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@seznam.cz)
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
17 using System.Collections;
19 using System.Globalization;
20 using System.Reflection;
25 /// Tokenizer for C# source code.
28 public class Tokenizer : yyParser.yyInput
30 SeekableStreamReader reader;
32 CompilationUnit file_name;
39 bool handle_get_set = false;
40 bool handle_remove_add = false;
41 bool handle_assembly = false;
42 bool handle_where = false;
43 bool handle_typeof = false;
44 bool lambda_arguments_parsing;
45 Location current_location;
46 Location current_comment_location = Location.Null;
47 ArrayList escaped_identifiers = new ArrayList ();
48 public int parsing_block;
49 internal int query_parsing;
52 // XML documentation buffer. The save point is used to divide
53 // comments on types and comments on members.
55 StringBuilder xml_comment_buffer;
58 // See comment on XmlCommentState enumeration.
60 XmlCommentState xml_doc_state = XmlCommentState.Allowed;
63 // Whether tokens have been seen on this line
65 bool tokens_seen = false;
68 // Whether a token has been seen on the file
69 // This is needed because `define' is not allowed to be used
70 // after a token has been seen.
72 bool any_token_seen = false;
74 static Hashtable token_values;
75 static readonly char[] simple_whitespaces = new char[] { ' ', '\t' };
77 private static Hashtable TokenValueName
80 if (token_values == null)
81 token_values = GetTokenValueNameHash ();
87 private static Hashtable GetTokenValueNameHash ()
89 Type t = typeof (Token);
90 FieldInfo [] fields = t.GetFields ();
91 Hashtable hash = new Hashtable ();
92 foreach (FieldInfo field in fields) {
93 if (field.IsLiteral && field.IsStatic && field.FieldType == typeof (int))
94 hash.Add (field.GetValue (null), field.Name);
100 // Returns a verbose representation of the current location
102 public string location {
106 if (current_token == Token.ERROR)
107 det = "detail: " + error_details;
111 // return "Line: "+line+" Col: "+col + "\n" +
112 // "VirtLine: "+ref_line +
113 // " Token: "+current_token + " " + det;
114 string current_token_name = TokenValueName [current_token] as string;
115 if (current_token_name == null)
116 current_token_name = current_token.ToString ();
118 return String.Format ("{0} ({1},{2}), Token: {3} {4}", ref_name.Name,
126 public bool PropertyParsing {
127 get { return handle_get_set; }
128 set { handle_get_set = value; }
131 public bool AssemblyTargetParsing {
132 get { return handle_assembly; }
133 set { handle_assembly = value; }
136 public bool EventParsing {
137 get { return handle_remove_add; }
138 set { handle_remove_add = value; }
141 public bool ConstraintsParsing {
142 get { return handle_where; }
143 set { handle_where = value; }
146 public bool TypeOfParsing {
147 get { return handle_typeof; }
148 set { handle_typeof = value; }
151 public XmlCommentState doc_state {
152 get { return xml_doc_state; }
154 if (value == XmlCommentState.Allowed) {
155 check_incorrect_doc_comment ();
156 reset_doc_comment ();
158 xml_doc_state = value;
162 public bool IsEscapedIdentifier (Location loc)
164 foreach (LocatedToken lt in escaped_identifiers)
165 if (lt.Location.Equals (loc))
173 static CharArrayHashtable[] keywords;
174 static Hashtable keyword_strings;
175 static NumberStyles styles;
176 static NumberFormatInfo csharp_format_info;
179 // Values for the associated token returned
187 const int TAKING = 1;
188 const int ELSE_SEEN = 4;
189 const int PARENT_TAKING = 8;
190 const int REGION = 16;
193 // pre-processor if stack state:
197 static System.Text.StringBuilder string_builder;
199 const int max_id_size = 512;
200 static char [] id_builder = new char [max_id_size];
202 static CharArrayHashtable [] identifiers = new CharArrayHashtable [max_id_size + 1];
204 const int max_number_size = 512;
205 static char [] number_builder = new char [max_number_size];
206 static int number_pos;
209 // Details about the error encoutered by the tokenizer
211 string error_details;
213 public string error {
215 return error_details;
232 // This is used when the tokenizer needs to save
233 // the current position as it needs to do some parsing
234 // on its own to deamiguate a token in behalf of the
237 Stack position_stack = new Stack (2);
244 public int putback_char;
245 public int previous_col;
246 public Stack ifstack;
247 public int parsing_generic_less_than;
248 public int current_token;
250 public Position (Tokenizer t)
252 position = t.reader.Position;
254 ref_line = t.ref_line;
257 putback_char = t.putback_char;
258 previous_col = t.previous_col;
259 if (t.ifstack != null && t.ifstack.Count != 0)
260 ifstack = (Stack)t.ifstack.Clone ();
261 parsing_generic_less_than = t.parsing_generic_less_than;
262 current_token = t.current_token;
266 public void PushPosition ()
268 position_stack.Push (new Position (this));
271 public void PopPosition ()
273 Position p = (Position) position_stack.Pop ();
275 reader.Position = p.position;
276 ref_line = p.ref_line;
280 putback_char = p.putback_char;
281 previous_col = p.previous_col;
283 parsing_generic_less_than = p.parsing_generic_less_than;
284 current_token = p.current_token;
287 // Do not reset the position, ignore it.
