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;
46 // XML documentation buffer. The save point is used to divide
47 // comments on types and comments on members.
49 StringBuilder xml_comment_buffer;
52 // See comment on XmlCommentState enumeration.
54 XmlCommentState xmlDocState = XmlCommentState.Allowed;
57 // Whether tokens have been seen on this line
59 bool tokens_seen = false;
62 // Whether a token has been seen on the file
63 // This is needed because `define' is not allowed to be used
64 // after a token has been seen.
66 bool any_token_seen = false;
68 static Hashtable tokenValues;
70 private static Hashtable TokenValueName
73 if (tokenValues == null)
74 tokenValues = GetTokenValueNameHash ();
80 private static Hashtable GetTokenValueNameHash ()
82 Type t = typeof (Token);
83 FieldInfo [] fields = t.GetFields ();
84 Hashtable hash = new Hashtable ();
85 foreach (FieldInfo field in fields) {
86 if (field.IsLiteral && field.IsStatic && field.FieldType == typeof (int))
87 hash.Add (field.GetValue (null), field.Name);
93 // Returns a verbose representation of the current location
95 public string location {
99 if (current_token == Token.ERROR)
100 det = "detail: " + error_details;
104 // return "Line: "+line+" Col: "+col + "\n" +
105 // "VirtLine: "+ref_line +
106 // " Token: "+current_token + " " + det;
107 string current_token_name = TokenValueName [current_token] as string;
108 if (current_token_name == null)
109 current_token_name = current_token.ToString ();
111 return String.Format ("{0} ({1},{2}), Token: {3} {4}", ref_name.Name,
119 public bool PropertyParsing {
121 return handle_get_set;
125 handle_get_set = value;
129 public bool AssemblyTargetParsing {
131 return handle_assembly;
135 handle_assembly = value;
139 public bool EventParsing {
141 return handle_remove_add;
145 handle_remove_add = value;
149 public XmlCommentState doc_state {
150 get { return xmlDocState; }
152 if (value == XmlCommentState.Allowed) {
153 check_incorrect_doc_comment ();
154 consume_doc_comment ();
164 static CharArrayHashtable[] keywords;
165 static Hashtable keywordStrings = new Hashtable ();
166 static NumberStyles styles;
167 static NumberFormatInfo csharp_format_info;
170 // Values for the associated token returned
180 const int TAKING = 1;
181 const int TAKEN_BEFORE = 2;
182 const int ELSE_SEEN = 4;
183 const int PARENT_TAKING = 8;
184 const int REGION = 16;
187 // pre-processor if stack state:
191 static System.Text.StringBuilder string_builder;
193 const int max_id_size = 512;
194 static char [] id_builder = new char [max_id_size];
196 static CharArrayHashtable [] identifiers = new CharArrayHashtable [max_id_size + 1];
198 const int max_number_size = 512;
199 static char [] number_builder = new char [max_number_size];
200 static int number_pos;
203 // Details about the error encoutered by the tokenizer
205 string error_details;
207 public string error {
209 return error_details;
225 static void AddKeyword (string kw, int token) {
226 keywordStrings.Add (kw, kw);
227 if (keywords [kw.Length] == null) {
228 keywords [kw.Length] = new CharArrayHashtable (kw.Length);
230 keywords [kw.Length] [kw.ToCharArray ()] = token;
233 static void InitTokens ()
235 keywords = new CharArrayHashtable [64];
237 AddKeyword ("__arglist", Token.ARGLIST);
238 AddKeyword ("abstract", Token.ABSTRACT);
239 AddKeyword ("as", Token.AS);
240 AddKeyword ("add", Token.ADD);
241 AddKeyword ("assembly", Token.ASSEMBLY);
242 AddKeyword ("base", Token.BASE);
243 AddKeyword ("bool", Token.BOOL);
244 AddKeyword ("break", Token.BREAK);
245 AddKeyword ("byte", Token.BYTE);
246 AddKeyword ("case", Token.CASE);
247 AddKeyword ("catch", Token.CATCH);
248 AddKeyword ("char", Token.CHAR);
249 AddKeyword ("checked", Token.CHECKED);
250 AddKeyword ("class", Token.CLASS);
251 AddKeyword ("const", Token.CONST);
252 AddKeyword ("continue", Token.CONTINUE);
253 AddKeyword ("decimal", Token.DECIMAL);
254 AddKeyword ("default", Token.DEFAULT);
255 AddKeyword ("delegate", Token.DELEGATE);
256 AddKeyword ("do", Token.DO);
257 AddKeyword ("double", Token.DOUBLE);
258 AddKeyword ("else", Token.ELSE);
259 AddKeyword ("enum", Token.ENUM);
260 AddKeyword ("event", Token.EVENT);
261 AddKeyword ("explicit", Token.EXPLICIT);
262 AddKeyword ("extern", Token.EXTERN);
