1 /* YACC grammar for Modula-2 expressions, for GDB.
2 Copyright (C) 1986-2023 Free Software Foundation, Inc.
3 Generated from expread.y (now c-exp.y) and contributed by the Department
4 of Computer Science at the State University of New York at Buffalo, 1991.
6 This file is part of GDB.
8 This program is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 3 of the License, or
11 (at your option) any later version.
13 This program is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with this program. If not, see <http://www.gnu.org/licenses/>. */
21 /* Parse a Modula-2 expression from text in a string,
22 and return the result as a struct expression pointer.
23 That structure contains arithmetic operations in reverse polish,
24 with constants represented by operations that are followed by special data.
25 See expression.h for the details of the format.
26 What is important here is that it can be built up sequentially
27 during the process of parsing; the lower levels of the tree always
28 come first in the result.
30 Note that malloc's and realloc's in this file are transformed to
31 xmalloc and xrealloc respectively by the same sed command in the
32 makefile that remaps any other malloc/realloc inserted by the parser
33 generator. Doing this with #defines and trying to control the interaction
34 with include files (<malloc.h> and <stdlib.h> for example) just became
35 too messy, particularly when such includes can be inserted at random
36 times by the parser generator. */
41 #include "expression.h"
44 #include "parser-defs.h"
46 #include "bfd.h" /* Required by objfiles.h. */
47 #include "symfile.h" /* Required by objfiles.h. */
48 #include "objfiles.h" /* For have_full_symbols and have_partial_symbols */
52 #define parse_type(ps) builtin_type (ps->gdbarch ())
53 #define parse_m2_type(ps) builtin_m2_type (ps->gdbarch ())
55 /* Remap normal yacc parser interface names (yyparse, yylex, yyerror,
57 #define GDB_YY_REMAP_PREFIX m2_
60 /* The state of the parser, used internally when we are parsing the
63 static struct parser_state
*pstate
= NULL
;
67 static int yylex (void);
69 static void yyerror (const char *);
71 static int parse_number
(int);
73 /* The sign of the number being parsed. */
74 static int number_sign
= 1;
79 /* Although the yacc "value" of an expression is not used,
80 since the result is stored in the structure being created,
81 other node types do have values. */
92 const struct block
*bval
;
93 enum exp_opcode opcode
;
94 struct internalvar
*ivar
;
100 %type
<voidval
> exp type_exp start set
101 %type
<voidval
> variable
106 %token
<lval
> INT HEX ERROR
107 %token
<ulval
> UINT M2_TRUE M2_FALSE CHAR
110 /* Both NAME and TYPENAME tokens represent symbols in the input,
111 and both convey their data as strings.
112 But a TYPENAME is a string that happens to be defined as a typedef
113 or builtin type name (such as int or char)
114 and a NAME is any other symbol.
116 Contexts where this distinction is not important can use the
117 nonterminal "name", which matches either NAME or TYPENAME. */
120 %token
<sval
> NAME BLOCKNAME IDENT VARNAME
121 %token
<sval
> TYPENAME
123 %token SIZE CAP ORD HIGH ABS MIN_FUNC MAX_FUNC FLOAT_FUNC VAL CHR ODD TRUNC
125 %token INC DEC INCL EXCL
127 /* The GDB scope operator */
130 %token
<sval
> DOLLAR_VARIABLE
136 %left
'<' '>' LEQ GEQ
'=' NOTEQUAL
'#' IN
138 %left LOGICAL_AND
'&'
141 %left
'*' '/' DIV MOD
143 %right
'^' DOT
'[' '('
146 /* This is not an actual token ; it is used for precedence.
