_start = reader.Position - 1;
1694 // Caller did peek_char
1704 int peek = peek_char ();
1706 if (peek == 'x' || peek == 'X') {
1707 val = res = handle_hex (loc);
1709 res.ParsedValue = reader.ReadChars (read_start, reader.Position - 1);
1712 return Token.LITERAL;
1720 // We need to handle the case of
1721 // "1.1" vs "1.string" (LITERAL_FLOAT vs NUMBER DOT IDENTIFIER)
1723 bool is_real = false;
1728 if (decimal_digits ('.')){
1734 val = res = adjust_int (-1, loc);
1737 res.ParsedValue = reader.ReadChars (read_start, reader.Position - 1);
1739 return Token.LITERAL;
1743 if (c == 'e' || c == 'E'){
1746 if (number_pos == MaxNumberLength)
1747 Error_NumericConstantTooLong ();
1748 number_builder [number_pos++] = (char) c;
1752 if (number_pos == MaxNumberLength)
1753 Error_NumericConstantTooLong ();
1754 number_builder [number_pos++] = '+';
1756 } else if (c == '-') {
1757 if (number_pos == MaxNumberLength)
1758 Error_NumericConstantTooLong ();
1759 number_builder [number_pos++] = '-';
1762 if (number_pos == MaxNumberLength)
1763 Error_NumericConstantTooLong ();
1764 number_builder [number_pos++] = '+';
1771 var type = real_type_suffix (c);
1772 if (type == TypeCode.Empty && !is_real) {
1773 res = adjust_int (c, loc);
1777 if (type != TypeCode.Empty) {
1781 res = adjust_real (type, loc);
1787 var chars = reader.ReadChars (read_start, reader.Position - (type == TypeCode.Empty && c > 0 ? 1 : 0));
1788 if (chars[chars.Length - 1] == '\r')
1789 Array.Resize (ref chars, chars.Length - 1);
1790 res.ParsedValue = chars;
1793 return Token.LITERAL;
1797 // Accepts exactly count (4 or 8) hex, no more no less
1799 int getHex (int count, out int surrogate, out bool error)
1804 int top = count != -1 ? count : 4;
1809 for (i = 0; i < top; i++){
1812 if (c >= '0' && c <= '9')
1813 c = (int) c - (int) '0';
1814 else if (c >= 'A' && c <= 'F')
1815 c = (int) c - (int) 'A' + 10;
1816 else if (c >= 'a' && c <= 'f')
1817 c = (int) c - (int) 'a' + 10;
1823 total = (total * 16) + c;
1825 int p = peek_char ();
1828 if (!is_hex ((char)p))
1834 if (total > 0x0010FFFF) {
1839 if (total >= 0x00010000) {
1840 surrogate = ((total - 0x00010000) % 0x0400 + 0xDC00);
1841 total = ((total - 0x00010000) / 0x0400 + 0xD800);
1848 int escape (int c, out int surrogate)
1884 v = getHex (-1, out surrogate, out error);
1890 return EscapeUnicode (d, out surrogate);
1893 Report.Error (1009, Location, "Unrecognized escape sequence `\\{0}'", ((char)d).ToString ());
1902 int EscapeUnicode (int ch, out int surrogate)
1906 ch = getHex (8, out surrogate, out error);
1908 ch = getHex (4, out surrogate, out error);
1912 Report.Error (1009, Location, "Unrecognized escape sequence");
1920 if (putback_char != -1) {
1929 if (peek_char () == '\n') {
1935 } else if (x == '\n') {
1940 } else if (x >= UnicodeLS && x <= UnicodePS) {
1949 void advance_line ()
1959 if (putback_char == -1)
1960 putback_char = reader.Read ();
1961 return putback_char;
1966 if (putback_char != -1)
1967 return putback_char;
1968 return reader.Peek ();
1971 public void putback (int c)
1973 if (putback_char != -1) {
1974 throw new InternalErrorException (string.Format ("Secondary putback [{0}] putting back [{1}] is not allowed", (char)putback_char, (char) c), Location);
1977 if (c == '\n' || col == 0 || (c >= UnicodeLS && c <= UnicodePS)) {
1978 // It won't happen though.
