1 /* Mainly the interface between cpplib and the C front ends.
2 Copyright (C) 1987, 1988, 1989, 1992, 1994, 1995, 1996, 1997
3 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005
4 Free Software Foundation, Inc.
6 This file is part of GCC.
8 GCC is free software; you can redistribute it and/or modify it under
9 the terms of the GNU General Public License as published by the Free
10 Software Foundation; either version 2, or (at your option) any later
13 GCC is distributed in the hope that it will be useful, but WITHOUT ANY
14 WARRANTY; without even the implied warranty of MERCHANTABILITY or
15 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
18 You should have received a copy of the GNU General Public License
19 along with GCC; see the file COPYING. If not, write to the Free
20 Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA
25 #include "coretypes.h"
42 #include "splay-tree.h"
45 /* We may keep statistics about how long which files took to compile. */
46 static int header_time
, body_time
;
47 static splay_tree file_info_tree
;
49 int pending_lang_change
; /* If we need to switch languages - C++ only */
50 int c_header_level
; /* depth in C headers - C++ only */
52 /* If we need to translate characters received. This is tri-state:
53 0 means use only the untranslated string; 1 means use only
54 the translated string; -1 means chain the translated string
55 to the untranslated one. */
56 int c_lex_string_translate
= 1;
58 /* True if strings should be passed to the caller of c_lex completely
59 unmolested (no concatenation, no translation). */
60 bool c_lex_return_raw_strings
= false;
62 static tree
interpret_integer (const cpp_token
*, unsigned int);
63 static tree
interpret_float (const cpp_token
*, unsigned int);
64 static enum integer_type_kind narrowest_unsigned_type
65 (unsigned HOST_WIDE_INT
, unsigned HOST_WIDE_INT
, unsigned int);
66 static enum integer_type_kind narrowest_signed_type
67 (unsigned HOST_WIDE_INT
, unsigned HOST_WIDE_INT
, unsigned int);
68 static enum cpp_ttype
lex_string (const cpp_token
*, tree
*, bool);
69 static tree
lex_charconst (const cpp_token
*);
70 static void update_header_times (const char *);
71 static int dump_one_header (splay_tree_node
, void *);
72 static void cb_line_change (cpp_reader
*, const cpp_token
*, int);
73 static void cb_ident (cpp_reader
*, unsigned int, const cpp_string
*);
74 static void cb_def_pragma (cpp_reader
*, unsigned int);
75 static void cb_define (cpp_reader
*, unsigned int, cpp_hashnode
*);
76 static void cb_undef (cpp_reader
*, unsigned int, cpp_hashnode
*);
81 struct cpp_callbacks
*cb
;
82 struct c_fileinfo
*toplevel
;
84 /* The get_fileinfo data structure must be initialized before
85 cpp_read_main_file is called. */
86 toplevel
= get_fileinfo ("<top level>");
87 if (flag_detailed_statistics
)
90 body_time
= get_run_time ();
91 toplevel
->time
= body_time
;
94 cb
= cpp_get_callbacks (parse_in
);
96 cb
->line_change
= cb_line_change
;
98 cb
->def_pragma
= cb_def_pragma
;
99 cb
->valid_pch
= c_common_valid_pch
;
100 cb
->read_pch
= c_common_read_pch
;
102 /* Set the debug callbacks if we can use them. */
103 if (debug_info_level
== DINFO_LEVEL_VERBOSE
104 && (write_symbols
== DWARF2_DEBUG
105 || write_symbols
== VMS_AND_DWARF2_DEBUG
))
107 cb
->define
