2 // cs-tokenizer.cs: The Tokenizer for the C# compiler
3 // This also implements the preprocessor
5 // Author: Miguel de Icaza (miguel@gnu.org)
7 // Licensed under the terms of the GNU GPL
9 // (C) 2001, 2002 Ximian, Inc (http://www.ximian.com)
10 // (C) 2004 Novell, Inc
15 * Make sure we accept the proper Unicode ranges, per the spec.
21 using System.Collections;
23 using System.Globalization;
24 using System.Reflection;
29 /// Tokenizer for C# source code.
32 public class Tokenizer : yyParser.yyInput
34 SeekableStreamReader reader;
35 public SourceFile ref_name;
36 public SourceFile file_name;
37 public int ref_line = 1;
40 public int current_token;
41 bool handle_get_set = false;
42 bool handle_remove_add = false;
43 bool handle_assembly = false;
44 bool handle_constraints = false;
45 bool handle_typeof = false;
48 // XML documentation buffer. The save point is used to divide
49 // comments on types and comments on members.
51 StringBuilder xml_comment_buffer;
54 // See comment on XmlCommentState enumeration.
56 XmlCommentState xmlDocState = XmlCommentState.Allowed;
59 // Whether tokens have been seen on this line
61 bool tokens_seen = false;
64 // Whether a token has been seen on the file
65 // This is needed because `define' is not allowed to be used
66 // after a token has been seen.
68 bool any_token_seen = false;
70 static Hashtable tokenValues;
72 private static Hashtable TokenValueName
75 if (tokenValues == null)
76 tokenValues = GetTokenValueNameHash ();
82 private static Hashtable GetTokenValueNameHash ()
84 Type t = typeof (Token);
85 FieldInfo [] fields = t.GetFields ();
86 Hashtable hash = new Hashtable ();
87 foreach (FieldInfo field in fields) {
88 if (field.IsLiteral && field.IsStatic && field.FieldType == typeof (int))
89 hash.Add (field.GetValue (null), field.Name);
95 // Returns a verbose representation of the current location
97 public string location {
101 if (current_token == Token.ERROR)
102 det = "detail: " + error_details;
106 // return "Line: "+line+" Col: "+col + "\n" +
107 // "VirtLine: "+ref_line +
108 // " Token: "+current_token + " " + det;
109 string current_token_name = TokenValueName [current_token] as string;
110 if (current_token_name == null)
111 current_token_name = current_token.ToString ();
113 return String.Format ("{0} ({1},{2}), Token: {3} {4}", ref_name.Name,
121 public bool PropertyParsing {
123 return handle_get_set;
127 handle_get_set = value;
131 public bool AssemblyTargetParsing {
133 return handle_assembly;
137 handle_assembly = value;
141 public bool EventParsing {
143 return handle_remove_add;
147 handle_remove_add = value;
151 public bool ConstraintsParsing {
153 return handle_constraints;
157 handle_constraints = value;
161 public bool TypeOfParsing {
163 return handle_typeof;
167 handle_typeof = value;
171 public XmlCommentState doc_state {
172 get { return xmlDocState; }
174 if (value == XmlCommentState.Allowed) {
175 check_incorrect_doc_comment ();
176 consume_doc_comment ();
185 static CharArrayHashtable[] keywords;
186 static NumberStyles styles;
187 static NumberFormatInfo csharp_format_info;
190 // Values for the associated token returned
200 const int TAKING = 1;
201 const int TAKEN_BEFORE = 2;
202 const int ELSE_SEEN = 4;
203 const int PARENT_TAKING = 8;
204 const int REGION = 16;
207 // pre-processor if stack state:
211 static System.Text.StringBuilder string_builder;
213 const int max_id_size = 512;
214 static char [] id_builder = new char [max_id_size];
216 static CharArrayHashtable [] identifiers = new CharArrayHashtable [max_id_size + 1];
218 const int max_number_size = 128;
219 static char [] number_builder = new char [max_number_size];
220 static int number_pos;
223 // Details about the error encoutered by the tokenizer
225 string error_details;
227 public string error {
229 return error_details;
245 static void AddKeyword (string kw, int token) {
246 if (keywords [kw.Length] == null) {
247 keywords [kw.Length] = new CharArrayHashtable (kw.Length);
249 keywords [kw.Length] [kw.ToCharArray ()] = token;
252 static void InitTokens ()
254 keywords = new CharArrayHashtable [64];
256 AddKeyword ("__arglist", Token.ARGLIST);
257 AddKeyword ("abstract", Token.ABSTRACT);
258 AddKeyword ("as", Token.AS);
259 AddKeyword ("add", Token.ADD);
260 AddKeyword ("assembly", Token.ASSEMBLY);
261 AddKeyword ("base", Token.BASE);
262 AddKeyword ("bool", Token.BOOL);
263 AddKeyword ("break", Token.BREAK);
264 AddKeyword ("byte", Token.BYTE);
265 AddKeyword ("case", Token.CASE);
266 AddKeyword ("catch", Token.CATCH);
267 AddKeyword ("char", Token.CHAR);
268 AddKeyword ("checked", Token.CHECKED);
269 AddKeyword ("class", Token.CLASS);
270 AddKeyword ("const", Token.CONST);
271 AddKeyword ("continue", Token.CONTINUE);
272 AddKeyword ("decimal", Token.DECIMAL);
273 AddKeyword ("default", Token.DEFAULT);
274 AddKeyword ("delegate", Token.DELEGATE);
275 AddKeyword ("do", Token.DO);
276 AddKeyword ("double", Token.DOUBLE);
277 AddKeyword ("else", Token.ELSE);
278 AddKeyword ("enum", Token.ENUM);
279 AddKeyword ("event", Token.EVENT);
280 AddKeyword ("explicit", Token.EXPLICIT);
