1 ------------------------------------------------------------------------------
3 -- GNAT COMPILER COMPONENTS --
9 -- Copyright (C) 1992-2011, 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 Csets; use Csets;
27 with Namet.Sp; use Namet.Sp;
28 with Stylesw; use Stylesw;
29 with Uintp; use Uintp;
31 with GNAT.Spelling_Checker; use GNAT.Spelling_Checker;
40 function Bad_Spelling_Of (T : Token_Type) return Boolean is
41 Tname : constant String := Token_Type'Image (T);
42 -- Characters of token name
44 S : String (1 .. Tname'Last - 4);
45 -- Characters of token name folded to lower case, omitting TOK_ at start
47 M1 : String (1 .. 42) := "incorrect spelling of keyword ************";
48 M2 : String (1 .. 44) := "illegal abbreviation of keyword ************";
49 -- Buffers used to construct error message
53 -- Starting subscripts in M1, M2 for keyword name
55 SL : constant Natural := S'Length;
56 -- Length of expected token name excluding TOK_ at start
59 if Token /= Tok_Identifier then
64 S (J) := Fold_Lower (Tname (J + 4));
67 Get_Name_String (Token_Name);
69 -- A special check for case of PROGRAM used for PROCEDURE
73 and then Name_Buffer (1 .. 7) = "program"
75 Error_Msg_SC -- CODEFIX
76 ("PROCEDURE expected");
80 -- A special check for an illegal abbreviation
82 elsif Name_Len < S'Length
83 and then Name_Len >= 4
84 and then Name_Buffer (1 .. Name_Len) = S (1 .. Name_Len)
86 for J in 1 .. S'Last loop
87 M2 (P2 + J - 1) := Fold_Upper (S (J));
90 Error_Msg_SC (M2 (1 .. P2 - 1 + S'Last));
95 -- Now we go into the full circuit to check for a misspelling
97 -- Never consider something a misspelling if either the actual or
98 -- expected string is less than 3 characters (before this check we
99 -- used to consider i to be a misspelled if in some cases!)
101 if SL < 3 or else Name_Len < 3 then
104 -- Special case: prefix matches, i.e. the leading characters of the
105 -- token that we have exactly match the required keyword. If there
106 -- are at least two characters left over, assume that we have a case
107 -- of two keywords joined together which should not be joined.
109 elsif Name_Len > SL + 1
110 and then S = Name_Buffer (1 .. SL)
112 Scan_Ptr := Token_Ptr + S'Length;
113 Error_Msg_S ("|missing space");
118 if Is_Bad_Spelling_Of (Name_Buffer (1 .. Name_Len), S) then
119 for J in 1 .. S'Last loop
120 M1 (P1 + J - 1) := Fold_Upper (S (J));
123 Error_Msg_SC -- CODFIX
124 (M1 (1 .. P1 - 1 + S'Last));
133 ----------------------
134 -- Check_95_Keyword --
135 ----------------------
137 -- On entry, the caller has checked that current token is an identifier
138 -- whose name matches the name of the 95 keyword New_Tok.
140 procedure Check_95_Keyword (Token_95, Next : Token_Type) is
141 Scan_State : Saved_Scan_State;
144 Save_Scan_State (Scan_State); -- at identifier/keyword
145 Scan; -- past identifier/keyword
148 Restore_Scan_State (Scan_State); -- to identifier
149 Error_Msg_Name_1 := Token_Name;
150 Error_Msg_SC ("(Ada 83) keyword* cannot be used!");
153 Restore_Scan_State (Scan_State); -- to identifier
155 end Check_95_Keyword;
157 ----------------------
158 -- Check_Bad_Layout --
159 ----------------------
161 procedure Check_Bad_Layout is
163 if RM_Column_Check and then Token_Is_At_Start_Of_Line
164 and then Start_Column <= Scope.Table (Scope.Last).Ecol
166 Error_Msg_BC -- CODEFIX
167 ("(style) incorrect layout");
169 end Check_Bad_Layout;
171 --------------------------
172 -- Check_Future_Keyword --
173 --------------------------
175 procedure Check_Future_Keyword is
177 -- Ada 2005 (AI-284): Compiling in Ada 95 mode we warn that INTERFACE,
178 -- OVERRIDING, and SYNCHRONIZED are new reserved words.
