1 ------------------------------------------------------------------------------
3 -- GNAT COMPILER COMPONENTS --
9 -- Copyright (C) 1992-2006, Free Software Foundation, Inc. --
11 -- GNAT is free software; you can redistribute it and/or modify it under --
12 -- terms of the GNU General Public License as published by the Free Soft- --
13 -- ware Foundation; either version 2, or (at your option) any later ver- --
14 -- sion. GNAT is distributed in the hope that it will be useful, but WITH- --
15 -- OUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY --
16 -- or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License --
17 -- for more details. You should have received a copy of the GNU General --
18 -- Public License distributed with GNAT; see file COPYING. If not, write --
19 -- to the Free Software Foundation, 51 Franklin Street, Fifth Floor, --
20 -- Boston, MA 02110-1301, USA. --
22 -- GNAT was originally developed by the GNAT team at New York University. --
23 -- Extensive contributions were provided by Ada Core Technologies Inc. --
25 ------------------------------------------------------------------------------
27 with Csets; use Csets;
28 with Err_Vars; use Err_Vars;
29 with Hostparm; use Hostparm;
30 with Namet; use Namet;
32 with Scans; use Scans;
33 with Sinput; use Sinput;
34 with Snames; use Snames;
35 with Stringt; use Stringt;
36 with Stylesw; use Stylesw;
37 with Uintp; use Uintp;
38 with Urealp; use Urealp;
39 with Widechar; use Widechar;
42 with System.WCh_Con; use System.WCh_Con;
44 with GNAT.UTF_32; use GNAT.UTF_32;
49 -- Make control characters visible
51 Special_Characters : array (Character) of Boolean := (others => False);
52 -- For characters that are Special token, the value is True
54 Comment_Is_Token : Boolean := False;
55 -- True if comments are tokens
57 End_Of_Line_Is_Token : Boolean := False;
58 -- True if End_Of_Line is a token
60 -----------------------
61 -- Local Subprograms --
62 -----------------------
64 procedure Accumulate_Token_Checksum;
65 pragma Inline (Accumulate_Token_Checksum);
67 procedure Accumulate_Checksum (C : Character);
68 pragma Inline (Accumulate_Checksum);
69 -- This routine accumulates the checksum given character C. During the
70 -- scanning of a source file, this routine is called with every character
71 -- in the source, excluding blanks, and all control characters (except
72 -- that ESC is included in the checksum). Upper case letters not in string
73 -- literals are folded by the caller. See Sinput spec for the documentation
74 -- of the checksum algorithm. Note: checksum values are only used if we
75 -- generate code, so it is not necessary to worry about making the right
76 -- sequence of calls in any error situation.
78 procedure Accumulate_Checksum (C : Char_Code);
79 pragma Inline (Accumulate_Checksum);
80 -- This version is identical, except that the argument, C, is a character
81 -- code value instead of a character. This is used when wide characters
82 -- are scanned. We use the character code rather than the ASCII characters
83 -- so that the checksum is independent of wide character encoding method.
85 procedure Initialize_Checksum;
86 pragma Inline (Initialize_Checksum);
87 -- Initialize checksum value
89 -------------------------
90 -- Accumulate_Checksum --
91 -------------------------
93 procedure Accumulate_Checksum (C : Character) is
95 System.CRC32.Update (System.CRC32.CRC32 (Checksum), C);
96 end Accumulate_Checksum;
98 procedure Accumulate_Checksum (C : Char_Code) is
101 Accumulate_Checksum (Character'Val (C / 2 ** 24));
102 Accumulate_Checksum (Character'Val ((C / 2 ** 16) mod 256));
103 Accumulate_Checksum (Character'Val ((C / 256) mod 256));
105 Accumulate_Checksum (Character'Val (C / 256));
108 Accumulate_Checksum (Character'Val (C mod 256));
109 end Accumulate_Checksum;
111 -------------------------------
112 -- Accumulate_Token_Checksum --
113 -------------------------------
115 procedure Accumulate_Token_Checksum is
118 (System.CRC32.CRC32 (Checksum),
119 Character'Val (Token_Type'Pos (Token)));
120 end Accumulate_Token_Checksum;
122 ----------------------------
123 -- Determine_Token_Casing --
124 ----------------------------
126 function Determine_Token_Casing return Casing_Type is
128 return Determine_Casing (Source (Token_Ptr .. Scan_Ptr - 1));
129 end Determine_Token_Casing;
131 -------------------------
132 -- Initialize_Checksum --
133 -------------------------
135 procedure Initialize_Checksum is
137 System.CRC32.Initialize (System.CRC32.CRC32 (Checksum));
138 end Initialize_Checksum;
140 ------------------------
141 -- Initialize_Scanner --
142 ------------------------
144 procedure Initialize_Scanner (Index : Source_File_Index) is
146 -- Establish reserved words
148 Scans.Initialize_Ada_Keywords;
150 -- Initialize scan control variables
152 Current_Source_File := Index;
153 Source := Source_Text (Current_Source_File);
154 Scan_Ptr := Source_First (Current_Source_File);
156 Token_Ptr := Scan_Ptr;
157 Current_Line_Start := Scan_Ptr;
159 Token_Name := No_Name;
160 Start_Column := Set_Start_Column;
161 First_Non_Blank_Location := Scan_Ptr;
164 Wide_Char_Byte_Count := 0;
166 -- Do not call Scan, otherwise the License stuff does not work in Scn
168 end Initialize_Scanner;
170 ------------------------------
171 -- Reset_Special_Characters --
172 ------------------------------
174 procedure Reset_Special_Characters is
176 Special_Characters := (others => False);
177 end Reset_Special_Characters;
185 Start_Of_Comment : Source_Ptr;
186 -- Record start of comment position
188 Underline_Found : Boolean;
189 -- During scanning of an identifier, set to True if last character
190 -- scanned was an underline or other punctuation character. This
191 -- is used to flag the error of two underlines/punctuations in a
192 -- row or ending an identifier with a underline/punctuation. Here
193 -- punctuation means any UTF_32 character in the Unicode category
194 -- Punctuation,Connector.
197 -- Used to remember start of last wide character scanned
199 procedure Check_End_Of_Line;
200 -- Called when end of line encountered. Checks that line is not too
201 -- long, and that other style checks for the end of line are met.
203 function Double_Char_Token (C : Character) return Boolean;
204 -- This function is used for double character tokens like := or <>. It
205 -- checks if the character following Source (Scan_Ptr) is C, and if so
206 -- bumps Scan_Ptr past the pair of characters and returns True. A space
207 -- between the two characters is also recognized with an appropriate
208 -- error message being issued. If C is not present, False is returned.
209 -- Note that Double_Char_Token can only be used for tokens defined in
210 -- the Ada syntax (it's use for error cases like && is not appropriate
211 -- since we do not want a junk message for a case like &-space-&).
213 procedure Error_Illegal_Character;
214 -- Give illegal character error, Scan_Ptr points to character. On
215 -- return, Scan_Ptr is bumped past the illegal character.
217 procedure Error_Illegal_Wide_Character;
218 -- Give illegal wide character message. On return, Scan_Ptr is bumped
219 -- past the illegal character, which may still leave us pointing to
220 -- junk, not much we can do if the escape sequence is messed up!
222 procedure Error_Long_Line;
223 -- Signal error of excessively long line
225 procedure Error_No_Double_Underline;
226 -- Signal error of two underline or punctuation characters in a row.
227 -- Called with Scan_Ptr pointing to second underline/punctuation char.
230 -- This is the procedure for scanning out numeric literals. On entry,
231 -- Scan_Ptr points to the digit that starts the numeric literal (the
232 -- checksum for this character has not been accumulated yet). On return
233 -- Scan_Ptr points past the last character of the numeric literal, Token
234 -- and Token_Node are set appropriately, and the checksum is updated.
237 -- This is the procedure for scanning out string literals. On entry,
238 -- Scan_Ptr points to the opening string quote (the checksum for this
239 -- character has not been accumulated yet). On return Scan_Ptr points
240 -- past the closing quote of the string literal, Token and Token_Node
241 -- are set appropriately, and the checksum is upated.