263 AddKeyword ("false", Token.FALSE);
264 AddKeyword ("finally", Token.FINALLY);
265 AddKeyword ("fixed", Token.FIXED);
266 AddKeyword ("float", Token.FLOAT);
267 AddKeyword ("for", Token.FOR);
268 AddKeyword ("foreach", Token.FOREACH);
269 AddKeyword ("goto", Token.GOTO);
270 AddKeyword ("get", Token.GET);
271 AddKeyword ("if", Token.IF);
272 AddKeyword ("implicit", Token.IMPLICIT);
273 AddKeyword ("in", Token.IN);
274 AddKeyword ("int", Token.INT);
275 AddKeyword ("interface", Token.INTERFACE);
276 AddKeyword ("internal", Token.INTERNAL);
277 AddKeyword ("is", Token.IS);
278 AddKeyword ("lock", Token.LOCK);
279 AddKeyword ("long", Token.LONG);
280 AddKeyword ("namespace", Token.NAMESPACE);
281 AddKeyword ("new", Token.NEW);
282 AddKeyword ("null", Token.NULL);
283 AddKeyword ("object", Token.OBJECT);
284 AddKeyword ("operator", Token.OPERATOR);
285 AddKeyword ("out", Token.OUT);
286 AddKeyword ("override", Token.OVERRIDE);
287 AddKeyword ("params", Token.PARAMS);
288 AddKeyword ("private", Token.PRIVATE);
289 AddKeyword ("protected", Token.PROTECTED);
290 AddKeyword ("public", Token.PUBLIC);
291 AddKeyword ("readonly", Token.READONLY);
292 AddKeyword ("ref", Token.REF);
293 AddKeyword ("remove", Token.REMOVE);
294 AddKeyword ("return", Token.RETURN);
295 AddKeyword ("sbyte", Token.SBYTE);
296 AddKeyword ("sealed", Token.SEALED);
297 AddKeyword ("set", Token.SET);
298 AddKeyword ("short", Token.SHORT);
299 AddKeyword ("sizeof", Token.SIZEOF);
300 AddKeyword ("stackalloc", Token.STACKALLOC);
301 AddKeyword ("static", Token.STATIC);
302 AddKeyword ("string", Token.STRING);
303 AddKeyword ("struct", Token.STRUCT);
304 AddKeyword ("switch", Token.SWITCH);
305 AddKeyword ("this", Token.THIS);
306 AddKeyword ("throw", Token.THROW);
307 AddKeyword ("true", Token.TRUE);
308 AddKeyword ("try", Token.TRY);
309 AddKeyword ("typeof", Token.TYPEOF);
310 AddKeyword ("uint", Token.UINT);
311 AddKeyword ("ulong", Token.ULONG);
312 AddKeyword ("unchecked", Token.UNCHECKED);
313 AddKeyword ("unsafe", Token.UNSAFE);
314 AddKeyword ("ushort", Token.USHORT);
315 AddKeyword ("using", Token.USING);
316 AddKeyword ("virtual", Token.VIRTUAL);
317 AddKeyword ("void", Token.VOID);
318 AddKeyword ("volatile", Token.VOLATILE);
319 AddKeyword ("while", Token.WHILE);
320 AddKeyword ("partial", Token.PARTIAL);
329 csharp_format_info = NumberFormatInfo.InvariantInfo;
330 styles = NumberStyles.Float;
332 string_builder = new System.Text.StringBuilder ();
335 int GetKeyword (char[] id, int id_len)
338 * Keywords are stored in an array of hashtables grouped by their
342 if ((id_len >= keywords.Length) || (keywords [id_len] == null))
344 object o = keywords [id_len] [id];
351 if (handle_get_set == false && (res == Token.GET || res == Token.SET))
353 if (handle_remove_add == false && (res == Token.REMOVE || res == Token.ADD))
355 if (handle_assembly == false && res == Token.ASSEMBLY)
362 public Location Location {
364 return new Location (ref_line);
368 void define (string def)
370 if (!RootContext.AllDefines.Contains (def)){
371 RootContext.AllDefines [def] = true;
373 if (defines.Contains (def))
375 defines [def] = true;
378 public Tokenizer (SeekableStreamReader input, SourceFile file, ArrayList defs)
380 this.ref_name = file;
381 this.file_name = file;
387 defines = new Hashtable ();
388 foreach (string def in defs)
392 xml_comment_buffer = new StringBuilder ();
395 // FIXME: This could be `Location.Push' but we have to
396 // find out why the MS compiler allows this
398 Mono.CSharp.Location.Push (file);
401 static bool is_identifier_start_character (char c)
403 return (c >= 'A' && c <= 'Z') || (c >= 'a' && c <= 'z') || c == '_' || Char.IsLetter (c);
406 static bool is_identifier_part_character (char c)
408 return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_' || (c >= '0' && c <= '9') || Char.IsLetter (c);
411 public static bool IsKeyword (string s)
413 return keywordStrings [s] != null;
416 public static bool IsValidIdentifier (string s)
418 if (s == null || s.Length == 0)
421 if (!is_identifier_start_character (s [0]))
424 for (int i = 1; i < s.Length; i ++)
425 if (! is_identifier_part_character (s [i]))
431 int is_punct (char c, ref bool doread)
440 return Token.OPEN_BRACE;
442 return Token.CLOSE_BRACE;
444 // To block doccomment inside attribute declaration.
