1 ------------------------------------------------------------------------------
3 -- GNAT COMPILER COMPONENTS --
9 -- Copyright (C) 1992-2013, 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 3, 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 COPYING3. If not, go to --
19 -- http://www.gnu.org/licenses for a complete copy of the license. --
21 -- GNAT was originally developed by the GNAT team at New York University. --
22 -- Extensive contributions were provided by Ada Core Technologies Inc. --
24 ------------------------------------------------------------------------------
26 with Atree
; use Atree
;
27 with Csets
; use Csets
;
28 with Hostparm
; use Hostparm
;
29 with Namet
; use Namet
;
31 with Restrict
; use Restrict
;
32 with Rident
; use Rident
;
33 with Scans
; use Scans
;
34 with Sinput
; use Sinput
;
35 with Snames
; use Snames
;
36 with Stringt
; use Stringt
;
37 with Stylesw
; use Stylesw
;
38 with Uintp
; use Uintp
;
39 with Urealp
; use Urealp
;
40 with Widechar
; use Widechar
;
42 pragma Warnings
(Off
);
43 -- This package is used also by gnatcoll
45 with System
.UTF_32
; use System
.UTF_32
;
46 with System
.WCh_Con
; use System
.WCh_Con
;
52 -- Make control characters visible
54 Special_Characters
: array (Character) of Boolean := (others => False);
55 -- For characters that are Special token, the value is True
57 Comment_Is_Token
: Boolean := False;
58 -- True if comments are tokens
60 End_Of_Line_Is_Token
: Boolean := False;
61 -- True if End_Of_Line is a token
63 -----------------------
64 -- Local Subprograms --
65 -----------------------
67 procedure Accumulate_Token_Checksum
;
68 pragma Inline
(Accumulate_Token_Checksum
);
69 -- Called after each numeric literal and identifier/keyword. For keywords,
70 -- the token used is Tok_Identifier. This allows to detect additional
71 -- spaces added in sources when using the builder switch -m.
73 procedure Accumulate_Token_Checksum_GNAT_6_3
;
74 -- Used in place of Accumulate_Token_Checksum for GNAT versions 5.04 to
75 -- 6.3, when Tok_Some was not included in Token_Type and the actual
76 -- Token_Type was used for keywords. This procedure is never used in the
77 -- compiler or gnatmake, only in gprbuild.
79 procedure Accumulate_Token_Checksum_GNAT_5_03
;
80 -- Used in place of Accumulate_Token_Checksum for GNAT version 5.03, when
81 -- Tok_Interface, Tok_Some, Tok_Synchronized and Tok_Overriding were not
82 -- included in Token_Type and the actual Token_Type was used for keywords.
83 -- This procedure is never used in the compiler or gnatmake, only in
86 procedure Accumulate_Checksum
(C
: Character);
87 pragma Inline
(Accumulate_Checksum
);
88 -- This routine accumulates the checksum given character C. During the
89 -- scanning of a source file, this routine is called with every character
90 -- in the source, excluding blanks, and all control characters (except
91 -- that ESC is included in the checksum). Upper case letters not in string
92 -- literals are folded by the caller. See Sinput spec for the documentation
93 -- of the checksum algorithm. Note: checksum values are only used if we
94 -- generate code, so it is not necessary to worry about making the right
95 -- sequence of calls in any error situation.
97 procedure Accumulate_Checksum
(C
: Char_Code
);
98 pragma Inline
(Accumulate_Checksum
);
99 -- This version is identical, except that the argument, C, is a character
100 -- code value instead of a character. This is used when wide characters
101 -- are scanned. We use the character code rather than the ASCII characters
102 -- so that the checksum is independent of wide character encoding method.
104 procedure Initialize_Checksum
;
105 pragma Inline
(Initialize_Checksum
);
106 -- Initialize checksum value
108 -------------------------
109 -- Accumulate_Checksum --
110 -------------------------
112 procedure Accumulate_Checksum
(C
: Character) is
114 System
.CRC32
.Update
(System
.CRC32
.CRC32
(Checksum
), C
);
115 end Accumulate_Checksum
;
117 procedure Accumulate_Checksum
(C
: Char_Code
) is
120 Accumulate_Checksum
(Character'Val (C
/ 2 ** 24));
121 Accumulate_Checksum
(Character'Val ((C
/ 2 ** 16) mod 256));
122 Accumulate_Checksum
(Character'Val ((C
/ 256) mod 256));
124 Accumulate_Checksum
(Character'Val (C
/ 256));
127 Accumulate_Checksum
(Character'Val (C
mod 256));
128 end Accumulate_Checksum
;
130 -------------------------------
131 -- Accumulate_Token_Checksum --
132 -------------------------------
134 procedure Accumulate_Token_Checksum
is
137 (System
.CRC32
.CRC32
(Checksum
),
138 Character'Val (Token_Type
'Pos (Token
)));
139 end Accumulate_Token_Checksum
;
141 ----------------------------------------
142 -- Accumulate_Token_Checksum_GNAT_6_3 --
143 ----------------------------------------
145 procedure Accumulate_Token_Checksum_GNAT_6_3
is
147 -- Individual values of Token_Type are used, instead of subranges, so
148 -- that additions or suppressions of enumerated values in type
149 -- Token_Type are detected by the compiler.
152 when Tok_Integer_Literal | Tok_Real_Literal | Tok_String_Literal |
153 Tok_Char_Literal | Tok_Operator_Symbol | Tok_Identifier |
154 Tok_Double_Asterisk | Tok_Ampersand | Tok_Minus | Tok_Plus |
155 Tok_Asterisk | Tok_Mod | Tok_Rem | Tok_Slash | Tok_New |
156 Tok_Abs | Tok_Others | Tok_Null | Tok_Dot | Tok_Apostrophe |
157 Tok_Left_Paren | Tok_Delta | Tok_Digits | Tok_Range |
158 Tok_Right_Paren | Tok_Comma | Tok_And | Tok_Or | Tok_Xor |
159 Tok_Less | Tok_Equal | Tok_Greater | Tok_Not_Equal |
160 Tok_Greater_Equal | Tok_Less_Equal | Tok_In | Tok_Not |
161 Tok_Box | Tok_Colon_Equal | Tok_Colon | Tok_Greater_Greater |
162 Tok_Abstract | Tok_Access | Tok_Aliased | Tok_All | Tok_Array |
163 Tok_At | Tok_Body | Tok_Constant | Tok_Do | Tok_Is |
164 Tok_Interface | Tok_Limited | Tok_Of | Tok_Out | Tok_Record |
165 Tok_Renames | Tok_Reverse
=>
168 (System
.CRC32
.CRC32
(Checksum
),
169 Character'Val (Token_Type
'Pos (Token
)));
174 (System
.CRC32
.CRC32
(Checksum
),
175 Character'Val (Token_Type
'Pos (Tok_Identifier
)));
177 when Tok_Tagged | Tok_Then | Tok_Less_Less | Tok_Abort | Tok_Accept |
178 Tok_Case | Tok_Delay | Tok_Else | Tok_Elsif | Tok_End |
179 Tok_Exception | Tok_Exit | Tok_Goto | Tok_If | Tok_Pragma |
180 Tok_Raise | Tok_Requeue | Tok_Return | Tok_Select |
181 Tok_Terminate | Tok_Until | Tok_When | Tok_Begin | Tok_Declare |
182 Tok_For | Tok_Loop | Tok_While | Tok_Entry | Tok_Protected |
183 Tok_Task | Tok_Type | Tok_Subtype | Tok_Overriding |
184 Tok_Synchronized | Tok_Use | Tok_Function | Tok_Generic |
185 Tok_Package | Tok_Procedure | Tok_Private | Tok_With |
186 Tok_Separate | Tok_EOF | Tok_Semicolon | Tok_Arrow |
187 Tok_Vertical_Bar | Tok_Dot_Dot | Tok_Project | Tok_Extends |
188 Tok_External | Tok_External_As_List | Tok_Comment |
189 Tok_End_Of_Line | Tok_Special | Tok_SPARK_Hide | No_Token
=>
192 (System
.CRC32
.CRC32
(Checksum
),
193 Character'Val (Token_Type
'Pos (Token_Type
'Pred (Token
))));
195 end Accumulate_Token_Checksum_GNAT_6_3
;
197 -----------------------------------------
198 -- Accumulate_Token_Checksum_GNAT_5_03 --
199 -----------------------------------------
201 procedure Accumulate_Token_Checksum_GNAT_5_03
is
203 -- Individual values of Token_Type are used, instead of subranges, so
204 -- that additions or suppressions of enumerated values in type
205 -- Token_Type are detected by the compiler.
