Fix unused warnings.
[official-gcc/graphite-test-results.git] / gcc / ada / par-ch4.adb
blobb679e20334873e62493c750dbd5b4014580d704f
1 ------------------------------------------------------------------------------
2 -- --
3 -- GNAT COMPILER COMPONENTS --
4 -- --
5 -- P A R . C H 4 --
6 -- --
7 -- B o d y --
8 -- --
9 -- Copyright (C) 1992-2010, Free Software Foundation, Inc. --
10 -- --
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. --
20 -- --
21 -- GNAT was originally developed by the GNAT team at New York University. --
22 -- Extensive contributions were provided by Ada Core Technologies Inc. --
23 -- --
24 ------------------------------------------------------------------------------
26 pragma Style_Checks (All_Checks);
27 -- Turn off subprogram body ordering check. Subprograms are in order
28 -- by RM section rather than alphabetical
30 with Stringt; use Stringt;
32 separate (Par)
33 package body Ch4 is
35 -- Attributes that cannot have arguments
37 Is_Parameterless_Attribute : constant Attribute_Class_Array :=
38 (Attribute_Body_Version => True,
39 Attribute_External_Tag => True,
40 Attribute_Img => True,
41 Attribute_Version => True,
42 Attribute_Base => True,
43 Attribute_Class => True,
44 Attribute_Stub_Type => True,
45 Attribute_Type_Key => True,
46 others => False);
47 -- This map contains True for parameterless attributes that return a
48 -- string or a type. For those attributes, a left parenthesis after
49 -- the attribute should not be analyzed as the beginning of a parameters
50 -- list because it may denote a slice operation (X'Img (1 .. 2)) or
51 -- a type conversion (X'Class (Y)).
53 -- Note that this map designates the minimum set of attributes where a
54 -- construct in parentheses that is not an argument can appear right
55 -- after the attribute. For attributes like 'Size, we do not put them
56 -- in the map. If someone writes X'Size (3), that's illegal in any case,
57 -- but we get a better error message by parsing the (3) as an illegal
58 -- argument to the attribute, rather than some meaningless junk that
59 -- follows the attribute.
61 -----------------------
62 -- Local Subprograms --
63 -----------------------
65 function P_Aggregate_Or_Paren_Expr return Node_Id;
66 function P_Allocator return Node_Id;
67 function P_Case_Expression_Alternative return Node_Id;
68 function P_Record_Or_Array_Component_Association return Node_Id;
69 function P_Factor return Node_Id;
70 function P_Primary return Node_Id;
71 function P_Relation return Node_Id;
72 function P_Term return Node_Id;
74 function P_Binary_Adding_Operator return Node_Kind;
75 function P_Logical_Operator return Node_Kind;
76 function P_Multiplying_Operator return Node_Kind;
77 function P_Relational_Operator return Node_Kind;
78 function P_Unary_Adding_Operator return Node_Kind;
80 procedure Bad_Range_Attribute (Loc : Source_Ptr);
81 -- Called to place complaint about bad range attribute at the given
82 -- source location. Terminates by raising Error_Resync.
84 procedure P_Membership_Test (N : Node_Id);
85 -- N is the node for a N_In or N_Not_In node whose right operand has not
86 -- yet been processed. It is called just after scanning out the IN keyword.
87 -- On return, either Right_Opnd or Alternatives is set, as appropriate.
89 function P_Range_Attribute_Reference (Prefix_Node : Node_Id) return Node_Id;
90 -- Scan a range attribute reference. The caller has scanned out the
91 -- prefix. The current token is known to be an apostrophe and the
92 -- following token is known to be RANGE.
94 -------------------------
95 -- Bad_Range_Attribute --
96 -------------------------
98 procedure Bad_Range_Attribute (Loc : Source_Ptr) is
99 begin
100 Error_Msg ("range attribute cannot be used in expression!", Loc);
101 Resync_Expression;
102 end Bad_Range_Attribute;
104 --------------------------
105 -- 4.1 Name (also 6.4) --
106 --------------------------
108 -- NAME ::=
109 -- DIRECT_NAME | EXPLICIT_DEREFERENCE
110 -- | INDEXED_COMPONENT | SLICE
111 -- | SELECTED_COMPONENT | ATTRIBUTE
112 -- | TYPE_CONVERSION | FUNCTION_CALL
113 -- | CHARACTER_LITERAL
115 -- DIRECT_NAME ::= IDENTIFIER | OPERATOR_SYMBOL
117 -- PREFIX ::= NAME | IMPLICIT_DEREFERENCE
119 -- EXPLICIT_DEREFERENCE ::= NAME . all
121 -- IMPLICIT_DEREFERENCE ::= NAME
123 -- INDEXED_COMPONENT ::= PREFIX (EXPRESSION {, EXPRESSION})
125 -- SLICE ::= PREFIX (DISCRETE_RANGE)
127 -- SELECTED_COMPONENT ::= PREFIX . SELECTOR_NAME
129 -- SELECTOR_NAME ::= IDENTIFIER | CHARACTER_LITERAL | OPERATOR_SYMBOL
131 -- ATTRIBUTE_REFERENCE ::= PREFIX ' ATTRIBUTE_DESIGNATOR
133 -- ATTRIBUTE_DESIGNATOR ::=
134 -- IDENTIFIER [(static_EXPRESSION)]
135 -- | access | delta | digits
137 -- FUNCTION_CALL ::=
138 -- function_NAME
139 -- | function_PREFIX ACTUAL_PARAMETER_PART
141 -- ACTUAL_PARAMETER_PART ::=
142 -- (PARAMETER_ASSOCIATION {,PARAMETER_ASSOCIATION})
144 -- PARAMETER_ASSOCIATION ::=
145 -- [formal_parameter_SELECTOR_NAME =>] EXPLICIT_ACTUAL_PARAMETER
147 -- EXPLICIT_ACTUAL_PARAMETER ::= EXPRESSION | variable_NAME
149 -- Note: syntactically a procedure call looks just like a function call,
150 -- so this routine is in practice used to scan out procedure calls as well.
152 -- On return, Expr_Form is set to either EF_Name or EF_Simple_Name
154 -- Error recovery: can raise Error_Resync
156 -- Note: if on return Token = Tok_Apostrophe, then the apostrophe must be
157 -- followed by either a left paren (qualified expression case), or by
158 -- range (range attribute case). All other uses of apostrophe (i.e. all
159 -- other attributes) are handled in this routine.
161 -- Error recovery: can raise Error_Resync
163 function P_Name return Node_Id is
164 Scan_State : Saved_Scan_State;
165 Name_Node : Node_Id;
166 Prefix_Node : Node_Id;
167 Ident_Node : Node_Id;
168 Expr_Node : Node_Id;
169 Range_Node : Node_Id;
170 Arg_Node : Node_Id;
172 Arg_List : List_Id := No_List; -- kill junk warning
173 Attr_Name : Name_Id := No_Name; -- kill junk warning
175 begin
176 -- Case of not a name
178 if Token not in Token_Class_Name then
180 -- If it looks like start of expression, complain and scan expression
182 if Token in Token_Class_Literal
183 or else Token = Tok_Left_Paren
184 then
185 Error_Msg_SC ("name expected");
186 return P_Expression;
188 -- Otherwise some other junk, not much we can do
190 else
191 Error_Msg_AP ("name expected");
192 raise Error_Resync;
193 end if;
194 end if;
196 -- Loop through designators in qualified name
198 Name_Node := Token_Node;
200 loop
201 Scan; -- past designator
202 exit when Token /= Tok_Dot;
203 Save_Scan_State (Scan_State); -- at dot
204 Scan; -- past dot
206 -- If we do not have another designator after the dot, then join
207 -- the normal circuit to handle a dot extension (may be .all or
208 -- character literal case). Otherwise loop back to scan the next
209 -- designator.
211 if Token not in Token_Class_Desig then
212 goto Scan_Name_Extension_Dot;
213 else
214 Prefix_Node := Name_Node;
215 Name_Node := New_Node (N_Selected_Component, Prev_Token_Ptr);
216 Set_Prefix (Name_Node, Prefix_Node);
217 Set_Selector_Name (Name_Node, Token_Node);
218 end if;
219 end loop;
221 -- We have now scanned out a qualified designator. If the last token is
222 -- an operator symbol, then we certainly do not have the Snam case, so
223 -- we can just use the normal name extension check circuit
225 if Prev_Token = Tok_Operator_Symbol then
226 goto Scan_Name_Extension;
227 end if;
229 -- We have scanned out a qualified simple name, check for name extension
230 -- Note that we know there is no dot here at this stage, so the only
231 -- possible cases of name extension are apostrophe and left paren.
233 if Token = Tok_Apostrophe then
234 Save_Scan_State (Scan_State); -- at apostrophe
235 Scan; -- past apostrophe
237 -- Qualified expression in Ada 2012 mode (treated as a name)
239 if Ada_Version >= Ada_2012 and then Token = Tok_Left_Paren then
240 goto Scan_Name_Extension_Apostrophe;
242 -- If left paren not in Ada 2012, then it is not part of the name,
243 -- since qualified expressions are not names in prior versions of
244 -- Ada, so return with Token backed up to point to the apostrophe.
245 -- The treatment for the range attribute is similar (we do not
246 -- consider x'range to be a name in this grammar).
248 elsif Token = Tok_Left_Paren or else Token = Tok_Range then
249 Restore_Scan_State (Scan_State); -- to apostrophe
250 Expr_Form := EF_Simple_Name;
251 return Name_Node;
253 -- Otherwise we have the case of a name extended by an attribute
255 else
256 goto Scan_Name_Extension_Apostrophe;
257 end if;
259 -- Check case of qualified simple name extended by a left parenthesis
261 elsif Token = Tok_Left_Paren then
262 Scan; -- past left paren
263 goto Scan_Name_Extension_Left_Paren;
265 -- Otherwise the qualified simple name is not extended, so return
267 else
268 Expr_Form := EF_Simple_Name;
269 return Name_Node;
270 end if;
272 -- Loop scanning past name extensions. A label is used for control
273 -- transfer for this loop for ease of interfacing with the finite state
274 -- machine in the parenthesis scanning circuit, and also to allow for
275 -- passing in control to the appropriate point from the above code.
277 <<Scan_Name_Extension>>
279 -- Character literal used as name cannot be extended. Also this
280 -- cannot be a call, since the name for a call must be a designator.
281 -- Return in these cases, or if there is no name extension
283 if Token not in Token_Class_Namext
284 or else Prev_Token = Tok_Char_Literal
285 then
286 Expr_Form := EF_Name;
287 return Name_Node;
288 end if;
290 -- Merge here when we know there is a name extension
292 <<Scan_Name_Extension_OK>>
294 if Token = Tok_Left_Paren then
295 Scan; -- past left paren
296 goto Scan_Name_Extension_Left_Paren;
298 elsif Token = Tok_Apostrophe then
299 Save_Scan_State (Scan_State); -- at apostrophe
300 Scan; -- past apostrophe
301 goto Scan_Name_Extension_Apostrophe;
303 else -- Token = Tok_Dot
304 Save_Scan_State (Scan_State); -- at dot
305 Scan; -- past dot
306 goto Scan_Name_Extension_Dot;
307 end if;
309 -- Case of name extended by dot (selection), dot is already skipped
310 -- and the scan state at the point of the dot is saved in Scan_State.
312 <<Scan_Name_Extension_Dot>>
314 -- Explicit dereference case
316 if Token = Tok_All then
317 Prefix_Node := Name_Node;
318 Name_Node := New_Node (N_Explicit_Dereference, Token_Ptr);
319 Set_Prefix (Name_Node, Prefix_Node);
320 Scan; -- past ALL
321 goto Scan_Name_Extension;
323 -- Selected component case
325 elsif Token in Token_Class_Name then
326 Prefix_Node := Name_Node;
327 Name_Node := New_Node (N_Selected_Component, Prev_Token_Ptr);
328 Set_Prefix (Name_Node, Prefix_Node);
329 Set_Selector_Name (Name_Node, Token_Node);
330 Scan; -- past selector
331 goto Scan_Name_Extension;
333 -- Reserved identifier as selector
335 elsif Is_Reserved_Identifier then
336 Scan_Reserved_Identifier (Force_Msg => False);
337 Prefix_Node := Name_Node;
338 Name_Node := New_Node (N_Selected_Component, Prev_Token_Ptr);
339 Set_Prefix (Name_Node, Prefix_Node);
340 Set_Selector_Name (Name_Node, Token_Node);
341 Scan; -- past identifier used as selector
342 goto Scan_Name_Extension;
344 -- If dot is at end of line and followed by nothing legal,
345 -- then assume end of name and quit (dot will be taken as
346 -- an erroneous form of some other punctuation by our caller).