445 if (doc_state == XmlCommentState.Allowed)
446 doc_state = XmlCommentState.NotAllowed;
447 return Token.OPEN_BRACKET;
449 return Token.CLOSE_BRACKET;
451 return Token.OPEN_PARENS;
453 if (deambiguate_close_parens == 0)
454 return Token.CLOSE_PARENS;
456 --deambiguate_close_parens;
458 // Save current position and parse next token.
459 int old = reader.Position;
460 int new_token = token ();
461 reader.Position = old;
464 if (new_token == Token.OPEN_PARENS)
465 return Token.CLOSE_PARENS_OPEN_PARENS;
466 else if (new_token == Token.MINUS)
467 return Token.CLOSE_PARENS_MINUS;
468 else if (IsCastToken (new_token))
469 return Token.CLOSE_PARENS_CAST;
471 return Token.CLOSE_PARENS_NO_CAST;
479 return Token.SEMICOLON;
492 t = Token.OP_ADD_ASSIGN;
502 t = Token.OP_SUB_ASSIGN;
531 } else if (d == '='){
533 return Token.OP_AND_ASSIGN;
535 return Token.BITWISE_AND;
542 } else if (d == '='){
544 return Token.OP_OR_ASSIGN;
546 return Token.BITWISE_OR;
552 return Token.OP_MULT_ASSIGN;
560 return Token.OP_DIV_ASSIGN;
568 return Token.OP_MOD_ASSIGN;
570 return Token.PERCENT;
576 return Token.OP_XOR_ASSIGN;
588 return Token.OP_SHIFT_LEFT_ASSIGN;
590 return Token.OP_SHIFT_LEFT;
591 } else if (d == '='){
605 return Token.OP_SHIFT_RIGHT_ASSIGN;
607 return Token.OP_SHIFT_RIGHT;
608 } else if (d == '='){
617 int deambiguate_close_parens = 0;
619 public void Deambiguate_CloseParens ()
622 deambiguate_close_parens++;
625 void Error_NumericConstantTooLong ()
627 Report.Error (1021, Location, "Numeric constant too long");
630 bool decimal_digits (int c)
633 bool seen_digits = false;
636 if (number_pos == max_number_size)
637 Error_NumericConstantTooLong ();
638 number_builder [number_pos++] = (char) c;
642 // We use peekChar2, because decimal_digits needs to do a
643 // 2-character look-ahead (5.ToString for example).
645 while ((d = peekChar2 ()) != -1){
646 if (d >= '0' && d <= '9'){
647 if (number_pos == max_number_size)
648 Error_NumericConstantTooLong ();
649 number_builder [number_pos++] = (char) d;
661 return (e >= '0' && e <= '9') || (e >= 'A' && e <= 'F') || (e >= 'a' && e <= 'f');
664 void hex_digits (int c)
667 number_builder [number_pos++] = (char) c;
671 int real_type_suffix (int c)
677 t = Token.LITERAL_FLOAT;
680 t = Token.LITERAL_DOUBLE;
683 t= Token.LITERAL_DECIMAL;
691 int integer_type_suffix (ulong ul, int c)
693 bool is_unsigned = false;
694 bool is_long = false;
697 bool scanning = true;
708 if (!is_unsigned && (RootContext.WarningLevel >= 4)){
710 // if we have not seen anything in between
713 Report.Warning (78, Location, "The 'l' suffix is easily confused with the digit '1' (use 'L' for clarity)");
716 // This goto statement causes the MS CLR 2.0 beta 1 csc to report an error, so
741 if (is_long && is_unsigned){
743 return Token.LITERAL_INTEGER;
744 } else if (is_unsigned){
745 // uint if possible, or ulong else.