180 if Ada_Version = Ada_95
181 and then Warn_On_Ada_2005_Compatibility
183 if Token_Name = Name_Overriding
184 or else Token_Name = Name_Synchronized
185 or else (Token_Name = Name_Interface
186 and then Prev_Token /= Tok_Pragma)
188 Error_Msg_N ("& is a reserved word in Ada 2005?", Token_Node);
192 -- Similarly, warn about Ada 2012 reserved words
194 if Ada_Version in Ada_95 .. Ada_2005
195 and then Warn_On_Ada_2012_Compatibility
197 if Token_Name = Name_Some then
198 Error_Msg_N ("& is a reserved word in Ada 2012?", Token_Node);
202 -- Note: we deliberately do not emit these warnings when operating in
203 -- Ada 83 mode because in that case we assume the user is building
204 -- legacy code anyway and is not interested in updating Ada versions.
206 end Check_Future_Keyword;
208 --------------------------
209 -- Check_Misspelling_Of --
210 --------------------------
212 procedure Check_Misspelling_Of (T : Token_Type) is
214 if Bad_Spelling_Of (T) then
217 end Check_Misspelling_Of;
219 -----------------------------
220 -- Check_Simple_Expression --
221 -----------------------------
223 procedure Check_Simple_Expression (E : Node_Id) is
225 if Expr_Form = EF_Non_Simple then
226 Error_Msg_N ("this expression must be parenthesized", E);
228 end Check_Simple_Expression;
230 ---------------------------------------
231 -- Check_Simple_Expression_In_Ada_83 --
232 ---------------------------------------
234 procedure Check_Simple_Expression_In_Ada_83 (E : Node_Id) is
236 if Expr_Form = EF_Non_Simple then
237 if Ada_Version = Ada_83 then
238 Error_Msg_N ("(Ada 83) this expression must be parenthesized!", E);
241 end Check_Simple_Expression_In_Ada_83;
243 ------------------------
244 -- Check_Subtype_Mark --
245 ------------------------
247 function Check_Subtype_Mark (Mark : Node_Id) return Node_Id is
249 if Nkind (Mark) = N_Identifier
250 or else Nkind (Mark) = N_Selected_Component
251 or else (Nkind (Mark) = N_Attribute_Reference
252 and then Is_Type_Attribute_Name (Attribute_Name (Mark)))
257 Error_Msg ("subtype mark expected", Sloc (Mark));
260 end Check_Subtype_Mark;
266 function Comma_Present return Boolean is
267 Scan_State : Saved_Scan_State;
271 -- First check, if a comma is present, then a comma is present!
273 if Token = Tok_Comma then
277 -- If we have a right paren, then that is taken as ending the list
278 -- i.e. no comma is present.
280 elsif Token = Tok_Right_Paren then
283 -- If pragmas, then get rid of them and make a recursive call
284 -- to process what follows these pragmas.
286 elsif Token = Tok_Pragma then
288 return Comma_Present;
290 -- At this stage we have an error, and the goal is to decide on whether
291 -- or not we should diagnose an error and report a (non-existent)
292 -- comma as being present, or simply to report no comma is present
294 -- If we are a semicolon, then the question is whether we have a missing
295 -- right paren, or whether the semicolon should have been a comma. To
296 -- guess the right answer, we scan ahead keeping track of the paren
297 -- level, looking for a clue that helps us make the right decision.
299 -- This approach is highly accurate in the single error case, and does
300 -- not make bad mistakes in the multiple error case (indeed we can't
301 -- really make a very bad decision at this point in any case).
303 elsif Token = Tok_Semicolon then
304 Save_Scan_State (Scan_State);
305 Scan; -- past semicolon
307 -- Check for being followed by identifier => which almost certainly
308 -- means we are still in a parameter list and the comma should have
309 -- been a semicolon (such a sequence could not follow a semicolon)
311 if Token = Tok_Identifier then
314 if Token = Tok_Arrow then
319 -- If that test didn't work, loop ahead looking for a comma or
320 -- semicolon at the same parenthesis level. Always remember that
321 -- we can't go badly wrong in an error situation like this!
325 -- Here is the look ahead loop, Paren_Count tells us whether the
326 -- token we are looking at is at the same paren level as the
327 -- suspicious semicolon that we are trying to figure out.
331 -- If we hit another semicolon or an end of file, and we have
332 -- not seen a right paren or another comma on the way, then
333 -- probably the semicolon did end the list. Indeed that is
334 -- certainly the only single error correction possible here.
336 if Token = Tok_Semicolon or else Token = Tok_EOF then
337 Restore_Scan_State (Scan_State);
340 -- A comma at the same paren level as the semicolon is a strong
341 -- indicator that the semicolon should have been a comma, indeed
342 -- again this is the only possible single error correction.
