[NDS32] new attribute no_prologue and new option -mret-in-naked-func.
[official-gcc.git] / gcc / ada / par-ch4.adb
blobd0b46740ada8c3f0836dd2902a93633fe7949044
1 -----------------------------------------------------------------------------
2 -- --
3 -- GNAT COMPILER COMPONENTS --
4 -- --
5 -- P A R . C H 4 --
6 -- --
7 -- B o d y --
8 -- --
9 -- Copyright (C) 1992-2018, 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_Base => True,
39 Attribute_Body_Version => True,
40 Attribute_Class => True,
41 Attribute_External_Tag => True,
42 Attribute_Img => True,
43 Attribute_Loop_Entry => True,
44 Attribute_Old => True,
45 Attribute_Result => True,
46 Attribute_Stub_Type => True,
47 Attribute_Version => True,
48 Attribute_Type_Key => True,
49 others => False);
50 -- This map contains True for parameterless attributes that return a string
51 -- or a type. For those attributes, a left parenthesis after the attribute
52 -- should not be analyzed as the beginning of a parameters list because it
53 -- may denote a slice operation (X'Img (1 .. 2)) or a type conversion
54 -- (X'Class (Y)). The Ada 2012 attribute 'Old is in this category.
56 -- Note: Loop_Entry is in this list because, although it can take an
57 -- optional argument (the loop name), we can't distinguish that at parse
58 -- time from the case where no loop name is given and a legitimate index
59 -- expression is present. So we parse the argument as an indexed component
60 -- and the semantic analysis sorts out this syntactic ambiguity based on
61 -- the type and form of the expression.
63 -- Note that this map designates the minimum set of attributes where a
64 -- construct in parentheses that is not an argument can appear right
65 -- after the attribute. For attributes like 'Size, we do not put them
66 -- in the map. If someone writes X'Size (3), that's illegal in any case,
67 -- but we get a better error message by parsing the (3) as an illegal
68 -- argument to the attribute, rather than some meaningless junk that
69 -- follows the attribute.
71 -----------------------
72 -- Local Subprograms --
73 -----------------------
75 function P_Aggregate_Or_Paren_Expr return Node_Id;
76 function P_Allocator return Node_Id;
77 function P_Case_Expression_Alternative return Node_Id;
78 function P_Iterated_Component_Association return Node_Id;
79 function P_Record_Or_Array_Component_Association return Node_Id;
80 function P_Factor return Node_Id;
81 function P_Primary return Node_Id;
82 function P_Relation return Node_Id;
83 function P_Term return Node_Id;
85 function P_Binary_Adding_Operator return Node_Kind;
86 function P_Logical_Operator return Node_Kind;
87 function P_Multiplying_Operator return Node_Kind;
88 function P_Relational_Operator return Node_Kind;
89 function P_Unary_Adding_Operator return Node_Kind;
91 procedure Bad_Range_Attribute (Loc : Source_Ptr);
92 -- Called to place complaint about bad range attribute at the given
93 -- source location. Terminates by raising Error_Resync.
95 procedure Check_Bad_Exp;
96 -- Called after scanning a**b, posts error if ** detected
98 procedure P_Membership_Test (N : Node_Id);
99 -- N is the node for a N_In or N_Not_In node whose right operand has not
100 -- yet been processed. It is called just after scanning out the IN keyword.
101 -- On return, either Right_Opnd or Alternatives is set, as appropriate.
103 function P_Range_Attribute_Reference (Prefix_Node : Node_Id) return Node_Id;
104 -- Scan a range attribute reference. The caller has scanned out the
105 -- prefix. The current token is known to be an apostrophe and the
106 -- following token is known to be RANGE.
108 function P_Unparen_Cond_Case_Quant_Expression return Node_Id;
109 -- This function is called with Token pointing to IF, CASE, or FOR, in a
110 -- context that allows a case, conditional, or quantified expression if
111 -- it is surrounded by parentheses. If not surrounded by parentheses, the
112 -- expression is still returned, but an error message is issued.
114 -------------------------
115 -- Bad_Range_Attribute --
116 -------------------------
118 procedure Bad_Range_Attribute (Loc : Source_Ptr) is
119 begin
120 Error_Msg ("range attribute cannot be used in expression!", Loc);
121 Resync_Expression;
122 end Bad_Range_Attribute;
124 -------------------
125 -- Check_Bad_Exp --
126 -------------------
128 procedure Check_Bad_Exp is
129 begin
130 if Token = Tok_Double_Asterisk then
131 Error_Msg_SC ("parenthesization required for '*'*");
132 Scan; -- past **
133 Discard_Junk_Node (P_Primary);
134 Check_Bad_Exp;
135 end if;
136 end Check_Bad_Exp;
138 --------------------------
139 -- 4.1 Name (also 6.4) --
140 --------------------------
142 -- NAME ::=
143 -- DIRECT_NAME | EXPLICIT_DEREFERENCE
144 -- | INDEXED_COMPONENT | SLICE
145 -- | SELECTED_COMPONENT | ATTRIBUTE
146 -- | TYPE_CONVERSION | FUNCTION_CALL
147 -- | CHARACTER_LITERAL | TARGET_NAME
149 -- DIRECT_NAME ::= IDENTIFIER | OPERATOR_SYMBOL
151 -- PREFIX ::= NAME | IMPLICIT_DEREFERENCE
153 -- EXPLICIT_DEREFERENCE ::= NAME . all
155 -- IMPLICIT_DEREFERENCE ::= NAME
157 -- INDEXED_COMPONENT ::= PREFIX (EXPRESSION {, EXPRESSION})
159 -- SLICE ::= PREFIX (DISCRETE_RANGE)
161 -- SELECTED_COMPONENT ::= PREFIX . SELECTOR_NAME
163 -- SELECTOR_NAME ::= IDENTIFIER | CHARACTER_LITERAL | OPERATOR_SYMBOL
165 -- ATTRIBUTE_REFERENCE ::= PREFIX ' ATTRIBUTE_DESIGNATOR
167 -- ATTRIBUTE_DESIGNATOR ::=
168 -- IDENTIFIER [(static_EXPRESSION)]
169 -- | access | delta | digits
171 -- FUNCTION_CALL ::=
172 -- function_NAME
173 -- | function_PREFIX ACTUAL_PARAMETER_PART
175 -- ACTUAL_PARAMETER_PART ::=
176 -- (PARAMETER_ASSOCIATION {,PARAMETER_ASSOCIATION})
178 -- PARAMETER_ASSOCIATION ::=
179 -- [formal_parameter_SELECTOR_NAME =>] EXPLICIT_ACTUAL_PARAMETER
181 -- EXPLICIT_ACTUAL_PARAMETER ::= EXPRESSION | variable_NAME
183 -- TARGET_NAME ::= @ (AI12-0125-3: abbreviation for LHS)
185 -- Note: syntactically a procedure call looks just like a function call,
186 -- so this routine is in practice used to scan out procedure calls as well.
188 -- On return, Expr_Form is set to either EF_Name or EF_Simple_Name
190 -- Error recovery: can raise Error_Resync
192 -- Note: if on return Token = Tok_Apostrophe, then the apostrophe must be
193 -- followed by either a left paren (qualified expression case), or by
194 -- range (range attribute case). All other uses of apostrophe (i.e. all
195 -- other attributes) are handled in this routine.
197 -- Error recovery: can raise Error_Resync
199 function P_Name return Node_Id is
200 Scan_State : Saved_Scan_State;
201 Name_Node : Node_Id;
202 Prefix_Node : Node_Id;
203 Ident_Node : Node_Id;
204 Expr_Node : Node_Id;
205 Range_Node : Node_Id;
206 Arg_Node : Node_Id;
208 Arg_List : List_Id := No_List; -- kill junk warning
209 Attr_Name : Name_Id := No_Name; -- kill junk warning
211 begin
212 -- Case of not a name
214 if Token not in Token_Class_Name then
216 -- If it looks like start of expression, complain and scan expression
218 if Token in Token_Class_Literal
219 or else Token = Tok_Left_Paren
220 then
221 Error_Msg_SC ("name expected");
222 return P_Expression;
224 -- Otherwise some other junk, not much we can do
226 else
227 Error_Msg_AP ("name expected");
228 raise Error_Resync;
229 end if;
230 end if;
232 -- Loop through designators in qualified name
233 -- AI12-0125 : target_name
235 if Token = Tok_At_Sign then
236 Scan_Reserved_Identifier (Force_Msg => False);
238 if Present (Current_Assign_Node) then
239 Set_Has_Target_Names (Current_Assign_Node);
240 end if;
241 end if;
243 Name_Node := Token_Node;
245 loop
246 Scan; -- past designator
247 exit when Token /= Tok_Dot;
248 Save_Scan_State (Scan_State); -- at dot
249 Scan; -- past dot
251 -- If we do not have another designator after the dot, then join
252 -- the normal circuit to handle a dot extension (may be .all or
253 -- character literal case). Otherwise loop back to scan the next
254 -- designator.
256 if Token not in Token_Class_Desig then
257 goto Scan_Name_Extension_Dot;
258 else
259 Prefix_Node := Name_Node;
260 Name_Node := New_Node (N_Selected_Component, Prev_Token_Ptr);
261 Set_Prefix (Name_Node, Prefix_Node);
262 Set_Selector_Name (Name_Node, Token_Node);
263 end if;
264 end loop;
266 -- We have now scanned out a qualified designator. If the last token is
267 -- an operator symbol, then we certainly do not have the Snam case, so
268 -- we can just use the normal name extension check circuit
270 if Prev_Token = Tok_Operator_Symbol then
271 goto Scan_Name_Extension;
272 end if;
274 -- We have scanned out a qualified simple name, check for name extension
275 -- Note that we know there is no dot here at this stage, so the only
276 -- possible cases of name extension are apostrophe and left paren.
278 if Token = Tok_Apostrophe then
279 Save_Scan_State (Scan_State); -- at apostrophe
280 Scan; -- past apostrophe
282 -- Qualified expression in Ada 2012 mode (treated as a name)
284 if Ada_Version >= Ada_2012 and then Token = Tok_Left_Paren then
285 goto Scan_Name_Extension_Apostrophe;
287 -- If left paren not in Ada 2012, then it is not part of the name,
288 -- since qualified expressions are not names in prior versions of
289 -- Ada, so return with Token backed up to point to the apostrophe.
290 -- The treatment for the range attribute is similar (we do not
291 -- consider x'range to be a name in this grammar).
293 elsif Token = Tok_Left_Paren or else Token = Tok_Range then
294 Restore_Scan_State (Scan_State); -- to apostrophe
295 Expr_Form := EF_Simple_Name;
296 return Name_Node;
298 -- Otherwise we have the case of a name extended by an attribute
300 else
301 goto Scan_Name_Extension_Apostrophe;
302 end if;
304 -- Check case of qualified simple name extended by a left parenthesis
306 elsif Token = Tok_Left_Paren then
307 Scan; -- past left paren
308 goto Scan_Name_Extension_Left_Paren;
310 -- Otherwise the qualified simple name is not extended, so return
312 else
313 Expr_Form := EF_Simple_Name;
314 return Name_Node;
315 end if;
317 -- Loop scanning past name extensions. A label is used for control
318 -- transfer for this loop for ease of interfacing with the finite state
319 -- machine in the parenthesis scanning circuit, and also to allow for
320 -- passing in control to the appropriate point from the above code.
322 <<Scan_Name_Extension>>
324 -- Character literal used as name cannot be extended. Also this
325 -- cannot be a call, since the name for a call must be a designator.
326 -- Return in these cases, or if there is no name extension
328 if Token not in Token_Class_Namext
329 or else Prev_Token = Tok_Char_Literal
330 then
331 Expr_Form := EF_Name;
332 return Name_Node;
333 end if;
335 -- Merge here when we know there is a name extension
337 <<Scan_Name_Extension_OK>>
339 if Token = Tok_Left_Paren then
340 Scan; -- past left paren
341 goto Scan_Name_Extension_Left_Paren;
343 elsif Token = Tok_Apostrophe then
344 Save_Scan_State (Scan_State); -- at apostrophe
345 Scan; -- past apostrophe
346 goto Scan_Name_Extension_Apostrophe;
348 else -- Token = Tok_Dot
349 Save_Scan_State (Scan_State); -- at dot
350 Scan; -- past dot
351 goto Scan_Name_Extension_Dot;
352 end if;
354 -- Case of name extended by dot (selection), dot is already skipped
355 -- and the scan state at the point of the dot is saved in Scan_State.
357 <<Scan_Name_Extension_Dot>>
359 -- Explicit dereference case
361 if Token = Tok_All then
362 Prefix_Node := Name_Node;
363 Name_Node := New_Node (N_Explicit_Dereference, Token_Ptr);
364 Set_Prefix (Name_Node, Prefix_Node);
365 Scan; -- past ALL
366 goto Scan_Name_Extension;
368 -- Selected component case
370 elsif Token in Token_Class_Name then
371 Prefix_Node := Name_Node;
372 Name_Node := New_Node (N_Selected_Component, Prev_Token_Ptr);
373 Set_Prefix (Name_Node, Prefix_Node);
374 Set_Selector_Name (Name_Node, Token_Node);
375 Scan; -- past selector
376 goto Scan_Name_Extension;
378 -- Reserved identifier as selector
380 elsif Is_Reserved_Identifier then
381 Scan_Reserved_Identifier (Force_Msg => False);
382 Prefix_Node := Name_Node;
383 Name_Node := New_Node (N_Selected_Component, Prev_Token_Ptr);
384 Set_Prefix (Name_Node, Prefix_Node);
385 Set_Selector_Name (Name_Node, Token_Node);
386 Scan; -- past identifier used as selector
387 goto Scan_Name_Extension;
389 -- If dot is at end of line and followed by nothing legal,
390 -- then assume end of name and quit (dot will be taken as
391 -- an incorrect form of some other punctuation by our caller).
393 elsif Token_Is_At_Start_Of_Line then
394 Restore_Scan_State (Scan_State);
395 return Name_Node;
397 -- Here if nothing legal after the dot
399 else
400 Error_Msg_AP ("selector expected");
401 raise Error_Resync;
402 end if;
404 -- Here for an apostrophe as name extension. The scan position at the
405 -- apostrophe has already been saved, and the apostrophe scanned out.
407 <<Scan_Name_Extension_Apostrophe>>
409 Scan_Apostrophe : declare
410 function Apostrophe_Should_Be_Semicolon return Boolean;
411 -- Checks for case where apostrophe should probably be
412 -- a semicolon, and if so, gives appropriate message,
413 -- resets the scan pointer to the apostrophe, changes
414 -- the current token to Tok_Semicolon, and returns True.
415 -- Otherwise returns False.