281 AddKeyword ("extern", Token.EXTERN);
282 AddKeyword ("false", Token.FALSE);
283 AddKeyword ("finally", Token.FINALLY);
284 AddKeyword ("fixed", Token.FIXED);
285 AddKeyword ("float", Token.FLOAT);
286 AddKeyword ("for", Token.FOR);
287 AddKeyword ("foreach", Token.FOREACH);
288 AddKeyword ("goto", Token.GOTO);
289 AddKeyword ("get", Token.GET);
290 AddKeyword ("if", Token.IF);
291 AddKeyword ("implicit", Token.IMPLICIT);
292 AddKeyword ("in", Token.IN);
293 AddKeyword ("int", Token.INT);
294 AddKeyword ("interface", Token.INTERFACE);
295 AddKeyword ("internal", Token.INTERNAL);
296 AddKeyword ("is", Token.IS);
297 AddKeyword ("lock", Token.LOCK);
298 AddKeyword ("long", Token.LONG);
299 AddKeyword ("namespace", Token.NAMESPACE);
300 AddKeyword ("new", Token.NEW);
301 AddKeyword ("null", Token.NULL);
302 AddKeyword ("object", Token.OBJECT);
303 AddKeyword ("operator", Token.OPERATOR);
304 AddKeyword ("out", Token.OUT);
305 AddKeyword ("override", Token.OVERRIDE);
306 AddKeyword ("params", Token.PARAMS);
307 AddKeyword ("private", Token.PRIVATE);
308 AddKeyword ("protected", Token.PROTECTED);
309 AddKeyword ("public", Token.PUBLIC);
310 AddKeyword ("readonly", Token.READONLY);
311 AddKeyword ("ref", Token.REF);
312 AddKeyword ("remove", Token.REMOVE);
313 AddKeyword ("return", Token.RETURN);
314 AddKeyword ("sbyte", Token.SBYTE);
315 AddKeyword ("sealed", Token.SEALED);
316 AddKeyword ("set", Token.SET);
317 AddKeyword ("short", Token.SHORT);
318 AddKeyword ("sizeof", Token.SIZEOF);
319 AddKeyword ("stackalloc", Token.STACKALLOC);
320 AddKeyword ("static", Token.STATIC);
321 AddKeyword ("string", Token.STRING);
322 AddKeyword ("struct", Token.STRUCT);
323 AddKeyword ("switch", Token.SWITCH);
324 AddKeyword ("this", Token.THIS);
325 AddKeyword ("throw", Token.THROW);
326 AddKeyword ("true", Token.TRUE);
327 AddKeyword ("try", Token.TRY);
328 AddKeyword ("typeof", Token.TYPEOF);
329 AddKeyword ("uint", Token.UINT);
330 AddKeyword ("ulong", Token.ULONG);
331 AddKeyword ("unchecked", Token.UNCHECKED);
332 AddKeyword ("unsafe", Token.UNSAFE);
333 AddKeyword ("ushort", Token.USHORT);
334 AddKeyword ("using", Token.USING);
335 AddKeyword ("virtual", Token.VIRTUAL);
336 AddKeyword ("void", Token.VOID);
337 AddKeyword ("volatile", Token.VOLATILE);
338 AddKeyword ("where", Token.WHERE);
339 AddKeyword ("while", Token.WHILE);
340 AddKeyword ("partial", Token.PARTIAL);
349 csharp_format_info = NumberFormatInfo.InvariantInfo;
350 styles = NumberStyles.Float;
352 string_builder = new System.Text.StringBuilder ();
355 int GetKeyword (char[] id, int id_len)
358 * Keywords are stored in an array of hashtables grouped by their
362 if ((id_len >= keywords.Length) || (keywords [id_len] == null))
364 object o = keywords [id_len] [id];
371 if (handle_get_set == false && (res == Token.GET || res == Token.SET))
373 if (handle_remove_add == false && (res == Token.REMOVE || res == Token.ADD))
375 if (handle_assembly == false && res == Token.ASSEMBLY)
377 if (handle_constraints == false && res == Token.WHERE)
384 public Location Location {
386 return new Location (ref_line);
390 void define (string def)
392 if (!RootContext.AllDefines.Contains (def)){
393 RootContext.AllDefines [def] = true;
395 if (defines.Contains (def))
397 defines [def] = true;
400 public Tokenizer (SeekableStreamReader input, SourceFile file, ArrayList defs)
402 this.ref_name = file;
403 this.file_name = file;
409 defines = new Hashtable ();
410 foreach (string def in defs)
414 xml_comment_buffer = new StringBuilder ();
417 // FIXME: This could be `Location.Push' but we have to
418 // find out why the MS compiler allows this
420 Mono.CSharp.Location.Push (file);
423 public static void Cleanup () {
427 static bool is_identifier_start_character (char c)
429 return (c >= 'A' && c <= 'Z') || (c >= 'a' && c <= 'z') || c == '_' || Char.IsLetter (c);
432 static bool is_identifier_part_character (char c)
434 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || (c >= '0' && c <= '9') || Char.IsLetter (c);
437 public static bool IsValidIdentifier (string s)
439 if (s == null || s.Length == 0)
442 if (!is_identifier_start_character (s [0]))
445 for (int i = 1; i < s.Length; i ++)
446 if (! is_identifier_part_character (s [i]))
452 bool parse_generic_dimension (out int dimension)
457 int the_token = token ();
458 if (the_token == Token.OP_GENERICS_GT)
460 else if (the_token == Token.COMMA) {
468 bool parse_less_than ()
471 int the_token = token ();
473 case Token.IDENTIFIER:
496 the_token = token ();
498 if (the_token == Token.OP_GENERICS_GT)
500 else if ((the_token == Token.COMMA) || (the_token == Token.DOT))
502 else if (the_token == Token.OP_GENERICS_LT) {
503 if (!parse_less_than ())
506 } else if (the_token == Token.OPEN_BRACKET) {
508 the_token = token ();
509 if (the_token == Token.CLOSE_BRACKET)
511 else if (the_token == Token.COMMA)
512 goto rank_specifiers;
519 bool parsing_less_than = false;
520 int parsing_generic_less_than = 0;
522 int is_punct (char c, ref bool doread)
531 return Token.OPEN_BRACE;
533 return Token.CLOSE_BRACE;
535 // To block doccomment inside attribute declaration.