747 if ((ul & 0xffffffff00000000) == 0)
752 // long if possible, ulong otherwise
753 if ((ul & 0x8000000000000000) != 0)
758 // int, uint, long or ulong in that order
759 if ((ul & 0xffffffff00000000) == 0){
762 if ((ui & 0x80000000) != 0)
767 if ((ul & 0x8000000000000000) != 0)
773 return Token.LITERAL_INTEGER;
777 // given `c' as the next char in the input decide whether
778 // we need to convert to a special type, and then choose
779 // the best representation for the integer
781 int adjust_int (int c)
785 ulong ul = (uint) (number_builder [0] - '0');
787 for (int i = 1; i < number_pos; i++){
788 ul = checked ((ul * 10) + ((uint)(number_builder [i] - '0')));
790 return integer_type_suffix (ul, c);
792 uint ui = (uint) (number_builder [0] - '0');
794 for (int i = 1; i < number_pos; i++){
795 ui = checked ((ui * 10) + ((uint)(number_builder [i] - '0')));
797 return integer_type_suffix (ui, c);
799 } catch (OverflowException) {
800 error_details = "Integral constant is too large";
801 Report.Error (1021, Location, error_details);
803 return Token.LITERAL_INTEGER;
807 int adjust_real (int t)
809 string s = new String (number_builder, 0, number_pos);
812 case Token.LITERAL_DECIMAL:
814 val = System.Decimal.Parse (s, styles, csharp_format_info);
815 } catch (OverflowException) {
817 error_details = "Floating-point constant is outside the range of the type 'decimal'";
818 Report.Error (594, Location, error_details);
821 case Token.LITERAL_FLOAT:
823 val = (float) System.Double.Parse (s, styles, csharp_format_info);
824 } catch (OverflowException) {
826 error_details = "Floating-point constant is outside the range of the type 'float'";
827 Report.Error (594, Location, error_details);
831 case Token.LITERAL_DOUBLE:
833 t = Token.LITERAL_DOUBLE;
835 val = System.Double.Parse (s, styles, csharp_format_info);
836 } catch (OverflowException) {
838 error_details = "Floating-point constant is outside the range of the type 'double'";
839 Report.Error (594, Location, error_details);
852 while ((d = peekChar ()) != -1){
854 number_builder [number_pos++] = (char) d;
860 string s = new String (number_builder, 0, number_pos);
863 ul = System.UInt32.Parse (s, NumberStyles.HexNumber);
865 ul = System.UInt64.Parse (s, NumberStyles.HexNumber);
866 } catch (OverflowException){
867 error_details = "Integral constant is too large";
868 Report.Error (1021, Location, error_details);
870 return Token.LITERAL_INTEGER;
872 catch (FormatException) {
873 Report.Error (1013, Location, "Invalid number");
875 return Token.LITERAL_INTEGER;
878 return integer_type_suffix (ul, peekChar ());
882 // Invoked if we know we have .digits or digits
884 int is_number (int c)
886 bool is_real = false;
891 if (c >= '0' && c <= '9'){
893 int peek = peekChar ();
895 if (peek == 'x' || peek == 'X')
896 return handle_hex ();
903 // We need to handle the case of
904 // "1.1" vs "1.string" (LITERAL_FLOAT vs NUMBER DOT IDENTIFIER)
907 if (decimal_digits ('.')){
913 return adjust_int (-1);
917 if (c == 'e' || c == 'E'){
919 if (number_pos == max_number_size)
920 Error_NumericConstantTooLong ();
921 number_builder [number_pos++] = 'e';
925 if (number_pos == max_number_size)
926 Error_NumericConstantTooLong ();
927 number_builder [number_pos++] = '+';
929 } else if (c == '-') {
930 if (number_pos == max_number_size)
931 Error_NumericConstantTooLong ();
932 number_builder [number_pos++] = '-';
935 if (number_pos == max_number_size)
936 Error_NumericConstantTooLong ();
937 number_builder [number_pos++] = '+';
944 type = real_type_suffix (c);
945 if (type == Token.NONE && !is_real){
947 return adjust_int (c);
951 if (type == Token.NONE){
956 return adjust_real (type);
958 Console.WriteLine ("This should not be reached");
959 throw new Exception ("Is Number should never reach this point");
963 // Accepts exactly count (4 or 8) hex, no more no less
965 int getHex (int count, out bool error)
970 int top = count != -1 ? count : 4;
974 for (i = 0; i < top; i++){
977 if (c >= '0' && c <= '9')
978 c = (int) c - (int) '0';
979 else if (c >= 'A' && c <= 'F')
980 c = (int) c - (int) 'A' + 10;
981 else if (c >= 'a' && c <= 'f')
982 c = (int) c - (int) 'a' + 10;
988 total = (total * 16) + c;
993 if (!is_hex ((char)p))
1034 v = getHex (-1, out error);
1039 v = getHex (4, out error);
1044 v = getHex (8, out error);
1049 Report.Error (1009, Location, "Unrecognized escape sequence in " + (char)d);
1058 if (putback_char != -1){
1059 int x = putback_char;
1064 return reader.Read ();
1069 if (putback_char != -1)
1070 return putback_char;
1071 putback_char = reader.Read ();
1072 return putback_char;
1077 if (putback_char != -1)
1078 return putback_char;
