| blob | 4e10dd9049ca5acb08f2590e915021a7b68a657a |
1 ------------------------------------------------------------------------------
2 -- --
3 -- GNAT COMPILER COMPONENTS --
4 -- --
5 -- P A R --
6 -- --
7 -- B o d y --
8 -- --
9 -- Copyright (C) 1992-2023, 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 ------------------------------------------------------------------------------
57 ---------
58 -- Par --
59 ---------
64 -- True within a record definition. Used to control warning for
65 -- redefinition of standard entities (not issued for field names).
68 -- Counter used for constructing loop/block names (see the routine
69 -- Par.Ch5.Get_Loop_Block_Name).
72 -- Count number of units parsed (relevant only in syntax check only mode,
73 -- since in semantics check mode only a single unit is permitted anyway).
76 -- Variable used to save values of config switches while we parse the
77 -- new unit, to be restored on exit for proper recursive behavior.
80 -- True within a delta aggregate (but only after the "delta" token has
81 -- been scanned). Used to distinguish syntax errors from syntactically
82 -- correct "deep" delta aggregates (enabled via -gnatX0).
85 -- Variables for storing the original state of whether style checks should
86 -- be active in general and which particular ones should be checked.
88 --------------------
89 -- Error Recovery --
90 --------------------
92 -- When an error is encountered, a call is made to one of the Error_Msg
93 -- routines to record the error. If the syntax scan is not derailed by the
94 -- error (e.g. a complaint that logical operators are inconsistent in an
95 -- EXPRESSION), then control returns from the Error_Msg call, and the
96 -- parse continues unimpeded.
98 -- If on the other hand, the Error_Msg represents a situation from which
99 -- the parser cannot recover locally, the exception Error_Resync is raised
100 -- immediately after the call to Error_Msg. Handlers for Error_Resync
101 -- are located at strategic points to resynchronize the parse. For example,
102 -- when an error occurs in a statement, the handler skips to the next
103 -- semicolon and continues the scan from there.
105 -- Each parsing procedure contains a note with the heading "Error recovery"
106 -- which shows if it can propagate the Error_Resync exception. In order
107 -- not to propagate the exception, a procedure must either contain its own
108 -- handler for this exception, or it must not call any other routines which
109 -- propagate the exception.
111 -- Note: the arrangement of Error_Resync handlers is such that it should
112 -- never be possible to transfer control through a procedure which made
113 -- an entry in the scope stack, invalidating the contents of the stack.
116 -- Exception raised on error that is not handled locally, see above
119 -- The resynchronization routines in Par.Sync run a risk of getting
120 -- stuck in an infinite loop if they do not skip a token, and the caller
121 -- keeps repeating the same resync call. On the other hand, if they skip
122 -- a token unconditionally, some recovery opportunities are missed. The
123 -- variable Last_Resync_Point records the token location previously set
124 -- by a Resync call, and if a subsequent Resync call occurs at the same
125 -- location, then the Resync routine does guarantee to skip a token.
127 --------------------------------------------
128 -- Handling Semicolon Used in Place of IS --
129 --------------------------------------------
131 -- The following global variables are used in handling the error situation
132 -- of using a semicolon in place of IS in a subprogram declaration as in:
134 -- procedure X (Y : Integer);
135 -- Q : Integer;
136 -- begin
137 -- ...
138 -- end;
140 -- The two contexts in which this can appear are at the outer level, and
141 -- within a declarative region. At the outer level, we know something is
142 -- wrong as soon as we see the Q (or begin, if there are no declarations),
143 -- and we can immediately decide that the semicolon should have been IS.
145 -- The situation in a declarative region is more complex. The declaration
146 -- of Q could belong to the outer region, and we do not know that we have
147 -- an error until we hit the begin. It is still not clear at this point
148 -- from a syntactic point of view that something is wrong, because the
149 -- begin could belong to the enclosing subprogram or package. However, we
150 -- can incorporate a bit of semantic knowledge and note that the body of
151 -- X is missing, so we definitely DO have an error. We diagnose this error
152 -- as semicolon in place of IS on the subprogram line.
154 -- There are two styles for this diagnostic. If the begin immediately
155 -- follows the semicolon, then we can place a flag (IS expected) right
156 -- on the semicolon. Otherwise we do not detect the error until we hit
157 -- the begin which refers back to the line with the semicolon.
159 -- To control the process in the second case, the following global
160 -- variables are set to indicate that we have a subprogram declaration
161 -- whose body is required and has not yet been found. The prefix SIS
162 -- stands for "Subprogram IS" handling.
165 -- Set True to indicate that an entry is active (i.e. that a subprogram
166 -- declaration has been encountered, and no body for this subprogram
167 -- has been encountered). The remaining variables other than
168 -- SIS_Aspect_Import_Seen are valid only if this is True.
171 -- If this is True when a subprogram declaration has been encountered, we
172 -- do not set SIS_Entry_Active, because the Import means there is no body.
173 -- Set False at the start of P_Subprogram, set True when an Import aspect
174 -- specification is seen, and used when P_Subprogram finds a subprogram
175 -- declaration. This is necessary because the aspects are parsed before
176 -- we know we have a subprogram declaration.
179 -- Subprogram designator
182 -- Source location of FUNCTION/PROCEDURE keyword
185 -- Column number of FUNCTION/PROCEDURE keyword
188 -- Source location of semicolon at end of subprogram declaration
191 -- Pointer to tree node for subprogram declaration
194 -- Used to save message ID of missing semicolon message (which will be
195 -- modified to missing IS if necessary). Set to No_Error_Msg in the
196 -- normal (non-error) case.
198 -- Five things can happen to an active SIS entry
200 -- 1. If a BEGIN is encountered with an SIS entry active, then we have
201 -- exactly the situation in which we know the body of the subprogram is
202 -- missing. After posting an error message, we change the spec to a body,
203 -- rechaining the declarations that intervened between the spec and BEGIN.
205 -- 2. Another subprogram declaration or body is encountered. In this
206 -- case the entry gets overwritten with the information for the new
207 -- subprogram declaration. We don't catch some nested cases this way,
208 -- but it doesn't seem worth the effort.
210 -- 3. A nested declarative region (e.g. package declaration or package
211 -- body) is encountered. The SIS active indication is reset at the start
212 -- of such a nested region. Again, like case 2, this causes us to miss
213 -- some nested cases, but it doesn't seen worth the effort to stack and
214 -- unstack the SIS information. Maybe we will reconsider this if we ever
215 -- get a complaint about a missed case.