288 public void DiscardPosition ()
290 position_stack.Pop ();
293 static void AddKeyword (string kw, int token)
295 keyword_strings.Add (kw, kw);
296 if (keywords [kw.Length] == null) {
297 keywords [kw.Length] = new CharArrayHashtable (kw.Length);
299 keywords [kw.Length] [kw.ToCharArray ()] = token;
302 static void InitTokens ()
304 keyword_strings = new Hashtable ();
305 keywords = new CharArrayHashtable [64];
307 AddKeyword ("__arglist", Token.ARGLIST);
308 AddKeyword ("abstract", Token.ABSTRACT);
309 AddKeyword ("as", Token.AS);
310 AddKeyword ("add", Token.ADD);
311 AddKeyword ("assembly", Token.ASSEMBLY);
312 AddKeyword ("base", Token.BASE);
313 AddKeyword ("bool", Token.BOOL);
314 AddKeyword ("break", Token.BREAK);
315 AddKeyword ("byte", Token.BYTE);
316 AddKeyword ("case", Token.CASE);
317 AddKeyword ("catch", Token.CATCH);
318 AddKeyword ("char", Token.CHAR);
319 AddKeyword ("checked", Token.CHECKED);
320 AddKeyword ("class", Token.CLASS);
321 AddKeyword ("const", Token.CONST);
322 AddKeyword ("continue", Token.CONTINUE);
323 AddKeyword ("decimal", Token.DECIMAL);
324 AddKeyword ("default", Token.DEFAULT);
325 AddKeyword ("delegate", Token.DELEGATE);
326 AddKeyword ("do", Token.DO);
327 AddKeyword ("double", Token.DOUBLE);
328 AddKeyword ("else", Token.ELSE);
329 AddKeyword ("enum", Token.ENUM);
330 AddKeyword ("event", Token.EVENT);
331 AddKeyword ("explicit", Token.EXPLICIT);
332 AddKeyword ("extern", Token.EXTERN);
333 AddKeyword ("false", Token.FALSE);
334 AddKeyword ("finally", Token.FINALLY);
335 AddKeyword ("fixed", Token.FIXED);
336 AddKeyword ("float", Token.FLOAT);
337 AddKeyword ("for", Token.FOR);
338 AddKeyword ("foreach", Token.FOREACH);
339 AddKeyword ("goto", Token.GOTO);
340 AddKeyword ("get", Token.GET);
341 AddKeyword ("if", Token.IF);
342 AddKeyword ("implicit", Token.IMPLICIT);
343 AddKeyword ("in", Token.IN);
344 AddKeyword ("int", Token.INT);
345 AddKeyword ("interface", Token.INTERFACE);
346 AddKeyword ("internal", Token.INTERNAL);
347 AddKeyword ("is", Token.IS);
348 AddKeyword ("lock", Token.LOCK);
349 AddKeyword ("long", Token.LONG);
350 AddKeyword ("namespace", Token.NAMESPACE);
351 AddKeyword ("new", Token.NEW);
352 AddKeyword ("null", Token.NULL);
353 AddKeyword ("object", Token.OBJECT);
354 AddKeyword ("operator", Token.OPERATOR);
355 AddKeyword ("out", Token.OUT);
356 AddKeyword ("override", Token.OVERRIDE);
357 AddKeyword ("params", Token.PARAMS);
358 AddKeyword ("private", Token.PRIVATE);
359 AddKeyword ("protected", Token.PROTECTED);
360 AddKeyword ("public", Token.PUBLIC);
361 AddKeyword ("readonly", Token.READONLY);
362 AddKeyword ("ref", Token.REF);
363 AddKeyword ("remove", Token.REMOVE);
364 AddKeyword ("return", Token.RETURN);
365 AddKeyword ("sbyte", Token.SBYTE);
366 AddKeyword ("sealed", Token.SEALED);
367 AddKeyword ("set", Token.SET);
368 AddKeyword ("short", Token.SHORT);
369 AddKeyword ("sizeof", Token.SIZEOF);
370 AddKeyword ("stackalloc", Token.STACKALLOC);
371 AddKeyword ("static", Token.STATIC);
372 AddKeyword ("string", Token.STRING);
373 AddKeyword ("struct", Token.STRUCT);
374 AddKeyword ("switch", Token.SWITCH);
375 AddKeyword ("this", Token.THIS);
376 AddKeyword ("throw", Token.THROW);
377 AddKeyword ("true", Token.TRUE);
378 AddKeyword ("try", Token.TRY);
379 AddKeyword ("typeof", Token.TYPEOF);
380 AddKeyword ("uint", Token.UINT);
381 AddKeyword ("ulong", Token.ULONG);
382 AddKeyword ("unchecked", Token.UNCHECKED);
383 AddKeyword ("unsafe", Token.UNSAFE);
384 AddKeyword ("ushort", Token.USHORT);
385 AddKeyword ("using", Token.USING);
386 AddKeyword ("virtual", Token.VIRTUAL);
387 AddKeyword ("void", Token.VOID);
388 AddKeyword ("volatile", Token.VOLATILE);
389 AddKeyword ("while", Token.WHILE);
390 AddKeyword ("partial", Token.PARTIAL);
391 AddKeyword ("where", Token.WHERE);
394 AddKeyword ("from", Token.FROM);
395 AddKeyword ("join", Token.JOIN);
396 AddKeyword ("on", Token.ON);
397 AddKeyword ("equals", Token.EQUALS);
398 AddKeyword ("select", Token.SELECT);
399 AddKeyword ("group", Token.GROUP);
400 AddKeyword ("by", Token.BY);
401 AddKeyword ("let", Token.LET);
402 AddKeyword ("orderby", Token.ORDERBY);
403 AddKeyword ("ascending", Token.ASCENDING);
404 AddKeyword ("descending", Token.DESCENDING);
405 AddKeyword ("into", Token.INTO);
416 public static void Reset ()
419 csharp_format_info = NumberFormatInfo.InvariantInfo;
420 styles = NumberStyles.Float;
422 string_builder = new System.Text.StringBuilder ();
425 int GetKeyword (char[] id, int id_len)
428 * Keywords are stored in an array of hashtables grouped by their
432 if ((id_len >= keywords.Length) || (keywords [id_len] == null))
434 object o = keywords [id_len] [id];
441 if (!handle_get_set && (res == Token.GET || res == Token.SET))
443 if (!handle_remove_add && (res == Token.REMOVE || res == Token.ADD))
445 if (!handle_assembly && res == Token.ASSEMBLY)
449 // A query expression is any expression that starts with `from identifier'
450 // followed by any token except ; , =
452 if (query_parsing == 0) {
453 if (res == Token.FROM && !lambda_arguments_parsing) {
455 // HACK: to disable generics micro-parser, because PushPosition does not
456 // store identifiers array
457 parsing_generic_less_than = 1;
459 case Token.IDENTIFIER:
471 int next_token = xtoken ();
472 if (next_token == Token.SEMICOLON || next_token == Token.COMMA || next_token == Token.EQUALS)
476 if (RootContext.Version <= LanguageVersion.ISO_2)
477 Report.FeatureIsNotAvailable (Location, "query expressions");
480 Expression.Error_VoidInvalidInTheContext (Location);
484 // HACK: A token is not a keyword so we need to restore identifiers buffer
485 // which has been overwritten before we grabbed the identifier
486 id_builder [0] = 'f'; id_builder [1] = 'r'; id_builder [2] = 'o'; id_builder [3] = 'm';
493 if (res > Token.QUERY_FIRST_TOKEN && res < Token.QUERY_LAST_TOKEN)
497 if (res == Token.WHERE && !handle_where && query_parsing == 0)
503 public Location Location {
504 get { return current_location; }
507 public Tokenizer (SeekableStreamReader input, CompilationUnit file)
509 this.ref_name = file;
510 this.file_name = file;
515 xml_comment_buffer = new StringBuilder ();
518 // FIXME: This could be `Location.Push' but we have to
519 // find out why the MS compiler allows this
521 Mono.CSharp.Location.Push (file, file);
524 static bool is_identifier_start_character (char c)