208 when Tok_Integer_Literal | Tok_Real_Literal | Tok_String_Literal |
209 Tok_Char_Literal | Tok_Operator_Symbol | Tok_Identifier |
210 Tok_Double_Asterisk | Tok_Ampersand | Tok_Minus | Tok_Plus |
211 Tok_Asterisk | Tok_Mod | Tok_Rem | Tok_Slash | Tok_New |
212 Tok_Abs | Tok_Others | Tok_Null | Tok_Dot | Tok_Apostrophe |
213 Tok_Left_Paren | Tok_Delta | Tok_Digits | Tok_Range |
214 Tok_Right_Paren | Tok_Comma | Tok_And | Tok_Or | Tok_Xor |
215 Tok_Less | Tok_Equal | Tok_Greater | Tok_Not_Equal |
216 Tok_Greater_Equal | Tok_Less_Equal | Tok_In | Tok_Not |
217 Tok_Box | Tok_Colon_Equal | Tok_Colon | Tok_Greater_Greater |
218 Tok_Abstract | Tok_Access | Tok_Aliased | Tok_All | Tok_Array |
219 Tok_At | Tok_Body | Tok_Constant | Tok_Do | Tok_Is
=>
222 (System
.CRC32
.CRC32
(Checksum
),
223 Character'Val (Token_Type
'Pos (Token
)));
225 when Tok_Interface | Tok_Some | Tok_Overriding | Tok_Synchronized
=>
227 (System
.CRC32
.CRC32
(Checksum
),
228 Character'Val (Token_Type
'Pos (Tok_Identifier
)));
230 when Tok_Limited | Tok_Of | Tok_Out | Tok_Record |
231 Tok_Renames | Tok_Reverse
=>
234 (System
.CRC32
.CRC32
(Checksum
),
235 Character'Val (Token_Type
'Pos (Token
) - 1));
237 when Tok_Tagged | Tok_Then | Tok_Less_Less | Tok_Abort | Tok_Accept |
238 Tok_Case | Tok_Delay | Tok_Else | Tok_Elsif | Tok_End |
239 Tok_Exception | Tok_Exit | Tok_Goto | Tok_If | Tok_Pragma |
240 Tok_Raise | Tok_Requeue | Tok_Return | Tok_Select |
241 Tok_Terminate | Tok_Until | Tok_When | Tok_Begin | Tok_Declare |
242 Tok_For | Tok_Loop | Tok_While | Tok_Entry | Tok_Protected |
243 Tok_Task | Tok_Type | Tok_Subtype
=>
246 (System
.CRC32
.CRC32
(Checksum
),
247 Character'Val (Token_Type
'Pos (Token
) - 2));
249 when Tok_Use | Tok_Function | Tok_Generic |
250 Tok_Package | Tok_Procedure | Tok_Private | Tok_With |
251 Tok_Separate | Tok_EOF | Tok_Semicolon | Tok_Arrow |
252 Tok_Vertical_Bar | Tok_Dot_Dot | Tok_Project | Tok_Extends |
253 Tok_External | Tok_External_As_List | Tok_Comment |
254 Tok_End_Of_Line | Tok_Special | Tok_SPARK_Hide | No_Token
=>
257 (System
.CRC32
.CRC32
(Checksum
),
258 Character'Val (Token_Type
'Pos (Token
) - 4));
260 end Accumulate_Token_Checksum_GNAT_5_03
;
262 ----------------------------
263 -- Determine_Token_Casing --
264 ----------------------------
266 function Determine_Token_Casing
return Casing_Type
is
268 return Determine_Casing
(Source
(Token_Ptr
.. Scan_Ptr
- 1));
269 end Determine_Token_Casing
;
271 -------------------------
272 -- Initialize_Checksum --
273 -------------------------
275 procedure Initialize_Checksum
is
277 System
.CRC32
.Initialize
(System
.CRC32
.CRC32
(Checksum
));
278 end Initialize_Checksum
;
280 ------------------------
281 -- Initialize_Scanner --
282 ------------------------
284 procedure Initialize_Scanner
(Index
: Source_File_Index
) is
286 -- Establish reserved words
288 Scans
.Initialize_Ada_Keywords
;
290 -- Initialize scan control variables
292 Current_Source_File
:= Index
;
293 Source
:= Source_Text
(Current_Source_File
);
294 Scan_Ptr
:= Source_First
(Current_Source_File
);
296 Token_Ptr
:= Scan_Ptr
;
297 Current_Line_Start
:= Scan_Ptr
;
299 Token_Name
:= No_Name
;
300 Start_Column
:= Set_Start_Column
;
301 First_Non_Blank_Location
:= Scan_Ptr
;
304 Wide_Char_Byte_Count
:= 0;
306 -- Do not call Scan, otherwise the License stuff does not work in Scn
308 end Initialize_Scanner
;
310 ------------------------------
311 -- Reset_Special_Characters --
312 ------------------------------
314 procedure Reset_Special_Characters
is
316 Special_Characters
:= (others => False);
317 end Reset_Special_Characters
;
325 Start_Of_Comment
: Source_Ptr
;
326 -- Record start of comment position
328 Underline_Found
: Boolean;
329 -- During scanning of an identifier, set to True if last character
330 -- scanned was an underline or other punctuation character. This
331 -- is used to flag the error of two underlines/punctuations in a
332 -- row or ending an identifier with a underline/punctuation. Here
333 -- punctuation means any UTF_32 character in the Unicode category
334 -- Punctuation,Connector.
337 -- Used to remember start of last wide character scanned
339 procedure Check_End_Of_Line
;
340 -- Called when end of line encountered. Checks that line is not too
341 -- long, and that other style checks for the end of line are met.
343 function Double_Char_Token
(C
: Character) return Boolean;
344 -- This function is used for double character tokens like := or <>. It
345 -- checks if the character following Source (Scan_Ptr) is C, and if so
346 -- bumps Scan_Ptr past the pair of characters and returns True. A space
347 -- between the two characters is also recognized with an appropriate
348 -- error message being issued. If C is not present, False is returned.
349 -- Note that Double_Char_Token can only be used for tokens defined in
350 -- the Ada syntax (it's use for error cases like && is not appropriate
351 -- since we do not want a junk message for a case like &-space-&).
353 procedure Error_Illegal_Character
;
354 -- Give illegal character error, Scan_Ptr points to character. On
355 -- return, Scan_Ptr is bumped past the illegal character.
357 procedure Error_Illegal_Wide_Character
;
358 -- Give illegal wide character message. On return, Scan_Ptr is bumped
359 -- past the illegal character, which may still leave us pointing to
360 -- junk, not much we can do if the escape sequence is messed up!
362 procedure Error_Long_Line
;
363 -- Signal error of excessively long line
365 procedure Error_No_Double_Underline
;
366 -- Signal error of two underline or punctuation characters in a row.
367 -- Called with Scan_Ptr pointing to second underline/punctuation char.
370 -- This is the procedure for scanning out numeric literals. On entry,
371 -- Scan_Ptr points to the digit that starts the numeric literal (the
372 -- checksum for this character has not been accumulated yet). On return
373 -- Scan_Ptr points past the last character of the numeric literal, Token
374 -- and Token_Node are set appropriately, and the checksum is updated.
377 -- This is the procedure for scanning out string literals. On entry,
378 -- Scan_Ptr points to the opening string quote (the checksum for this
379 -- character has not been accumulated yet). On return Scan_Ptr points
380 -- past the closing quote of the string literal, Token and Token_Node
381 -- are set appropriately, and the checksum is updated.
383 procedure Skip_Other_Format_Characters
;
384 -- Skips past any "other format" category characters at the current
385 -- cursor location (does not skip past spaces or any other characters).
387 function Start_Of_Wide_Character
return Boolean;
388 -- Returns True if the scan pointer is pointing to the start of a wide
389 -- character sequence, does not modify the scan pointer in any case.
391 -----------------------
392 -- Check_End_Of_Line --
393 -----------------------
395 procedure Check_End_Of_Line
is
396 Len
: constant Int
:=
398 Int
(Current_Line_Start
) -
399 Wide_Char_Byte_Count
;
403 Style
.Check_Line_Terminator
(Len
);
406 -- Deal with checking maximum line length
408 if Style_Check
and Style_Check_Max_Line_Length
then
409 Style
.Check_Line_Max_Length
(Len
);
411 -- If style checking is inactive, check maximum line length against
414 elsif Len
> Max_Line_Length
then
418 -- Now one more checking circuit. Normally we are only enforcing a
419 -- limit of physical characters, with tabs counting as one character.
420 -- But if after tab expansion we would have a total line length that
421 -- exceeded 32766, that would really cause trouble, because column
422 -- positions would exceed the maximum we allow for a column count.
423 -- Note: the limit is 32766 rather than 32767, since we use a value
424 -- of 32767 for special purposes (see Sinput). Now we really do not
425 -- want to go messing with tabs in the normal case, so what we do is
426 -- to check for a line that has more than 4096 physical characters.
427 -- Any shorter line could not be a problem, even if it was all tabs.
436 Ptr
:= Current_Line_Start
;
438 exit when Ptr
= Scan_Ptr
;
440 if Source
(Ptr
) = ASCII
.HT
then
441 Col
:= (Col
- 1 + 8) / 8 * 8 + 1;
448 ("this line is longer than 32766 characters",
450 raise Unrecoverable_Error
;
458 -- Reset wide character byte count for next line
460 Wide_Char_Byte_Count
:= 0;
461 end Check_End_Of_Line
;
463 -----------------------
464 -- Double_Char_Token --
465 -----------------------
467 function Double_Char_Token
(C
: Character) return Boolean is
469 if Source
(Scan_Ptr
+ 1) = C
then
470 Accumulate_Checksum
(C
);
471 Scan_Ptr
:= Scan_Ptr
+ 2;
474 elsif Source
(Scan_Ptr
+ 1) = ' '
475 and then Source
(Scan_Ptr
+ 2) = C
477 Scan_Ptr
:= Scan_Ptr
+ 1;
478 Error_Msg_S
-- CODEFIX
479 ("no space allowed here");
480 Scan_Ptr
:= Scan_Ptr
+ 2;
486 end Double_Char_Token
;
488 -----------------------------
489 -- Error_Illegal_Character --
490 -----------------------------
492 procedure Error_Illegal_Character
is
494 Error_Msg_S
("illegal character");
495 Scan_Ptr
:= Scan_Ptr
+ 1;
496 end Error_Illegal_Character
;
498 ----------------------------------
499 -- Error_Illegal_Wide_Character --
500 ----------------------------------
502 procedure Error_Illegal_Wide_Character
is
504 Scan_Ptr
:= Scan_Ptr
+ 1;
505 Error_Msg
("illegal wide character", Wptr
);
506 end Error_Illegal_Wide_Character
;
508 ---------------------
509 -- Error_Long_Line --
510 ---------------------
512 procedure Error_Long_Line
is
515 ("this line is too long",
516 Current_Line_Start
+ Source_Ptr
(Max_Line_Length
));
519 -------------------------------
520 -- Error_No_Double_Underline --
521 -------------------------------
523 procedure Error_No_Double_Underline
is
525 Underline_Found
:= False;
527 -- There are four cases, and we special case the messages
529 if Source
(Scan_Ptr
) = '_' then
530 if Source
(Scan_Ptr
- 1) = '_' then
531 Error_Msg_S
-- CODEFIX
532 ("two consecutive underlines not permitted");
534 Error_Msg_S
("underline cannot follow punctuation character");
538 if Source
(Scan_Ptr
- 1) = '_' then
539 Error_Msg_S
("punctuation character cannot follow underline");
542 ("two consecutive punctuation characters not permitted");
545 end Error_No_Double_Underline
;
554 -- Current source program character
556 Base_Char
: Character;
557 -- Either # or : (character at start of based number)
563 -- Value of base in Uint format
566 -- Value of integer scanned by Scan_Integer in Uint format
569 -- Value of integer in numeric value being scanned
572 -- Scale value for real literal
575 -- Scale in Uint format
577 Exponent_Is_Negative
: Boolean;
578 -- Set true for negative exponent
580 Extended_Digit_Value
: Int
;
581 -- Extended digit value
583 Point_Scanned
: Boolean;
584 -- Flag for decimal point scanned in numeric literal
586 -----------------------
587 -- Local Subprograms --
588 -----------------------
590 procedure Error_Digit_Expected
;
591 -- Signal error of bad digit, Scan_Ptr points to the location at
592 -- which the digit was expected on input, and is unchanged on return.
