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 Free Software Foundation, Inc.
5 This file is part of GCC.
7 GCC is free software; you can redistribute it and/or modify it under
8 the terms of the GNU General Public License as published by the Free
9 Software Foundation; either version 2, or (at your option) any later
12 GCC is distributed in the hope that it will be useful, but WITHOUT ANY
13 WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
17 You should have received a copy of the GNU General Public License
18 along with GCC; see the file COPYING. If not, write to the Free
19 Software Foundation, 59 Temple Place - Suite 330, Boston, MA
24 #include "coretypes.h"
41 #include "splay-tree.h"
44 /* We may keep statistics about how long which files took to compile. */
45 static int header_time
, body_time
;
46 static splay_tree file_info_tree
;
48 #undef WCHAR_TYPE_SIZE
49 #define WCHAR_TYPE_SIZE TYPE_PRECISION (wchar_type_node)
51 /* Number of bytes in a wide character. */
52 #define WCHAR_BYTES (WCHAR_TYPE_SIZE / BITS_PER_UNIT)
54 int pending_lang_change
; /* If we need to switch languages - C++ only */
55 int c_header_level
; /* depth in C headers - C++ only */
57 /* If we need to translate characters received. This is tri-state:
58 0 means use only the untranslated string; 1 means use only
59 the translated string; -1 means chain the translated string
60 to the untranslated one. */
61 int c_lex_string_translate
= 1;
63 /* True if strings should be passed to the caller of c_lex completely
64 unmolested (no concatenation, no translation). */
65 bool c_lex_return_raw_strings
= false;
67 static tree
interpret_integer (const cpp_token
*, unsigned int);
68 static tree
interpret_float (const cpp_token
*, unsigned int);
69 static enum integer_type_kind narrowest_unsigned_type
70 (unsigned HOST_WIDE_INT
, unsigned HOST_WIDE_INT
, unsigned int);
71 static enum integer_type_kind narrowest_signed_type
72 (unsigned HOST_WIDE_INT
, unsigned HOST_WIDE_INT
, unsigned int);
73 static enum cpp_ttype
lex_string (const cpp_token
*, tree
*, bool);
74 static tree
lex_charconst (const cpp_token
*);
75 static void update_header_times (const char *);
76 static int dump_one_header (splay_tree_node
, void *);
77 static void cb_line_change (cpp_reader
*, const cpp_token
*, int);
78 static void cb_ident (cpp_reader
*, unsigned int, const cpp_string
*);
79 static void cb_def_pragma (cpp_reader
*, unsigned int);
80 static void cb_define (cpp_reader
*, unsigned int, cpp_hashnode
*);
81 static void cb_undef (cpp_reader
*, unsigned int, cpp_hashnode
*);
86 struct cpp_callbacks
*cb
;
87 struct c_fileinfo
*toplevel
;
89 /* The get_fileinfo data structure must be initialized before
90 cpp_read_main_file is called. */
91 toplevel
= get_fileinfo ("<top level>");
92 if (flag_detailed_statistics
)
95 body_time
= get_run_time ();
96 toplevel
->time
= body_time
;
99 cb
= cpp_get_callbacks (parse_in
);
101 cb
->line_change
= cb_line_change
;
102 cb
->ident
= cb_ident
;
103 cb
->def_pragma
= cb_def_pragma
;
104 cb
->valid_pch
= c_common_valid_pch
;
105 cb
->read_pch
= c_common_read_pch
;
107 /* Set the debug callbacks if we can use them. */
108 if (debug_info_level
== DINFO_LEVEL_VERBOSE
109 && (write_symbols