417 ------------------------------------
418 -- Apostrophe_Should_Be_Semicolon --
419 ------------------------------------
421 function Apostrophe_Should_Be_Semicolon return Boolean is
422 begin
423 if Token_Is_At_Start_Of_Line then
424 Restore_Scan_State (Scan_State); -- to apostrophe
425 Error_Msg_SC ("|""''"" should be "";""");
426 Token := Tok_Semicolon;
427 return True;
428 else
429 return False;
430 end if;
431 end Apostrophe_Should_Be_Semicolon;
433 -- Start of processing for Scan_Apostrophe
435 begin
436 -- Check for qualified expression case in Ada 2012 mode
438 if Ada_Version >= Ada_2012 and then Token = Tok_Left_Paren then
439 Name_Node := P_Qualified_Expression (Name_Node);
440 goto Scan_Name_Extension;
442 -- If range attribute after apostrophe, then return with Token
443 -- pointing to the apostrophe. Note that in this case the prefix
444 -- need not be a simple name (cases like A.all'range). Similarly
445 -- if there is a left paren after the apostrophe, then we also
446 -- return with Token pointing to the apostrophe (this is the
447 -- aggregate case, or some error case).
449 elsif Token = Tok_Range or else Token = Tok_Left_Paren then
450 Restore_Scan_State (Scan_State); -- to apostrophe
451 Expr_Form := EF_Name;
452 return Name_Node;
454 -- Here for cases where attribute designator is an identifier
456 elsif Token = Tok_Identifier then
457 Attr_Name := Token_Name;
459 if not Is_Attribute_Name (Attr_Name) then
460 if Apostrophe_Should_Be_Semicolon then
461 Expr_Form := EF_Name;
462 return Name_Node;
464 -- Here for a bad attribute name
466 else
467 Signal_Bad_Attribute;
468 Scan; -- past bad identifier
470 if Token = Tok_Left_Paren then
471 Scan; -- past left paren
473 loop
474 Discard_Junk_Node (P_Expression_If_OK);
475 exit when not Comma_Present;
476 end loop;
478 T_Right_Paren;
479 end if;
481 return Error;
482 end if;
483 end if;
485 if Style_Check then
486 Style.Check_Attribute_Name (False);
487 end if;
489 -- Here for case of attribute designator is not an identifier
491 else
492 if Token = Tok_Delta then
493 Attr_Name := Name_Delta;
495 elsif Token = Tok_Digits then
496 Attr_Name := Name_Digits;
498 elsif Token = Tok_Access then
499 Attr_Name := Name_Access;
501 elsif Token = Tok_Mod and then Ada_Version >= Ada_95 then
502 Attr_Name := Name_Mod;
504 elsif Apostrophe_Should_Be_Semicolon then
505 Expr_Form := EF_Name;
506 return Name_Node;
508 else
509 Error_Msg_AP ("attribute designator expected");
510 raise Error_Resync;
511 end if;
513 if Style_Check then
514 Style.Check_Attribute_Name (True);
515 end if;
516 end if;
518 -- We come here with an OK attribute scanned, and corresponding
519 -- Attribute identifier node stored in Ident_Node.
521 Prefix_Node := Name_Node;
522 Name_Node := New_Node (N_Attribute_Reference, Prev_Token_Ptr);
523 Scan; -- past attribute designator
524 Set_Prefix (Name_Node, Prefix_Node);
525 Set_Attribute_Name (Name_Node, Attr_Name);
527 -- Scan attribute arguments/designator. We skip this if we know
528 -- that the attribute cannot have an argument (see documentation
529 -- of Is_Parameterless_Attribute for further details).
531 if Token = Tok_Left_Paren
532 and then not
533 Is_Parameterless_Attribute (Get_Attribute_Id (Attr_Name))
534 then
535 -- Attribute Update contains an array or record association
536 -- list which provides new values for various components or
537 -- elements. The list is parsed as an aggregate, and we get
538 -- better error handling by knowing that in the parser.
540 if Attr_Name = Name_Update then
541 Set_Expressions (Name_Node, New_List);
542 Append (P_Aggregate, Expressions (Name_Node));
544 -- All other cases of parsing attribute arguments
546 else
547 Set_Expressions (Name_Node, New_List);
548 Scan; -- past left paren
550 loop
551 declare
552 Expr : constant Node_Id := P_Expression_If_OK;
553 Rnam : Node_Id;
555 begin
556 -- Case of => for named notation
558 if Token = Tok_Arrow then
560 -- Named notation allowed only for the special
561 -- case of System'Restriction_Set (No_Dependence =>
562 -- unit_NAME), in which case construct a parameter
563 -- assocation node and append to the arguments.
565 if Attr_Name = Name_Restriction_Set
566 and then Nkind (Expr) = N_Identifier
567 and then Chars (Expr) = Name_No_Dependence
568 then
569 Scan; -- past arrow
570 Rnam := P_Name;
571 Append_To (Expressions (Name_Node),
572 Make_Parameter_Association (Sloc (Rnam),
573 Selector_Name => Expr,
574 Explicit_Actual_Parameter => Rnam));
575 exit;
577 -- For all other cases named notation is illegal
579 else
580 Error_Msg_SC
581 ("named parameters not permitted "
582 & "for attributes");
583 Scan; -- past junk arrow
584 end if;
586 -- Here for normal case (not => for named parameter)
588 else
589 -- Special handling for 'Image in Ada 2012, where
590 -- the attribute can be parameterless and its value
591 -- can be the prefix of a slice. Rewrite name as a
592 -- slice, Expr is its low bound.
594 if Token = Tok_Dot_Dot
595 and then Attr_Name = Name_Image
596 and then Ada_Version >= Ada_2012
597 then
598 Set_Expressions (Name_Node, No_List);
599 Prefix_Node := Name_Node;
600 Name_Node :=
601 New_Node (N_Slice, Sloc (Prefix_Node));
602 Set_Prefix (Name_Node, Prefix_Node);
603 Range_Node := New_Node (N_Range, Token_Ptr);
604 Set_Low_Bound (Range_Node, Expr);
605 Scan; -- past ..
606 Expr_Node := P_Expression;
607 Check_Simple_Expression (Expr_Node);
608 Set_High_Bound (Range_Node, Expr_Node);
609 Set_Discrete_Range (Name_Node, Range_Node);
610 T_Right_Paren;
612 goto Scan_Name_Extension;
614 else
615 Append (Expr, Expressions (Name_Node));
616 exit when not Comma_Present;
617 end if;
618 end if;
619 end;
620 end loop;
622 T_Right_Paren;
623 end if;
624 end if;
626 goto Scan_Name_Extension;
627 end Scan_Apostrophe;
629 -- Here for left parenthesis extending name (left paren skipped)
631 <<Scan_Name_Extension_Left_Paren>>
633 -- We now have to scan through a list of items, terminated by a
634 -- right parenthesis. The scan is handled by a finite state
635 -- machine. The possibilities are:
637 -- (discrete_range)
639 -- This is a slice. This case is handled in LP_State_Init
641 -- (expression, expression, ..)
643 -- This is interpreted as an indexed component, i.e. as a
644 -- case of a name which can be extended in the normal manner.
645 -- This case is handled by LP_State_Name or LP_State_Expr.
647 -- Note: if and case expressions (without an extra level of
648 -- parentheses) are permitted in this context).
650 -- (..., identifier => expression , ...)
652 -- If there is at least one occurrence of identifier => (but
653 -- none of the other cases apply), then we have a call.
655 -- Test for Id => case
657 if Token = Tok_Identifier then
658 Save_Scan_State (Scan_State); -- at Id
659 Scan; -- past Id
661 -- Test for => (allow := as an error substitute)
663 if Token = Tok_Arrow or else Token = Tok_Colon_Equal then
664 Restore_Scan_State (Scan_State); -- to Id
665 Arg_List := New_List;
666 goto LP_State_Call;
668 else
669 Restore_Scan_State (Scan_State); -- to Id
670 end if;
671 end if;
673 -- Here we have an expression after all
675 Expr_Node := P_Expression_Or_Range_Attribute_If_OK;
677 -- Check cases of discrete range for a slice
679 -- First possibility: Range_Attribute_Reference
681 if Expr_Form = EF_Range_Attr then
682 Range_Node := Expr_Node;
684 -- Second possibility: Simple_expression .. Simple_expression
686 elsif Token = Tok_Dot_Dot then
687 Check_Simple_Expression (Expr_Node);
688 Range_Node := New_Node (N_Range, Token_Ptr);
689 Set_Low_Bound (Range_Node, Expr_Node);
690 Scan; -- past ..
691 Expr_Node := P_Expression;
692 Check_Simple_Expression (Expr_Node);
693 Set_High_Bound (Range_Node, Expr_Node);
695 -- Third possibility: Type_name range Range
697 elsif Token = Tok_Range then
698 if Expr_Form /= EF_Simple_Name then
699 Error_Msg_SC ("subtype mark must precede RANGE");
700 raise Error_Resync;
701 end if;
703 Range_Node := P_Subtype_Indication (Expr_Node);
705 -- Otherwise we just have an expression. It is true that we might
706 -- have a subtype mark without a range constraint but this case
707 -- is syntactically indistinguishable from the expression case.
709 else
710 Arg_List := New_List;
711 goto LP_State_Expr;
712 end if;
714 -- Fall through here with unmistakable Discrete range scanned,
715 -- which means that we definitely have the case of a slice. The
716 -- Discrete range is in Range_Node.
718 if Token = Tok_Comma then
719 Error_Msg_SC ("slice cannot have more than one dimension");
720 raise Error_Resync;
722 elsif Token /= Tok_Right_Paren then
723 if Token = Tok_Arrow then
725 -- This may be an aggregate that is missing a qualification
727 Error_Msg_SC
728 ("context of aggregate must be a qualified expression");
729 raise Error_Resync;
731 else
732 T_Right_Paren;
733 raise Error_Resync;
734 end if;
736 else
737 Scan; -- past right paren
738 Prefix_Node := Name_Node;
739 Name_Node := New_Node (N_Slice, Sloc (Prefix_Node));
740 Set_Prefix (Name_Node, Prefix_Node);
741 Set_Discrete_Range (Name_Node, Range_Node);
743 -- An operator node is legal as a prefix to other names,
744 -- but not for a slice.
746 if Nkind (Prefix_Node) = N_Operator_Symbol then
747 Error_Msg_N ("illegal prefix for slice", Prefix_Node);
748 end if;
750 -- If we have a name extension, go scan it
752 if Token in Token_Class_Namext then
753 goto Scan_Name_Extension_OK;
755 -- Otherwise return (a slice is a name, but is not a call)
757 else
758 Expr_Form := EF_Name;
759 return Name_Node;
760 end if;
761 end if;
763 -- In LP_State_Expr, we have scanned one or more expressions, and
764 -- so we have a call or an indexed component which is a name. On
765 -- entry we have the expression just scanned in Expr_Node and
766 -- Arg_List contains the list of expressions encountered so far
768 <<LP_State_Expr>>
769 Append (Expr_Node, Arg_List);
771 if Token = Tok_Arrow then
772 Error_Msg
773 ("expect identifier in parameter association", Sloc (Expr_Node));
774 Scan; -- past arrow
776 elsif not Comma_Present then
777 T_Right_Paren;
779 Prefix_Node := Name_Node;
780 Name_Node := New_Node (N_Indexed_Component, Sloc (Prefix_Node));
781 Set_Prefix (Name_Node, Prefix_Node);
782 Set_Expressions (Name_Node, Arg_List);
784 goto Scan_Name_Extension;
785 end if;
787 -- Comma present (and scanned out), test for identifier => case
788 -- Test for identifier => case
790 if Token = Tok_Identifier then
791 Save_Scan_State (Scan_State); -- at Id
792 Scan; -- past Id
794 -- Test for => (allow := as error substitute)
796 if Token = Tok_Arrow or else Token = Tok_Colon_Equal then
797 Restore_Scan_State (Scan_State); -- to Id
798 goto LP_State_Call;
800 -- Otherwise it's just an expression after all, so backup
802 else
803 Restore_Scan_State (Scan_State); -- to Id
804 end if;
805 end if;
807 -- Here we have an expression after all, so stay in this state
809 Expr_Node := P_Expression_If_OK;
810 goto LP_State_Expr;
812 -- LP_State_Call corresponds to the situation in which at least one
813 -- instance of Id => Expression has been encountered, so we know that
814 -- we do not have a name, but rather a call. We enter it with the
815 -- scan pointer pointing to the next argument to scan, and Arg_List
816 -- containing the list of arguments scanned so far.
818 <<LP_State_Call>>
820 -- Test for case of Id => Expression (named parameter)
822 if Token = Tok_Identifier then
823 Save_Scan_State (Scan_State); -- at Id
824 Ident_Node := Token_Node;
825 Scan; -- past Id
827 -- Deal with => (allow := as incorrect substitute)
829 if Token = Tok_Arrow or else Token = Tok_Colon_Equal then
830 Arg_Node := New_Node (N_Parameter_Association, Prev_Token_Ptr);
831 Set_Selector_Name (Arg_Node, Ident_Node);
832 T_Arrow;
833 Set_Explicit_Actual_Parameter (Arg_Node, P_Expression);
834 Append (Arg_Node, Arg_List);
836 -- If a comma follows, go back and scan next entry
838 if Comma_Present then
839 goto LP_State_Call;
841 -- Otherwise we have the end of a call
843 else
844 Prefix_Node := Name_Node;
845 Name_Node := New_Node (N_Function_Call, Sloc (Prefix_Node));
846 Set_Name (Name_Node, Prefix_Node);
847 Set_Parameter_Associations (Name_Node, Arg_List);
848 T_Right_Paren;
850 if Token in Token_Class_Namext then
851 goto Scan_Name_Extension_OK;
853 -- This is a case of a call which cannot be a name
855 else
856 Expr_Form := EF_Name;
857 return Name_Node;
858 end if;
859 end if;
861 -- Not named parameter: Id started an expression after all
863 else
864 Restore_Scan_State (Scan_State); -- to Id
865 end if;
866 end if;
868 -- Here if entry did not start with Id => which means that it
869 -- is a positional parameter, which is not allowed, since we
870 -- have seen at least one named parameter already.
872 Error_Msg_SC
873 ("positional parameter association " &
874 "not allowed after named one");
876 Expr_Node := P_Expression_If_OK;
878 -- Leaving the '>' in an association is not unusual, so suggest
879 -- a possible fix.
881 if Nkind (Expr_Node) = N_Op_Eq then
882 Error_Msg_N ("\maybe `='>` was intended", Expr_Node);
883 end if;
885 -- We go back to scanning out expressions, so that we do not get
886 -- multiple error messages when several positional parameters
887 -- follow a named parameter.
889 goto LP_State_Expr;
891 -- End of treatment for name extensions starting with left paren
893 -- End of loop through name extensions
895 end P_Name;
897 -- This function parses a restricted form of Names which are either
898 -- designators, or designators preceded by a sequence of prefixes
899 -- that are direct names.