536 if (doc_state == XmlCommentState.Allowed)
537 doc_state = XmlCommentState.NotAllowed;
538 return Token.OPEN_BRACKET;
540 return Token.CLOSE_BRACKET;
542 return Token.OPEN_PARENS;
544 if (deambiguate_close_parens == 0)
545 return Token.CLOSE_PARENS;
547 --deambiguate_close_parens;
549 // Save current position and parse next token.
550 int old = reader.Position;
551 int new_token = token ();
552 reader.Position = old;
555 if (new_token == Token.OPEN_PARENS)
556 return Token.CLOSE_PARENS_OPEN_PARENS;
557 else if (new_token == Token.MINUS)
558 return Token.CLOSE_PARENS_MINUS;
559 else if (IsCastToken (new_token))
560 return Token.CLOSE_PARENS_CAST;
562 return Token.CLOSE_PARENS_NO_CAST;
570 return Token.SEMICOLON;
578 if (parsing_generic_less_than++ > 0)
579 return Token.OP_GENERICS_LT;
581 int old = reader.Position;
584 if (parse_generic_dimension (out dimension)) {
586 return Token.GENERIC_DIMENSION;
588 reader.Position = old;
592 // Save current position and parse next token.
593 old = reader.Position;
594 bool is_generic_lt = parse_less_than ();
595 reader.Position = old;
599 parsing_generic_less_than++;
600 return Token.OP_GENERICS_LT;
602 parsing_generic_less_than = 0;
611 return Token.OP_SHIFT_LEFT_ASSIGN;
613 return Token.OP_SHIFT_LEFT;
614 } else if (d == '='){
619 } else if (c == '>') {
620 if (parsing_generic_less_than > 0) {
621 parsing_generic_less_than--;
622 return Token.OP_GENERICS_GT;
632 return Token.OP_SHIFT_RIGHT_ASSIGN;
634 return Token.OP_SHIFT_RIGHT;
635 } else if (d == '='){
648 t = Token.OP_ADD_ASSIGN;
658 t = Token.OP_SUB_ASSIGN;
687 } else if (d == '='){
689 return Token.OP_AND_ASSIGN;
691 return Token.BITWISE_AND;
698 } else if (d == '='){
700 return Token.OP_OR_ASSIGN;
702 return Token.BITWISE_OR;
708 return Token.OP_MULT_ASSIGN;
716 return Token.OP_DIV_ASSIGN;
724 return Token.OP_MOD_ASSIGN;
726 return Token.PERCENT;
732 return Token.OP_XOR_ASSIGN;
740 int deambiguate_close_parens = 0;
742 public void Deambiguate_CloseParens ()
745 deambiguate_close_parens++;
748 void Error_NumericConstantTooLong ()
750 Report.Error (1021, Location, "Numeric constant too long");
753 bool decimal_digits (int c)
756 bool seen_digits = false;
759 if (number_pos == max_number_size)
760 Error_NumericConstantTooLong ();
761 number_builder [number_pos++] = (char) c;
765 // We use peekChar2, because decimal_digits needs to do a
766 // 2-character look-ahead (5.ToString for example).
768 while ((d = peekChar2 ()) != -1){
769 if (d >= '0' && d <= '9'){
770 if (number_pos == max_number_size)
771 Error_NumericConstantTooLong ();
772 number_builder [number_pos++] = (char) d;
784 return (e >= '0' && e <= '9') || (e >= 'A' && e <= 'F') || (e >= 'a' && e <= 'f');
787 void hex_digits (int c)
790 number_builder [number_pos++] = (char) c;
794 int real_type_suffix (int c)
800 t = Token.LITERAL_FLOAT;
803 t = Token.LITERAL_DOUBLE;
806 t= Token.LITERAL_DECIMAL;
814 int integer_type_suffix (ulong ul, int c)
816 bool is_unsigned = false;
817 bool is_long = false;
820 bool scanning = true;
831 if (!is_unsigned && (RootContext.WarningLevel >= 4)){
833 // if we have not seen anything in between
836 Report.Warning (78, Location, "The 'l' suffix is easily confused with the digit '1' (use 'L' for clarity)");
839 // This goto statement causes the MS CLR 2.0 beta 1 csc to report an error, so
864 if (is_long && is_unsigned){
866 return Token.LITERAL_INTEGER;
867 } else if (is_unsigned){
868 // uint if possible, or ulong else.