594 procedure Scan_Integer
;
595 -- Scan integer literal. On entry, Scan_Ptr points to a digit, on
596 -- exit Scan_Ptr points past the last character of the integer.
598 -- For each digit encountered, UI_Int_Value is multiplied by 10, and
599 -- the value of the digit added to the result. In addition, the value
600 -- in Scale is decremented by one for each actual digit scanned.
602 --------------------------
603 -- Error_Digit_Expected --
604 --------------------------
606 procedure Error_Digit_Expected
is
608 Error_Msg_S
("digit expected");
609 end Error_Digit_Expected
;
615 procedure Scan_Integer
is
617 -- Next character scanned
620 C
:= Source
(Scan_Ptr
);
622 -- Loop through digits (allowing underlines)
625 Accumulate_Checksum
(C
);
627 UI_Int_Value
* 10 + (Character'Pos (C
) - Character'Pos ('0'));
628 Scan_Ptr
:= Scan_Ptr
+ 1;
630 C
:= Source
(Scan_Ptr
);
632 -- Case of underline encountered
636 -- We do not accumulate the '_' in the checksum, so that
637 -- 1_234 is equivalent to 1234, and does not trigger
638 -- compilation for "minimal recompilation" (gnatmake -m).
641 Scan_Ptr
:= Scan_Ptr
+ 1;
642 C
:= Source
(Scan_Ptr
);
644 Error_No_Double_Underline
;
647 if C
not in '0' .. '9' then
648 Error_Digit_Expected
;
653 exit when C
not in '0' .. '9';
658 -- Start of Processing for Nlit
663 UI_Int_Value
:= Uint_0
;
664 Based_Literal_Uses_Colon
:= False;
667 Point_Scanned
:= False;
668 UI_Num_Value
:= UI_Int_Value
;
670 -- Various possibilities now for continuing the literal are period,
671 -- E/e (for exponent), or :/# (for based literal).
674 C
:= Source
(Scan_Ptr
);
678 -- Scan out point, but do not scan past .. which is a range
679 -- sequence, and must not be eaten up scanning a numeric literal.
681 while C
= '.' and then Source
(Scan_Ptr
+ 1) /= '.' loop
682 Accumulate_Checksum
('.');
684 if Point_Scanned
then
685 Error_Msg_S
("duplicate point ignored");
688 Point_Scanned
:= True;
689 Scan_Ptr
:= Scan_Ptr
+ 1;
690 C
:= Source
(Scan_Ptr
);
692 if C
not in '0' .. '9' then
694 ("real literal cannot end with point", Scan_Ptr
- 1);
697 UI_Num_Value
:= UI_Int_Value
;
701 -- Based literal case. The base is the value we already scanned.
702 -- In the case of colon, we insist that the following character
703 -- is indeed an extended digit or a period. This catches a number
704 -- of common errors, as well as catching the well known tricky
705 -- bug otherwise arising from "x : integer range 1 .. 10:= 6;"
708 or else (C
= ':' and then
709 (Source
(Scan_Ptr
+ 1) = '.'
711 Source
(Scan_Ptr
+ 1) in '0' .. '9'
713 Source
(Scan_Ptr
+ 1) in 'A' .. 'Z'
715 Source
(Scan_Ptr
+ 1) in 'a' .. 'z'))
717 Accumulate_Checksum
(C
);
719 UI_Base
:= UI_Int_Value
;
721 if Base_Char
= ':' then
722 Based_Literal_Uses_Colon
:= True;
725 if UI_Base
< 2 or else UI_Base
> 16 then
726 Error_Msg_SC
("base not 2-16");
730 Base
:= UI_To_Int
(UI_Base
);
731 Scan_Ptr
:= Scan_Ptr
+ 1;
733 -- Scan out extended integer [. integer]
735 C
:= Source
(Scan_Ptr
);
736 UI_Int_Value
:= Uint_0
;
740 if C
in '0' .. '9' then
741 Accumulate_Checksum
(C
);
742 Extended_Digit_Value
:=
743 Int
'(Character'Pos (C)) - Int'(Character'Pos ('0'));
745 elsif C
in 'A' .. 'F' then
746 Accumulate_Checksum
(Character'Val (Character'Pos (C
) + 32));
747 Extended_Digit_Value
:=
748 Int
'(Character'Pos (C)) - Int'(Character'Pos ('A')) + 10;
750 elsif C
in 'a' .. 'f' then
751 Accumulate_Checksum
(C
);
752 Extended_Digit_Value
:=
753 Int
'(Character'Pos (C)) - Int'(Character'Pos ('a')) + 10;
756 Error_Msg_S
("extended digit expected");
760 if Extended_Digit_Value
>= Base
then
761 Error_Msg_S
("digit '>= base");
764 UI_Int_Value
:= UI_Int_Value
* UI_Base
+ Extended_Digit_Value
;
766 Scan_Ptr
:= Scan_Ptr
+ 1;
767 C
:= Source
(Scan_Ptr
);
771 Accumulate_Checksum
('_');
772 Scan_Ptr
:= Scan_Ptr
+ 1;
773 C
:= Source
(Scan_Ptr
);
775 Error_No_Double_Underline
;
779 Accumulate_Checksum
('.');
781 if Point_Scanned
then
782 Error_Msg_S
("duplicate point ignored");
785 Scan_Ptr
:= Scan_Ptr
+ 1;
786 C
:= Source
(Scan_Ptr
);
787 Point_Scanned
:= True;
790 elsif C
= Base_Char
then
791 Accumulate_Checksum
(C
);
792 Scan_Ptr
:= Scan_Ptr
+ 1;
795 elsif C
= '#' or else C
= ':' then
796 Error_Msg_S
("based number delimiters must match");
797 Scan_Ptr
:= Scan_Ptr
+ 1;
800 elsif not Identifier_Char
(C
) then
801 if Base_Char
= '#' then
802 Error_Msg_S
-- CODEFIX
805 Error_Msg_S
-- CODEFIX
814 UI_Num_Value
:= UI_Int_Value
;
819 if not Point_Scanned
then
823 UI_Scale
:= UI_From_Int
(Scale
);
826 if Source
(Scan_Ptr
) = 'e' or else Source
(Scan_Ptr
) = 'E' then
827 Accumulate_Checksum
('e');
828 Scan_Ptr
:= Scan_Ptr
+ 1;
829 Exponent_Is_Negative
:= False;
831 if Source
(Scan_Ptr
) = '+' then
832 Accumulate_Checksum
('+');
833 Scan_Ptr
:= Scan_Ptr
+ 1;
835 elsif Source
(Scan_Ptr
) = '-' then
836 Accumulate_Checksum
('-');
838 if not Point_Scanned
then
840 ("negative exponent not allowed for integer literal");
842 Exponent_Is_Negative
:= True;
845 Scan_Ptr
:= Scan_Ptr
+ 1;
848 UI_Int_Value
:= Uint_0
;
850 if Source
(Scan_Ptr
) in '0' .. '9' then
853 Error_Digit_Expected
;
856 if Exponent_Is_Negative
then
857 UI_Scale
:= UI_Scale
- UI_Int_Value
;
859 UI_Scale
:= UI_Scale
+ UI_Int_Value
;
863 -- Case of real literal to be returned
865 if Point_Scanned
then
866 Token
:= Tok_Real_Literal
;
867 Real_Literal_Value
:=
873 -- Case of integer literal to be returned
876 Token
:= Tok_Integer_Literal
;
879 Int_Literal_Value
:= UI_Num_Value
;
881 -- Avoid doing possibly expensive calculations in cases like
882 -- parsing 163E800_000# when semantics will not be done anyway.
883 -- This is especially useful when parsing garbled input.
885 elsif Operating_Mode
/= Check_Syntax
886 and then (Serious_Errors_Detected
= 0 or else Try_Semantics
)
888 Int_Literal_Value
:= UI_Num_Value
* UI_Base
** UI_Scale
;
891 Int_Literal_Value
:= No_Uint
;
895 if Checksum_Accumulate_Token_Checksum
then
896 Accumulate_Token_Checksum
;
908 Delimiter
: Character;
909 -- Delimiter (first character of string)
912 -- Current source program character
915 -- Current character code value
918 -- Error flag for Scan_Wide call
920 String_Start
: Source_Ptr
;
921 -- Point to first character of string
923 procedure Error_Bad_String_Char
;
924 -- Signal bad character in string/character literal. On entry
925 -- Scan_Ptr points to the improper character encountered during the
926 -- scan. Scan_Ptr is not modified, so it still points to the bad
927 -- character on return.
929 procedure Error_Unterminated_String
;
930 -- Procedure called if a line terminator character is encountered
931 -- during scanning a string, meaning that the string is not properly
934 procedure Set_String
;
935 -- Procedure used to distinguish between string and operator symbol.
936 -- On entry the string has been scanned out, and its characters start
937 -- at Token_Ptr and end one character before Scan_Ptr. On exit Token
938 -- is set to Tok_String_Literal/Tok_Operator_Symbol as appropriate,
939 -- and Token_Node is appropriately initialized. In addition, in the
940 -- operator symbol case, Token_Name is appropriately set, and the
941 -- flags [Wide_]Wide_Character_Found are set appropriately.
943 ---------------------------
944 -- Error_Bad_String_Char --
945 ---------------------------
947 procedure Error_Bad_String_Char
is
948 C
: constant Character := Source
(Scan_Ptr
);
952 Error_Msg_S
("horizontal tab not allowed in string");
954 elsif C
= VT
or else C
= FF
then
955 Error_Msg_S
("format effector not allowed in string");
957 elsif C
in Upper_Half_Character
then
958 Error_Msg_S
("(Ada 83) upper half character not allowed");
961 Error_Msg_S
("control character not allowed in string");
963 end Error_Bad_String_Char
;
965 -------------------------------
966 -- Error_Unterminated_String --
967 -------------------------------
969 procedure Error_Unterminated_String
is
973 -- An interesting little refinement. Consider the following
976 -- A := "this is an unterminated string;
977 -- A := "this is an unterminated string &
978 -- P(A, "this is a parameter that didn't get terminated);
979 -- P("this is a parameter that didn't get terminated, A);
981 -- We fiddle a little to do slightly better placement in these
982 -- cases also if there is white space at the end of the line we
983 -- place the flag at the start of this white space, not at the
984 -- end. Note that we only have to test for blanks, since tabs
985 -- aren't allowed in strings in the first place and would have
986 -- caused an error message.