344 elsif Token = Tok_Comma then
345 exit when Paren_Count = 0;
347 -- A left paren just bumps the paren count
349 elsif Token = Tok_Left_Paren then
350 Paren_Count := Paren_Count + 1;
352 -- A right paren that is at the same paren level as the semicolon
353 -- also means that the only possible single error correction is
354 -- to assume that the semicolon should have been a comma. If we
355 -- are not at the same paren level, then adjust the paren level.
357 elsif Token = Tok_Right_Paren then
358 exit when Paren_Count = 0;
359 Paren_Count := Paren_Count - 1;
362 -- Keep going, we haven't made a decision yet
367 -- If we fall through the loop, it means that we found a terminating
368 -- right paren or another comma. In either case it is reasonable to
369 -- assume that the semicolon was really intended to be a comma. Also
370 -- come here for the identifier arrow case.
373 Restore_Scan_State (Scan_State);
374 Error_Msg_SC -- CODEFIX
375 ("|"";"" should be "",""");
376 Scan; -- past the semicolon
379 -- If we are not at semicolon or a right paren, then we base the
380 -- decision on whether or not the next token can be part of an
381 -- expression. If not, then decide that no comma is present (the
382 -- caller will eventually generate a missing right parent message)
384 elsif Token in Token_Class_Eterm then
387 -- Otherwise we assume a comma is present, even if none is present,
388 -- since the next token must be part of an expression, so if we were
389 -- at the end of the list, then there is more than one error present.
392 T_Comma; -- to give error
397 -----------------------
398 -- Discard_Junk_List --
399 -----------------------
401 procedure Discard_Junk_List (L : List_Id) is
402 pragma Warnings (Off, L);
405 end Discard_Junk_List;
407 -----------------------
408 -- Discard_Junk_Node --
409 -----------------------
411 procedure Discard_Junk_Node (N : Node_Id) is
412 pragma Warnings (Off, N);
415 end Discard_Junk_Node;
421 procedure Ignore (T : Token_Type) is
424 if T = Tok_Comma then
425 Error_Msg_SC -- CODEFIX
426 ("|extra "","" ignored");
428 elsif T = Tok_Left_Paren then
429 Error_Msg_SC -- CODEFIX
430 ("|extra ""("" ignored");
432 elsif T = Tok_Right_Paren then
433 Error_Msg_SC -- CODEFIX
434 ("|extra "")"" ignored");
436 elsif T = Tok_Semicolon then
437 Error_Msg_SC -- CODEFIX
438 ("|extra "";"" ignored");
440 elsif T = Tok_Colon then
441 Error_Msg_SC -- CODEFIX
442 ("|extra "":"" ignored");
446 Tname : constant String := Token_Type'Image (Token);
448 Error_Msg_SC ("|extra " & Tname (5 .. Tname'Last) & "ignored");
452 Scan; -- Scan past ignored token
456 ----------------------------
457 -- Is_Reserved_Identifier --
458 ----------------------------
460 function Is_Reserved_Identifier (C : Id_Check := None) return Boolean is
462 if not Is_Reserved_Keyword (Token) then
467 Ident_Casing : constant Casing_Type :=
468 Identifier_Casing (Current_Source_File);
469 Key_Casing : constant Casing_Type :=
470 Keyword_Casing (Current_Source_File);
473 -- If the casing of identifiers and keywords is different in
474 -- this source file, and the casing of this token matches the
475 -- keyword casing, then we return False, since it is pretty
476 -- clearly intended to be a keyword.
478 if Ident_Casing = Unknown
479 or else Key_Casing = Unknown
480 or else Ident_Casing = Key_Casing
481 or else Determine_Token_Casing /= Key_Casing
485 -- Here we have a keyword written clearly with keyword casing.