870 if ((ul & 0xffffffff00000000) == 0)
875 // long if possible, ulong otherwise
876 if ((ul & 0x8000000000000000) != 0)
881 // int, uint, long or ulong in that order
882 if ((ul & 0xffffffff00000000) == 0){
885 if ((ui & 0x80000000) != 0)
890 if ((ul & 0x8000000000000000) != 0)
896 return Token.LITERAL_INTEGER;
900 // given `c' as the next char in the input decide whether
901 // we need to convert to a special type, and then choose
902 // the best representation for the integer
904 int adjust_int (int c)
908 ulong ul = (uint) (number_builder [0] - '0');
910 for (int i = 1; i < number_pos; i++){
911 ul = checked ((ul * 10) + ((uint)(number_builder [i] - '0')));
913 return integer_type_suffix (ul, c);
915 uint ui = (uint) (number_builder [0] - '0');
917 for (int i = 1; i < number_pos; i++){
918 ui = checked ((ui * 10) + ((uint)(number_builder [i] - '0')));
920 return integer_type_suffix (ui, c);
922 } catch (OverflowException) {
923 error_details = "Integral constant is too large";
924 Report.Error (1021, Location, error_details);
926 return Token.LITERAL_INTEGER;
928 catch (FormatException) {
929 Report.Error (1013, Location, "Invalid number");
931 return Token.LITERAL_INTEGER;
935 int adjust_real (int t)
937 string s = new String (number_builder, 0, number_pos);
940 case Token.LITERAL_DECIMAL:
942 val = System.Decimal.Parse (s, styles, csharp_format_info);
943 } catch (OverflowException) {
945 error_details = "Floating-point constant is outside the range of the type 'decimal'";
946 Report.Error (594, Location, error_details);
949 case Token.LITERAL_FLOAT:
951 val = (float) System.Double.Parse (s, styles, csharp_format_info);
952 } catch (OverflowException) {
954 error_details = "Floating-point constant is outside the range of the type 'float'";
955 Report.Error (594, Location, error_details);
959 case Token.LITERAL_DOUBLE:
961 t = Token.LITERAL_DOUBLE;
963 val = System.Double.Parse (s, styles, csharp_format_info);
964 } catch (OverflowException) {
966 error_details = "Floating-point constant is outside the range of the type 'double'";
967 Report.Error (594, Location, error_details);
980 while ((d = peekChar ()) != -1){
982 number_builder [number_pos++] = (char) d;
988 string s = new String (number_builder, 0, number_pos);
991 ul = System.UInt32.Parse (s, NumberStyles.HexNumber);
993 ul = System.UInt64.Parse (s, NumberStyles.HexNumber);
994 } catch (OverflowException){
995 error_details = "Integral constant is too large";
996 Report.Error (1021, Location, error_details);
998 return Token.LITERAL_INTEGER;
1001 return integer_type_suffix (ul, peekChar ());
1005 // Invoked if we know we have .digits or digits
1007 int is_number (int c)
1009 bool is_real = false;
1014 if (c >= '0' && c <= '9'){
1016 int peek = peekChar ();
1018 if (peek == 'x' || peek == 'X')
1019 return handle_hex ();
1026 // We need to handle the case of
1027 // "1.1" vs "1.string" (LITERAL_FLOAT vs NUMBER DOT IDENTIFIER)
1030 if (decimal_digits ('.')){
1036 return adjust_int (-1);
1040 if (c == 'e' || c == 'E'){
1042 if (number_pos == max_number_size)
1043 Error_NumericConstantTooLong ();
1044 number_builder [number_pos++] = 'e';
1048 if (number_pos == max_number_size)
1049 Error_NumericConstantTooLong ();
1050 number_builder [number_pos++] = '+';
1052 } else if (c == '-') {
1053 if (number_pos == max_number_size)
1054 Error_NumericConstantTooLong ();
1055 number_builder [number_pos++] = '-';
1058 if (number_pos == max_number_size)
1059 Error_NumericConstantTooLong ();
1060 number_builder [number_pos++] = '+';
1067 type = real_type_suffix (c);
1068 if (type == Token.NONE && !is_real){
1070 return adjust_int (c);
1074 if (type == Token.NONE){
1079 return adjust_real (type);
1081 Console.WriteLine ("This should not be reached");
1082 throw new Exception ("Is Number should never reach this point");
1086 // Accepts exactly count (4 or 8) hex, no more no less
1088 int getHex (int count, out bool error)
1093 int top = count != -1 ? count : 4;
1097 for (i = 0; i < top; i++){
1100 if (c >= '0' && c <= '9')
1101 c = (int) c - (int) '0';
1102 else if (c >= 'A' && c <= 'F')
1103 c = (int) c - (int) 'A' + 10;
1104 else if (c >= 'a' && c <= 'f')
1105 c = (int) c - (int) 'a' + 10;
1111 total = (total * 16) + c;
1113 int p = peekChar ();
1116 if (!is_hex ((char)p))
1157 v = getHex (-1, out error);
1162 v = getHex (4, out error);
1167 v = getHex (8, out error);
1172 Report.Error (1009, Location, "Unrecognized escape sequence in " + (char)d);