988 -- Two more cases that we treat specially are:
990 -- A := "this string uses the wrong terminator'
991 -- A := "this string uses the wrong terminator' &
993 -- In these cases we give a different error message as well
995 -- We actually reposition the scan pointer to the point where we
996 -- place the flag in these cases, since it seems a better bet on
997 -- the original intention.
999 while Source
(Scan_Ptr
- 1) = ' '
1000 or else Source
(Scan_Ptr
- 1) = '&'
1002 Scan_Ptr
:= Scan_Ptr
- 1;
1003 Unstore_String_Char
;
1006 -- Check for case of incorrect string terminator, but single quote
1007 -- is not considered incorrect if the opening terminator misused
1008 -- a single quote (error message already given).
1011 and then Source
(Scan_Ptr
- 1) = '''
1013 Unstore_String_Char
;
1015 ("incorrect string terminator character", Scan_Ptr
- 1);
1019 -- Backup over semicolon or right-paren/semicolon sequence
1021 if Source
(Scan_Ptr
- 1) = ';' then
1022 Scan_Ptr
:= Scan_Ptr
- 1;
1023 Unstore_String_Char
;
1025 if Source
(Scan_Ptr
- 1) = ')' then
1026 Scan_Ptr
:= Scan_Ptr
- 1;
1027 Unstore_String_Char
;
1031 -- See if there is a comma in the string, if so, guess that
1032 -- the first comma terminates the string.
1035 while S
< Scan_Ptr
loop
1036 if Source
(S
) = ',' then
1037 while Scan_Ptr
> S
loop
1038 Scan_Ptr
:= Scan_Ptr
- 1;
1039 Unstore_String_Char
;
1048 -- Now we have adjusted the scan pointer, give message
1050 Error_Msg_S
-- CODEFIX
1051 ("missing string quote");
1052 end Error_Unterminated_String
;
1058 procedure Set_String
is
1059 Slen
: constant Int
:= Int
(Scan_Ptr
- Token_Ptr
- 2);
1065 -- Token_Name is currently set to Error_Name. The following
1066 -- section of code resets Token_Name to the proper Name_Op_xx
1067 -- value if the string is a valid operator symbol, otherwise it is
1068 -- left set to Error_Name.
1071 C1
:= Source
(Token_Ptr
+ 1);
1075 Token_Name
:= Name_Op_Eq
;
1078 Token_Name
:= Name_Op_Gt
;
1081 Token_Name
:= Name_Op_Lt
;
1084 Token_Name
:= Name_Op_Add
;
1087 Token_Name
:= Name_Op_Subtract
;
1090 Token_Name
:= Name_Op_Concat
;
1093 Token_Name
:= Name_Op_Multiply
;
1096 Token_Name
:= Name_Op_Divide
;
1103 C1
:= Source
(Token_Ptr
+ 1);
1104 C2
:= Source
(Token_Ptr
+ 2);
1106 if C1
= '*' and then C2
= '*' then
1107 Token_Name
:= Name_Op_Expon
;
1112 Token_Name
:= Name_Op_Ne
;
1114 Token_Name
:= Name_Op_Le
;
1116 Token_Name
:= Name_Op_Ge
;
1119 elsif (C1
= 'O' or else C1
= 'o') and then -- OR
1120 (C2
= 'R' or else C2
= 'r')
1122 Token_Name
:= Name_Op_Or
;
1126 C1
:= Source
(Token_Ptr
+ 1);
1127 C2
:= Source
(Token_Ptr
+ 2);
1128 C3
:= Source
(Token_Ptr
+ 3);
1130 if (C1
= 'A' or else C1
= 'a') and then -- AND
1131 (C2
= 'N' or else C2
= 'n') and then
1132 (C3
= 'D' or else C3
= 'd')
1134 Token_Name
:= Name_Op_And
;
1136 elsif (C1
= 'A' or else C1
= 'a') and then -- ABS
1137 (C2
= 'B' or else C2
= 'b') and then
1138 (C3
= 'S' or else C3
= 's')
1140 Token_Name
:= Name_Op_Abs
;
1142 elsif (C1
= 'M' or else C1
= 'm') and then -- MOD
1143 (C2
= 'O' or else C2
= 'o') and then
1144 (C3
= 'D' or else C3
= 'd')
1146 Token_Name
:= Name_Op_Mod
;
1148 elsif (C1
= 'N' or else C1
= 'n') and then -- NOT
1149 (C2
= 'O' or else C2
= 'o') and then
1150 (C3
= 'T' or else C3
= 't')
1152 Token_Name
:= Name_Op_Not
;
1154 elsif (C1
= 'R' or else C1
= 'r') and then -- REM
1155 (C2
= 'E' or else C2
= 'e') and then
1156 (C3
= 'M' or else C3
= 'm')
1158 Token_Name
:= Name_Op_Rem
;
1160 elsif (C1
= 'X' or else C1
= 'x') and then -- XOR
1161 (C2
= 'O' or else C2
= 'o') and then
1162 (C3
= 'R' or else C3
= 'r')
1164 Token_Name
:= Name_Op_Xor
;
1169 -- If it is an operator symbol, then Token_Name is set. If it is
1170 -- some other string value, then Token_Name still contains
1173 if Token_Name
= Error_Name
then
1174 Token
:= Tok_String_Literal
;
1177 Token
:= Tok_Operator_Symbol
;
1181 -- Start of processing for Slit
1184 -- On entry, Scan_Ptr points to the opening character of the string
1185 -- which is either a percent, double quote, or apostrophe (single
1186 -- quote). The latter case is an error detected by the character
1189 String_Start
:= Scan_Ptr
;
1191 Delimiter
:= Source
(Scan_Ptr
);
1192 Accumulate_Checksum
(Delimiter
);
1195 Wide_Character_Found
:= False;
1196 Wide_Wide_Character_Found
:= False;
1197 Scan_Ptr
:= Scan_Ptr
+ 1;
1199 -- Loop to scan out characters of string literal
1202 C
:= Source
(Scan_Ptr
);
1204 if C
= Delimiter
then
1205 Accumulate_Checksum
(C
);
1206 Scan_Ptr
:= Scan_Ptr
+ 1;
1207 exit when Source
(Scan_Ptr
) /= Delimiter
;
1208 Code
:= Get_Char_Code
(C
);
1209 Accumulate_Checksum
(C
);
1210 Scan_Ptr
:= Scan_Ptr
+ 1;
1213 if C
= '"' and then Delimiter
= '%' then
1215 ("quote not allowed in percent delimited string");
1216 Code
:= Get_Char_Code
(C
);
1217 Scan_Ptr
:= Scan_Ptr
+ 1;
1219 elsif Start_Of_Wide_Character
then
1221 Scan_Wide
(Source
, Scan_Ptr
, Code
, Err
);
1224 Error_Illegal_Wide_Character
;
1225 Code
:= Get_Char_Code
(' ');
1228 Accumulate_Checksum
(Code
);
1230 -- In Ada 95 mode we allow any wide characters in a string
1231 -- but in Ada 2005, the set of characters allowed has been
1232 -- restricted to graphic characters.
1234 if Ada_Version
>= Ada_2005
1235 and then Is_UTF_32_Non_Graphic
(UTF_32
(Code
))
1238 ("(Ada 2005) non-graphic character not permitted " &
1239 "in string literal", Wptr
);
1243 Accumulate_Checksum
(C
);
1245 if C
not in Graphic_Character
then
1246 if C
in Line_Terminator
then
1247 Error_Unterminated_String
;
1250 elsif C
in Upper_Half_Character
then
1251 if Ada_Version
= Ada_83
then
1252 Error_Bad_String_Char
;
1256 Error_Bad_String_Char
;
1260 Code
:= Get_Char_Code
(C
);
1261 Scan_Ptr
:= Scan_Ptr
+ 1;
1265 Store_String_Char
(Code
);
1267 if not In_Character_Range
(Code
) then
1268 if In_Wide_Character_Range
(Code
) then
1269 Wide_Character_Found
:= True;
1271 Wide_Wide_Character_Found
:= True;
1276 String_Literal_Id
:= End_String
;
1281 ----------------------------------
1282 -- Skip_Other_Format_Characters --
1283 ----------------------------------
1285 procedure Skip_Other_Format_Characters
is
1291 while Start_Of_Wide_Character
loop
1293 Scan_Wide
(Source
, Scan_Ptr
, Code
, Err
);
1295 if not Is_UTF_32_Other
(UTF_32
(Code
)) then
1300 end Skip_Other_Format_Characters
;
1302 -----------------------------
1303 -- Start_Of_Wide_Character --
1304 -----------------------------
1306 function Start_Of_Wide_Character
return Boolean is
1307 C
: constant Character := Source
(Scan_Ptr
);
1310 -- ESC encoding method with ESC present
1313 and then Wide_Character_Encoding_Method
in WC_ESC_Encoding_Method
1317 -- Upper half character with upper half encoding
1319 elsif C
in Upper_Half_Character
and then Upper_Half_Encoding
then
1322 -- Brackets encoding
1325 and then Source
(Scan_Ptr
+ 1) = '"'
1326 and then Identifier_Char
(Source
(Scan_Ptr
+ 2))
1330 -- Not the start of a wide character
1335 end Start_Of_Wide_Character
;
1337 -- Start of processing for Scan
1340 Prev_Token
:= Token
;
1341 Prev_Token_Ptr
:= Token_Ptr
;
1342 Token_Name
:= Error_Name
;
1344 -- The following loop runs more than once only if a format effector
1345 -- (tab, vertical tab, form feed, line feed, carriage return) is
1346 -- encountered and skipped, or some error situation, such as an
1347 -- illegal character, is encountered.
1349 <<Scan_Next_Character
>>
1352 -- Skip past blanks, loop is opened up for speed
1354 while Source
(Scan_Ptr
) = ' ' loop
1355 if Source
(Scan_Ptr
+ 1) /= ' ' then
1356 Scan_Ptr
:= Scan_Ptr
+ 1;
1360 if Source
(Scan_Ptr
+ 2) /= ' ' then
1361 Scan_Ptr
:= Scan_Ptr
+ 2;
1365 if Source
(Scan_Ptr
+ 3) /= ' ' then
1366 Scan_Ptr
:= Scan_Ptr
+ 3;
1370 if Source
(Scan_Ptr
+ 4) /= ' ' then
1371 Scan_Ptr
:= Scan_Ptr
+ 4;
1375 if Source
(Scan_Ptr
+ 5) /= ' ' then
1376 Scan_Ptr
:= Scan_Ptr
+ 5;
1380 if Source
(Scan_Ptr
+ 6) /= ' ' then
1381 Scan_Ptr
:= Scan_Ptr
+ 6;
1385 if Source
(Scan_Ptr
+ 7) /= ' ' then
1386 Scan_Ptr
:= Scan_Ptr
+ 7;
1390 Scan_Ptr
:= Scan_Ptr
+ 8;
1393 -- We are now at a non-blank character, which is the first character
1394 -- of the token we will scan, and hence the value of Token_Ptr.