1079 return reader.Peek ();
1082 void putback (int c)
1084 if (putback_char != -1){
1085 Console.WriteLine ("Col: " + col);
1086 Console.WriteLine ("Row: " + line);
1087 Console.WriteLine ("Name: " + ref_name.Name);
1088 Console.WriteLine ("Current [{0}] putting back [{1}] ", putback_char, c);
1089 throw new Exception ("This should not happen putback on putback");
1094 public bool advance ()
1096 return peekChar () != -1;
1099 public Object Value {
1105 public Object value ()
1110 bool IsCastToken (int token)
1115 case Token.IDENTIFIER:
1116 case Token.LITERAL_INTEGER:
1117 case Token.LITERAL_FLOAT:
1118 case Token.LITERAL_DOUBLE:
1119 case Token.LITERAL_DECIMAL:
1120 case Token.LITERAL_CHARACTER:
1121 case Token.LITERAL_STRING:
1133 case Token.UNCHECKED:
1137 // These can be part of a member access
1157 current_token = xtoken ();
1158 return current_token;
1161 static StringBuilder static_cmd_arg = new System.Text.StringBuilder ();
1163 void get_cmd_arg (out string cmd, out string arg)
1167 tokens_seen = false;
1169 static_cmd_arg.Length = 0;
1171 // skip over white space
1172 while ((c = getChar ()) != -1 && (c != '\n') && ((c == '\r') || (c == ' ') || (c == '\t')))
1176 while ((c != -1) && (c != '\n') && (c != ' ') && (c != '\t') && (c != '\r')){
1177 if (is_identifier_part_character ((char) c)){
1178 static_cmd_arg.Append ((char) c);
1186 cmd = static_cmd_arg.ToString ();
1192 } else if (c == '\r')
1195 // skip over white space
1196 while ((c = getChar ()) != -1 && (c != '\n') && ((c == '\r') || (c == ' ') || (c == '\t')))
1203 } else if (c == '\r'){
1208 static_cmd_arg.Length = 0;
1209 static_cmd_arg.Append ((char) c);
1211 while ((c = getChar ()) != -1 && (c != '\n') && (c != '\r')){
1212 static_cmd_arg.Append ((char) c);
1218 } else if (c == '\r')
1220 arg = static_cmd_arg.ToString ().Trim ();
1224 // Handles the #line directive
1226 bool PreProcessLine (string arg)
1231 if (arg == "default"){
1233 ref_name = file_name;
1234 Location.Push (ref_name);
1236 } else if (arg == "hidden"){
1238 // We ignore #line hidden
1246 if ((pos = arg.IndexOf (' ')) != -1 && pos != 0){
1247 ref_line = System.Int32.Parse (arg.Substring (0, pos));
1250 char [] quotes = { '\"' };
1252 string name = arg.Substring (pos). Trim (quotes);
1253 ref_name = Location.LookupFile (name);
1254 file_name.HasLineDirective = true;
1255 ref_name.HasLineDirective = true;
1256 Location.Push (ref_name);
1258 ref_line = System.Int32.Parse (arg);
1268 // Handles #define and #undef
1270 void PreProcessDefinition (bool is_define, string arg)
1272 if (arg == "" || arg == "true" || arg == "false"){
1273 Report.Error (1001, Location, "Missing identifer to pre-processor directive");
1277 char[] whitespace = { ' ', '\t' };
1278 if (arg.IndexOfAny (whitespace) != -1){
1279 Report.Error (1025, Location, "Single-line comment or end-of-line expected");
1283 if (!is_identifier_start_character (arg [0]))
1284 Report.Error (1001, Location, "Identifier expected: " + arg);
1286 foreach (char c in arg.Substring (1)){
1287 if (!is_identifier_part_character (c)){
1288 Report.Error (1001, Location, "Identifier expected: " + arg);
1294 if (defines == null)
1295 defines = new Hashtable ();
1298 if (defines == null)
1300 if (defines.Contains (arg))
1301 defines.Remove (arg);
1306 /// Handles #pragma directive
1308 void PreProcessPragma (string arg)
1310 const string warning = "warning";
1311 const string w_disable = "warning disable";
1312 const string w_restore = "warning restore";
1314 if (arg == w_disable) {
1315 Report.RegisterWarningRegion (Location).WarningDisable (line);
1319 if (arg == w_restore) {
1320 Report.RegisterWarningRegion (Location).WarningEnable (line);
1324 if (arg.StartsWith (w_disable)) {
1325 int[] codes = ParseNumbers (arg.Substring (w_disable.Length));
1326 foreach (int code in codes) {
1328 Report.RegisterWarningRegion (Location).WarningDisable (Location, code);
1333 if (arg.StartsWith (w_restore)) {
1334 int[] codes = ParseNumbers (arg.Substring (w_restore.Length));
1335 Hashtable w_table = Report.warning_ignore_table;
1336 foreach (int code in codes) {
1337 if (w_table != null && w_table.Contains (code))
1338 Report.Warning (1635, 1, Location, "Cannot restore warning 'CS{0:0000}' because it was disabled globally", code);
1339 Report.RegisterWarningRegion (Location).WarningEnable (Location, code);
1344 if (arg.StartsWith (warning)) {
1345 Report.Warning (1634, 1, Location, "Expected disable or restore");
1349 Report.Warning (1633, 1, Location, "Unrecognized #pragma directive");
1352 int[] ParseNumbers (string text)
1354 string[] string_array = text.Split (',');
1355 int[] values = new int [string_array.Length];