243 -----------------------
244 -- Check_End_Of_Line --
245 -----------------------
247 procedure Check_End_Of_Line is
248 Len : constant Int :=
250 Int (Current_Line_Start) -
251 Wide_Char_Byte_Count;
255 Style.Check_Line_Terminator (Len);
258 -- Deal with checking maximum line length
260 if Style_Check and Style_Check_Max_Line_Length then
261 Style.Check_Line_Max_Length (Len);
263 -- If style checking is inactive, check maximum line length against
266 elsif Len > Max_Line_Length then
270 -- Reset wide character byte count for next line
272 Wide_Char_Byte_Count := 0;
273 end Check_End_Of_Line;
275 -----------------------
276 -- Double_Char_Token --
277 -----------------------
279 function Double_Char_Token (C : Character) return Boolean is
281 if Source (Scan_Ptr + 1) = C then
282 Accumulate_Checksum (C);
283 Scan_Ptr := Scan_Ptr + 2;
286 elsif Source (Scan_Ptr + 1) = ' '
287 and then Source (Scan_Ptr + 2) = C
289 Scan_Ptr := Scan_Ptr + 1;
290 Error_Msg_S ("no space allowed here");
291 Scan_Ptr := Scan_Ptr + 2;
297 end Double_Char_Token;
299 -----------------------------
300 -- Error_Illegal_Character --
301 -----------------------------
303 procedure Error_Illegal_Character is
305 Error_Msg_S ("illegal character");
306 Scan_Ptr := Scan_Ptr + 1;
307 end Error_Illegal_Character;
309 ----------------------------------
310 -- Error_Illegal_Wide_Character --
311 ----------------------------------
313 procedure Error_Illegal_Wide_Character is
315 Error_Msg ("illegal wide character", Wptr);
316 end Error_Illegal_Wide_Character;
318 ---------------------
319 -- Error_Long_Line --
320 ---------------------
322 procedure Error_Long_Line is
325 ("this line is too long",
326 Current_Line_Start + Source_Ptr (Max_Line_Length));
329 -------------------------------
330 -- Error_No_Double_Underline --
331 -------------------------------
333 procedure Error_No_Double_Underline is
335 Underline_Found := False;
337 -- There are four cases, and we special case the messages
339 if Source (Scan_Ptr) = '_' then
340 if Source (Scan_Ptr - 1) = '_' then
342 ("two consecutive underlines not permitted");
345 ("underline cannot follow punctuation character");
349 if Source (Scan_Ptr - 1) = '_' then
351 ("punctuation character cannot follow underline");
354 ("two consecutive punctuation characters not permitted");
357 end Error_No_Double_Underline;
366 -- Current source program character
368 Base_Char : Character;
369 -- Either # or : (character at start of based number)
375 -- Value of base in Uint format
378 -- Value of integer scanned by Scan_Integer in Uint format
381 -- Value of integer in numeric value being scanned
384 -- Scale value for real literal
387 -- Scale in Uint format
389 Exponent_Is_Negative : Boolean;
390 -- Set true for negative exponent
392 Extended_Digit_Value : Int;
393 -- Extended digit value
395 Point_Scanned : Boolean;
396 -- Flag for decimal point scanned in numeric literal
398 -----------------------
399 -- Local Subprograms --
400 -----------------------
402 procedure Error_Digit_Expected;
403 -- Signal error of bad digit, Scan_Ptr points to the location at
404 -- which the digit was expected on input, and is unchanged on return.
406 procedure Scan_Integer;
407 -- Procedure to scan integer literal. On entry, Scan_Ptr points to a
408 -- digit, on exit Scan_Ptr points past the last character of the
411 -- For each digit encountered, UI_Int_Value is multiplied by 10, and
412 -- the value of the digit added to the result. In addition, the
413 -- value in Scale is decremented by one for each actual digit
416 --------------------------
417 -- Error_Digit_Expected --
418 --------------------------
420 procedure Error_Digit_Expected is
422 Error_Msg_S ("digit expected");
423 end Error_Digit_Expected;
429 procedure Scan_Integer is
431 -- Next character scanned
434 C := Source (Scan_Ptr);
436 -- Loop through digits (allowing underlines)
439 Accumulate_Checksum (C);
441 UI_Int_Value * 10 + (Character'Pos (C) - Character'Pos ('0'));
442 Scan_Ptr := Scan_Ptr + 1;
444 C := Source (Scan_Ptr);
446 -- Case of underline encountered
450 -- We do not accumulate the '_' in the checksum, so that
451 -- 1_234 is equivalent to 1234, and does not trigger
452 -- compilation for "minimal recompilation" (gnatmake -m).
455 Scan_Ptr := Scan_Ptr + 1;
456 C := Source (Scan_Ptr);
458 Error_No_Double_Underline;
461 if C not in '0' .. '9' then
462 Error_Digit_Expected;
467 exit when C not in '0' .. '9';
472 -- Start of Processing for Nlit
477 UI_Int_Value := Uint_0;
480 Point_Scanned := False;
481 UI_Num_Value := UI_Int_Value;
483 -- Various possibilities now for continuing the literal are period,
484 -- E/e (for exponent), or :/# (for based literal).
487 C := Source (Scan_Ptr);
491 -- Scan out point, but do not scan past .. which is a range
492 -- sequence, and must not be eaten up scanning a numeric literal.
494 while C = '.' and then Source (Scan_Ptr + 1) /= '.' loop
495 Accumulate_Checksum ('.');
497 if Point_Scanned then
498 Error_Msg_S ("duplicate point ignored");
501 Point_Scanned := True;
502 Scan_Ptr := Scan_Ptr + 1;
503 C := Source (Scan_Ptr);
505 if C not in '0' .. '9' then
507 ("real literal cannot end with point", Scan_Ptr - 1);
510 UI_Num_Value := UI_Int_Value;
514 -- Based literal case. The base is the value we already scanned.
515 -- In the case of colon, we insist that the following character
516 -- is indeed an extended digit or a period. This catches a number
517 -- of common errors, as well as catching the well known tricky
518 -- bug otherwise arising from "x : integer range 1 .. 10:= 6;"
521 or else (C = ':' and then
522 (Source (Scan_Ptr + 1) = '.'