1396 Token_Ptr
:= Scan_Ptr
;
1398 -- Here begins the main case statement which transfers control on the
1399 -- basis of the non-blank character we have encountered.
1401 case Source
(Scan_Ptr
) is
1403 -- Line terminator characters
1405 when CR | LF | FF | VT
=>
1406 goto Scan_Line_Terminator
;
1408 -- Horizontal tab, just skip past it
1415 Scan_Ptr
:= Scan_Ptr
+ 1;
1417 -- End of file character, treated as an end of file only if it is
1418 -- the last character in the buffer, otherwise it is ignored.
1421 if Scan_Ptr
= Source_Last
(Current_Source_File
) then
1431 Scan_Ptr
:= Scan_Ptr
+ 1;
1437 Accumulate_Checksum
('&');
1439 if Source
(Scan_Ptr
+ 1) = '&' then
1440 Error_Msg_S
-- CODEFIX
1441 ("'&'& should be `AND THEN`");
1442 Scan_Ptr
:= Scan_Ptr
+ 2;
1447 Scan_Ptr
:= Scan_Ptr
+ 1;
1448 Token
:= Tok_Ampersand
;
1452 -- Asterisk (can be multiplication operator or double asterisk which
1453 -- is the exponentiation compound delimiter).
1456 Accumulate_Checksum
('*');
1458 if Source
(Scan_Ptr
+ 1) = '*' then
1459 Accumulate_Checksum
('*');
1460 Scan_Ptr
:= Scan_Ptr
+ 2;
1461 Token
:= Tok_Double_Asterisk
;
1465 Scan_Ptr
:= Scan_Ptr
+ 1;
1466 Token
:= Tok_Asterisk
;
1470 -- Colon, which can either be an isolated colon, or part of an
1471 -- assignment compound delimiter.
1474 Accumulate_Checksum
(':');
1476 if Double_Char_Token
('=') then
1477 Token
:= Tok_Colon_Equal
;
1480 Style
.Check_Colon_Equal
;
1485 elsif Source
(Scan_Ptr
+ 1) = '-'
1486 and then Source
(Scan_Ptr
+ 2) /= '-'
1488 Token
:= Tok_Colon_Equal
;
1489 Error_Msg
-- CODEFIX
1490 (":- should be :=", Scan_Ptr
);
1491 Scan_Ptr
:= Scan_Ptr
+ 2;
1495 Scan_Ptr
:= Scan_Ptr
+ 1;
1508 Accumulate_Checksum
('(');
1509 Scan_Ptr
:= Scan_Ptr
+ 1;
1510 Token
:= Tok_Left_Paren
;
1513 Style
.Check_Left_Paren
;
1521 if Source
(Scan_Ptr
+ 1) = '"' then
1522 goto Scan_Wide_Character
;
1525 Error_Msg_S
("illegal character, replaced by ""(""");
1526 Scan_Ptr
:= Scan_Ptr
+ 1;
1527 Token
:= Tok_Left_Paren
;
1534 Error_Msg_S
("illegal character, replaced by ""(""");
1535 Scan_Ptr
:= Scan_Ptr
+ 1;
1536 Token
:= Tok_Left_Paren
;
1542 Accumulate_Checksum
(',');
1543 Scan_Ptr
:= Scan_Ptr
+ 1;
1552 -- Dot, which is either an isolated period, or part of a double dot
1553 -- compound delimiter sequence. We also check for the case of a
1554 -- digit following the period, to give a better error message.
1557 Accumulate_Checksum
('.');
1559 if Double_Char_Token
('.') then
1560 Token
:= Tok_Dot_Dot
;
1563 Style
.Check_Dot_Dot
;
1568 elsif Source
(Scan_Ptr
+ 1) in '0' .. '9' then
1569 Error_Msg_S
("numeric literal cannot start with point");
1570 Scan_Ptr
:= Scan_Ptr
+ 1;
1573 Scan_Ptr
:= Scan_Ptr
+ 1;
1578 -- Equal, which can either be an equality operator, or part of the
1579 -- arrow (=>) compound delimiter.
1582 Accumulate_Checksum
('=');
1584 if Double_Char_Token
('>') then
1593 elsif Source
(Scan_Ptr
+ 1) = '=' then
1594 Error_Msg_S
-- CODEFIX
1596 Scan_Ptr
:= Scan_Ptr
+ 1;
1599 Scan_Ptr
:= Scan_Ptr
+ 1;
1603 -- Greater than, which can be a greater than operator, greater than
1604 -- or equal operator, or first character of a right label bracket.
1607 Accumulate_Checksum
('>');
1609 if Double_Char_Token
('=') then
1610 Token
:= Tok_Greater_Equal
;
1613 elsif Double_Char_Token
('>') then
1614 Token
:= Tok_Greater_Greater
;
1618 Scan_Ptr
:= Scan_Ptr
+ 1;
1619 Token
:= Tok_Greater
;
1623 -- Less than, which can be a less than operator, less than or equal
1624 -- operator, or the first character of a left label bracket, or the
1625 -- first character of a box (<>) compound delimiter.
1628 Accumulate_Checksum
('<');
1630 if Double_Char_Token
('=') then
1631 Token
:= Tok_Less_Equal
;
1634 elsif Double_Char_Token
('>') then
1643 elsif Double_Char_Token
('<') then
1644 Token
:= Tok_Less_Less
;
1648 Scan_Ptr
:= Scan_Ptr
+ 1;
1653 -- Minus, which is either a subtraction operator, or the first
1654 -- character of double minus starting a comment
1656 when '-' => Minus_Case
: begin
1657 if Source
(Scan_Ptr
+ 1) = '>' then
1658 Error_Msg_S
("invalid token");
1659 Scan_Ptr
:= Scan_Ptr
+ 2;
1663 elsif Source
(Scan_Ptr
+ 1) /= '-' then
1664 Accumulate_Checksum
('-');
1665 Scan_Ptr
:= Scan_Ptr
+ 1;
1671 else -- Source (Scan_Ptr + 1) = '-' then
1673 Style
.Check_Comment
;
1676 Scan_Ptr
:= Scan_Ptr
+ 2;
1678 -- If we are in preprocessor mode with Replace_In_Comments set,
1679 -- then we return the "--" as a token on its own.
1681 if Replace_In_Comments
then
1682 Token
:= Tok_Comment
;
1686 -- Otherwise scan out the comment
1688 Start_Of_Comment
:= Scan_Ptr
;
1690 -- Loop to scan comment (this loop runs more than once only if
1691 -- a horizontal tab or other non-graphic character is scanned)
1694 -- Scan to non graphic character (opened up for speed)
1696 -- Note that we just eat left brackets, which means that
1697 -- bracket notation cannot be used for end of line
1698 -- characters in comments. This seems a reasonable choice,
1699 -- since no one would ever use brackets notation in a real
1700 -- program in this situation, and if we allow brackets
1701 -- notation, we forbid some valid comments which contain a
1702 -- brackets sequence that happens to match an end of line
1706 exit when Source
(Scan_Ptr
) not in Graphic_Character
;
1707 Scan_Ptr
:= Scan_Ptr
+ 1;
1708 exit when Source
(Scan_Ptr
) not in Graphic_Character
;
1709 Scan_Ptr
:= Scan_Ptr
+ 1;
1710 exit when Source
(Scan_Ptr
) not in Graphic_Character
;
1711 Scan_Ptr
:= Scan_Ptr
+ 1;
1712 exit when Source
(Scan_Ptr
) not in Graphic_Character
;
1713 Scan_Ptr
:= Scan_Ptr
+ 1;
1714 exit when Source
(Scan_Ptr
) not in Graphic_Character
;
1715 Scan_Ptr
:= Scan_Ptr
+ 1;
1718 -- Keep going if horizontal tab
1720 if Source
(Scan_Ptr
) = HT
then
1725 Scan_Ptr
:= Scan_Ptr
+ 1;
1727 -- Terminate scan of comment if line terminator
1729 elsif Source
(Scan_Ptr
) in Line_Terminator
then
1732 -- Terminate scan of comment if end of file encountered
1733 -- (embedded EOF character or real last character in file)
1735 elsif Source
(Scan_Ptr
) = EOF
then
1738 -- If we have a wide character, we have to scan it out,
1739 -- because it might be a legitimate line terminator
1741 elsif Start_Of_Wide_Character
then
1743 Wptr
: constant Source_Ptr
:= Scan_Ptr
;
1748 Scan_Wide
(Source
, Scan_Ptr
, Code
, Err
);
1750 -- If not well formed wide character, then just skip
1751 -- past it and ignore it.
1754 Scan_Ptr
:= Wptr
+ 1;
1756 -- If UTF_32 terminator, terminate comment scan
1758 elsif Is_UTF_32_Line_Terminator
(UTF_32
(Code
)) then
1764 -- Keep going if character in 80-FF range, or is ESC. These
1765 -- characters are allowed in comments by RM-2.1(1), 2.7(2).
1766 -- They are allowed even in Ada 83 mode according to the
1767 -- approved AI. ESC was added to the AI in June 93.
1769 elsif Source
(Scan_Ptr
) in Upper_Half_Character
1770 or else Source
(Scan_Ptr
) = ESC
1772 Scan_Ptr
:= Scan_Ptr
+ 1;
1774 -- Otherwise we have an illegal comment character
1777 Error_Illegal_Character
;
1781 -- Note that, except when comments are tokens, we do NOT
1782 -- execute a return here, instead we fall through to reexecute
1783 -- the scan loop to look for a token.
1785 if Comment_Is_Token
then
1786 Name_Len
:= Integer (Scan_Ptr
- Start_Of_Comment
);
1787 Name_Buffer
(1 .. Name_Len
) :=
1788 String (Source
(Start_Of_Comment
.. Scan_Ptr
- 1));
1789 Comment_Id
:= Name_Find
;
1790 Token
:= Tok_Comment
;
1794 -- If the SPARK restriction is set for this unit, then generate
1795 -- a token Tok_SPARK_Hide for a SPARK HIDE directive.