158 { pstate
->push_new
<type_operation
> ($1); }
163 exp
: exp
'^' %prec UNARY
164 { pstate
->wrap
<unop_ind_operation
> (); }
168 { number_sign
= -1; }
171 pstate
->wrap
<unary_neg_operation
> (); }
174 exp
: '+' exp %prec UNARY
175 { pstate
->wrap
<unary_plus_operation
> (); }
178 exp
: not_exp exp %prec UNARY
179 { pstate
->wrap
<unary_logical_not_operation
> (); }
186 exp
: CAP
'(' exp
')'
187 { error (_
("CAP function is not implemented")); }
190 exp
: ORD
'(' exp
')'
191 { error (_
("ORD function is not implemented")); }
194 exp
: ABS
'(' exp
')'
195 { error (_
("ABS function is not implemented")); }
198 exp
: HIGH
'(' exp
')'
199 { pstate
->wrap
<m2_unop_high_operation
> (); }
202 exp
: MIN_FUNC
'(' type
')'
203 { error (_
("MIN function is not implemented")); }
206 exp
: MAX_FUNC
'(' type
')'
207 { error (_
("MAX function is not implemented")); }
210 exp
: FLOAT_FUNC
'(' exp
')'
211 { error (_
("FLOAT function is not implemented")); }
214 exp
: VAL
'(' type
',' exp
')'
215 { error (_
("VAL function is not implemented")); }
218 exp
: CHR
'(' exp
')'
219 { error (_
("CHR function is not implemented")); }
222 exp
: ODD
'(' exp
')'
223 { error (_
("ODD function is not implemented")); }
226 exp
: TRUNC
'(' exp
')'
227 { error (_
("TRUNC function is not implemented")); }
230 exp
: TSIZE
'(' exp
')'
231 { pstate
->wrap
<unop_sizeof_operation
> (); }
234 exp
: SIZE exp %prec UNARY
235 { pstate
->wrap
<unop_sizeof_operation
> (); }
239 exp
: INC
'(' exp
')'
240 { pstate
->wrap
<preinc_operation
> (); }
243 exp
: INC
'(' exp
',' exp
')'
245 operation_up rhs
= pstate
->pop
();
246 operation_up lhs
= pstate
->pop
();
247 pstate
->push_new
<assign_modify_operation
>
248 (BINOP_ADD
, std
::move
(lhs
), std
::move
(rhs
));
252 exp
: DEC
'(' exp
')'
253 { pstate
->wrap
<predec_operation
> (); }
256 exp
: DEC
'(' exp
',' exp
')'
258 operation_up rhs
= pstate
->pop
();
259 operation_up lhs
= pstate
->pop
();
260 pstate
->push_new
<assign_modify_operation
>
261 (BINOP_SUB
, std
::move
(lhs
), std
::move
(rhs
));
267 pstate
->push_new
<structop_operation
>
268 (pstate
->pop
(), copy_name
($3));
276 { error (_
("Sets are not implemented."));}
279 exp
: INCL
'(' exp
',' exp
')'
280 { error (_
("Sets are not implemented."));}
283 exp
: EXCL
'(' exp
',' exp
')'
284 { error (_
("Sets are not implemented."));}
287 set
: '{' arglist
'}'
288 { error (_
("Sets are not implemented."));}
289 | type
'{' arglist
'}'
290 { error (_
("Sets are not implemented."));}
294 /* Modula-2 array subscript notation [a,b,c...]. */
296 /* This function just saves the number of arguments
297 that follow in the list. It is *not* specific to
299 { pstate
->start_arglist
(); }
300 non_empty_arglist
']' %prec DOT
302 gdb_assert
(pstate
->arglist_len
> 0);
303 std
::vector
<operation_up
> args
304 = pstate
->pop_vector
(pstate
->end_arglist
());
305 pstate
->push_new
<multi_subscript_operation
>
306 (pstate
->pop
(), std
::move
(args
));
311 /* This is to save the value of arglist_len
312 being accumulated by an outer function call. */
313 { pstate
->start_arglist
(); }
314 arglist
')' %prec DOT
316 std
::vector
<operation_up
> args
317 = pstate
->pop_vector
(pstate
->end_arglist
());
318 pstate
->push_new
<funcall_operation
>
319 (pstate
->pop
(), std
::move
(args
));
327 { pstate
->arglist_len
= 1; }
330 arglist
: arglist
',' exp %prec ABOVE_COMMA
331 { pstate
->arglist_len
++; }
336 { pstate
->arglist_len
= 1; }
340 : non_empty_arglist
',' exp %prec ABOVE_COMMA
341 { pstate
->arglist_len
++; }
345 exp