1988 public bool advance ()
1990 return peek_char () != -1 || CompleteOnEOF;
1993 public Object Value {
1999 public Object value ()
2006 current_token = xtoken ();
2007 return current_token;
2010 int TokenizePreprocessorKeyword (out int c)
2012 // skip over white space
2015 } while (c == ' ' || c == '\t');
2019 while (c != -1 && c >= 'a' && c <= 'z') {
2020 id_builder[pos++] = (char) c;
2023 int peek = peek_char ();
2024 if (peek == 'U' || peek == 'u') {
2026 c = EscapeUnicode (c, out surrogate);
2027 if (surrogate != 0) {
2028 if (is_identifier_part_character ((char) c)) {
2029 id_builder[pos++] = (char) c;
2040 PreprocessorDirective get_cmd_arg (out string arg)
2044 tokens_seen = false;
2047 var cmd = GetPreprocessorDirective (id_builder, TokenizePreprocessorKeyword (out c));
2049 if ((cmd & PreprocessorDirective.CustomArgumentsParsing) != 0)
2052 // skip over white space
2053 while (c == ' ' || c == '\t')
2056 int has_identifier_argument = (int)(cmd & PreprocessorDirective.RequiresArgument);
2059 while (c != -1 && c != '\n' && c != UnicodeLS && c != UnicodePS) {
2060 if (c == '\\' && has_identifier_argument >= 0) {
2061 if (has_identifier_argument != 0) {
2062 has_identifier_argument = 1;
2064 int peek = peek_char ();
2065 if (peek == 'U' || peek == 'u') {
2067 c = EscapeUnicode (c, out surrogate);
2068 if (surrogate != 0) {
2069 if (is_identifier_part_character ((char) c)) {
2070 if (pos == value_builder.Length)
2071 Array.Resize (ref value_builder, pos * 2);
2073 value_builder[pos++] = (char) c;
2079 has_identifier_argument = -1;
2081 } else if (c == '/' && peek_char () == '/') {
2083 // Eat single-line comments
2090 if (pos == value_builder.Length)
2091 Array.Resize (ref value_builder, pos * 2);
2093 value_builder[pos++] = (char) c;
2098 if (pos > MaxIdentifierLength)
2099 arg = new string (value_builder, 0, pos);
2101 arg = InternIdentifier (value_builder, pos);
2103 // Eat any trailing whitespaces
2104 arg = arg.Trim (simple_whitespaces);
2111 // Handles the #line directive
2113 bool PreProcessLine ()
2115 Location loc = Location;
2119 int length = TokenizePreprocessorKeyword (out c);
2120 if (length == line_default.Length) {
2121 if (!IsTokenIdentifierEqual (line_default))
2124 current_source = source_file.SourceFile;
2125 if (!hidden_block_start.IsNull) {
2126 current_source.RegisterHiddenScope (hidden_block_start, loc);
2127 hidden_block_start = Location.Null;
2134 if (length == line_hidden.Length) {
2135 if (!IsTokenIdentifierEqual (line_hidden))
2138 if (hidden_block_start.IsNull)
2139 hidden_block_start = loc;
2144 if (length != 0 || c < '0' || c > '9') {
2146 // Eat any remaining characters to continue parsing on next line
2152 int new_line = TokenizeNumber (c);
2155 // Eat any remaining characters to continue parsing on next line
2158 return new_line != 0;
2163 // skip over white space
2166 } while (c == ' ' || c == '\t');
2167 } else if (c == '"') {
2171 if (c != '\n' && c != '/' && c != '"' && c != UnicodeLS && c != UnicodePS) {
2173 // Eat any remaining characters to continue parsing on next line
2177 Report.Error (1578, loc, "Filename, single-line comment or end-of-line expected");
2181 string new_file_name = null;
2183 new_file_name = TokenizeFileName (ref c);
2185 // skip over white space
2186 while (c == ' ' || c == '\t') {
2191 if (c == '\n' || c == UnicodeLS || c == UnicodePS) {
2193 } else if (c == '/') {
2194 ReadSingleLineComment ();
2197 // Eat any remaining characters to continue parsing on next line
2201 Error_EndLineExpected ();
2205 if (new_file_name != null) {
2206 current_source = context.LookupFile (source_file, new_file_name);
2207 source_file.AddIncludeFile (current_source);
2210 if (!hidden_block_start.IsNull) {
2211 current_source.RegisterHiddenScope (hidden_block_start, loc);
2212 hidden_block_start = Location.Null;
2215 ref_line = new_line;
2220 // Handles #define and #undef
2222 void PreProcessDefinition (bool is_define, string ident, bool caller_is_taking)
2224 if (ident.Length == 0 || ident == "true" || ident == "false"){
2225 Report.Error (1001, Location, "Missing identifier to pre-processor directive");
2229 if (ident.IndexOfAny (simple_whitespaces) != -1){
2230 Error_EndLineExpected ();
2234 if (!is_identifier_start_character (ident [0]))
2235 Report.Error (1001, Location, "Identifier expected: {0}", ident);
2237 foreach (char c in ident.Substring (1)){
2238 if (!is_identifier_part_character (c)){
2239 Report.Error (1001, Location, "Identifier expected: {0}", ident);
2244 if (!caller_is_taking)
2251 if (context.Settings.IsConditionalSymbolDefined (ident))
2254 source_file.AddDefine (ident);
2259 source_file.AddUndefine (ident);
2263 byte read_hex (out bool error)
2266 int c = get_char ();
2268 if ((c >= '0') && (c <= '9'))
2269 total = (int) c - (int) '0';
2270 else if ((c >= 'A') && (c <= 'F'))
2271 total = (int) c - (int) 'A' + 10;
2272 else if ((c >= 'a') && (c <= 'f'))
2273 total = (int) c - (int) 'a' + 10;
2282 if ((c >= '0') && (c <= '9'))
2283 total += (int) c - (int) '0';
2284 else if ((c >= 'A') && (c <= 'F'))
2285 total += (int) c - (int) 'A' + 10;
2286 else if ((c >= 'a') && (c <= 'f'))
2287 total += (int) c - (int) 'a' + 10;
2294 return (byte) total;
2298 // Parses #pragma checksum