217 -- 4. We encounter a valid pragma INTERFACE or IMPORT that effectively
218 -- supplies the missing body. In this case we reset the entry.
220 -- 5. We encounter the end of the declarative region without encountering
221 -- a BEGIN first. In this situation we simply reset the entry. We know
222 -- that there is a missing body, but it seems more reasonable to let the
223 -- later semantic checking discover this.
225 ----------------------------------------------------
226 -- Handling of Reserved Words Used as Identifiers --
227 ----------------------------------------------------
229 -- Note: throughout the parser, the terms reserved word and keyword are
230 -- used interchangeably to refer to the same set of reserved keywords
231 -- (including until, protected, etc).
233 -- If a reserved word is used in place of an identifier, the parser where
234 -- possible tries to recover gracefully. In particular, if the keyword is
235 -- clearly spelled using identifier casing, e.g. Until in a source program
236 -- using mixed case identifiers and lower case keywords, then the keyword
237 -- is treated as an identifier if it appears in a place where an identifier
238 -- is required.
240 -- The situation is more complex if the keyword is spelled with normal
241 -- keyword casing. In this case, the parser is more reluctant to consider
242 -- it to be intended as an identifier, unless it has some further
243 -- confirmation.
245 -- In the case of an identifier appearing in the identifier list of a
246 -- declaration, the appearance of a comma or colon right after the keyword
247 -- on the same line is taken as confirmation. For an enumeration literal,
248 -- a comma or right paren right after the identifier is also treated as
249 -- adequate confirmation.
251 -- The following type is used in calls to Is_Reserved_Identifier and
252 -- also to P_Defining_Identifier and P_Identifier. The default for all
253 -- these functions is that reserved words in reserved word case are not
254 -- considered to be reserved identifiers. The Id_Check value indicates
255 -- tokens, which if they appear immediately after the identifier, are
256 -- taken as confirming that the use of an identifier was expected
260 -- Default, no special token test
262 C_Comma_Right_Paren,
263 -- Consider as identifier if followed by comma or right paren
265 C_Comma_Colon,
266 -- Consider as identifier if followed by comma or colon
268 C_Do,
269 -- Consider as identifier if followed by DO
271 C_Dot,
272 -- Consider as identifier if followed by period
274 C_Greater_Greater,
275 -- Consider as identifier if followed by >>
277 C_In,
278 -- Consider as identifier if followed by IN
280 C_Is,
281 -- Consider as identifier if followed by IS
283 C_Left_Paren_Semicolon,
284 -- Consider as identifier if followed by left paren or semicolon
286 C_Use,
287 -- Consider as identifier if followed by USE
289 C_Vertical_Bar_Arrow);
290 -- Consider as identifier if followed by | or =>
292 --------------------------------------------
293 -- Handling IS Used in Place of Semicolon --
294 --------------------------------------------
296 -- This is a somewhat trickier situation, and we can't catch it in all
297 -- cases, but we do our best to detect common situations resulting from
298 -- a "cut and paste" operation which forgets to change the IS to semicolon.
299 -- Consider the following example:
301 -- package body X is
302 -- procedure A;
303 -- procedure B is
304 -- procedure C;
305 -- ...
306 -- procedure D is
307 -- begin
308 -- ...
309 -- end;
310 -- begin
311 -- ...
312 -- end;
314 -- The trouble is that the section of text from PROCEDURE B through END;
315 -- constitutes a valid procedure body, and the danger is that we find out
316 -- far too late that something is wrong (indeed most compilers will behave
317 -- uncomfortably on the above example).
319 -- We have two approaches to helping to control this situation. First we
320 -- make every attempt to avoid swallowing the last END; if we can be sure
321 -- that some error will result from doing so. In particular, we won't
322 -- accept the END; unless it is exactly correct (in particular it must not
323 -- have incorrect name tokens), and we won't accept it if it is immediately
324 -- followed by end of file, WITH or SEPARATE (all tokens that unmistakeably
325 -- signal the start of a compilation unit, and which therefore allow us to
326 -- reserve the END; for the outer level.) For more details on this aspect
327 -- of the handling, see package Par.Endh.
329 -- If we can avoid eating up the END; then the result in the absence of
330 -- any additional steps would be to post a missing END referring back to
331 -- the subprogram with the bogus IS. Similarly, if the enclosing package
332 -- has no BEGIN, then the result is a missing BEGIN message, which again
333 -- refers back to the subprogram header.
335 -- Such an error message is not too bad (it's already a big improvement
336 -- over what many parsers do), but it's not ideal, because the declarations
337 -- following the IS have been absorbed into the wrong scope. In the above
338 -- case, this could result for example in a bogus complaint that the body
339 -- of D was missing from the package.
341 -- To catch at least some of these cases, we take the following additional
342 -- steps. First, a subprogram body is marked as having a suspicious IS if
343 -- the declaration line is followed by a line which starts with a symbol
344 -- that can start a declaration in the same column, or to the left of the
345 -- column in which the FUNCTION or PROCEDURE starts (normal style is to
346 -- indent any declarations which really belong a subprogram). If such a
347 -- subprogram encounters a missing BEGIN or missing END, then we decide
348 -- that the IS should have been a semicolon, and the subprogram body node
349 -- is marked (by setting the Bad_Is_Detected flag true. Note that we do
350 -- not do this for library level procedures, only for nested procedures,
351 -- since for library level procedures, we must have a body.
353 -- The processing for a declarative part checks to see if the last
354 -- declaration scanned is marked in this way, and if it is, the tree
355 -- is modified to reflect the IS being interpreted as a semicolon.
357 ---------------------------------------------------
358 -- Parser Type Definitions and Control Variables --
359 ---------------------------------------------------
361 -- The following variable and associated type declaration are used by the
362 -- expression parsing routines to return more detailed information about
363 -- the categorization of a parsed expression.
374 -- The following type is used by P_Subprogram, P_Package, to indicate which
375 -- of several possibilities is acceptable.
397 -- The above are the only allowed values of Pf_Rec arguments
422 -- List of goto nodes appearing in the current compilation. Used to
423 -- recognize natural loops and convert them into bona fide loops for
424 -- optimization purposes.
427 -- List of label nodes for labels appearing in the current compilation.