: '{' type
'}' exp %prec UNARY
347 pstate
->push_new
<unop_memval_operation
>
348 (pstate
->pop
(), $2);
352 exp
: type
'(' exp
')' %prec UNARY
354 pstate
->push_new
<unop_cast_operation
>
355 (pstate
->pop
(), $1);
363 /* Binary operators in order of decreasing precedence. Note that some
364 of these operators are overloaded! (ie. sets) */
368 { pstate
->wrap2
<repeat_operation
> (); }
372 { pstate
->wrap2
<mul_operation
> (); }
376 { pstate
->wrap2
<div_operation
> (); }
380 { pstate
->wrap2
<intdiv_operation
> (); }
384 { pstate
->wrap2
<rem_operation
> (); }
388 { pstate
->wrap2
<add_operation
> (); }
392 { pstate
->wrap2
<sub_operation
> (); }
396 { pstate
->wrap2
<equal_operation
> (); }
399 exp
: exp NOTEQUAL exp
400 { pstate
->wrap2
<notequal_operation
> (); }
402 { pstate
->wrap2
<notequal_operation
> (); }
406 { pstate
->wrap2
<leq_operation
> (); }
410 { pstate
->wrap2
<geq_operation
> (); }
414 { pstate
->wrap2
<less_operation
> (); }
418 { pstate
->wrap2
<gtr_operation
> (); }
421 exp
: exp LOGICAL_AND exp
422 { pstate
->wrap2
<logical_and_operation
> (); }
426 { pstate
->wrap2
<logical_or_operation
> (); }
430 { pstate
->wrap2
<assign_operation
> (); }
437 { pstate
->push_new
<bool_operation
> ($1); }
441 { pstate
->push_new
<bool_operation
> ($1); }
446 pstate
->push_new
<long_const_operation
>
447 (parse_m2_type
(pstate
)->builtin_int
, $1);
453 pstate
->push_new
<long_const_operation
>
454 (parse_m2_type
(pstate
)->builtin_card
, $1);
460 pstate
->push_new
<long_const_operation
>
461 (parse_m2_type
(pstate
)->builtin_char
, $1);
469 std
::copy
(std
::begin
($1), std
::end
($1),
471 pstate
->push_new
<float_const_operation
>
472 (parse_m2_type
(pstate
)->builtin_real
, data
);
479 exp
: SIZE
'(' type
')' %prec UNARY
481 pstate
->push_new
<long_const_operation
>
482 (parse_m2_type
(pstate
)->builtin_int
,
488 { error (_
("strings are not implemented")); }
491 /* This will be used for extensions later. Like adding modules. */
493 { $$
= $1->value_block
(); }
498 = lookup_symbol
(copy_name
($1).c_str
(),
499 pstate
->expression_context_block
,
500 VAR_DOMAIN
, 0).symbol
;
505 /* GDB scope operator */
506 fblock
: block COLONCOLON BLOCKNAME
508 = lookup_symbol
(copy_name
($3).c_str
(), $1,
509 VAR_DOMAIN
, 0).symbol
;
510 if
(!tem || tem
->aclass
() != LOC_BLOCK
)
511 error (_
("No function \"%s\" in specified context."),
512 copy_name
($3).c_str
());
517 /* Useful for assigning to PROCEDURE variables */
520 block_symbol sym
{ $1, nullptr
};
521 pstate
->push_new
<var_value_operation
> (sym
);
525 /* GDB internal ($foo) variable */
526 variable: DOLLAR_VARIABLE
527 { pstate
->push_dollar
($1); }
530 /* GDB scope operator */
531 variable: block COLONCOLON NAME
532 { struct block_symbol sym
533 = lookup_symbol
(copy_name
($3).c_str
(), $1,
537 error (_
("No symbol \"%s\" in specified context."),
538 copy_name
($3).c_str
());
539 if
(symbol_read_needs_frame
(sym.symbol
))
540 pstate
->block_tracker
->update
(sym
);
542 pstate
->push_new
<var_value_operation
> (sym
);
546 /* Base case for variables. */
548 { struct block_symbol sym
;
549 struct field_of_this_result is_a_field_of_this
;
551 std
::string name
= copy_name
($1);
553 = lookup_symbol
(name.c_str
(),
554 pstate
->expression_context_block
,
556 &is_a_field_of_this
);
558 pstate
->push_symbol
(name.c_str
(), sym
);
565 = lookup_typename
(pstate
->language
(),
566 copy_name
($1).c_str
(),
567 pstate
->expression_context_block
,
575 /* Take care of parsing a number (anything that starts with a digit).