348 elsif Token_Is_At_Start_Of_Line then
349 Restore_Scan_State (Scan_State);
350 return Name_Node;
352 -- Here if nothing legal after the dot
354 else
355 Error_Msg_AP ("selector expected");
356 raise Error_Resync;
357 end if;
359 -- Here for an apostrophe as name extension. The scan position at the
360 -- apostrophe has already been saved, and the apostrophe scanned out.
362 <<Scan_Name_Extension_Apostrophe>>
364 Scan_Apostrophe : declare
365 function Apostrophe_Should_Be_Semicolon return Boolean;
366 -- Checks for case where apostrophe should probably be
367 -- a semicolon, and if so, gives appropriate message,
368 -- resets the scan pointer to the apostrophe, changes
369 -- the current token to Tok_Semicolon, and returns True.
370 -- Otherwise returns False.
372 ------------------------------------
373 -- Apostrophe_Should_Be_Semicolon --
374 ------------------------------------
376 function Apostrophe_Should_Be_Semicolon return Boolean is
377 begin
378 if Token_Is_At_Start_Of_Line then
379 Restore_Scan_State (Scan_State); -- to apostrophe
380 Error_Msg_SC ("|""''"" should be "";""");
381 Token := Tok_Semicolon;
382 return True;
383 else
384 return False;
385 end if;
386 end Apostrophe_Should_Be_Semicolon;
388 -- Start of processing for Scan_Apostrophe
390 begin
391 -- Check for qualified expression case in Ada 2012 mode
393 if Ada_Version >= Ada_2012 and then Token = Tok_Left_Paren then
394 Name_Node := P_Qualified_Expression (Name_Node);
395 goto Scan_Name_Extension;
397 -- If range attribute after apostrophe, then return with Token
398 -- pointing to the apostrophe. Note that in this case the prefix
399 -- need not be a simple name (cases like A.all'range). Similarly
400 -- if there is a left paren after the apostrophe, then we also
401 -- return with Token pointing to the apostrophe (this is the
402 -- aggregate case, or some error case).
404 elsif Token = Tok_Range or else Token = Tok_Left_Paren then
405 Restore_Scan_State (Scan_State); -- to apostrophe
406 Expr_Form := EF_Name;
407 return Name_Node;
409 -- Here for cases where attribute designator is an identifier
411 elsif Token = Tok_Identifier then
412 Attr_Name := Token_Name;
414 if not Is_Attribute_Name (Attr_Name) then
415 if Apostrophe_Should_Be_Semicolon then
416 Expr_Form := EF_Name;
417 return Name_Node;
419 -- Here for a bad attribute name
421 else
422 Signal_Bad_Attribute;
423 Scan; -- past bad identifier
425 if Token = Tok_Left_Paren then
426 Scan; -- past left paren
428 loop
429 Discard_Junk_Node (P_Expression_If_OK);
430 exit when not Comma_Present;
431 end loop;
433 T_Right_Paren;
434 end if;
436 return Error;
437 end if;
438 end if;
440 if Style_Check then
441 Style.Check_Attribute_Name (False);
442 end if;
444 -- Here for case of attribute designator is not an identifier
446 else
447 if Token = Tok_Delta then
448 Attr_Name := Name_Delta;
450 elsif Token = Tok_Digits then
451 Attr_Name := Name_Digits;
453 elsif Token = Tok_Access then
454 Attr_Name := Name_Access;
456 elsif Token = Tok_Mod and then Ada_Version >= Ada_95 then
457 Attr_Name := Name_Mod;
459 elsif Apostrophe_Should_Be_Semicolon then
460 Expr_Form := EF_Name;
461 return Name_Node;
463 else
464 Error_Msg_AP ("attribute designator expected");
465 raise Error_Resync;
466 end if;
468 if Style_Check then
469 Style.Check_Attribute_Name (True);
470 end if;
471 end if;
473 -- We come here with an OK attribute scanned, and the
474 -- corresponding Attribute identifier node stored in Ident_Node.
476 Prefix_Node := Name_Node;
477 Name_Node := New_Node (N_Attribute_Reference, Prev_Token_Ptr);
478 Scan; -- past attribute designator
479 Set_Prefix (Name_Node, Prefix_Node);
480 Set_Attribute_Name (Name_Node, Attr_Name);
482 -- Scan attribute arguments/designator. We skip this if we know
483 -- that the attribute cannot have an argument.
485 if Token = Tok_Left_Paren
486 and then not
487 Is_Parameterless_Attribute (Get_Attribute_Id (Attr_Name))
488 then
489 Set_Expressions (Name_Node, New_List);
490 Scan; -- past left paren
492 loop
493 declare
494 Expr : constant Node_Id := P_Expression_If_OK;
496 begin
497 if Token = Tok_Arrow then
498 Error_Msg_SC
499 ("named parameters not permitted for attributes");
500 Scan; -- past junk arrow
502 else
503 Append (Expr, Expressions (Name_Node));
504 exit when not Comma_Present;
505 end if;
506 end;
507 end loop;
509 T_Right_Paren;
510 end if;
512 goto Scan_Name_Extension;
513 end Scan_Apostrophe;
515 -- Here for left parenthesis extending name (left paren skipped)
517 <<Scan_Name_Extension_Left_Paren>>
519 -- We now have to scan through a list of items, terminated by a
520 -- right parenthesis. The scan is handled by a finite state
521 -- machine. The possibilities are:
523 -- (discrete_range)
525 -- This is a slice. This case is handled in LP_State_Init
527 -- (expression, expression, ..)
529 -- This is interpreted as an indexed component, i.e. as a
530 -- case of a name which can be extended in the normal manner.
531 -- This case is handled by LP_State_Name or LP_State_Expr.
533 -- Note: conditional expressions (without an extra level of
534 -- parentheses) are permitted in this context).
536 -- (..., identifier => expression , ...)
538 -- If there is at least one occurrence of identifier => (but
539 -- none of the other cases apply), then we have a call.
541 -- Test for Id => case
543 if Token = Tok_Identifier then
544 Save_Scan_State (Scan_State); -- at Id
545 Scan; -- past Id
547 -- Test for => (allow := as an error substitute)
549 if Token = Tok_Arrow or else Token = Tok_Colon_Equal then
550 Restore_Scan_State (Scan_State); -- to Id
551 Arg_List := New_List;
552 goto LP_State_Call;
554 else
555 Restore_Scan_State (Scan_State); -- to Id
556 end if;
557 end if;
559 -- Here we have an expression after all
561 Expr_Node := P_Expression_Or_Range_Attribute_If_OK;
563 -- Check cases of discrete range for a slice
565 -- First possibility: Range_Attribute_Reference
567 if Expr_Form = EF_Range_Attr then
568 Range_Node := Expr_Node;
570 -- Second possibility: Simple_expression .. Simple_expression
572 elsif Token = Tok_Dot_Dot then
573 Check_Simple_Expression (Expr_Node);
574 Range_Node := New_Node (N_Range, Token_Ptr);
575 Set_Low_Bound (Range_Node, Expr_Node);
576 Scan; -- past ..
577 Expr_Node := P_Expression;
578 Check_Simple_Expression (Expr_Node);
579 Set_High_Bound (Range_Node, Expr_Node);
581 -- Third possibility: Type_name range Range
583 elsif Token = Tok_Range then
584 if Expr_Form /= EF_Simple_Name then
585 Error_Msg_SC ("subtype mark must precede RANGE");
586 raise Error_Resync;
587 end if;
589 Range_Node := P_Subtype_Indication (Expr_Node);
591 -- Otherwise we just have an expression. It is true that we might
592 -- have a subtype mark without a range constraint but this case
593 -- is syntactically indistinguishable from the expression case.
595 else
596 Arg_List := New_List;
597 goto LP_State_Expr;
598 end if;
600 -- Fall through here with unmistakable Discrete range scanned,
601 -- which means that we definitely have the case of a slice. The
602 -- Discrete range is in Range_Node.
604 if Token = Tok_Comma then
605 Error_Msg_SC ("slice cannot have more than one dimension");
606 raise Error_Resync;
608 elsif Token /= Tok_Right_Paren then
609 T_Right_Paren;
610 raise Error_Resync;
612 else
613 Scan; -- past right paren
614 Prefix_Node := Name_Node;
615 Name_Node := New_Node (N_Slice, Sloc (Prefix_Node));
616 Set_Prefix (Name_Node, Prefix_Node);
617 Set_Discrete_Range (Name_Node, Range_Node);
619 -- An operator node is legal as a prefix to other names,
620 -- but not for a slice.
622 if Nkind (Prefix_Node) = N_Operator_Symbol then
623 Error_Msg_N ("illegal prefix for slice", Prefix_Node);
624 end if;
626 -- If we have a name extension, go scan it
628 if Token in Token_Class_Namext then
629 goto Scan_Name_Extension_OK;
631 -- Otherwise return (a slice is a name, but is not a call)
633 else
634 Expr_Form := EF_Name;
635 return Name_Node;
636 end if;
637 end if;
639 -- In LP_State_Expr, we have scanned one or more expressions, and
640 -- so we have a call or an indexed component which is a name. On
641 -- entry we have the expression just scanned in Expr_Node and
642 -- Arg_List contains the list of expressions encountered so far
644 <<LP_State_Expr>>
645 Append (Expr_Node, Arg_List);
647 if Token = Tok_Arrow then
648 Error_Msg
649 ("expect identifier in parameter association",
650 Sloc (Expr_Node));
651 Scan; -- past arrow
653 elsif not Comma_Present then
654 T_Right_Paren;
655 Prefix_Node := Name_Node;
656 Name_Node := New_Node (N_Indexed_Component, Sloc (Prefix_Node));
657 Set_Prefix (Name_Node, Prefix_Node);
658 Set_Expressions (Name_Node, Arg_List);
659 goto Scan_Name_Extension;
660 end if;
662 -- Comma present (and scanned out), test for identifier => case
663 -- Test for identifier => case
665 if Token = Tok_Identifier then
666 Save_Scan_State (Scan_State); -- at Id
667 Scan; -- past Id
669 -- Test for => (allow := as error substitute)
671 if Token = Tok_Arrow or else Token = Tok_Colon_Equal then
672 Restore_Scan_State (Scan_State); -- to Id
673 goto LP_State_Call;
675 -- Otherwise it's just an expression after all, so backup
677 else
678 Restore_Scan_State (Scan_State); -- to Id
679 end if;
680 end if;
682 -- Here we have an expression after all, so stay in this state
684 Expr_Node := P_Expression_If_OK;
685 goto LP_State_Expr;
687 -- LP_State_Call corresponds to the situation in which at least
688 -- one instance of Id => Expression has been encountered, so we
689 -- know that we do not have a name, but rather a call. We enter
690 -- it with the scan pointer pointing to the next argument to scan,
691 -- and Arg_List containing the list of arguments scanned so far.
693 <<LP_State_Call>>
695 -- Test for case of Id => Expression (named parameter)
697 if Token = Tok_Identifier then
698 Save_Scan_State (Scan_State); -- at Id
699 Ident_Node := Token_Node;
700 Scan; -- past Id
702 -- Deal with => (allow := as erroneous substitute)
704 if Token = Tok_Arrow or else Token = Tok_Colon_Equal then
705 Arg_Node := New_Node (N_Parameter_Association, Prev_Token_Ptr);
706 Set_Selector_Name (Arg_Node, Ident_Node);
707 T_Arrow;
708 Set_Explicit_Actual_Parameter (Arg_Node, P_Expression);
709 Append (Arg_Node, Arg_List);
711 -- If a comma follows, go back and scan next entry
713 if Comma_Present then
714 goto LP_State_Call;
716 -- Otherwise we have the end of a call
718 else
719 Prefix_Node := Name_Node;
720 Name_Node := New_Node (N_Function_Call, Sloc (Prefix_Node));
721 Set_Name (Name_Node, Prefix_Node);
722 Set_Parameter_Associations (Name_Node, Arg_List);
723 T_Right_Paren;
725 if Token in Token_Class_Namext then
726 goto Scan_Name_Extension_OK;
728 -- This is a case of a call which cannot be a name
730 else
731 Expr_Form := EF_Name;
732 return Name_Node;
733 end if;
734 end if;
736 -- Not named parameter: Id started an expression after all
738 else
739 Restore_Scan_State (Scan_State); -- to Id
740 end if;
741 end if;
743 -- Here if entry did not start with Id => which means that it
744 -- is a positional parameter, which is not allowed, since we
745 -- have seen at least one named parameter already.