1181 if (putback_char != -1){
1182 int x = putback_char;
1187 return reader.Read ();
1192 if (putback_char != -1)
1193 return putback_char;
1194 putback_char = reader.Read ();
1195 return putback_char;
1200 if (putback_char != -1)
1201 return putback_char;
1202 return reader.Peek ();
1205 void putback (int c)
1207 if (putback_char != -1){
1208 Console.WriteLine ("Col: " + col);
1209 Console.WriteLine ("Row: " + line);
1210 Console.WriteLine ("Name: " + ref_name.Name);
1211 Console.WriteLine ("Current [{0}] putting back [{1}] ", putback_char, c);
1212 throw new Exception ("This should not happen putback on putback");
1217 public bool advance ()
1219 return peekChar () != -1;
1222 public Object Value {
1228 public Object value ()
1233 bool IsCastToken (int token)
1238 case Token.IDENTIFIER:
1239 case Token.LITERAL_INTEGER:
1240 case Token.LITERAL_FLOAT:
1241 case Token.LITERAL_DOUBLE:
1242 case Token.LITERAL_DECIMAL:
1243 case Token.LITERAL_CHARACTER:
1244 case Token.LITERAL_STRING:
1256 case Token.UNCHECKED:
1260 // These can be part of a member access
1280 current_token = xtoken ();
1282 if (current_token != Token.DEFAULT)
1283 return current_token;
1285 int c = consume_whitespace ();
1287 current_token = Token.ERROR;
1289 current_token = Token.DEFAULT_OPEN_PARENS;
1293 return current_token;
1296 static StringBuilder static_cmd_arg = new System.Text.StringBuilder ();
1298 void get_cmd_arg (out string cmd, out string arg)
1302 tokens_seen = false;
1304 static_cmd_arg.Length = 0;
1306 // skip over white space
1307 while ((c = getChar ()) != -1 && (c != '\n') && ((c == '\r') || (c == ' ') || (c == '\t')))
1310 while ((c != -1) && (c != '\n') && (c != ' ') && (c != '\t') && (c != '\r')){
1311 if (is_identifier_part_character ((char) c)){
1312 static_cmd_arg.Append ((char) c);
1320 cmd = static_cmd_arg.ToString ();
1326 } else if (c == '\r')
1329 // skip over white space
1330 while ((c = getChar ()) != -1 && (c != '\n') && ((c == '\r') || (c == ' ') || (c == '\t')))
1337 } else if (c == '\r'){
1342 static_cmd_arg.Length = 0;
1343 static_cmd_arg.Append ((char) c);
1345 while ((c = getChar ()) != -1 && (c != '\n') && (c != '\r')){
1346 static_cmd_arg.Append ((char) c);
1352 } else if (c == '\r')
1354 arg = static_cmd_arg.ToString ().Trim ();
1358 // Handles the #line directive
1360 bool PreProcessLine (string arg)
1365 if (arg == "default"){
1367 ref_name = file_name;
1368 Location.Push (ref_name);
1370 } else if (arg == "hidden"){
1372 // We ignore #line hidden
1380 if ((pos = arg.IndexOf (' ')) != -1 && pos != 0){
1381 ref_line = System.Int32.Parse (arg.Substring (0, pos));
1384 char [] quotes = { '\"' };
1386 string name = arg.Substring (pos). Trim (quotes);
1387 ref_name = Location.LookupFile (name);
1388 file_name.HasLineDirective = true;
1389 ref_name.HasLineDirective = true;
1390 Location.Push (ref_name);
1392 ref_line = System.Int32.Parse (arg);
1402 // Handles #define and #undef
1404 void PreProcessDefinition (bool is_define, string arg)
1406 if (arg == "" || arg == "true" || arg == "false"){
1407 Report.Error (1001, Location, "Missing identifer to pre-processor directive");
1411 char[] whitespace = { ' ', '\t' };
1412 if (arg.IndexOfAny (whitespace) != -1){
1413 Report.Error (1025, Location, "Single-line comment or end-of-line expected");
1417 if (!is_identifier_start_character (arg [0]))
1418 Report.Error (1001, Location, "Identifier expected: " + arg);
1420 foreach (char c in arg.Substring (1)){
1421 if (!is_identifier_part_character (c)){
1422 Report.Error (1001, Location, "Identifier expected: " + arg);
1428 if (defines == null)
1429 defines = new Hashtable ();
1432 if (defines == null)
1434 if (defines.Contains (arg))
1435 defines.Remove (arg);
1440 /// Handles #pragma directive
1442 void PreProcessPragma (string arg)
1444 const string disable = "warning disable";
1445 const string restore = "warning restore";
1447 if (arg == disable) {
1448 Report.RegisterWarningRegion (Location).WarningDisable (line);
1452 if (arg == restore) {
1453 Report.RegisterWarningRegion (Location).WarningEnable (line);
1457 if (arg.StartsWith (disable)) {
1458 int[] codes = ParseNumbers (arg.Substring (disable.Length));
1459 foreach (int code in codes) {
1461 Report.RegisterWarningRegion (Location).WarningDisable (Location, code);
1466 if (arg.StartsWith (restore)) {
1467 int[] codes = ParseNumbers (arg.Substring (restore.Length));