1357 foreach (string string_code in string_array) {
1359 values[index++] = int.Parse (string_code, System.Globalization.CultureInfo.InvariantCulture);
1361 catch (FormatException) {
1362 Report.Warning (1692, Location, "Invalid number");
1368 bool eval_val (string s)
1375 if (defines == null)
1377 if (defines.Contains (s))
1383 bool pp_primary (ref string s)
1392 s = s.Substring (1);
1393 bool val = pp_expr (ref s);
1394 if (s.Length > 0 && s [0] == ')'){
1395 s = s.Substring (1);
1398 Error_InvalidDirective ();
1402 if (is_identifier_start_character (c)){
1408 if (is_identifier_part_character (c)){
1412 bool v = eval_val (s.Substring (0, j));
1413 s = s.Substring (j);
1416 bool vv = eval_val (s);
1421 Error_InvalidDirective ();
1425 bool pp_unary (ref string s)
1432 if (len > 1 && s [1] == '='){
1433 Error_InvalidDirective ();
1436 s = s.Substring (1);
1437 return ! pp_primary (ref s);
1439 return pp_primary (ref s);
1441 Error_InvalidDirective ();
1446 bool pp_eq (ref string s)
1448 bool va = pp_unary (ref s);
1454 if (len > 2 && s [1] == '='){
1455 s = s.Substring (2);
1456 return va == pp_unary (ref s);
1458 Error_InvalidDirective ();
1461 } else if (s [0] == '!' && len > 1 && s [1] == '='){
1462 s = s.Substring (2);
1464 return va != pp_unary (ref s);
1473 bool pp_and (ref string s)
1475 bool va = pp_eq (ref s);
1481 if (len > 2 && s [1] == '&'){
1482 s = s.Substring (2);
1483 return (va & pp_and (ref s));
1485 Error_InvalidDirective ();
1494 // Evaluates an expression for `#if' or `#elif'
1496 bool pp_expr (ref string s)
1498 bool va = pp_and (ref s);
1505 if (len > 2 && s [1] == '|'){
1506 s = s.Substring (2);
1507 return va | pp_expr (ref s);
1509 Error_InvalidDirective ();
1518 bool eval (string s)
1520 bool v = pp_expr (ref s);
1523 Error_InvalidDirective ();
1530 void Error_InvalidDirective ()
1532 Report.Error (1517, Location, "Invalid pre-processor directive");
1535 void Error_UnexpectedDirective (string extra)
1539 "Unexpected processor directive (" + extra + ")");
1542 void Error_TokensSeen ()
1546 "Cannot define or undefine pre-processor symbols after a token in the file");
1550 // if true, then the code continues processing the code
1551 // if false, the code stays in a loop until another directive is
1554 bool handle_preprocessing_directive (bool caller_is_taking)
1557 bool region_directive = false;
1559 get_cmd_arg (out cmd, out arg);
1561 // Eat any trailing whitespaces and single-line comments
1562 if (arg.IndexOf ("//") != -1)
1563 arg = arg.Substring (0, arg.IndexOf ("//"));
1564 arg = arg.TrimEnd (' ', '\t');
1567 // The first group of pre-processing instructions is always processed
1571 if (RootContext.Version == LanguageVersion.ISO_1) {
1572 Report.FeatureIsNotStandardized (Location, "#pragma");
1573 return caller_is_taking;
1576 PreProcessPragma (arg);
1577 return caller_is_taking;
1580 if (!PreProcessLine (arg))
1583 "Argument to #line directive is missing or invalid");
1584 return caller_is_taking;
1587 region_directive = true;
1592 region_directive = true;
1597 Error_InvalidDirective ();
1600 bool taking = false;
1601 if (ifstack == null)
1602 ifstack = new Stack ();
1604 if (ifstack.Count == 0){
1607 int state = (int) ifstack.Peek ();
1608 if ((state & TAKING) != 0)
1612 if (eval (arg) && taking){
1613 int push = TAKING | TAKEN_BEFORE | PARENT_TAKING;
1614 if (region_directive)
1616 ifstack.Push (push);
1619 int push = (taking ? PARENT_TAKING : 0);
1620 if (region_directive)
1622 ifstack.Push (push);
1627 if (ifstack == null || ifstack.Count == 0){
1628 Error_UnexpectedDirective ("no #if for this #endif");
1631 int pop = (int) ifstack.Pop ();
1633 if (region_directive && ((pop & REGION) == 0))
1634 Report.Error (1027, Location, "#endif directive expected");
1635 else if (!region_directive && ((pop & REGION) != 0))
1636 Report.Error (1038, Location, "#endregion directive expected");
1638 if (ifstack.Count == 0)
1641 int state = (int) ifstack.Peek ();
1643 if ((state & TAKING) != 0)
1651 if (ifstack == null || ifstack.Count == 0){
1652 Error_UnexpectedDirective ("no #if for this #elif");
1655 int state = (int) ifstack.Peek ();
1657 if ((state & REGION) != 0) {
1658 Report.Error (1038, Location, "#endregion directive expected");
1662 if ((state & ELSE_SEEN) != 0){
1663 Error_UnexpectedDirective ("#elif not valid after #else");
1667 if ((state & (TAKEN_BEFORE | TAKING)) != 0)
1670 if (eval (arg) && ((state & PARENT_TAKING) != 0)){
1671 state = (int) ifstack.Pop ();
1672 ifstack.Push (state | TAKING | TAKEN_BEFORE);
1679 if (ifstack == null || ifstack.Count == 0){
1682 "Unexpected processor directive (no #if for this #else)");