2300 bool ParsePragmaChecksum ()
2303 // The syntax is ` "foo.txt" "{guid}" "hash"'
2305 // guid is predefined hash algorithm guid {406ea660-64cf-4c82-b6f0-42d48172a799} for md5
2307 int c = get_char ();
2312 string file_name = TokenizeFileName (ref c);
2314 // TODO: Any white-spaces count
2318 SourceFile file = context.LookupFile (source_file, file_name);
2320 if (get_char () != '"' || get_char () != '{')
2324 byte[] guid_bytes = new byte [16];
2327 for (; i < 4; i++) {
2328 guid_bytes [i] = read_hex (out error);
2333 if (get_char () != '-')
2336 for (; i < 10; i++) {
2337 guid_bytes [i] = read_hex (out error);
2341 guid_bytes [i++] = read_hex (out error);
2345 if (get_char () != '-')
2349 for (; i < 16; i++) {
2350 guid_bytes [i] = read_hex (out error);
2355 if (get_char () != '}' || get_char () != '"')
2358 // TODO: Any white-spaces count
2363 if (get_char () != '"')
2366 // Any length of checksum
2367 List<byte> checksum_bytes = new List<byte> (16);
2369 var checksum_location = Location;
2371 while (c != '"' && c != -1) {
2372 checksum_bytes.Add (read_hex (out error));
2380 ReadSingleLineComment ();
2381 } else if (get_char () != '"') {
2385 if (context.Settings.GenerateDebugInfo) {
2386 var chsum = checksum_bytes.ToArray ();
2388 if (file.HasChecksum) {
2389 if (!ArrayComparer.IsEqual (file.Checksum, chsum)) {
2390 // TODO: Report.SymbolRelatedToPreviousError
2391 Report.Warning (1697, 1, checksum_location, "Different checksum values specified for file `{0}'", file.Name);
2395 file.SetChecksum (guid_bytes, chsum);
2396 current_source.AutoGenerated = true;
2402 bool IsTokenIdentifierEqual (char[] identifier)
2404 for (int i = 0; i < identifier.Length; ++i) {
2405 if (identifier[i] != id_builder[i])
2412 bool ScanClosingInterpolationBrace ()
2419 var c = reader.Read ();
2428 if (str_quote % 2 == 1) {
2434 } while (res == null);
2440 int TokenizeNumber (int value)
2444 decimal_digits (value);
2445 uint ui = (uint) (number_builder[0] - '0');
2448 for (int i = 1; i < number_pos; i++) {
2449 ui = checked ((ui * 10) + ((uint) (number_builder[i] - '0')));
2453 } catch (OverflowException) {
2454 Error_NumericConstantTooLong ();
2459 string TokenizeFileName (ref int c)
2461 var string_builder = new StringBuilder ();
2462 while (c != -1 && c != '\n' && c != UnicodeLS && c != UnicodePS) {
2469 string_builder.Append ((char) c);
2472 if (string_builder.Length == 0) {
2473 Report.Warning (1709, 1, Location, "Filename specified for preprocessor directive is empty");
2477 return string_builder.ToString ();
2480 int TokenizePragmaWarningIdentifier (ref int c, ref bool identifier)
2483 if ((c >= '0' && c <= '9') || is_identifier_start_character (c)) {
2486 if (c >= '0' && c <= '9') {
2488 number = TokenizeNumber (c);
2492 if (c != ' ' && c != '\t' && c != ',' && c != '\n' && c != -1 && c != UnicodeLS && c != UnicodePS) {
2493 return ReadPragmaWarningComment (c);
2497 // LAMESPEC v6: No spec what identifier really is in this context, it seems keywords are allowed too
2501 id_builder [pos++] = (char)c;
2502 while (c < MaxIdentifierLength) {
2504 id_builder [pos] = (char)c;
2506 if (c >= '0' && c <= '9') {
2507 if (pos == 6 && id_builder [0] == 'C' && id_builder [1] == 'S') {
2508 // Recognize CSXXXX as C# XXXX warning
2511 for (int i = 0; i < 4; ++i) {
2512 var ch = id_builder [i + 2];
2513 if (ch < '0' || ch > '9') {
2518 number += (ch - '0') * pow;
2522 } else if ((c < 'a' || c > 'z') && (c < 'A' || c > 'Z') && c != '_') {
2535 // skip over white space
2536 while (c == ' ' || c == '\t')
2543 // skip over white space
2544 while (c == ' ' || c == '\t')
2550 return ReadPragmaWarningComment (c);
2553 int ReadPragmaWarningComment (int c)
2556 ReadSingleLineComment ();
2558 Report.Warning (1692, 1, Location, "Invalid number");
2560 // Read everything till the end of the line or file
2567 void ReadToEndOfLine ()
2572 } while (c != -1 && c != '\n' && c != UnicodeLS && c != UnicodePS);
2575 void ReadSingleLineComment ()
2577 if (peek_char () != '/')
2578 Report.Warning (1696, 1, Location, "Single-line comment or end-of-line expected");
2580 // Read everything till the end of the line or file
2585 /// Handles #pragma directive
2587 void ParsePragmaDirective ()
2590 int length = TokenizePreprocessorKeyword (out c);
2591 if (length == pragma_warning.Length && IsTokenIdentifierEqual (pragma_warning)) {
2592 length = TokenizePreprocessorKeyword (out c);
2595 // #pragma warning disable
2596 // #pragma warning restore
2598 if (length == pragma_warning_disable.Length) {
2599 bool disable = IsTokenIdentifierEqual (pragma_warning_disable);
2600 if (disable || IsTokenIdentifierEqual (pragma_warning_restore)) {
2601 // skip over white space
2602 while (c == ' ' || c == '\t')
2607 if (c == '\n' || c == '/' || c == UnicodeLS || c == UnicodePS) {
2609 ReadSingleLineComment ();
2612 // Disable/Restore all warnings
2615 Report.RegisterWarningRegion (loc).WarningDisable (loc.Row);
2617 Report.RegisterWarningRegion (loc).WarningEnable (loc.Row);
2621 // Disable/Restore a warning or group of warnings
2625 bool identifier = false;
2626 code = TokenizePragmaWarningIdentifier (ref c, ref identifier);
2629 // no-op, custom warnings cannot occur in mcs
2630 } else if (disable) {