428 -- Used by Par.Labl to construct the corresponding implicit declarations.
430 -----------------
431 -- Scope Table --
432 -----------------
434 -- The scope table, also referred to as the scope stack, is used to record
435 -- the current scope context. It is organized as a stack, with inner nested
436 -- entries corresponding to higher entries on the stack. An entry is made
437 -- when the parser encounters the opening of a nested construct (such as a
438 -- record, task, package etc.), and then package Par.Endh uses this stack
439 -- to deal with END lines (including properly dealing with END nesting
440 -- errors).
443 -- Type of end entry required for this scope. The last two entries are
444 -- used only in the subprogram body case to mark the case of a suspicious
445 -- IS, or a bad IS (i.e. suspicions confirmed by missing BEGIN or END).
446 -- See separate section on dealing with IS used in place of semicolon.
447 -- Note that for many purposes E_Name, E_Suspicious_Is and E_Bad_Is are
448 -- treated the same (E_Suspicious_Is and E_Bad_Is are simply special cases
449 -- of E_Name). They are placed at the end of the enumeration so that a
450 -- test for >= E_Name catches all three cases efficiently.
463 -- The following describes a single entry in the scope table
467 -- Type of end entry, as per above description
470 -- A flag indicating whether the label, if present, is required to
471 -- appear on the end line. It is referenced only in the case of Etyp is
472 -- equal to E_Name or E_Suspicious_Is where the name may or may not be
473 -- required (yes for labeled block, no in other cases). Note that for
474 -- all cases except begin, the question of whether a label is required
475 -- can be determined from the other fields (for loop, it is required if
476 -- it is present, and for the other constructs it is never required or
477 -- allowed).
480 -- Contains the absolute column number (with tabs expanded) of the
481 -- expected column of the end assuming normal Ada indentation usage. If
482 -- the RM_Column_Check mode is set, this value is used for generating
483 -- error messages about indentation. Otherwise it is used only to
484 -- control heuristic error recovery actions. This value is zero origin.
487 -- This field is used to provide the name of the construct being parsed
488 -- and indirectly its kind. For loops and blocks, the field contains the
489 -- source name or the generated one. For package specifications, bodies,
490 -- subprogram specifications and bodies the field holds the
491 -- corresponding program unit name. For task declarations and bodies,
492 -- protected types and bodies, and accept statements the field hold the
493 -- name of the type or operation. For if-statements, case-statements,
494 -- return statements, and selects, the field is initialized to Error.
496 -- Note: this is a bit of an odd (mis)use of Error, since there is no
497 -- Error, but we use this value as a place holder to indicate that it
498 -- is an error to have a label on the end line.
500 -- Whenever the field is a name, it is attached to the parent node of
501 -- the construct being parsed. Thus the parent node indicates the kind
502 -- of construct whose parse tree is being built. This is used in error
503 -- recovery.
506 -- Points to the list of declarations (i.e. the declarative part)
507 -- associated with this construct. It is set only in the END [name]
508 -- cases, and is set to No_List for all other cases which do not have a
509 -- declarative unit associated with them. This is used for determining
510 -- the proper location for implicit label declarations.
513 -- Empty except in the case of entries for IF and CASE statements, in
514 -- which case it contains the N_If_Statement or N_Case_Statement node.
515 -- This is used for setting the End_Span field.
518 -- Source location of the opening token of the construct. This is used
519 -- to refer back to this line in error messages (such as missing or
520 -- incorrect end lines). The Sloc field is not used, and is not set, if
521 -- a label is present (the Labl field provides the text name of the
522 -- label in this case, which is fine for error messages).
525 -- S_Is is relevant only if Etyp is set to E_Suspicious_Is or E_Bad_Is.
526 -- It records the location of the IS that is considered to be
527 -- suspicious.
530 -- A boolean flag that is set true if the opening entry is the dubious
531 -- result of some prior error, e.g. a record entry where the record
532 -- keyword was missing. It is used to suppress the issuing of a
533 -- corresponding junk complaint about the end line (we do not want
534 -- to complain about a missing end record when there was no record).
537 -- The following declares the scope table itself. The Last field is the
538 -- stack pointer, so that Scope.Table (Scope.Last) is the top entry. The
539 -- oldest entry, at Scope_Stack (0), is a dummy entry with Etyp set to
540 -- E_Dummy, and the other fields undefined. This dummy entry ensures that
541 -- Scope_Stack (Scope_Stack_Ptr).Etyp can always be tested, and that the
542 -- scope stack pointer is always in range.
555 -- Return the indicated Scope_Table_Entry. We use a pointer for
556 -- efficiency. Callers should not save the pointer, but should do things
557 -- like Scopes (Scope.Last).Something. Note that there is one place in
558 -- Par.Ch5 that indexes the stack out of bounds, and can't call this.
561 begin
566 ------------------------------------------
567 -- Table for Handling Suspicious Labels --
568 ------------------------------------------
570 -- This is a special data structure which is used to deal very specifically
571 -- with the following error case
573 -- label;
574 -- loop
575 -- ...
576 -- end loop label;
578 -- Similar cases apply to FOR, WHILE, DECLARE, or BEGIN
580 -- In each case the opening line looks like a procedure call because of
581 -- the semicolon. And the end line looks illegal because of an unexpected
582 -- label. If we did nothing special, we would just diagnose the label on
583 -- the end as unexpected. But that does not help point to the real error
584 -- which is that the semicolon after label should be a colon.
586 -- To deal with this, we build an entry in the Suspicious_Labels table
587 -- whenever we encounter an identifier followed by a semicolon, followed
588 -- by one of LOOP, FOR, WHILE, DECLARE, BEGIN. Then this entry is used to
589 -- issue the right message when we hit the END that confirms that this was
590 -- a bad label.
594 -- Node for the procedure call statement built for the label; construct
597 -- Location of the possibly wrong semicolon
600 -- Source location of the LOOP, FOR, WHILE, DECLARE, BEGIN token
611 -- Now when we are about to issue a message complaining about an END label
612 -- that should not be there because it appears to end a construct that has
613 -- no label, we first search the suspicious labels table entry, using the
614 -- source location stored in the scope table as a key. If we find a match,
615 -- then we check that the label on the end matches the name in the call,
616 -- and if so, we issue a message saying the semicolon should be a colon.