1468 foreach (int code in codes) {
1469 Report.RegisterWarningRegion (Location).WarningEnable (Location, code);
1477 int[] ParseNumbers (string text)
1479 string[] string_array = text.Split (',');
1480 int[] values = new int [string_array.Length];
1482 foreach (string string_code in string_array) {
1484 values[index++] = int.Parse (string_code, System.Globalization.CultureInfo.InvariantCulture);
1486 catch (FormatException) {
1487 Report.Warning (1692, Location, "Invalid number");
1493 bool eval_val (string s)
1500 if (defines == null)
1502 if (defines.Contains (s))
1508 bool pp_primary (ref string s)
1517 s = s.Substring (1);
1518 bool val = pp_expr (ref s);
1519 if (s.Length > 0 && s [0] == ')'){
1520 s = s.Substring (1);
1523 Error_InvalidDirective ();
1527 if (is_identifier_start_character (c)){
1533 if (is_identifier_part_character (c)){
1537 bool v = eval_val (s.Substring (0, j));
1538 s = s.Substring (j);
1541 bool vv = eval_val (s);
1546 Error_InvalidDirective ();
1550 bool pp_unary (ref string s)
1557 if (len > 1 && s [1] == '='){
1558 Error_InvalidDirective ();
1561 s = s.Substring (1);
1562 return ! pp_primary (ref s);
1564 return pp_primary (ref s);
1566 Error_InvalidDirective ();
1571 bool pp_eq (ref string s)
1573 bool va = pp_unary (ref s);
1579 if (len > 2 && s [1] == '='){
1580 s = s.Substring (2);
1581 return va == pp_unary (ref s);
1583 Error_InvalidDirective ();
1586 } else if (s [0] == '!' && len > 1 && s [1] == '='){
1587 s = s.Substring (2);
1589 return va != pp_unary (ref s);
1598 bool pp_and (ref string s)
1600 bool va = pp_eq (ref s);
1606 if (len > 2 && s [1] == '&'){
1607 s = s.Substring (2);
1608 return (va & pp_eq (ref s));
1610 Error_InvalidDirective ();
1619 // Evaluates an expression for `#if' or `#elif'
1621 bool pp_expr (ref string s)
1623 bool va = pp_and (ref s);
1630 if (len > 2 && s [1] == '|'){
1631 s = s.Substring (2);
1632 return va | pp_expr (ref s);
1634 Error_InvalidDirective ();
1643 bool eval (string s)
1645 bool v = pp_expr (ref s);
1648 Error_InvalidDirective ();
1655 void Error_InvalidDirective ()
1657 Report.Error (1517, Location, "Invalid pre-processor directive");
1660 void Error_UnexpectedDirective (string extra)
1664 "Unexpected processor directive (" + extra + ")");
1667 void Error_TokensSeen ()
1671 "Cannot define or undefine pre-processor symbols after a token in the file");
1675 // if true, then the code continues processing the code
1676 // if false, the code stays in a loop until another directive is
1679 bool handle_preprocessing_directive (bool caller_is_taking)
1682 bool region_directive = false;
1684 get_cmd_arg (out cmd, out arg);
1686 // Eat any trailing whitespaces and single-line comments
1687 if (arg.IndexOf ("//") != -1)
1688 arg = arg.Substring (0, arg.IndexOf ("//"));
1689 arg = arg.TrimEnd (' ', '\t');
1692 // The first group of pre-processing instructions is always processed
1696 if (RootContext.Version == LanguageVersion.ISO_1) {
1697 Report.FeatureIsNotStandardized (Location, "#pragma");
1698 return caller_is_taking;
1701 PreProcessPragma (arg);
1702 return caller_is_taking;
1705 if (!PreProcessLine (arg))
1708 "Argument to #line directive is missing or invalid");
1709 return caller_is_taking;
1712 region_directive = true;
1717 region_directive = true;
1722 Error_InvalidDirective ();
1725 bool taking = false;
1726 if (ifstack == null)
1727 ifstack = new Stack ();
1729 if (ifstack.Count == 0){
1732 int state = (int) ifstack.Peek ();
1733 if ((state & TAKING) != 0)
1737 if (eval (arg) && taking){
1738 int push = TAKING | TAKEN_BEFORE | PARENT_TAKING;
1739 if (region_directive)
1741 ifstack.Push (push);
1744 int push = (taking ? PARENT_TAKING : 0);
1745 if (region_directive)
1747 ifstack.Push (push);
1752 if (ifstack == null || ifstack.Count == 0){
1753 Error_UnexpectedDirective ("no #if for this #endif");
1756 int pop = (int) ifstack.Pop ();
1758 if (region_directive && ((pop & REGION) == 0))
1759 Report.Error (1027, Location, "#endif directive expected");
1760 else if (!region_directive && ((pop & REGION) != 0))
1761 Report.Error (1038, Location, "#endregion directive expected");
1763 if (ifstack.Count == 0)
1766 int state = (int) ifstack.Peek ();
1768 if ((state & TAKING) != 0)
1776 if (ifstack == null || ifstack.Count == 0){
1777 Error_UnexpectedDirective ("no #if for this #elif");
1780 int state = (int) ifstack.Peek ();
1782 if ((state & REGION) != 0) {