== DWARF2_DEBUG
110 || write_symbols
== VMS_AND_DWARF2_DEBUG
))
112 cb
->define
= cb_define
;
113 cb
->undef
= cb_undef
;
118 get_fileinfo (const char *name
)
121 struct c_fileinfo
*fi
;
124 file_info_tree
= splay_tree_new ((splay_tree_compare_fn
) strcmp
,
126 (splay_tree_delete_value_fn
) free
);
128 n
= splay_tree_lookup (file_info_tree
, (splay_tree_key
) name
);
130 return (struct c_fileinfo
*) n
->value
;
132 fi
= XNEW (struct c_fileinfo
);
134 fi
->interface_only
= 0;
135 fi
->interface_unknown
= 1;
136 splay_tree_insert (file_info_tree
, (splay_tree_key
) name
,
137 (splay_tree_value
) fi
);
142 update_header_times (const char *name
)
144 /* Changing files again. This means currently collected time
145 is charged against header time, and body time starts back at 0. */
146 if (flag_detailed_statistics
)
148 int this_time
= get_run_time ();
149 struct c_fileinfo
*file
= get_fileinfo (name
);
150 header_time
+= this_time
- body_time
;
151 file
->time
+= this_time
- body_time
;
152 body_time
= this_time
;
157 dump_one_header (splay_tree_node n
, void * ARG_UNUSED (dummy
))
159 print_time ((const char *) n
->key
,
160 ((struct c_fileinfo
*) n
->value
)->time
);
165 dump_time_statistics (void)
167 struct c_fileinfo
*file
= get_fileinfo (input_filename
);
168 int this_time
= get_run_time ();
169 file
->time
+= this_time
- body_time
;
171 fprintf (stderr
, "\n******\n");
172 print_time ("header files (total)", header_time
);
173 print_time ("main file (total)", this_time
- body_time
);
174 fprintf (stderr
, "ratio = %g : 1\n",
175 (double) header_time
/ (double) (this_time
- body_time
));
176 fprintf (stderr
, "\n******\n");
178 splay_tree_foreach (file_info_tree
, dump_one_header
, 0);
182 cb_ident (cpp_reader
* ARG_UNUSED (pfile
),
183 unsigned int ARG_UNUSED (line
),
184 const cpp_string
* ARG_UNUSED (str
))
186 #ifdef ASM_OUTPUT_IDENT
189 /* Convert escapes in the string. */
190 cpp_string cstr
= { 0, 0 };
191 if (cpp_interpret_string (pfile
, str
, 1, &cstr
, false))
193 ASM_OUTPUT_IDENT (asm_out_file
, (const char *) cstr
.text
);
194 free ((void *) cstr
.text
);
200 /* Called at the start of every non-empty line. TOKEN is the first
201 lexed token on the line. Used for diagnostic line numbers. */
203 cb_line_change (cpp_reader
* ARG_UNUSED (pfile
), const cpp_token
*token
,
206 if (token
->type
!= CPP_EOF
&& !parsing_args
)
207 #ifdef USE_MAPPED_LOCATION
208 input_location
= token
->src_loc
;
211 source_location loc
= token
->src_loc
;
212 const struct line_map
*map
= linemap_lookup (&line_table
, loc
);
213 input_line
= SOURCE_LINE (map
, loc
);
219 fe_file_change (const struct line_map
*new_map
)
224 if (new_map
->reason
== LC_ENTER
)
226 /* Don't stack the main buffer on the input stack;
227 we already did in compile_file. */
228 if (!MAIN_FILE_P (new_map
))
230 #ifdef USE_MAPPED_LOCATION
231 int included_at
= LAST_SOURCE_LINE_LOCATION (new_map
- 1);
233 input_location
= included_at
;
234 push_srcloc (new_map
->start_location
);
236 int included_at
= LAST_SOURCE_LINE (new_map
- 1);
238 input_line
= included_at
;
239 push_srcloc (new_map
->to_file
, 1);
241 (*debug_hooks
->start_source_file
) (included_at
, new_map
->to_file
);
242 #ifndef NO_IMPLICIT_EXTERN_C
245 else if (new_map