= cb_define
;
108 cb
->undef
= cb_undef
;
113 get_fileinfo (const char *name
)
116 struct c_fileinfo
*fi
;
119 file_info_tree
= splay_tree_new ((splay_tree_compare_fn
) strcmp
,
121 (splay_tree_delete_value_fn
) free
);
123 n
= splay_tree_lookup (file_info_tree
, (splay_tree_key
) name
);
125 return (struct c_fileinfo
*) n
->value
;
127 fi
= XNEW (struct c_fileinfo
);
129 fi
->interface_only
= 0;
130 fi
->interface_unknown
= 1;
131 splay_tree_insert (file_info_tree
, (splay_tree_key
) name
,
132 (splay_tree_value
) fi
);
137 update_header_times (const char *name
)
139 /* Changing files again. This means currently collected time
140 is charged against header time, and body time starts back at 0. */
141 if (flag_detailed_statistics
)
143 int this_time
= get_run_time ();
144 struct c_fileinfo
*file
= get_fileinfo (name
);
145 header_time
+= this_time
- body_time
;
146 file
->time
+= this_time
- body_time
;
147 body_time
= this_time
;
152 dump_one_header (splay_tree_node n
, void * ARG_UNUSED (dummy
))
154 print_time ((const char *) n
->key
,
155 ((struct c_fileinfo
*) n
->value
)->time
);
160 dump_time_statistics (void)
162 struct c_fileinfo
*file
= get_fileinfo (input_filename
);
163 int this_time
= get_run_time ();
164 file
->time
+= this_time
- body_time
;
166 fprintf (stderr
, "\n******\n");
167 print_time ("header files (total)", header_time
);
168 print_time ("main file (total)", this_time
- body_time
);
169 fprintf (stderr
, "ratio = %g : 1\n",
170 (double) header_time
/ (double) (this_time
- body_time
));
171 fprintf (stderr
, "\n******\n");
173 splay_tree_foreach (file_info_tree
, dump_one_header
, 0);
177 cb_ident (cpp_reader
* ARG_UNUSED (pfile
),
178 unsigned int ARG_UNUSED (line
),
179 const cpp_string
* ARG_UNUSED (str
))
181 #ifdef ASM_OUTPUT_IDENT
184 /* Convert escapes in the string. */
185 cpp_string cstr
= { 0, 0 };
186 if (cpp_interpret_string (pfile
, str
, 1, &cstr
, false))
188 ASM_OUTPUT_IDENT (asm_out_file
, (const char *) cstr
.text
);
189 free ((void *) cstr
.text
);
195 /* Called at the start of every non-empty line. TOKEN is the first
196 lexed token on the line. Used for diagnostic line numbers. */
198 cb_line_change (cpp_reader
* ARG_UNUSED (pfile
), const cpp_token
*token
,
201 if (token
->type
!= CPP_EOF
&& !parsing_args
)
202 #ifdef USE_MAPPED_LOCATION
203 input_location
= token
->src_loc
;
206 source_location loc
= token
->src_loc
;
207 const struct line_map
*map
= linemap_lookup (&line_table
, loc
);
208 input_line
= SOURCE_LINE (map
, loc
);
214 fe_file_change (const struct line_map
*new_map
)
219 if (new_map
->reason
== LC_ENTER
)
221 /* Don't stack the main buffer on the input stack;
222 we already did in compile_file. */
223 if (!MAIN_FILE_P (new_map
))
225 #ifdef USE_MAPPED_LOCATION
226 int included_at
= LAST_SOURCE_LINE_LOCATION (new_map
- 1);
228 input_location
= included_at
;
229 push_srcloc (new_map
->start_location
);
231 int included_at
= LAST_SOURCE_LINE (new_map
- 1);
233 input_line
= included_at
;
234 push_srcloc (new_map
->to_file
, 1);
236 (*debug_hooks
->start_source_file
) (included_at
, new_map
->to_file
);
237 #ifndef NO_IMPLICIT_EXTERN_C
240 else if (new_map
->sysp
== 2)
243 ++pending_lang_change
;
248 else if (new_map
->reason
== LC_LEAVE
)
250 #ifndef NO_IMPLICIT_EXTERN_C
251 if (c_header_level