486 -- In default mode, we would not be willing to consider this as
487 -- a reserved identifier, but if C is set, we may still accept it
491 Scan_State : Saved_Scan_State;
492 OK_Next_Tok : Boolean;
495 Save_Scan_State (Scan_State);
498 if Token_Is_At_Start_Of_Line then
506 when C_Comma_Right_Paren =>
508 Token = Tok_Comma or else Token = Tok_Right_Paren;
510 when C_Comma_Colon =>
512 Token = Tok_Comma or else Token = Tok_Colon;
522 when C_Greater_Greater =>
524 Token = Tok_Greater_Greater;
534 when C_Left_Paren_Semicolon =>
536 Token = Tok_Left_Paren or else Token = Tok_Semicolon;
542 when C_Vertical_Bar_Arrow =>
544 Token = Tok_Vertical_Bar or else Token = Tok_Arrow;
547 Restore_Scan_State (Scan_State);
557 -- If we fall through it is not a reserved identifier
560 end Is_Reserved_Identifier;
562 ----------------------
563 -- Merge_Identifier --
564 ----------------------
566 procedure Merge_Identifier (Prev : Node_Id; Nxt : Token_Type) is
568 if Token /= Tok_Identifier then
573 S : Saved_Scan_State;
580 Restore_Scan_State (S);
587 -- Check exactly one space between identifiers
589 if Source (Token_Ptr - 1) /= ' '
590 or else Int (Token_Ptr) /=
591 Int (Prev_Token_Ptr) + Length_Of_Name (Chars (Prev)) + 1
598 Get_Name_String (Chars (Token_Node));
601 Buf : constant String (1 .. Name_Len) :=
602 Name_Buffer (1 .. Name_Len);
605 Get_Name_String (Chars (Prev));
606 Add_Char_To_Name_Buffer ('_');
607 Add_Str_To_Name_Buffer (Buf);
608 Set_Chars (Prev, Name_Find);
611 Error_Msg_Node_1 := Prev;
612 Error_Msg_SC ("unexpected identifier, possibly & was meant here");
614 end Merge_Identifier;
620 function Next_Token_Is (Tok : Token_Type) return Boolean is
621 Scan_State : Saved_Scan_State;
624 Save_Scan_State (Scan_State);
626 Result := (Token = Tok);
627 Restore_Scan_State (Scan_State);
635 procedure No_Constraint is
637 if Token in Token_Class_Consk then
638 Error_Msg_SC ("constraint not allowed here");
639 Discard_Junk_Node (P_Constraint_Opt);
643 ---------------------
644 -- Pop_Scope_Stack --
645 ---------------------
647 procedure Pop_Scope_Stack is
649 pragma Assert (Scope.Last > 0);
650 Scope.Decrement_Last;
653 Error_Msg_Uint_1 := UI_From_Int (Scope.Last);
654 Error_Msg_SC ("decrement scope stack ptr, new value = ^!");
658 ----------------------
659 -- Push_Scope_Stack --
660 ----------------------
662 procedure Push_Scope_Stack is
664 Scope.Increment_Last;
666 if Style_Check_Max_Nesting_Level
667 and then Scope.Last = Style_Max_Nesting_Level + 1
670 ("(style) maximum nesting level exceeded",
671 First_Non_Blank_Location);
674 Scope.Table (Scope.Last).Junk := False;
675 Scope.Table (Scope.Last).Node := Empty;
678 Error_Msg_Uint_1 := UI_From_Int (Scope.Last);
679 Error_Msg_SC ("increment scope stack ptr, new value = ^!");
681 end Push_Scope_Stack;
683 ----------------------
684 -- Separate_Present --
685 ----------------------
687 function Separate_Present return Boolean is
688 Scan_State : Saved_Scan_State;
691 if Token = Tok_Separate then
694 elsif Token /= Tok_Identifier then
698 Save_Scan_State (Scan_State);
699 Scan; -- past identifier
701 if Token = Tok_Semicolon then
702 Restore_Scan_State (Scan_State);
703 return Bad_Spelling_Of (Tok_Separate);
706 Restore_Scan_State (Scan_State);
710 end Separate_Present;
712 --------------------------
713 -- Signal_Bad_Attribute --
714 --------------------------
716 procedure Signal_Bad_Attribute is
718 Error_Msg_N ("unrecognized attribute&", Token_Node);
720 -- Check for possible misspelling
722 Error_Msg_Name_1 := First_Attribute_Name;
723 while Error_Msg_Name_1 <= Last_Attribute_Name loop
724 if Is_Bad_Spelling_Of (Token_Name, Error_Msg_Name_1) then
725 Error_Msg_N -- CODEFIX
726 ("\possible misspelling of %", Token_Node);
730 Error_Msg_Name_1 := Error_Msg_Name_1 + 1;
732 end Signal_Bad_Attribute;
734 -----------------------------
735 -- Token_Is_At_End_Of_Line --
736 -----------------------------
738 function Token_Is_At_End_Of_Line return Boolean is
742 -- Skip past blanks and horizontal tabs
745 while Source (S) = ' ' or else Source (S) = ASCII.HT loop
749 -- We are at end of line if at a control character (CR/LF/VT/FF/EOF)
750 -- or if we are at the start of an end of line comment sequence.
752 return Source (S) < ' '
753 or else (Source (S) = '-' and then Source (S + 1) = '-');
754 end Token_Is_At_End_Of_Line;
756 -------------------------------
757 -- Token_Is_At_Start_Of_Line --
758 -------------------------------
760 function Token_Is_At_Start_Of_Line return Boolean is
762 return (Token_Ptr = First_Non_Blank_Location or else Token = Tok_EOF);
763 end Token_Is_At_Start_Of_Line;