901 -- Error recovery: cannot raise Error_Resync
903 function P_Function_Name return Node_Id is
904 Designator_Node : Node_Id;
905 Prefix_Node : Node_Id;
906 Selector_Node : Node_Id;
907 Dot_Sloc : Source_Ptr := No_Location;
909 begin
910 -- Prefix_Node is set to the gathered prefix so far, Empty means that
911 -- no prefix has been scanned. This allows us to build up the result
912 -- in the required right recursive manner.
914 Prefix_Node := Empty;
916 -- Loop through prefixes
918 loop
919 Designator_Node := Token_Node;
921 if Token not in Token_Class_Desig then
922 return P_Identifier; -- let P_Identifier issue the error message
924 else -- Token in Token_Class_Desig
925 Scan; -- past designator
926 exit when Token /= Tok_Dot;
927 end if;
929 -- Here at a dot, with token just before it in Designator_Node
931 if No (Prefix_Node) then
932 Prefix_Node := Designator_Node;
933 else
934 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
935 Set_Prefix (Selector_Node, Prefix_Node);
936 Set_Selector_Name (Selector_Node, Designator_Node);
937 Prefix_Node := Selector_Node;
938 end if;
940 Dot_Sloc := Token_Ptr;
941 Scan; -- past dot
942 end loop;
944 -- Fall out of the loop having just scanned a designator
946 if No (Prefix_Node) then
947 return Designator_Node;
948 else
949 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
950 Set_Prefix (Selector_Node, Prefix_Node);
951 Set_Selector_Name (Selector_Node, Designator_Node);
952 return Selector_Node;
953 end if;
955 exception
956 when Error_Resync =>
957 return Error;
958 end P_Function_Name;
960 -- This function parses a restricted form of Names which are either
961 -- identifiers, or identifiers preceded by a sequence of prefixes
962 -- that are direct names.
964 -- Error recovery: cannot raise Error_Resync
966 function P_Qualified_Simple_Name return Node_Id is
967 Designator_Node : Node_Id;
968 Prefix_Node : Node_Id;
969 Selector_Node : Node_Id;
970 Dot_Sloc : Source_Ptr := No_Location;
972 begin
973 -- Prefix node is set to the gathered prefix so far, Empty means that
974 -- no prefix has been scanned. This allows us to build up the result
975 -- in the required right recursive manner.
977 Prefix_Node := Empty;
979 -- Loop through prefixes
981 loop
982 Designator_Node := Token_Node;
984 if Token = Tok_Identifier then
985 Scan; -- past identifier
986 exit when Token /= Tok_Dot;
988 elsif Token not in Token_Class_Desig then
989 return P_Identifier; -- let P_Identifier issue the error message
991 else
992 Scan; -- past designator
994 if Token /= Tok_Dot then
995 Error_Msg_SP ("identifier expected");
996 return Error;
997 end if;
998 end if;
1000 -- Here at a dot, with token just before it in Designator_Node
1002 if No (Prefix_Node) then
1003 Prefix_Node := Designator_Node;
1004 else
1005 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
1006 Set_Prefix (Selector_Node, Prefix_Node);
1007 Set_Selector_Name (Selector_Node, Designator_Node);
1008 Prefix_Node := Selector_Node;
1009 end if;
1011 Dot_Sloc := Token_Ptr;
1012 Scan; -- past dot
1013 end loop;
1015 -- Fall out of the loop having just scanned an identifier
1017 if No (Prefix_Node) then
1018 return Designator_Node;
1019 else
1020 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
1021 Set_Prefix (Selector_Node, Prefix_Node);
1022 Set_Selector_Name (Selector_Node, Designator_Node);
1023 return Selector_Node;
1024 end if;
1026 exception
1027 when Error_Resync =>
1028 return Error;
1029 end P_Qualified_Simple_Name;
1031 -- This procedure differs from P_Qualified_Simple_Name only in that it
1032 -- raises Error_Resync if any error is encountered. It only returns after
1033 -- scanning a valid qualified simple name.
1035 -- Error recovery: can raise Error_Resync
1037 function P_Qualified_Simple_Name_Resync return Node_Id is
1038 Designator_Node : Node_Id;
1039 Prefix_Node : Node_Id;
1040 Selector_Node : Node_Id;
1041 Dot_Sloc : Source_Ptr := No_Location;
1043 begin
1044 Prefix_Node := Empty;
1046 -- Loop through prefixes
1048 loop
1049 Designator_Node := Token_Node;
1051 if Token = Tok_Identifier then
1052 Scan; -- past identifier
1053 exit when Token /= Tok_Dot;
1055 elsif Token not in Token_Class_Desig then
1056 Discard_Junk_Node (P_Identifier); -- to issue the error message
1057 raise Error_Resync;
1059 else
1060 Scan; -- past designator
1062 if Token /= Tok_Dot then
1063 Error_Msg_SP ("identifier expected");
1064 raise Error_Resync;
1065 end if;
1066 end if;
1068 -- Here at a dot, with token just before it in Designator_Node
1070 if No (Prefix_Node) then
1071 Prefix_Node := Designator_Node;
1072 else
1073 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
1074 Set_Prefix (Selector_Node, Prefix_Node);
1075 Set_Selector_Name (Selector_Node, Designator_Node);
1076 Prefix_Node := Selector_Node;
1077 end if;
1079 Dot_Sloc := Token_Ptr;
1080 Scan; -- past period
1081 end loop;
1083 -- Fall out of the loop having just scanned an identifier
1085 if No (Prefix_Node) then
1086 return Designator_Node;
1087 else
1088 Selector_Node := New_Node (N_Selected_Component, Dot_Sloc);
1089 Set_Prefix (Selector_Node, Prefix_Node);
1090 Set_Selector_Name (Selector_Node, Designator_Node);
1091 return Selector_Node;
1092 end if;
1093 end P_Qualified_Simple_Name_Resync;
1095 ----------------------
1096 -- 4.1 Direct_Name --
1097 ----------------------
1099 -- Parsed by P_Name and other functions in section 4.1
1101 -----------------
1102 -- 4.1 Prefix --
1103 -----------------
1105 -- Parsed by P_Name (4.1)
1107 -------------------------------
1108 -- 4.1 Explicit Dereference --
1109 -------------------------------
1111 -- Parsed by P_Name (4.1)
1113 -------------------------------
1114 -- 4.1 Implicit_Dereference --
1115 -------------------------------
1117 -- Parsed by P_Name (4.1)
1119 ----------------------------
1120 -- 4.1 Indexed Component --
1121 ----------------------------
1123 -- Parsed by P_Name (4.1)
1125 ----------------
1126 -- 4.1 Slice --
1127 ----------------
1129 -- Parsed by P_Name (4.1)
1131 -----------------------------
1132 -- 4.1 Selected_Component --
1133 -----------------------------
1135 -- Parsed by P_Name (4.1)
1137 ------------------------
1138 -- 4.1 Selector Name --
1139 ------------------------
1141 -- Parsed by P_Name (4.1)
1143 ------------------------------
1144 -- 4.1 Attribute Reference --
1145 ------------------------------
1147 -- Parsed by P_Name (4.1)
1149 -------------------------------
1150 -- 4.1 Attribute Designator --
1151 -------------------------------
1153 -- Parsed by P_Name (4.1)
1155 --------------------------------------
1156 -- 4.1.4 Range Attribute Reference --
1157 --------------------------------------
1159 -- RANGE_ATTRIBUTE_REFERENCE ::= PREFIX ' RANGE_ATTRIBUTE_DESIGNATOR
1161 -- RANGE_ATTRIBUTE_DESIGNATOR ::= range [(static_EXPRESSION)]
1163 -- In the grammar, a RANGE attribute is simply a name, but its use is
1164 -- highly restricted, so in the parser, we do not regard it as a name.
1165 -- Instead, P_Name returns without scanning the 'RANGE part of the
1166 -- attribute, and the caller uses the following function to construct
1167 -- a range attribute in places where it is appropriate.
1169 -- Note that RANGE here is treated essentially as an identifier,
1170 -- rather than a reserved word.
1172 -- The caller has parsed the prefix, i.e. a name, and Token points to
1173 -- the apostrophe. The token after the apostrophe is known to be RANGE
1174 -- at this point. The prefix node becomes the prefix of the attribute.
1176 -- Error_Recovery: Cannot raise Error_Resync
1178 function P_Range_Attribute_Reference
1179 (Prefix_Node : Node_Id)
1180 return Node_Id
1182 Attr_Node : Node_Id;
1184 begin
1185 Attr_Node := New_Node (N_Attribute_Reference, Token_Ptr);
1186 Set_Prefix (Attr_Node, Prefix_Node);
1187 Scan; -- past apostrophe
1189 if Style_Check then
1190 Style.Check_Attribute_Name (True);
1191 end if;
1193 Set_Attribute_Name (Attr_Node, Name_Range);
1194 Scan; -- past RANGE
1196 if Token = Tok_Left_Paren then
1197 Scan; -- past left paren
1198 Set_Expressions (Attr_Node, New_List (P_Expression_If_OK));
1199 T_Right_Paren;
1200 end if;
1202 return Attr_Node;
1203 end P_Range_Attribute_Reference;
1205 ---------------------------------------
1206 -- 4.1.4 Range Attribute Designator --
1207 ---------------------------------------
1209 -- Parsed by P_Range_Attribute_Reference (4.4)
1211 --------------------
1212 -- 4.3 Aggregate --
1213 --------------------
1215 -- AGGREGATE ::= RECORD_AGGREGATE | EXTENSION_AGGREGATE | ARRAY_AGGREGATE
1217 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3), except in the case where
1218 -- an aggregate is known to be required (code statement, extension
1219 -- aggregate), in which cases this routine performs the necessary check
1220 -- that we have an aggregate rather than a parenthesized expression
1222 -- Error recovery: can raise Error_Resync
1224 function P_Aggregate return Node_Id is
1225 Aggr_Sloc : constant Source_Ptr := Token_Ptr;
1226 Aggr_Node : constant Node_Id := P_Aggregate_Or_Paren_Expr;
1228 begin
1229 if Nkind (Aggr_Node) /= N_Aggregate
1230 and then
1231 Nkind (Aggr_Node) /= N_Extension_Aggregate
1232 then
1233 Error_Msg
1234 ("aggregate may not have single positional component", Aggr_Sloc);
1235 return Error;
1236 else
1237 return Aggr_Node;
1238 end if;
1239 end P_Aggregate;
1241 ------------------------------------------------
1242 -- 4.3 Aggregate or Parenthesized Expression --
1243 ------------------------------------------------
1245 -- This procedure parses out either an aggregate or a parenthesized
1246 -- expression (these two constructs are closely related, since a
1247 -- parenthesized expression looks like an aggregate with a single
1248 -- positional component).
1250 -- AGGREGATE ::=
1251 -- RECORD_AGGREGATE | EXTENSION_AGGREGATE | ARRAY_AGGREGATE
1253 -- RECORD_AGGREGATE ::= (RECORD_COMPONENT_ASSOCIATION_LIST)
1255 -- RECORD_COMPONENT_ASSOCIATION_LIST ::=
1256 -- RECORD_COMPONENT_ASSOCIATION {, RECORD_COMPONENT_ASSOCIATION}
1257 -- | null record
1259 -- RECORD_COMPONENT_ASSOCIATION ::=
1260 -- [COMPONENT_CHOICE_LIST =>] EXPRESSION
1262 -- COMPONENT_CHOICE_LIST ::=
1263 -- component_SELECTOR_NAME {| component_SELECTOR_NAME}
1264 -- | others
1266 -- EXTENSION_AGGREGATE ::=
1267 -- (ANCESTOR_PART with RECORD_COMPONENT_ASSOCIATION_LIST)
1269 -- ANCESTOR_PART ::= EXPRESSION | SUBTYPE_MARK
1271 -- ARRAY_AGGREGATE ::=
1272 -- POSITIONAL_ARRAY_AGGREGATE | NAMED_ARRAY_AGGREGATE
1274 -- POSITIONAL_ARRAY_AGGREGATE ::=
1275 -- (EXPRESSION, EXPRESSION {, EXPRESSION})
1276 -- | (EXPRESSION {, EXPRESSION}, others => EXPRESSION)
1277 -- | (EXPRESSION {, EXPRESSION}, others => <>)
1279 -- NAMED_ARRAY_AGGREGATE ::=
1280 -- (ARRAY_COMPONENT_ASSOCIATION {, ARRAY_COMPONENT_ASSOCIATION})
1282 -- PRIMARY ::= (EXPRESSION);
1284 -- Error recovery: can raise Error_Resync
1286 -- Note: POSITIONAL_ARRAY_AGGREGATE rule has been extended to give support
1287 -- to Ada 2005 limited aggregates (AI-287)
1289 function P_Aggregate_Or_Paren_Expr return Node_Id is
1290 Aggregate_Node : Node_Id;
1291 Expr_List : List_Id;
1292 Assoc_List : List_Id;
1293 Expr_Node : Node_Id;
1294 Lparen_Sloc : Source_Ptr;
1295 Scan_State : Saved_Scan_State;
1297 procedure Box_Error;
1298 -- Called if <> is encountered as positional aggregate element. Issues
1299 -- error message and sets Expr_Node to Error.
1301 function Is_Quantified_Expression return Boolean;
1302 -- The presence of iterated component associations requires a one
1303 -- token lookahead to distinguish it from quantified expressions.
1305 ---------------
1306 -- Box_Error --
1307 ---------------
1309 procedure Box_Error is
1310 begin
1311 if Ada_Version < Ada_2005 then
1312 Error_Msg_SC ("box in aggregate is an Ada 2005 extension");
1313 end if;
1315 -- Ada 2005 (AI-287): The box notation is allowed only with named
1316 -- notation because positional notation might be error prone. For
1317 -- example, in "(X, <>, Y, <>)", there is no type associated with
1318 -- the boxes, so you might not be leaving out the components you
1319 -- thought you were leaving out.
1321 Error_Msg_SC ("(Ada 2005) box only allowed with named notation");
1322 Scan; -- past box
1323 Expr_Node := Error;
1324 end Box_Error;
1326 ------------------------------
1327 -- Is_Quantified_Expression --
1328 ------------------------------
1330 function Is_Quantified_Expression return Boolean is
1331 Maybe : Boolean;
1332 Scan_State : Saved_Scan_State;
1334 begin
1335 Save_Scan_State (Scan_State);
1336 Scan; -- past FOR
1337 Maybe := Token = Tok_All or else Token = Tok_Some;
1338 Restore_Scan_State (Scan_State); -- to FOR
1339 return Maybe;
1340 end Is_Quantified_Expression;
1342 -- Start of processing for P_Aggregate_Or_Paren_Expr
1344 begin
1345 Lparen_Sloc := Token_Ptr;
1346 T_Left_Paren;
1348 -- Note on parentheses count. For cases like an if expression, the
1349 -- parens here really count as real parentheses for the paren count,
1350 -- so we adjust the paren count accordingly after scanning the expr.