->sysp
== 2)
248 ++pending_lang_change
;
253 else if (new_map
->reason
== LC_LEAVE
)
255 #ifndef NO_IMPLICIT_EXTERN_C
256 if (c_header_level
&& --c_header_level
== 0)
258 if (new_map
->sysp
== 2)
259 warning ("badly nested C headers from preprocessor");
260 --pending_lang_change
;
265 (*debug_hooks
->end_source_file
) (new_map
->to_line
);
268 update_header_times (new_map
->to_file
);
269 in_system_header
= new_map
->sysp
!= 0;
270 #ifdef USE_MAPPED_LOCATION
271 input_location
= new_map
->start_location
;
273 input_filename
= new_map
->to_file
;
274 input_line
= new_map
->to_line
;
279 cb_def_pragma (cpp_reader
*pfile
, source_location loc
)
281 /* Issue a warning message if we have been asked to do so. Ignore
282 unknown pragmas in system headers unless an explicit
283 -Wunknown-pragmas has been given. */
284 if (warn_unknown_pragmas
> in_system_header
)
286 #ifndef USE_MAPPED_LOCATION
287 const struct line_map
*map
= linemap_lookup (&line_table
, loc
);
289 const unsigned char *space
, *name
;
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 #ifdef USE_MAPPED_LOCATION
303 input_location
= loc
;
305 input_line
= SOURCE_LINE (map
, loc
);
307 warning ("ignoring #pragma %s %s", space
, name
);
311 /* #define callback for DWARF and DWARF2 debug info. */
313 cb_define (cpp_reader
*pfile
, source_location loc
, cpp_hashnode
*node
)
315 const struct line_map
*map
= linemap_lookup (&line_table
, loc
);
316 (*debug_hooks
->define
) (SOURCE_LINE (map
, loc
),
317 (const char *) cpp_macro_definition (pfile
, node
));
320 /* #undef callback for DWARF and DWARF2 debug info. */
322 cb_undef (cpp_reader
* ARG_UNUSED (pfile
), source_location loc
,
325 const struct line_map
*map
= linemap_lookup (&line_table
, loc
);
326 (*debug_hooks
->undef
) (SOURCE_LINE (map
, loc
),
327 (const char *) NODE_NAME (node
));
330 /* Read a token and return its type. Fill *VALUE with its value, if
331 applicable. Fill *CPP_FLAGS with the token's flags, if it is
335 c_lex_with_flags (tree
*value
, unsigned char *cpp_flags
)
337 static bool no_more_pch
;
338 const cpp_token
*tok
;
341 timevar_push (TV_CPP
);
343 tok
= cpp_get_token (parse_in
);
353 *value
= HT_IDENT_TO_GCC_IDENT (HT_NODE (tok
->val
.node
));
358 unsigned int flags
= cpp_classify_number (parse_in
, tok
);
360 switch (flags
& CPP_N_CATEGORY
)
363 /* cpplib has issued an error. */
364 *value
= error_mark_node
;
368 *value
= interpret_integer (tok
, flags
);
372 *value
= interpret_float (tok
, flags
);
382 /* An @ may give the next token special significance in Objective-C. */
383 if (c_dialect_objc ())
385 location_t atloc
= input_location
;
388 tok
= cpp_get_token (parse_in
);
397 type
= lex_string (tok
, value
, true);
401 *value
= HT_IDENT_TO_GCC_IDENT (HT_NODE (tok
->val
.node
));
402 if (objc_is_reserved_word (*value
))
411 error ("%Hstray %<@%> in program", &atloc
);
421 unsigned char name
[4];
423 *cpp_spell_token (parse_in
, tok
, name
) = 0;
425 error ("stray %qs in program", name
);
432 cppchar_t c
= tok
->val
.str
.text
[0];
434 if (c
== '"' || c
== '\'')
435 error ("missing terminating %c character", (int) c
);
436 else if (ISGRAPH (c
))
437 error ("stray %qc in program", (int) c
);
439 error ("stray %<\\%o%> in program", (int) c
);
445 *value