524 Source (Scan_Ptr + 1) in '0' .. '9'
526 Source (Scan_Ptr + 1) in 'A' .. 'Z'
528 Source (Scan_Ptr + 1) in 'a' .. 'z'))
531 Obsolescent_Check (Scan_Ptr);
533 if Warn_On_Obsolescent_Feature then
535 ("use of "":"" is an obsolescent feature ('R'M 'J.2(3))?");
537 ("\use ""'#"" instead?");
541 Accumulate_Checksum (C);
543 UI_Base := UI_Int_Value;
545 if UI_Base < 2 or else UI_Base > 16 then
546 Error_Msg_SC ("base not 2-16");
550 Base := UI_To_Int (UI_Base);
551 Scan_Ptr := Scan_Ptr + 1;
553 -- Scan out extended integer [. integer]
555 C := Source (Scan_Ptr);
556 UI_Int_Value := Uint_0;
560 if C in '0' .. '9' then
561 Accumulate_Checksum (C);
562 Extended_Digit_Value :=
563 Int'(Character'Pos (C)) - Int'(Character'Pos ('0'));
565 elsif C in 'A' .. 'F' then
566 Accumulate_Checksum (Character'Val (Character'Pos (C) + 32));
567 Extended_Digit_Value :=
568 Int'(Character'Pos (C)) - Int'(Character'Pos ('A')) + 10;
570 elsif C in 'a' .. 'f' then
571 Accumulate_Checksum (C);
572 Extended_Digit_Value :=
573 Int'(Character'Pos (C)) - Int'(Character'Pos ('a')) + 10;
576 Error_Msg_S ("extended digit expected");
580 if Extended_Digit_Value >= Base then
581 Error_Msg_S ("digit '>= base");
584 UI_Int_Value := UI_Int_Value * UI_Base + Extended_Digit_Value;
586 Scan_Ptr := Scan_Ptr + 1;
587 C := Source (Scan_Ptr);
591 Accumulate_Checksum ('_');
592 Scan_Ptr := Scan_Ptr + 1;
593 C := Source (Scan_Ptr);
595 Error_No_Double_Underline;
599 Accumulate_Checksum ('.');
601 if Point_Scanned then
602 Error_Msg_S ("duplicate point ignored");
605 Scan_Ptr := Scan_Ptr + 1;
606 C := Source (Scan_Ptr);
607 Point_Scanned := True;
610 elsif C = Base_Char then
611 Accumulate_Checksum (C);
612 Scan_Ptr := Scan_Ptr + 1;
615 elsif C = '#' or else C = ':' then
616 Error_Msg_S ("based number delimiters must match");
617 Scan_Ptr := Scan_Ptr + 1;
620 elsif not Identifier_Char (C) then
621 if Base_Char = '#' then
622 Error_Msg_S ("missing '#");
624 Error_Msg_S ("missing ':");
632 UI_Num_Value := UI_Int_Value;
637 if not Point_Scanned then
641 UI_Scale := UI_From_Int (Scale);
644 if Source (Scan_Ptr) = 'e' or else Source (Scan_Ptr) = 'E' then
645 Accumulate_Checksum ('e');
646 Scan_Ptr := Scan_Ptr + 1;
647 Exponent_Is_Negative := False;
649 if Source (Scan_Ptr) = '+' then
650 Accumulate_Checksum ('+');
651 Scan_Ptr := Scan_Ptr + 1;
653 elsif Source (Scan_Ptr) = '-' then
654 Accumulate_Checksum ('-');
656 if not Point_Scanned then
658 ("negative exponent not allowed for integer literal");
660 Exponent_Is_Negative := True;
663 Scan_Ptr := Scan_Ptr + 1;
666 UI_Int_Value := Uint_0;
668 if Source (Scan_Ptr) in '0' .. '9' then
671 Error_Digit_Expected;
674 if Exponent_Is_Negative then
675 UI_Scale := UI_Scale - UI_Int_Value;
677 UI_Scale := UI_Scale + UI_Int_Value;
681 -- Case of real literal to be returned
683 if Point_Scanned then
684 Token := Tok_Real_Literal;
685 Real_Literal_Value :=
691 -- Case of integer literal to be returned
694 Token := Tok_Integer_Literal;
697 Int_Literal_Value := UI_Num_Value;
699 -- Avoid doing possibly expensive calculations in cases like
700 -- parsing 163E800_000# when semantics will not be done anyway.
701 -- This is especially useful when parsing garbled input.
703 elsif Operating_Mode /= Check_Syntax
704 and then (Serious_Errors_Detected = 0 or else Try_Semantics)
706 Int_Literal_Value := UI_Num_Value * UI_Base ** UI_Scale;
709 Int_Literal_Value := No_Uint;
713 Accumulate_Token_Checksum;
724 Delimiter : Character;
725 -- Delimiter (first character of string)
728 -- Current source program character
731 -- Current character code value
734 -- Error flag for Scan_Wide call
736 procedure Error_Bad_String_Char;
737 -- Signal bad character in string/character literal. On entry
738 -- Scan_Ptr points to the improper character encountered during the
739 -- scan. Scan_Ptr is not modified, so it still points to the bad
740 -- character on return.
742 procedure Error_Unterminated_String;
743 -- Procedure called if a line terminator character is encountered
744 -- during scanning a string, meaning that the string is not properly
747 procedure Set_String;
748 -- Procedure used to distinguish between string and operator symbol.
749 -- On entry the string has been scanned out, and its characters
750 -- start at Token_Ptr and end one character before Scan_Ptr. On exit
751 -- Token is set to Tok_String_Literal or Tok_Operator_Symbol as
752 -- appropriate, and Token_Node is appropriately initialized. In
753 -- addition, in the operator symbol case, Token_Name is
754 -- appropriately set.
756 ---------------------------
757 -- Error_Bad_String_Char --
758 ---------------------------
760 procedure Error_Bad_String_Char is
761 C : constant Character := Source (Scan_Ptr);
765 Error_Msg_S ("horizontal tab not allowed in string");
767 elsif C = VT or else C = FF then
768 Error_Msg_S ("format effector not allowed in string");
770 elsif C in Upper_Half_Character then
771 Error_Msg_S ("(Ada 83) upper half character not allowed");
774 Error_Msg_S ("control character not allowed in string");
776 end Error_Bad_String_Char;
778 -------------------------------
779 -- Error_Unterminated_String --
780 -------------------------------
782 procedure Error_Unterminated_String is
784 -- An interesting little refinement. Consider the following
787 -- A := "this is an unterminated string;
788 -- A := "this is an unterminated string &
789 -- P(A, "this is a parameter that didn't get terminated);
791 -- We fiddle a little to do slightly better placement in these
792 -- cases also if there is white space at the end of the line we
793 -- place the flag at the start of this white space, not at the
794 -- end. Note that we only have to test for blanks, since tabs
795 -- aren't allowed in strings in the first place and would have
796 -- caused an error message.
798 -- Two more cases that we treat specially are:
800 -- A := "this string uses the wrong terminator'
801 -- A := "this string uses the wrong terminator' &
803 -- In these cases we give a different error message as well
805 -- We actually reposition the scan pointer to the point where we
806 -- place the flag in these cases, since it seems a better bet on
807 -- the original intention.
809 while Source (Scan_Ptr - 1) = ' '
810 or else Source (Scan_Ptr - 1) = '&'
812 Scan_Ptr := Scan_Ptr - 1;
816 -- Check for case of incorrect string terminator, but single quote
817 -- is not considered incorrect if the opening terminator misused
818 -- a single quote (error message already given).
821 and then Source (Scan_Ptr - 1) = '''
825 ("incorrect string terminator character", Scan_Ptr - 1);
829 if Source (Scan_Ptr - 1) = ';' then
830 Scan_Ptr := Scan_Ptr - 1;
833 if Source (Scan_Ptr - 1) = ')' then
834 Scan_Ptr := Scan_Ptr - 1;
839 Error_Msg_S ("missing string quote");
840 end Error_Unterminated_String;
846 procedure Set_String is
847 Slen : constant Int := Int (Scan_Ptr - Token_Ptr - 2);
853 -- Token_Name is currently set to Error_Name. The following
854 -- section of code resets Token_Name to the proper Name_Op_xx
855 -- value if the string is a valid operator symbol, otherwise it is
856 -- left set to Error_Name.