&& --c_header_level
== 0)
253 if (new_map
->sysp
== 2)
254 warning (0, "badly nested C headers from preprocessor");
255 --pending_lang_change
;
260 (*debug_hooks
->end_source_file
) (new_map
->to_line
);
263 update_header_times (new_map
->to_file
);
264 in_system_header
= new_map
->sysp
!= 0;
265 #ifdef USE_MAPPED_LOCATION
266 input_location
= new_map
->start_location
;
268 input_filename
= new_map
->to_file
;
269 input_line
= new_map
->to_line
;
274 cb_def_pragma (cpp_reader
*pfile
, source_location loc
)
276 /* Issue a warning message if we have been asked to do so. Ignore
277 unknown pragmas in system headers unless an explicit
278 -Wunknown-pragmas has been given. */
279 if (warn_unknown_pragmas
> in_system_header
)
281 const unsigned char *space
, *name
;
283 #ifndef USE_MAPPED_LOCATION
285 const struct line_map
*map
= linemap_lookup (&line_table
, loc
);
286 fe_loc
.file
= map
->to_file
;
287 fe_loc
.line
= SOURCE_LINE (map
, loc
);
289 location_t fe_loc
= loc
;
292 space
= name
= (const unsigned char *) "";
293 s
= cpp_get_token (pfile
);
294 if (s
->type
!= CPP_EOF
)
296 space
= cpp_token_as_text (pfile
, s
);
297 s
= cpp_get_token (pfile
);
298 if (s
->type
== CPP_NAME
)
299 name
= cpp_token_as_text (pfile
, s
);
302 warning (OPT_Wunknown_pragmas
, "%Hignoring #pragma %s %s",
303 &fe_loc
, space
, name
);
307 /* #define callback for DWARF and DWARF2 debug info. */
309 cb_define (cpp_reader
*pfile
, source_location loc
, cpp_hashnode
*node
)
311 const struct line_map
*map
= linemap_lookup (&line_table
, loc
);
312 (*debug_hooks
->define
) (SOURCE_LINE (map
, loc
),
313 (const char *) cpp_macro_definition (pfile
, node
));
316 /* #undef callback for DWARF and DWARF2 debug info. */
318 cb_undef (cpp_reader
* ARG_UNUSED (pfile
), source_location loc
,
321 const struct line_map
*map
= linemap_lookup (&line_table
, loc
);
322 (*debug_hooks
->undef
) (SOURCE_LINE (map
, loc
),
323 (const char *) NODE_NAME (node
));
326 /* Read a token and return its type. Fill *VALUE with its value, if
327 applicable. Fill *CPP_FLAGS with the token's flags, if it is
331 c_lex_with_flags (tree
*value
, location_t
*loc
, unsigned char *cpp_flags
)
333 static bool no_more_pch
;
334 const cpp_token
*tok
;
337 timevar_push (TV_CPP
);
339 tok
= cpp_get_token (parse_in
);
343 #ifdef USE_MAPPED_LOCATION
346 *loc
= input_location
;
354 *value
= HT_IDENT_TO_GCC_IDENT (HT_NODE (tok
->val
.node
));
359 unsigned int flags
= cpp_classify_number (parse_in
, tok
);
361 switch (flags
& CPP_N_CATEGORY
)
364 /* cpplib has issued an error. */
365 *value
= error_mark_node
;
369 *value
= interpret_integer (tok
, flags
);
373 *value
= interpret_float (tok
, flags
);
383 /* An @ may give the next token special significance in Objective-C. */
384 if (c_dialect_objc ())
386 location_t atloc
= input_location
;
389 tok
= cpp_get_token (parse_in
);
398 type
= lex_string (tok
, value
, true);
402 *value
= HT_IDENT_TO_GCC_IDENT (HT_NODE (tok
->val
.node
));
403 if (objc_is_reserved_word (*value
))
412 error ("%Hstray %<@%> in program", &atloc
);
422 unsigned char name
[4];
424 *cpp_spell_token (parse_in
, tok
, name
, true) = 0;
426 error ("stray %qs in program", name
);
433 cppchar_t c
= tok
->val
.str
.text
[0];
435 if (c
== '"' || c
== '\'')