= lex_charconst (tok
);
450 if (!c_lex_return_raw_strings
)
452 type
= lex_string (tok
, value
, false);
459 *value
= build_string (tok
->val
.str
.len
, (char *) tok
->val
.str
.text
);
462 /* These tokens should not be visible outside cpplib. */
463 case CPP_HEADER_NAME
:
474 *cpp_flags
= tok
->flags
;
479 c_common_no_more_pch ();
482 timevar_pop (TV_CPP
);
490 return c_lex_with_flags (value
, NULL
);
493 /* Returns the narrowest C-visible unsigned type, starting with the
494 minimum specified by FLAGS, that can fit HIGH:LOW, or itk_none if
497 static enum integer_type_kind
498 narrowest_unsigned_type (unsigned HOST_WIDE_INT low
,
499 unsigned HOST_WIDE_INT high
,
502 enum integer_type_kind itk
;
504 if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
)
505 itk
= itk_unsigned_int
;
506 else if ((flags
& CPP_N_WIDTH
) == CPP_N_MEDIUM
)
507 itk
= itk_unsigned_long
;
509 itk
= itk_unsigned_long_long
;
511 for (; itk
< itk_none
; itk
+= 2 /* skip unsigned types */)
513 tree upper
= TYPE_MAX_VALUE (integer_types
[itk
]);
515 if ((unsigned HOST_WIDE_INT
) TREE_INT_CST_HIGH (upper
) > high
516 || ((unsigned HOST_WIDE_INT
) TREE_INT_CST_HIGH (upper
) == high
517 && TREE_INT_CST_LOW (upper
) >= low
))
524 /* Ditto, but narrowest signed type. */
525 static enum integer_type_kind
526 narrowest_signed_type (unsigned HOST_WIDE_INT low
,
527 unsigned HOST_WIDE_INT high
, unsigned int flags
)
529 enum integer_type_kind itk
;
531 if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
)
533 else if ((flags
& CPP_N_WIDTH
) == CPP_N_MEDIUM
)
539 for (; itk
< itk_none
; itk
+= 2 /* skip signed types */)
541 tree upper
= TYPE_MAX_VALUE (integer_types
[itk
]);
543 if ((unsigned HOST_WIDE_INT
) TREE_INT_CST_HIGH (upper
) > high
544 || ((unsigned HOST_WIDE_INT
) TREE_INT_CST_HIGH (upper
) == high
545 && TREE_INT_CST_LOW (upper
) >= low
))
552 /* Interpret TOKEN, an integer with FLAGS as classified by cpplib. */
554 interpret_integer (const cpp_token
*token
, unsigned int flags
)
557 enum integer_type_kind itk
;
559 cpp_options
*options
= cpp_get_options (parse_in
);
561 integer
= cpp_interpret_integer (parse_in
, token
, flags
);
562 integer
= cpp_num_sign_extend (integer
, options
->precision
);
564 /* The type of a constant with a U suffix is straightforward. */
565 if (flags
& CPP_N_UNSIGNED
)
566 itk
= narrowest_unsigned_type (integer
.low
, integer
.high
, flags
);
569 /* The type of a potentially-signed integer constant varies
570 depending on the base it's in, the standard in use, and the
572 enum integer_type_kind itk_u
573 = narrowest_unsigned_type (integer
.low
, integer
.high
, flags
);
574 enum integer_type_kind itk_s
575 = narrowest_signed_type (integer
.low
, integer
.high
, flags
);
577 /* In both C89 and C99, octal and hex constants may be signed or
578 unsigned, whichever fits tighter. We do not warn about this
579 choice differing from the traditional choice, as the constant
580 is probably a bit pattern and either way will work. */
581 if ((flags
& CPP_N_RADIX
) != CPP_N_DECIMAL
)
582 itk
= MIN (itk_u
, itk_s
);
585 /* In C99, decimal constants are always signed.
586 In C89, decimal constants that don't fit in long have
587 undefined behavior; we try to make them unsigned long.
