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 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
40 #include "splay-tree.h"
43 #ifdef MULTIBYTE_CHARS
46 #endif /* MULTIBYTE_CHARS */
47 #ifndef GET_ENVIRONMENT
48 #define GET_ENVIRONMENT(ENV_VALUE,ENV_NAME) ((ENV_VALUE) = getenv (ENV_NAME))
51 /* The current line map. */
52 static const struct line_map
*map
;
54 /* The line used to refresh the lineno global variable after each token. */
55 static unsigned int src_lineno
;
57 /* We may keep statistics about how long which files took to compile. */
58 static int header_time
, body_time
;
59 static splay_tree file_info_tree
;
61 /* File used for outputting assembler code. */
62 extern FILE *asm_out_file
;
64 #undef WCHAR_TYPE_SIZE
65 #define WCHAR_TYPE_SIZE TYPE_PRECISION (wchar_type_node)
67 /* Number of bytes in a wide character. */
68 #define WCHAR_BYTES (WCHAR_TYPE_SIZE / BITS_PER_UNIT)
70 int pending_lang_change
; /* If we need to switch languages - C++ only */
71 int c_header_level
; /* depth in C headers - C++ only */
73 /* Nonzero tells yylex to ignore \ in string constants. */
74 static int ignore_escape_flag
;
76 static tree interpret_integer
PARAMS ((const cpp_token
*, unsigned int));
77 static tree interpret_float
PARAMS ((const cpp_token
*, unsigned int));
78 static enum integer_type_kind
79 narrowest_unsigned_type
PARAMS ((tree
, unsigned int));
80 static enum integer_type_kind
81 narrowest_signed_type
PARAMS ((tree
, unsigned int));
82 static tree lex_string
PARAMS ((const unsigned char *, unsigned int,
84 static tree lex_charconst
PARAMS ((const cpp_token
*));
85 static void update_header_times
PARAMS ((const char *));
86 static int dump_one_header
PARAMS ((splay_tree_node
, void *));
87 static void cb_line_change
PARAMS ((cpp_reader
*, const cpp_token
*, int));
88 static void cb_ident
PARAMS ((cpp_reader
*, unsigned int,
90 static void cb_file_change
PARAMS ((cpp_reader
*, const struct line_map
*));
91 static void cb_def_pragma
PARAMS ((cpp_reader
*, unsigned int));
92 static void cb_define
PARAMS ((cpp_reader
*, unsigned int,
94 static void cb_undef
PARAMS ((cpp_reader
*, unsigned int,
101 struct cpp_callbacks
*cb
;
102 struct c_fileinfo
*toplevel
;
104 /* Set up filename timing. Must happen before cpp_read_main_file. */
105 file_info_tree
= splay_tree_new ((splay_tree_compare_fn
)strcmp
,
107 (splay_tree_delete_value_fn
)free
);
108 toplevel
= get_fileinfo ("<top level>");
109 if (flag_detailed_statistics
)
112 body_time
= get_run_time ();
113 toplevel
->time
= body_time
;
116 #ifdef MULTIBYTE_CHARS
117 /* Change to the native locale for multibyte conversions. */
118 setlocale (LC_CTYPE
, "");
119 GET_ENVIRONMENT (literal_codeset
, "LANG");
122 cb
= cpp_get_callbacks (parse_in
);
124 cb
->line_change
= cb_line_change
;
125 cb
->ident
= cb_ident
;
126 cb
->file_change
= cb_file_change
;
127 cb
->def_pragma
= cb_def_pragma
;
129 /* Set the debug callbacks if we can use them. */
130 if (debug_info_level
== DINFO_LEVEL_VERBOSE
131 && (write_symbols
== DWARF_DEBUG
|| write_symbols
== DWARF2_DEBUG
132 || write_symbols
== VMS_AND_DWARF2_DEBUG
))
134 cb
->define
= cb_define
;
135 cb
->undef
= cb_undef
;
141 if (filename
== NULL
|| !strcmp (filename
, "-"))
144 return cpp_read_main_file (parse_in
, filename
, ident_hash
);
147 /* A thin wrapper around the real parser that initializes the
148 integrated preprocessor after debug output has been initialized.