1797 if Restriction_Check_Required
(SPARK_05
)
1798 and then Source
(Start_Of_Comment
) = '#'
1801 Scan_SPARK_Ptr
: Source_Ptr
;
1804 Scan_SPARK_Ptr
:= Start_Of_Comment
+ 1;
1808 while Source
(Scan_SPARK_Ptr
) = ' '
1809 or else Source
(Scan_SPARK_Ptr
) = HT
1811 Scan_SPARK_Ptr
:= Scan_SPARK_Ptr
+ 1;
1814 -- Recognize HIDE directive. SPARK input cannot be
1815 -- encoded as wide characters, so only deal with
1816 -- lower/upper case.
1818 if (Source
(Scan_SPARK_Ptr
) = 'h'
1819 or else Source
(Scan_SPARK_Ptr
) = 'H')
1820 and then (Source
(Scan_SPARK_Ptr
+ 1) = 'i'
1821 or else Source
(Scan_SPARK_Ptr
+ 1) = 'I')
1822 and then (Source
(Scan_SPARK_Ptr
+ 2) = 'd'
1823 or else Source
(Scan_SPARK_Ptr
+ 2) = 'D')
1824 and then (Source
(Scan_SPARK_Ptr
+ 3) = 'e'
1825 or else Source
(Scan_SPARK_Ptr
+ 3) = 'E')
1826 and then (Source
(Scan_SPARK_Ptr
+ 4) = ' '
1827 or else Source
(Scan_SPARK_Ptr
+ 4) = HT
)
1829 Token
:= Tok_SPARK_Hide
;
1837 -- Double quote or percent starting a string literal
1844 -- Apostrophe. This can either be the start of a character literal,
1845 -- or an isolated apostrophe used in a qualified expression or an
1846 -- attribute. We treat it as a character literal if it does not
1847 -- follow a right parenthesis, identifier, the keyword ALL or
1848 -- a literal. This means that we correctly treat constructs like:
1850 -- A := CHARACTER'('A');
1852 -- Note that RM-2.2(7) does not require a separator between
1853 -- "CHARACTER" and "'" in the above.
1855 when ''' => Char_Literal_Case
: declare
1860 Accumulate_Checksum
(''');
1861 Scan_Ptr
:= Scan_Ptr
+ 1;
1863 -- Here is where we make the test to distinguish the cases. Treat
1864 -- as apostrophe if previous token is an identifier, right paren
1865 -- or the reserved word "all" (latter case as in A.all'Address)
1866 -- (or the reserved word "project" in project files). Also treat
1867 -- it as apostrophe after a literal (this catches some legitimate
1868 -- cases, like A."abs"'Address, and also gives better error
1869 -- behavior for impossible cases like 123'xxx).
1871 if Prev_Token
= Tok_Identifier
1872 or else Prev_Token
= Tok_Right_Paren
1873 or else Prev_Token
= Tok_All
1874 or else Prev_Token
= Tok_Project
1875 or else Prev_Token
in Token_Class_Literal
1877 Token
:= Tok_Apostrophe
;
1880 Style
.Check_Apostrophe
;
1885 -- Otherwise the apostrophe starts a character literal
1888 -- Case of wide character literal
1890 if Start_Of_Wide_Character
then
1892 Scan_Wide
(Source
, Scan_Ptr
, Code
, Err
);
1893 Accumulate_Checksum
(Code
);
1896 Error_Illegal_Wide_Character
;
1897 Code
:= Character'Pos (' ');
1899 -- In Ada 95 mode we allow any wide character in a character
1900 -- literal, but in Ada 2005, the set of characters allowed
1901 -- is restricted to graphic characters.
1903 elsif Ada_Version
>= Ada_2005
1904 and then Is_UTF_32_Non_Graphic
(UTF_32
(Code
))
1906 Error_Msg
-- CODEFIX????
1907 ("(Ada 2005) non-graphic character not permitted " &
1908 "in character literal", Wptr
);
1911 if Source
(Scan_Ptr
) /= ''' then
1912 Error_Msg_S
("missing apostrophe");
1914 Scan_Ptr
:= Scan_Ptr
+ 1;
1917 -- If we do not find a closing quote in the expected place then
1918 -- assume that we have a misguided attempt at a string literal.
1920 -- However, if previous token is RANGE, then we return an
1921 -- apostrophe instead since this gives better error recovery
1923 elsif Source
(Scan_Ptr
+ 1) /= ''' then
1924 if Prev_Token
= Tok_Range
then
1925 Token
:= Tok_Apostrophe
;
1929 Scan_Ptr
:= Scan_Ptr
- 1;
1931 ("strings are delimited by double quote character");
1937 -- Otherwise we have a (non-wide) character literal
1940 Accumulate_Checksum
(Source
(Scan_Ptr
));
1942 if Source
(Scan_Ptr
) not in Graphic_Character
then
1943 if Source
(Scan_Ptr
) in Upper_Half_Character
then
1944 if Ada_Version
= Ada_83
then
1945 Error_Illegal_Character
;
1949 Error_Illegal_Character
;
1953 Code
:= Get_Char_Code
(Source
(Scan_Ptr
));
1954 Scan_Ptr
:= Scan_Ptr
+ 2;
1957 -- Fall through here with Scan_Ptr updated past the closing
1958 -- quote, and Code set to the Char_Code value for the literal
1960 Accumulate_Checksum
(''');
1961 Token
:= Tok_Char_Literal
;
1962 Set_Character_Literal_Name
(Code
);
1963 Token_Name
:= Name_Find
;
1964 Character_Code
:= Code
;
1968 end Char_Literal_Case
;
1970 -- Right parenthesis
1973 Accumulate_Checksum
(')');
1974 Scan_Ptr
:= Scan_Ptr
+ 1;
1975 Token
:= Tok_Right_Paren
;
1978 Style
.Check_Right_Paren
;
1983 -- Right bracket or right brace, treated as right paren
1986 Error_Msg_S
("illegal character, replaced by "")""");
1987 Scan_Ptr
:= Scan_Ptr
+ 1;
1988 Token
:= Tok_Right_Paren
;
1991 -- Slash (can be division operator or first character of not equal)
1994 Accumulate_Checksum
('/');
1996 if Double_Char_Token
('=') then
1997 Token
:= Tok_Not_Equal
;
2000 Scan_Ptr
:= Scan_Ptr
+ 1;
2008 Accumulate_Checksum
(';');
2009 Scan_Ptr
:= Scan_Ptr
+ 1;
2010 Token
:= Tok_Semicolon
;
2013 Style
.Check_Semicolon
;
2020 when '|' => Vertical_Bar_Case
: begin
2021 Accumulate_Checksum
('|');
2023 -- Special check for || to give nice message
2025 if Source
(Scan_Ptr
+ 1) = '|' then
2026 Error_Msg_S
-- CODEFIX
2027 ("""'|'|"" should be `OR ELSE`");
2028 Scan_Ptr
:= Scan_Ptr
+ 2;
2033 Scan_Ptr
:= Scan_Ptr
+ 1;
2034 Token
:= Tok_Vertical_Bar
;
2037 Style
.Check_Vertical_Bar
;
2043 end Vertical_Bar_Case
;
2045 -- Exclamation, replacement character for vertical bar
2047 when '!' => Exclamation_Case
: begin
2048 Accumulate_Checksum
('!');
2050 if Source
(Scan_Ptr
+ 1) = '=' then
2051 Error_Msg_S
-- CODEFIX
2052 ("'!= should be /=");
2053 Scan_Ptr
:= Scan_Ptr
+ 2;
2054 Token
:= Tok_Not_Equal
;
2058 Scan_Ptr
:= Scan_Ptr
+ 1;
2059 Token
:= Tok_Vertical_Bar
;
2063 end Exclamation_Case
;
2067 when '+' => Plus_Case
: begin
2068 Accumulate_Checksum
('+');
2069 Scan_Ptr
:= Scan_Ptr
+ 1;
2074 -- Digits starting a numeric literal
2078 -- First a bit of a scan ahead to see if we have a case of an
2079 -- identifier starting with a digit (remembering exponent case).
2082 C
: constant Character := Source
(Scan_Ptr
+ 1);
2085 -- OK literal if digit followed by digit or underscore
2087 if C
in '0' .. '9' or else C
= '_' then
2090 -- OK literal if digit not followed by identifier char
2092 elsif not Identifier_Char
(C
) then
2095 -- OK literal if digit followed by e/E followed by digit/sign.
2096 -- We also allow underscore after the E, which is an error, but
2097 -- better handled by Nlit than deciding this is an identifier.
2099 elsif (C
= 'e' or else C
= 'E')
2100 and then (Source
(Scan_Ptr
+ 2) in '0' .. '9'
2101 or else Source
(Scan_Ptr
+ 2) = '+'
2102 or else Source
(Scan_Ptr
+ 2) = '-'
2103 or else Source
(Scan_Ptr
+ 2) = '_')
2107 -- Here we have what really looks like an identifier that
2108 -- starts with a digit, so give error msg.
2111 Error_Msg_S
("identifier may not start with digit");
2113 Underline_Found
:= False;
2114 Name_Buffer
(1) := Source
(Scan_Ptr
);
2115 Accumulate_Checksum
(Name_Buffer
(1));
2116 Scan_Ptr
:= Scan_Ptr
+ 1;
2117 goto Scan_Identifier
;
2121 -- Here we have an OK integer literal
2125 -- Check for proper delimiter, ignoring other format characters
2127 Skip_Other_Format_Characters
;
2129 if Identifier_Char
(Source
(Scan_Ptr
)) then
2131 ("delimiter required between literal and identifier");
2137 -- Lower case letters
2141 Underline_Found
:= False;
2142 Name_Buffer
(1) := Source
(Scan_Ptr
);
2143 Accumulate_Checksum
(Name_Buffer
(1));
2144 Scan_Ptr
:= Scan_Ptr
+ 1;
2145 goto Scan_Identifier
;
2147 -- Upper case letters
2151 Underline_Found
:= False;
2153 Character'Val (Character'Pos (Source
(Scan_Ptr
)) + 32);
2154 Accumulate_Checksum
(Name_Buffer
(1));
2155 Scan_Ptr
:= Scan_Ptr
+ 1;
2156 goto Scan_Identifier
;
2158 -- Underline character
2161 if Special_Characters
('_') then
2162 Token_Ptr
:= Scan_Ptr
;
2163 Scan_Ptr
:= Scan_Ptr
+ 1;
2164 Token
:= Tok_Special
;
2165 Special_Character
:= '_';
2169 Error_Msg_S
("identifier cannot start with underline");
2171 Name_Buffer
(1) := '_';
2172 Scan_Ptr
:= Scan_Ptr
+ 1;
2173 Underline_Found
:= False;
2174 goto Scan_Identifier
;
2176 -- Space (not possible, because we scanned past blanks)
2179 raise Program_Error
;
2181 -- Characters in top half of ASCII 8-bit chart
2183 when Upper_Half_Character
=>
2185 -- Wide character case
2187 if Upper_Half_Encoding
then
2188 goto Scan_Wide_Character
;
2190 -- Otherwise we have OK Latin-1 character
2193 -- Upper half characters may possibly be identifier letters
2194 -- but can never be digits, so Identifier_Char can be used to
2195 -- test for a valid start of identifier character.