2631 Report.RegisterWarningRegion (loc).WarningDisable (loc, code, context.Report);
2633 Report.RegisterWarningRegion (loc).WarningEnable (loc, code, context);
2636 } while (code >= 0 && c != '\n' && c != -1 && c != UnicodeLS && c != UnicodePS);
2643 Report.Warning (1634, 1, Location, "Expected disable or restore");
2645 // Eat any remaining characters on the line
2654 if (length == pragma_checksum.Length && IsTokenIdentifierEqual (pragma_checksum)) {
2655 if (c != ' ' || !ParsePragmaChecksum ()) {
2656 Report.Warning (1695, 1, Location,
2657 "Invalid #pragma checksum syntax. Expected \"filename\" \"{XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX}\" \"XXXX...\"");
2663 Report.Warning (1633, 1, Location, "Unrecognized #pragma directive");
2665 // Eat any remaining characters on the line
2669 bool eval_val (string s)
2676 return source_file.IsConditionalDefined (s);
2679 bool pp_primary (ref string s)
2688 s = s.Substring (1);
2689 bool val = pp_expr (ref s, false);
2690 if (s.Length > 0 && s [0] == ')'){
2691 s = s.Substring (1);
2694 Error_InvalidDirective ();
2698 if (is_identifier_start_character (c)){
2704 if (is_identifier_part_character (c)){
2708 bool v = eval_val (s.Substring (0, j));
2709 s = s.Substring (j);
2712 bool vv = eval_val (s);
2717 Error_InvalidDirective ();
2721 bool pp_unary (ref string s)
2728 if (len > 1 && s [1] == '='){
2729 Error_InvalidDirective ();
2732 s = s.Substring (1);
2733 return ! pp_primary (ref s);
2735 return pp_primary (ref s);
2737 Error_InvalidDirective ();
2742 bool pp_eq (ref string s)
2744 bool va = pp_unary (ref s);
2750 if (len > 2 && s [1] == '='){
2751 s = s.Substring (2);
2752 return va == pp_unary (ref s);
2754 Error_InvalidDirective ();
2757 } else if (s [0] == '!' && len > 1 && s [1] == '='){
2758 s = s.Substring (2);
2760 return va != pp_unary (ref s);
2769 bool pp_and (ref string s)
2771 bool va = pp_eq (ref s);
2777 if (len > 2 && s [1] == '&'){
2778 s = s.Substring (2);
2779 return (va & pp_and (ref s));
2781 Error_InvalidDirective ();
2790 // Evaluates an expression for `#if' or `#elif'
2792 bool pp_expr (ref string s, bool isTerm)
2794 bool va = pp_and (ref s);
2801 if (len > 2 && s [1] == '|'){
2802 s = s.Substring (2);
2803 return va | pp_expr (ref s, isTerm);
2805 Error_InvalidDirective ();
2810 Error_EndLineExpected ();
2818 bool eval (string s)
2820 bool v = pp_expr (ref s, true);
2829 void Error_NumericConstantTooLong ()
2831 Report.Error (1021, Location, "Integral constant is too large");
2834 void Error_InvalidDirective ()
2836 Report.Error (1517, Location, "Invalid preprocessor directive");
2839 void Error_UnexpectedDirective (string extra)
2843 "Unexpected processor directive ({0})", extra);
2846 void Error_TokensSeen ()
2848 Report.Error (1032, Location,
2849 "Cannot define or undefine preprocessor symbols after first token in file");
2852 void Eror_WrongPreprocessorLocation ()
2854 Report.Error (1040, Location,
2855 "Preprocessor directives must appear as the first non-whitespace character on a line");
2858 void Error_EndLineExpected ()
2860 Report.Error (1025, Location, "Single-line comment or end-of-line expected");
2864 // Raises a warning when tokenizer found documentation comment
2865 // on unexpected place
2867 void WarningMisplacedComment (Location loc)
2869 if (doc_state != XmlCommentState.Error) {
2870 doc_state = XmlCommentState.Error;
2871 Report.Warning (1587, 2, loc, "XML comment is not placed on a valid language element");
2876 // if true, then the code continues processing the code
2877 // if false, the code stays in a loop until another directive is
2879 // When caller_is_taking is false we ignore all directives except the ones
2880 // which can help us to identify where the #if block ends
2881 bool ParsePreprocessingDirective (bool caller_is_taking)
2884 bool region_directive = false;
2886 var directive = get_cmd_arg (out arg);
2889 // The first group of pre-processing instructions is always processed
2891 switch (directive) {
2892 case PreprocessorDirective.Region:
2893 region_directive = true;
2895 goto case PreprocessorDirective.If;
2897 case PreprocessorDirective.Endregion:
2898 if (ifstack == null || ifstack.Count == 0){
2899 Error_UnexpectedDirective ("no #region for this #endregion");
2902 int pop = ifstack.Pop ();
2904 if ((pop & REGION) == 0)
2905 Report.Error (1027, Location, "Expected `#endif' directive");
2907 return caller_is_taking;
2909 case PreprocessorDirective.If:
2910 if (ifstack == null)
2911 ifstack = new Stack<int> (2);
2913 int flags = region_directive ? REGION : 0;
2914 if (ifstack.Count == 0){
2915 flags |= PARENT_TAKING;
2917 int state = ifstack.Peek ();
2918 if ((state & TAKING) != 0) {
2919 flags |= PARENT_TAKING;
2923 if (eval (arg) && caller_is_taking) {
2924 ifstack.Push (flags | TAKING);
2927 ifstack.Push (flags);
2930 case PreprocessorDirective.Endif:
2931 if (ifstack == null || ifstack.Count == 0){
2932 Error_UnexpectedDirective ("no #if for this #endif");
2935 pop = ifstack.Pop ();
2937 if ((pop & REGION) != 0)
2938 Report.Error (1038, Location, "#endregion directive expected");
2940 if (arg.Length != 0) {
2941 Error_EndLineExpected ();
2944 if (ifstack.Count == 0)
2947 int state = ifstack.Peek ();