436 error ("missing terminating %c character", (int) c
);
437 else if (ISGRAPH (c
))
438 error ("stray %qc in program", (int) c
);
440 error ("stray %<\\%o%> in program", (int) c
);
446 *value
= lex_charconst (tok
);
451 if (!c_lex_return_raw_strings
)
453 type
= lex_string (tok
, value
, false);
460 *value
= build_string (tok
->val
.str
.len
, (char *) tok
->val
.str
.text
);
463 /* These tokens should not be visible outside cpplib. */
464 case CPP_HEADER_NAME
:
475 *cpp_flags
= tok
->flags
;
480 c_common_no_more_pch ();
483 timevar_pop (TV_CPP
);
492 return c_lex_with_flags (value
, &loc
, NULL
);
495 /* Returns the narrowest C-visible unsigned type, starting with the
496 minimum specified by FLAGS, that can fit HIGH:LOW, or itk_none if
499 static enum integer_type_kind
500 narrowest_unsigned_type (unsigned HOST_WIDE_INT low
,
501 unsigned HOST_WIDE_INT high
,
504 enum integer_type_kind itk
;
506 if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
)
507 itk
= itk_unsigned_int
;
508 else if ((flags
& CPP_N_WIDTH
) == CPP_N_MEDIUM
)
509 itk
= itk_unsigned_long
;
511 itk
= itk_unsigned_long_long
;
513 for (; itk
< itk_none
; itk
+= 2 /* skip unsigned types */)
515 tree upper
= TYPE_MAX_VALUE (integer_types
[itk
]);
517 if ((unsigned HOST_WIDE_INT
) TREE_INT_CST_HIGH (upper
) > high
518 || ((unsigned HOST_WIDE_INT
) TREE_INT_CST_HIGH (upper
) == high
519 && TREE_INT_CST_LOW (upper
) >= low
))
526 /* Ditto, but narrowest signed type. */
527 static enum integer_type_kind
528 narrowest_signed_type (unsigned HOST_WIDE_INT low
,
529 unsigned HOST_WIDE_INT high
, unsigned int flags
)
531 enum integer_type_kind itk
;
533 if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
)
535 else if ((flags
& CPP_N_WIDTH
) == CPP_N_MEDIUM
)
541 for (; itk
< itk_none
; itk
+= 2 /* skip signed types */)
543 tree upper
= TYPE_MAX_VALUE (integer_types
[itk
]);
545 if ((unsigned HOST_WIDE_INT
) TREE_INT_CST_HIGH (upper
) > high
546 || ((unsigned HOST_WIDE_INT
) TREE_INT_CST_HIGH (upper
) == high
547 && TREE_INT_CST_LOW (upper
) >= low
))
554 /* Interpret TOKEN, an integer with FLAGS as classified by cpplib. */
556 interpret_integer (const cpp_token
*token
, unsigned int flags
)
559 enum integer_type_kind itk
;
561 cpp_options
*options
= cpp_get_options (parse_in
);
563 integer
= cpp_interpret_integer (parse_in
, token
, flags
);
564 integer
= cpp_num_sign_extend (integer
, options
->precision
);
566 /* The type of a constant with a U suffix is straightforward. */
567 if (flags
& CPP_N_UNSIGNED
)
568 itk
= narrowest_unsigned_type (integer
.low
, integer
.high
, flags
);
571 /* The type of a potentially-signed integer constant varies
572 depending on the base it's in, the standard in use, and the
574 enum integer_type_kind itk_u
575 = narrowest_unsigned_type (integer
.low
, integer
.high
, flags
);
576 enum integer_type_kind itk_s
577 = narrowest_signed_type (integer
.low
, integer
.high
, flags
);
579 /* In both C89 and C99, octal and hex constants may be signed or
580 unsigned, whichever fits tighter. We do not warn about this
581 choice differing from the traditional choice, as the constant
582 is probably a bit pattern and either way will work. */
583 if ((flags
& CPP_N_RADIX
) != CPP_N_DECIMAL
)
584 itk
= MIN (itk_u
, itk_s
);
587 /* In C99, decimal constants are always signed.