526 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || Char.IsLetter (c);
529 static bool is_identifier_part_character (char c)
531 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || (c >= '0' && c <= '9') ||
532 Char.IsLetter (c) || Char.GetUnicodeCategory (c) == UnicodeCategory.ConnectorPunctuation;
535 public static bool IsKeyword (string s)
537 return keyword_strings [s] != null;
541 // Tests whether '(' is beggining of lambda parameters
543 bool IsLambdaOpenParens ()
546 while ((ntoken = xtoken ()) != Token.EOF) {
548 case Token.CLOSE_PARENS:
549 return xtoken () == Token.ARROW;
552 case Token.SEMICOLON:
553 case Token.OPEN_BRACE:
554 case Token.OPEN_PARENS:
555 case Token.LITERAL_STRING:
556 case Token.LITERAL_INTEGER:
557 case Token.LITERAL_FLOAT:
558 case Token.LITERAL_DOUBLE:
559 case Token.LITERAL_DECIMAL:
560 case Token.LITERAL_CHARACTER:
566 case Token.OP_SHIFT_LEFT:
567 case Token.OP_SHIFT_RIGHT:
574 case Token.BITWISE_AND:
575 case Token.BITWISE_OR:
587 Error_TokenExpected (",' or `)");
591 public static bool IsValidIdentifier (string s)
593 if (s == null || s.Length == 0)
596 if (!is_identifier_start_character (s [0]))
599 for (int i = 1; i < s.Length; i ++)
600 if (! is_identifier_part_character (s [i]))
606 bool parse_less_than ()
609 int the_token = token ();
610 if (the_token == Token.OPEN_BRACKET) {
612 the_token = token ();
613 } while (the_token != Token.CLOSE_BRACKET);
614 the_token = token ();
617 case Token.IDENTIFIER:
636 case Token.OP_GENERICS_GT:
643 the_token = token ();
645 if (the_token == Token.OP_GENERICS_GT)
647 else if (the_token == Token.COMMA || the_token == Token.DOT || the_token == Token.DOUBLE_COLON)
649 else if (the_token == Token.INTERR || the_token == Token.STAR)
651 else if (the_token == Token.OP_GENERICS_LT) {
652 if (!parse_less_than ())
655 } else if (the_token == Token.OPEN_BRACKET) {
657 the_token = token ();
658 if (the_token == Token.CLOSE_BRACKET)
660 else if (the_token == Token.COMMA)
661 goto rank_specifiers;
668 public void PutbackNullable ()
670 if (nullable_pos < 0)
671 throw new Exception ();
675 reader.Position = nullable_pos;
680 public void PutbackCloseParens ()
686 int nullable_pos = -1;
688 public void CheckNullable (bool is_nullable)
691 nullable_pos = reader.Position;
696 bool parse_generic_dimension (out int dimension)
701 int the_token = token ();
702 if (the_token == Token.OP_GENERICS_GT)
704 else if (the_token == Token.COMMA) {
712 public int peek_token ()
717 the_token = token ();
723 int parsing_generic_less_than = 0;
725 int is_punct (char c, ref bool doread)
735 return Token.OPEN_BRACE;
738 return Token.CLOSE_BRACE;
740 // To block doccomment inside attribute declaration.
741 if (doc_state == XmlCommentState.Allowed)
742 doc_state = XmlCommentState.NotAllowed;
743 return Token.OPEN_BRACKET;
745 return Token.CLOSE_BRACKET;
747 if (!lambda_arguments_parsing) {
748 lambda_arguments_parsing = true;
750 bool lambda_start = IsLambdaOpenParens ();
752 lambda_arguments_parsing = false;
754 if (RootContext.Version <= LanguageVersion.ISO_2)
755 Report.FeatureIsNotAvailable (Location, "lambda expressions");
757 return Token.OPEN_PARENS_LAMBDA;
760 return Token.OPEN_PARENS;
762 if (deambiguate_close_parens == 0)
763 return Token.CLOSE_PARENS;
765 --deambiguate_close_parens;
769 int new_token = xtoken ();
773 if (new_token == Token.OPEN_PARENS)
774 return Token.CLOSE_PARENS_OPEN_PARENS;
775 else if (new_token == Token.MINUS)
776 return Token.CLOSE_PARENS_MINUS;
777 else if (IsCastToken (new_token))
778 return Token.CLOSE_PARENS_CAST;
780 return Token.CLOSE_PARENS_NO_CAST;
787 return Token.SEMICOLON;
795 return Token.OP_COALESCING;
801 if (parsing_generic_less_than++ > 0)
802 return Token.OP_GENERICS_LT;
807 if (parse_generic_dimension (out dimension)) {
810 return Token.GENERIC_DIMENSION;
815 // Save current position and parse next token.
817 bool is_generic_lt = parse_less_than ();
821 return Token.OP_GENERICS_LT;
823 parsing_generic_less_than = 0;
832 return Token.OP_SHIFT_LEFT_ASSIGN;
834 return Token.OP_SHIFT_LEFT;
835 } else if (d == '='){
840 } else if (c == '>') {
848 if (parsing_generic_less_than > 1 || (parsing_generic_less_than == 1 && d != '>')) {
849 parsing_generic_less_than--;
850 return Token.OP_GENERICS_GT;
859 return Token.OP_SHIFT_RIGHT_ASSIGN;
861 return Token.OP_SHIFT_RIGHT;
875 t = Token.OP_ADD_ASSIGN;
889 t = Token.OP_SUB_ASSIGN;
927 } else if (d == '='){
929 return Token.OP_AND_ASSIGN;
932 return Token.BITWISE_AND;
939 } else if (d == '='){
941 return Token.OP_OR_ASSIGN;
943 return Token.BITWISE_OR;
949 return Token.OP_MULT_ASSIGN;
958 return Token.OP_DIV_ASSIGN;
966 return Token.OP_MOD_ASSIGN;
968 return Token.PERCENT;
974 return Token.OP_XOR_ASSIGN;
982 return Token.DOUBLE_COLON;
991 int deambiguate_close_parens = 0;
993 public void Deambiguate_CloseParens (object expression)
997 // When any binary operation, a conditional is used we are sure it is not a cast
1000 if (expression is Binary || expression is Conditional)
1003 deambiguate_close_parens++;
1006 bool decimal_digits (int c)
1009 bool seen_digits = false;
1012 if (number_pos == max_number_size)
1013 Error_NumericConstantTooLong ();
1014 number_builder [number_pos++] = (char) c;
1018 // We use peek_char2, because decimal_digits needs to do a
1019 // 2-character look-ahead (5.ToString for example).
1021 while ((d = peek_char2 ()) != -1){
1022 if (d >= '0' && d <= '9'){
1023 if (number_pos == max_number_size)
1024 Error_NumericConstantTooLong ();
1025 number_builder [number_pos++] = (char) d;
1035 static bool is_hex (int e)
1037 return (e >= '0' && e <= '9') || (e >= 'A' && e <= 'F') || (e >= 'a' && e <= 'f');
1040 static int real_type_suffix (int c)
1046 t = Token.LITERAL_FLOAT;
1049 t = Token.LITERAL_DOUBLE;
1052 t= Token.LITERAL_DECIMAL;
1060 int integer_type_suffix (ulong ul, int c)
1062 bool is_unsigned = false;
1063 bool is_long = false;
1066 bool scanning = true;
1079 // if we have not seen anything in between
1080 // report this error
1082 Report.Warning (78, 4, Location, "The 'l' suffix is easily confused with the digit '1' (use 'L' for clarity)");
1085 // This goto statement causes the MS CLR 2.0 beta 1 csc to report an error, so
1086 // work around that.
1110 if (is_long && is_unsigned){
1112 return Token.LITERAL_INTEGER;
1113 } else if (is_unsigned){
1114 // uint if possible, or ulong else.