859 C1 := Source (Token_Ptr + 1);
863 Token_Name := Name_Op_Eq;
866 Token_Name := Name_Op_Gt;
869 Token_Name := Name_Op_Lt;
872 Token_Name := Name_Op_Add;
875 Token_Name := Name_Op_Subtract;
878 Token_Name := Name_Op_Concat;
881 Token_Name := Name_Op_Multiply;
884 Token_Name := Name_Op_Divide;
891 C1 := Source (Token_Ptr + 1);
892 C2 := Source (Token_Ptr + 2);
894 if C1 = '*' and then C2 = '*' then
895 Token_Name := Name_Op_Expon;
900 Token_Name := Name_Op_Ne;
902 Token_Name := Name_Op_Le;
904 Token_Name := Name_Op_Ge;
907 elsif (C1 = 'O' or else C1 = 'o') and then -- OR
908 (C2 = 'R' or else C2 = 'r')
910 Token_Name := Name_Op_Or;
914 C1 := Source (Token_Ptr + 1);
915 C2 := Source (Token_Ptr + 2);
916 C3 := Source (Token_Ptr + 3);
918 if (C1 = 'A' or else C1 = 'a') and then -- AND
919 (C2 = 'N' or else C2 = 'n') and then
920 (C3 = 'D' or else C3 = 'd')
922 Token_Name := Name_Op_And;
924 elsif (C1 = 'A' or else C1 = 'a') and then -- ABS
925 (C2 = 'B' or else C2 = 'b') and then
926 (C3 = 'S' or else C3 = 's')
928 Token_Name := Name_Op_Abs;
930 elsif (C1 = 'M' or else C1 = 'm') and then -- MOD
931 (C2 = 'O' or else C2 = 'o') and then
932 (C3 = 'D' or else C3 = 'd')
934 Token_Name := Name_Op_Mod;
936 elsif (C1 = 'N' or else C1 = 'n') and then -- NOT
937 (C2 = 'O' or else C2 = 'o') and then
938 (C3 = 'T' or else C3 = 't')
940 Token_Name := Name_Op_Not;
942 elsif (C1 = 'R' or else C1 = 'r') and then -- REM
943 (C2 = 'E' or else C2 = 'e') and then
944 (C3 = 'M' or else C3 = 'm')
946 Token_Name := Name_Op_Rem;
948 elsif (C1 = 'X' or else C1 = 'x') and then -- XOR
949 (C2 = 'O' or else C2 = 'o') and then
950 (C3 = 'R' or else C3 = 'r')
952 Token_Name := Name_Op_Xor;
957 -- If it is an operator symbol, then Token_Name is set. If it is
958 -- some other string value, then Token_Name still contains
961 if Token_Name = Error_Name then
962 Token := Tok_String_Literal;
965 Token := Tok_Operator_Symbol;
969 -- Start of processing for Slit
972 -- On entry, Scan_Ptr points to the opening character of the string
973 -- which is either a percent, double quote, or apostrophe (single
974 -- quote). The latter case is an error detected by the character
977 Delimiter := Source (Scan_Ptr);
978 Accumulate_Checksum (Delimiter);
980 Scan_Ptr := Scan_Ptr + 1;
982 -- Loop to scan out characters of string literal
985 C := Source (Scan_Ptr);
987 if C = Delimiter then
988 Accumulate_Checksum (C);
989 Scan_Ptr := Scan_Ptr + 1;
990 exit when Source (Scan_Ptr) /= Delimiter;
991 Code := Get_Char_Code (C);
992 Accumulate_Checksum (C);
993 Scan_Ptr := Scan_Ptr + 1;
996 if C = '"' and then Delimiter = '%' then
998 ("quote not allowed in percent delimited string");
999 Code := Get_Char_Code (C);
1000 Scan_Ptr := Scan_Ptr + 1;
1003 and then Wide_Character_Encoding_Method
1004 in WC_ESC_Encoding_Method)
1005 or else (C in Upper_Half_Character
1006 and then Upper_Half_Encoding)
1008 and then Source (Scan_Ptr + 1) = '"'
1009 and then Identifier_Char (Source (Scan_Ptr + 2)))
1012 Scan_Wide (Source, Scan_Ptr, Code, Err);
1015 Error_Illegal_Wide_Character;
1016 Code := Get_Char_Code (' ');
1019 Accumulate_Checksum (Code);
1021 -- In Ada 95 mode we allow any wide characters in a string
1022 -- but in Ada 2005, the set of characters allowed has been
1023 -- restricted to graphic characters.
1025 if Ada_Version >= Ada_05
1026 and then Is_UTF_32_Non_Graphic (UTF_32 (Code))
1029 ("(Ada 2005) non-graphic character not permitted " &
1030 "in string literal", Wptr);
1034 Accumulate_Checksum (C);
1036 if C not in Graphic_Character then
1037 if C in Line_Terminator then
1038 Error_Unterminated_String;
1041 elsif C in Upper_Half_Character then
1042 if Ada_Version = Ada_83 then
1043 Error_Bad_String_Char;
1047 Error_Bad_String_Char;
1051 Code := Get_Char_Code (C);
1052 Scan_Ptr := Scan_Ptr + 1;
1056 Store_String_Char (Code);
1058 if not In_Character_Range (Code) then
1059 Wide_Character_Found := True;
1063 String_Literal_Id := End_String;
1068 -- Start of processing for Scan
1071 Prev_Token := Token;
1072 Prev_Token_Ptr := Token_Ptr;
1073 Token_Name := Error_Name;
1075 -- The following loop runs more than once only if a format effector
1076 -- (tab, vertical tab, form feed, line feed, carriage return) is
1077 -- encountered and skipped, or some error situation, such as an
1078 -- illegal character, is encountered.
1080 <<Scan_Next_Character>>
1083 -- Skip past blanks, loop is opened up for speed
1085 while Source (Scan_Ptr) = ' ' loop
1086 if Source (Scan_Ptr + 1) /= ' ' then
1087 Scan_Ptr := Scan_Ptr + 1;
1091 if Source (Scan_Ptr + 2) /= ' ' then
1092 Scan_Ptr := Scan_Ptr + 2;
1096 if Source (Scan_Ptr + 3) /= ' ' then
1097 Scan_Ptr := Scan_Ptr + 3;
1101 if Source (Scan_Ptr + 4) /= ' ' then
1102 Scan_Ptr := Scan_Ptr + 4;
1106 if Source (Scan_Ptr + 5) /= ' ' then
1107 Scan_Ptr := Scan_Ptr + 5;
1111 if Source (Scan_Ptr + 6) /= ' ' then
1112 Scan_Ptr := Scan_Ptr + 6;
1116 if Source (Scan_Ptr + 7) /= ' ' then
1117 Scan_Ptr := Scan_Ptr + 7;
1121 Scan_Ptr := Scan_Ptr + 8;
1124 -- We are now at a non-blank character, which is the first character
1125 -- of the token we will scan, and hence the value of Token_Ptr.
1127 Token_Ptr := Scan_Ptr;
1129 -- Here begins the main case statement which transfers control on the
1130 -- basis of the non-blank character we have encountered.
1132 case Source (Scan_Ptr) is
1134 -- Line terminator characters
1136 when CR | LF | FF | VT =>
1137 goto Scan_Line_Terminator;
1139 -- Horizontal tab, just skip past it
1142 if Style_Check then Style.Check_HT; end if;
1143 Scan_Ptr := Scan_Ptr + 1;
1145 -- End of file character, treated as an end of file only if it is
1146 -- the last character in the buffer, otherwise it is ignored.
1149 if Scan_Ptr = Source_Last (Current_Source_File) then
1151 if Style_Check then Style.Check_EOF; end if;
1155 Scan_Ptr := Scan_Ptr + 1;
1161 Accumulate_Checksum ('&');
1163 if Source (Scan_Ptr + 1) = '&' then
1164 Error_Msg_S ("'&'& should be `AND THEN`");
1165 Scan_Ptr := Scan_Ptr + 2;
1170 Scan_Ptr := Scan_Ptr + 1;
1171 Token := Tok_Ampersand;
1175 -- Asterisk (can be multiplication operator or double asterisk which
1176 -- is the exponentiation compound delimiter).
1179 Accumulate_Checksum ('*');
1181 if Source (Scan_Ptr + 1) = '*' then
1182 Accumulate_Checksum ('*');
1183 Scan_Ptr := Scan_Ptr + 2;
1184 Token := Tok_Double_Asterisk;
1188 Scan_Ptr := Scan_Ptr + 1;
1189 Token := Tok_Asterisk;
1193 -- Colon, which can either be an isolated colon, or part of an
1194 -- assignment compound delimiter.
1197 Accumulate_Checksum (':');
1199 if Double_Char_Token ('=') then
1200 Token := Tok_Colon_Equal;
1201 if Style_Check then Style.Check_Colon_Equal; end if;
1204 elsif Source (Scan_Ptr + 1) = '-'
1205 and then Source (Scan_Ptr + 2) /= '-'
1207 Token := Tok_Colon_Equal;
1208 Error_Msg (":- should be :=", Scan_Ptr);
1209 Scan_Ptr := Scan_Ptr + 2;
1213 Scan_Ptr := Scan_Ptr + 1;
1215 if Style_Check then Style.Check_Colon; end if;
1222 Accumulate_Checksum ('(');
1223 Scan_Ptr := Scan_Ptr + 1;
1224 Token := Tok_Left_Paren;
1225 if Style_Check then Style.Check_Left_Paren; end if;
1231 if Source (Scan_Ptr + 1) = '"' then
1232 goto Scan_Wide_Character;
1235 Error_Msg_S ("illegal character, replaced by ""(""");
1236 Scan_Ptr := Scan_Ptr + 1;
1237 Token := Tok_Left_Paren;
1244 Error_Msg_S ("illegal character, replaced by ""(""");
1245 Scan_Ptr := Scan_Ptr + 1;
1246 Token := Tok_Left_Paren;
1252 Accumulate_Checksum (',');
1253 Scan_Ptr := Scan_Ptr + 1;
1255 if Style_Check then Style.Check_Comma; end if;
1258 -- Dot, which is either an isolated period, or part of a double dot
1259 -- compound delimiter sequence. We also check for the case of a
1260 -- digit following the period, to give a better error message.