588 In GCC's extended C89, that last is true of decimal
589 constants that don't fit in long long, too. */
592 if (itk_s
> itk_u
&& itk_s
> itk_long
)
596 if (itk_u
< itk_unsigned_long
)
597 itk_u
= itk_unsigned_long
;
599 warning ("this decimal constant is unsigned only in ISO C90");
601 else if (warn_traditional
)
602 warning ("this decimal constant would be unsigned in ISO C90");
608 /* cpplib has already issued a warning for overflow. */
609 type
= ((flags
& CPP_N_UNSIGNED
)
610 ? widest_unsigned_literal_type_node
611 : widest_integer_literal_type_node
);
613 type
= integer_types
[itk
];
615 if (itk
> itk_unsigned_long
616 && (flags
& CPP_N_WIDTH
) != CPP_N_LARGE
617 && !in_system_header
&& !flag_isoc99
)
618 pedwarn ("integer constant is too large for %qs type",
619 (flags
& CPP_N_UNSIGNED
) ? "unsigned long" : "long");
621 value
= build_int_cst_wide (type
, integer
.low
, integer
.high
);
623 /* Convert imaginary to a complex type. */
624 if (flags
& CPP_N_IMAGINARY
)
625 value
= build_complex (NULL_TREE
, build_int_cst (type
, 0), value
);
630 /* Interpret TOKEN, a floating point number with FLAGS as classified
633 interpret_float (const cpp_token
*token
, unsigned int flags
)
637 REAL_VALUE_TYPE real
;
640 const char *type_name
;
642 /* FIXME: make %T work in error/warning, then we don't need type_name. */
643 if ((flags
& CPP_N_WIDTH
) == CPP_N_LARGE
)
645 type
= long_double_type_node
;
646 type_name
= "long double";
648 else if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
649 || flag_single_precision_constant
)
651 type
= float_type_node
;
656 type
= double_type_node
;
657 type_name
= "double";
660 /* Copy the constant to a nul-terminated buffer. If the constant
661 has any suffixes, cut them off; REAL_VALUE_ATOF/ REAL_VALUE_HTOF
662 can't handle them. */
663 copylen
= token
->val
.str
.len
;
664 if ((flags
& CPP_N_WIDTH
) != CPP_N_MEDIUM
)
665 /* Must be an F or L suffix. */
667 if (flags
& CPP_N_IMAGINARY
)
671 copy
= (char *) alloca (copylen
+ 1);
672 memcpy (copy
, token
->val
.str
.text
, copylen
);
673 copy
[copylen
] = '\0';
675 real_from_string (&real
, copy
);
676 real_convert (&real
, TYPE_MODE (type
), &real
);
678 /* Both C and C++ require a diagnostic for a floating constant
679 outside the range of representable values of its type. Since we
680 have __builtin_inf* to produce an infinity, it might now be
681 appropriate for this to be a mandatory pedwarn rather than
682 conditioned on -pedantic. */
683 if (REAL_VALUE_ISINF (real
) && pedantic
)
684 pedwarn ("floating constant exceeds range of %<%s%>", type_name
);
686 /* Create a node with determined type and value. */
687 value
= build_real (type
, real
);
688 if (flags
& CPP_N_IMAGINARY
)
689 value
= build_complex (NULL_TREE
, convert (type
, integer_zero_node
), value
);
694 /* Convert a series of STRING and/or WSTRING tokens into a tree,
695 performing string constant concatenation. TOK is the first of
696 these. VALP is the location to write the string into. OBJC_STRING
697 indicates whether an '@' token preceded the incoming token.