2948 return (state & TAKING) != 0;
2951 case PreprocessorDirective.Elif:
2952 if (ifstack == null || ifstack.Count == 0){
2953 Error_UnexpectedDirective ("no #if for this #elif");
2956 int state = ifstack.Pop ();
2958 if ((state & REGION) != 0) {
2959 Report.Error (1038, Location, "#endregion directive expected");
2963 if ((state & ELSE_SEEN) != 0){
2964 Error_UnexpectedDirective ("#elif not valid after #else");
2968 if ((state & TAKING) != 0) {
2973 if (eval (arg) && ((state & PARENT_TAKING) != 0)){
2974 ifstack.Push (state | TAKING);
2978 ifstack.Push (state);
2982 case PreprocessorDirective.Else:
2983 if (ifstack == null || ifstack.Count == 0){
2984 Error_UnexpectedDirective ("no #if for this #else");
2987 int state = ifstack.Peek ();
2989 if ((state & REGION) != 0) {
2990 Report.Error (1038, Location, "#endregion directive expected");
2994 if ((state & ELSE_SEEN) != 0){
2995 Error_UnexpectedDirective ("#else within #else");
3001 if (arg.Length != 0) {
3002 Error_EndLineExpected ();
3007 if ((state & PARENT_TAKING) != 0) {
3008 ret = (state & TAKING) == 0;
3016 ifstack.Push (state | ELSE_SEEN);
3020 case PreprocessorDirective.Define:
3021 if (any_token_seen){
3022 if (caller_is_taking)
3023 Error_TokensSeen ();
3024 return caller_is_taking;
3026 PreProcessDefinition (true, arg, caller_is_taking);
3027 return caller_is_taking;
3029 case PreprocessorDirective.Undef:
3030 if (any_token_seen){
3031 if (caller_is_taking)
3032 Error_TokensSeen ();
3033 return caller_is_taking;
3035 PreProcessDefinition (false, arg, caller_is_taking);
3036 return caller_is_taking;
3038 case PreprocessorDirective.Invalid:
3039 Report.Error (1024, Location, "Wrong preprocessor directive");
3044 // These are only processed if we are in a `taking' block
3046 if (!caller_is_taking)
3050 case PreprocessorDirective.Error:
3051 Report.Error (1029, Location, "#error: '{0}'", arg);
3054 case PreprocessorDirective.Warning:
3055 Report.Warning (1030, 1, Location, "#warning: `{0}'", arg);
3058 case PreprocessorDirective.Pragma:
3059 if (context.Settings.Version == LanguageVersion.ISO_1) {
3060 Report.FeatureIsNotAvailable (context, Location, "#pragma");
3063 ParsePragmaDirective ();
3066 case PreprocessorDirective.Line:
3067 Location loc = Location;
3068 if (!PreProcessLine ())
3069 Report.Error (1576, loc, "The line number specified for #line directive is missing or invalid");
3071 return caller_is_taking;
3074 throw new NotImplementedException (directive.ToString ());
3077 int consume_string (bool quoted)
3081 Location start_location = Location;
3083 start_location = start_location - 1;
3086 int reader_pos = reader.Position;
3090 // Cannot use get_char because of \r in quoted strings
3091 if (putback_char != -1) {
3101 if (quoted && peek_char () == '"') {
3102 if (pos == value_builder.Length)
3103 Array.Resize (ref value_builder, pos * 2);
3105 value_builder[pos++] = (char) c;
3110 ILiteralConstant res = new StringLiteral (context.BuiltinTypes, CreateStringFromBuilder (pos), start_location);
3113 res.ParsedValue = quoted ?
3114 reader.ReadChars (reader_pos - 2, reader.Position - 1) :
3115 reader.ReadChars (reader_pos - 1, reader.Position);
3118 return Token.LITERAL;
3121 if (c == '\n' || c == UnicodeLS || c == UnicodePS) {
3123 Report.Error (1010, Location, "Newline in constant");
3127 // Don't add \r to string literal
3128 if (pos > 1 && value_builder [pos - 1] == '\r')
3131 val = new StringLiteral (context.BuiltinTypes, new string (value_builder, 0, pos), start_location);
3132 return Token.LITERAL;
3136 } else if (c == '\\' && !quoted) {
3139 c = escape (c, out surrogate);
3142 if (surrogate != 0) {
3143 if (pos == value_builder.Length)
3144 Array.Resize (ref value_builder, pos * 2);
3146 value_builder[pos++] = (char) c;
3149 } else if (c == -1) {
3150 Report.Error (1039, Location, "Unterminated string literal");
3156 if (pos == value_builder.Length)
3157 Array.Resize (ref value_builder, pos * 2);
3159 value_builder[pos++] = (char) c;
3163 private int consume_identifier (int s)
3165 int res = consume_identifier (s, false);
3167 if (doc_state == XmlCommentState.Allowed)
3168 doc_state = XmlCommentState.NotAllowed;
3173 int consume_identifier (int c, bool quoted)
3176 // This method is very performance sensitive. It accounts
3177 // for approximately 25% of all parser time
3187 c = escape (c, out surrogate);
3188 if (quoted || is_identifier_start_character (c)) {
3189 // it's added bellow
3190 } else if (surrogate != 0) {
3191 id_builder [pos++] = (char)c;
3194 Report.Error (1056, Location, "Unexpected character `\\{0}'", c.ToString ("x4"));
3199 id_builder [pos++] = (char) c;
3205 if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || (c >= '0' && c <= '9')) {
3206 id_builder [pos++] = (char) c;
3213 c = escape (c, out surrogate);
3214 if (is_identifier_part_character ((char) c))
3215 id_builder[pos++] = (char) c;
3216 else if (surrogate != 0) {
3220 // TODO: Probably need more whitespace characters
3225 Report.Error (1056, Location, "Unexpected character `\\{0}'", c.ToString ("x4"));
3232 } else if (is_identifier_part_character_slow_part ((char) c)) {
3233 id_builder [pos++] = (char) c;
3240 } catch (IndexOutOfRangeException) {