618 -- Quite a bit of work, but really helpful in the case where it helps, and
619 -- the need for this is based on actual experience with tracking down this
620 -- kind of error (the eye often easily mistakes semicolon for colon).
622 -- Note: we actually have enough information to patch up the tree, but
623 -- this may not be worth the effort. Also we could deal with the same
624 -- situation for EXIT with a label, but for now don't bother with that.
627 -- This is the node of the current assignment statement being compiled.
628 -- It is used to record the presence of target_names on its RHS. This
629 -- context-dependent trick simplifies the analysis of such nodes, where
630 -- the RHS must first be analyzed with expansion disabled.
632 ---------------------------------
633 -- Parsing Routines by Chapter --
634 ---------------------------------
636 -- Uncommented declarations in this section simply parse the construct
637 -- corresponding to their name, and return an ID value for the Node or
638 -- List that is created.
640 -------------
641 -- Par.Ch2 --
642 -------------
646 -- Scan out a pragma. If Skipping is True, then the caller is skipping
647 -- the pragma in the context of illegal placement (this is used to avoid
648 -- some junk cascaded messages). Some pragmas must be dealt with during
649 -- the parsing phase (e.g. pragma Page, since we can generate a listing
650 -- in syntax only mode). It is possible that the parser uses the rescan
651 -- logic (using Save/Restore_Scan_State) with the effect of calling this
652 -- procedure more than once for the same pragma. All parse-time pragma
653 -- handling must be prepared to handle such multiple calls correctly.
655 function P_Identifier
658 -- Scans out an identifier. The parameter C determines the treatment
659 -- of reserved identifiers. See declaration of Id_Check for details.
661 -- An appropriate error message, pointing to the token, is also issued
662 -- if either this is the first occurrence of misuse of this identifier,
663 -- or if Force_Msg is True.
668 -- This function scans for a sequence of pragmas in other than a
669 -- declaration sequence or statement sequence context. All pragmas
670 -- can appear except pragmas Assert and Debug, which are only allowed
671 -- in a declaration or statement sequence context.
674 -- Skips misplaced pragmas with a complaint
677 -- Parses optional pragmas and appends them to the List
680 -------------
681 -- Par.Ch3 --
682 -------------
686 -- This variable is set by a call to P_Declarative_Part. Normally it
687 -- is set to No_Error_Msg, indicating that no special processing is
688 -- required by the caller. The special case arises when a statement
689 -- is found in the sequence of declarations. In this case the Id of
690 -- the message issued ("declaration expected") is preserved in this
691 -- variable, then the caller can change it to an appropriate missing
692 -- begin message if indeed the BEGIN is missing.
709 procedure P_Declarative_Items
714 -- Parses a sequence of zero or more declarative items, and appends them
715 -- to Decls. Done indicates whether or not there might be additional
716 -- declarative items to parse. If Done is True, then there are no more
717 -- to parse; otherwise there might be more.
718 --
719 -- Declare_Expression is true if we are parsing a declare_expression, in
720 -- which case we want to suppress certain style checking.
721 --
722 -- In_Spec is true if we are scanning a package declaration, and is used
723 -- to generate an appropriate message if a statement is encountered in
724 -- such a context.
725 --
726 -- In_Statements is true if we are called to parse declarative items in
727 -- a sequence of statements. In this case, we do not give an error upon
728 -- encountering a statement, but return to the caller with Done = True,
729 -- so the caller can resume parsing statements.
731 function P_Basic_Declarative_Items
733 -- Used to parse the declarative items in a package visible or
734 -- private part (in which case Declare_Expression is False), and
735 -- the declare_items of a declare_expression (in which case
736 -- Declare_Expression is True). Declare_Expression is used to
737 -- affect the wording of error messages, and to control style
738 -- checking.
740 function P_Access_Definition
742 -- Ada 2005 (AI-231/AI-254): The caller parses the null-exclusion part
743 -- and indicates if it was present
745 function P_Access_Type_Definition
747 -- Ada 2005 (AI-254): The formal is used to indicate if the caller has
748 -- parsed the null_exclusion part. In this case the caller has also
749 -- removed the ACCESS token
752 -- Scan out one or more component items and append them to the given
753 -- list. Only scans out more than one declaration in the case where the
754 -- source has a single declaration with multiple defining identifiers.
757 -- Scan out a defining identifier. The parameter C controls the
758 -- treatment of errors in case a reserved word is scanned. See the
759 -- declaration of this type for details.
761 function P_Interface_Type_Definition
763 -- Ada 2005 (AI-251): Parse the interface type definition part. Abstract
764 -- Present indicates if the reserved word "abstract" has been previously
765 -- found. It is used to report an error message because interface types
766 -- are by definition abstract tagged. We generate a record_definition
767 -- node if the list of interfaces is empty; otherwise we generate a
768 -- derived_type_definition node (the first interface in this list is the
769 -- ancestor interface).
771 function P_Null_Exclusion
773 -- Ada 2005 (AI-231): Parse the null-excluding part. A True result
774 -- indicates that the null-excluding part was present.
775 --
776 -- Allow_Anonymous_In_95 is True if we are in a context that allows
777 -- anonymous access types in Ada 95, in which case "not null" is legal
778 -- if it precedes "access".
780 function P_Subtype_Indication
782 -- Ada 2005 (AI-231): The flag Not_Null_Present indicates that the
783 -- null-excluding part has been scanned out and it was present.
785 function P_Range_Or_Subtype_Mark
787 -- Scans out a range or subtype mark, and also permits a general simple
788 -- expression if Allow_Simple_Expression is set to True.
791 -- If an initialization expression is present (:= expression), then
792 -- it is scanned out and returned, otherwise Empty is returned if no
793 -- initialization expression is present. This procedure also handles
794 -- certain common error cases cleanly. The parameter P indicates if
795 -- a right paren can follow the expression (default = no right paren
796 -- allowed).
799 -- Used when scanning statements to skip past a misplaced declaration
800 -- The declaration is scanned out and appended to the given list.
801 -- Token is known to be a declaration token (in Token_Class_Declk)
802 -- on entry, so there definition is a declaration to be scanned.
804 function P_Subtype_Indication
807 -- This version of P_Subtype_Indication is called when the caller has
808 -- already scanned out the subtype mark which is passed as a parameter.
809 -- Ada 2005 (AI-231): The flag Not_Null_Present indicates that the
810 -- null-excluding part has been scanned out and it was present.