588 In C89, decimal constants that don't fit in long have
589 undefined behavior; we try to make them unsigned long.
590 In GCC's extended C89, that last is true of decimal
591 constants that don't fit in long long, too. */
594 if (itk_s
> itk_u
&& itk_s
> itk_long
)
598 if (itk_u
< itk_unsigned_long
)
599 itk_u
= itk_unsigned_long
;
601 warning (0, "this decimal constant is unsigned only in ISO C90");
604 warning (OPT_Wtraditional
,
605 "this decimal constant would be unsigned in ISO C90");
611 /* cpplib has already issued a warning for overflow. */
612 type
= ((flags
& CPP_N_UNSIGNED
)
613 ? widest_unsigned_literal_type_node
614 : widest_integer_literal_type_node
);
616 type
= integer_types
[itk
];
618 if (itk
> itk_unsigned_long
619 && (flags
& CPP_N_WIDTH
) != CPP_N_LARGE
620 && !in_system_header
&& !flag_isoc99
)
621 pedwarn ("integer constant is too large for %qs type",
622 (flags
& CPP_N_UNSIGNED
) ? "unsigned long" : "long");
624 value
= build_int_cst_wide (type
, integer
.low
, integer
.high
);
626 /* Convert imaginary to a complex type. */
627 if (flags
& CPP_N_IMAGINARY
)
628 value
= build_complex (NULL_TREE
, build_int_cst (type
, 0), value
);
633 /* Interpret TOKEN, a floating point number with FLAGS as classified
636 interpret_float (const cpp_token
*token
, unsigned int flags
)
640 REAL_VALUE_TYPE real
;
643 const char *type_name
;
645 /* FIXME: make %T work in error/warning, then we don't need type_name. */
646 if ((flags
& CPP_N_WIDTH
) == CPP_N_LARGE
)
648 type
= long_double_type_node
;
649 type_name
= "long double";
651 else if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
652 || flag_single_precision_constant
)
654 type
= float_type_node
;
659 type
= double_type_node
;
660 type_name
= "double";
663 /* Copy the constant to a nul-terminated buffer. If the constant
664 has any suffixes, cut them off; REAL_VALUE_ATOF/ REAL_VALUE_HTOF
665 can't handle them. */
666 copylen
= token
->val
.str
.len
;
667 if ((flags
& CPP_N_WIDTH
) != CPP_N_MEDIUM
)
668 /* Must be an F or L suffix. */
670 if (flags
& CPP_N_IMAGINARY
)
674 copy
= (char *) alloca (copylen
+ 1);
675 memcpy (copy
, token
->val
.str
.text
, copylen
);
676 copy
[copylen
] = '\0';
678 real_from_string (&real
, copy
);
679 real_convert (&real
, TYPE_MODE (type
), &real
);
681 /* Both C and C++ require a diagnostic for a floating constant
682 outside the range of representable values of its type. Since we
683 have __builtin_inf* to produce an infinity, it might now be
684 appropriate for this to be a mandatory pedwarn rather than
685 conditioned on -pedantic. */
686 if (REAL_VALUE_ISINF (real
) && pedantic
)
687 pedwarn ("floating constant exceeds range of %<%s%>", type_name
);
689 /* Create a node with determined type and value. */
690 value
= build_real (type
, real
);
691 if (flags
& CPP_N_IMAGINARY
)
692 value
= build_complex (NULL_TREE
, convert (type
, integer_zero_node
), value
);
697 /* Convert a series of STRING and/or WSTRING tokens into a tree,
698 performing string constant concatenation. TOK is the first of
699 these. VALP is the location to write the string into. OBJC_STRING
700 indicates whether an '@' token preceded the incoming token.