1352 -- If expression
1354 if Token = Tok_If then
1355 Expr_Node := P_If_Expression;
1356 T_Right_Paren;
1357 Set_Paren_Count (Expr_Node, Paren_Count (Expr_Node) + 1);
1358 return Expr_Node;
1360 -- Case expression
1362 elsif Token = Tok_Case then
1363 Expr_Node := P_Case_Expression;
1364 T_Right_Paren;
1365 Set_Paren_Count (Expr_Node, Paren_Count (Expr_Node) + 1);
1366 return Expr_Node;
1368 -- Quantified expression
1370 elsif Token = Tok_For and then Is_Quantified_Expression then
1371 Expr_Node := P_Quantified_Expression;
1372 T_Right_Paren;
1373 Set_Paren_Count (Expr_Node, Paren_Count (Expr_Node) + 1);
1374 return Expr_Node;
1376 -- Note: the mechanism used here of rescanning the initial expression
1377 -- is distinctly unpleasant, but it saves a lot of fiddling in scanning
1378 -- out the discrete choice list.
1380 -- Deal with expression and extension aggregates first
1382 elsif Token /= Tok_Others then
1383 Save_Scan_State (Scan_State); -- at start of expression
1385 -- Deal with (NULL RECORD)
1387 if Token = Tok_Null then
1388 Scan; -- past NULL
1390 if Token = Tok_Record then
1391 Aggregate_Node := New_Node (N_Aggregate, Lparen_Sloc);
1392 Set_Null_Record_Present (Aggregate_Node, True);
1393 Scan; -- past RECORD
1394 T_Right_Paren;
1395 return Aggregate_Node;
1396 else
1397 Restore_Scan_State (Scan_State); -- to NULL that must be expr
1398 end if;
1400 elsif Token = Tok_For then
1401 Aggregate_Node := New_Node (N_Aggregate, Lparen_Sloc);
1402 Expr_Node := P_Iterated_Component_Association;
1403 goto Aggregate;
1404 end if;
1406 -- Scan expression, handling box appearing as positional argument
1408 if Token = Tok_Box then
1409 Box_Error;
1410 else
1411 Expr_Node := P_Expression_Or_Range_Attribute_If_OK;
1412 end if;
1414 -- Extension or Delta aggregate
1416 if Token = Tok_With then
1417 if Nkind (Expr_Node) = N_Attribute_Reference
1418 and then Attribute_Name (Expr_Node) = Name_Range
1419 then
1420 Bad_Range_Attribute (Sloc (Expr_Node));
1421 return Error;
1422 end if;
1424 if Ada_Version = Ada_83 then
1425 Error_Msg_SC ("(Ada 83) extension aggregate not allowed");
1426 end if;
1428 Scan; -- past WITH
1429 if Token = Tok_Delta then
1430 Scan; -- past DELTA
1431 Aggregate_Node := New_Node (N_Delta_Aggregate, Lparen_Sloc);
1432 Set_Expression (Aggregate_Node, Expr_Node);
1433 Expr_Node := Empty;
1435 goto Aggregate;
1437 else
1438 Aggregate_Node := New_Node (N_Extension_Aggregate, Lparen_Sloc);
1439 Set_Ancestor_Part (Aggregate_Node, Expr_Node);
1440 end if;
1442 -- Deal with WITH NULL RECORD case
1444 if Token = Tok_Null then
1445 Save_Scan_State (Scan_State); -- at NULL
1446 Scan; -- past NULL
1448 if Token = Tok_Record then
1449 Scan; -- past RECORD
1450 Set_Null_Record_Present (Aggregate_Node, True);
1451 T_Right_Paren;
1452 return Aggregate_Node;
1454 else
1455 Restore_Scan_State (Scan_State); -- to NULL that must be expr
1456 end if;
1457 end if;
1459 if Token /= Tok_Others then
1460 Save_Scan_State (Scan_State);
1461 Expr_Node := P_Expression;
1462 else
1463 Expr_Node := Empty;
1464 end if;
1466 -- Expression
1468 elsif Token = Tok_Right_Paren or else Token in Token_Class_Eterm then
1469 if Nkind (Expr_Node) = N_Attribute_Reference
1470 and then Attribute_Name (Expr_Node) = Name_Range
1471 then
1472 Error_Msg
1473 ("|parentheses not allowed for range attribute", Lparen_Sloc);
1474 Scan; -- past right paren
1475 return Expr_Node;
1476 end if;
1478 -- Bump paren count of expression
1480 if Expr_Node /= Error then
1481 Set_Paren_Count (Expr_Node, Paren_Count (Expr_Node) + 1);
1482 end if;
1484 T_Right_Paren; -- past right paren (error message if none)
1485 return Expr_Node;
1487 -- Normal aggregate
1489 else
1490 Aggregate_Node := New_Node (N_Aggregate, Lparen_Sloc);
1491 end if;
1493 -- Others
1495 else
1496 Aggregate_Node := New_Node (N_Aggregate, Lparen_Sloc);
1497 Expr_Node := Empty;
1498 end if;
1500 -- Prepare to scan list of component associations
1501 <<Aggregate>>
1502 Expr_List := No_List; -- don't set yet, maybe all named entries
1503 Assoc_List := No_List; -- don't set yet, maybe all positional entries
1505 -- This loop scans through component associations. On entry to the
1506 -- loop, an expression has been scanned at the start of the current
1507 -- association unless initial token was OTHERS, in which case
1508 -- Expr_Node is set to Empty.
1510 loop
1511 -- Deal with others association first. This is a named association
1513 if No (Expr_Node) then
1514 if No (Assoc_List) then
1515 Assoc_List := New_List;
1516 end if;
1518 Append (P_Record_Or_Array_Component_Association, Assoc_List);
1520 -- Improper use of WITH
1522 elsif Token = Tok_With then
1523 Error_Msg_SC ("WITH must be preceded by single expression in " &
1524 "extension aggregate");
1525 raise Error_Resync;
1527 -- Range attribute can only appear as part of a discrete choice list
1529 elsif Nkind (Expr_Node) = N_Attribute_Reference
1530 and then Attribute_Name (Expr_Node) = Name_Range
1531 and then Token /= Tok_Arrow
1532 and then Token /= Tok_Vertical_Bar
1533 then
1534 Bad_Range_Attribute (Sloc (Expr_Node));
1535 return Error;
1537 -- Assume positional case if comma, right paren, or literal or
1538 -- identifier or OTHERS follows (the latter cases are missing
1539 -- comma cases). Also assume positional if a semicolon follows,
1540 -- which can happen if there are missing parens.
1542 elsif Nkind (Expr_Node) = N_Iterated_Component_Association then
1543 if No (Assoc_List) then
1544 Assoc_List := New_List (Expr_Node);
1545 else
1546 Append_To (Assoc_List, Expr_Node);
1547 end if;
1549 elsif Token = Tok_Comma
1550 or else Token = Tok_Right_Paren
1551 or else Token = Tok_Others
1552 or else Token in Token_Class_Lit_Or_Name
1553 or else Token = Tok_Semicolon
1554 then
1555 if Present (Assoc_List) then
1556 Error_Msg_BC -- CODEFIX
1557 ("""='>"" expected (positional association cannot follow "
1558 & "named association)");
1559 end if;
1561 if No (Expr_List) then
1562 Expr_List := New_List;
1563 end if;
1565 Append (Expr_Node, Expr_List);
1567 -- Check for aggregate followed by left parent, maybe missing comma
1569 elsif Nkind (Expr_Node) = N_Aggregate
1570 and then Token = Tok_Left_Paren
1571 then
1572 T_Comma;
1574 if No (Expr_List) then
1575 Expr_List := New_List;
1576 end if;
1578 Append (Expr_Node, Expr_List);
1580 -- Anything else is assumed to be a named association
1582 else
1583 Restore_Scan_State (Scan_State); -- to start of expression
1585 if No (Assoc_List) then
1586 Assoc_List := New_List;
1587 end if;
1589 Append (P_Record_Or_Array_Component_Association, Assoc_List);
1590 end if;
1592 exit when not Comma_Present;
1594 -- If we are at an expression terminator, something is seriously
1595 -- wrong, so let's get out now, before we start eating up stuff
1596 -- that doesn't belong to us.
1598 if Token in Token_Class_Eterm and then Token /= Tok_For then
1599 Error_Msg_AP
1600 ("expecting expression or component association");
1601 exit;
1602 end if;
1604 -- Deal with misused box
1606 if Token = Tok_Box then
1607 Box_Error;
1609 -- Otherwise initiate for reentry to top of loop by scanning an
1610 -- initial expression, unless the first token is OTHERS or FOR,
1611 -- which indicates an iterated component association.
1613 elsif Token = Tok_Others then
1614 Expr_Node := Empty;
1616 elsif Token = Tok_For then
1617 Expr_Node := P_Iterated_Component_Association;
1619 else
1620 Save_Scan_State (Scan_State); -- at start of expression
1621 Expr_Node := P_Expression_Or_Range_Attribute_If_OK;
1623 end if;
1624 end loop;
1626 -- All component associations (positional and named) have been scanned
1628 T_Right_Paren;
1630 if Nkind (Aggregate_Node) /= N_Delta_Aggregate then
1631 Set_Expressions (Aggregate_Node, Expr_List);
1632 end if;
1634 Set_Component_Associations (Aggregate_Node, Assoc_List);
1635 return Aggregate_Node;
1636 end P_Aggregate_Or_Paren_Expr;
1638 ------------------------------------------------
1639 -- 4.3 Record or Array Component Association --
1640 ------------------------------------------------
1642 -- RECORD_COMPONENT_ASSOCIATION ::=
1643 -- [COMPONENT_CHOICE_LIST =>] EXPRESSION
1644 -- | COMPONENT_CHOICE_LIST => <>
1646 -- COMPONENT_CHOICE_LIST =>
1647 -- component_SELECTOR_NAME {| component_SELECTOR_NAME}
1648 -- | others
1650 -- ARRAY_COMPONENT_ASSOCIATION ::=
1651 -- DISCRETE_CHOICE_LIST => EXPRESSION
1652 -- | DISCRETE_CHOICE_LIST => <>
1653 -- | ITERATED_COMPONENT_ASSOCIATION
1655 -- Note: this routine only handles the named cases, including others.
1656 -- Cases where the component choice list is not present have already
1657 -- been handled directly.
1659 -- Error recovery: can raise Error_Resync
1661 -- Note: RECORD_COMPONENT_ASSOCIATION and ARRAY_COMPONENT_ASSOCIATION
1662 -- rules have been extended to give support to Ada 2005 limited
1663 -- aggregates (AI-287)
1665 function P_Record_Or_Array_Component_Association return Node_Id is
1666 Assoc_Node : Node_Id;
1668 begin
1669 -- A loop indicates an iterated_component_association
1671 if Token = Tok_For then
1672 return P_Iterated_Component_Association;
1673 end if;
1675 Assoc_Node := New_Node (N_Component_Association, Token_Ptr);
1676 Set_Choices (Assoc_Node, P_Discrete_Choice_List);
1677 Set_Sloc (Assoc_Node, Token_Ptr);
1678 TF_Arrow;
1680 if Token = Tok_Box then
1682 -- Ada 2005(AI-287): The box notation is used to indicate the
1683 -- default initialization of aggregate components
1685 if Ada_Version < Ada_2005 then
1686 Error_Msg_SP
1687 ("component association with '<'> is an Ada 2005 extension");
1688 Error_Msg_SP ("\unit must be compiled with -gnat05 switch");
1689 end if;
1691 Set_Box_Present (Assoc_Node);
1692 Scan; -- Past box
1693 else
1694 Set_Expression (Assoc_Node, P_Expression);
1695 end if;
1697 return Assoc_Node;
1698 end P_Record_Or_Array_Component_Association;
1700 -----------------------------
1701 -- 4.3.1 Record Aggregate --
1702 -----------------------------
1704 -- Case of enumeration aggregate is parsed by P_Aggregate (4.3)
1705 -- All other cases are parsed by P_Aggregate_Or_Paren_Expr (4.3)
1707 ----------------------------------------------
1708 -- 4.3.1 Record Component Association List --
1709 ----------------------------------------------
1711 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1713 ----------------------------------
1714 -- 4.3.1 Component Choice List --
1715 ----------------------------------
1717 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1719 --------------------------------
1720 -- 4.3.1 Extension Aggregate --
1721 --------------------------------
1723 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1725 --------------------------
1726 -- 4.3.1 Ancestor Part --
1727 --------------------------
1729 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1731 ----------------------------
1732 -- 4.3.1 Array Aggregate --
1733 ----------------------------
1735 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1737 ---------------------------------------
1738 -- 4.3.1 Positional Array Aggregate --
1739 ---------------------------------------
1741 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1743 ----------------------------------
1744 -- 4.3.1 Named Array Aggregate --
1745 ----------------------------------
1747 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1749 ----------------------------------------
1750 -- 4.3.1 Array Component Association --
1751 ----------------------------------------
1753 -- Parsed by P_Aggregate_Or_Paren_Expr (4.3)
1755 ---------------------
1756 -- 4.4 Expression --
1757 ---------------------
1759 -- This procedure parses EXPRESSION or CHOICE_EXPRESSION
1761 -- EXPRESSION ::=
1762 -- RELATION {LOGICAL_OPERATOR RELATION}
1764 -- CHOICE_EXPRESSION ::=
1765 -- CHOICE_RELATION {LOGICAL_OPERATOR CHOICE_RELATION}
1767 -- LOGICAL_OPERATOR ::= and | and then | or | or else | xor
1769 -- On return, Expr_Form indicates the categorization of the expression
1770 -- EF_Range_Attr is not a possible value (if a range attribute is found,
1771 -- an error message is given, and Error is returned).
1773 -- Error recovery: cannot raise Error_Resync
1775 function P_Expression return Node_Id is
1776 Logical_Op : Node_Kind;
1777 Prev_Logical_Op : Node_Kind;
1778 Op_Location : Source_Ptr;
1779 Node1 : Node_Id;
1780 Node2 : Node_Id;
1782 begin
1783 Node1 := P_Relation;
1785 if Token in Token_Class_Logop then
1786 Prev_Logical_Op := N_Empty;
1788 loop
1789 Op_Location := Token_Ptr;
1790 Logical_Op := P_Logical_Operator;
1792 if Prev_Logical_Op /= N_Empty and then
1793 Logical_Op /= Prev_Logical_Op
1794 then
1795 Error_Msg
1796 ("mixed logical operators in expression", Op_Location);
1797 Prev_Logical_Op := N_Empty;
1798 else
1799 Prev_Logical_Op := Logical_Op;
1800 end if;
1802 Node2 := Node1;
1803 Node1 := New_Op_Node (Logical_Op, Op_Location);
1804 Set_Left_Opnd (Node1, Node2);
1805 Set_Right_Opnd (Node1, P_Relation);
1807 -- Check for case of errant comma or semicolon
1809 if Token = Tok_Comma or else Token = Tok_Semicolon then
1810 declare
1811 Com : constant Boolean := Token = Tok_Comma;
1812 Scan_State : Saved_Scan_State;
1813 Logop : Node_Kind;
1815 begin
1816 Save_Scan_State (Scan_State); -- at comma/semicolon
1817 Scan; -- past comma/semicolon
1819 -- Check for AND THEN or OR ELSE after comma/semicolon. We
1820 -- do not deal with AND/OR because those cases get mixed up
1821 -- with the select alternatives case.