813 -- Parse a subtype mark attribute. The caller has already parsed the
814 -- subtype mark, which is passed in as the argument, and has checked
815 -- that the current token is apostrophe.
818 -------------
819 -- Par.Ch4 --
820 -------------
834 -- Scans out an expression allowing an unparenthesized case expression,
835 -- if expression, or quantified expression to appear without enclosing
836 -- parentheses. However, if such an expression is not preceded by a left
837 -- paren, and followed by a right paren, an error message will be output
838 -- noting that parenthesization is required.
841 -- Scans out an expression in contexts where the expression cannot be
842 -- terminated by a right paren (gives better error recovery if an errant
843 -- right paren is found after the expression).
846 -- Scans out an expression or range attribute where a conditional
847 -- expression is permitted to appear without surrounding parentheses.
848 -- However, if such an expression is not preceded by a left paren, and
849 -- followed by a right paren, an error message will be output noting
850 -- that parenthesization is required.
853 -- Scans out an if expression. Called with Token pointing to the
854 -- IF keyword, and returns pointing to the terminating right paren,
855 -- semicolon or comma, but does not consume this terminating token.
858 -- This routine scans out a qualified expression when the caller has
859 -- already scanned out the name and apostrophe of the construct.
862 -- This routine scans out a quantified expression when the caller has
863 -- already scanned out the keyword "for" of the construct.
866 -------------
867 -- Par.Ch5 --
868 -------------
872 -- Scan out and return a condition. Note that an error is given if
873 -- the condition is followed by a right parenthesis.
876 -- Similar to the above, but the caller has already scanned out the
877 -- conditional expression and passes it as an argument. This form of
878 -- the call does not check for a following right parenthesis.
881 -- Parse an iterator specification. The defining identifier has already
882 -- been scanned, as it is the common prefix between loop and iterator
883 -- specification.
886 -- Used in loop constructs and quantified expressions.
888 function P_Sequence_Of_Statements
890 -- SS_Flags indicates the acceptable termination tokens; see body for
891 -- details. Handled is true if we are parsing a handled sequence of
892 -- statements.
895 -- Parses declarations and handled statement sequence, setting
896 -- fields of Parent node appropriately.
899 -------------
900 -- Par.Ch6 --
901 -------------
912 -- Sets In_Present and/or Out_Present flags in Node scanning past IN,
913 -- OUT or IN OUT tokens in the source.
916 -- Scans out any construct starting with either of the keywords
917 -- PROCEDURE or FUNCTION. The parameter indicates which possible
918 -- possible kinds of construct (body, spec, instantiation etc.)
919 -- are permissible in the current context.
922 -------------
923 -- Par.Ch7 --
924 -------------
928 -- Scans out any construct starting with the keyword PACKAGE. The
929 -- parameter indicates which possible kinds of construct (body, spec,
930 -- instantiation etc.) are permissible in the current context.
933 -------------
934 -- Par.Ch8 --
935 -------------
941 -------------
942 -- Par.Ch9 --
943 -------------
958 --------------
959 -- Par.Ch10 --
960 --------------
964 -- Note: this function scans a single compilation unit, and checks that
965 -- an end of file follows this unit, diagnosing any unexpected input as
966 -- an error, and then skipping it, so that Token is set to Tok_EOF on
967 -- return. An exception is in syntax-only mode, where multiple
968 -- compilation units are permitted. In this case, P_Compilation_Unit
969 -- does not check for end of file and there may be more compilation
970 -- units to scan. The caller can uniquely detect this situation by the
971 -- fact that Token is not set to Tok_EOF on return.
972 --
973 -- What about multiple unit/file capability that now exists???
974 --
975 -- The Ignore parameter is normally set False. It is set True in the
976 -- multiple unit per file mode if we are skipping past a unit that we
977 -- are not interested in.
980 --------------
981 -- Par.Ch11 --
982 --------------
990 -- Parses the partial construct EXCEPTION followed by a list of
991 -- exception handlers which appears in a number of productions, and
992 -- returns the list of exception handlers.
995 --------------
996 -- Par.Ch12 --
997 --------------
1004 --------------
1005 -- Par.Ch13 --
1006 --------------
1011 function Aspect_Specifications_Present
1013 -- This function tests whether the next keyword is WITH followed by
1014 -- something that looks reasonably like an aspect specification. If so,
1015 -- True is returned. Otherwise False is returned. In either case control
1016 -- returns with the token pointer unchanged (i.e. pointing to the WITH
1017 -- token in the case where True is returned). This function takes care
1018 -- of generating appropriate messages if aspect specifications appear
1019 -- in versions of Ada prior to Ada 2012. The parameter strict can be
1020 -- set to True, to be rather strict about considering something to be
1021 -- an aspect specification. If Strict is False, then the circuitry is
1022 -- rather more generous in considering something ill-formed to be an
1023 -- attempt at an aspect specification. The default is more strict for
1024 -- Ada versions before Ada 2012 (where aspect specifications are not
1025 -- permitted). Note: this routine never checks the terminator token
1026 -- for aspects so it does not matter whether the aspect specifications
1027 -- are terminated by semicolon or some other character.
1028 --
1029 -- Note: This function also handles the case of WHEN used where WITH
1030 -- was intended, and in that case posts an error and returns True.
1032 procedure P_Aspect_Specifications
1035 -- This procedure scans out a series of aspect specifications. If
1036 -- argument Semicolon is True, a terminating semicolon is also scanned.
1037 -- If this argument is False, the scan pointer is left pointing past the
1038 -- aspects and the caller must check for a proper terminator.
1039 --
1040 -- P_Aspect_Specifications is called with the current token pointing
1041 -- to either a WITH keyword starting an aspect specification, or an
1042 -- instance of what shpould be a terminator token. In the former case,
1043 -- the aspect specifications are scanned out including the terminator
1044 -- token if it is a semicolon, and the Has_Aspect_Specifications
1045 -- flag is set in the given declaration node. A list of aspects
1046 -- is built and stored for this declaration node using a call to
1047 -- Set_Aspect_Specifications. If no WITH keyword is present, then this
1048 -- call has no effect other than scanning out the terminator if it is a
1049 -- semicolon (with the exception that it detects WHEN used in place of
1050 -- WITH).
1052 -- If Decl is Error on entry, any scanned aspect specifications are
1053 -- ignored and a message is output saying aspect specifications not
1054 -- permitted here. If Decl is Empty, then scanned aspect specifications
1055 -- are also ignored, but no error message is given (this is used when
1056 -- the caller has already taken care of the error message).