698 Returns the CPP token type of the result (CPP_STRING, CPP_WSTRING,
701 This is unfortunately more work than it should be. If any of the
702 strings in the series has an L prefix, the result is a wide string
703 (6.4.5p4). Whether or not the result is a wide string affects the
704 meaning of octal and hexadecimal escapes (6.4.4.4p6,9). But escape
705 sequences do not continue across the boundary between two strings in
706 a series (6.4.5p7), so we must not lose the boundaries. Therefore
707 cpp_interpret_string takes a vector of cpp_string structures, which
708 we must arrange to provide. */
710 static enum cpp_ttype
711 lex_string (const cpp_token
*tok
, tree
*valp
, bool objc_string
)
716 struct obstack str_ob
;
719 /* Try to avoid the overhead of creating and destroying an obstack
720 for the common case of just one string. */
721 cpp_string str
= tok
->val
.str
;
722 cpp_string
*strs
= &str
;
724 if (tok
->type
== CPP_WSTRING
)
728 tok
= cpp_get_token (parse_in
);
734 if (c_dialect_objc ())
751 gcc_obstack_init (&str_ob
);
752 obstack_grow (&str_ob
, &str
, sizeof (cpp_string
));
756 obstack_grow (&str_ob
, &tok
->val
.str
, sizeof (cpp_string
));
760 /* We have read one more token than we want. */
761 _cpp_backup_tokens (parse_in
, 1);
763 strs
= (cpp_string
*) obstack_finish (&str_ob
);
765 if (concats
&& !objc_string
&& warn_traditional
&& !in_system_header
)
766 warning ("traditional C rejects string constant concatenation");
768 if ((c_lex_string_translate
769 ? cpp_interpret_string
: cpp_interpret_string_notranslate
)
770 (parse_in
, strs
, concats
+ 1, &istr
, wide
))
772 value
= build_string (istr
.len
, (char *) istr
.text
);
773 free ((void *) istr
.text
);
775 if (c_lex_string_translate
== -1)
777 int xlated
= cpp_interpret_string_notranslate (parse_in
, strs
,
780 /* Assume that, if we managed to translate the string above,
781 then the untranslated parsing will always succeed. */
784 if (TREE_STRING_LENGTH (value
) != (int) istr
.len
785 || 0 != strncmp (TREE_STRING_POINTER (value
), (char *) istr
.text
,
788 /* Arrange for us to return the untranslated string in
789 *valp, but to set up the C type of the translated
791 *valp
= build_string (istr
.len
, (char *) istr
.text
);
792 valp
= &TREE_CHAIN (*valp
);
794 free ((void *) istr
.text
);
799 /* Callers cannot generally handle error_mark_node in this context,
800 so return the empty string instead. cpp_interpret_string has
803 value
= build_string (TYPE_PRECISION (wchar_type_node
)
804 / TYPE_PRECISION (char_type_node
),
805 "\0\0\0"); /* widest supported wchar_t
808 value
= build_string (1, "");
811 TREE_TYPE (value
) = wide
? wchar_array_type_node
: char_array_type_node
;
812 *valp
= fix_string_type (value
);
815 obstack_free (&str_ob
, 0);
817 return objc_string
? CPP_OBJC_STRING
: wide
? CPP_WSTRING
: CPP_STRING
;
820 /* Converts a (possibly wide) character constant token into a tree. */
822 lex_charconst (const cpp_token
*token
)
826 unsigned int chars_seen
;
829 result
= cpp_interpret_charconst (parse_in
, token
,
830 &chars_seen
, &unsignedp
);
832 if (token
->type
== CPP_WCHAR
)
833 type
= wchar_type_node
;
834 /* In C, a character constant has type 'int'.
835 In C++ 'char', but multi-char charconsts have type 'int'. */
836 else if (!c_dialect_cxx () || chars_seen
> 1)
837 type
= integer_type_node
;
839 type
= char_type_node
;
841 /* Cast to cppchar_signed_t to get correct sign-extension of RESULT
842 before possibly widening to HOST_WIDE_INT for build_int_cst. */
843 if (unsignedp
|| (cppchar_signed_t
) result
>= 0)
844 value
= build_int_cst_wide (type
, result
, 0);
846 value
= build_int_cst_wide (type
, (cppchar_signed_t
) result
, -1);