2197 if Identifier_Char
(Source
(Scan_Ptr
)) then
2199 Underline_Found
:= False;
2200 goto Scan_Identifier
;
2202 Error_Illegal_Character
;
2208 -- ESC character, possible start of identifier if wide characters
2209 -- using ESC encoding are allowed in identifiers, which we can
2210 -- tell by looking at the Identifier_Char flag for ESC, which is
2211 -- only true if these conditions are met. In Ada 2005 mode, may
2212 -- also be valid UTF_32 space or line terminator character.
2214 if Identifier_Char
(ESC
) then
2216 goto Scan_Wide_Character
;
2218 Error_Illegal_Character
;
2221 -- Invalid control characters
2223 when NUL | SOH | STX | ETX | EOT | ENQ | ACK | BEL | BS | ASCII
.SO |
2224 SI | DLE | DC1 | DC2 | DC3 | DC4 | NAK | SYN | ETB | CAN |
2225 EM | FS | GS | RS | US | DEL
2227 Error_Illegal_Character
;
2229 -- Invalid graphic characters
2231 when '#' |
'$' |
'?' |
'@' |
'`' |
'\' |
'^' |
'~' =>
2233 -- If Set_Special_Character has been called for this character,
2234 -- set Scans.Special_Character and return a Special token.
2236 if Special_Characters
(Source
(Scan_Ptr
)) then
2237 Token_Ptr
:= Scan_Ptr
;
2238 Token
:= Tok_Special
;
2239 Special_Character
:= Source
(Scan_Ptr
);
2240 Scan_Ptr
:= Scan_Ptr
+ 1;
2243 -- Check for something looking like a preprocessor directive
2245 elsif Source
(Scan_Ptr
) = '#'
2246 and then (Source
(Scan_Ptr
+ 1 .. Scan_Ptr
+ 2) = "if"
2248 Source
(Scan_Ptr
+ 1 .. Scan_Ptr
+ 5) = "elsif"
2250 Source
(Scan_Ptr
+ 1 .. Scan_Ptr
+ 4) = "else"
2252 Source
(Scan_Ptr
+ 1 .. Scan_Ptr
+ 3) = "end")
2255 ("preprocessor directive ignored, preprocessor not active");
2257 -- Skip to end of line
2260 if Source
(Scan_Ptr
) in Graphic_Character
2262 Source
(Scan_Ptr
) = HT
2264 Scan_Ptr
:= Scan_Ptr
+ 1;
2266 -- Done if line terminator or EOF
2268 elsif Source
(Scan_Ptr
) in Line_Terminator
2270 Source
(Scan_Ptr
) = EOF
2274 -- If we have a wide character, we have to scan it out,
2275 -- because it might be a legitimate line terminator
2277 elsif Start_Of_Wide_Character
then
2279 Wptr
: constant Source_Ptr
:= Scan_Ptr
;
2284 Scan_Wide
(Source
, Scan_Ptr
, Code
, Err
);
2286 -- If not well formed wide character, then just skip
2287 -- past it and ignore it.
2290 Scan_Ptr
:= Wptr
+ 1;
2292 -- If UTF_32 terminator, terminate comment scan
2294 elsif Is_UTF_32_Line_Terminator
(UTF_32
(Code
)) then
2300 -- Else keep going (don't worry about bad comment chars
2301 -- in this context, we just want to find the end of line.
2304 Scan_Ptr
:= Scan_Ptr
+ 1;
2308 -- Otherwise, this is an illegal character
2311 Error_Illegal_Character
;
2314 -- End switch on non-blank character
2318 -- End loop past format effectors. The exit from this loop is by
2319 -- executing a return statement following completion of token scan
2320 -- (control never falls out of this loop to the code which follows)
2324 -- Wide_Character scanning routine. On entry we have encountered the
2325 -- initial character of a wide character sequence.
2327 <<Scan_Wide_Character
>>
2336 Scan_Wide
(Source
, Scan_Ptr
, Code
, Err
);
2338 -- If bad wide character, signal error and continue scan
2341 Error_Illegal_Wide_Character
;
2342 goto Scan_Next_Character
;
2345 Cat
:= Get_Category
(UTF_32
(Code
));
2347 -- If OK letter, reset scan ptr and go scan identifier
2349 if Is_UTF_32_Letter
(Cat
) then
2352 Underline_Found
:= False;
2353 goto Scan_Identifier
;
2355 -- If OK wide space, ignore and keep scanning (we do not include
2356 -- any ignored spaces in checksum)
2358 elsif Is_UTF_32_Space
(Cat
) then
2359 goto Scan_Next_Character
;
2361 -- If other format character, ignore and keep scanning (again we
2362 -- do not include in the checksum) (this is for AI-0079).
2364 elsif Is_UTF_32_Other
(Cat
) then
2365 goto Scan_Next_Character
;
2367 -- If OK wide line terminator, terminate current line
2369 elsif Is_UTF_32_Line_Terminator
(UTF_32
(Code
)) then
2371 goto Scan_Line_Terminator
;
2373 -- Punctuation is an error (at start of identifier)
2375 elsif Is_UTF_32_Punctuation
(Cat
) then
2376 Error_Msg
("identifier cannot start with punctuation", Wptr
);
2379 Underline_Found
:= False;
2380 goto Scan_Identifier
;
2382 -- Mark character is an error (at start of identifier)
2384 elsif Is_UTF_32_Mark
(Cat
) then
2385 Error_Msg
("identifier cannot start with mark character", Wptr
);
2388 Underline_Found
:= False;
2389 goto Scan_Identifier
;
2391 -- Extended digit character is an error. Could be bad start of
2392 -- identifier or bad literal. Not worth doing too much to try to
2393 -- distinguish these cases, but we will do a little bit.
2395 elsif Is_UTF_32_Digit
(Cat
) then
2397 ("identifier cannot start with digit character", Wptr
);
2400 Underline_Found
:= False;
2401 goto Scan_Identifier
;
2403 -- All other wide characters are illegal here
2406 Error_Illegal_Wide_Character
;
2407 goto Scan_Next_Character
;
2411 -- Routine to scan line terminator. On entry Scan_Ptr points to a
2412 -- character which is one of FF,LR,CR,VT, or one of the wide characters
2413 -- that is treated as a line terminator.
2415 <<Scan_Line_Terminator
>>
2417 -- Check line too long
2421 -- Set Token_Ptr, if End_Of_Line is a token, for the case when it is
2424 if End_Of_Line_Is_Token
then
2425 Token_Ptr
:= Scan_Ptr
;
2432 Skip_Line_Terminators
(Scan_Ptr
, Physical
);
2434 -- If we are at start of physical line, update scan pointers to
2435 -- reflect the start of the new line.
2438 Current_Line_Start
:= Scan_Ptr
;
2439 Start_Column
:= Set_Start_Column
;
2440 First_Non_Blank_Location
:= Scan_Ptr
;
2442 -- If End_Of_Line is a token, we return it as it is a
2445 if End_Of_Line_Is_Token
then
2446 Token
:= Tok_End_Of_Line
;
2452 goto Scan_Next_Character
;
2454 -- Identifier scanning routine. On entry, some initial characters of
2455 -- the identifier may have already been stored in Name_Buffer. If so,
2456 -- Name_Len has the number of characters stored, otherwise Name_Len is
2457 -- set to zero on entry. Underline_Found is also set False on entry.
2461 -- This loop scans as fast as possible past lower half letters and
2462 -- digits, which we expect to be the most common characters.
2465 if Source
(Scan_Ptr
) in 'a' .. 'z'
2466 or else Source
(Scan_Ptr
) in '0' .. '9'
2468 Name_Buffer
(Name_Len
+ 1) := Source
(Scan_Ptr
);
2469 Accumulate_Checksum
(Source
(Scan_Ptr
));
2471 elsif Source
(Scan_Ptr
) in 'A' .. 'Z' then
2472 Name_Buffer
(Name_Len
+ 1) :=
2473 Character'Val (Character'Pos (Source
(Scan_Ptr
)) + 32);
2474 Accumulate_Checksum
(Name_Buffer
(Name_Len
+ 1));
2480 Underline_Found
:= False;
2481 Scan_Ptr
:= Scan_Ptr
+ 1;
2482 Name_Len
:= Name_Len
+ 1;
2485 -- If we fall through, then we have encountered either an underline
2486 -- character, or an extended identifier character (i.e. one from the
2487 -- upper half), or a wide character, or an identifier terminator. The
2488 -- initial test speeds us up in the most common case where we have
2489 -- an identifier terminator. Note that ESC is an identifier character
2490 -- only if a wide character encoding method that uses ESC encoding
2491 -- is active, so if we find an ESC character we know that we have a
2494 if Identifier_Char
(Source
(Scan_Ptr
))
2495 or else (Source
(Scan_Ptr
) in Upper_Half_Character
2496 and then Upper_Half_Encoding
)
2498 -- Case of underline
2500 if Source
(Scan_Ptr
) = '_' then
2501 Accumulate_Checksum
('_');
2503 if Underline_Found
then
2504 Error_No_Double_Underline
;
2506 Underline_Found
:= True;
2507 Name_Len
:= Name_Len
+ 1;
2508 Name_Buffer
(Name_Len
) := '_';
2511 Scan_Ptr
:= Scan_Ptr
+ 1;
2512 goto Scan_Identifier
;
2514 -- Upper half character
2516 elsif Source
(Scan_Ptr
) in Upper_Half_Character
2517 and then not Upper_Half_Encoding
2519 Accumulate_Checksum
(Source
(Scan_Ptr
));
2520 Store_Encoded_Character
2521 (Get_Char_Code
(Fold_Lower
(Source
(Scan_Ptr
))));
2522 Scan_Ptr
:= Scan_Ptr
+ 1;
2523 Underline_Found
:= False;
2524 goto Scan_Identifier
;
2526 -- Left bracket not followed by a quote terminates an identifier.
2527 -- This is an error, but we don't want to give a junk error msg
2528 -- about wide characters in this case!