747 Error_Msg_SC
748 ("positional parameter association " &
749 "not allowed after named one");
751 Expr_Node := P_Expression_If_OK;
753 -- Leaving the '>' in an association is not unusual, so suggest
754 -- a possible fix.
756 if Nkind (Expr_Node) = N_Op_Eq then
757 Error_Msg_N ("\maybe `='>` was intended", Expr_Node);
758 end if;
760 -- We go back to scanning out expressions, so that we do not get
761 -- multiple error messages when several positional parameters
762 -- follow a named parameter.
764 goto LP_State_Expr;
766 -- End of treatment for name extensions starting with left paren
768 -- End of loop through name extensions
770 end P_Name;
772 -- This function parses a restricted form of Names which are either
773 -- designators, or designators preceded by a sequence of prefixes
774 -- that are direct names.
776 -- Error recovery: cannot raise Error_Resync
778 function P_Function_Name return Node_Id is
779 Designator_Node : Node_Id;
780 Prefix_Node : Node_Id;
781 Selector_Node : Node_Id;
782 Dot_Sloc : Source_Ptr := No_Location;
784 begin
785 -- Prefix_Node is set to the gathered prefix so far, Empty means that
786 -- no prefix has been scanned. This allows us to build up the result
787 -- in the required right recursive manner.
789 Prefix_Node := Empty;
791 -- Loop through prefixes
793 loop
794 Designator_Node := Token_Node;
796 if Token not in Token_Class_Desig then
797 return P_Identifier; -- let P_Identifier issue the error message
799 else -- Token in Token_Class_Desig
800 Scan; -- past designator
801 exit when Token /= Tok_Dot;
802 end if;
804 -- Here at a dot, with token just before it in Designator_Node
806 if No (Prefix_Node) then
807 Prefix_Node := Designator_Node;
808 else
809 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
810 Set_Prefix (Selector_Node, Prefix_Node);
811 Set_Selector_Name (Selector_Node, Designator_Node);
812 Prefix_Node := Selector_Node;
813 end if;
815 Dot_Sloc := Token_Ptr;
816 Scan; -- past dot
817 end loop;
819 -- Fall out of the loop having just scanned a designator
821 if No (Prefix_Node) then
822 return Designator_Node;
823 else
824 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
825 Set_Prefix (Selector_Node, Prefix_Node);
826 Set_Selector_Name (Selector_Node, Designator_Node);
827 return Selector_Node;
828 end if;
830 exception
831 when Error_Resync =>
832 return Error;
833 end P_Function_Name;
835 -- This function parses a restricted form of Names which are either
836 -- identifiers, or identifiers preceded by a sequence of prefixes
837 -- that are direct names.
839 -- Error recovery: cannot raise Error_Resync
841 function P_Qualified_Simple_Name return Node_Id is
842 Designator_Node : Node_Id;
843 Prefix_Node : Node_Id;
844 Selector_Node : Node_Id;
845 Dot_Sloc : Source_Ptr := No_Location;
847 begin
848 -- Prefix node is set to the gathered prefix so far, Empty means that
849 -- no prefix has been scanned. This allows us to build up the result
850 -- in the required right recursive manner.
852 Prefix_Node := Empty;
854 -- Loop through prefixes
856 loop
857 Designator_Node := Token_Node;
859 if Token = Tok_Identifier then
860 Scan; -- past identifier
861 exit when Token /= Tok_Dot;
863 elsif Token not in Token_Class_Desig then
864 return P_Identifier; -- let P_Identifier issue the error message
866 else
867 Scan; -- past designator
869 if Token /= Tok_Dot then
870 Error_Msg_SP ("identifier expected");
871 return Error;
872 end if;
873 end if;
875 -- Here at a dot, with token just before it in Designator_Node
877 if No (Prefix_Node) then
878 Prefix_Node := Designator_Node;
879 else
880 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
881 Set_Prefix (Selector_Node, Prefix_Node);
882 Set_Selector_Name (Selector_Node, Designator_Node);
883 Prefix_Node := Selector_Node;
884 end if;
886 Dot_Sloc := Token_Ptr;
887 Scan; -- past dot
888 end loop;
890 -- Fall out of the loop having just scanned an identifier
892 if No (Prefix_Node) then
893 return Designator_Node;
894 else
895 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
896 Set_Prefix (Selector_Node, Prefix_Node);
897 Set_Selector_Name (Selector_Node, Designator_Node);
898 return Selector_Node;
899 end if;
901 exception
902 when Error_Resync =>
903 return Error;
904 end P_Qualified_Simple_Name;
906 -- This procedure differs from P_Qualified_Simple_Name only in that it
907 -- raises Error_Resync if any error is encountered. It only returns after
908 -- scanning a valid qualified simple name.
910 -- Error recovery: can raise Error_Resync
912 function P_Qualified_Simple_Name_Resync return Node_Id is
913 Designator_Node : Node_Id;
914 Prefix_Node : Node_Id;
915 Selector_Node : Node_Id;
916 Dot_Sloc : Source_Ptr := No_Location;
918 begin
919 Prefix_Node := Empty;
921 -- Loop through prefixes
923 loop
924 Designator_Node := Token_Node;
926 if Token = Tok_Identifier then
927 Scan; -- past identifier
928 exit when Token /= Tok_Dot;
930 elsif Token not in Token_Class_Desig then
931 Discard_Junk_Node (P_Identifier); -- to issue the error message
932 raise Error_Resync;
934 else
935 Scan; -- past designator
937 if Token /= Tok_Dot then
938 Error_Msg_SP ("identifier expected");
939 raise Error_Resync;
940 end if;
941 end if;
943 -- Here at a dot, with token just before it in Designator_Node
945 if No (Prefix_Node) then
946 Prefix_Node := Designator_Node;
947 else
948 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
949 Set_Prefix (Selector_Node, Prefix_Node);
950 Set_Selector_Name (Selector_Node, Designator_Node);
951 Prefix_Node := Selector_Node;
952 end if;
954 Dot_Sloc := Token_Ptr;
955 Scan; -- past period
956 end loop;
958 -- Fall out of the loop having just scanned an identifier
960 if No (Prefix_Node) then
961 return Designator_Node;
962 else
963 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
964 Set_Prefix (Selector_Node, Prefix_Node);
965 Set_Selector_Name (Selector_Node, Designator_Node);
966 return Selector_Node;
967 end if;
968 end P_Qualified_Simple_Name_Resync;
970 ----------------------
971 -- 4.1 Direct_Name --
972 ----------------------
974 -- Parsed by P_Name and other functions in section 4.1
976 -----------------
977 -- 4.1 Prefix --
978 -----------------
980 -- Parsed by P_Name (4.1)
982 -------------------------------
983 -- 4.1 Explicit Dereference --
984 -------------------------------
986 -- Parsed by P_Name (4.1)
988 -------------------------------
989 -- 4.1 Implicit_Dereference --
990 -------------------------------
992 -- Parsed by P_Name (4.1)
994 ----------------------------
995 -- 4.1 Indexed Component --
996 ----------------------------
998 -- Parsed by P_Name (4.1)
1000 ----------------
1001 -- 4.1 Slice --
1002 ----------------
1004 -- Parsed by P_Name (4.1)
1006 -----------------------------
1007 -- 4.1 Selected_Component --
1008 -----------------------------
1010 -- Parsed by P_Name (4.1)
1012 ------------------------
1013 -- 4.1 Selector Name --
1014 ------------------------
1016 -- Parsed by P_Name (4.1)
1018 ------------------------------
1019 -- 4.1 Attribute Reference --
1020 ------------------------------
1022 -- Parsed by P_Name (4.1)
1024 -------------------------------
1025 -- 4.1 Attribute Designator --
1026 -------------------------------
1028 -- Parsed by P_Name (4.1)
1030 --------------------------------------
1031 -- 4.1.4 Range Attribute Reference --
1032 --------------------------------------
1034 -- RANGE_ATTRIBUTE_REFERENCE ::= PREFIX ' RANGE_ATTRIBUTE_DESIGNATOR
1036 -- RANGE_ATTRIBUTE_DESIGNATOR ::= range [(static_EXPRESSION)]
1038 -- In the grammar, a RANGE attribute is simply a name, but its use is
1039 -- highly restricted, so in the parser, we do not regard it as a name.
1040 -- Instead, P_Name returns without scanning the 'RANGE part of the
1041 -- attribute, and the caller uses the following function to construct
1042 -- a range attribute in places where it is appropriate.
1044 -- Note that RANGE here is treated essentially as an identifier,
1045 -- rather than a reserved word.
1047 -- The caller has parsed the prefix, i.e. a name, and Token points to
1048 -- the apostrophe. The token after the apostrophe is known to be RANGE
1049 -- at this point. The prefix node becomes the prefix of the attribute.
1051 -- Error_Recovery: Cannot raise Error_Resync
1053 function P_Range_Attribute_Reference
1054 (Prefix_Node : Node_Id)
1055 return Node_Id
1057 Attr_Node : Node_Id;
1059 begin
1060 Attr_Node := New_Node (N_Attribute_Reference, Token_Ptr);
1061 Set_Prefix (Attr_Node, Prefix_Node);
1062 Scan; -- past apostrophe
1064 if Style_Check then
1065 Style.Check_Attribute_Name (True);
1066 end if;
1068 Set_Attribute_Name (Attr_Node, Name_Range);
1069 Scan; -- past RANGE
1071 if Token = Tok_Left_Paren then
1072 Scan; -- past left paren
1073 Set_Expressions (Attr_Node, New_List (P_Expression_If_OK));
1074 T_Right_Paren;
1075 end if;
1077 return Attr_Node;
1078 end P_Range_Attribute_Reference;
1080 ---------------------------------------
1081 -- 4.1.4 Range Attribute Designator --
1082 ---------------------------------------
1084 -- Parsed by P_Range_Attribute_Reference (4.4)
1086 --------------------
1087 -- 4.3 Aggregate --
1088 --------------------
1090 -- AGGREGATE ::= RECORD_AGGREGATE | EXTENSION_AGGREGATE | ARRAY_AGGREGATE
1092 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3), except in the case where
1093 -- an aggregate is known to be required (code statement, extension
1094 -- aggregate), in which cases this routine performs the necessary check
1095 -- that we have an aggregate rather than a parenthesized expression
1097 -- Error recovery: can raise Error_Resync
1099 function P_Aggregate return Node_Id is
1100 Aggr_Sloc : constant Source_Ptr := Token_Ptr;
1101 Aggr_Node : constant Node_Id := P_Aggregate_Or_Paren_Expr;
1103 begin
1104 if Nkind (Aggr_Node) /= N_Aggregate
1105 and then
1106 Nkind (Aggr_Node) /= N_Extension_Aggregate
1107 then
1108 Error_Msg
1109 ("aggregate may not have single positional component", Aggr_Sloc);
1110 return Error;
1111 else
1112 return Aggr_Node;
1113 end if;
1114 end P_Aggregate;
1116 ------------------------------------------------
1117 -- 4.3 Aggregate or Parenthesized Expression --
1118 ------------------------------------------------
1120 -- This procedure parses out either an aggregate or a parenthesized
1121 -- expression (these two constructs are closely related, since a
1122 -- parenthesized expression looks like an aggregate with a single
1123 -- positional component).