1058 function Get_Aspect_Specifications
1060 -- Parse a list of aspects but do not attach them to a declaration node.
1061 -- Subsidiary to P_Aspect_Specifications procedure. Used when parsing
1062 -- a subprogram specification that may be a declaration or a body.
1063 -- Semicolon has the same meaning as for P_Aspect_Specifications above.
1066 -- Function to parse a code statement. The caller has scanned out
1067 -- the name to be used as the subtype mark (but has not checked that
1068 -- it is suitable for use as a subtype mark, i.e. is either an
1069 -- identifier or a selected component). The current token is an
1070 -- apostrophe and the following token is either a left paren or
1071 -- RANGE (the latter being an error to be caught by P_Code_Statement.
1074 -- Note: the parsing for annexe J features (i.e. obsolescent features)
1075 -- is found in the logical section where these features would be if
1076 -- they were not obsolescent. In particular:
1078 -- Delta constraint is parsed by P_Delta_Constraint (3.5.9)
1079 -- At clause is parsed by P_At_Clause (13.1)
1080 -- Mod clause is parsed by P_Mod_Clause (13.5.1)
1082 --------------
1083 -- Par.Endh --
1084 --------------
1086 -- Routines for handling end lines, including scope recovery
1089 function Check_End
1092 -- Called when an end sequence is required. In the absence of an error
1093 -- situation, Token contains Tok_End on entry, but in a missing end
1094 -- case, this may not be the case. Pop_End_Context is used to determine
1095 -- the appropriate action to be taken. The returned result is True if
1096 -- an End sequence was encountered and False if no End sequence was
1097 -- present. This occurs if the END keyword encountered was determined
1098 -- to be improper and deleted (i.e. Pop_End_Context set End_Action to
1099 -- Skip_And_Reject). Note that the END sequence includes a semicolon,
1100 -- except in the case of END RECORD, where a semicolon follows the END
1101 -- RECORD, but is not part of the record type definition itself.
1102 --
1103 -- If Decl is non-empty, then aspect specifications are permitted
1104 -- following the end, and Decl is the declaration node with which
1105 -- these aspect specifications are to be associated. If Decl is empty,
1106 -- then aspect specifications are not permitted and will generate an
1107 -- error message.
1108 --
1109 -- Is_Loc is set to other than the default only for the case of a
1110 -- package declaration. It points to the IS keyword of the declaration,
1111 -- and is used to specialize the error messages for misplaced aspect
1112 -- specifications in this case. Note that Decl is always Empty if Is_Loc
1113 -- is set.
1116 -- Skip past an end sequence. On entry Token contains Tok_End, and we
1117 -- we know that the end sequence is syntactically incorrect, and that
1118 -- an appropriate error message has already been posted. The mission
1119 -- is simply to position the scan pointer to be the best guess of the
1120 -- position after the end sequence. We do not issue any additional
1121 -- error messages while carrying this out.
1123 procedure End_Statements
1127 -- Called when an end is required or expected to terminate a sequence
1128 -- of statements. The caller has already made an appropriate entry in
1129 -- the Scope.Table to describe the expected form of the end. This can
1130 -- only be used in cases where the only appropriate terminator is end.
1131 -- If Parent is non-empty, then if a correct END line is encountered,
1132 -- the End_Label field of Parent is set appropriately.
1133 --
1134 -- If Decl is non-null, then it is a declaration node, and aspect
1135 -- specifications are permitted after the end statement. These aspect
1136 -- specifications, if present, are stored in this declaration node.
1137 -- If Decl is null, then aspect specifications are not permitted after
1138 -- the end statement.
1139 --
1140 -- In the case where Decl is null, Is_Sloc determines the handling. If
1141 -- it is set to No_Location, then aspect specifications are ignored and
1142 -- an error message is given. Is_Sloc is used in the package declaration
1143 -- case to point to the IS, and is used to specialize the error emssages
1144 -- issued in this case.
1147 --------------
1148 -- Par.Sync --
1149 --------------
1151 -- These procedures are used to resynchronize after errors. Following an
1152 -- error which is not immediately locally recoverable, the exception
1153 -- Error_Resync is raised. The handler for Error_Resync typically calls
1154 -- one of these recovery procedures to resynchronize the source position
1155 -- to a point from which parsing can be restarted.
1157 -- Note: these procedures output an information message that tokens are
1158 -- being skipped, but this message is output only if the option for
1159 -- Multiple_Errors_Per_Line is set in Options.
1163 -- Used if an error occurs scanning a choice. The scan pointer is
1164 -- advanced to the next vertical bar, arrow, or semicolon, whichever
1165 -- comes first. We also quit if we encounter an end of file.
1168 -- Synchronize to next token which could be the start of a compilation
1169 -- unit, or to the end of file token.
1172 -- Used if an error is detected during the parsing of an expression.
1173 -- It skips past tokens until either a token which cannot be part of
1174 -- an expression is encountered (an expression terminator), or if a
1175 -- comma or right parenthesis or vertical bar is encountered at the
1176 -- current parenthesis level (a parenthesis level counter is maintained
1177 -- to carry out this test).
1180 -- Used when parsing aspect specifications to skip a malformed aspect.
1181 -- The scan pointer is positioned next to a comma, a semicolon or "is"
1182 -- when the aspect applies to a body.
1185 -- Used if an error occurs while scanning a sequence of declarations.
1186 -- The scan pointer is positioned past the next semicolon and the scan
1187 -- resumes. The scan is also resumed on encountering a token which
1188 -- starts a declaration (but we make sure to skip at least one token
1189 -- in this case, to avoid getting stuck in a loop).
1192 -- Used if an error occurs while scanning a sequence of statements. The
1193 -- scan pointer is positioned past the next semicolon, or to the next
1194 -- occurrence of either then or loop, and the scan resumes.
1197 -- Used if an error occurs while scanning a parenthesized list of items
1198 -- separated by semicolons. The scan pointer is advanced to the next
1199 -- semicolon or right parenthesis at the outer parenthesis level, or
1200 -- to the next is or RETURN keyword occurrence, whichever comes first.
1203 -- Similar to Resync_Past_Semicolon, except that the scan pointer is
1204 -- left pointing to the semicolon rather than past it.