1263 Accumulate_Checksum ('.');
1265 if Double_Char_Token ('.') then
1266 Token := Tok_Dot_Dot;
1267 if Style_Check then Style.Check_Dot_Dot; end if;
1270 elsif Source (Scan_Ptr + 1) in '0' .. '9' then
1271 Error_Msg_S ("numeric literal cannot start with point");
1272 Scan_Ptr := Scan_Ptr + 1;
1275 Scan_Ptr := Scan_Ptr + 1;
1280 -- Equal, which can either be an equality operator, or part of the
1281 -- arrow (=>) compound delimiter.
1284 Accumulate_Checksum ('=');
1286 if Double_Char_Token ('>') then
1288 if Style_Check then Style.Check_Arrow; end if;
1291 elsif Source (Scan_Ptr + 1) = '=' then
1292 Error_Msg_S ("== should be =");
1293 Scan_Ptr := Scan_Ptr + 1;
1296 Scan_Ptr := Scan_Ptr + 1;
1300 -- Greater than, which can be a greater than operator, greater than
1301 -- or equal operator, or first character of a right label bracket.
1304 Accumulate_Checksum ('>');
1306 if Double_Char_Token ('=') then
1307 Token := Tok_Greater_Equal;
1310 elsif Double_Char_Token ('>') then
1311 Token := Tok_Greater_Greater;
1315 Scan_Ptr := Scan_Ptr + 1;
1316 Token := Tok_Greater;
1320 -- Less than, which can be a less than operator, less than or equal
1321 -- operator, or the first character of a left label bracket, or the
1322 -- first character of a box (<>) compound delimiter.
1325 Accumulate_Checksum ('<');
1327 if Double_Char_Token ('=') then
1328 Token := Tok_Less_Equal;
1331 elsif Double_Char_Token ('>') then
1333 if Style_Check then Style.Check_Box; end if;
1336 elsif Double_Char_Token ('<') then
1337 Token := Tok_Less_Less;
1341 Scan_Ptr := Scan_Ptr + 1;
1346 -- Minus, which is either a subtraction operator, or the first
1347 -- character of double minus starting a comment
1349 when '-' => Minus_Case : begin
1350 if Source (Scan_Ptr + 1) = '>' then
1351 Error_Msg_S ("invalid token");
1352 Scan_Ptr := Scan_Ptr + 2;
1356 elsif Source (Scan_Ptr + 1) /= '-' then
1357 Accumulate_Checksum ('-');
1358 Scan_Ptr := Scan_Ptr + 1;
1364 else -- Source (Scan_Ptr + 1) = '-' then
1365 if Style_Check then Style.Check_Comment; end if;
1366 Scan_Ptr := Scan_Ptr + 2;
1368 -- If we are in preprocessor mode with Replace_In_Comments set,
1369 -- then we return the "--" as a token on its own.
1371 if Replace_In_Comments then
1372 Token := Tok_Comment;
1376 -- Otherwise scan out the comment
1378 Start_Of_Comment := Scan_Ptr;
1380 -- Loop to scan comment (this loop runs more than once only if
1381 -- a horizontal tab or other non-graphic character is scanned)
1384 -- Scan to non graphic character (opened up for speed)
1386 -- Note that we just eat left brackets, which means that
1387 -- bracket notation cannot be used for end of line
1388 -- characters in comments. This seems a reasonable choice,
1389 -- since no one would ever use brackets notation in a real
1390 -- program in this situation, and if we allow brackets
1391 -- notation, we forbid some valid comments which contain a
1392 -- brackets sequence that happens to match an end of line
1396 exit when Source (Scan_Ptr) not in Graphic_Character;
1397 Scan_Ptr := Scan_Ptr + 1;
1398 exit when Source (Scan_Ptr) not in Graphic_Character;
1399 Scan_Ptr := Scan_Ptr + 1;
1400 exit when Source (Scan_Ptr) not in Graphic_Character;
1401 Scan_Ptr := Scan_Ptr + 1;
1402 exit when Source (Scan_Ptr) not in Graphic_Character;
1403 Scan_Ptr := Scan_Ptr + 1;
1404 exit when Source (Scan_Ptr) not in Graphic_Character;
1405 Scan_Ptr := Scan_Ptr + 1;
1408 -- Keep going if horizontal tab
1410 if Source (Scan_Ptr) = HT then
1411 if Style_Check then Style.Check_HT; end if;
1412 Scan_Ptr := Scan_Ptr + 1;
1414 -- Terminate scan of comment if line terminator
1416 elsif Source (Scan_Ptr) in Line_Terminator then
1419 -- Terminate scan of comment if end of file encountered
1420 -- (embedded EOF character or real last character in file)
1422 elsif Source (Scan_Ptr) = EOF then
1425 -- If we have a wide character, we have to scan it out,
1426 -- because it might be a legitimate line terminator
1428 elsif (Source (Scan_Ptr) = ESC
1429 and then Identifier_Char (ESC))
1431 (Source (Scan_Ptr) in Upper_Half_Character
1432 and then Upper_Half_Encoding)
1435 Wptr : constant Source_Ptr := Scan_Ptr;
1440 Scan_Wide (Source, Scan_Ptr, Code, Err);
1442 -- If not well formed wide character, then just skip
1443 -- past it and ignore it.
1446 Scan_Ptr := Wptr + 1;
1448 -- If UTF_32 terminator, terminate comment scan
1450 elsif Is_UTF_32_Line_Terminator (UTF_32 (Code)) then
1456 -- Keep going if character in 80-FF range, or is ESC. These
1457 -- characters are allowed in comments by RM-2.1(1), 2.7(2).
1458 -- They are allowed even in Ada 83 mode according to the
1459 -- approved AI. ESC was added to the AI in June 93.
1461 elsif Source (Scan_Ptr) in Upper_Half_Character
1462 or else Source (Scan_Ptr) = ESC
1464 Scan_Ptr := Scan_Ptr + 1;
1466 -- Otherwise we have an illegal comment character
1469 Error_Illegal_Character;
1473 -- Note that, except when comments are tokens, we do NOT
1474 -- execute a return here, instead we fall through to reexecute
1475 -- the scan loop to look for a token.
1477 if Comment_Is_Token then
1478 Name_Len := Integer (Scan_Ptr - Start_Of_Comment);
1479 Name_Buffer (1 .. Name_Len) :=
1480 String (Source (Start_Of_Comment .. Scan_Ptr - 1));
1481 Comment_Id := Name_Find;
1482 Token := Tok_Comment;
1488 -- Double quote starting a string literal
1495 -- Percent starting a string literal
1498 Obsolescent_Check (Token_Ptr);
1500 if Warn_On_Obsolescent_Feature then
1502 ("use of ""'%"" is an obsolescent feature ('R'M 'J.2(4))?");
1504 ("\use """""" instead?");
1511 -- Apostrophe. This can either be the start of a character literal,
1512 -- or an isolated apostrophe used in a qualified expression or an
1513 -- attribute. We treat it as a character literal if it does not
1514 -- follow a right parenthesis, identifier, the keyword ALL or
1515 -- a literal. This means that we correctly treat constructs like:
1517 -- A := CHARACTER'('A');
1519 -- Note that RM-2.2(7) does not require a separator between
1520 -- "CHARACTER" and "'" in the above.
1522 when ''' => Char_Literal_Case : declare
1527 Accumulate_Checksum (''');
1528 Scan_Ptr := Scan_Ptr + 1;
1530 -- Here is where we make the test to distinguish the cases. Treat
1531 -- as apostrophe if previous token is an identifier, right paren
1532 -- or the reserved word "all" (latter case as in A.all'Address)
1533 -- (or the reserved word "project" in project files). Also treat
1534 -- it as apostrophe after a literal (this catches some legitimate
1535 -- cases, like A."abs"'Address, and also gives better error
1536 -- behavior for impossible cases like 123'xxx).