701 Returns the CPP token type of the result (CPP_STRING, CPP_WSTRING,
704 This is unfortunately more work than it should be. If any of the
705 strings in the series has an L prefix, the result is a wide string
706 (6.4.5p4). Whether or not the result is a wide string affects the
707 meaning of octal and hexadecimal escapes (6.4.4.4p6,9). But escape
708 sequences do not continue across the boundary between two strings in
709 a series (6.4.5p7), so we must not lose the boundaries. Therefore
710 cpp_interpret_string takes a vector of cpp_string structures, which
711 we must arrange to provide. */
713 static enum cpp_ttype
714 lex_string (const cpp_token
*tok
, tree
*valp
, bool objc_string
)
719 struct obstack str_ob
;
722 /* Try to avoid the overhead of creating and destroying an obstack
723 for the common case of just one string. */
724 cpp_string str
= tok
->val
.str
;
725 cpp_string
*strs
= &str
;
727 if (tok
->type
== CPP_WSTRING
)
731 tok
= cpp_get_token (parse_in
);
737 if (c_dialect_objc ())
754 gcc_obstack_init (&str_ob
);
755 obstack_grow (&str_ob
, &str
, sizeof (cpp_string
));
759 obstack_grow (&str_ob
, &tok
->val
.str
, sizeof (cpp_string
));
763 /* We have read one more token than we want. */
764 _cpp_backup_tokens (parse_in
, 1);
766 strs
= XOBFINISH (&str_ob
, cpp_string
*);
768 if (concats
&& !objc_string
&& !in_system_header
)
769 warning (OPT_Wtraditional
,
770 "traditional C rejects string constant concatenation");
772 if ((c_lex_string_translate
773 ? cpp_interpret_string
: cpp_interpret_string_notranslate
)
774 (parse_in
, strs
, concats
+ 1, &istr
, wide
))
776 value
= build_string (istr
.len
, (char *) istr
.text
);
777 free ((void *) istr
.text
);
779 if (c_lex_string_translate
== -1)
781 int xlated
= cpp_interpret_string_notranslate (parse_in
, strs
,
784 /* Assume that, if we managed to translate the string above,
785 then the untranslated parsing will always succeed. */
788 if (TREE_STRING_LENGTH (value
) != (int) istr
.len
789 || 0 != strncmp (TREE_STRING_POINTER (value
), (char *) istr
.text
,
792 /* Arrange for us to return the untranslated string in
793 *valp, but to set up the C type of the translated
795 *valp
= build_string (istr
.len
, (char *) istr
.text
);
796 valp
= &TREE_CHAIN (*valp
);
798 free ((void *) istr
.text
);
803 /* Callers cannot generally handle error_mark_node in this context,
804 so return the empty string instead. cpp_interpret_string has
807 value
= build_string (TYPE_PRECISION (wchar_type_node
)
808 / TYPE_PRECISION (char_type_node
),
809 "\0\0\0"); /* widest supported wchar_t
812 value
= build_string (1, "");
815 TREE_TYPE (value
) = wide
? wchar_array_type_node
: char_array_type_node
;
816 *valp
= fix_string_type (value
);
819 obstack_free (&str_ob
, 0);
821 return objc_string
? CPP_OBJC_STRING
: wide
? CPP_WSTRING
: CPP_STRING
;
824 /* Converts a (possibly wide) character constant token into a tree. */
826 lex_charconst (const cpp_token
*token
)
830 unsigned int chars_seen
;
833 result
= cpp_interpret_charconst (parse_in
, token
,
834 &chars_seen
, &unsignedp
);
836 if (token
->type
== CPP_WCHAR
)
837 type
= wchar_type_node
;
838 /* In C, a character constant has type 'int'.
839 In C++ 'char', but multi-char charconsts have type 'int'. */
840 else if (!c_dialect_cxx () || chars_seen
> 1)
841 type
= integer_type_node
;
843 type
= char_type_node
;
845 /* Cast to cppchar_signed_t to get correct sign-extension of RESULT
846 before possibly widening to HOST_WIDE_INT for build_int_cst. */
847 if (unsignedp
|| (cppchar_signed_t
) result
>= 0)
848 value
= build_int_cst_wide (type
, result
, 0);
850 value
= build_int_cst_wide (type
, (cppchar_signed_t
) result
, -1);