1207 -- Used when an error occurs scanning an entry index specification. The
1208 -- scan pointer is positioned to the next WHEN (or to IS or semicolon if
1209 -- either of these appear before WHEN, indicating another error has
1210 -- occurred).
1213 --------------
1214 -- Par.Tchk --
1215 --------------
1217 -- Routines to check for expected tokens
1221 -- Procedures with names of the form T_xxx, where Tok_xxx is a token
1222 -- name, check that the current token matches the required token, and
1223 -- if so, scan past it. If not, an error is issued indicating that
1224 -- the required token is not present (xxx expected). In most cases, the
1225 -- scan pointer is not moved in the not-found case, but there are some
1226 -- exceptions to this, see for example T_Id, where the scan pointer is
1227 -- moved across a literal appearing where an identifier is expected.
1262 -- Procedures having names of the form TF_xxx, where Tok_xxx is a token
1263 -- name check that the current token matches the required token, and
1264 -- if so, scan past it. If not, an error message is issued indicating
1265 -- that the required token is not present (xxx expected).
1267 -- If the missing token is at the end of the line, then control returns
1268 -- immediately after posting the message. If there are remaining tokens
1269 -- on the current line, a search is conducted to see if the token
1270 -- appears later on the current line, as follows:
1272 -- A call to Scan_Save is issued and a forward search for the token
1273 -- is carried out. If the token is found on the current line before a
1274 -- semicolon, then it is scanned out and the scan continues from that
1275 -- point. If not the scan is restored to the point where it was missing.
1285 -- Procedures with names of the form U_xxx, where Tok_xxx is a token
1286 -- name, are just like the corresponding T_xxx procedures except that
1287 -- an error message, if given, is unconditional.
1293 --------------
1294 -- Par.Util --
1295 --------------
1299 -- This function is called in an error situation. It checks if the
1300 -- current token is an identifier whose name is a plausible bad
1301 -- spelling of the given keyword token, and if so, issues an error
1302 -- message, sets Token from T, and returns True. Otherwise Token is
1303 -- unchanged, and False is returned.
1306 -- Check for bad indentation in RM checking mode. Used for statements
1307 -- and declarations. Checks if current token is at start of line and
1308 -- is exdented from the current expected end column, and if so an
1309 -- error message is generated.
1313 -- This is similar to the function above, except that it does not
1314 -- return a result. It is typically used in a situation where any
1315 -- identifier is an error, and it makes sense to simply convert it
1316 -- to the given token if it is a plausible misspelling of it.
1319 -- This routine checks if the token after the current one matches the
1320 -- Next argument. If so, the scan is backed up to the current token
1321 -- and Token_Type is changed to Token_95 after issuing an appropriate
1322 -- error message ("(Ada 83) keyword xx cannot be used"). If not,
1323 -- the scan is backed up with Token_Type unchanged. This routine
1324 -- is used to deal with an attempt to use a 95 keyword in Ada 83
1325 -- mode. The caller has typically checked that the current token,
1326 -- an identifier, matches one of the 95 keywords.
1329 -- Emit a warning if the current token is a valid identifier in the
1330 -- language version in use, but is a reserved word in a later language
1331 -- version (unless the language version in use is Ada 83).
1334 -- Given an expression E, that has just been scanned, so that Expr_Form
1335 -- is still set, outputs an error if E is a non-simple expression. E is
1336 -- not modified by this call.
1339 -- Like Check_Simple_Expression, except that the error message is only
1340 -- given when operating in Ada 83 mode, and includes "in Ada 83".
1343 -- Called to check that a node representing a name (or call) is
1344 -- suitable for a subtype mark, i.e, that it is an identifier or
1345 -- a selected component. If so, or if it is already Error, then
1346 -- it is returned unchanged. Otherwise an error message is issued
1347 -- and Error is returned.
1350 -- Used in comma delimited lists to determine if a comma is present, or
1351 -- can reasonably be assumed to have been present (an error message is
1352 -- generated in the latter case). If True is returned, the scan has been
1353 -- positioned past the comma. If False is returned, the scan position
1354 -- is unchanged. Note that all comma-delimited lists are terminated by
1355 -- a right paren, so the only legitimate tokens when Comma_Present is
1356 -- called are right paren and comma. If some other token is found, then
1357 -- Comma_Present has the job of deciding whether it is better to pretend
1358 -- a comma was present, post a message for a missing comma and return
1359 -- True, or return False and let the caller diagnose the missing right
1360 -- parenthesis.
1366 -- These procedures do nothing at all, their effect is simply to discard
1367 -- the argument. A typical use is to skip by some junk that is not
1368 -- expected in the current context.
1371 -- If current token matches T, then give an error message and skip
1372 -- past it, otherwise the call has no effect at all. T may be any
1373 -- reserved word token, or comma, left or right paren, or semicolon.
1376 -- Test if current token is a reserved identifier. This test is based
1377 -- on the token being a keyword and being spelled in typical identifier
1378 -- style (i.e. starting with an upper case letter). The parameter C
1379 -- determines the special treatment if a reserved word is encountered
1380 -- that has the normal casing of a reserved word.
1383 -- Called when the previous token is an identifier (whose Token_Node
1384 -- value is given by Prev) to check if current token is an identifier
1385 -- that can be merged with the previous one adding an underscore. The
1386 -- merge is only attempted if the following token matches Nxt. If all
1387 -- conditions are met, an error message is issued, and the merge is
1388 -- carried out, modifying the Chars field of Prev.
1391 -- This function deals with the following specialized situations
1392 --
1393 -- when 'x' =>
1394 -- exit/return [identifier]
1395 -- when 'y' =>
1396 --
1397 -- This looks like a messed up EXIT WHEN or RETURN WHEN, when in fact
1398 -- the problem is a missing semicolon. It is called with Token pointing
1399 -- to the WHEN token, and returns True if a semicolon is missing before
1400 -- the WHEN as in the above example.
1403 -- Looks at token after current one and returns True if the token type
1404 -- matches Tok. The scan is unconditionally restored on return.
1407 -- Called in a place where no constraint is allowed, but one might
1408 -- appear due to a common error (e.g. after the type mark in a procedure
1409 -- parameter. If a constraint is present, an error message is posted,
1410 -- and the constraint is scanned and discarded.
1414 -- Push a new entry onto the scope stack. Scope.Last (the stack pointer)
1415 -- is incremented. The Junk field is preinitialized to False. The caller
1416 -- is expected to fill in all remaining entries of the new top stack
1417 -- entry at Scopes (Scope.Last).