1783 Report.Error (1038, Location, "#endregion directive expected");
1787 if ((state & ELSE_SEEN) != 0){
1788 Error_UnexpectedDirective ("#elif not valid after #else");
1792 if ((state & (TAKEN_BEFORE | TAKING)) != 0)
1795 if (eval (arg) && ((state & PARENT_TAKING) != 0)){
1796 state = (int) ifstack.Pop ();
1797 ifstack.Push (state | TAKING | TAKEN_BEFORE);
1804 if (ifstack == null || ifstack.Count == 0){
1807 "Unexpected processor directive (no #if for this #else)");
1810 int state = (int) ifstack.Peek ();
1812 if ((state & REGION) != 0) {
1813 Report.Error (1038, Location, "#endregion directive expected");
1817 if ((state & ELSE_SEEN) != 0){
1818 Error_UnexpectedDirective ("#else within #else");
1825 if ((state & TAKEN_BEFORE) == 0){
1826 ret = ((state & PARENT_TAKING) != 0);
1835 ifstack.Push (state | ELSE_SEEN);
1842 // These are only processed if we are in a `taking' block
1844 if (!caller_is_taking)
1849 if (any_token_seen){
1850 Error_TokensSeen ();
1853 PreProcessDefinition (true, arg);
1857 if (any_token_seen){
1858 Error_TokensSeen ();
1861 PreProcessDefinition (false, arg);
1865 Report.Error (1029, Location, "#error: '" + arg + "'");
1869 Report.Warning (1030, Location, "#warning: '{0}'", arg);
1873 Report.Error (1024, Location, "Preprocessor directive expected (got: " + cmd + ")");
1878 private int consume_string (bool quoted)
1881 string_builder.Length = 0;
1883 while ((c = getChar ()) != -1){
1885 if (quoted && peekChar () == '"'){
1886 string_builder.Append ((char) c);
1890 val = string_builder.ToString ();
1891 return Token.LITERAL_STRING;
1897 Report.Error (1010, Location, "Newline in constant");
1909 string_builder.Append ((char) c);
1912 Report.Error (1039, Location, "Unterminated string literal");
1916 private int consume_identifier (int s)
1918 int res = consume_identifier (s, false);
1920 if (doc_state == XmlCommentState.Allowed)
1921 doc_state = XmlCommentState.NotAllowed;
1924 case Token.NAMESPACE:
1925 check_incorrect_doc_comment ();
1929 if (res == Token.PARTIAL) {
1930 // Save current position and parse next token.
1931 int old = reader.Position;
1932 int old_putback = putback_char;
1936 int next_token = token ();
1937 bool ok = (next_token == Token.CLASS) ||
1938 (next_token == Token.STRUCT) ||
1939 (next_token == Token.INTERFACE);
1941 reader.Position = old;
1942 putback_char = old_putback;
1948 return Token.IDENTIFIER;
1955 private int consume_identifier (int s, bool quoted)
1960 id_builder [0] = (char) s;
1962 while ((c = reader.Read ()) != -1) {
1963 if (is_identifier_part_character ((char) c)){
1964 if (pos == max_id_size){
1965 Report.Error (645, Location, "Identifier too long (limit is 512 chars)");
1969 id_builder [pos++] = (char) c;
1979 // Optimization: avoids doing the keyword lookup
1980 // on uppercase letters and _
1982 if (!quoted && (s >= 'a' || s == '_')){
1983 int keyword = GetKeyword (id_builder, pos);
1989 // Keep identifiers in an array of hashtables to avoid needless
1993 if (identifiers [pos] != null) {
1994 val = identifiers [pos][id_builder];
1996 return Token.IDENTIFIER;
2000 identifiers [pos] = new CharArrayHashtable (pos);
2002 val = new String (id_builder, 0, pos);
2004 char [] chars = new char [pos];
2005 Array.Copy (id_builder, chars, pos);
2007 identifiers [pos] [chars] = val;
2009 return Token.IDENTIFIER;
2012 int consume_whitespace ()
2015 bool doread = false;
2018 // Whether we have seen comments on the current line
2019 bool comments_seen = false;
2022 // optimization: eliminate col and implement #directive semantic correctly.
2023 for (;(c = getChar ()) != -1; col++) {
2028 col = (((col + 8) / 8) * 8) - 1;
2032 if (c == ' ' || c == '\f' || c == '\v' || c == 0xa0)
2036 if (peekChar () == '\n')
2042 any_token_seen |= tokens_seen;
2043 tokens_seen = false;
2047 // Handle double-slash comments.
2049 int d = peekChar ();
2053 if (RootContext.Documentation != null && peekChar () == '/') {
2055 // Don't allow ////.
2056 if ((d = peekChar ()) != '/') {
2057 if (doc_state == XmlCommentState.Allowed)
2058 handle_one_line_xml_comment ();
2059 else if (doc_state == XmlCommentState.NotAllowed)
2060 warn_incorrect_doc_comment ();
2063 while ((d = getChar ()) != -1 && (d != '\n') && d != '\r')
2070 any_token_seen |= tokens_seen;
2071 tokens_seen = false;
2073 } else if (d == '*'){
2075 bool docAppend = false;
2076 if (RootContext.Documentation != null && peekChar () == '*') {
2078 // But when it is /**/, just do nothing.