1116 if ((ul & 0xffffffff00000000) == 0)
1120 } else if (is_long){
1121 // long if possible, ulong otherwise
1122 if ((ul & 0x8000000000000000) != 0)
1127 // int, uint, long or ulong in that order
1128 if ((ul & 0xffffffff00000000) == 0){
1129 uint ui = (uint) ul;
1131 if ((ui & 0x80000000) != 0)
1136 if ((ul & 0x8000000000000000) != 0)
1142 return Token.LITERAL_INTEGER;
1146 // given `c' as the next char in the input decide whether
1147 // we need to convert to a special type, and then choose
1148 // the best representation for the integer
1150 int adjust_int (int c)
1153 if (number_pos > 9){
1154 ulong ul = (uint) (number_builder [0] - '0');
1156 for (int i = 1; i < number_pos; i++){
1157 ul = checked ((ul * 10) + ((uint)(number_builder [i] - '0')));
1159 return integer_type_suffix (ul, c);
1161 uint ui = (uint) (number_builder [0] - '0');
1163 for (int i = 1; i < number_pos; i++){
1164 ui = checked ((ui * 10) + ((uint)(number_builder [i] - '0')));
1166 return integer_type_suffix (ui, c);
1168 } catch (OverflowException) {
1169 error_details = "Integral constant is too large";
1170 Report.Error (1021, Location, error_details);
1172 return Token.LITERAL_INTEGER;
1174 catch (FormatException) {
1175 Report.Error (1013, Location, "Invalid number");
1177 return Token.LITERAL_INTEGER;
1181 int adjust_real (int t)
1183 string s = new String (number_builder, 0, number_pos);
1184 const string error_details = "Floating-point constant is outside the range of type `{0}'";
1187 case Token.LITERAL_DECIMAL:
1189 val = System.Decimal.Parse (s, styles, csharp_format_info);
1190 } catch (OverflowException) {
1192 Report.Error (594, Location, error_details, "decimal");
1195 case Token.LITERAL_FLOAT:
1197 val = float.Parse (s, styles, csharp_format_info);
1198 } catch (OverflowException) {
1200 Report.Error (594, Location, error_details, "float");
1204 case Token.LITERAL_DOUBLE:
1206 t = Token.LITERAL_DOUBLE;
1208 val = System.Double.Parse (s, styles, csharp_format_info);
1209 } catch (OverflowException) {
1211 Report.Error (594, Location, error_details, "double");
1224 while ((d = peek_char ()) != -1){
1226 number_builder [number_pos++] = (char) d;
1232 string s = new String (number_builder, 0, number_pos);
1234 if (number_pos <= 8)
1235 ul = System.UInt32.Parse (s, NumberStyles.HexNumber);
1237 ul = System.UInt64.Parse (s, NumberStyles.HexNumber);
1238 } catch (OverflowException){
1239 error_details = "Integral constant is too large";
1240 Report.Error (1021, Location, error_details);
1242 return Token.LITERAL_INTEGER;
1244 catch (FormatException) {
1245 Report.Error (1013, Location, "Invalid number");
1247 return Token.LITERAL_INTEGER;
1250 return integer_type_suffix (ul, peek_char ());
1254 // Invoked if we know we have .digits or digits
1256 int is_number (int c)
1258 bool is_real = false;
1263 if (c >= '0' && c <= '9'){
1265 int peek = peek_char ();
1267 if (peek == 'x' || peek == 'X')
1268 return handle_hex ();
1275 // We need to handle the case of
1276 // "1.1" vs "1.string" (LITERAL_FLOAT vs NUMBER DOT IDENTIFIER)
1279 if (decimal_digits ('.')){
1285 return adjust_int (-1);
1289 if (c == 'e' || c == 'E'){
1291 if (number_pos == max_number_size)
1292 Error_NumericConstantTooLong ();
1293 number_builder [number_pos++] = 'e';
1297 if (number_pos == max_number_size)
1298 Error_NumericConstantTooLong ();
1299 number_builder [number_pos++] = '+';
1301 } else if (c == '-') {
1302 if (number_pos == max_number_size)
1303 Error_NumericConstantTooLong ();
1304 number_builder [number_pos++] = '-';
1307 if (number_pos == max_number_size)
1308 Error_NumericConstantTooLong ();
1309 number_builder [number_pos++] = '+';
1316 type = real_type_suffix (c);
1317 if (type == Token.NONE && !is_real){
1319 return adjust_int (c);
1323 if (type == Token.NONE){
1328 return adjust_real (type);
1330 Console.WriteLine ("This should not be reached");
1331 throw new Exception ("Is Number should never reach this point");
1335 // Accepts exactly count (4 or 8) hex, no more no less
1337 int getHex (int count, out int surrogate, out bool error)
1342 int top = count != -1 ? count : 4;
1347 for (i = 0; i < top; i++){
1350 if (c >= '0' && c <= '9')
1351 c = (int) c - (int) '0';
1352 else if (c >= 'A' && c <= 'F')
1353 c = (int) c - (int) 'A' + 10;
1354 else if (c >= 'a' && c <= 'f')
1355 c = (int) c - (int) 'a' + 10;
1361 total = (total * 16) + c;
1363 int p = peek_char ();
1366 if (!is_hex ((char)p))
1372 if (total > 0x0010FFFF) {
1377 if (total >= 0x00010000) {
1378 total = ((total - 0x00010000) / 0x0400 + 0xD800);
1379 surrogate = ((total - 0x00010000) % 0x0400 + 0xDC00);
1386 int escape (int c, out int surrogate)
1422 v = getHex (-1, out surrogate, out error);
1428 return EscapeUnicode (d, out surrogate);
1431 Report.Error (1009, Location, "Unrecognized escape sequence `\\{0}'", ((char)d).ToString ());
1440 int EscapeUnicode (int ch, out int surrogate)
1444 ch = getHex (8, out surrogate, out error);
1446 ch = getHex (4, out surrogate, out error);
1450 Report.Error (1009, Location, "Unrecognized escape sequence");
1458 if (putback_char != -1) {
1471 void advance_line ()
1481 if (putback_char != -1)
1482 return putback_char;
1483 putback_char = reader.Read ();
1484 return putback_char;
1489 if (putback_char != -1)
1490 return putback_char;
1491 return reader.Peek ();
1494 void putback (int c)
1496 if (putback_char != -1){
1497 Console.WriteLine ("Col: " + col);
1498 Console.WriteLine ("Row: " + line);
1499 Console.WriteLine ("Name: " + ref_name.Name);
1500 Console.WriteLine ("Current [{0}] putting back [{1}] ", putback_char, c);
1501 throw new Exception ("This should not happen putback on putback");
1503 if (c == '\n' || col == 0) {
1504 // It won't happen though.