1125 -- AGGREGATE ::=
1126 -- RECORD_AGGREGATE | EXTENSION_AGGREGATE | ARRAY_AGGREGATE
1128 -- RECORD_AGGREGATE ::= (RECORD_COMPONENT_ASSOCIATION_LIST)
1130 -- RECORD_COMPONENT_ASSOCIATION_LIST ::=
1131 -- RECORD_COMPONENT_ASSOCIATION {, RECORD_COMPONENT_ASSOCIATION}
1132 -- | null record
1134 -- RECORD_COMPONENT_ASSOCIATION ::=
1135 -- [COMPONENT_CHOICE_LIST =>] EXPRESSION
1137 -- COMPONENT_CHOICE_LIST ::=
1138 -- component_SELECTOR_NAME {| component_SELECTOR_NAME}
1139 -- | others
1141 -- EXTENSION_AGGREGATE ::=
1142 -- (ANCESTOR_PART with RECORD_COMPONENT_ASSOCIATION_LIST)
1144 -- ANCESTOR_PART ::= EXPRESSION | SUBTYPE_MARK
1146 -- ARRAY_AGGREGATE ::=
1147 -- POSITIONAL_ARRAY_AGGREGATE | NAMED_ARRAY_AGGREGATE
1149 -- POSITIONAL_ARRAY_AGGREGATE ::=
1150 -- (EXPRESSION, EXPRESSION {, EXPRESSION})
1151 -- | (EXPRESSION {, EXPRESSION}, others => EXPRESSION)
1152 -- | (EXPRESSION {, EXPRESSION}, others => <>)
1154 -- NAMED_ARRAY_AGGREGATE ::=
1155 -- (ARRAY_COMPONENT_ASSOCIATION {, ARRAY_COMPONENT_ASSOCIATION})
1157 -- PRIMARY ::= (EXPRESSION);
1159 -- Error recovery: can raise Error_Resync
1161 -- Note: POSITIONAL_ARRAY_AGGREGATE rule has been extended to give support
1162 -- to Ada 2005 limited aggregates (AI-287)
1164 function P_Aggregate_Or_Paren_Expr return Node_Id is
1165 Aggregate_Node : Node_Id;
1166 Expr_List : List_Id;
1167 Assoc_List : List_Id;
1168 Expr_Node : Node_Id;
1169 Lparen_Sloc : Source_Ptr;
1170 Scan_State : Saved_Scan_State;
1172 procedure Box_Error;
1173 -- Called if <> is encountered as positional aggregate element. Issues
1174 -- error message and sets Expr_Node to Error.
1176 ---------------
1177 -- Box_Error --
1178 ---------------
1180 procedure Box_Error is
1181 begin
1182 if Ada_Version < Ada_2005 then
1183 Error_Msg_SC ("box in aggregate is an Ada 2005 extension");
1184 end if;
1186 -- Ada 2005 (AI-287): The box notation is allowed only with named
1187 -- notation because positional notation might be error prone. For
1188 -- example, in "(X, <>, Y, <>)", there is no type associated with
1189 -- the boxes, so you might not be leaving out the components you
1190 -- thought you were leaving out.
1192 Error_Msg_SC ("(Ada 2005) box only allowed with named notation");
1193 Scan; -- past box
1194 Expr_Node := Error;
1195 end Box_Error;
1197 -- Start of processsing for P_Aggregate_Or_Paren_Expr
1199 begin
1200 Lparen_Sloc := Token_Ptr;
1201 T_Left_Paren;
1203 -- Conditional expression case
1205 if Token = Tok_If then
1206 Expr_Node := P_Conditional_Expression;
1207 T_Right_Paren;
1208 return Expr_Node;
1210 -- Case expression case
1212 elsif Token = Tok_Case then
1213 Expr_Node := P_Case_Expression;
1214 T_Right_Paren;
1215 return Expr_Node;
1217 -- Quantified expression case
1219 elsif Token = Tok_For then
1220 Expr_Node := P_Quantified_Expression;
1221 T_Right_Paren;
1222 return Expr_Node;
1224 -- Note: the mechanism used here of rescanning the initial expression
1225 -- is distinctly unpleasant, but it saves a lot of fiddling in scanning
1226 -- out the discrete choice list.
1228 -- Deal with expression and extension aggregate cases first
1230 elsif Token /= Tok_Others then
1231 Save_Scan_State (Scan_State); -- at start of expression
1233 -- Deal with (NULL RECORD) case
1235 if Token = Tok_Null then
1236 Scan; -- past NULL
1238 if Token = Tok_Record then
1239 Aggregate_Node := New_Node (N_Aggregate, Lparen_Sloc);
1240 Set_Null_Record_Present (Aggregate_Node, True);
1241 Scan; -- past RECORD
1242 T_Right_Paren;
1243 return Aggregate_Node;
1244 else
1245 Restore_Scan_State (Scan_State); -- to NULL that must be expr
1246 end if;
1247 end if;
1249 -- Scan expression, handling box appearing as positional argument
1251 if Token = Tok_Box then
1252 Box_Error;
1253 else
1254 Expr_Node := P_Expression_Or_Range_Attribute_If_OK;
1255 end if;
1257 -- Extension aggregate case
1259 if Token = Tok_With then
1260 if Nkind (Expr_Node) = N_Attribute_Reference
1261 and then Attribute_Name (Expr_Node) = Name_Range
1262 then
1263 Bad_Range_Attribute (Sloc (Expr_Node));
1264 return Error;
1265 end if;
1267 if Ada_Version = Ada_83 then
1268 Error_Msg_SC ("(Ada 83) extension aggregate not allowed");
1269 end if;
1271 Aggregate_Node := New_Node (N_Extension_Aggregate, Lparen_Sloc);
1272 Set_Ancestor_Part (Aggregate_Node, Expr_Node);
1273 Scan; -- past WITH
1275 -- Deal with WITH NULL RECORD case
1277 if Token = Tok_Null then
1278 Save_Scan_State (Scan_State); -- at NULL
1279 Scan; -- past NULL
1281 if Token = Tok_Record then
1282 Scan; -- past RECORD
1283 Set_Null_Record_Present (Aggregate_Node, True);
1284 T_Right_Paren;
1285 return Aggregate_Node;
1287 else
1288 Restore_Scan_State (Scan_State); -- to NULL that must be expr
1289 end if;
1290 end if;
1292 if Token /= Tok_Others then
1293 Save_Scan_State (Scan_State);
1294 Expr_Node := P_Expression;
1295 else
1296 Expr_Node := Empty;
1297 end if;
1299 -- Expression case
1301 elsif Token = Tok_Right_Paren or else Token in Token_Class_Eterm then
1302 if Nkind (Expr_Node) = N_Attribute_Reference
1303 and then Attribute_Name (Expr_Node) = Name_Range
1304 then
1305 Error_Msg
1306 ("|parentheses not allowed for range attribute", Lparen_Sloc);
1307 Scan; -- past right paren
1308 return Expr_Node;
1309 end if;
1311 -- Bump paren count of expression
1313 if Expr_Node /= Error then
1314 Set_Paren_Count (Expr_Node, Paren_Count (Expr_Node) + 1);
1315 end if;
1317 T_Right_Paren; -- past right paren (error message if none)
1318 return Expr_Node;
1320 -- Normal aggregate case
1322 else
1323 Aggregate_Node := New_Node (N_Aggregate, Lparen_Sloc);
1324 end if;
1326 -- Others case
1328 else
1329 Aggregate_Node := New_Node (N_Aggregate, Lparen_Sloc);
1330 Expr_Node := Empty;
1331 end if;
1333 -- Prepare to scan list of component associations
1335 Expr_List := No_List; -- don't set yet, maybe all named entries
1336 Assoc_List := No_List; -- don't set yet, maybe all positional entries
1338 -- This loop scans through component associations. On entry to the
1339 -- loop, an expression has been scanned at the start of the current
1340 -- association unless initial token was OTHERS, in which case
1341 -- Expr_Node is set to Empty.
1343 loop
1344 -- Deal with others association first. This is a named association
1346 if No (Expr_Node) then
1347 if No (Assoc_List) then
1348 Assoc_List := New_List;
1349 end if;
1351 Append (P_Record_Or_Array_Component_Association, Assoc_List);
1353 -- Improper use of WITH
1355 elsif Token = Tok_With then
1356 Error_Msg_SC ("WITH must be preceded by single expression in " &
1357 "extension aggregate");
1358 raise Error_Resync;
1360 -- Range attribute can only appear as part of a discrete choice list
1362 elsif Nkind (Expr_Node) = N_Attribute_Reference
1363 and then Attribute_Name (Expr_Node) = Name_Range
1364 and then Token /= Tok_Arrow
1365 and then Token /= Tok_Vertical_Bar
1366 then
1367 Bad_Range_Attribute (Sloc (Expr_Node));
1368 return Error;
1370 -- Assume positional case if comma, right paren, or literal or
1371 -- identifier or OTHERS follows (the latter cases are missing
1372 -- comma cases). Also assume positional if a semicolon follows,
1373 -- which can happen if there are missing parens
1375 elsif Token = Tok_Comma
1376 or else Token = Tok_Right_Paren
1377 or else Token = Tok_Others
1378 or else Token in Token_Class_Lit_Or_Name
1379 or else Token = Tok_Semicolon
1380 then
1381 if Present (Assoc_List) then
1382 Error_Msg_BC -- CODEFIX
1383 ("""='>"" expected (positional association cannot follow " &
1384 "named association)");
1385 end if;
1387 if No (Expr_List) then
1388 Expr_List := New_List;
1389 end if;
1391 Append (Expr_Node, Expr_List);
1393 -- Check for aggregate followed by left parent, maybe missing comma
1395 elsif Nkind (Expr_Node) = N_Aggregate
1396 and then Token = Tok_Left_Paren
1397 then
1398 T_Comma;
1400 if No (Expr_List) then
1401 Expr_List := New_List;
1402 end if;
1404 Append (Expr_Node, Expr_List);
1406 -- Anything else is assumed to be a named association
1408 else
1409 Restore_Scan_State (Scan_State); -- to start of expression
1411 if No (Assoc_List) then
1412 Assoc_List := New_List;
1413 end if;
1415 Append (P_Record_Or_Array_Component_Association, Assoc_List);
1416 end if;
1418 exit when not Comma_Present;
1420 -- If we are at an expression terminator, something is seriously
1421 -- wrong, so let's get out now, before we start eating up stuff
1422 -- that doesn't belong to us!
1424 if Token in Token_Class_Eterm then
1426 -- If Some becomes a keyword, the following is needed to make it
1427 -- acceptable in older versions of Ada.
1429 if Token = Tok_Some
1430 and then Ada_Version < Ada_2012
1431 then
1432 Scan_Reserved_Identifier (False);
1433 else
1434 Error_Msg_AP
1435 ("expecting expression or component association");
1436 exit;
1437 end if;
1438 end if;
1440 -- Deal with misused box
1442 if Token = Tok_Box then
1443 Box_Error;
1445 -- Otherwise initiate for reentry to top of loop by scanning an
1446 -- initial expression, unless the first token is OTHERS.
1448 elsif Token = Tok_Others then
1449 Expr_Node := Empty;
1451 else
1452 Save_Scan_State (Scan_State); -- at start of expression
1453 Expr_Node := P_Expression_Or_Range_Attribute_If_OK;
1455 end if;
1456 end loop;
1458 -- All component associations (positional and named) have been scanned
1460 T_Right_Paren;
1461 Set_Expressions (Aggregate_Node, Expr_List);
1462 Set_Component_Associations (Aggregate_Node, Assoc_List);
1463 return Aggregate_Node;
1464 end P_Aggregate_Or_Paren_Expr;
1466 ------------------------------------------------
1467 -- 4.3 Record or Array Component Association --
1468 ------------------------------------------------
1470 -- RECORD_COMPONENT_ASSOCIATION ::=
1471 -- [COMPONENT_CHOICE_LIST =>] EXPRESSION
1472 -- | COMPONENT_CHOICE_LIST => <>
1474 -- COMPONENT_CHOICE_LIST =>
1475 -- component_SELECTOR_NAME {| component_SELECTOR_NAME}
1476 -- | others
1478 -- ARRAY_COMPONENT_ASSOCIATION ::=
1479 -- DISCRETE_CHOICE_LIST => EXPRESSION
1480 -- | DISCRETE_CHOICE_LIST => <>
1482 -- Note: this routine only handles the named cases, including others.
1483 -- Cases where the component choice list is not present have already
1484 -- been handled directly.