1823 if Token = Tok_And or else Token = Tok_Or then
1824 Logop := P_Logical_Operator;
1825 Restore_Scan_State (Scan_State); -- to comma/semicolon
1827 if Nkind_In (Logop, N_And_Then, N_Or_Else) then
1828 Scan; -- past comma/semicolon
1830 if Com then
1831 Error_Msg_SP -- CODEFIX
1832 ("|extra "","" ignored");
1833 else
1834 Error_Msg_SP -- CODEFIX
1835 ("|extra "";"" ignored");
1836 end if;
1838 else
1839 Restore_Scan_State (Scan_State); -- to comma/semicolon
1840 end if;
1842 else
1843 Restore_Scan_State (Scan_State); -- to comma/semicolon
1844 end if;
1845 end;
1846 end if;
1848 exit when Token not in Token_Class_Logop;
1849 end loop;
1851 Expr_Form := EF_Non_Simple;
1852 end if;
1854 if Token = Tok_Apostrophe then
1855 Bad_Range_Attribute (Token_Ptr);
1856 return Error;
1857 else
1858 return Node1;
1859 end if;
1860 end P_Expression;
1862 -- This function is identical to the normal P_Expression, except that it
1863 -- also permits the appearance of a case, conditional, or quantified
1864 -- expression if the call immediately follows a left paren, and followed
1865 -- by a right parenthesis. These forms are allowed if these conditions
1866 -- are not met, but an error message will be issued.
1868 function P_Expression_If_OK return Node_Id is
1869 begin
1870 -- Case of conditional, case or quantified expression
1872 if Token = Tok_Case or else Token = Tok_If or else Token = Tok_For then
1873 return P_Unparen_Cond_Case_Quant_Expression;
1875 -- Normal case, not case/conditional/quantified expression
1877 else
1878 return P_Expression;
1879 end if;
1880 end P_Expression_If_OK;
1882 -- This function is identical to the normal P_Expression, except that it
1883 -- checks that the expression scan did not stop on a right paren. It is
1884 -- called in all contexts where a right parenthesis cannot legitimately
1885 -- follow an expression.
1887 -- Error recovery: can not raise Error_Resync
1889 function P_Expression_No_Right_Paren return Node_Id is
1890 Expr : constant Node_Id := P_Expression;
1891 begin
1892 Ignore (Tok_Right_Paren);
1893 return Expr;
1894 end P_Expression_No_Right_Paren;
1896 ----------------------------------------
1897 -- 4.4 Expression_Or_Range_Attribute --
1898 ----------------------------------------
1900 -- EXPRESSION ::=
1901 -- RELATION {and RELATION} | RELATION {and then RELATION}
1902 -- | RELATION {or RELATION} | RELATION {or else RELATION}
1903 -- | RELATION {xor RELATION}
1905 -- RANGE_ATTRIBUTE_REFERENCE ::= PREFIX ' RANGE_ATTRIBUTE_DESIGNATOR
1907 -- RANGE_ATTRIBUTE_DESIGNATOR ::= range [(static_EXPRESSION)]
1909 -- On return, Expr_Form indicates the categorization of the expression
1910 -- and EF_Range_Attr is one of the possibilities.
1912 -- Error recovery: cannot raise Error_Resync
1914 -- In the grammar, a RANGE attribute is simply a name, but its use is
1915 -- highly restricted, so in the parser, we do not regard it as a name.
1916 -- Instead, P_Name returns without scanning the 'RANGE part of the
1917 -- attribute, and P_Expression_Or_Range_Attribute handles the range
1918 -- attribute reference. In the normal case where a range attribute is
1919 -- not allowed, an error message is issued by P_Expression.
1921 function P_Expression_Or_Range_Attribute return Node_Id is
1922 Logical_Op : Node_Kind;
1923 Prev_Logical_Op : Node_Kind;
1924 Op_Location : Source_Ptr;
1925 Node1 : Node_Id;
1926 Node2 : Node_Id;
1927 Attr_Node : Node_Id;
1929 begin
1930 Node1 := P_Relation;
1932 if Token = Tok_Apostrophe then
1933 Attr_Node := P_Range_Attribute_Reference (Node1);
1934 Expr_Form := EF_Range_Attr;
1935 return Attr_Node;
1937 elsif Token in Token_Class_Logop then
1938 Prev_Logical_Op := N_Empty;
1940 loop
1941 Op_Location := Token_Ptr;
1942 Logical_Op := P_Logical_Operator;
1944 if Prev_Logical_Op /= N_Empty and then
1945 Logical_Op /= Prev_Logical_Op
1946 then
1947 Error_Msg
1948 ("mixed logical operators in expression", Op_Location);
1949 Prev_Logical_Op := N_Empty;
1950 else
1951 Prev_Logical_Op := Logical_Op;
1952 end if;
1954 Node2 := Node1;
1955 Node1 := New_Op_Node (Logical_Op, Op_Location);
1956 Set_Left_Opnd (Node1, Node2);
1957 Set_Right_Opnd (Node1, P_Relation);
1958 exit when Token not in Token_Class_Logop;
1959 end loop;
1961 Expr_Form := EF_Non_Simple;
1962 end if;
1964 if Token = Tok_Apostrophe then
1965 Bad_Range_Attribute (Token_Ptr);
1966 return Error;
1967 else
1968 return Node1;
1969 end if;
1970 end P_Expression_Or_Range_Attribute;
1972 -- Version that allows a non-parenthesized case, conditional, or quantified
1973 -- expression if the call immediately follows a left paren, and followed
1974 -- by a right parenthesis. These forms are allowed if these conditions
1975 -- are not met, but an error message will be issued.
1977 function P_Expression_Or_Range_Attribute_If_OK return Node_Id is
1978 begin
1979 -- Case of conditional, case or quantified expression
1981 if Token = Tok_Case or else Token = Tok_If or else Token = Tok_For then
1982 return P_Unparen_Cond_Case_Quant_Expression;
1984 -- Normal case, not one of the above expression types
1986 else
1987 return P_Expression_Or_Range_Attribute;
1988 end if;
1989 end P_Expression_Or_Range_Attribute_If_OK;
1991 -------------------
1992 -- 4.4 Relation --
1993 -------------------
1995 -- This procedure scans both relations and choice relations
1997 -- CHOICE_RELATION ::=
1998 -- SIMPLE_EXPRESSION [RELATIONAL_OPERATOR SIMPLE_EXPRESSION]
2000 -- RELATION ::=
2001 -- SIMPLE_EXPRESSION [not] in MEMBERSHIP_CHOICE_LIST
2002 -- | RAISE_EXPRESSION
2004 -- MEMBERSHIP_CHOICE_LIST ::=
2005 -- MEMBERSHIP_CHOICE {'|' MEMBERSHIP CHOICE}
2007 -- MEMBERSHIP_CHOICE ::=
2008 -- CHOICE_EXPRESSION | RANGE | SUBTYPE_MARK
2010 -- RAISE_EXPRESSION ::= raise exception_NAME [with string_EXPRESSION]
2012 -- On return, Expr_Form indicates the categorization of the expression
2014 -- Note: if Token = Tok_Apostrophe on return, then Expr_Form is set to
2015 -- EF_Simple_Name and the following token is RANGE (range attribute case).
2017 -- Error recovery: cannot raise Error_Resync. If an error occurs within an
2018 -- expression, then tokens are scanned until either a non-expression token,
2019 -- a right paren (not matched by a left paren) or a comma, is encountered.
2021 function P_Relation return Node_Id is
2022 Node1, Node2 : Node_Id;
2023 Optok : Source_Ptr;
2025 begin
2026 -- First check for raise expression
2028 if Token = Tok_Raise then
2029 Expr_Form := EF_Non_Simple;
2030 return P_Raise_Expression;
2031 end if;
2033 -- All other cases
2035 Node1 := P_Simple_Expression;
2037 if Token not in Token_Class_Relop then
2038 return Node1;
2040 else
2041 -- Here we have a relational operator following. If so then scan it
2042 -- out. Note that the assignment symbol := is treated as a relational
2043 -- operator to improve the error recovery when it is misused for =.
2044 -- P_Relational_Operator also parses the IN and NOT IN operations.
2046 Optok := Token_Ptr;
2047 Node2 := New_Op_Node (P_Relational_Operator, Optok);
2048 Set_Left_Opnd (Node2, Node1);
2050 -- Case of IN or NOT IN
2052 if Prev_Token = Tok_In then
2053 P_Membership_Test (Node2);
2055 -- Case of relational operator (= /= < <= > >=)
2057 else
2058 Set_Right_Opnd (Node2, P_Simple_Expression);
2059 end if;
2061 Expr_Form := EF_Non_Simple;
2063 if Token in Token_Class_Relop then
2064 Error_Msg_SC ("unexpected relational operator");
2065 raise Error_Resync;
2066 end if;
2068 return Node2;
2069 end if;
2071 -- If any error occurs, then scan to the next expression terminator symbol
2072 -- or comma or right paren at the outer (i.e. current) parentheses level.
2073 -- The flags are set to indicate a normal simple expression.
2075 exception
2076 when Error_Resync =>
2077 Resync_Expression;
2078 Expr_Form := EF_Simple;
2079 return Error;
2080 end P_Relation;
2082 ----------------------------
2083 -- 4.4 Simple Expression --
2084 ----------------------------
2086 -- SIMPLE_EXPRESSION ::=
2087 -- [UNARY_ADDING_OPERATOR] TERM {BINARY_ADDING_OPERATOR TERM}
2089 -- On return, Expr_Form indicates the categorization of the expression
2091 -- Note: if Token = Tok_Apostrophe on return, then Expr_Form is set to
2092 -- EF_Simple_Name and the following token is RANGE (range attribute case).
2094 -- Error recovery: cannot raise Error_Resync. If an error occurs within an
2095 -- expression, then tokens are scanned until either a non-expression token,
2096 -- a right paren (not matched by a left paren) or a comma, is encountered.
2098 -- Note: P_Simple_Expression is called only internally by higher level
2099 -- expression routines. In cases in the grammar where a simple expression
2100 -- is required, the approach is to scan an expression, and then post an
2101 -- appropriate error message if the expression obtained is not simple. This
2102 -- gives better error recovery and treatment.
2104 function P_Simple_Expression return Node_Id is
2105 Scan_State : Saved_Scan_State;
2106 Node1 : Node_Id;
2107 Node2 : Node_Id;
2108 Tokptr : Source_Ptr;
2110 function At_Start_Of_Attribute return Boolean;
2111 -- Tests if we have quote followed by attribute name, if so, return True
2112 -- otherwise return False.
2114 ---------------------------
2115 -- At_Start_Of_Attribute --
2116 ---------------------------
2118 function At_Start_Of_Attribute return Boolean is
2119 begin
2120 if Token /= Tok_Apostrophe then
2121 return False;
2123 else
2124 declare
2125 Scan_State : Saved_Scan_State;
2127 begin
2128 Save_Scan_State (Scan_State);
2129 Scan; -- past quote
2131 if Token = Tok_Identifier
2132 and then Is_Attribute_Name (Chars (Token_Node))
2133 then
2134 Restore_Scan_State (Scan_State);
2135 return True;
2136 else
2137 Restore_Scan_State (Scan_State);
2138 return False;
2139 end if;
2140 end;
2141 end if;
2142 end At_Start_Of_Attribute;
2144 -- Start of processing for P_Simple_Expression
2146 begin
2147 -- Check for cases starting with a name. There are two reasons for
2148 -- special casing. First speed things up by catching a common case
2149 -- without going through several routine layers. Second the caller must
2150 -- be informed via Expr_Form when the simple expression is a name.
2152 if Token in Token_Class_Name then
2153 Node1 := P_Name;
2155 -- Deal with apostrophe cases
2157 if Token = Tok_Apostrophe then
2158 Save_Scan_State (Scan_State); -- at apostrophe
2159 Scan; -- past apostrophe
2161 -- If qualified expression, scan it out and fall through
2163 if Token = Tok_Left_Paren then
2164 Node1 := P_Qualified_Expression (Node1);
2165 Expr_Form := EF_Simple;
2167 -- If range attribute, then we return with Token pointing to the
2168 -- apostrophe. Note: avoid the normal error check on exit. We
2169 -- know that the expression really is complete in this case.
2171 else -- Token = Tok_Range then
2172 Restore_Scan_State (Scan_State); -- to apostrophe
2173 Expr_Form := EF_Simple_Name;
2174 return Node1;
2175 end if;
2176 end if;
2178 -- If an expression terminator follows, the previous processing
2179 -- completely scanned out the expression (a common case), and
2180 -- left Expr_Form set appropriately for returning to our caller.
2182 if Token in Token_Class_Sterm then
2183 null;
2185 -- If we do not have an expression terminator, then complete the
2186 -- scan of a simple expression. This code duplicates the code
2187 -- found in P_Term and P_Factor.
2189 else
2190 if Token = Tok_Double_Asterisk then
2191 if Style_Check then
2192 Style.Check_Exponentiation_Operator;
2193 end if;
2195 Node2 := New_Op_Node (N_Op_Expon, Token_Ptr);
2196 Scan; -- past **
2197 Set_Left_Opnd (Node2, Node1);
2198 Set_Right_Opnd (Node2, P_Primary);
2199 Check_Bad_Exp;
2200 Node1 := Node2;
2201 end if;
2203 loop
2204 exit when Token not in Token_Class_Mulop;
2205 Tokptr := Token_Ptr;
2206 Node2 := New_Op_Node (P_Multiplying_Operator, Tokptr);
2208 if Style_Check then
2209 Style.Check_Binary_Operator;
2210 end if;
2212 Scan; -- past operator
2213 Set_Left_Opnd (Node2, Node1);
2214 Set_Right_Opnd (Node2, P_Factor);
2215 Node1 := Node2;
2216 end loop;
2218 loop
2219 exit when Token not in Token_Class_Binary_Addop;
2220 Tokptr := Token_Ptr;
2221 Node2 := New_Op_Node (P_Binary_Adding_Operator, Tokptr);
2223 if Style_Check then
2224 Style.Check_Binary_Operator;
2225 end if;
2227 Scan; -- past operator
2228 Set_Left_Opnd (Node2, Node1);
2229 Set_Right_Opnd (Node2, P_Term);
2230 Node1 := Node2;
2231 end loop;
2233 Expr_Form := EF_Simple;
2234 end if;
2236 -- Cases where simple expression does not start with a name
2238 else
2239 -- Scan initial sign and initial Term
2241 if Token in Token_Class_Unary_Addop then
2242 Tokptr := Token_Ptr;
2243 Node1 := New_Op_Node (P_Unary_Adding_Operator, Tokptr);
2245 if Style_Check then
2246 Style.Check_Unary_Plus_Or_Minus (Inside_Depends);
2247 end if;
2249 Scan; -- past operator
2250 Set_Right_Opnd (Node1, P_Term);
2251 else
2252 Node1 := P_Term;
2253 end if;
2255 -- In the following, we special-case a sequence of concatenations of
2256 -- string literals, such as "aaa" & "bbb" & ... & "ccc", with nothing
2257 -- else mixed in. For such a sequence, we return a tree representing
2258 -- "" & "aaabbb...ccc" (a single concatenation). This is done only if
2259 -- the number of concatenations is large. If semantic analysis
2260 -- resolves the "&" to a predefined one, then this folding gives the
2261 -- right answer. Otherwise, semantic analysis will complain about a
2262 -- capacity-exceeded error. The purpose of this trick is to avoid
2263 -- creating a deeply nested tree, which would cause deep recursion
2264 -- during semantics, causing stack overflow. This way, we can handle
2265 -- enormous concatenations in the normal case of predefined "&". We
2266 -- first build up the normal tree, and then rewrite it if
2267 -- appropriate.