2079 if (peekChar () == '/') {
2083 if (doc_state == XmlCommentState.Allowed)
2085 else if (doc_state == XmlCommentState.NotAllowed)
2086 warn_incorrect_doc_comment ();
2089 int current_comment_start = 0;
2091 current_comment_start = xml_comment_buffer.Length;
2092 xml_comment_buffer.Append (Environment.NewLine);
2095 while ((d = getChar ()) != -1){
2096 if (d == '*' && peekChar () == '/'){
2099 comments_seen = true;
2103 xml_comment_buffer.Append ((char) d);
2109 any_token_seen |= tokens_seen;
2110 tokens_seen = false;
2114 update_formatted_doc_comment (current_comment_start);
2117 goto is_punct_label;
2126 any_token_seen |= tokens_seen;
2127 tokens_seen = false;
2128 comments_seen = false;
2132 /* For now, ignore pre-processor commands */
2133 // FIXME: In C# the '#' is not limited to appear
2134 // on the first column.
2138 if (tokens_seen || comments_seen) {
2139 error_details = "Preprocessor directives must appear as the first non-whitespace " +
2140 "character on a line.";
2142 Report.Error (1040, Location, error_details);
2149 cont = handle_preprocessing_directive (cont);
2157 bool skipping = false;
2158 for (;(c = getChar ()) != -1; col++){
2164 } else if (c == ' ' || c == '\t' || c == '\v' || c == '\r' || c == 0xa0)
2168 if (c == '#' && !skipping)
2171 any_token_seen |= tokens_seen;
2172 tokens_seen = false;
2174 Report.Error (1027, Location, "#endif/#endregion expected");
2184 public int xtoken ()
2187 bool doread = false;
2191 // optimization: eliminate col and implement #directive semantic correctly.
2193 c = consume_whitespace ();
2197 if (is_identifier_start_character ((char)c)){
2199 return consume_identifier (c);
2203 if ((t = is_punct ((char)c, ref doread)) != Token.ERROR){
2212 if (c >= '0' && c <= '9'){
2214 return is_number (c);
2219 int peek = peekChar ();
2220 if (peek >= '0' && peek <= '9')
2221 return is_number (c);
2226 return consume_string (false);
2232 error_details = "Empty character literal";
2233 Report.Error (1011, Location, error_details);
2239 val = new System.Char ();
2244 error_details = "Too many characters in character literal";
2245 Report.Error (1012, Location, error_details);
2247 // Try to recover, read until newline or next "'"
2248 while ((c = getChar ()) != -1){
2249 if (c == '\n' || c == '\''){
2259 return Token.LITERAL_CHARACTER;
2266 return consume_string (true);
2267 } else if (is_identifier_start_character ((char) c)){
2268 return consume_identifier (c, true);
2270 Report.Error (1033, Location, "'@' must be followed by string constant or identifier");
2275 error_details = "Preprocessor directives must appear as the first non-whitespace " +
2276 "character on a line.";
2278 Report.Error (1040, Location, error_details);
2283 error_details = ((char)c).ToString ();
2289 // Handles one line xml comment
2291 private void handle_one_line_xml_comment ()
2294 while ((c = peekChar ()) == ' ')
2295 getChar (); // skip heading whitespaces.
2296 while ((c = peekChar ()) != -1 && c != '\n' && c != '\r') {
2298 xml_comment_buffer.Append ((char) getChar ());
2300 if (c == '\r' || c == '\n')
2301 xml_comment_buffer.Append (Environment.NewLine);
2305 // Remove heading "*" in Javadoc-like xml documentation.
2307 private void update_formatted_doc_comment (int current_comment_start)
2309 int length = xml_comment_buffer.Length - current_comment_start;
2310 string [] lines = xml_comment_buffer.ToString (
2311 current_comment_start,
2312 length).Replace ("\r", "").Split ('\n');
2314 // The first line starts with /**, thus it is not target
2315 // for the format check.
2316 for (int i = 1; i < lines.Length; i++) {
2317 string s = lines [i];
2318 int idx = s.IndexOf ('*');
2321 if (i < lines.Length - 1)
2325 head = s.Substring (0, idx);
2326 foreach (char c in head)
2329 lines [i] = s.Substring (idx + 1);
2331 xml_comment_buffer.Remove (current_comment_start, length);
2332 xml_comment_buffer.Insert (current_comment_start, String.Join (Environment.NewLine, lines));
2336 // Checks if there was incorrect doc comments and raise
2339 public void check_incorrect_doc_comment ()
2341 if (xml_comment_buffer.Length > 0)
2342 warn_incorrect_doc_comment ();
2346 // Raises a warning when tokenizer found incorrect doccomment
2349 private void warn_incorrect_doc_comment ()
2351 doc_state = XmlCommentState.Error;
2352 // in csc, it is 'XML comment is not placed on a valid
2353 // language element'. But that does not make sense.
2354 Report.Warning (1587, 2, Location, "XML comment is placed on an invalid language element which can not accept it.");
2358 // Consumes the saved xml comment lines (if any)
2359 // as for current target member or type.
2361 public string consume_doc_comment ()
2363 if (xml_comment_buffer.Length > 0) {
2364 string ret = xml_comment_buffer.ToString ();
2365 xml_comment_buffer.Length = 0;
2371 public void cleanup ()
2373 if (ifstack != null && ifstack.Count >= 1) {
2374 int state = (int) ifstack.Pop ();
2375 if ((state & REGION) != 0)
2376 Report.Error (1038, Location, "#endregion directive expected");
2378 Report.Error (1027, Location, "#endif directive expected");
2385 // Indicates whether it accepts XML documentation or not.
2387 public enum XmlCommentState {
2388 // comment is allowed in this state.
2390 // comment is not allowed in this state.
2392 // once comments appeared when it is NotAllowed, then the
2393 // state is changed to it, until the state is changed to