149 Also, make sure the start_source_file debug hook gets called for
150 the primary source file. */
153 c_common_parse_file (set_yydebug
)
154 int set_yydebug ATTRIBUTE_UNUSED
;
157 yydebug
= set_yydebug
;
159 warning ("YYDEBUG not defined");
162 (*debug_hooks
->start_source_file
) (lineno
, input_filename
);
163 cpp_finish_options (parse_in
);
166 free_parser_stacks ();
174 struct c_fileinfo
*fi
;
176 n
= splay_tree_lookup (file_info_tree
, (splay_tree_key
) name
);
178 return (struct c_fileinfo
*) n
->value
;
180 fi
= (struct c_fileinfo
*) xmalloc (sizeof (struct c_fileinfo
));
182 fi
->interface_only
= 0;
183 fi
->interface_unknown
= 1;
184 splay_tree_insert (file_info_tree
, (splay_tree_key
) name
,
185 (splay_tree_value
) fi
);
190 update_header_times (name
)
193 /* Changing files again. This means currently collected time
194 is charged against header time, and body time starts back at 0. */
195 if (flag_detailed_statistics
)
197 int this_time
= get_run_time ();
198 struct c_fileinfo
*file
= get_fileinfo (name
);
199 header_time
+= this_time
- body_time
;
200 file
->time
+= this_time
- body_time
;
201 body_time
= this_time
;
206 dump_one_header (n
, dummy
)
208 void *dummy ATTRIBUTE_UNUSED
;
210 print_time ((const char *) n
->key
,
211 ((struct c_fileinfo
*) n
->value
)->time
);
216 dump_time_statistics ()
218 struct c_fileinfo
*file
= get_fileinfo (input_filename
);
219 int this_time
= get_run_time ();
220 file
->time
+= this_time
- body_time
;
222 fprintf (stderr
, "\n******\n");
223 print_time ("header files (total)", header_time
);
224 print_time ("main file (total)", this_time
- body_time
);
225 fprintf (stderr
, "ratio = %g : 1\n",
226 (double)header_time
/ (double)(this_time
- body_time
));
227 fprintf (stderr
, "\n******\n");
229 splay_tree_foreach (file_info_tree
, dump_one_header
, 0);
233 cb_ident (pfile
, line
, str
)
234 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
235 unsigned int line ATTRIBUTE_UNUSED
;
236 const cpp_string
*str ATTRIBUTE_UNUSED
;
238 #ifdef ASM_OUTPUT_IDENT
241 /* Convert escapes in the string. */
242 tree value
= lex_string (str
->text
, str
->len
, 0);
243 ASM_OUTPUT_IDENT (asm_out_file
, TREE_STRING_POINTER (value
));
248 /* Called at the start of every non-empty line. TOKEN is the first
249 lexed token on the line. Used for diagnostic line numbers. */
251 cb_line_change (pfile
, token
, parsing_args
)
252 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
253 const cpp_token
*token
;
254 int parsing_args ATTRIBUTE_UNUSED
;
256 src_lineno
= SOURCE_LINE (map
, token
->line
);
260 cb_file_change (pfile
, new_map
)
261 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
262 const struct line_map
*new_map
;
264 unsigned int to_line
= SOURCE_LINE (new_map
, new_map
->to_line
);
266 if (new_map
->reason
== LC_ENTER
)
268 /* Don't stack the main buffer on the input stack;
269 we already did in compile_file. */
271 main_input_filename