1486 -- Error recovery: can raise Error_Resync
1488 -- Note: RECORD_COMPONENT_ASSOCIATION and ARRAY_COMPONENT_ASSOCIATION
1489 -- rules have been extended to give support to Ada 2005 limited
1490 -- aggregates (AI-287)
1492 function P_Record_Or_Array_Component_Association return Node_Id is
1493 Assoc_Node : Node_Id;
1495 begin
1496 Assoc_Node := New_Node (N_Component_Association, Token_Ptr);
1497 Set_Choices (Assoc_Node, P_Discrete_Choice_List);
1498 Set_Sloc (Assoc_Node, Token_Ptr);
1499 TF_Arrow;
1501 if Token = Tok_Box then
1503 -- Ada 2005(AI-287): The box notation is used to indicate the
1504 -- default initialization of aggregate components
1506 if Ada_Version < Ada_2005 then
1507 Error_Msg_SP
1508 ("component association with '<'> is an Ada 2005 extension");
1509 Error_Msg_SP ("\unit must be compiled with -gnat05 switch");
1510 end if;
1512 Set_Box_Present (Assoc_Node);
1513 Scan; -- Past box
1514 else
1515 Set_Expression (Assoc_Node, P_Expression);
1516 end if;
1518 return Assoc_Node;
1519 end P_Record_Or_Array_Component_Association;
1521 -----------------------------
1522 -- 4.3.1 Record Aggregate --
1523 -----------------------------
1525 -- Case of enumeration aggregate is parsed by P_Aggregate (4.3)
1526 -- All other cases are parsed by P_Aggregate_Or_Paren_Expr (4.3)
1528 ----------------------------------------------
1529 -- 4.3.1 Record Component Association List --
1530 ----------------------------------------------
1532 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1534 ----------------------------------
1535 -- 4.3.1 Component Choice List --
1536 ----------------------------------
1538 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1540 --------------------------------
1541 -- 4.3.1 Extension Aggregate --
1542 --------------------------------
1544 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1546 --------------------------
1547 -- 4.3.1 Ancestor Part --
1548 --------------------------
1550 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1552 ----------------------------
1553 -- 4.3.1 Array Aggregate --
1554 ----------------------------
1556 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1558 ---------------------------------------
1559 -- 4.3.1 Positional Array Aggregate --
1560 ---------------------------------------
1562 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1564 ----------------------------------
1565 -- 4.3.1 Named Array Aggregate --
1566 ----------------------------------
1568 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1570 ----------------------------------------
1571 -- 4.3.1 Array Component Association --
1572 ----------------------------------------
1574 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1576 ---------------------
1577 -- 4.4 Expression --
1578 ---------------------
1580 -- EXPRESSION ::=
1581 -- RELATION {and RELATION} | RELATION {and then RELATION}
1582 -- | RELATION {or RELATION} | RELATION {or else RELATION}
1583 -- | RELATION {xor RELATION}
1585 -- On return, Expr_Form indicates the categorization of the expression
1586 -- EF_Range_Attr is not a possible value (if a range attribute is found,
1587 -- an error message is given, and Error is returned).
1589 -- Error recovery: cannot raise Error_Resync
1591 function P_Expression return Node_Id is
1592 Logical_Op : Node_Kind;
1593 Prev_Logical_Op : Node_Kind;
1594 Op_Location : Source_Ptr;
1595 Node1 : Node_Id;
1596 Node2 : Node_Id;
1598 begin
1599 Node1 := P_Relation;
1601 if Token in Token_Class_Logop then
1602 Prev_Logical_Op := N_Empty;
1604 loop
1605 Op_Location := Token_Ptr;
1606 Logical_Op := P_Logical_Operator;
1608 if Prev_Logical_Op /= N_Empty and then
1609 Logical_Op /= Prev_Logical_Op
1610 then
1611 Error_Msg
1612 ("mixed logical operators in expression", Op_Location);
1613 Prev_Logical_Op := N_Empty;
1614 else
1615 Prev_Logical_Op := Logical_Op;
1616 end if;
1618 Node2 := Node1;
1619 Node1 := New_Op_Node (Logical_Op, Op_Location);
1620 Set_Left_Opnd (Node1, Node2);
1621 Set_Right_Opnd (Node1, P_Relation);
1622 exit when Token not in Token_Class_Logop;
1623 end loop;
1625 Expr_Form := EF_Non_Simple;
1626 end if;
1628 if Token = Tok_Apostrophe then
1629 Bad_Range_Attribute (Token_Ptr);
1630 return Error;
1631 else
1632 return Node1;
1633 end if;
1634 end P_Expression;
1636 -- This function is identical to the normal P_Expression, except that it
1637 -- also permits the appearance of a case, conditional, or quantified
1638 -- expression without the usual surrounding parentheses.
1640 function P_Expression_If_OK return Node_Id is
1641 begin
1642 if Token = Tok_Case then
1643 return P_Case_Expression;
1645 elsif Token = Tok_If then
1646 return P_Conditional_Expression;
1648 elsif Token = Tok_For then
1649 return P_Quantified_Expression;
1651 else
1652 return P_Expression;
1653 end if;
1654 end P_Expression_If_OK;
1656 -- This function is identical to the normal P_Expression, except that it
1657 -- checks that the expression scan did not stop on a right paren. It is
1658 -- called in all contexts where a right parenthesis cannot legitimately
1659 -- follow an expression.
1661 -- Error recovery: can not raise Error_Resync
1663 function P_Expression_No_Right_Paren return Node_Id is
1664 Expr : constant Node_Id := P_Expression;
1665 begin
1666 Ignore (Tok_Right_Paren);
1667 return Expr;
1668 end P_Expression_No_Right_Paren;
1670 ----------------------------------------
1671 -- 4.4 Expression_Or_Range_Attribute --
1672 ----------------------------------------
1674 -- EXPRESSION ::=
1675 -- RELATION {and RELATION} | RELATION {and then RELATION}
1676 -- | RELATION {or RELATION} | RELATION {or else RELATION}
1677 -- | RELATION {xor RELATION}
1679 -- RANGE_ATTRIBUTE_REFERENCE ::= PREFIX ' RANGE_ATTRIBUTE_DESIGNATOR
1681 -- RANGE_ATTRIBUTE_DESIGNATOR ::= range [(static_EXPRESSION)]
1683 -- On return, Expr_Form indicates the categorization of the expression
1684 -- and EF_Range_Attr is one of the possibilities.
1686 -- Error recovery: cannot raise Error_Resync
1688 -- In the grammar, a RANGE attribute is simply a name, but its use is
1689 -- highly restricted, so in the parser, we do not regard it as a name.
1690 -- Instead, P_Name returns without scanning the 'RANGE part of the
1691 -- attribute, and P_Expression_Or_Range_Attribute handles the range
1692 -- attribute reference. In the normal case where a range attribute is
1693 -- not allowed, an error message is issued by P_Expression.
1695 function P_Expression_Or_Range_Attribute return Node_Id is
1696 Logical_Op : Node_Kind;
1697 Prev_Logical_Op : Node_Kind;
1698 Op_Location : Source_Ptr;
1699 Node1 : Node_Id;
1700 Node2 : Node_Id;
1701 Attr_Node : Node_Id;
1703 begin
1704 Node1 := P_Relation;
1706 if Token = Tok_Apostrophe then
1707 Attr_Node := P_Range_Attribute_Reference (Node1);
1708 Expr_Form := EF_Range_Attr;
1709 return Attr_Node;
1711 elsif Token in Token_Class_Logop then
1712 Prev_Logical_Op := N_Empty;
1714 loop
1715 Op_Location := Token_Ptr;
1716 Logical_Op := P_Logical_Operator;
1718 if Prev_Logical_Op /= N_Empty and then
1719 Logical_Op /= Prev_Logical_Op
1720 then
1721 Error_Msg
1722 ("mixed logical operators in expression", Op_Location);
1723 Prev_Logical_Op := N_Empty;
1724 else
1725 Prev_Logical_Op := Logical_Op;
1726 end if;
1728 Node2 := Node1;
1729 Node1 := New_Op_Node (Logical_Op, Op_Location);
1730 Set_Left_Opnd (Node1, Node2);
1731 Set_Right_Opnd (Node1, P_Relation);
1732 exit when Token not in Token_Class_Logop;
1733 end loop;
1735 Expr_Form := EF_Non_Simple;
1736 end if;
1738 if Token = Tok_Apostrophe then
1739 Bad_Range_Attribute (Token_Ptr);
1740 return Error;
1741 else
1742 return Node1;
1743 end if;
1744 end P_Expression_Or_Range_Attribute;
1746 -- Version that allows a non-parenthesized case, conditional, or quantified
1747 -- expression
1749 function P_Expression_Or_Range_Attribute_If_OK return Node_Id is
1750 begin
1751 if Token = Tok_Case then
1752 return P_Case_Expression;
1754 elsif Token = Tok_If then
1755 return P_Conditional_Expression;
1757 elsif Token = Tok_For then
1758 return P_Quantified_Expression;
1760 else
1761 return P_Expression_Or_Range_Attribute;
1762 end if;
1763 end P_Expression_Or_Range_Attribute_If_OK;
1765 -------------------
1766 -- 4.4 Relation --
1767 -------------------
1769 -- RELATION ::=
1770 -- SIMPLE_EXPRESSION [RELATIONAL_OPERATOR SIMPLE_EXPRESSION]
1771 -- | SIMPLE_EXPRESSION [not] in MEMBERSHIP_CHOICE_LIST
1773 -- On return, Expr_Form indicates the categorization of the expression
1775 -- Note: if Token = Tok_Apostrophe on return, then Expr_Form is set to
1776 -- EF_Simple_Name and the following token is RANGE (range attribute case).
1778 -- Error recovery: cannot raise Error_Resync. If an error occurs within an
1779 -- expression, then tokens are scanned until either a non-expression token,
1780 -- a right paren (not matched by a left paren) or a comma, is encountered.
1782 function P_Relation return Node_Id is
1783 Node1, Node2 : Node_Id;
1784 Optok : Source_Ptr;
1786 begin
1787 Node1 := P_Simple_Expression;
1789 if Token not in Token_Class_Relop then
1790 return Node1;
1792 else
1793 -- Here we have a relational operator following. If so then scan it
1794 -- out. Note that the assignment symbol := is treated as a relational
1795 -- operator to improve the error recovery when it is misused for =.
1796 -- P_Relational_Operator also parses the IN and NOT IN operations.
1798 Optok := Token_Ptr;
1799 Node2 := New_Op_Node (P_Relational_Operator, Optok);
1800 Set_Left_Opnd (Node2, Node1);
1802 -- Case of IN or NOT IN
1804 if Prev_Token = Tok_In then
1805 P_Membership_Test (Node2);
1807 -- Case of relational operator (= /= < <= > >=)
1809 else
1810 Set_Right_Opnd (Node2, P_Simple_Expression);
1811 end if;
1813 Expr_Form := EF_Non_Simple;
1815 if Token in Token_Class_Relop then
1816 Error_Msg_SC ("unexpected relational operator");
1817 raise Error_Resync;
1818 end if;
1820 return Node2;
1821 end if;
1823 -- If any error occurs, then scan to the next expression terminator symbol
1824 -- or comma or right paren at the outer (i.e. current) parentheses level.
1825 -- The flags are set to indicate a normal simple expression.
1827 exception
1828 when Error_Resync =>
1829 Resync_Expression;
1830 Expr_Form := EF_Simple;
1831 return Error;
1832 end P_Relation;
1834 ----------------------------
1835 -- 4.4 Simple Expression --
1836 ----------------------------
1838 -- SIMPLE_EXPRESSION ::=
1839 -- [UNARY_ADDING_OPERATOR] TERM {BINARY_ADDING_OPERATOR TERM}
1841 -- On return, Expr_Form indicates the categorization of the expression
1843 -- Note: if Token = Tok_Apostrophe on return, then Expr_Form is set to
1844 -- EF_Simple_Name and the following token is RANGE (range attribute case).
1846 -- Error recovery: cannot raise Error_Resync. If an error occurs within an
1847 -- expression, then tokens are scanned until either a non-expression token,
1848 -- a right paren (not matched by a left paren) or a comma, is encountered.
1850 -- Note: P_Simple_Expression is called only internally by higher level
1851 -- expression routines. In cases in the grammar where a simple expression
1852 -- is required, the approach is to scan an expression, and then post an
1853 -- appropriate error message if the expression obtained is not simple. This
1854 -- gives better error recovery and treatment.
1856 function P_Simple_Expression return Node_Id is
1857 Scan_State : Saved_Scan_State;
1858 Node1 : Node_Id;
1859 Node2 : Node_Id;
1860 Tokptr : Source_Ptr;
1862 begin
1863 -- Check for cases starting with a name. There are two reasons for
1864 -- special casing. First speed things up by catching a common case
1865 -- without going through several routine layers. Second the caller must
1866 -- be informed via Expr_Form when the simple expression is a name.
1868 if Token in Token_Class_Name then
1869 Node1 := P_Name;
1871 -- Deal with apostrophe cases
1873 if Token = Tok_Apostrophe then
1874 Save_Scan_State (Scan_State); -- at apostrophe
1875 Scan; -- past apostrophe
1877 -- If qualified expression, scan it out and fall through
1879 if Token = Tok_Left_Paren then
1880 Node1 := P_Qualified_Expression (Node1);
1881 Expr_Form := EF_Simple;
1883 -- If range attribute, then we return with Token pointing to the
1884 -- apostrophe. Note: avoid the normal error check on exit. We
1885 -- know that the expression really is complete in this case!