1514 public bool advance ()
1516 return peek_char () != -1;
1519 public Object Value {
1525 public Object value ()
1530 static bool IsCastToken (int token)
1535 case Token.IDENTIFIER:
1536 case Token.LITERAL_INTEGER:
1537 case Token.LITERAL_FLOAT:
1538 case Token.LITERAL_DOUBLE:
1539 case Token.LITERAL_DECIMAL:
1540 case Token.LITERAL_CHARACTER:
1541 case Token.LITERAL_STRING:
1544 case Token.DELEGATE:
1554 case Token.UNCHECKED:
1559 // These can be part of a member access
1581 current_token = xtoken ();
1583 if (current_token != Token.DEFAULT)
1584 return current_token;
1589 current_token = Token.ERROR;
1590 else if (c == Token.OPEN_PARENS)
1591 current_token = Token.DEFAULT_OPEN_PARENS;
1592 else if (c == Token.COLON)
1593 current_token = Token.DEFAULT_COLON;
1597 return current_token;
1600 static StringBuilder static_cmd_arg = new System.Text.StringBuilder ();
1602 void get_cmd_arg (out string cmd, out string arg)
1606 tokens_seen = false;
1609 // skip over white space
1612 } while (c == '\r' || c == ' ' || c == '\t');
1614 static_cmd_arg.Length = 0;
1615 while (c != -1 && is_identifier_part_character ((char)c)) {
1616 static_cmd_arg.Append ((char)c);
1619 int peek = peek_char ();
1620 if (peek == 'U' || peek == 'u') {
1622 c = EscapeUnicode (c, out surrogate);
1623 if (surrogate != 0) {
1624 if (is_identifier_part_character ((char) c))
1625 static_cmd_arg.Append ((char) c);
1632 cmd = static_cmd_arg.ToString ();
1634 // skip over white space
1635 while (c == '\r' || c == ' ' || c == '\t')
1638 static_cmd_arg.Length = 0;
1639 while (c != -1 && c != '\n' && c != '\r') {
1641 int peek = peek_char ();
1642 if (peek == 'U' || peek == 'u') {
1644 c = EscapeUnicode (c, out surrogate);
1645 if (surrogate != 0) {
1646 if (is_identifier_part_character ((char) c))
1647 static_cmd_arg.Append ((char) c);
1652 static_cmd_arg.Append ((char) c);
1656 if (static_cmd_arg.Length != 0)
1657 arg = static_cmd_arg.ToString ();
1661 // Handles the #line directive
1663 bool PreProcessLine (string arg)
1665 if (arg.Length == 0)
1668 if (arg == "default"){
1670 ref_name = file_name;
1672 Location.Push (file_name, ref_name);
1674 } else if (arg == "hidden"){
1682 if ((pos = arg.IndexOf (' ')) != -1 && pos != 0){
1683 ref_line = System.Int32.Parse (arg.Substring (0, pos));
1686 char [] quotes = { '\"' };
1688 string name = arg.Substring (pos). Trim (quotes);
1689 ref_name = Location.LookupFile (file_name, name);
1690 file_name.AddFile (ref_name);
1692 Location.Push (file_name, ref_name);
1694 ref_line = System.Int32.Parse (arg);
1705 // Handles #define and #undef
1707 void PreProcessDefinition (bool is_define, string ident, bool caller_is_taking)
1709 if (ident.Length == 0 || ident == "true" || ident == "false"){
1710 Report.Error (1001, Location, "Missing identifer to pre-processor directive");
1714 if (ident.IndexOfAny (simple_whitespaces) != -1){
1715 Error_EndLineExpected ();
1719 if (!is_identifier_start_character (ident [0]))
1720 Report.Error (1001, Location, "Identifier expected: " + ident);
1722 foreach (char c in ident.Substring (1)){
1723 if (!is_identifier_part_character (c)){
1724 Report.Error (1001, Location, "Identifier expected: " + ident);
1729 if (!caller_is_taking)
1736 if (RootContext.IsConditionalDefined (ident))
1739 file_name.AddDefine (ident);
1744 file_name.AddUndefine (ident);
1748 static byte read_hex (string arg, int pos, out bool error)
1755 if ((c >= '0') && (c <= '9'))
1756 total = (int) c - (int) '0';
1757 else if ((c >= 'A') && (c <= 'F'))
1758 total = (int) c - (int) 'A' + 10;
1759 else if ((c >= 'a') && (c <= 'f'))
1760 total = (int) c - (int) 'a' + 10;
1769 if ((c >= '0') && (c <= '9'))
1770 total += (int) c - (int) '0';
1771 else if ((c >= 'A') && (c <= 'F'))
1772 total += (int) c - (int) 'A' + 10;
1773 else if ((c >= 'a') && (c <= 'f'))
1774 total += (int) c - (int) 'a' + 10;
1780 return (byte) total;
1784 /// Handles #pragma checksum
1786 bool PreProcessPragmaChecksum (string arg)
1788 if ((arg [0] != ' ') && (arg [0] != '\t'))
1791 arg = arg.Trim (simple_whitespaces);
1792 if ((arg.Length < 2) || (arg [0] != '"'))
1795 StringBuilder file_sb = new StringBuilder ();
1799 while ((ch = arg [pos++]) != '"') {
1800 if (pos >= arg.Length)
1804 if (pos+1 >= arg.Length)
1809 file_sb.Append (ch);
1812 if ((pos+2 >= arg.Length) || ((arg [pos] != ' ') && (arg [pos] != '\t')))
1815 arg = arg.Substring (pos).Trim (simple_whitespaces);
1816 if ((arg.Length < 42) || (arg [0] != '"') || (arg [1] != '{') ||
1817 (arg [10] != '-') || (arg [15] != '-') || (arg [20] != '-') ||
1818 (arg [25] != '-') || (arg [38] != '}') || (arg [39] != '"'))
1822 byte[] guid_bytes = new byte [16];
1824 for (int i = 0; i < 4; i++) {
1825 guid_bytes [i] = read_hex (arg, 2+2*i, out error);
1829 for (int i = 0; i < 2; i++) {
1830 guid_bytes [i+4] = read_hex (arg, 11+2*i, out error);
1833 guid_bytes [i+6] = read_hex (arg, 16+2*i, out error);
1836 guid_bytes [i+8] = read_hex (arg, 21+2*i, out error);
1841 for (int i = 0; i < 6; i++) {
1842 guid_bytes [i+10] = read_hex (arg, 26+2*i, out error);
1847 arg = arg.Substring (40).Trim (simple_whitespaces);
1848 if ((arg.Length < 34) || (arg [0] != '"') || (arg [33] != '"'))
1851 byte[] checksum_bytes = new byte [16];
1852 for (int i = 0; i < 16; i++) {
1853 checksum_bytes [i] = read_hex (arg, 1+2*i, out error);
1858 arg = arg.Substring (34).Trim (simple_whitespaces);
1862 SourceFile file = Location.LookupFile (file_name, file_sb.ToString ());
1863 file.SetChecksum (guid_bytes, checksum_bytes);
1864 ref_name.AutoGenerated = true;
1869 /// Handles #pragma directive
1871 void PreProcessPragma (string arg)
1873 const string warning = "warning";
1874 const string w_disable = "warning disable";
1875 const string w_restore = "warning restore";
1876 const string checksum = "checksum";
1878 if (arg == w_disable) {
1879 Report.RegisterWarningRegion (Location).WarningDisable (Location.Row);
1883 if (arg == w_restore) {
1884 Report.RegisterWarningRegion (Location).WarningEnable (Location.Row);
1888 if (arg.StartsWith (w_disable)) {