= new_map
->to_file
;
274 int included_at
= SOURCE_LINE (new_map
- 1, new_map
->from_line
- 1);
276 lineno
= included_at
;
277 push_srcloc (new_map
->to_file
, 1);
278 (*debug_hooks
->start_source_file
) (included_at
, new_map
->to_file
);
279 #ifndef NO_IMPLICIT_EXTERN_C
282 else if (new_map
->sysp
== 2)
285 ++pending_lang_change
;
290 else if (new_map
->reason
== LC_LEAVE
)
292 #ifndef NO_IMPLICIT_EXTERN_C
293 if (c_header_level
&& --c_header_level
== 0)
295 if (new_map
->sysp
== 2)
296 warning ("badly nested C headers from preprocessor");
297 --pending_lang_change
;
302 (*debug_hooks
->end_source_file
) (to_line
);
305 update_header_times (new_map
->to_file
);
306 in_system_header
= new_map
->sysp
!= 0;
307 input_filename
= new_map
->to_file
;
312 extract_interface_info ();
316 cb_def_pragma (pfile
, line
)
320 /* Issue a warning message if we have been asked to do so. Ignore
321 unknown pragmas in system headers unless an explicit
322 -Wunknown-pragmas has been given. */
323 if (warn_unknown_pragmas
> in_system_header
)
325 const unsigned char *space
, *name
= 0;
328 s
= cpp_get_token (pfile
);
329 space
= cpp_token_as_text (pfile
, s
);
330 s
= cpp_get_token (pfile
);
331 if (s
->type
== CPP_NAME
)
332 name
= cpp_token_as_text (pfile
, s
);
334 lineno
= SOURCE_LINE (map
, line
);
336 warning ("ignoring #pragma %s %s", space
, name
);
338 warning ("ignoring #pragma %s", space
);
342 /* #define callback for DWARF and DWARF2 debug info. */
344 cb_define (pfile
, line
, node
)
349 (*debug_hooks
->define
) (SOURCE_LINE (map
, line
),
350 (const char *) cpp_macro_definition (pfile
, node
));
353 /* #undef callback for DWARF and DWARF2 debug info. */
355 cb_undef (pfile
, line
, node
)
356 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
360 (*debug_hooks
->undef
) (SOURCE_LINE (map
, line
),
361 (const char *) NODE_NAME (node
));
365 /* Returns nonzero if C is a universal-character-name. Give an error if it
366 is not one which may appear in an identifier, as per [extendid].
368 Note that extended character support in identifiers has not yet been
369 implemented. It is my personal opinion that this is not a desirable
370 feature. Portable code cannot count on support for more than the basic
371 identifier character set. */
384 /* None of the valid chars are outside the Basic Multilingual Plane (the
388 error ("universal-character-name '\\U%08x' not valid in identifier", c
);
393 if ((c
>= 0x00c0 && c
<= 0x00d6)
394 || (c
>= 0x00d8 && c
<= 0x00f6)
395 || (c
>= 0x00f8 && c
<= 0x01f5)
396 || (c
>= 0x01fa && c
<= 0x0217)
397 || (c
>= 0x0250 && c
<= 0x02a8)
398 || (c
>= 0x1e00 && c
<= 0x1e9a)
399 || (c
>= 0x1ea0 && c
<= 0x1ef9))
404 || (c
>= 0x0388 && c
<= 0x038a)
406 || (c
>= 0x038e && c
<= 0x03a1)
407 || (c
>= 0x03a3 && c
<= 0x03ce)
408 || (c
>= 0x03d0 && c
<= 0x03d6)
413 || (c
>= 0x03e2 && c
<= 0x03f3)
414 || (c
>= 0x1f00 && c
<= 0x1f15)
415 || (c
>= 0x1f18 && c
<= 0x1f1d)