576 Set yylval and return the token type; update lexptr.
577 LEN is the number of characters in it. */
579 /*** Needs some error checking for the float case ***/
582 parse_number
(int olen
)
584 const char *p
= pstate
->lexptr
;
588 int base
= input_radix
;
596 else if
(p
[len
-1] == 'C' || p
[len
-1] == 'B')
599 ischar
= p
[len
-1] == 'C';
603 /* Scan the number */
604 for
(c
= 0; c
< len
; c
++)
606 if
(p
[c
] == '.' && base
== 10)
608 /* It's a float since it contains a point. */
609 if
(!parse_float
(p
, len
,
610 parse_m2_type
(pstate
)->builtin_real
,
614 pstate
->lexptr
+= len
;
617 if
(p
[c
] == '.' && base
!= 10)
618 error (_
("Floating point numbers must be base 10."));
619 if
(base
== 10 && (p
[c
] < '0' || p
[c
] > '9'))
620 error (_
("Invalid digit \'%c\' in number."),p
[c
]);
627 if
( base
== 8 && (c
== '8' || c
== '9'))
628 error (_
("Invalid digit \'%c\' in octal number."),c
);
629 if
(c
>= '0' && c
<= '9')
633 if
(base
== 16 && c
>= 'A' && c
<= 'F')
641 if
(n
== 0 && prevn
== 0)
643 else if
(RANGE_CHECK
&& prevn
>= n
)
644 range_error
(_
("Overflow on numeric constant."));
650 if
(*p
== 'B' ||
*p
== 'C' ||
*p
== 'H')
651 pstate
->lexptr
++; /* Advance past B,C or H */
659 int int_bits
= gdbarch_int_bit
(pstate
->gdbarch
());
660 bool have_signed
= number_sign
== -1;
661 bool have_unsigned
= number_sign
== 1;
662 if
(have_signed
&& fits_in_type
(number_sign
, n
, int_bits
, true
))
667 else if
(have_unsigned
&& fits_in_type
(number_sign
, n
, int_bits
, false
))
673 error (_
("Overflow on numeric constant."));
685 { {'<', '>'}, NOTEQUAL
},
686 { {':', '='}, ASSIGN
},
689 { {':', ':'}, COLONCOLON
},
693 /* Some specific keywords */
700 static struct keyword keytab
[] =
703 {"IN", IN
},/* Note space after IN */
704 {"AND", LOGICAL_AND
},
722 {"FLOAT", FLOAT_FUNC
},
728 /* Depth of parentheses. */
729 static int paren_depth
;
731 /* Read one token, getting characters through lexptr. */
733 /* This is where we will check to make sure that the language and the
734 operators used are compatible */
742 const char *tokstart
;
747 pstate
->prev_lexptr
= pstate
->lexptr
;
749 tokstart
= pstate
->lexptr
;
752 /* See if it is a special token of length 2 */
753 for
( i
= 0 ; i
< (int) (sizeof tokentab2
/ sizeof tokentab2
[0]) ; i
++)
754 if
(strncmp
(tokentab2
[i
].name
, tokstart
, 2) == 0)
757 return tokentab2
[i
].token
;
760 switch
(c
= *tokstart
)
777 if
(paren_depth
== 0)
784 if
(pstate
->comma_terminates
&& paren_depth
== 0)
790 /* Might be a floating point number. */
791 if
(pstate
->lexptr
[1] >= '0' && pstate
->lexptr
[1] <= '9')
792 break
; /* Falls into number code. */
799 /* These are character tokens that appear as-is in the YACC grammar */
822 for
(namelen
= 1; (c
= tokstart
[namelen
]) != quote
&& c
!= '\0'; namelen
++)
825 c
= tokstart
[++namelen
];
826 if
(c
>= '0' && c
<= '9')
828 c
= tokstart
[++namelen
];
829 if
(c
>= '0' && c
<= '9')
830 c
= tokstart
[++namelen
];
834 error (_
("Unterminated string or character constant."));
835 yylval.sval.ptr
= tokstart
+ 1;
836 yylval.sval.length
= namelen
- 1;
837 pstate
->lexptr
+= namelen
+ 1;
839 if
(namelen
== 2) /* Single character */
841 yylval.ulval
= tokstart
[1];
848 /* Is it a number? */
849 /* Note: We have already dealt with the case of the token '.'.
850 See case '.' above. */
851 if
((c
>= '0' && c
<= '9'))
854 int got_dot
= 0, got_e
= 0;
855 const char *p
= tokstart
;
860 if
(!got_e
&& (*p
== 'e' ||
*p
== 'E'))
862 else if
(!got_dot
&& *p
== '.')