3241 Report.Error (645, Location, "Identifier too long (limit is 512 chars)");
3249 // Optimization: avoids doing the keyword lookup
3250 // on uppercase letters
3252 if (id_builder [0] >= '_' && !quoted) {
3253 int keyword = GetKeyword (id_builder, pos);
3254 if (keyword != -1) {
3255 val = ltb.Create (keyword == Token.AWAIT ? "await" : null, current_source, ref_line, column);
3260 string s = InternIdentifier (id_builder, pos);
3261 val = ltb.Create (s, current_source, ref_line, column);
3262 if (quoted && parsing_attribute_section)
3263 AddEscapedIdentifier (((LocatedToken) val).Location);
3265 return Token.IDENTIFIER;
3268 string InternIdentifier (char[] charBuffer, int length)
3271 // Keep identifiers in an array of hashtables to avoid needless
3274 var identifiers_group = identifiers[length];
3276 if (identifiers_group != null) {
3277 if (identifiers_group.TryGetValue (charBuffer, out s)) {
3281 // TODO: this should be number of files dependant
3282 // corlib compilation peaks at 1000 and System.Core at 150
3283 int capacity = length > 20 ? 10 : 100;
3284 identifiers_group = new Dictionary<char[], string> (capacity, new IdentifiersComparer (length));
3285 identifiers[length] = identifiers_group;
3288 char[] chars = new char[length];
3289 Array.Copy (charBuffer, chars, length);
3291 s = new string (charBuffer, 0, length);
3292 identifiers_group.Add (chars, s);
3296 public int xtoken ()
3298 if (parsing_interpolation_format) {
3299 return TokenizeInterpolationFormat ();
3304 // Whether we have seen comments on the current line
3305 bool comments_seen = false;
3306 while ((c = get_char ()) != -1) {
3309 col = ((col - 1 + tab_size) / tab_size) * tab_size;
3317 case 0xFEFF: // Ignore BOM anywhere in the file
3320 /* This is required for compatibility with .NET
3322 if (peek_char () == 0xBB) {
3325 if (get_char () == 0xBF)
3333 return consume_identifier (c);
3336 val = ltb.Create (current_source, ref_line, col);
3337 return Token.OPEN_BRACE;
3339 if (parsing_string_interpolation > 0) {
3340 if (peek_char () != '}') {
3341 --parsing_string_interpolation;
3342 return TokenizeInterpolatedString ();
3348 val = ltb.Create (current_source, ref_line, col);
3349 return Token.CLOSE_BRACE;
3351 // To block doccomment inside attribute declaration.
3352 if (doc_state == XmlCommentState.Allowed)
3353 doc_state = XmlCommentState.NotAllowed;
3355 val = ltb.Create (current_source, ref_line, col);
3357 if (parsing_block == 0 || lambda_arguments_parsing)
3358 return Token.OPEN_BRACKET;
3360 int next = peek_char ();
3364 return Token.OPEN_BRACKET;
3374 next = peek_token ();
3375 if (next == Token.COMMA || next == Token.CLOSE_BRACKET)
3376 return Token.OPEN_BRACKET;
3378 return Token.OPEN_BRACKET_EXPR;
3380 return Token.OPEN_BRACKET_EXPR;
3383 ltb.CreateOptional (current_source, ref_line, col, ref val);
3384 return Token.CLOSE_BRACKET;
3386 val = ltb.Create (current_source, ref_line, col);
3388 // An expression versions of parens can appear in block context only
3390 if (parsing_block != 0 && !lambda_arguments_parsing) {
3393 // Optmize most common case where we know that parens
3396 switch (current_token) {
3397 case Token.IDENTIFIER:
3406 case Token.DELEGATE:
3407 case Token.OP_GENERICS_GT:
3408 return Token.OPEN_PARENS;
3411 // Optimize using peek
3412 int xx = peek_char ();
3419 return Token.OPEN_PARENS;
3422 lambda_arguments_parsing = true;
3424 d = TokenizeOpenParens ();
3426 lambda_arguments_parsing = false;
3430 return Token.OPEN_PARENS;
3432 ltb.CreateOptional (current_source, ref_line, col, ref val);
3433 return Token.CLOSE_PARENS;
3435 ltb.CreateOptional (current_source, ref_line, col, ref val);
3438 ltb.CreateOptional (current_source, ref_line, col, ref val);
3439 return Token.SEMICOLON;
3441 val = ltb.Create (current_source, ref_line, col);
3444 val = ltb.Create (current_source, ref_line, col);
3445 return TokenizePossibleNullableType ();
3447 val = ltb.Create (current_source, ref_line, col);
3448 if (parsing_generic_less_than++ > 0)
3449 return Token.OP_GENERICS_LT;
3451 return TokenizeLessThan ();
3454 val = ltb.Create (current_source, ref_line, col);
3462 if (parsing_generic_less_than > 1 || (parsing_generic_less_than == 1 && d != '>')) {
3463 parsing_generic_less_than--;
3464 return Token.OP_GENERICS_GT;
3473 return Token.OP_SHIFT_RIGHT_ASSIGN;
3475 return Token.OP_SHIFT_RIGHT;
3481 val = ltb.Create (current_source, ref_line, col);
3485 } else if (d == '=') {
3486 d = Token.OP_ADD_ASSIGN;
3494 val = ltb.Create (current_source, ref_line, col);
3498 } else if (d == '=')
3499 d = Token.OP_SUB_ASSIGN;
3509 val = ltb.Create (current_source, ref_line, col);
3510 if (peek_char () == '='){
3517 val = ltb.Create (current_source, ref_line, col);
3528 return Token.ASSIGN;
3531 val = ltb.Create (current_source, ref_line, col);
3535 return Token.OP_AND;
3539 return Token.OP_AND_ASSIGN;
3541 return Token.BITWISE_AND;
3544 val = ltb.Create (current_source, ref_line, col);
3552 return Token.OP_OR_ASSIGN;
3554 return Token.BITWISE_OR;
3557 val = ltb.Create (current_source, ref_line, col);
3558 if (peek_char () == '='){
3560 return Token.OP_MULT_ASSIGN;
3567 val = ltb.Create (current_source, ref_line, col);
3569 return Token.OP_DIV_ASSIGN;
3572 // Handle double-slash comments.