1538 if Prev_Token = Tok_Identifier
1539 or else Prev_Token = Tok_Right_Paren
1540 or else Prev_Token = Tok_All
1541 or else Prev_Token = Tok_Project
1542 or else Prev_Token in Token_Class_Literal
1544 Token := Tok_Apostrophe;
1545 if Style_Check then Style.Check_Apostrophe; end if;
1548 -- Otherwise the apostrophe starts a character literal
1551 -- Case of wide character literal
1553 if (Source (Scan_Ptr) = ESC
1555 Wide_Character_Encoding_Method in WC_ESC_Encoding_Method)
1557 (Source (Scan_Ptr) in Upper_Half_Character
1559 Upper_Half_Encoding)
1561 (Source (Scan_Ptr) = '['
1563 Source (Scan_Ptr + 1) = '"')
1566 Scan_Wide (Source, Scan_Ptr, Code, Err);
1567 Accumulate_Checksum (Code);
1570 Error_Illegal_Wide_Character;
1571 Code := Character'Pos (' ');
1573 -- In Ada 95 mode we allow any wide character in a character
1574 -- literal, but in Ada 2005, the set of characters allowed
1575 -- is restricted to graphic characters.
1577 elsif Ada_Version >= Ada_05
1578 and then Is_UTF_32_Non_Graphic (UTF_32 (Code))
1581 ("(Ada 2005) non-graphic character not permitted " &
1582 "in character literal", Wptr);
1585 if Source (Scan_Ptr) /= ''' then
1586 Error_Msg_S ("missing apostrophe");
1588 Scan_Ptr := Scan_Ptr + 1;
1591 -- If we do not find a closing quote in the expected place then
1592 -- assume that we have a misguided attempt at a string literal.
1594 -- However, if previous token is RANGE, then we return an
1595 -- apostrophe instead since this gives better error recovery
1597 elsif Source (Scan_Ptr + 1) /= ''' then
1598 if Prev_Token = Tok_Range then
1599 Token := Tok_Apostrophe;
1603 Scan_Ptr := Scan_Ptr - 1;
1605 ("strings are delimited by double quote character");
1611 -- Otherwise we have a (non-wide) character literal
1614 Accumulate_Checksum (Source (Scan_Ptr));
1616 if Source (Scan_Ptr) not in Graphic_Character then
1617 if Source (Scan_Ptr) in Upper_Half_Character then
1618 if Ada_Version = Ada_83 then
1619 Error_Illegal_Character;
1623 Error_Illegal_Character;
1627 Code := Get_Char_Code (Source (Scan_Ptr));
1628 Scan_Ptr := Scan_Ptr + 2;
1631 -- Fall through here with Scan_Ptr updated past the closing
1632 -- quote, and Code set to the Char_Code value for the literal
1634 Accumulate_Checksum (''');
1635 Token := Tok_Char_Literal;
1636 Set_Character_Literal_Name (Code);
1637 Token_Name := Name_Find;
1638 Character_Code := Code;
1642 end Char_Literal_Case;
1644 -- Right parenthesis
1647 Accumulate_Checksum (')');
1648 Scan_Ptr := Scan_Ptr + 1;
1649 Token := Tok_Right_Paren;
1650 if Style_Check then Style.Check_Right_Paren; end if;
1653 -- Right bracket or right brace, treated as right paren
1656 Error_Msg_S ("illegal character, replaced by "")""");
1657 Scan_Ptr := Scan_Ptr + 1;
1658 Token := Tok_Right_Paren;
1661 -- Slash (can be division operator or first character of not equal)
1664 Accumulate_Checksum ('/');
1666 if Double_Char_Token ('=') then
1667 Token := Tok_Not_Equal;
1670 Scan_Ptr := Scan_Ptr + 1;
1678 Accumulate_Checksum (';');
1679 Scan_Ptr := Scan_Ptr + 1;
1680 Token := Tok_Semicolon;
1681 if Style_Check then Style.Check_Semicolon; end if;
1686 when '|' => Vertical_Bar_Case : begin
1687 Accumulate_Checksum ('|');
1689 -- Special check for || to give nice message
1691 if Source (Scan_Ptr + 1) = '|' then
1692 Error_Msg_S ("""'|'|"" should be `OR ELSE`");
1693 Scan_Ptr := Scan_Ptr + 2;
1698 Scan_Ptr := Scan_Ptr + 1;
1699 Token := Tok_Vertical_Bar;
1700 if Style_Check then Style.Check_Vertical_Bar; end if;
1703 end Vertical_Bar_Case;
1705 -- Exclamation, replacement character for vertical bar
1707 when '!' => Exclamation_Case : begin
1708 Accumulate_Checksum ('!');
1709 Obsolescent_Check (Token_Ptr);
1711 if Warn_On_Obsolescent_Feature then
1713 ("use of ""'!"" is an obsolescent feature ('R'M 'J.2(2))?");
1715 ("\use ""'|"" instead?");
1718 if Source (Scan_Ptr + 1) = '=' then
1719 Error_Msg_S ("'!= should be /=");
1720 Scan_Ptr := Scan_Ptr + 2;
1721 Token := Tok_Not_Equal;
1725 Scan_Ptr := Scan_Ptr + 1;
1726 Token := Tok_Vertical_Bar;
1729 end Exclamation_Case;
1733 when '+' => Plus_Case : begin
1734 Accumulate_Checksum ('+');
1735 Scan_Ptr := Scan_Ptr + 1;
1740 -- Digits starting a numeric literal
1744 -- First a bit of a scan ahead to see if we have a case of an
1745 -- identifier starting with a digit (remembering exponent case).
1748 C : constant Character := Source (Scan_Ptr + 1);
1751 -- OK literal if digit followed by digit or underscore
1753 if C in '0' .. '9' or else C = '_' then
1756 -- OK literal if digit not followed by identifier char
1758 elsif not Identifier_Char (C) then
1761 -- OK literal if digit followed by e/E followed by digit/sign.
1762 -- We also allow underscore after the E, which is an error, but
1763 -- better handled by Nlit than deciding this is an identifier.
1765 elsif (C = 'e' or else C = 'E')
1766 and then (Source (Scan_Ptr + 2) in '0' .. '9'
1767 or else Source (Scan_Ptr + 2) = '+'
1768 or else Source (Scan_Ptr + 2) = '-'
1769 or else Source (Scan_Ptr + 2) = '_')
1773 -- Here we have what really looks like an identifier that
1774 -- starts with a digit, so give error msg.
1777 Error_Msg_S ("identifier may not start with digit");
1779 Underline_Found := False;
1780 Name_Buffer (1) := Source (Scan_Ptr);
1781 Accumulate_Checksum (Name_Buffer (1));
1782 Scan_Ptr := Scan_Ptr + 1;
1783 goto Scan_Identifier;
1787 -- Here we have an OK integer literal
1791 if Identifier_Char (Source (Scan_Ptr)) then
1793 ("delimiter required between literal and identifier");
1799 -- Lower case letters
1803 Underline_Found := False;
1804 Name_Buffer (1) := Source (Scan_Ptr);
1805 Accumulate_Checksum (Name_Buffer (1));
1806 Scan_Ptr := Scan_Ptr + 1;
1807 goto Scan_Identifier;
1809 -- Upper case letters
1813 Underline_Found := False;
1815 Character'Val (Character'Pos (Source (Scan_Ptr)) + 32);
1816 Accumulate_Checksum (Name_Buffer (1));
1817 Scan_Ptr := Scan_Ptr + 1;
1818 goto Scan_Identifier;
1820 -- Underline character
1823 if Special_Characters ('_') then
1824 Token_Ptr := Scan_Ptr;
1825 Scan_Ptr := Scan_Ptr + 1;
1826 Token := Tok_Special;
1827 Special_Character := '_';
1831 Error_Msg_S ("identifier cannot start with underline");
1833 Name_Buffer (1) := '_';
1834 Scan_Ptr := Scan_Ptr + 1;
1835 Underline_Found := False;
1836 goto Scan_Identifier;
1838 -- Space (not possible, because we scanned past blanks)
1841 raise Program_Error;
1843 -- Characters in top half of ASCII 8-bit chart
1845 when Upper_Half_Character =>
1847 -- Wide character case
1849 if Upper_Half_Encoding then
1850 goto Scan_Wide_Character;
1852 -- Otherwise we have OK Latin-1 character
1855 -- Upper half characters may possibly be identifier letters
1856 -- but can never be digits, so Identifier_Char can be used to
1857 -- test for a valid start of identifier character.