1889 int[] codes = ParseNumbers (arg.Substring (w_disable.Length));
1890 foreach (int code in codes) {
1892 Report.RegisterWarningRegion (Location).WarningDisable (Location, code);
1897 if (arg.StartsWith (w_restore)) {
1898 int[] codes = ParseNumbers (arg.Substring (w_restore.Length));
1899 Hashtable w_table = Report.warning_ignore_table;
1900 foreach (int code in codes) {
1901 if (w_table != null && w_table.Contains (code))
1902 Report.Warning (1635, 1, Location, String.Format ("Cannot restore warning `CS{0:0000}' because it was disabled globally", code));
1903 Report.RegisterWarningRegion (Location).WarningEnable (Location, code);
1908 if (arg.StartsWith (warning)) {
1909 Report.Warning (1634, 1, Location, "Expected disable or restore");
1913 if (arg.StartsWith (checksum)) {
1914 if (!PreProcessPragmaChecksum (arg.Substring (checksum.Length)))
1915 Warning_InvalidPragmaChecksum ();
1919 Report.Warning (1633, 1, Location, "Unrecognized #pragma directive");
1922 int[] ParseNumbers (string text)
1924 string[] string_array = text.Split (',');
1925 int[] values = new int [string_array.Length];
1927 foreach (string string_code in string_array) {
1929 values[index++] = int.Parse (string_code, System.Globalization.CultureInfo.InvariantCulture);
1931 catch (FormatException) {
1932 Report.Warning (1692, 1, Location, "Invalid number");
1938 bool eval_val (string s)
1945 return file_name.IsConditionalDefined (s);
1948 bool pp_primary (ref string s)
1957 s = s.Substring (1);
1958 bool val = pp_expr (ref s, false);
1959 if (s.Length > 0 && s [0] == ')'){
1960 s = s.Substring (1);
1963 Error_InvalidDirective ();
1967 if (is_identifier_start_character (c)){
1973 if (is_identifier_part_character (c)){
1977 bool v = eval_val (s.Substring (0, j));
1978 s = s.Substring (j);
1981 bool vv = eval_val (s);
1986 Error_InvalidDirective ();
1990 bool pp_unary (ref string s)
1997 if (len > 1 && s [1] == '='){
1998 Error_InvalidDirective ();
2001 s = s.Substring (1);
2002 return ! pp_primary (ref s);
2004 return pp_primary (ref s);
2006 Error_InvalidDirective ();
2011 bool pp_eq (ref string s)
2013 bool va = pp_unary (ref s);
2019 if (len > 2 && s [1] == '='){
2020 s = s.Substring (2);
2021 return va == pp_unary (ref s);
2023 Error_InvalidDirective ();
2026 } else if (s [0] == '!' && len > 1 && s [1] == '='){
2027 s = s.Substring (2);
2029 return va != pp_unary (ref s);
2038 bool pp_and (ref string s)
2040 bool va = pp_eq (ref s);
2046 if (len > 2 && s [1] == '&'){
2047 s = s.Substring (2);
2048 return (va & pp_and (ref s));
2050 Error_InvalidDirective ();
2059 // Evaluates an expression for `#if' or `#elif'
2061 bool pp_expr (ref string s, bool isTerm)
2063 bool va = pp_and (ref s);
2070 if (len > 2 && s [1] == '|'){
2071 s = s.Substring (2);
2072 return va | pp_expr (ref s, isTerm);
2074 Error_InvalidDirective ();
2079 Error_EndLineExpected ();
2087 bool eval (string s)
2089 bool v = pp_expr (ref s, true);
2098 void Error_NumericConstantTooLong ()
2100 Report.Error (1021, Location, "Numeric constant too long");
2103 void Error_InvalidDirective ()
2105 Report.Error (1517, Location, "Invalid preprocessor directive");
2108 void Error_UnexpectedDirective (string extra)
2112 "Unexpected processor directive (" + extra + ")");
2115 void Error_TokenExpected (string token)
2117 Report.Error (1026, Location, "Expected `{0}'", token);
2120 void Error_TokensSeen ()
2122 Report.Error (1032, Location,
2123 "Cannot define or undefine preprocessor symbols after first token in file");
2126 void Eror_WrongPreprocessorLocation ()
2128 Report.Error (1040, Location,
2129 "Preprocessor directives must appear as the first non-whitespace character on a line");
2132 void Error_EndLineExpected ()
2134 Report.Error (1025, Location, "Single-line comment or end-of-line expected");
2137 void Warning_InvalidPragmaChecksum ()
2139 Report.Warning (1695, 1, Location,
2140 "Invalid #pragma checksum syntax; should be " +
2141 "#pragma checksum \"filename\" " +
2142 "\"{XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX}\" \"XXXX...\"");
2145 // if true, then the code continues processing the code
2146 // if false, the code stays in a loop until another directive is
2148 // When caller_is_taking is false we ignore all directives except the ones
2149 // which can help us to identify where the #if block ends
2150 bool handle_preprocessing_directive (bool caller_is_taking)
2153 bool region_directive = false;
2155 get_cmd_arg (out cmd, out arg);
2157 // Eat any trailing whitespaces and single-line comments
2158 if (arg.IndexOf ("//") != -1)
2159 arg = arg.Substring (0, arg.IndexOf ("//"));
2160 arg = arg.Trim (simple_whitespaces);
2163 // The first group of pre-processing instructions is always processed
2167 region_directive = true;
2172 if (ifstack == null || ifstack.Count == 0){
2173 Error_UnexpectedDirective ("no #region for this #endregion");
2176 int pop = (int) ifstack.Pop ();
2178 if ((pop & REGION) == 0)
2179 Report.Error (1027, Location, "Expected `#endif' directive");
2181 return caller_is_taking;
2184 if (ifstack == null)
2185 ifstack = new Stack (2);
2187 int flags = region_directive ? REGION : 0;
2188 if (ifstack.Count == 0){
2189 flags |= PARENT_TAKING;
2191 int state = (int) ifstack.Peek ();
2192 if ((state & TAKING) != 0) {
2193 flags |= PARENT_TAKING;
2197 if (caller_is_taking && eval (arg)) {
2198 ifstack.Push (flags | TAKING);
2201 ifstack.Push (flags);
2205 if (ifstack == null || ifstack.Count == 0){
2206 Error_UnexpectedDirective ("no #if for this #endif");
2209 pop = (int) ifstack.Pop ();
2211 if ((pop & REGION) != 0)
2212 Report.Error (1038, Location, "#endregion directive expected");
2214 if (arg.Length != 0) {
2215 Error_EndLineExpected ();
2218 if (ifstack.Count == 0)
2221 int state = (int) ifstack.Peek ();
2222 return (state & TAKING) != 0;
2226 if (ifstack == null || ifstack.Count == 0){
2227 Error_UnexpectedDirective ("no #if for this #elif");
2230 int state = (int) ifstack.Pop ();
2232 if ((state & REGION) != 0) {