1685 int state = (int) ifstack.Peek ();
1687 if ((state & REGION) != 0) {
1688 Report.Error (1038, Location, "#endregion directive expected");
1692 if ((state & ELSE_SEEN) != 0){
1693 Error_UnexpectedDirective ("#else within #else");
1700 if ((state & TAKEN_BEFORE) == 0){
1701 ret = ((state & PARENT_TAKING) != 0);
1710 ifstack.Push (state | ELSE_SEEN);
1717 // These are only processed if we are in a `taking' block
1719 if (!caller_is_taking)
1724 if (any_token_seen){
1725 Error_TokensSeen ();
1728 PreProcessDefinition (true, arg);
1732 if (any_token_seen){
1733 Error_TokensSeen ();
1736 PreProcessDefinition (false, arg);
1740 Report.Error (1029, Location, "#error: '" + arg + "'");
1744 Report.Warning (1030, Location, "#warning: '{0}'", arg);
1748 Report.Error (1024, Location, "Preprocessor directive expected (got: " + cmd + ")");
1753 private int consume_string (bool quoted)
1756 string_builder.Length = 0;
1758 while ((c = getChar ()) != -1){
1760 if (quoted && peekChar () == '"'){
1761 string_builder.Append ((char) c);
1765 val = string_builder.ToString ();
1766 return Token.LITERAL_STRING;
1772 Report.Error (1010, Location, "Newline in constant");
1784 string_builder.Append ((char) c);
1787 Report.Error (1039, Location, "Unterminated string literal");
1791 private int consume_identifier (int s)
1793 int res = consume_identifier (s, false);
1795 if (doc_state == XmlCommentState.Allowed)
1796 doc_state = XmlCommentState.NotAllowed;
1799 case Token.NAMESPACE:
1800 check_incorrect_doc_comment ();
1804 if (res == Token.PARTIAL) {
1805 // Save current position and parse next token.
1806 int old = reader.Position;
1807 int old_putback = putback_char;
1811 int next_token = token ();
1812 bool ok = (next_token == Token.CLASS) ||
1813 (next_token == Token.STRUCT) ||
1814 (next_token == Token.INTERFACE);
1816 reader.Position = old;
1817 putback_char = old_putback;
1823 return Token.IDENTIFIER;
1830 private int consume_identifier (int s, bool quoted)
1835 id_builder [0] = (char) s;
1837 while ((c = reader.Read ()) != -1) {
1838 if (is_identifier_part_character ((char) c)){
1839 if (pos == max_id_size){
1840 Report.Error (645, Location, "Identifier too long (limit is 512 chars)");
1844 id_builder [pos++] = (char) c;
1854 // Optimization: avoids doing the keyword lookup
1855 // on uppercase letters and _
1857 if (!quoted && (s >= 'a' || s == '_')){
1858 int keyword = GetKeyword (id_builder, pos);
1864 // Keep identifiers in an array of hashtables to avoid needless
1868 if (identifiers [pos] != null) {
1869 val = identifiers [pos][id_builder];
1871 return Token.IDENTIFIER;
1875 identifiers [pos] = new CharArrayHashtable (pos);
1877 val = new String (id_builder, 0, pos);
1879 char [] chars = new char [pos];
1880 Array.Copy (id_builder, chars, pos);
1882 identifiers [pos] [chars] = val;
1884 return Token.IDENTIFIER;
1887 public int xtoken ()
1890 bool doread = false;
1893 // Whether we have seen comments on the current line
1894 bool comments_seen = false;
1897 // optimization: eliminate col and implement #directive semantic correctly.
1898 for (;(c = getChar ()) != -1; col++) {
1903 col = (((col + 8) / 8) * 8) - 1;
1907 if (c == ' ' || c == '\f' || c == '\v' || c == 0xa0)
1911 if (peekChar () == '\n')
1917 any_token_seen |= tokens_seen;
1918 tokens_seen = false;
1919 comments_seen = false;
1923 // Handle double-slash comments.
1925 int d = peekChar ();
1929 if (RootContext.Documentation != null && peekChar () == '/') {
1931 // Don't allow ////.
1932 if ((d = peekChar ()) != '/') {
1933 if (doc_state == XmlCommentState.Allowed)
1934 handle_one_line_xml_comment ();
1935 else if (doc_state == XmlCommentState.NotAllowed)
1936 warn_incorrect_doc_comment ();
1939 while ((d = getChar ()) != -1 && (d != '\n') && d != '\r')
1946 any_token_seen |= tokens_seen;
1947 tokens_seen = false;
1948 comments_seen = false;
1950 } else if (d == '*'){
1952 bool docAppend = false;
1953 if (RootContext.Documentation != null && peekChar () == '*') {
1955 // But when it is /**/, just do nothing.
1956 if (peekChar () == '/') {
1960 if (doc_state == XmlCommentState.Allowed)
1962 else if (doc_state == XmlCommentState.NotAllowed)
1963 warn_incorrect_doc_comment ();
1966 int current_comment_start = 0;
1968 current_comment_start = xml_comment_buffer.Length;
1969 xml_comment_buffer.Append (Environment.NewLine);
1972 Location start_location = Location;
1974 while ((d = getChar ()) != -1){
1975 if (d == '*' && peekChar () == '/'){
1978 comments_seen = true;
1982 xml_comment_buffer.Append ((char) d);
1988 any_token_seen |= tokens_seen;
1989 tokens_seen = false;
1991 // Reset 'comments_seen' just to be consistent.