3574 if (parsing_string_interpolation > 0) {
3575 Report.Error (8077, Location, "A single-line comment may not be used in an interpolated string");
3580 if (doc_processing) {
3581 if (peek_char () == '/') {
3583 // Don't allow ////.
3584 if ((d = peek_char ()) != '/') {
3585 if (doc_state == XmlCommentState.Allowed)
3586 handle_one_line_xml_comment ();
3587 else if (doc_state == XmlCommentState.NotAllowed)
3588 WarningMisplacedComment (Location - 3);
3591 if (xml_comment_buffer.Length > 0)
3592 doc_state = XmlCommentState.NotAllowed;
3598 any_token_seen |= tokens_seen;
3599 tokens_seen = false;
3600 comments_seen = false;
3602 } else if (d == '*'){
3604 bool docAppend = false;
3605 if (doc_processing && peek_char () == '*') {
3607 // But when it is /**/, just do nothing.
3608 if (peek_char () == '/') {
3612 if (doc_state == XmlCommentState.Allowed)
3614 else if (doc_state == XmlCommentState.NotAllowed) {
3615 WarningMisplacedComment (Location - 2);
3619 int current_comment_start = 0;
3621 current_comment_start = xml_comment_buffer.Length;
3622 xml_comment_buffer.Append (Environment.NewLine);
3625 while ((d = get_char ()) != -1){
3626 if (d == '*' && peek_char () == '/'){
3628 comments_seen = true;
3632 xml_comment_buffer.Append ((char) d);
3634 if (d == '\n' || d == UnicodeLS || d == UnicodePS){
3635 any_token_seen |= tokens_seen;
3636 tokens_seen = false;
3638 // Reset 'comments_seen' just to be consistent.
3639 // It doesn't matter either way, here.
3641 comments_seen = false;
3645 Report.Error (1035, Location, "End-of-file found, '*/' expected");
3648 update_formatted_doc_comment (current_comment_start);
3651 val = ltb.Create (current_source, ref_line, col);
3655 val = ltb.Create (current_source, ref_line, col);
3656 if (peek_char () == '='){
3658 return Token.OP_MOD_ASSIGN;
3660 return Token.PERCENT;
3663 val = ltb.Create (current_source, ref_line, col);
3664 if (peek_char () == '='){
3666 return Token.OP_XOR_ASSIGN;
3668 return Token.CARRET;
3671 val = ltb.Create (current_source, ref_line, col);
3672 if (peek_char () == ':') {
3674 return Token.DOUBLE_COLON;
3678 case '0': case '1': case '2': case '3': case '4':
3679 case '5': case '6': case '7': case '8': case '9':
3681 return is_number (c, false);
3683 case '\n': // white space
3686 any_token_seen |= tokens_seen;
3687 tokens_seen = false;
3688 comments_seen = false;
3694 if (d >= '0' && d <= '9')
3695 return is_number (c, true);
3697 ltb.CreateOptional (current_source, ref_line, col, ref val);
3701 if (tokens_seen || comments_seen) {
3702 Eror_WrongPreprocessorLocation ();
3706 if (ParsePreprocessingDirective (true))
3709 bool directive_expected = false;
3710 while ((c = get_char ()) != -1) {
3712 directive_expected = true;
3713 } else if (!directive_expected) {
3714 // TODO: Implement comment support for disabled code and uncomment this code
3716 // Eror_WrongPreprocessorLocation ();
3717 // return Token.ERROR;
3722 if (c == ' ' || c == '\t' || c == '\n' || c == '\f' || c == '\v' || c == UnicodeLS || c == UnicodePS)
3726 if (ParsePreprocessingDirective (false))
3729 directive_expected = false;
3733 tokens_seen = false;
3740 if (parsing_string_interpolation > 0 && !ScanClosingInterpolationBrace ()) {
3741 parsing_string_interpolation = 0;
3742 Report.Error (8076, Location, "Missing close delimiter `}' for interpolated expression");
3743 val = new StringLiteral (context.BuiltinTypes, "", Location);
3744 return Token.INTERPOLATED_STRING_END;
3747 return consume_string (false);
3750 return TokenizeBackslash ();
3756 return consume_string (true);
3759 if (is_identifier_start_character (c)){
3760 return consume_identifier (c, true);
3763 Report.Error (1646, Location, "Keyword, identifier, or string expected after verbatim specifier: @");
3767 if (peek_char () == '"') {
3769 return TokenizeInterpolatedString ();
3773 case EvalStatementParserCharacter:
3774 return Token.EVAL_STATEMENT_PARSER;
3775 case EvalCompilationUnitParserCharacter:
3776 return Token.EVAL_COMPILATION_UNIT_PARSER;
3777 case EvalUsingDeclarationsParserCharacter:
3778 return Token.EVAL_USING_DECLARATIONS_UNIT_PARSER;
3779 case DocumentationXref:
3780 return Token.DOC_SEE;
3783 if (is_identifier_start_character (c)) {
3785 return consume_identifier (c);
3788 if (char.IsWhiteSpace ((char) c))
3791 Report.Error (1056, Location, "Unexpected character `{0}'", ((char) c).ToString ());
3796 return Token.COMPLETE_COMPLETION;
3799 return Token.GENERATE_COMPLETION;
3806 int TokenizeBackslash ()
3809 int read_start = reader.Position;
3811 Location start_location = Location;
3812 int c = get_char ();
3815 val = new CharLiteral (context.BuiltinTypes, (char) c, start_location);
3816 Report.Error (1011, start_location, "Empty character literal");
3817 return Token.LITERAL;
3820 if (c == '\n' || c == UnicodeLS || c == UnicodePS) {
3821 Report.Error (1010, start_location, "Newline in constant");
3826 c = escape (c, out d);
3830 throw new NotImplementedException ();
3832 ILiteralConstant res = new CharLiteral (context.BuiltinTypes, (char) c, start_location);
3837 Report.Error (1012, start_location, "Too many characters in character literal");
3839 // Try to recover, read until newline or next "'"
3840 while ((c = get_char ()) != -1) {
3841 if (c == '\n' || c == '\'' || c == UnicodeLS || c == UnicodePS)
3847 res.ParsedValue = reader.ReadChars (read_start - 1, reader.Position);
3850 return Token.LITERAL;
3853 int TokenizeLessThan ()
3857 // Save current position and parse next token.