2233 Report.Error (1038, Location, "#endregion directive expected");
2237 if ((state & ELSE_SEEN) != 0){
2238 Error_UnexpectedDirective ("#elif not valid after #else");
2242 if ((state & TAKING) != 0) {
2247 if (eval (arg) && ((state & PARENT_TAKING) != 0)){
2248 ifstack.Push (state | TAKING);
2252 ifstack.Push (state);
2257 if (ifstack == null || ifstack.Count == 0){
2258 Error_UnexpectedDirective ("no #if for this #else");
2261 int state = (int) ifstack.Peek ();
2263 if ((state & REGION) != 0) {
2264 Report.Error (1038, Location, "#endregion directive expected");
2268 if ((state & ELSE_SEEN) != 0){
2269 Error_UnexpectedDirective ("#else within #else");
2275 if (arg.Length != 0) {
2276 Error_EndLineExpected ();
2281 if ((state & PARENT_TAKING) != 0) {
2282 ret = (state & TAKING) == 0;
2290 ifstack.Push (state | ELSE_SEEN);
2295 if (any_token_seen){
2296 Error_TokensSeen ();
2297 return caller_is_taking;
2299 PreProcessDefinition (true, arg, caller_is_taking);
2300 return caller_is_taking;
2303 if (any_token_seen){
2304 Error_TokensSeen ();
2305 return caller_is_taking;
2307 PreProcessDefinition (false, arg, caller_is_taking);
2308 return caller_is_taking;
2312 // These are only processed if we are in a `taking' block
2314 if (!caller_is_taking)
2319 Report.Error (1029, Location, "#error: '" + arg + "'");
2323 Report.Warning (1030, 1, Location, "#warning: `{0}'", arg);
2327 if (RootContext.Version == LanguageVersion.ISO_1) {
2328 Report.FeatureIsNotAvailable (Location, "#pragma");
2332 PreProcessPragma (arg);
2336 if (!PreProcessLine (arg))
2339 "The line number specified for #line directive is missing or invalid");
2340 return caller_is_taking;
2343 Report.Error (1024, Location, "Wrong preprocessor directive");
2348 private int consume_string (bool quoted)
2351 string_builder.Length = 0;
2354 // No need to parse full string when parsing lambda arguments
2356 if (lambda_arguments_parsing)
2357 return Token.LITERAL_STRING;
2359 while ((c = get_char ()) != -1){
2361 if (quoted && peek_char () == '"'){
2362 string_builder.Append ((char) c);
2366 val = string_builder.ToString ();
2367 return Token.LITERAL_STRING;
2373 Report.Error (1010, Location, "Newline in constant");
2378 c = escape (c, out surrogate);
2381 if (surrogate != 0) {
2382 string_builder.Append ((char) c);
2386 string_builder.Append ((char) c);
2389 Report.Error (1039, Location, "Unterminated string literal");
2393 private int consume_identifier (int s)
2395 int res = consume_identifier (s, false);
2397 if (doc_state == XmlCommentState.Allowed)
2398 doc_state = XmlCommentState.NotAllowed;
2401 case Token.NAMESPACE:
2402 check_incorrect_doc_comment ();
2406 if (res == Token.PARTIAL) {
2407 if (parsing_block > 0) {
2408 val = new LocatedToken (Location, "partial");
2409 return Token.IDENTIFIER;
2412 // Save current position and parse next token.
2415 int next_token = token ();
2416 bool ok = (next_token == Token.CLASS) ||
2417 (next_token == Token.STRUCT) ||
2418 (next_token == Token.INTERFACE) ||
2419 (next_token == Token.VOID);
2424 if (next_token == Token.VOID) {
2425 if (RootContext.Version <= LanguageVersion.ISO_2)
2426 Report.FeatureIsNotAvailable (Location, "partial methods");
2427 } else if (RootContext.Version == LanguageVersion.ISO_1)
2428 Report.FeatureIsNotAvailable (Location, "partial types");
2433 if (next_token < Token.LAST_KEYWORD)
2434 Report.Error (267, Location, "The `partial' modifier can be used only immediately before `class', `struct', `interface', or `void' keyword");
2436 val = new LocatedToken (Location, "partial");
2437 return Token.IDENTIFIER;
2443 private int consume_identifier (int c, bool quoted)
2449 c = escape (c, out surrogate);
2450 if (surrogate != 0) {
2451 id_builder [pos++] = (char) c;
2456 id_builder [pos++] = (char) c;
2457 current_location = new Location (ref_line, hidden ? -1 : Col);
2459 while ((c = get_char ()) != -1) {
2461 if (is_identifier_part_character ((char) c)){
2462 if (pos == max_id_size){
2463 Report.Error (645, Location, "Identifier too long (limit is 512 chars)");
2467 id_builder [pos++] = (char) c;
2468 // putback_char = -1;
2469 } else if (c == '\\') {
2471 c = escape (c, out surrogate);
2472 if (surrogate != 0) {
2473 if (is_identifier_part_character ((char) c))
2474 id_builder [pos++] = (char) c;
2479 // putback_char = c;
2486 // Optimization: avoids doing the keyword lookup
2487 // on uppercase letters
2489 if (id_builder [0] >= '_' && !quoted) {
2490 int keyword = GetKeyword (id_builder, pos);
2491 if (keyword != -1) {
2498 // Keep identifiers in an array of hashtables to avoid needless
2502 if (identifiers [pos] != null) {
2503 val = identifiers [pos][id_builder];
2505 val = new LocatedToken (Location, (string) val);
2507 escaped_identifiers.Add (val);
2508 return Token.IDENTIFIER;
2512 identifiers [pos] = new CharArrayHashtable (pos);
2514 val = new String (id_builder, 0, pos);
2515 if (RootContext.Version == LanguageVersion.ISO_1) {
2516 for (int i = 1; i < id_builder.Length; i += 3) {
2517 if (id_builder [i] == '_' && (id_builder [i - 1] == '_' || id_builder [i + 1] == '_')) {
2518 Report.Error (1638, Location,
2519 "`{0}': Any identifier with double underscores cannot be used when ISO language version mode is specified", val.ToString ());
2525 char [] chars = new char [pos];
2526 Array.Copy (id_builder, chars, pos);
2528 identifiers [pos] [chars] = val;
2530 val = new LocatedToken (Location, (string) val);
2532 escaped_identifiers.Add (val);
2533 return Token.IDENTIFIER;
2536 public int xtoken ()
2539 bool doread = false;
2542 // Whether we have seen comments on the current line
2543 bool comments_seen = false;
2545 for (;(c = get_char ()) != -1;) {
2547 col = ((col + 8) / 8) * 8;
2551 if (c == ' ' || c == '\f' || c == '\v' || c == 0xa0 || c == 0)
2555 if (peek_char () != '\n')
2560 any_token_seen |= tokens_seen;
2561 tokens_seen = false;
2562 comments_seen = false;
2566 // Handle double-slash comments.
2568 int d = peek_char ();
2572 if (RootContext.Documentation != null && peek_char () == '/') {
2574 // Don't allow ////.