1887 else -- Token = Tok_Range then
1888 Restore_Scan_State (Scan_State); -- to apostrophe
1889 Expr_Form := EF_Simple_Name;
1890 return Node1;
1891 end if;
1892 end if;
1894 -- If an expression terminator follows, the previous processing
1895 -- completely scanned out the expression (a common case), and
1896 -- left Expr_Form set appropriately for returning to our caller.
1898 if Token in Token_Class_Sterm then
1899 null;
1901 -- If we do not have an expression terminator, then complete the
1902 -- scan of a simple expression. This code duplicates the code
1903 -- found in P_Term and P_Factor.
1905 else
1906 if Token = Tok_Double_Asterisk then
1907 if Style_Check then
1908 Style.Check_Exponentiation_Operator;
1909 end if;
1911 Node2 := New_Op_Node (N_Op_Expon, Token_Ptr);
1912 Scan; -- past **
1913 Set_Left_Opnd (Node2, Node1);
1914 Set_Right_Opnd (Node2, P_Primary);
1915 Node1 := Node2;
1916 end if;
1918 loop
1919 exit when Token not in Token_Class_Mulop;
1920 Tokptr := Token_Ptr;
1921 Node2 := New_Op_Node (P_Multiplying_Operator, Tokptr);
1923 if Style_Check then
1924 Style.Check_Binary_Operator;
1925 end if;
1927 Scan; -- past operator
1928 Set_Left_Opnd (Node2, Node1);
1929 Set_Right_Opnd (Node2, P_Factor);
1930 Node1 := Node2;
1931 end loop;
1933 loop
1934 exit when Token not in Token_Class_Binary_Addop;
1935 Tokptr := Token_Ptr;
1936 Node2 := New_Op_Node (P_Binary_Adding_Operator, Tokptr);
1938 if Style_Check then
1939 Style.Check_Binary_Operator;
1940 end if;
1942 Scan; -- past operator
1943 Set_Left_Opnd (Node2, Node1);
1944 Set_Right_Opnd (Node2, P_Term);
1945 Node1 := Node2;
1946 end loop;
1948 Expr_Form := EF_Simple;
1949 end if;
1951 -- Cases where simple expression does not start with a name
1953 else
1954 -- Scan initial sign and initial Term
1956 if Token in Token_Class_Unary_Addop then
1957 Tokptr := Token_Ptr;
1958 Node1 := New_Op_Node (P_Unary_Adding_Operator, Tokptr);
1960 if Style_Check then
1961 Style.Check_Unary_Plus_Or_Minus;
1962 end if;
1964 Scan; -- past operator
1965 Set_Right_Opnd (Node1, P_Term);
1966 else
1967 Node1 := P_Term;
1968 end if;
1970 -- In the following, we special-case a sequence of concatenations of
1971 -- string literals, such as "aaa" & "bbb" & ... & "ccc", with nothing
1972 -- else mixed in. For such a sequence, we return a tree representing
1973 -- "" & "aaabbb...ccc" (a single concatenation). This is done only if
1974 -- the number of concatenations is large. If semantic analysis
1975 -- resolves the "&" to a predefined one, then this folding gives the
1976 -- right answer. Otherwise, semantic analysis will complain about a
1977 -- capacity-exceeded error. The purpose of this trick is to avoid
1978 -- creating a deeply nested tree, which would cause deep recursion
1979 -- during semantics, causing stack overflow. This way, we can handle
1980 -- enormous concatenations in the normal case of predefined "&". We
1981 -- first build up the normal tree, and then rewrite it if
1982 -- appropriate.
1984 declare
1985 Num_Concats_Threshold : constant Positive := 1000;
1986 -- Arbitrary threshold value to enable optimization
1988 First_Node : constant Node_Id := Node1;
1989 Is_Strlit_Concat : Boolean;
1990 -- True iff we've parsed a sequence of concatenations of string
1991 -- literals, with nothing else mixed in.
1993 Num_Concats : Natural;
1994 -- Number of "&" operators if Is_Strlit_Concat is True
1996 begin
1997 Is_Strlit_Concat :=
1998 Nkind (Node1) = N_String_Literal
1999 and then Token = Tok_Ampersand;
2000 Num_Concats := 0;
2002 -- Scan out sequence of terms separated by binary adding operators
2004 loop
2005 exit when Token not in Token_Class_Binary_Addop;
2006 Tokptr := Token_Ptr;
2007 Node2 := New_Op_Node (P_Binary_Adding_Operator, Tokptr);
2008 Scan; -- past operator
2009 Set_Left_Opnd (Node2, Node1);
2010 Node1 := P_Term;
2011 Set_Right_Opnd (Node2, Node1);
2013 -- Check if we're still concatenating string literals
2015 Is_Strlit_Concat :=
2016 Is_Strlit_Concat
2017 and then Nkind (Node2) = N_Op_Concat
2018 and then Nkind (Node1) = N_String_Literal;
2020 if Is_Strlit_Concat then
2021 Num_Concats := Num_Concats + 1;
2022 end if;
2024 Node1 := Node2;
2025 end loop;
2027 -- If we have an enormous series of concatenations of string
2028 -- literals, rewrite as explained above. The Is_Folded_In_Parser
2029 -- flag tells semantic analysis that if the "&" is not predefined,
2030 -- the folded value is wrong.
2032 if Is_Strlit_Concat
2033 and then Num_Concats >= Num_Concats_Threshold
2034 then
2035 declare
2036 Empty_String_Val : String_Id;
2037 -- String_Id for ""
2039 Strlit_Concat_Val : String_Id;
2040 -- Contains the folded value (which will be correct if the
2041 -- "&" operators are the predefined ones).
2043 Cur_Node : Node_Id;
2044 -- For walking up the tree
2046 New_Node : Node_Id;
2047 -- Folded node to replace Node1
2049 Loc : constant Source_Ptr := Sloc (First_Node);
2051 begin
2052 -- Walk up the tree starting at the leftmost string literal
2053 -- (First_Node), building up the Strlit_Concat_Val as we
2054 -- go. Note that we do not use recursion here -- the whole
2055 -- point is to avoid recursively walking that enormous tree.
2057 Start_String;
2058 Store_String_Chars (Strval (First_Node));
2060 Cur_Node := Parent (First_Node);
2061 while Present (Cur_Node) loop
2062 pragma Assert (Nkind (Cur_Node) = N_Op_Concat and then
2063 Nkind (Right_Opnd (Cur_Node)) = N_String_Literal);
2065 Store_String_Chars (Strval (Right_Opnd (Cur_Node)));
2066 Cur_Node := Parent (Cur_Node);
2067 end loop;
2069 Strlit_Concat_Val := End_String;
2071 -- Create new folded node, and rewrite result with a concat-
2072 -- enation of an empty string literal and the folded node.
2074 Start_String;
2075 Empty_String_Val := End_String;
2076 New_Node :=
2077 Make_Op_Concat (Loc,
2078 Make_String_Literal (Loc, Empty_String_Val),
2079 Make_String_Literal (Loc, Strlit_Concat_Val,
2080 Is_Folded_In_Parser => True));
2081 Rewrite (Node1, New_Node);
2082 end;
2083 end if;
2084 end;
2086 -- All done, we clearly do not have name or numeric literal so this
2087 -- is a case of a simple expression which is some other possibility.
2089 Expr_Form := EF_Simple;
2090 end if;
2092 -- Come here at end of simple expression, where we do a couple of
2093 -- special checks to improve error recovery.
2095 -- Special test to improve error recovery. If the current token
2096 -- is a period, then someone is trying to do selection on something
2097 -- that is not a name, e.g. a qualified expression.
2099 if Token = Tok_Dot then
2100 Error_Msg_SC ("prefix for selection is not a name");
2102 -- If qualified expression, comment and continue, otherwise something
2103 -- is pretty nasty so do an Error_Resync call.
2105 if Ada_Version < Ada_2012
2106 and then Nkind (Node1) = N_Qualified_Expression
2107 then
2108 Error_Msg_SC ("\would be legal in Ada 2012 mode");
2109 else
2110 raise Error_Resync;
2111 end if;
2112 end if;
2114 -- Special test to improve error recovery: If the current token is
2115 -- not the first token on a line (as determined by checking the
2116 -- previous token position with the start of the current line),
2117 -- then we insist that we have an appropriate terminating token.
2118 -- Consider the following two examples:
2120 -- 1) if A nad B then ...
2122 -- 2) A := B
2123 -- C := D
2125 -- In the first example, we would like to issue a binary operator
2126 -- expected message and resynchronize to the then. In the second
2127 -- example, we do not want to issue a binary operator message, so
2128 -- that instead we will get the missing semicolon message. This
2129 -- distinction is of course a heuristic which does not always work,
2130 -- but in practice it is quite effective.
2132 -- Note: the one case in which we do not go through this circuit is
2133 -- when we have scanned a range attribute and want to return with
2134 -- Token pointing to the apostrophe. The apostrophe is not normally
2135 -- an expression terminator, and is not in Token_Class_Sterm, but
2136 -- in this special case we know that the expression is complete.
2138 if not Token_Is_At_Start_Of_Line
2139 and then Token not in Token_Class_Sterm
2140 then
2141 -- Normally the right error message is indeed that we expected a
2142 -- binary operator, but in the case of being between a right and left
2143 -- paren, e.g. in an aggregate, a more likely error is missing comma.
2145 if Prev_Token = Tok_Right_Paren and then Token = Tok_Left_Paren then
2146 T_Comma;
2147 else
2148 Error_Msg_AP ("binary operator expected");
2149 end if;
2151 raise Error_Resync;
2153 else
2154 return Node1;
2155 end if;
2157 -- If any error occurs, then scan to next expression terminator symbol
2158 -- or comma, right paren or vertical bar at the outer (i.e. current) paren
2159 -- level. Expr_Form is set to indicate a normal simple expression.
2161 exception
2162 when Error_Resync =>
2163 Resync_Expression;
2164 Expr_Form := EF_Simple;
2165 return Error;
2166 end P_Simple_Expression;
2168 -----------------------------------------------
2169 -- 4.4 Simple Expression or Range Attribute --
2170 -----------------------------------------------
2172 -- SIMPLE_EXPRESSION ::=
2173 -- [UNARY_ADDING_OPERATOR] TERM {BINARY_ADDING_OPERATOR TERM}
2175 -- RANGE_ATTRIBUTE_REFERENCE ::= PREFIX ' RANGE_ATTRIBUTE_DESIGNATOR
2177 -- RANGE_ATTRIBUTE_DESIGNATOR ::= range [(static_EXPRESSION)]
2179 -- Error recovery: cannot raise Error_Resync
2181 function P_Simple_Expression_Or_Range_Attribute return Node_Id is
2182 Sexpr : Node_Id;
2183 Attr_Node : Node_Id;
2185 begin
2186 -- We don't just want to roar ahead and call P_Simple_Expression
2187 -- here, since we want to handle the case of a parenthesized range
2188 -- attribute cleanly.