2269 declare
2270 Num_Concats_Threshold : constant Positive := 1000;
2271 -- Arbitrary threshold value to enable optimization
2273 First_Node : constant Node_Id := Node1;
2274 Is_Strlit_Concat : Boolean;
2275 -- True iff we've parsed a sequence of concatenations of string
2276 -- literals, with nothing else mixed in.
2278 Num_Concats : Natural;
2279 -- Number of "&" operators if Is_Strlit_Concat is True
2281 begin
2282 Is_Strlit_Concat :=
2283 Nkind (Node1) = N_String_Literal
2284 and then Token = Tok_Ampersand;
2285 Num_Concats := 0;
2287 -- Scan out sequence of terms separated by binary adding operators
2289 loop
2290 exit when Token not in Token_Class_Binary_Addop;
2291 Tokptr := Token_Ptr;
2292 Node2 := New_Op_Node (P_Binary_Adding_Operator, Tokptr);
2294 if Style_Check and then not Debug_Flag_Dot_QQ then
2295 Style.Check_Binary_Operator;
2296 end if;
2298 Scan; -- past operator
2299 Set_Left_Opnd (Node2, Node1);
2300 Node1 := P_Term;
2301 Set_Right_Opnd (Node2, Node1);
2303 -- Check if we're still concatenating string literals
2305 Is_Strlit_Concat :=
2306 Is_Strlit_Concat
2307 and then Nkind (Node2) = N_Op_Concat
2308 and then Nkind (Node1) = N_String_Literal;
2310 if Is_Strlit_Concat then
2311 Num_Concats := Num_Concats + 1;
2312 end if;
2314 Node1 := Node2;
2315 end loop;
2317 -- If we have an enormous series of concatenations of string
2318 -- literals, rewrite as explained above. The Is_Folded_In_Parser
2319 -- flag tells semantic analysis that if the "&" is not predefined,
2320 -- the folded value is wrong.
2322 if Is_Strlit_Concat
2323 and then Num_Concats >= Num_Concats_Threshold
2324 then
2325 declare
2326 Empty_String_Val : String_Id;
2327 -- String_Id for ""
2329 Strlit_Concat_Val : String_Id;
2330 -- Contains the folded value (which will be correct if the
2331 -- "&" operators are the predefined ones).
2333 Cur_Node : Node_Id;
2334 -- For walking up the tree
2336 New_Node : Node_Id;
2337 -- Folded node to replace Node1
2339 Loc : constant Source_Ptr := Sloc (First_Node);
2341 begin
2342 -- Walk up the tree starting at the leftmost string literal
2343 -- (First_Node), building up the Strlit_Concat_Val as we
2344 -- go. Note that we do not use recursion here -- the whole
2345 -- point is to avoid recursively walking that enormous tree.
2347 Start_String;
2348 Store_String_Chars (Strval (First_Node));
2350 Cur_Node := Parent (First_Node);
2351 while Present (Cur_Node) loop
2352 pragma Assert (Nkind (Cur_Node) = N_Op_Concat and then
2353 Nkind (Right_Opnd (Cur_Node)) = N_String_Literal);
2355 Store_String_Chars (Strval (Right_Opnd (Cur_Node)));
2356 Cur_Node := Parent (Cur_Node);
2357 end loop;
2359 Strlit_Concat_Val := End_String;
2361 -- Create new folded node, and rewrite result with a concat-
2362 -- enation of an empty string literal and the folded node.
2364 Start_String;
2365 Empty_String_Val := End_String;
2366 New_Node :=
2367 Make_Op_Concat (Loc,
2368 Make_String_Literal (Loc, Empty_String_Val),
2369 Make_String_Literal (Loc, Strlit_Concat_Val,
2370 Is_Folded_In_Parser => True));
2371 Rewrite (Node1, New_Node);
2372 end;
2373 end if;
2374 end;
2376 -- All done, we clearly do not have name or numeric literal so this
2377 -- is a case of a simple expression which is some other possibility.
2379 Expr_Form := EF_Simple;
2380 end if;
2382 -- Come here at end of simple expression, where we do a couple of
2383 -- special checks to improve error recovery.
2385 -- Special test to improve error recovery. If the current token is a
2386 -- period, then someone is trying to do selection on something that is
2387 -- not a name, e.g. a qualified expression.
2389 if Token = Tok_Dot then
2390 Error_Msg_SC ("prefix for selection is not a name");
2392 -- If qualified expression, comment and continue, otherwise something
2393 -- is pretty nasty so do an Error_Resync call.
2395 if Ada_Version < Ada_2012
2396 and then Nkind (Node1) = N_Qualified_Expression
2397 then
2398 Error_Msg_SC ("\would be legal in Ada 2012 mode");
2399 else
2400 raise Error_Resync;
2401 end if;
2402 end if;
2404 -- Special test to improve error recovery: If the current token is
2405 -- not the first token on a line (as determined by checking the
2406 -- previous token position with the start of the current line),
2407 -- then we insist that we have an appropriate terminating token.
2408 -- Consider the following two examples:
2410 -- 1) if A nad B then ...
2412 -- 2) A := B
2413 -- C := D
2415 -- In the first example, we would like to issue a binary operator
2416 -- expected message and resynchronize to the then. In the second
2417 -- example, we do not want to issue a binary operator message, so
2418 -- that instead we will get the missing semicolon message. This
2419 -- distinction is of course a heuristic which does not always work,
2420 -- but in practice it is quite effective.
2422 -- Note: the one case in which we do not go through this circuit is
2423 -- when we have scanned a range attribute and want to return with
2424 -- Token pointing to the apostrophe. The apostrophe is not normally
2425 -- an expression terminator, and is not in Token_Class_Sterm, but
2426 -- in this special case we know that the expression is complete.
2428 if not Token_Is_At_Start_Of_Line
2429 and then Token not in Token_Class_Sterm
2430 then
2431 -- Normally the right error message is indeed that we expected a
2432 -- binary operator, but in the case of being between a right and left
2433 -- paren, e.g. in an aggregate, a more likely error is missing comma.
2435 if Prev_Token = Tok_Right_Paren and then Token = Tok_Left_Paren then
2436 T_Comma;
2438 -- And if we have a quote, we may have a bad attribute
2440 elsif At_Start_Of_Attribute then
2441 Error_Msg_SC ("prefix of attribute must be a name");
2443 if Ada_Version >= Ada_2012 then
2444 Error_Msg_SC ("\qualify expression to turn it into a name");
2445 end if;
2447 -- Normal case for binary operator expected message
2449 else
2450 Error_Msg_AP ("binary operator expected");
2451 end if;
2453 raise Error_Resync;
2455 else
2456 return Node1;
2457 end if;
2459 -- If any error occurs, then scan to next expression terminator symbol
2460 -- or comma, right paren or vertical bar at the outer (i.e. current) paren
2461 -- level. Expr_Form is set to indicate a normal simple expression.
2463 exception
2464 when Error_Resync =>
2465 Resync_Expression;
2466 Expr_Form := EF_Simple;
2467 return Error;
2468 end P_Simple_Expression;
2470 -----------------------------------------------
2471 -- 4.4 Simple Expression or Range Attribute --
2472 -----------------------------------------------
2474 -- SIMPLE_EXPRESSION ::=
2475 -- [UNARY_ADDING_OPERATOR] TERM {BINARY_ADDING_OPERATOR TERM}
2477 -- RANGE_ATTRIBUTE_REFERENCE ::= PREFIX ' RANGE_ATTRIBUTE_DESIGNATOR
2479 -- RANGE_ATTRIBUTE_DESIGNATOR ::= range [(static_EXPRESSION)]
2481 -- Error recovery: cannot raise Error_Resync
2483 function P_Simple_Expression_Or_Range_Attribute return Node_Id is
2484 Sexpr : Node_Id;
2485 Attr_Node : Node_Id;
2487 begin
2488 -- We don't just want to roar ahead and call P_Simple_Expression
2489 -- here, since we want to handle the case of a parenthesized range
2490 -- attribute cleanly.
2492 if Token = Tok_Left_Paren then
2493 declare
2494 Lptr : constant Source_Ptr := Token_Ptr;
2495 Scan_State : Saved_Scan_State;
2497 begin
2498 Save_Scan_State (Scan_State);
2499 Scan; -- past left paren
2500 Sexpr := P_Simple_Expression;
2502 if Token = Tok_Apostrophe then
2503 Attr_Node := P_Range_Attribute_Reference (Sexpr);
2504 Expr_Form := EF_Range_Attr;
2506 if Token = Tok_Right_Paren then
2507 Scan; -- scan past right paren if present
2508 end if;
2510 Error_Msg ("parentheses not allowed for range attribute", Lptr);
2512 return Attr_Node;
2513 end if;
2515 Restore_Scan_State (Scan_State);
2516 end;
2517 end if;
2519 -- Here after dealing with parenthesized range attribute
2521 Sexpr := P_Simple_Expression;
2523 if Token = Tok_Apostrophe then
2524 Attr_Node := P_Range_Attribute_Reference (Sexpr);
2525 Expr_Form := EF_Range_Attr;
2526 return Attr_Node;
2528 else
2529 return Sexpr;
2530 end if;
2531 end P_Simple_Expression_Or_Range_Attribute;
2533 ---------------
2534 -- 4.4 Term --
2535 ---------------
2537 -- TERM ::= FACTOR {MULTIPLYING_OPERATOR FACTOR}
2539 -- Error recovery: can raise Error_Resync
2541 function P_Term return Node_Id is
2542 Node1, Node2 : Node_Id;
2543 Tokptr : Source_Ptr;
2545 begin
2546 Node1 := P_Factor;
2548 loop
2549 exit when Token not in Token_Class_Mulop;
2550 Tokptr := Token_Ptr;
2551 Node2 := New_Op_Node (P_Multiplying_Operator, Tokptr);
2553 if Style_Check and then not Debug_Flag_Dot_QQ then
2554 Style.Check_Binary_Operator;
2555 end if;
2557 Scan; -- past operator
2558 Set_Left_Opnd (Node2, Node1);
2559 Set_Right_Opnd (Node2, P_Factor);
2560 Node1 := Node2;
2561 end loop;
2563 return Node1;
2564 end P_Term;
2566 -----------------
2567 -- 4.4 Factor --
2568 -----------------
2570 -- FACTOR ::= PRIMARY [** PRIMARY] | abs PRIMARY | not PRIMARY
2572 -- Error recovery: can raise Error_Resync
2574 function P_Factor return Node_Id is
2575 Node1 : Node_Id;
2576 Node2 : Node_Id;
2578 begin
2579 if Token = Tok_Abs then
2580 Node1 := New_Op_Node (N_Op_Abs, Token_Ptr);
2582 if Style_Check then
2583 Style.Check_Abs_Not;
2584 end if;
2586 Scan; -- past ABS
2587 Set_Right_Opnd (Node1, P_Primary);
2588 return Node1;
2590 elsif Token = Tok_Not then
2591 Node1 := New_Op_Node (N_Op_Not, Token_Ptr);
2593 if Style_Check then
2594 Style.Check_Abs_Not;
2595 end if;
2597 Scan; -- past NOT
2598 Set_Right_Opnd (Node1, P_Primary);
2599 return Node1;
2601 else
2602 Node1 := P_Primary;
2604 if Token = Tok_Double_Asterisk then
2605 Node2 := New_Op_Node (N_Op_Expon, Token_Ptr);
2606 Scan; -- past **
2607 Set_Left_Opnd (Node2, Node1);
2608 Set_Right_Opnd (Node2, P_Primary);
2609 Check_Bad_Exp;
2610 return Node2;
2611 else
2612 return Node1;
2613 end if;
2614 end if;
2615 end P_Factor;
2617 ------------------
2618 -- 4.4 Primary --
2619 ------------------
2621 -- PRIMARY ::=
2622 -- NUMERIC_LITERAL | null
2623 -- | STRING_LITERAL | AGGREGATE
2624 -- | NAME | QUALIFIED_EXPRESSION
2625 -- | ALLOCATOR | (EXPRESSION) | QUANTIFIED_EXPRESSION
2627 -- Error recovery: can raise Error_Resync
2629 function P_Primary return Node_Id is
2630 Scan_State : Saved_Scan_State;
2631 Node1 : Node_Id;
2633 Lparen : constant Boolean := Prev_Token = Tok_Left_Paren;
2634 -- Remember if previous token is a left parenthesis. This is used to
2635 -- deal with checking whether IF/CASE/FOR expressions appearing as
2636 -- primaries require extra parenthesization.
2638 begin
2639 -- The loop runs more than once only if misplaced pragmas are found
2640 -- or if a misplaced unary minus is skipped.
2642 loop
2643 case Token is
2645 -- Name token can start a name, call or qualified expression, all
2646 -- of which are acceptable possibilities for primary. Note also
2647 -- that string literal is included in name (as operator symbol)
2648 -- and type conversion is included in name (as indexed component).
2650 when Tok_Char_Literal
2651 | Tok_Identifier
2652 | Tok_Operator_Symbol
2654 Node1 := P_Name;
2656 -- All done unless apostrophe follows
2658 if Token /= Tok_Apostrophe then
2659 return Node1;
2661 -- Apostrophe following means that we have either just parsed
2662 -- the subtype mark of a qualified expression, or the prefix
2663 -- or a range attribute.
2665 else -- Token = Tok_Apostrophe
2666 Save_Scan_State (Scan_State); -- at apostrophe
2667 Scan; -- past apostrophe
2669 -- If range attribute, then this is always an error, since
2670 -- the only legitimate case (where the scanned expression is
2671 -- a qualified simple name) is handled at the level of the
2672 -- Simple_Expression processing. This case corresponds to a
2673 -- usage such as 3 + A'Range, which is always illegal.