1859 if Identifier_Char (Source (Scan_Ptr)) then
1861 Underline_Found := False;
1862 goto Scan_Identifier;
1864 Error_Illegal_Character;
1870 -- ESC character, possible start of identifier if wide characters
1871 -- using ESC encoding are allowed in identifiers, which we can
1872 -- tell by looking at the Identifier_Char flag for ESC, which is
1873 -- only true if these conditions are met. In Ada 2005 mode, may
1874 -- also be valid UTF_32 space or line terminator character.
1876 if Identifier_Char (ESC) then
1878 goto Scan_Wide_Character;
1880 Error_Illegal_Character;
1883 -- Invalid control characters
1885 when NUL | SOH | STX | ETX | EOT | ENQ | ACK | BEL | BS | ASCII.SO |
1886 SI | DLE | DC1 | DC2 | DC3 | DC4 | NAK | SYN | ETB | CAN |
1887 EM | FS | GS | RS | US | DEL
1889 Error_Illegal_Character;
1891 -- Invalid graphic characters
1893 when '#' | '$' | '?' | '@' | '`' | '\' | '^' | '~' =>
1895 -- If Set_Special_Character has been called for this character,
1896 -- set Scans.Special_Character and return a Special token.
1898 if Special_Characters (Source (Scan_Ptr)) then
1899 Token_Ptr := Scan_Ptr;
1900 Token := Tok_Special;
1901 Special_Character := Source (Scan_Ptr);
1902 Scan_Ptr := Scan_Ptr + 1;
1905 -- Otherwise, this is an illegal character
1908 Error_Illegal_Character;
1911 -- End switch on non-blank character
1915 -- End loop past format effectors. The exit from this loop is by
1916 -- executing a return statement following completion of token scan
1917 -- (control never falls out of this loop to the code which follows)
1921 -- Wide_Character scanning routine. On entry we have encountered the
1922 -- initial character of a wide character sequence.
1924 <<Scan_Wide_Character>>
1933 Scan_Wide (Source, Scan_Ptr, Code, Err);
1935 -- If bad wide character, signal error and continue scan
1938 Error_Illegal_Wide_Character;
1939 goto Scan_Next_Character;
1942 Cat := Get_Category (UTF_32 (Code));
1944 -- If OK letter, reset scan ptr and go scan identifier
1946 if Is_UTF_32_Letter (Cat) then
1949 Underline_Found := False;
1950 goto Scan_Identifier;
1952 -- If OK wide space, ignore and keep scanning (we do not include
1953 -- any ignored spaces in checksum)
1955 elsif Is_UTF_32_Space (Cat) then
1956 goto Scan_Next_Character;
1958 -- If OK wide line terminator, terminate current line
1960 elsif Is_UTF_32_Line_Terminator (UTF_32 (Code)) then
1962 goto Scan_Line_Terminator;
1964 -- Punctuation is an error (at start of identifier)
1966 elsif Is_UTF_32_Punctuation (Cat) then
1968 ("identifier cannot start with punctuation", Wptr);
1971 Underline_Found := False;
1972 goto Scan_Identifier;
1974 -- Mark character is an error (at start of identifer)
1976 elsif Is_UTF_32_Mark (Cat) then
1978 ("identifier cannot start with mark character", Wptr);
1981 Underline_Found := False;
1982 goto Scan_Identifier;
1984 -- Other format character is an error (at start of identifer)
1986 elsif Is_UTF_32_Other (Cat) then
1988 ("identifier cannot start with other format character", Wptr);
1991 Underline_Found := False;
1992 goto Scan_Identifier;
1994 -- Extended digit character is an error. Could be bad start of
1995 -- identifier or bad literal. Not worth doing too much to try to
1996 -- distinguish these cases, but we will do a little bit.
1998 elsif Is_UTF_32_Digit (Cat) then
2000 ("identifier cannot start with digit character", Wptr);
2003 Underline_Found := False;
2004 goto Scan_Identifier;
2006 -- All other wide characters are illegal here
2009 Error_Illegal_Wide_Character;
2010 goto Scan_Next_Character;
2014 -- Routine to scan line terminator. On entry Scan_Ptr points to a
2015 -- character which is one of FF,LR,CR,VT, or one of the wide characters
2016 -- that is treated as a line termiantor.
2018 <<Scan_Line_Terminator>>
2020 -- Check line too long
2024 -- Set Token_Ptr, if End_Of_Line is a token, for the case when it is
2027 if End_Of_Line_Is_Token then
2028 Token_Ptr := Scan_Ptr;
2035 Skip_Line_Terminators (Scan_Ptr, Physical);
2037 -- If we are at start of physical line, update scan pointers to
2038 -- reflect the start of the new line.
2041 Current_Line_Start := Scan_Ptr;
2042 Start_Column := Set_Start_Column;
2043 First_Non_Blank_Location := Scan_Ptr;
2045 -- If End_Of_Line is a token, we return it as it is a
2048 if End_Of_Line_Is_Token then
2049 Token := Tok_End_Of_Line;
2055 goto Scan_Next_Character;
2057 -- Identifier scanning routine. On entry, some initial characters of
2058 -- the identifier may have already been stored in Name_Buffer. If so,
2059 -- Name_Len has the number of characters stored. otherwise Name_Len is
2060 -- set to zero on entry. Underline_Found is also set False on entry.
2064 -- This loop scans as fast as possible past lower half letters and
2065 -- digits, which we expect to be the most common characters.
2068 if Source (Scan_Ptr) in 'a' .. 'z'
2069 or else Source (Scan_Ptr) in '0' .. '9'
2071 Name_Buffer (Name_Len + 1) := Source (Scan_Ptr);
2072 Accumulate_Checksum (Source (Scan_Ptr));
2074 elsif Source (Scan_Ptr) in 'A' .. 'Z' then
2075 Name_Buffer (Name_Len + 1) :=
2076 Character'Val (Character'Pos (Source (Scan_Ptr)) + 32);
2077 Accumulate_Checksum (Name_Buffer (Name_Len + 1));
2083 Underline_Found := False;
2084 Scan_Ptr := Scan_Ptr + 1;
2085 Name_Len := Name_Len + 1;
2088 -- If we fall through, then we have encountered either an underline
2089 -- character, or an extended identifier character (i.e. one from the
2090 -- upper half), or a wide character, or an identifier terminator. The
2091 -- initial test speeds us up in the most common case where we have
2092 -- an identifier terminator. Note that ESC is an identifier character
2093 -- only if a wide character encoding method that uses ESC encoding
2094 -- is active, so if we find an ESC character we know that we have a
2097 if Identifier_Char (Source (Scan_Ptr)) then
2099 -- Case of underline
2101 if Source (Scan_Ptr) = '_' then
2102 Accumulate_Checksum ('_');
2104 if Underline_Found then
2105 Error_No_Double_Underline;
2107 Underline_Found := True;
2108 Name_Len := Name_Len + 1;
2109 Name_Buffer (Name_Len) := '_';
2112 Scan_Ptr := Scan_Ptr + 1;
2113 goto Scan_Identifier;
2115 -- Upper half character
2117 elsif Source (Scan_Ptr) in Upper_Half_Character
2118 and then not Upper_Half_Encoding
2120 Accumulate_Checksum (Source (Scan_Ptr));
2121 Store_Encoded_Character
2122 (Get_Char_Code (Fold_Lower (Source (Scan_Ptr))));
2123 Scan_Ptr := Scan_Ptr + 1;
2124 Underline_Found := False;
2125 goto Scan_Identifier;
2127 -- Left bracket not followed by a quote terminates an identifier.
2128 -- This is an error, but we don't want to give a junk error msg
2129 -- about wide characters in this case!
2131 elsif Source (Scan_Ptr) = '['
2132 and then Source (Scan_Ptr + 1) /= '"'
2136 -- We know we have a wide character encoding here (the current
2137 -- character is either ESC, left bracket, or an upper half
2138 -- character depending on the encoding method).