2190 if Token = Tok_Left_Paren then
2191 declare
2192 Lptr : constant Source_Ptr := Token_Ptr;
2193 Scan_State : Saved_Scan_State;
2195 begin
2196 Save_Scan_State (Scan_State);
2197 Scan; -- past left paren
2198 Sexpr := P_Simple_Expression;
2200 if Token = Tok_Apostrophe then
2201 Attr_Node := P_Range_Attribute_Reference (Sexpr);
2202 Expr_Form := EF_Range_Attr;
2204 if Token = Tok_Right_Paren then
2205 Scan; -- scan past right paren if present
2206 end if;
2208 Error_Msg ("parentheses not allowed for range attribute", Lptr);
2210 return Attr_Node;
2211 end if;
2213 Restore_Scan_State (Scan_State);
2214 end;
2215 end if;
2217 -- Here after dealing with parenthesized range attribute
2219 Sexpr := P_Simple_Expression;
2221 if Token = Tok_Apostrophe then
2222 Attr_Node := P_Range_Attribute_Reference (Sexpr);
2223 Expr_Form := EF_Range_Attr;
2224 return Attr_Node;
2226 else
2227 return Sexpr;
2228 end if;
2229 end P_Simple_Expression_Or_Range_Attribute;
2231 ---------------
2232 -- 4.4 Term --
2233 ---------------
2235 -- TERM ::= FACTOR {MULTIPLYING_OPERATOR FACTOR}
2237 -- Error recovery: can raise Error_Resync
2239 function P_Term return Node_Id is
2240 Node1, Node2 : Node_Id;
2241 Tokptr : Source_Ptr;
2243 begin
2244 Node1 := P_Factor;
2246 loop
2247 exit when Token not in Token_Class_Mulop;
2248 Tokptr := Token_Ptr;
2249 Node2 := New_Op_Node (P_Multiplying_Operator, Tokptr);
2250 Scan; -- past operator
2251 Set_Left_Opnd (Node2, Node1);
2252 Set_Right_Opnd (Node2, P_Factor);
2253 Node1 := Node2;
2254 end loop;
2256 return Node1;
2257 end P_Term;
2259 -----------------
2260 -- 4.4 Factor --
2261 -----------------
2263 -- FACTOR ::= PRIMARY [** PRIMARY] | abs PRIMARY | not PRIMARY
2265 -- Error recovery: can raise Error_Resync
2267 function P_Factor return Node_Id is
2268 Node1 : Node_Id;
2269 Node2 : Node_Id;
2271 begin
2272 if Token = Tok_Abs then
2273 Node1 := New_Op_Node (N_Op_Abs, Token_Ptr);
2275 if Style_Check then
2276 Style.Check_Abs_Not;
2277 end if;
2279 Scan; -- past ABS
2280 Set_Right_Opnd (Node1, P_Primary);
2281 return Node1;
2283 elsif Token = Tok_Not then
2284 Node1 := New_Op_Node (N_Op_Not, Token_Ptr);
2286 if Style_Check then
2287 Style.Check_Abs_Not;
2288 end if;
2290 Scan; -- past NOT
2291 Set_Right_Opnd (Node1, P_Primary);
2292 return Node1;
2294 else
2295 Node1 := P_Primary;
2297 if Token = Tok_Double_Asterisk then
2298 Node2 := New_Op_Node (N_Op_Expon, Token_Ptr);
2299 Scan; -- past **
2300 Set_Left_Opnd (Node2, Node1);
2301 Set_Right_Opnd (Node2, P_Primary);
2302 return Node2;
2303 else
2304 return Node1;
2305 end if;
2306 end if;
2307 end P_Factor;
2309 ------------------
2310 -- 4.4 Primary --
2311 ------------------
2313 -- PRIMARY ::=
2314 -- NUMERIC_LITERAL | null
2315 -- | STRING_LITERAL | AGGREGATE
2316 -- | NAME | QUALIFIED_EXPRESSION
2317 -- | ALLOCATOR | (EXPRESSION) | QUANTIFIED_EXPRESSION
2319 -- Error recovery: can raise Error_Resync
2321 function P_Primary return Node_Id is
2322 Scan_State : Saved_Scan_State;
2323 Node1 : Node_Id;
2325 begin
2326 -- The loop runs more than once only if misplaced pragmas are found
2328 loop
2329 case Token is
2331 -- Name token can start a name, call or qualified expression, all
2332 -- of which are acceptable possibilities for primary. Note also
2333 -- that string literal is included in name (as operator symbol)
2334 -- and type conversion is included in name (as indexed component).
2336 when Tok_Char_Literal | Tok_Operator_Symbol | Tok_Identifier =>
2337 Node1 := P_Name;
2339 -- All done unless apostrophe follows
2341 if Token /= Tok_Apostrophe then
2342 return Node1;
2344 -- Apostrophe following means that we have either just parsed
2345 -- the subtype mark of a qualified expression, or the prefix
2346 -- or a range attribute.
2348 else -- Token = Tok_Apostrophe
2349 Save_Scan_State (Scan_State); -- at apostrophe
2350 Scan; -- past apostrophe
2352 -- If range attribute, then this is always an error, since
2353 -- the only legitimate case (where the scanned expression is
2354 -- a qualified simple name) is handled at the level of the
2355 -- Simple_Expression processing. This case corresponds to a
2356 -- usage such as 3 + A'Range, which is always illegal.
2358 if Token = Tok_Range then
2359 Restore_Scan_State (Scan_State); -- to apostrophe
2360 Bad_Range_Attribute (Token_Ptr);
2361 return Error;
2363 -- If left paren, then we have a qualified expression.
2364 -- Note that P_Name guarantees that in this case, where
2365 -- Token = Tok_Apostrophe on return, the only two possible
2366 -- tokens following the apostrophe are left paren and
2367 -- RANGE, so we know we have a left paren here.
2369 else -- Token = Tok_Left_Paren
2370 return P_Qualified_Expression (Node1);
2372 end if;
2373 end if;
2375 -- Numeric or string literal
2377 when Tok_Integer_Literal |
2378 Tok_Real_Literal |
2379 Tok_String_Literal =>
2381 Node1 := Token_Node;
2382 Scan; -- past number
2383 return Node1;
2385 -- Left paren, starts aggregate or parenthesized expression
2387 when Tok_Left_Paren =>
2388 declare
2389 Expr : constant Node_Id := P_Aggregate_Or_Paren_Expr;
2391 begin
2392 if Nkind (Expr) = N_Attribute_Reference
2393 and then Attribute_Name (Expr) = Name_Range
2394 then
2395 Bad_Range_Attribute (Sloc (Expr));
2396 end if;
2398 return Expr;
2399 end;
2401 -- Allocator
2403 when Tok_New =>
2404 return P_Allocator;
2406 -- Null
2408 when Tok_Null =>
2409 Scan; -- past NULL
2410 return New_Node (N_Null, Prev_Token_Ptr);
2412 -- Pragma, not allowed here, so just skip past it
2414 when Tok_Pragma =>
2415 P_Pragmas_Misplaced;
2417 -- Deal with IF (possible unparenthesized conditional expression)
2419 when Tok_If =>
2421 -- If this looks like a real if, defined as an IF appearing at
2422 -- the start of a new line, then we consider we have a missing
2423 -- operand.
2425 if Token_Is_At_Start_Of_Line then
2426 Error_Msg_AP ("missing operand");
2427 return Error;
2429 -- If this looks like a conditional expression, then treat it
2430 -- that way with an error message.
2432 elsif Ada_Version >= Ada_2012 then
2433 Error_Msg_SC
2434 ("conditional expression must be parenthesized");
2435 return P_Conditional_Expression;
2437 -- Otherwise treat as misused identifier
2439 else
2440 return P_Identifier;
2441 end if;
2443 -- Deal with CASE (possible unparenthesized case expression)
2445 when Tok_Case =>
2447 -- If this looks like a real case, defined as a CASE appearing
2448 -- the start of a new line, then we consider we have a missing
2449 -- operand.
2451 if Token_Is_At_Start_Of_Line then
2452 Error_Msg_AP ("missing operand");
2453 return Error;
2455 -- If this looks like a case expression, then treat it that way
2456 -- with an error message.
2458 elsif Ada_Version >= Ada_2012 then
2459 Error_Msg_SC ("case expression must be parenthesized");
2460 return P_Case_Expression;
2462 -- Otherwise treat as misused identifier
2464 else
2465 return P_Identifier;
2466 end if;
2468 -- For [all | some] indicates a quantified expression
2470 when Tok_For =>
2472 if Token_Is_At_Start_Of_Line then
2473 Error_Msg_AP ("misplaced loop");
2474 return Error;
2476 elsif Ada_Version >= Ada_2012 then
2477 Error_Msg_SC ("quantified expression must be parenthesized");
2478 return P_Quantified_Expression;
2480 else
2482 -- Otherwise treat as misused identifier
2484 return P_Identifier;
2485 end if;
2487 -- Anything else is illegal as the first token of a primary, but
2488 -- we test for a reserved identifier so that it is treated nicely
2490 when others =>
2491 if Is_Reserved_Identifier then
2492 return P_Identifier;
2494 elsif Prev_Token = Tok_Comma then
2495 Error_Msg_SP -- CODEFIX
2496 ("|extra "","" ignored");
2497 raise Error_Resync;
2499 else
2500 Error_Msg_AP ("missing operand");
2501 raise Error_Resync;
2502 end if;
2504 end case;
2505 end loop;
2506 end P_Primary;
2508 -------------------------------
2509 -- 4.4 Quantified_Expression --
2510 -------------------------------
2512 -- QUANTIFIED_EXPRESSION ::=
2513 -- for QUANTIFIER LOOP_PARAMETER_SPECIFICATION => PREDICATE |
2514 -- for QUANTIFIER ITERATOR_SPECIFICATION => PREDICATE
2516 function P_Quantified_Expression return Node_Id is
2517 Node1 : Node_Id;
2519 begin
2520 Scan; -- past FOR
2522 Node1 := New_Node (N_Quantified_Expression, Prev_Token_Ptr);
2524 if Token = Tok_All then
2525 Set_All_Present (Node1);
2527 -- We treat Some as a non-reserved keyword, so it appears to
2528 -- the scanner as an identifier. If Some is made into a reserved
2529 -- work, the check below is against Tok_Some.
2531 elsif Token /= Tok_Identifier
2532 or else Chars (Token_Node) /= Name_Some
2533 then
2534 Error_Msg_AP ("missing quantifier");
2535 raise Error_Resync;
2536 end if;
2538 Scan;
2539 Set_Loop_Parameter_Specification (Node1, P_Loop_Parameter_Specification);
2540 if Token = Tok_Arrow then
2541 Scan;
2542 Set_Condition (Node1, P_Expression);
2543 return Node1;
2544 else
2545 Error_Msg_AP ("missing arrow");
2546 raise Error_Resync;
2547 end if;
2548 end P_Quantified_Expression;
2550 ---------------------------
2551 -- 4.5 Logical Operator --
2552 ---------------------------
2554 -- LOGICAL_OPERATOR ::= and | or | xor
2556 -- Note: AND THEN and OR ELSE are also treated as logical operators
2557 -- by the parser (even though they are not operators semantically)
2559 -- The value returned is the appropriate Node_Kind code for the operator
2560 -- On return, Token points to the token following the scanned operator.
2562 -- The caller has checked that the first token is a legitimate logical
2563 -- operator token (i.e. is either XOR, AND, OR).
2565 -- Error recovery: cannot raise Error_Resync
2567 function P_Logical_Operator return Node_Kind is
2568 begin
2569 if Token = Tok_And then
2570 if Style_Check then
2571 Style.Check_Binary_Operator;
2572 end if;
2574 Scan; -- past AND
2576 if Token = Tok_Then then
2577 Scan; -- past THEN
2578 return N_And_Then;
2579 else
2580 return N_Op_And;
2581 end if;
2583 elsif Token = Tok_Or then
2584 if Style_Check then
2585 Style.Check_Binary_Operator;
2586 end if;
2588 Scan; -- past OR
2590 if Token = Tok_Else then
2591 Scan; -- past ELSE
2592 return N_Or_Else;
2593 else
2594 return N_Op_Or;
2595 end if;
2597 else -- Token = Tok_Xor
2598 if Style_Check then
2599 Style.Check_Binary_Operator;
2600 end if;
2602 Scan; -- past XOR
2603 return N_Op_Xor;
2604 end if;
2605 end P_Logical_Operator;
2607 ------------------------------
2608 -- 4.5 Relational Operator --
2609 ------------------------------
2611 -- RELATIONAL_OPERATOR ::= = | /= | < | <= | > | >=
2613 -- The value returned is the appropriate Node_Kind code for the operator.
2614 -- On return, Token points to the operator token, NOT past it.
2616 -- The caller has checked that the first token is a legitimate relational
2617 -- operator token (i.e. is one of the operator tokens listed above).
2619 -- Error recovery: cannot raise Error_Resync
2621 function P_Relational_Operator return Node_Kind is
2622 Op_Kind : Node_Kind;
2623 Relop_Node : constant array (Token_Class_Relop) of Node_Kind :=
2624 (Tok_Less => N_Op_Lt,
2625 Tok_Equal => N_Op_Eq,
2626 Tok_Greater => N_Op_Gt,
2627 Tok_Not_Equal => N_Op_Ne,
2628 Tok_Greater_Equal => N_Op_Ge,
2629 Tok_Less_Equal => N_Op_Le,
2630 Tok_In => N_In,
2631 Tok_Not => N_Not_In,
2632 Tok_Box => N_Op_Ne);
2634 begin
2635 if Token = Tok_Box then
2636 Error_Msg_SC -- CODEFIX
2637 ("|""'<'>"" should be ""/=""");
2638 end if;
2640 Op_Kind := Relop_Node (Token);
2642 if Style_Check then
2643 Style.Check_Binary_Operator;
2644 end if;
2646 Scan; -- past operator token
2648 if Prev_Token = Tok_Not then
2649 T_In;
2650 end if;
2652 return Op_Kind;
2653 end P_Relational_Operator;
2655 ---------------------------------
2656 -- 4.5 Binary Adding Operator --
2657 ---------------------------------
2659 -- BINARY_ADDING_OPERATOR ::= + | - | &
2661 -- The value returned is the appropriate Node_Kind code for the operator.