2675 if Token = Tok_Range then
2676 Restore_Scan_State (Scan_State); -- to apostrophe
2677 Bad_Range_Attribute (Token_Ptr);
2678 return Error;
2680 -- If left paren, then we have a qualified expression.
2681 -- Note that P_Name guarantees that in this case, where
2682 -- Token = Tok_Apostrophe on return, the only two possible
2683 -- tokens following the apostrophe are left paren and
2684 -- RANGE, so we know we have a left paren here.
2686 else -- Token = Tok_Left_Paren
2687 return P_Qualified_Expression (Node1);
2689 end if;
2690 end if;
2692 -- Numeric or string literal
2694 when Tok_Integer_Literal
2695 | Tok_Real_Literal
2696 | Tok_String_Literal
2698 Node1 := Token_Node;
2699 Scan; -- past number
2700 return Node1;
2702 -- Left paren, starts aggregate or parenthesized expression
2704 when Tok_Left_Paren =>
2705 declare
2706 Expr : constant Node_Id := P_Aggregate_Or_Paren_Expr;
2708 begin
2709 if Nkind (Expr) = N_Attribute_Reference
2710 and then Attribute_Name (Expr) = Name_Range
2711 then
2712 Bad_Range_Attribute (Sloc (Expr));
2713 end if;
2715 return Expr;
2716 end;
2718 -- Allocator
2720 when Tok_New =>
2721 return P_Allocator;
2723 -- Null
2725 when Tok_Null =>
2726 Scan; -- past NULL
2727 return New_Node (N_Null, Prev_Token_Ptr);
2729 -- Pragma, not allowed here, so just skip past it
2731 when Tok_Pragma =>
2732 P_Pragmas_Misplaced;
2734 -- Deal with IF (possible unparenthesized if expression)
2736 when Tok_If =>
2738 -- If this looks like a real if, defined as an IF appearing at
2739 -- the start of a new line, then we consider we have a missing
2740 -- operand. If in Ada 2012 and the IF is not properly indented
2741 -- for a statement, we prefer to issue a message about an ill-
2742 -- parenthesized if expression.
2744 if Token_Is_At_Start_Of_Line
2745 and then not
2746 (Ada_Version >= Ada_2012
2747 and then Style_Check_Indentation /= 0
2748 and then Start_Column rem Style_Check_Indentation /= 0)
2749 then
2750 Error_Msg_AP ("missing operand");
2751 return Error;
2753 -- If this looks like an if expression, then treat it that way
2754 -- with an error message if not explicitly surrounded by
2755 -- parentheses.
2757 elsif Ada_Version >= Ada_2012 then
2758 Node1 := P_If_Expression;
2760 if not (Lparen and then Token = Tok_Right_Paren) then
2761 Error_Msg
2762 ("if expression must be parenthesized", Sloc (Node1));
2763 end if;
2765 return Node1;
2767 -- Otherwise treat as misused identifier
2769 else
2770 return P_Identifier;
2771 end if;
2773 -- Deal with CASE (possible unparenthesized case expression)
2775 when Tok_Case =>
2777 -- If this looks like a real case, defined as a CASE appearing
2778 -- the start of a new line, then we consider we have a missing
2779 -- operand. If in Ada 2012 and the CASE is not properly
2780 -- indented for a statement, we prefer to issue a message about
2781 -- an ill-parenthesized case expression.
2783 if Token_Is_At_Start_Of_Line
2784 and then not
2785 (Ada_Version >= Ada_2012
2786 and then Style_Check_Indentation /= 0
2787 and then Start_Column rem Style_Check_Indentation /= 0)
2788 then
2789 Error_Msg_AP ("missing operand");
2790 return Error;
2792 -- If this looks like a case expression, then treat it that way
2793 -- with an error message if not within parentheses.
2795 elsif Ada_Version >= Ada_2012 then
2796 Node1 := P_Case_Expression;
2798 if not (Lparen and then Token = Tok_Right_Paren) then
2799 Error_Msg
2800 ("case expression must be parenthesized", Sloc (Node1));
2801 end if;
2803 return Node1;
2805 -- Otherwise treat as misused identifier
2807 else
2808 return P_Identifier;
2809 end if;
2811 -- For [all | some] indicates a quantified expression
2813 when Tok_For =>
2814 if Token_Is_At_Start_Of_Line then
2815 Error_Msg_AP ("misplaced loop");
2816 return Error;
2818 elsif Ada_Version >= Ada_2012 then
2819 Save_Scan_State (Scan_State);
2820 Scan; -- past FOR
2822 if Token = Tok_All or else Token = Tok_Some then
2823 Restore_Scan_State (Scan_State); -- To FOR
2824 Node1 := P_Quantified_Expression;
2826 if not (Lparen and then Token = Tok_Right_Paren) then
2827 Error_Msg
2828 ("quantified expression must be parenthesized",
2829 Sloc (Node1));
2830 end if;
2831 else
2832 Restore_Scan_State (Scan_State); -- To FOR
2833 Node1 := P_Iterated_Component_Association;
2834 end if;
2836 return Node1;
2838 -- Otherwise treat as misused identifier
2840 else
2841 return P_Identifier;
2842 end if;
2844 -- Minus may well be an improper attempt at a unary minus. Give
2845 -- a message, skip the minus and keep going.
2847 when Tok_Minus =>
2848 Error_Msg_SC ("parentheses required for unary minus");
2849 Scan; -- past minus
2851 when Tok_At_Sign => -- AI12-0125 : target_name
2852 if Ada_Version < Ada_2020 then
2853 Error_Msg_SC ("target name is an Ada 2020 extension");
2854 Error_Msg_SC ("\compile with -gnatX");
2855 end if;
2857 Node1 := P_Name;
2858 return Node1;
2860 -- Anything else is illegal as the first token of a primary, but
2861 -- we test for some common errors, to improve error messages.
2863 when others =>
2864 if Is_Reserved_Identifier then
2865 return P_Identifier;
2867 elsif Prev_Token = Tok_Comma then
2868 Error_Msg_SP -- CODEFIX
2869 ("|extra "","" ignored");
2870 raise Error_Resync;
2872 else
2873 Error_Msg_AP ("missing operand");
2874 raise Error_Resync;
2875 end if;
2876 end case;
2877 end loop;
2878 end P_Primary;
2880 -------------------------------
2881 -- 4.4 Quantified_Expression --
2882 -------------------------------
2884 -- QUANTIFIED_EXPRESSION ::=
2885 -- for QUANTIFIER LOOP_PARAMETER_SPECIFICATION => PREDICATE |
2886 -- for QUANTIFIER ITERATOR_SPECIFICATION => PREDICATE
2888 function P_Quantified_Expression return Node_Id is
2889 I_Spec : Node_Id;
2890 Node1 : Node_Id;
2892 begin
2893 Error_Msg_Ada_2012_Feature ("quantified expression", Token_Ptr);
2894 Scan; -- past FOR
2895 Node1 := New_Node (N_Quantified_Expression, Prev_Token_Ptr);
2897 if Token = Tok_All then
2898 Set_All_Present (Node1);
2899 elsif Token /= Tok_Some then
2900 Error_Msg_AP ("missing quantifier");
2901 raise Error_Resync;
2902 end if;
2904 Scan; -- past ALL or SOME
2905 I_Spec := P_Loop_Parameter_Specification;
2907 if Nkind (I_Spec) = N_Loop_Parameter_Specification then
2908 Set_Loop_Parameter_Specification (Node1, I_Spec);
2909 else
2910 Set_Iterator_Specification (Node1, I_Spec);
2911 end if;
2913 if Token = Tok_Arrow then
2914 Scan;
2915 Set_Condition (Node1, P_Expression);
2916 return Node1;
2917 else
2918 Error_Msg_AP ("missing arrow");
2919 raise Error_Resync;
2920 end if;
2921 end P_Quantified_Expression;
2923 ---------------------------
2924 -- 4.5 Logical Operator --
2925 ---------------------------
2927 -- LOGICAL_OPERATOR ::= and | or | xor
2929 -- Note: AND THEN and OR ELSE are also treated as logical operators
2930 -- by the parser (even though they are not operators semantically)
2932 -- The value returned is the appropriate Node_Kind code for the operator
2933 -- On return, Token points to the token following the scanned operator.
2935 -- The caller has checked that the first token is a legitimate logical
2936 -- operator token (i.e. is either XOR, AND, OR).
2938 -- Error recovery: cannot raise Error_Resync
2940 function P_Logical_Operator return Node_Kind is
2941 begin
2942 if Token = Tok_And then
2943 if Style_Check then
2944 Style.Check_Binary_Operator;
2945 end if;
2947 Scan; -- past AND
2949 if Token = Tok_Then then
2950 Scan; -- past THEN
2951 return N_And_Then;
2952 else
2953 return N_Op_And;
2954 end if;
2956 elsif Token = Tok_Or then
2957 if Style_Check then
2958 Style.Check_Binary_Operator;
2959 end if;
2961 Scan; -- past OR
2963 if Token = Tok_Else then
2964 Scan; -- past ELSE
2965 return N_Or_Else;
2966 else
2967 return N_Op_Or;
2968 end if;
2970 else -- Token = Tok_Xor
2971 if Style_Check then
2972 Style.Check_Binary_Operator;
2973 end if;
2975 Scan; -- past XOR
2976 return N_Op_Xor;
2977 end if;
2978 end P_Logical_Operator;
2980 ------------------------------
2981 -- 4.5 Relational Operator --
2982 ------------------------------
2984 -- RELATIONAL_OPERATOR ::= = | /= | < | <= | > | >=
2986 -- The value returned is the appropriate Node_Kind code for the operator.
2987 -- On return, Token points to the operator token, NOT past it.
2989 -- The caller has checked that the first token is a legitimate relational
2990 -- operator token (i.e. is one of the operator tokens listed above).
2992 -- Error recovery: cannot raise Error_Resync
2994 function P_Relational_Operator return Node_Kind is
2995 Op_Kind : Node_Kind;
2996 Relop_Node : constant array (Token_Class_Relop) of Node_Kind :=
2997 (Tok_Less => N_Op_Lt,
2998 Tok_Equal => N_Op_Eq,
2999 Tok_Greater => N_Op_Gt,
3000 Tok_Not_Equal => N_Op_Ne,
3001 Tok_Greater_Equal => N_Op_Ge,
3002 Tok_Less_Equal => N_Op_Le,
3003 Tok_In => N_In,
3004 Tok_Not => N_Not_In,
3005 Tok_Box => N_Op_Ne);
3007 begin
3008 if Token = Tok_Box then
3009 Error_Msg_SC -- CODEFIX
3010 ("|""'<'>"" should be ""/=""");
3011 end if;
3013 Op_Kind := Relop_Node (Token);
3015 if Style_Check then
3016 Style.Check_Binary_Operator;
3017 end if;
3019 Scan; -- past operator token
3021 -- Deal with NOT IN, if previous token was NOT, we must have IN now
3023 if Prev_Token = Tok_Not then
3025 -- Style check, for NOT IN, we require one space between NOT and IN
3027 if Style_Check and then Token = Tok_In then
3028 Style.Check_Not_In;
3029 end if;
3031 T_In;
3032 end if;
3034 return Op_Kind;
3035 end P_Relational_Operator;
3037 ---------------------------------
3038 -- 4.5 Binary Adding Operator --
3039 ---------------------------------
3041 -- BINARY_ADDING_OPERATOR ::= + | - | &
3043 -- The value returned is the appropriate Node_Kind code for the operator.
3044 -- On return, Token points to the operator token (NOT past it).
3046 -- The caller has checked that the first token is a legitimate adding
3047 -- operator token (i.e. is one of the operator tokens listed above).
3049 -- Error recovery: cannot raise Error_Resync
3051 function P_Binary_Adding_Operator return Node_Kind is
3052 Addop_Node : constant array (Token_Class_Binary_Addop) of Node_Kind :=
3053 (Tok_Ampersand => N_Op_Concat,
3054 Tok_Minus => N_Op_Subtract,
3055 Tok_Plus => N_Op_Add);
3056 begin
3057 return Addop_Node (Token);
3058 end P_Binary_Adding_Operator;
3060 --------------------------------
3061 -- 4.5 Unary Adding Operator --
3062 --------------------------------
3064 -- UNARY_ADDING_OPERATOR ::= + | -
3066 -- The value returned is the appropriate Node_Kind code for the operator.
3067 -- On return, Token points to the operator token (NOT past it).
3069 -- The caller has checked that the first token is a legitimate adding
3070 -- operator token (i.e. is one of the operator tokens listed above).
3072 -- Error recovery: cannot raise Error_Resync
3074 function P_Unary_Adding_Operator return Node_Kind is
3075 Addop_Node : constant array (Token_Class_Unary_Addop) of Node_Kind :=
3076 (Tok_Minus => N_Op_Minus,
3077 Tok_Plus => N_Op_Plus);
3078 begin
3079 return Addop_Node (Token);
3080 end P_Unary_Adding_Operator;
3082 -------------------------------
3083 -- 4.5 Multiplying Operator --
3084 -------------------------------
3086 -- MULTIPLYING_OPERATOR ::= * | / | mod | rem
3088 -- The value returned is the appropriate Node_Kind code for the operator.
3089 -- On return, Token points to the operator token (NOT past it).
3091 -- The caller has checked that the first token is a legitimate multiplying
3092 -- operator token (i.e. is one of the operator tokens listed above).
3094 -- Error recovery: cannot raise Error_Resync
3096 function P_Multiplying_Operator return Node_Kind is
3097 Mulop_Node : constant array (Token_Class_Mulop) of Node_Kind :=
3098 (Tok_Asterisk => N_Op_Multiply,
3099 Tok_Mod => N_Op_Mod,
3100 Tok_Rem => N_Op_Rem,
3101 Tok_Slash => N_Op_Divide);
3102 begin
3103 return Mulop_Node (Token);
3104 end P_Multiplying_Operator;
3106 --------------------------------------
3107 -- 4.5 Highest Precedence Operator --
3108 --------------------------------------
3110 -- Parsed by P_Factor (4.4)
3112 -- Note: this rule is not in fact used by the grammar at any point
3114 --------------------------
3115 -- 4.6 Type Conversion --
3116 --------------------------
3118 -- Parsed by P_Primary as a Name (4.1)
3120 -------------------------------
3121 -- 4.7 Qualified Expression --
3122 -------------------------------
3124 -- QUALIFIED_EXPRESSION ::=
3125 -- SUBTYPE_MARK ' (EXPRESSION) | SUBTYPE_MARK ' AGGREGATE
3127 -- The caller has scanned the name which is the Subtype_Mark parameter
3128 -- and scanned past the single quote following the subtype mark. The
3129 -- caller has not checked that this name is in fact appropriate for
3130 -- a subtype mark name (i.e. it is a selected component or identifier).