2530 elsif Source
(Scan_Ptr
) = '['
2531 and then Source
(Scan_Ptr
+ 1) /= '"'
2535 -- We know we have a wide character encoding here (the current
2536 -- character is either ESC, left bracket, or an upper half
2537 -- character depending on the encoding method).
2540 -- Scan out the wide character and insert the appropriate
2541 -- encoding into the name table entry for the identifier.
2551 Scan_Wide
(Source
, Scan_Ptr
, Code
, Err
);
2553 -- If error, signal error
2556 Error_Illegal_Wide_Character
;
2558 -- If the character scanned is a normal identifier
2559 -- character, then we treat it that way.
2561 elsif In_Character_Range
(Code
)
2562 and then Identifier_Char
(Get_Character
(Code
))
2564 Chr
:= Get_Character
(Code
);
2565 Accumulate_Checksum
(Chr
);
2566 Store_Encoded_Character
2567 (Get_Char_Code
(Fold_Lower
(Chr
)));
2568 Underline_Found
:= False;
2570 -- Here if not a normal identifier character
2573 Cat
:= Get_Category
(UTF_32
(Code
));
2575 -- Wide character in Unicode category "Other, Format"
2576 -- is not accepted in an identifier. This is because it
2577 -- it is considered a security risk (AI-0091).
2579 -- However, it is OK for such a character to appear at
2580 -- the end of an identifier.
2582 if Is_UTF_32_Other
(Cat
) then
2583 if not Identifier_Char
(Source
(Scan_Ptr
)) then
2584 goto Scan_Identifier_Complete
;
2587 ("identifier cannot contain other_format "
2588 & "character", Wptr
);
2589 goto Scan_Identifier
;
2592 -- Wide character in category Separator,Space terminates
2594 elsif Is_UTF_32_Space
(Cat
) then
2595 goto Scan_Identifier_Complete
;
2598 -- Here if wide character is part of the identifier
2600 -- Make sure we are allowing wide characters in
2601 -- identifiers. Note that we allow wide character
2602 -- notation for an OK identifier character. This in
2603 -- particular allows bracket or other notation to be
2604 -- used for upper half letters.
2606 -- Wide characters are always allowed in Ada 2005
2608 if Identifier_Character_Set
/= 'w'
2609 and then Ada_Version
< Ada_2005
2612 ("wide character not allowed in identifier", Wptr
);
2615 -- If OK letter, store it folding to upper case. Note
2616 -- that we include the folded letter in the checksum.
2618 if Is_UTF_32_Letter
(Cat
) then
2620 Char_Code
(UTF_32_To_Upper_Case
(UTF_32
(Code
)));
2621 Accumulate_Checksum
(Code
);
2622 Store_Encoded_Character
(Code
);
2623 Underline_Found
:= False;
2625 -- If OK extended digit or mark, then store it
2627 elsif Is_UTF_32_Digit
(Cat
)
2628 or else Is_UTF_32_Mark
(Cat
)
2630 Accumulate_Checksum
(Code
);
2631 Store_Encoded_Character
(Code
);
2632 Underline_Found
:= False;
2634 -- Wide punctuation is also stored, but counts as an
2635 -- underline character for error checking purposes.
2637 elsif Is_UTF_32_Punctuation
(Cat
) then
2638 Accumulate_Checksum
(Code
);
2640 if Underline_Found
then
2642 Cend
: constant Source_Ptr
:= Scan_Ptr
;
2645 Error_No_Double_Underline
;
2650 Store_Encoded_Character
(Code
);
2651 Underline_Found
:= True;
2654 -- Any other wide character is not acceptable
2658 ("invalid wide character in identifier", Wptr
);
2662 goto Scan_Identifier
;
2667 -- Scan of identifier is complete. The identifier is stored in
2668 -- Name_Buffer, and Scan_Ptr points past the last character.
2670 <<Scan_Identifier_Complete
>>
2671 Token_Name
:= Name_Find
;
2673 -- Check for identifier ending with underline or punctuation char
2675 if Underline_Found
then
2676 Underline_Found
:= False;
2678 if Source
(Scan_Ptr
- 1) = '_' then
2680 ("identifier cannot end with underline", Scan_Ptr
- 1);
2683 ("identifier cannot end with punctuation character", Wptr
);
2687 -- We will assume it is an identifier, not a keyword, so that the
2688 -- checksum is independent of the Ada version.
2690 Token
:= Tok_Identifier
;
2692 -- Here is where we check if it was a keyword
2694 if Is_Keyword_Name
(Token_Name
) then
2695 if Opt
.Checksum_GNAT_6_3
then
2696 Token
:= Token_Type
'Val (Get_Name_Table_Byte
(Token_Name
));
2698 if Checksum_Accumulate_Token_Checksum
then
2699 if Checksum_GNAT_5_03
then
2700 Accumulate_Token_Checksum_GNAT_5_03
;
2702 Accumulate_Token_Checksum_GNAT_6_3
;
2707 Accumulate_Token_Checksum
;
2708 Token
:= Token_Type
'Val (Get_Name_Table_Byte
(Token_Name
));
2711 -- Keyword style checks
2715 -- Deal with possible style check for non-lower case keyword,
2716 -- but we don't treat ACCESS, DELTA, DIGITS, RANGE as keywords
2717 -- for this purpose if they appear as attribute designators.
2718 -- Actually we only check the first character for speed.
2720 -- Ada 2005 (AI-284): Do not apply the style check in case of
2721 -- "pragma Interface"
2723 -- Ada 2005 (AI-340): Do not apply the style check in case of
2726 if Source
(Token_Ptr
) <= 'Z'
2727 and then (Prev_Token
/= Tok_Apostrophe
2729 (Token
/= Tok_Access
and then
2730 Token
/= Tok_Delta
and then
2731 Token
/= Tok_Digits
and then
2732 Token
/= Tok_Mod
and then
2733 Token
/= Tok_Range
))
2734 and then (Token
/= Tok_Interface
2736 (Token
= Tok_Interface
2737 and then Prev_Token
/= Tok_Pragma
))
2739 Style
.Non_Lower_Case_Keyword
;
2742 -- Check THEN/ELSE style rules. These do not apply to AND THEN
2743 -- or OR ELSE, and do not apply in if expressions.
2745 if (Token
= Tok_Then
and then Prev_Token
/= Tok_And
)
2747 (Token
= Tok_Else
and then Prev_Token
/= Tok_Or
)
2749 if Inside_If_Expression
= 0 then
2750 Style
.Check_Separate_Stmt_Lines
;
2755 -- We must reset Token_Name since this is not an identifier and
2756 -- if we leave Token_Name set, the parser gets confused because
2757 -- it thinks it is dealing with an identifier instead of the
2758 -- corresponding keyword.
2760 Token_Name
:= No_Name
;
2763 -- It is an identifier after all
2766 if Checksum_Accumulate_Token_Checksum
then
2767 Accumulate_Token_Checksum
;
2775 --------------------------
2776 -- Set_Comment_As_Token --
2777 --------------------------
2779 procedure Set_Comment_As_Token
(Value
: Boolean) is
2781 Comment_Is_Token
:= Value
;
2782 end Set_Comment_As_Token
;
2784 ------------------------------
2785 -- Set_End_Of_Line_As_Token --
2786 ------------------------------
2788 procedure Set_End_Of_Line_As_Token
(Value
: Boolean) is
2790 End_Of_Line_Is_Token
:= Value
;
2791 end Set_End_Of_Line_As_Token
;
2793 ---------------------------
2794 -- Set_Special_Character --
2795 ---------------------------
2797 procedure Set_Special_Character
(C
: Character) is
2800 when '#' |
'$' |
'_' |
'?' |
'@' |
'`' |
'\' |
'^' |
'~' =>
2801 Special_Characters
(C
) := True;
2806 end Set_Special_Character
;
2808 ----------------------
2809 -- Set_Start_Column --
2810 ----------------------
2812 -- Note: it seems at first glance a little expensive to compute this value
2813 -- for every source line (since it is certainly not used for all source
2814 -- lines). On the other hand, it doesn't take much more work to skip past
2815 -- the initial white space on the line counting the columns than it would
2816 -- to scan past the white space using the standard scanning circuits.
2818 function Set_Start_Column
return Column_Number
is
2819 Start_Column
: Column_Number
:= 0;
2822 -- Outer loop scans past horizontal tab characters
2826 -- Inner loop scans past blanks as fast as possible, bumping Scan_Ptr
2827 -- past the blanks and adjusting Start_Column to account for them.
2830 if Source
(Scan_Ptr
) = ' ' then
2831 if Source
(Scan_Ptr
+ 1) = ' ' then
2832 if Source
(Scan_Ptr
+ 2) = ' ' then
2833 if Source
(Scan_Ptr
+ 3) = ' ' then
2834 if Source
(Scan_Ptr
+ 4) = ' ' then
2835 if Source
(Scan_Ptr
+ 5) = ' ' then
2836 if Source
(Scan_Ptr
+ 6) = ' ' then
2837 Scan_Ptr
:= Scan_Ptr
+ 7;
2838 Start_Column
:= Start_Column
+ 7;
2840 Scan_Ptr
:= Scan_Ptr
+ 6;
2841 Start_Column
:= Start_Column
+ 6;
2845 Scan_Ptr
:= Scan_Ptr
+ 5;
2846 Start_Column
:= Start_Column
+ 5;
2850 Scan_Ptr
:= Scan_Ptr
+ 4;
2851 Start_Column
:= Start_Column
+ 4;
2855 Scan_Ptr
:= Scan_Ptr
+ 3;
2856 Start_Column
:= Start_Column
+ 3;
2860 Scan_Ptr
:= Scan_Ptr
+ 2;
2861 Start_Column
:= Start_Column
+ 2;
2865 Scan_Ptr
:= Scan_Ptr
+ 1;
2866 Start_Column
:= Start_Column
+ 1;
2872 end loop Blanks_Loop
;
2874 -- Outer loop keeps going only if a horizontal tab follows
2876 if Source
(Scan_Ptr
) = HT
then
2881 Scan_Ptr
:= Scan_Ptr
+ 1;
2882 Start_Column
:= (Start_Column
/ 8) * 8 + 8;
2888 return Start_Column
;
2890 -- A constraint error can happen only if we have a compiler with checks on
2891 -- and a line with a ludicrous number of tabs or spaces at the start. In
2892 -- such a case, we really don't care if Start_Column is right or not.
2895 when Constraint_Error
=>
2896 return Start_Column
;
2897 end Set_Start_Column
;