1420 -- Pop an entry off the top of the scope stack. Scope_Last (the scope
1421 -- table stack pointer) is decremented by one. It is a fatal error to
1422 -- try to pop off the dummy entry at the bottom of the stack (i.e.
1423 -- Scope.Last must be non-zero at the time of call).
1426 -- Determines if the current token is either Tok_Separate, or an
1427 -- identifier that is a possible misspelling of "separate" followed
1428 -- by a semicolon. True is returned if so, otherwise False.
1431 -- The current token is an identifier that is supposed to be an
1432 -- attribute identifier but is not. This routine posts appropriate
1433 -- error messages, including a check for a near misspelling.
1437 -- Determines if the current token is the first token on the line
1440 -- Determines if the current token is the last token on the line
1443 -- Issues a warning if Warn_On_Standard_Redefinition is set True, and
1444 -- the Node N (which is a Defining_Identifier node with the Chars field
1445 -- set) is a renaming of an entity in package Standard.
1449 --------------
1450 -- Par.Prag --
1451 --------------
1453 -- The processing for pragmas is split off from chapter 2
1456 -- This function is passed a tree for a pragma that has been scanned out.
1457 -- The pragma is syntactically well formed according to the general syntax
1458 -- for pragmas and the pragma identifier is for one of the recognized
1459 -- pragmas. It performs specific syntactic checks for specific pragmas.
1460 -- The result is the input node if it is OK, or Error otherwise. The
1461 -- reason that this is separated out is to facilitate the addition
1462 -- of implementation defined pragmas. The second parameter records the
1463 -- location of the semicolon following the pragma (this is needed for
1464 -- correct processing of the List and Page pragmas). The returned value
1465 -- is a copy of Pragma_Node, or Error if an error is found. Note that
1466 -- at the point where Prag is called, the right paren ending the pragma
1467 -- has been scanned out, and except in the case of pragma Style_Checks,
1468 -- so has the following semicolon. For Style_Checks, the caller delays
1469 -- the scanning of the semicolon so that it will be scanned using the
1470 -- settings from the Style_Checks pragma preceding it.
1472 --------------
1473 -- Par.Labl --
1474 --------------
1477 -- This procedure creates implicit label declarations for all labels that
1478 -- are declared in the current unit. Note that this could conceptually be
1479 -- done at the point where the labels are declared, but it is tricky to do
1480 -- it then, since the tree is not hooked up at the point where the label is
1481 -- declared (e.g. a sequence of statements is not yet attached to its
1482 -- containing scope at the point a label in the sequence is found).
1484 --------------
1485 -- Par.Load --
1486 --------------
1489 -- This procedure loads all subsidiary units that are required by this
1490 -- unit, including with'ed units, specs for bodies, and parents for child
1491 -- units. It does not load bodies for inlined procedures and generics,
1492 -- since we don't know till semantic analysis is complete what is needed.
1494 -----------
1495 -- Stubs --
1496 -----------
1498 -- The package bodies can see all routines defined in all other subpackages
1544 -- Start of processing for Par
1546 begin
1549 -- Deal with configuration pragmas case first
1552 declare
1556 begin
1557 loop
1567 else
1572 -- Give error if bad pragma
1574 if not Is_Configuration_Pragma_Name
1576 and then
1578 then
1580 Error_Msg_N
1583 else
1584 Error_Msg_N
1586 P_Node);
1589 -- Pragma is OK config pragma, so collect it
1591 else
1600 Complete_Source_File_Entry;
1603 -- Normal case of compilation unit
1605 else
1608 -- Store the state of Style_Checks pragamas
1613 -- The following loop runs more than once in syntax check mode
1614 -- where we allow multiple compilation units in the same file
1615 -- and in Multiple_Unit_Per_file mode where we skip units till
1616 -- we get to the unit we want.
1619 Set_Config_Switches
1623 -- Initialize scope table and other parser control variables
1635 -- If in multiple unit per file mode, skip past ignored unit
1639 -- We skip in syntax check only mode, since we don't want to do
1640 -- anything more than skip past the unit and ignore it. This means
1641 -- we skip processing like setting up a unit table entry.
1643 declare
1645 Operating_Mode;
1649 begin
1656 -- If we are at an end of file, and not yet at the right unit,
1657 -- then we have a fatal error. The unit is missing.
1665 -- Here if we are not skipping a file in multiple unit per file mode.
1666 -- Parse the unit that we are interested in. Note that in check
1667 -- syntax mode we are interested in all units in the file.
1669 else
1671 declare
1674 begin
1675 -- If parsing was successful and we are not in check syntax
1676 -- mode, check that language-defined units are compiled in GNAT
1677 -- mode. For this purpose we do NOT consider renamings in annex
1678 -- J as predefined. That allows users to compile their own
1679 -- versions of these files. Another exception is System.RPC
1680 -- and its children. This allows a user to supply their own
1681 -- communication layer.
1682 -- Similarly, we do not generate an error in CodePeer mode,
1683 -- to allow users to analyze third-party compiler packages.
1688 and then not GNAT_Mode
1689 and then not CodePeer_Mode
1690 then
1691 declare
1693 Get_Name_String
1697 -- Because Unit_Name includes "%s"/"%b", we need to strip
1698 -- the last two characters to get the real unit name.
1700 begin
1704 then
1705 Error_Msg
1710 and then
1712 then
1713 Error_Msg
1719 and then
1721 then
1722 Error_Msg
1730 and then
1734 then
1735 Error_Msg
1744 -- All done if at end of file
1748 -- If we are not at an end of file, it means we are in syntax
1749 -- check only mode, and we keep the loop going to parse all
1750 -- remaining units in the file.
1757 -- Restore the state of Style_Checks after parsing the unit to
1758 -- avoid parsed pragmas affecting other units.
1760 Reset_Style_Check_Options;
1764 -- Now that we have completely parsed the source file, we can complete
1765 -- the source file table entry.
1767 Complete_Source_File_Entry;
1769 -- An internal error check, the scope stack should now be empty
1773 -- Here we make the SCO table entries for the main unit
1779 -- Remaining steps are to create implicit label declarations and to load
1780 -- required subsidiary sources. These steps are required only if we are
1781 -- doing semantic checking.
1788 -- Restore settings of switches saved on entry