1992 // It doesn't matter either way, here.
1994 comments_seen = false;
1998 Report.Error (1035, start_location, "End-of-file found, '*/' expected");
2001 update_formatted_doc_comment (current_comment_start);
2004 goto is_punct_label;
2008 if (is_identifier_start_character ((char)c)){
2010 return consume_identifier (c);
2014 if ((t = is_punct ((char)c, ref doread)) != Token.ERROR){
2028 any_token_seen |= tokens_seen;
2029 tokens_seen = false;
2030 comments_seen = false;
2034 if (c >= '0' && c <= '9'){
2036 return is_number (c);
2041 int peek = peekChar ();
2042 if (peek >= '0' && peek <= '9')
2043 return is_number (c);
2047 /* For now, ignore pre-processor commands */
2048 // FIXME: In C# the '#' is not limited to appear
2049 // on the first column.
2053 if (tokens_seen || comments_seen) {
2054 error_details = "Preprocessor directives must appear as the first non-whitespace " +
2055 "character on a line.";
2057 Report.Error (1040, Location, error_details);
2064 cont = handle_preprocessing_directive (cont);
2072 bool skipping = false;
2073 for (;(c = getChar ()) != -1; col++){
2079 } else if (c == ' ' || c == '\t' || c == '\v' || c == '\r' || c == 0xa0)
2083 if (c == '#' && !skipping)
2086 any_token_seen |= tokens_seen;
2087 tokens_seen = false;
2089 Report.Error (1027, Location, "#endif/#endregion expected");
2094 return consume_string (false);
2100 error_details = "Empty character literal";
2101 Report.Error (1011, Location, error_details);
2107 val = new System.Char ();
2112 error_details = "Too many characters in character literal";
2113 Report.Error (1012, Location, error_details);
2115 // Try to recover, read until newline or next "'"
2116 while ((c = getChar ()) != -1){
2117 if (c == '\n' || c == '\''){
2128 return Token.LITERAL_CHARACTER;
2135 return consume_string (true);
2136 } else if (is_identifier_start_character ((char) c)){
2137 return consume_identifier (c, true);
2139 Report.Error (1646, Location, "Keyword, identifier, or string expected after verbatim specifier: @");
2143 error_details = ((char)c).ToString ();
2152 // Handles one line xml comment
2154 private void handle_one_line_xml_comment ()
2157 while ((c = peekChar ()) == ' ')
2158 getChar (); // skip heading whitespaces.
2159 while ((c = peekChar ()) != -1 && c != '\n' && c != '\r') {
2161 xml_comment_buffer.Append ((char) getChar ());
2163 if (c == '\r' || c == '\n')
2164 xml_comment_buffer.Append (Environment.NewLine);
2168 // Remove heading "*" in Javadoc-like xml documentation.
2170 private void update_formatted_doc_comment (int current_comment_start)
2172 int length = xml_comment_buffer.Length - current_comment_start;
2173 string [] lines = xml_comment_buffer.ToString (
2174 current_comment_start,
2175 length).Replace ("\r", "").Split ('\n');
2177 // The first line starts with /**, thus it is not target
2178 // for the format check.
2179 for (int i = 1; i < lines.Length; i++) {
2180 string s = lines [i];
2181 int idx = s.IndexOf ('*');
2184 if (i < lines.Length - 1)
2188 head = s.Substring (0, idx);
2189 foreach (char c in head)
2192 lines [i] = s.Substring (idx + 1);
2194 xml_comment_buffer.Remove (current_comment_start, length);
2195 xml_comment_buffer.Insert (current_comment_start, String.Join (Environment.NewLine, lines));
2199 // Checks if there was incorrect doc comments and raise
2202 public void check_incorrect_doc_comment ()
2204 if (xml_comment_buffer.Length > 0)
2205 warn_incorrect_doc_comment ();
2209 // Raises a warning when tokenizer found incorrect doccomment
2212 private void warn_incorrect_doc_comment ()
2214 doc_state = XmlCommentState.Error;
2215 // in csc, it is 'XML comment is not placed on a valid
2216 // language element'. But that does not make sense.
2217 Report.Warning (1587, 2, Location, "XML comment is placed on an invalid language element which can not accept it.");
2221 // Consumes the saved xml comment lines (if any)
2222 // as for current target member or type.
2224 public string consume_doc_comment ()
2226 if (xml_comment_buffer.Length > 0) {
2227 string ret = xml_comment_buffer.ToString ();
2228 xml_comment_buffer.Length = 0;
2234 public void cleanup ()
2236 if (ifstack != null && ifstack.Count >= 1) {
2237 int state = (int) ifstack.Pop ();
2238 if ((state & REGION) != 0)
2239 Report.Error (1038, Location, "#endregion directive expected");
2241 Report.Error (1027, "#endif directive expected");
2248 // Indicates whether it accepts XML documentation or not.
2250 public enum XmlCommentState {
2251 // comment is allowed in this state.
2253 // comment is not allowed in this state.
2255 // once comments appeared when it is NotAllowed, then the
2256 // state is changed to it, until the state is changed to