2141 -- Scan out the wide character and insert the appropriate
2142 -- encoding into the name table entry for the identifier.
2152 Scan_Wide (Source, Scan_Ptr, Code, Err);
2154 -- If error, signal error
2157 Error_Illegal_Wide_Character;
2159 -- If the character scanned is a normal identifier
2160 -- character, then we treat it that way.
2162 elsif In_Character_Range (Code)
2163 and then Identifier_Char (Get_Character (Code))
2165 Chr := Get_Character (Code);
2166 Accumulate_Checksum (Chr);
2167 Store_Encoded_Character
2168 (Get_Char_Code (Fold_Lower (Chr)));
2169 Underline_Found := False;
2171 -- Here if not a normal identifier character
2174 -- Make sure we are allowing wide characters in
2175 -- identifiers. Note that we allow wide character
2176 -- notation for an OK identifier character. This in
2177 -- particular allows bracket or other notation to be
2178 -- used for upper half letters.
2180 -- Wide characters are always allowed in Ada 2005
2182 if Identifier_Character_Set /= 'w'
2183 and then Ada_Version < Ada_05
2186 ("wide character not allowed in identifier", Wptr);
2189 Cat := Get_Category (UTF_32 (Code));
2191 -- If OK letter, store it folding to upper case. Note
2192 -- that we include the folded letter in the checksum.
2194 if Is_UTF_32_Letter (Cat) then
2196 Char_Code (UTF_32_To_Upper_Case (UTF_32 (Code)));
2197 Accumulate_Checksum (Code);
2198 Store_Encoded_Character (Code);
2199 Underline_Found := False;
2201 -- If OK extended digit or mark, then store it
2203 elsif Is_UTF_32_Digit (Cat)
2204 or else Is_UTF_32_Mark (Cat)
2206 Accumulate_Checksum (Code);
2207 Store_Encoded_Character (Code);
2208 Underline_Found := False;
2210 -- Wide punctuation is also stored, but counts as an
2211 -- underline character for error checking purposes.
2213 elsif Is_UTF_32_Punctuation (Cat) then
2214 Accumulate_Checksum (Code);
2216 if Underline_Found then
2218 Cend : constant Source_Ptr := Scan_Ptr;
2221 Error_No_Double_Underline;
2226 Store_Encoded_Character (Code);
2227 Underline_Found := True;
2230 -- Wide character in Unicode cateogory "Other, Format"
2231 -- is accepted in an identifier, but is ignored and not
2232 -- stored. It seems reasonable to exclude it from the
2235 -- Note that it is correct (see AI-395) to simply strip
2236 -- other format characters, before testing for double
2237 -- underlines, or for reserved words).
2239 elsif Is_UTF_32_Other (Cat) then
2242 -- Wide character in category Separator,Space terminates
2244 elsif Is_UTF_32_Space (Cat) then
2245 goto Scan_Identifier_Complete;
2247 -- Any other wide character is not acceptable
2251 ("invalid wide character in identifier", Wptr);
2255 goto Scan_Identifier;
2260 -- Scan of identifier is complete. The identifier is stored in
2261 -- Name_Buffer, and Scan_Ptr points past the last character.
2263 <<Scan_Identifier_Complete>>
2264 Token_Name := Name_Find;
2266 -- Check for identifier ending with underline or punctuation char
2268 if Underline_Found then
2269 Underline_Found := False;
2271 if Source (Scan_Ptr - 1) = '_' then
2273 ("identifier cannot end with underline", Scan_Ptr - 1);
2276 ("identifier cannot end with punctuation character", Wptr);
2280 -- Here is where we check if it was a keyword
2282 if Is_Keyword_Name (Token_Name) then
2283 Token := Token_Type'Val (Get_Name_Table_Byte (Token_Name));
2285 -- Deal with possible style check for non-lower case keyword, but
2286 -- we don't treat ACCESS, DELTA, DIGITS, RANGE as keywords for
2287 -- this purpose if they appear as attribute designators. Actually
2288 -- we only check the first character for speed.
2290 -- Ada 2005 (AI-284): Do not apply the style check in case of
2291 -- "pragma Interface"
2293 -- Ada 2005 (AI-340): Do not apply the style check in case of
2297 and then Source (Token_Ptr) <= 'Z'
2298 and then (Prev_Token /= Tok_Apostrophe
2300 (Token /= Tok_Access and then
2301 Token /= Tok_Delta and then
2302 Token /= Tok_Digits and then
2303 Token /= Tok_Mod and then
2304 Token /= Tok_Range))
2305 and then (Token /= Tok_Interface
2307 (Token = Tok_Interface
2308 and then Prev_Token /= Tok_Pragma))
2310 Style.Non_Lower_Case_Keyword;
2313 -- We must reset Token_Name since this is not an identifier and
2314 -- if we leave Token_Name set, the parser gets confused because
2315 -- it thinks it is dealing with an identifier instead of the
2316 -- corresponding keyword.
2318 Token_Name := No_Name;
2319 Accumulate_Token_Checksum;
2322 -- It is an identifier after all
2325 Token := Tok_Identifier;
2326 Accumulate_Token_Checksum;
2332 --------------------------
2333 -- Set_Comment_As_Token --
2334 --------------------------
2336 procedure Set_Comment_As_Token (Value : Boolean) is
2338 Comment_Is_Token := Value;
2339 end Set_Comment_As_Token;
2341 ------------------------------
2342 -- Set_End_Of_Line_As_Token --
2343 ------------------------------
2345 procedure Set_End_Of_Line_As_Token (Value : Boolean) is
2347 End_Of_Line_Is_Token := Value;
2348 end Set_End_Of_Line_As_Token;
2350 ---------------------------
2351 -- Set_Special_Character --
2352 ---------------------------
2354 procedure Set_Special_Character (C : Character) is
2357 when '#' | '$' | '_' | '?' | '@' | '`' | '\' | '^' | '~' =>
2358 Special_Characters (C) := True;
2363 end Set_Special_Character;
2365 ----------------------
2366 -- Set_Start_Column --
2367 ----------------------
2369 -- Note: it seems at first glance a little expensive to compute this value
2370 -- for every source line (since it is certainly not used for all source
2371 -- lines). On the other hand, it doesn't take much more work to skip past
2372 -- the initial white space on the line counting the columns than it would
2373 -- to scan past the white space using the standard scanning circuits.
2375 function Set_Start_Column return Column_Number is
2376 Start_Column : Column_Number := 0;
2379 -- Outer loop scans past horizontal tab characters
2383 -- Inner loop scans past blanks as fast as possible, bumping Scan_Ptr
2384 -- past the blanks and adjusting Start_Column to account for them.
2387 if Source (Scan_Ptr) = ' ' then
2388 if Source (Scan_Ptr + 1) = ' ' then
2389 if Source (Scan_Ptr + 2) = ' ' then
2390 if Source (Scan_Ptr + 3) = ' ' then
2391 if Source (Scan_Ptr + 4) = ' ' then
2392 if Source (Scan_Ptr + 5) = ' ' then
2393 if Source (Scan_Ptr + 6) = ' ' then
2394 Scan_Ptr := Scan_Ptr + 7;
2395 Start_Column := Start_Column + 7;
2397 Scan_Ptr := Scan_Ptr + 6;
2398 Start_Column := Start_Column + 6;
2402 Scan_Ptr := Scan_Ptr + 5;
2403 Start_Column := Start_Column + 5;
2407 Scan_Ptr := Scan_Ptr + 4;
2408 Start_Column := Start_Column + 4;
2412 Scan_Ptr := Scan_Ptr + 3;
2413 Start_Column := Start_Column + 3;
2417 Scan_Ptr := Scan_Ptr + 2;
2418 Start_Column := Start_Column + 2;
2422 Scan_Ptr := Scan_Ptr + 1;
2423 Start_Column := Start_Column + 1;
2429 end loop Blanks_Loop;
2431 -- Outer loop keeps going only if a horizontal tab follows
2433 if Source (Scan_Ptr) = HT then
2434 if Style_Check then Style.Check_HT; end if;
2435 Scan_Ptr := Scan_Ptr + 1;
2436 Start_Column := (Start_Column / 8) * 8 + 8;
2443 return Start_Column;
2444 end Set_Start_Column;