2662 -- On return, Token points to the operator token (NOT past it).
2664 -- The caller has checked that the first token is a legitimate adding
2665 -- operator token (i.e. is one of the operator tokens listed above).
2667 -- Error recovery: cannot raise Error_Resync
2669 function P_Binary_Adding_Operator return Node_Kind is
2670 Addop_Node : constant array (Token_Class_Binary_Addop) of Node_Kind :=
2671 (Tok_Ampersand => N_Op_Concat,
2672 Tok_Minus => N_Op_Subtract,
2673 Tok_Plus => N_Op_Add);
2674 begin
2675 return Addop_Node (Token);
2676 end P_Binary_Adding_Operator;
2678 --------------------------------
2679 -- 4.5 Unary Adding Operator --
2680 --------------------------------
2682 -- UNARY_ADDING_OPERATOR ::= + | -
2684 -- The value returned is the appropriate Node_Kind code for the operator.
2685 -- On return, Token points to the operator token (NOT past it).
2687 -- The caller has checked that the first token is a legitimate adding
2688 -- operator token (i.e. is one of the operator tokens listed above).
2690 -- Error recovery: cannot raise Error_Resync
2692 function P_Unary_Adding_Operator return Node_Kind is
2693 Addop_Node : constant array (Token_Class_Unary_Addop) of Node_Kind :=
2694 (Tok_Minus => N_Op_Minus,
2695 Tok_Plus => N_Op_Plus);
2696 begin
2697 return Addop_Node (Token);
2698 end P_Unary_Adding_Operator;
2700 -------------------------------
2701 -- 4.5 Multiplying Operator --
2702 -------------------------------
2704 -- MULTIPLYING_OPERATOR ::= * | / | mod | rem
2706 -- The value returned is the appropriate Node_Kind code for the operator.
2707 -- On return, Token points to the operator token (NOT past it).
2709 -- The caller has checked that the first token is a legitimate multiplying
2710 -- operator token (i.e. is one of the operator tokens listed above).
2712 -- Error recovery: cannot raise Error_Resync
2714 function P_Multiplying_Operator return Node_Kind is
2715 Mulop_Node : constant array (Token_Class_Mulop) of Node_Kind :=
2716 (Tok_Asterisk => N_Op_Multiply,
2717 Tok_Mod => N_Op_Mod,
2718 Tok_Rem => N_Op_Rem,
2719 Tok_Slash => N_Op_Divide);
2720 begin
2721 return Mulop_Node (Token);
2722 end P_Multiplying_Operator;
2724 --------------------------------------
2725 -- 4.5 Highest Precedence Operator --
2726 --------------------------------------
2728 -- Parsed by P_Factor (4.4)
2730 -- Note: this rule is not in fact used by the grammar at any point!
2732 --------------------------
2733 -- 4.6 Type Conversion --
2734 --------------------------
2736 -- Parsed by P_Primary as a Name (4.1)
2738 -------------------------------
2739 -- 4.7 Qualified Expression --
2740 -------------------------------
2742 -- QUALIFIED_EXPRESSION ::=
2743 -- SUBTYPE_MARK ' (EXPRESSION) | SUBTYPE_MARK ' AGGREGATE
2745 -- The caller has scanned the name which is the Subtype_Mark parameter
2746 -- and scanned past the single quote following the subtype mark. The
2747 -- caller has not checked that this name is in fact appropriate for
2748 -- a subtype mark name (i.e. it is a selected component or identifier).
2750 -- Error_Recovery: cannot raise Error_Resync
2752 function P_Qualified_Expression (Subtype_Mark : Node_Id) return Node_Id is
2753 Qual_Node : Node_Id;
2754 begin
2755 Qual_Node := New_Node (N_Qualified_Expression, Prev_Token_Ptr);
2756 Set_Subtype_Mark (Qual_Node, Check_Subtype_Mark (Subtype_Mark));
2757 Set_Expression (Qual_Node, P_Aggregate_Or_Paren_Expr);
2758 return Qual_Node;
2759 end P_Qualified_Expression;
2761 --------------------
2762 -- 4.8 Allocator --
2763 --------------------
2765 -- ALLOCATOR ::=
2766 -- new [NULL_EXCLUSION] SUBTYPE_INDICATION | new QUALIFIED_EXPRESSION
2768 -- The caller has checked that the initial token is NEW
2770 -- Error recovery: can raise Error_Resync
2772 function P_Allocator return Node_Id is
2773 Alloc_Node : Node_Id;
2774 Type_Node : Node_Id;
2775 Null_Exclusion_Present : Boolean;
2777 begin
2778 Alloc_Node := New_Node (N_Allocator, Token_Ptr);
2779 T_New;
2781 -- Scan Null_Exclusion if present (Ada 2005 (AI-231))
2783 Null_Exclusion_Present := P_Null_Exclusion;
2784 Set_Null_Exclusion_Present (Alloc_Node, Null_Exclusion_Present);
2785 Type_Node := P_Subtype_Mark_Resync;
2787 if Token = Tok_Apostrophe then
2788 Scan; -- past apostrophe
2789 Set_Expression (Alloc_Node, P_Qualified_Expression (Type_Node));
2790 else
2791 Set_Expression
2792 (Alloc_Node,
2793 P_Subtype_Indication (Type_Node, Null_Exclusion_Present));
2794 end if;
2796 return Alloc_Node;
2797 end P_Allocator;
2799 -----------------------
2800 -- P_Case_Expression --
2801 -----------------------
2803 function P_Case_Expression return Node_Id is
2804 Loc : constant Source_Ptr := Token_Ptr;
2805 Case_Node : Node_Id;
2806 Save_State : Saved_Scan_State;
2808 begin
2809 if Ada_Version < Ada_2012 then
2810 Error_Msg_SC ("|case expression is an Ada 2012 feature");
2811 Error_Msg_SC ("\|unit must be compiled with -gnat2012 switch");
2812 end if;
2814 Scan; -- past CASE
2815 Case_Node :=
2816 Make_Case_Expression (Loc,
2817 Expression => P_Expression_No_Right_Paren,
2818 Alternatives => New_List);
2819 T_Is;
2821 -- We now have scanned out CASE expression IS, scan alternatives
2823 loop
2824 T_When;
2825 Append_To (Alternatives (Case_Node), P_Case_Expression_Alternative);
2827 -- Missing comma if WHEN (more alternatives present)
2829 if Token = Tok_When then
2830 T_Comma;
2832 -- If comma/WHEN, skip comma and we have another alternative
2834 elsif Token = Tok_Comma then
2835 Save_Scan_State (Save_State);
2836 Scan; -- past comma
2838 if Token /= Tok_When then
2839 Restore_Scan_State (Save_State);
2840 exit;
2841 end if;
2843 -- If no comma or WHEN, definitely done
2845 else
2846 exit;
2847 end if;
2848 end loop;
2850 -- If we have an END CASE, diagnose as not needed
2852 if Token = Tok_End then
2853 Error_Msg_SC ("`END CASE` not allowed at end of case expression");
2854 Scan; -- past END
2856 if Token = Tok_Case then
2857 Scan; -- past CASE;
2858 end if;
2859 end if;
2861 -- Return the Case_Expression node
2863 return Case_Node;
2864 end P_Case_Expression;
2866 -----------------------------------
2867 -- P_Case_Expression_Alternative --
2868 -----------------------------------
2870 -- CASE_STATEMENT_ALTERNATIVE ::=
2871 -- when DISCRETE_CHOICE_LIST =>
2872 -- EXPRESSION
2874 -- The caller has checked that and scanned past the initial WHEN token
2875 -- Error recovery: can raise Error_Resync
2877 function P_Case_Expression_Alternative return Node_Id is
2878 Case_Alt_Node : Node_Id;
2879 begin
2880 Case_Alt_Node := New_Node (N_Case_Expression_Alternative, Token_Ptr);
2881 Set_Discrete_Choices (Case_Alt_Node, P_Discrete_Choice_List);
2882 TF_Arrow;
2883 Set_Expression (Case_Alt_Node, P_Expression);
2884 return Case_Alt_Node;
2885 end P_Case_Expression_Alternative;
2887 ------------------------------
2888 -- P_Conditional_Expression --
2889 ------------------------------
2891 function P_Conditional_Expression return Node_Id is
2892 Exprs : constant List_Id := New_List;
2893 Loc : constant Source_Ptr := Token_Ptr;
2894 Expr : Node_Id;
2895 State : Saved_Scan_State;
2897 begin
2898 Inside_Conditional_Expression := Inside_Conditional_Expression + 1;
2900 if Token = Tok_If and then Ada_Version < Ada_2012 then
2901 Error_Msg_SC ("|conditional expression is an Ada 2012 feature");
2902 Error_Msg_SC ("\|unit must be compiled with -gnat2012 switch");
2903 end if;
2905 Scan; -- past IF or ELSIF
2906 Append_To (Exprs, P_Condition);
2907 TF_Then;
2908 Append_To (Exprs, P_Expression);
2910 -- We now have scanned out IF expr THEN expr
2912 -- Check for common error of semicolon before the ELSE
2914 if Token = Tok_Semicolon then
2915 Save_Scan_State (State);
2916 Scan; -- past semicolon
2918 if Token = Tok_Else or else Token = Tok_Elsif then
2919 Error_Msg_SP -- CODEFIX
2920 ("|extra "";"" ignored");
2922 else
2923 Restore_Scan_State (State);
2924 end if;
2925 end if;
2927 -- Scan out ELSIF sequence if present
2929 if Token = Tok_Elsif then
2930 Expr := P_Conditional_Expression;
2931 Set_Is_Elsif (Expr);
2932 Append_To (Exprs, Expr);
2934 -- Scan out ELSE phrase if present
2936 elsif Token = Tok_Else then
2938 -- Scan out ELSE expression
2940 Scan; -- Past ELSE
2941 Append_To (Exprs, P_Expression);
2943 -- Two expression case (implied True, filled in during semantics)
2945 else
2946 null;
2947 end if;
2949 -- If we have an END IF, diagnose as not needed
2951 if Token = Tok_End then
2952 Error_Msg_SC
2953 ("`END IF` not allowed at end of conditional expression");
2954 Scan; -- past END
2956 if Token = Tok_If then
2957 Scan; -- past IF;
2958 end if;
2959 end if;
2961 Inside_Conditional_Expression := Inside_Conditional_Expression - 1;
2963 -- Return the Conditional_Expression node
2965 return
2966 Make_Conditional_Expression (Loc,
2967 Expressions => Exprs);
2968 end P_Conditional_Expression;
2970 -----------------------
2971 -- P_Membership_Test --
2972 -----------------------
2974 -- MEMBERSHIP_CHOICE_LIST ::= MEMBERHIP_CHOICE {'|' MEMBERSHIP_CHOICE}
2975 -- MEMBERSHIP_CHOICE ::= CHOICE_EXPRESSION | range | subtype_mark
2977 procedure P_Membership_Test (N : Node_Id) is
2978 Alt : constant Node_Id :=
2979 P_Range_Or_Subtype_Mark
2980 (Allow_Simple_Expression => (Ada_Version >= Ada_2012));
2982 begin
2983 -- Set case
2985 if Token = Tok_Vertical_Bar then
2986 if Ada_Version < Ada_2012 then
2987 Error_Msg_SC ("set notation is an Ada 2012 feature");
2988 Error_Msg_SC ("\|unit must be compiled with -gnat2012 switch");
2989 end if;
2991 Set_Alternatives (N, New_List (Alt));
2992 Set_Right_Opnd (N, Empty);
2994 -- Loop to accumulate alternatives
2996 while Token = Tok_Vertical_Bar loop
2997 Scan; -- past vertical bar
2998 Append_To
2999 (Alternatives (N),
3000 P_Range_Or_Subtype_Mark (Allow_Simple_Expression => True));
3001 end loop;
3003 -- Not set case
3005 else
3006 Set_Right_Opnd (N, Alt);
3007 Set_Alternatives (N, No_List);
3008 end if;
3009 end P_Membership_Test;
3011 end Ch4;