3132 -- Error_Recovery: cannot raise Error_Resync
3134 function P_Qualified_Expression (Subtype_Mark : Node_Id) return Node_Id is
3135 Qual_Node : Node_Id;
3136 begin
3137 Qual_Node := New_Node (N_Qualified_Expression, Prev_Token_Ptr);
3138 Set_Subtype_Mark (Qual_Node, Check_Subtype_Mark (Subtype_Mark));
3139 Set_Expression (Qual_Node, P_Aggregate_Or_Paren_Expr);
3140 return Qual_Node;
3141 end P_Qualified_Expression;
3143 --------------------
3144 -- 4.8 Allocator --
3145 --------------------
3147 -- ALLOCATOR ::=
3148 -- new [SUBPOOL_SPECIFICATION] SUBTYPE_INDICATION
3149 -- | new [SUBPOOL_SPECIFICATION] QUALIFIED_EXPRESSION
3151 -- SUBPOOL_SPECIFICATION ::= (subpool_handle_NAME)
3153 -- The caller has checked that the initial token is NEW
3155 -- Error recovery: can raise Error_Resync
3157 function P_Allocator return Node_Id is
3158 Alloc_Node : Node_Id;
3159 Type_Node : Node_Id;
3160 Null_Exclusion_Present : Boolean;
3162 begin
3163 Alloc_Node := New_Node (N_Allocator, Token_Ptr);
3164 T_New;
3166 -- Scan subpool_specification if present (Ada 2012 (AI05-0111-3))
3168 -- Scan Null_Exclusion if present (Ada 2005 (AI-231))
3170 if Token = Tok_Left_Paren then
3171 Scan; -- past (
3172 Set_Subpool_Handle_Name (Alloc_Node, P_Name);
3173 T_Right_Paren;
3175 Error_Msg_Ada_2012_Feature
3176 ("|subpool specification",
3177 Sloc (Subpool_Handle_Name (Alloc_Node)));
3178 end if;
3180 Null_Exclusion_Present := P_Null_Exclusion;
3181 Set_Null_Exclusion_Present (Alloc_Node, Null_Exclusion_Present);
3182 Type_Node := P_Subtype_Mark_Resync;
3184 if Token = Tok_Apostrophe then
3185 Scan; -- past apostrophe
3186 Set_Expression (Alloc_Node, P_Qualified_Expression (Type_Node));
3187 else
3188 Set_Expression
3189 (Alloc_Node,
3190 P_Subtype_Indication (Type_Node, Null_Exclusion_Present));
3192 -- AI05-0104: An explicit null exclusion is not allowed for an
3193 -- allocator without initialization. In previous versions of the
3194 -- language it just raises constraint error.
3196 if Ada_Version >= Ada_2012 and then Null_Exclusion_Present then
3197 Error_Msg_N
3198 ("an allocator with a subtype indication "
3199 & "cannot have a null exclusion", Alloc_Node);
3200 end if;
3201 end if;
3203 return Alloc_Node;
3204 end P_Allocator;
3206 -----------------------
3207 -- P_Case_Expression --
3208 -----------------------
3210 function P_Case_Expression return Node_Id is
3211 Loc : constant Source_Ptr := Token_Ptr;
3212 Case_Node : Node_Id;
3213 Save_State : Saved_Scan_State;
3215 begin
3216 Error_Msg_Ada_2012_Feature ("|case expression", Token_Ptr);
3217 Scan; -- past CASE
3218 Case_Node :=
3219 Make_Case_Expression (Loc,
3220 Expression => P_Expression_No_Right_Paren,
3221 Alternatives => New_List);
3222 T_Is;
3224 -- We now have scanned out CASE expression IS, scan alternatives
3226 loop
3227 T_When;
3228 Append_To (Alternatives (Case_Node), P_Case_Expression_Alternative);
3230 -- Missing comma if WHEN (more alternatives present)
3232 if Token = Tok_When then
3233 T_Comma;
3235 -- A semicolon followed by "when" is probably meant to be a comma
3237 elsif Token = Tok_Semicolon then
3238 Save_Scan_State (Save_State);
3239 Scan; -- past the semicolon
3241 if Token /= Tok_When then
3242 Restore_Scan_State (Save_State);
3243 exit;
3244 end if;
3246 Error_Msg_SP -- CODEFIX
3247 ("|"";"" should be "",""");
3249 -- If comma/WHEN, skip comma and we have another alternative
3251 elsif Token = Tok_Comma then
3252 Save_Scan_State (Save_State);
3253 Scan; -- past comma
3255 if Token /= Tok_When then
3256 Restore_Scan_State (Save_State);
3257 exit;
3258 end if;
3260 -- If no comma or WHEN, definitely done
3262 else
3263 exit;
3264 end if;
3265 end loop;
3267 -- If we have an END CASE, diagnose as not needed
3269 if Token = Tok_End then
3270 Error_Msg_SC ("`END CASE` not allowed at end of case expression");
3271 Scan; -- past END
3273 if Token = Tok_Case then
3274 Scan; -- past CASE;
3275 end if;
3276 end if;
3278 -- Return the Case_Expression node
3280 return Case_Node;
3281 end P_Case_Expression;
3283 -----------------------------------
3284 -- P_Case_Expression_Alternative --
3285 -----------------------------------
3287 -- CASE_STATEMENT_ALTERNATIVE ::=
3288 -- when DISCRETE_CHOICE_LIST =>
3289 -- EXPRESSION
3291 -- The caller has checked that and scanned past the initial WHEN token
3292 -- Error recovery: can raise Error_Resync
3294 function P_Case_Expression_Alternative return Node_Id is
3295 Case_Alt_Node : Node_Id;
3296 begin
3297 Case_Alt_Node := New_Node (N_Case_Expression_Alternative, Token_Ptr);
3298 Set_Discrete_Choices (Case_Alt_Node, P_Discrete_Choice_List);
3299 TF_Arrow;
3300 Set_Expression (Case_Alt_Node, P_Expression);
3301 return Case_Alt_Node;
3302 end P_Case_Expression_Alternative;
3304 --------------------------------------
3305 -- P_Iterated_Component_Association --
3306 --------------------------------------
3308 -- ITERATED_COMPONENT_ASSOCIATION ::=
3309 -- for DEFINING_IDENTIFIER in DISCRETE_CHOICE_LIST => EXPRESSION
3311 function P_Iterated_Component_Association return Node_Id is
3312 Assoc_Node : Node_Id;
3314 -- Start of processing for P_Iterated_Component_Association
3316 begin
3317 Scan; -- past FOR
3318 Assoc_Node :=
3319 New_Node (N_Iterated_Component_Association, Prev_Token_Ptr);
3321 Set_Defining_Identifier (Assoc_Node, P_Defining_Identifier);
3322 T_In;
3323 Set_Discrete_Choices (Assoc_Node, P_Discrete_Choice_List);
3324 TF_Arrow;
3325 Set_Expression (Assoc_Node, P_Expression);
3327 if Ada_Version < Ada_2020 then
3328 Error_Msg_SC ("iterated component is an Ada 2020 extension");
3329 Error_Msg_SC ("\compile with -gnatX");
3330 end if;
3332 return Assoc_Node;
3333 end P_Iterated_Component_Association;
3335 ---------------------
3336 -- P_If_Expression --
3337 ---------------------
3339 -- IF_EXPRESSION ::=
3340 -- if CONDITION then DEPENDENT_EXPRESSION
3341 -- {elsif CONDITION then DEPENDENT_EXPRESSION}
3342 -- [else DEPENDENT_EXPRESSION]
3344 -- DEPENDENT_EXPRESSION ::= EXPRESSION
3346 function P_If_Expression return Node_Id is
3347 function P_If_Expression_Internal
3348 (Loc : Source_Ptr;
3349 Cond : Node_Id) return Node_Id;
3350 -- This is the internal recursive routine that does all the work, it is
3351 -- recursive since it is used to process ELSIF parts, which internally
3352 -- are N_If_Expression nodes with the Is_Elsif flag set. The calling
3353 -- sequence is like the outer function except that the caller passes
3354 -- the conditional expression (scanned using P_Expression), and the
3355 -- scan pointer points just past this expression. Loc points to the
3356 -- IF or ELSIF token.
3358 ------------------------------
3359 -- P_If_Expression_Internal --
3360 ------------------------------
3362 function P_If_Expression_Internal
3363 (Loc : Source_Ptr;
3364 Cond : Node_Id) return Node_Id
3366 Exprs : constant List_Id := New_List;
3367 Expr : Node_Id;
3368 State : Saved_Scan_State;
3369 Eptr : Source_Ptr;
3371 begin
3372 -- All cases except where we are at right paren
3374 if Token /= Tok_Right_Paren then
3375 TF_Then;
3376 Append_To (Exprs, P_Condition (Cond));
3377 Append_To (Exprs, P_Expression);
3379 -- Case of right paren (missing THEN phrase). Note that we know this
3380 -- is the IF case, since the caller dealt with this possibility in
3381 -- the ELSIF case.
3383 else
3384 Error_Msg_BC ("missing THEN phrase");
3385 Append_To (Exprs, P_Condition (Cond));
3386 end if;
3388 -- We now have scanned out IF expr THEN expr
3390 -- Check for common error of semicolon before the ELSE
3392 if Token = Tok_Semicolon then
3393 Save_Scan_State (State);
3394 Scan; -- past semicolon
3396 if Token = Tok_Else or else Token = Tok_Elsif then
3397 Error_Msg_SP -- CODEFIX
3398 ("|extra "";"" ignored");
3400 else
3401 Restore_Scan_State (State);
3402 end if;
3403 end if;
3405 -- Scan out ELSIF sequence if present
3407 if Token = Tok_Elsif then
3408 Eptr := Token_Ptr;
3409 Scan; -- past ELSIF
3410 Expr := P_Expression;
3412 -- If we are at a right paren, we assume the ELSIF should be ELSE
3414 if Token = Tok_Right_Paren then
3415 Error_Msg ("ELSIF should be ELSE", Eptr);
3416 Append_To (Exprs, Expr);
3418 -- Otherwise we have an OK ELSIF
3420 else
3421 Expr := P_If_Expression_Internal (Eptr, Expr);
3422 Set_Is_Elsif (Expr);
3423 Append_To (Exprs, Expr);
3424 end if;
3426 -- Scan out ELSE phrase if present
3428 elsif Token = Tok_Else then
3430 -- Scan out ELSE expression
3432 Scan; -- Past ELSE
3433 Append_To (Exprs, P_Expression);
3435 -- Skip redundant ELSE parts
3437 while Token = Tok_Else loop
3438 Error_Msg_SC ("only one ELSE part is allowed");
3439 Scan; -- past ELSE
3440 Discard_Junk_Node (P_Expression);
3441 end loop;
3443 -- Two expression case (implied True, filled in during semantics)
3445 else
3446 null;
3447 end if;
3449 -- If we have an END IF, diagnose as not needed
3451 if Token = Tok_End then
3452 Error_Msg_SC ("`END IF` not allowed at end of if expression");
3453 Scan; -- past END
3455 if Token = Tok_If then
3456 Scan; -- past IF;
3457 end if;
3458 end if;
3460 -- Return the If_Expression node
3462 return Make_If_Expression (Loc, Expressions => Exprs);
3463 end P_If_Expression_Internal;
3465 -- Local variables
3467 Loc : constant Source_Ptr := Token_Ptr;
3468 If_Expr : Node_Id;
3470 -- Start of processing for P_If_Expression
3472 begin
3473 Error_Msg_Ada_2012_Feature ("|if expression", Token_Ptr);
3474 Scan; -- past IF
3475 Inside_If_Expression := Inside_If_Expression + 1;
3476 If_Expr := P_If_Expression_Internal (Loc, P_Expression);
3477 Inside_If_Expression := Inside_If_Expression - 1;
3478 return If_Expr;
3479 end P_If_Expression;
3481 -----------------------
3482 -- P_Membership_Test --
3483 -----------------------
3485 -- MEMBERSHIP_CHOICE_LIST ::= MEMBERHIP_CHOICE {'|' MEMBERSHIP_CHOICE}
3486 -- MEMBERSHIP_CHOICE ::= CHOICE_EXPRESSION | range | subtype_mark
3488 procedure P_Membership_Test (N : Node_Id) is
3489 Alt : constant Node_Id :=
3490 P_Range_Or_Subtype_Mark
3491 (Allow_Simple_Expression => (Ada_Version >= Ada_2012));
3493 begin
3494 -- Set case
3496 if Token = Tok_Vertical_Bar then
3497 Error_Msg_Ada_2012_Feature ("set notation", Token_Ptr);
3498 Set_Alternatives (N, New_List (Alt));
3499 Set_Right_Opnd (N, Empty);
3501 -- Loop to accumulate alternatives
3503 while Token = Tok_Vertical_Bar loop
3504 Scan; -- past vertical bar
3505 Append_To
3506 (Alternatives (N),
3507 P_Range_Or_Subtype_Mark (Allow_Simple_Expression => True));
3508 end loop;
3510 -- Not set case
3512 else
3513 Set_Right_Opnd (N, Alt);
3514 Set_Alternatives (N, No_List);
3515 end if;
3516 end P_Membership_Test;
3518 ------------------------------------------
3519 -- P_Unparen_Cond_Case_Quant_Expression --
3520 ------------------------------------------
3522 function P_Unparen_Cond_Case_Quant_Expression return Node_Id is
3523 Lparen : constant Boolean := Prev_Token = Tok_Left_Paren;
3525 Result : Node_Id;
3526 Scan_State : Saved_Scan_State;
3528 begin
3529 -- Case expression
3531 if Token = Tok_Case then
3532 Result := P_Case_Expression;
3534 if not (Lparen and then Token = Tok_Right_Paren) then
3535 Error_Msg_N ("case expression must be parenthesized!", Result);
3536 end if;
3538 -- If expression
3540 elsif Token = Tok_If then
3541 Result := P_If_Expression;
3543 if not (Lparen and then Token = Tok_Right_Paren) then
3544 Error_Msg_N ("if expression must be parenthesized!", Result);
3545 end if;
3547 -- Quantified expression or iterated component association
3549 elsif Token = Tok_For then
3551 Save_Scan_State (Scan_State);
3552 Scan; -- past FOR
3554 if Token = Tok_All or else Token = Tok_Some then
3555 Restore_Scan_State (Scan_State);
3556 Result := P_Quantified_Expression;
3558 if not (Lparen and then Token = Tok_Right_Paren) then
3559 Error_Msg_N
3560 ("quantified expression must be parenthesized!", Result);
3561 end if;
3563 else
3564 -- If no quantifier keyword, this is an iterated component in
3565 -- an aggregate.
3567 Restore_Scan_State (Scan_State);
3568 Result := P_Iterated_Component_Association;
3569 end if;
3571 -- No other possibility should exist (caller was supposed to check)
3573 else
3574 raise Program_Error;
3575 end if;
3577 -- Return expression (possibly after having given message)
3579 return Result;
3580 end P_Unparen_Cond_Case_Quant_Expression;
3582 end Ch4;