864 else if
(got_e
&& (p
[-1] == 'e' || p
[-1] == 'E')
865 && (*p
== '-' ||
*p
== '+'))
866 /* This is the sign of the exponent, not the end of the
869 else if
((*p
< '0' ||
*p
> '9') &&
870 (*p
< 'A' ||
*p
> 'F') &&
871 (*p
!= 'H')) /* Modula-2 hexadecimal number */
874 toktype
= parse_number
(p
- tokstart
);
875 if
(toktype
== ERROR
)
877 char *err_copy
= (char *) alloca
(p
- tokstart
+ 1);
879 memcpy
(err_copy
, tokstart
, p
- tokstart
);
880 err_copy
[p
- tokstart
] = 0;
881 error (_
("Invalid number \"%s\"."), err_copy
);
887 if
(!(c
== '_' || c
== '$'
888 ||
(c
>= 'a' && c
<= 'z') ||
(c
>= 'A' && c
<= 'Z')))
889 /* We must have come across a bad character (e.g. ';'). */
890 error (_
("Invalid character '%c' in expression."), c
);
892 /* It's a name. See how long it is. */
894 for
(c
= tokstart
[namelen
];
895 (c
== '_' || c
== '$' ||
(c
>= '0' && c
<= '9')
896 ||
(c
>= 'a' && c
<= 'z') ||
(c
>= 'A' && c
<= 'Z'));
897 c
= tokstart
[++namelen
])
900 /* The token "if" terminates the expression and is NOT
901 removed from the input stream. */
902 if
(namelen
== 2 && tokstart
[0] == 'i' && tokstart
[1] == 'f')
907 pstate
->lexptr
+= namelen
;
909 /* Lookup special keywords */
910 for
(i
= 0 ; i
< (int) (sizeof
(keytab
) / sizeof
(keytab
[0])) ; i
++)
911 if
(namelen
== strlen
(keytab
[i
].keyw
)
912 && strncmp
(tokstart
, keytab
[i
].keyw
, namelen
) == 0)
913 return keytab
[i
].token
;
915 yylval.sval.ptr
= tokstart
;
916 yylval.sval.length
= namelen
;
918 if
(*tokstart
== '$')
919 return DOLLAR_VARIABLE
;
921 /* Use token-type BLOCKNAME for symbols that happen to be defined as
922 functions. If this is not so, then ...
923 Use token-type TYPENAME for symbols that happen to be defined
924 currently as names of types; NAME for other symbols.
925 The caller is not constrained to care about the distinction. */
927 std
::string tmp
= copy_name
(yylval.sval
);
930 if
(lookup_symtab
(tmp.c_str
()))
932 sym
= lookup_symbol
(tmp.c_str
(), pstate
->expression_context_block
,
933 VAR_DOMAIN
, 0).symbol
;
934 if
(sym
&& sym
->aclass
() == LOC_BLOCK
)
936 if
(lookup_typename
(pstate
->language
(),
937 tmp.c_str
(), pstate
->expression_context_block
, 1))
942 switch
(sym
->aclass
())
948 case LOC_REGPARM_ADDR
:
951 case LOC_CONST_BYTES
:
952 case LOC_OPTIMIZED_OUT
:
963 error (_
("internal: Undefined class in m2lex()"));
967 error (_
("internal: Unforseen case in m2lex()"));
970 error (_
("unhandled token in m2lex()"));
976 /* Built-in BOOLEAN type. This is sort of a hack. */
977 if
(startswith
(tokstart
, "TRUE"))
982 else if
(startswith
(tokstart
, "FALSE"))
989 /* Must be another type of name... */
995 m2_language::parser
(struct parser_state
*par_state
) const
997 /* Setting up the parser state. */
998 scoped_restore pstate_restore
= make_scoped_restore
(&pstate
);
999 gdb_assert
(par_state
!= NULL
);
1003 int result
= yyparse ();
1005 pstate
->set_operation
(pstate
->pop
());
1010 yyerror (const char *msg
)
1012 if
(pstate
->prev_lexptr
)
1013 pstate
->lexptr
= pstate
->prev_lexptr
;
1015 error (_
("A %s in expression, near `%s'."), msg
, pstate
->lexptr
);