3859 int generic_dimension = 0;
3860 if (parse_less_than (ref generic_dimension)) {
3861 if (parsing_generic_declaration && (parsing_generic_declaration_doc || token () != Token.DOT)) {
3862 d = Token.OP_GENERICS_LT_DECL;
3864 if (generic_dimension > 0) {
3865 val = generic_dimension;
3867 return Token.GENERIC_DIMENSION;
3870 d = Token.OP_GENERICS_LT;
3877 parsing_generic_less_than = 0;
3886 return Token.OP_SHIFT_LEFT_ASSIGN;
3888 return Token.OP_SHIFT_LEFT;
3898 int TokenizeInterpolatedString ()
3901 var start_location = Location;
3904 var ch = get_char ();
3907 val = new StringLiteral (context.BuiltinTypes, CreateStringFromBuilder (pos), start_location);
3908 return Token.INTERPOLATED_STRING_END;
3910 if (peek_char () == '{') {
3911 value_builder [pos++] = (char)ch;
3916 ++parsing_string_interpolation;
3917 val = new StringLiteral (context.BuiltinTypes, CreateStringFromBuilder (pos), start_location);
3918 return Token.INTERPOLATED_STRING;
3922 ch = escape (ch, out surrogate);
3926 if (ch == '{' || ch == '}') {
3927 Report.Error (8087, Location, "A `{0}' character may only be escaped by doubling `{0}{0}' in an interpolated string", ((char) ch).ToString ());
3930 if (surrogate != 0) {
3931 if (pos == value_builder.Length)
3932 Array.Resize (ref value_builder, pos * 2);
3934 value_builder [pos++] = (char)ch;
3944 value_builder[pos++] = (char) ch;
3948 int TokenizeInterpolationFormat ()
3953 var ch = get_char ();
3962 Report.Error (8089, Location, "Empty interpolated expression format specifier");
3963 } else if (Array.IndexOf (simple_whitespaces, value_builder [pos - 1]) >= 0) {
3964 Report.Error (8088, Location, "A interpolated expression format specifier may not contain trailing whitespace");
3967 val = CreateStringFromBuilder (pos);
3968 return Token.LITERAL;
3976 ch = escape (ch, out surrogate);
3980 if (ch == '{' || ch == '}') {
3981 Report.Error (8087, Location, "A `{0}' character may only be escaped by doubling `{0}{0}' in an interpolated string", ((char) ch).ToString ());
3984 if (surrogate != 0) {
3985 if (pos == value_builder.Length)
3986 Array.Resize (ref value_builder, pos * 2);
3988 value_builder [pos++] = (char)ch;
3998 value_builder[pos++] = (char) ch;
4002 string CreateStringFromBuilder (int pos)
4005 return string.Empty;
4007 return InternIdentifier (value_builder, pos);
4009 return new string (value_builder, 0, pos);
4013 // Handles one line xml comment
4015 private void handle_one_line_xml_comment ()
4018 while ((c = peek_char ()) == ' ')
4019 get_char (); // skip heading whitespaces.
4020 while ((c = peek_char ()) != -1 && c != '\n' && c != '\r') {
4021 xml_comment_buffer.Append ((char) get_char ());
4023 if (c == '\r' || c == '\n')
4024 xml_comment_buffer.Append (Environment.NewLine);
4028 // Remove heading "*" in Javadoc-like xml documentation.
4030 private void update_formatted_doc_comment (int current_comment_start)
4032 int length = xml_comment_buffer.Length - current_comment_start;
4033 string [] lines = xml_comment_buffer.ToString (
4034 current_comment_start,
4035 length).Replace ("\r", "").Split ('\n');
4037 // The first line starts with /**, thus it is not target
4038 // for the format check.
4039 for (int i = 1; i < lines.Length; i++) {
4040 string s = lines [i];
4041 int idx = s.IndexOf ('*');
4044 if (i < lines.Length - 1)
4048 head = s.Substring (0, idx);
4049 foreach (char c in head)
4052 lines [i] = s.Substring (idx + 1);
4054 xml_comment_buffer.Remove (current_comment_start, length);
4055 xml_comment_buffer.Insert (current_comment_start, String.Join (Environment.NewLine, lines));
4059 // Checks if there was incorrect doc comments and raise
4062 public void check_incorrect_doc_comment ()
4064 if (xml_comment_buffer.Length > 0)
4065 WarningMisplacedComment (Location);
4069 // Consumes the saved xml comment lines (if any)
4070 // as for current target member or type.
4072 public string consume_doc_comment ()
4074 if (xml_comment_buffer.Length > 0) {
4075 string ret = xml_comment_buffer.ToString ();
4076 reset_doc_comment ();
4082 void reset_doc_comment ()
4084 xml_comment_buffer.Length = 0;
4087 public void cleanup ()
4089 if (ifstack != null && ifstack.Count >= 1) {
4090 int state = ifstack.Pop ();
4091 if ((state & REGION) != 0)
4092 Report.Error (1038, Location, "#endregion directive expected");
4094 Report.Error (1027, Location, "Expected `#endif' directive");
4100 // Indicates whether it accepts XML documentation or not.
4102 public enum XmlCommentState {
4103 // comment is allowed in this state.
4105 // comment is not allowed in this state.
4107 // once comments appeared when it is NotAllowed, then the
4108 // state is changed to it, until the state is changed to
projects / mono.git / blob