2575 if ((d = peek_char ()) != '/') {
2576 update_comment_location ();
2577 if (doc_state == XmlCommentState.Allowed)
2578 handle_one_line_xml_comment ();
2579 else if (doc_state == XmlCommentState.NotAllowed)
2580 warn_incorrect_doc_comment ();
2583 while ((d = get_char ()) != -1 && (d != '\n') && d != '\r');
2585 any_token_seen |= tokens_seen;
2586 tokens_seen = false;
2587 comments_seen = false;
2589 } else if (d == '*'){
2591 bool docAppend = false;
2592 if (RootContext.Documentation != null && peek_char () == '*') {
2594 update_comment_location ();
2595 // But when it is /**/, just do nothing.
2596 if (peek_char () == '/') {
2600 if (doc_state == XmlCommentState.Allowed)
2602 else if (doc_state == XmlCommentState.NotAllowed)
2603 warn_incorrect_doc_comment ();
2606 int current_comment_start = 0;
2608 current_comment_start = xml_comment_buffer.Length;
2609 xml_comment_buffer.Append (Environment.NewLine);
2612 Location start_location = Location;
2614 while ((d = get_char ()) != -1){
2615 if (d == '*' && peek_char () == '/'){
2617 comments_seen = true;
2621 xml_comment_buffer.Append ((char) d);
2624 any_token_seen |= tokens_seen;
2625 tokens_seen = false;
2627 // Reset 'comments_seen' just to be consistent.
2628 // It doesn't matter either way, here.
2630 comments_seen = false;
2634 Report.Error (1035, start_location, "End-of-file found, '*/' expected");
2637 update_formatted_doc_comment (current_comment_start);
2640 goto is_punct_label;
2644 if (c == '\\' || is_identifier_start_character ((char)c)){
2646 return consume_identifier (c);
2650 current_location = new Location (ref_line, hidden ? -1 : Col);
2651 if ((t = is_punct ((char)c, ref doread)) != Token.ERROR){
2661 any_token_seen |= tokens_seen;
2662 tokens_seen = false;
2663 comments_seen = false;
2667 if (c >= '0' && c <= '9'){
2669 return is_number (c);
2674 int peek = peek_char ();
2675 if (peek >= '0' && peek <= '9')
2676 return is_number (c);
2681 if (tokens_seen || comments_seen) {
2682 Eror_WrongPreprocessorLocation ();
2686 if (handle_preprocessing_directive (true))
2689 bool directive_expected = false;
2690 while ((c = get_char ()) != -1) {
2692 directive_expected = true;
2693 } else if (!directive_expected) {
2694 // TODO: Implement comment support for disabled code and uncomment this code
2696 // Eror_WrongPreprocessorLocation ();
2697 // return Token.ERROR;
2702 if (c == ' ' || c == '\t' || c == '\r' || c == '\n' || c == '\f' || c == '\v' )
2706 if (handle_preprocessing_directive (false))
2709 directive_expected = false;
2713 tokens_seen = false;
2721 return consume_string (false);
2727 error_details = "Empty character literal";
2728 Report.Error (1011, Location, error_details);
2731 if (c == '\r' || c == '\n') {
2732 Report.Error (1010, Location, "Newline in constant");
2737 c = escape (c, out surrogate);
2741 throw new NotImplementedException ();
2747 error_details = "Too many characters in character literal";
2748 Report.Error (1012, Location, error_details);
2750 // Try to recover, read until newline or next "'"
2751 while ((c = get_char ()) != -1){
2760 return Token.LITERAL_CHARACTER;
2767 return consume_string (true);
2768 } else if (is_identifier_start_character ((char) c)){
2769 return consume_identifier (c, true);
2771 Report.Error (1646, Location, "Keyword, identifier, or string expected after verbatim specifier: @");
2775 error_details = ((char)c).ToString ();
2784 // Handles one line xml comment
2786 private void handle_one_line_xml_comment ()
2789 while ((c = peek_char ()) == ' ')
2790 get_char (); // skip heading whitespaces.
2791 while ((c = peek_char ()) != -1 && c != '\n' && c != '\r') {
2792 xml_comment_buffer.Append ((char) get_char ());
2794 if (c == '\r' || c == '\n')
2795 xml_comment_buffer.Append (Environment.NewLine);
2799 // Remove heading "*" in Javadoc-like xml documentation.
2801 private void update_formatted_doc_comment (int current_comment_start)
2803 int length = xml_comment_buffer.Length - current_comment_start;
2804 string [] lines = xml_comment_buffer.ToString (
2805 current_comment_start,
2806 length).Replace ("\r", "").Split ('\n');
2808 // The first line starts with /**, thus it is not target
2809 // for the format check.
2810 for (int i = 1; i < lines.Length; i++) {
2811 string s = lines [i];
2812 int idx = s.IndexOf ('*');
2815 if (i < lines.Length - 1)
2819 head = s.Substring (0, idx);
2820 foreach (char c in head)
2823 lines [i] = s.Substring (idx + 1);
2825 xml_comment_buffer.Remove (current_comment_start, length);
2826 xml_comment_buffer.Insert (current_comment_start, String.Join (Environment.NewLine, lines));
2830 // Updates current comment location.
2832 private void update_comment_location ()
2834 if (current_comment_location.IsNull) {
2835 // "-2" is for heading "//" or "/*"
2836 current_comment_location =
2837 new Location (ref_line, hidden ? -1 : col - 2);
2842 // Checks if there was incorrect doc comments and raise
2845 public void check_incorrect_doc_comment ()
2847 if (xml_comment_buffer.Length > 0)
2848 warn_incorrect_doc_comment ();
2852 // Raises a warning when tokenizer found incorrect doccomment
2855 private void warn_incorrect_doc_comment ()
2857 if (doc_state != XmlCommentState.Error) {
2858 doc_state = XmlCommentState.Error;
2859 // in csc, it is 'XML comment is not placed on
2860 // a valid language element'. But that does not
2862 Report.Warning (1587, 2, Location, "XML comment is not placed on a valid language element");
2867 // Consumes the saved xml comment lines (if any)
2868 // as for current target member or type.
2870 public string consume_doc_comment ()
2872 if (xml_comment_buffer.Length > 0) {
2873 string ret = xml_comment_buffer.ToString ();
2874 reset_doc_comment ();
2880 void reset_doc_comment ()
2882 xml_comment_buffer.Length = 0;
2883 current_comment_location = Location.Null;
2886 public void cleanup ()
2888 if (ifstack != null && ifstack.Count >= 1) {
2889 current_location = new Location (ref_line, hidden ? -1 : Col);
2890 int state = (int) ifstack.Pop ();
2891 if ((state & REGION) != 0)
2892 Report.Error (1038, Location, "#endregion directive expected");
2894 Report.Error (1027, Location, "Expected `#endif' directive");
2900 // Indicates whether it accepts XML documentation or not.
2902 public enum XmlCommentState {
2903 // comment is allowed in this state.
2905 // comment is not allowed in this state.
2907 // once comments appeared when it is NotAllowed, then the
2908 // state is changed to it, until the state is changed to