416 || (c
>= 0x1f20 && c
<= 0x1f45)
417 || (c
>= 0x1f48 && c
<= 0x1f4d)
418 || (c
>= 0x1f50 && c
<= 0x1f57)
422 || (c
>= 0x1f5f && c
<= 0x1f7d)
423 || (c
>= 0x1f80 && c
<= 0x1fb4)
424 || (c
>= 0x1fb6 && c
<= 0x1fbc)
425 || (c
>= 0x1fc2 && c
<= 0x1fc4)
426 || (c
>= 0x1fc6 && c
<= 0x1fcc)
427 || (c
>= 0x1fd0 && c
<= 0x1fd3)
428 || (c
>= 0x1fd6 && c
<= 0x1fdb)
429 || (c
>= 0x1fe0 && c
<= 0x1fec)
430 || (c
>= 0x1ff2 && c
<= 0x1ff4)
431 || (c
>= 0x1ff6 && c
<= 0x1ffc))
435 if ((c
>= 0x0401 && c
<= 0x040d)
436 || (c
>= 0x040f && c
<= 0x044f)
437 || (c
>= 0x0451 && c
<= 0x045c)
438 || (c
>= 0x045e && c
<= 0x0481)
439 || (c
>= 0x0490 && c
<= 0x04c4)
440 || (c
>= 0x04c7 && c
<= 0x04c8)
441 || (c
>= 0x04cb && c
<= 0x04cc)
442 || (c
>= 0x04d0 && c
<= 0x04eb)
443 || (c
>= 0x04ee && c
<= 0x04f5)
444 || (c
>= 0x04f8 && c
<= 0x04f9))
448 if ((c
>= 0x0531 && c
<= 0x0556)
449 || (c
>= 0x0561 && c
<= 0x0587))
453 if ((c
>= 0x05d0 && c
<= 0x05ea)
454 || (c
>= 0x05f0 && c
<= 0x05f4))
458 if ((c
>= 0x0621 && c
<= 0x063a)
459 || (c
>= 0x0640 && c
<= 0x0652)
460 || (c
>= 0x0670 && c
<= 0x06b7)
461 || (c
>= 0x06ba && c
<= 0x06be)
462 || (c
>= 0x06c0 && c
<= 0x06ce)
463 || (c
>= 0x06e5 && c
<= 0x06e7))
467 if ((c
>= 0x0905 && c
<= 0x0939)
468 || (c
>= 0x0958 && c
<= 0x0962))
472 if ((c
>= 0x0985 && c
<= 0x098c)
473 || (c
>= 0x098f && c
<= 0x0990)
474 || (c
>= 0x0993 && c
<= 0x09a8)
475 || (c
>= 0x09aa && c
<= 0x09b0)
477 || (c
>= 0x09b6 && c
<= 0x09b9)
478 || (c
>= 0x09dc && c
<= 0x09dd)
479 || (c
>= 0x09df && c
<= 0x09e1)
480 || (c
>= 0x09f0 && c
<= 0x09f1))
484 if ((c
>= 0x0a05 && c
<= 0x0a0a)
485 || (c
>= 0x0a0f && c
<= 0x0a10)
486 || (c
>= 0x0a13 && c
<= 0x0a28)
487 || (c
>= 0x0a2a && c
<= 0x0a30)
488 || (c
>= 0x0a32 && c
<= 0x0a33)
489 || (c
>= 0x0a35 && c
<= 0x0a36)
490 || (c
>= 0x0a38 && c
<= 0x0a39)
491 || (c
>= 0x0a59 && c
<= 0x0a5c)
496 if ((c
>= 0x0a85 && c
<= 0x0a8b)
498 || (c
>= 0x0a8f && c
<= 0x0a91)
499 || (c
>= 0x0a93 && c
<= 0x0aa8)
500 || (c
>= 0x0aaa && c
<= 0x0ab0)
501 || (c
>= 0x0ab2 && c
<= 0x0ab3)
502 || (c
>= 0x0ab5 && c
<= 0x0ab9)
507 if ((c
>= 0x0b05 && c
<= 0x0b0c)
508 || (c
>= 0x0b0f && c
<= 0x0b10)
509 || (c
>= 0x0b13 && c
<= 0x0b28)
510 || (c
>= 0x0b2a && c
<= 0x0b30)
511 || (c
>= 0x0b32 && c
<= 0x0b33)
512 || (c
>= 0x0b36 && c
<= 0x0b39)
513 || (c
>= 0x0b5c && c
<= 0x0b5d)
514 || (c
>= 0x0b5f && c
<= 0x0b61))
518 if ((c
>= 0x0b85 && c
<= 0x0b8a)
519 || (c
>= 0x0b8e && c
<= 0x0b90)
520 || (c
>= 0x0b92 && c
<= 0x0b95)
521 || (c
>= 0x0b99 && c
<= 0x0b9a)
523 || (c
>= 0x0b9e && c
<= 0x0b9f)
524 || (c
>= 0x0ba3 && c
<= 0x0ba4)
525 || (c
>= 0x0ba8 && c
<= 0x0baa)
526 || (c
>= 0x0bae && c
<= 0x0bb5)
527 || (c
>= 0x0bb7 && c
<= 0x0bb9))
531 if ((c
>= 0x0c05 && c
<= 0x0c0c)
532 || (c
>= 0x0c0e && c
<= 0x0c10)
533 || (c
>= 0x0c12 && c
<= 0x0c28)
534 || (c
>= 0x0c2a && c
<= 0x0c33)
535 || (c
>= 0x0c35 && c
<= 0x0c39)
536 || (c
>= 0x0c60 && c
<= 0x0c61))
540 if ((c
>= 0x0c85 && c
<= 0x0c8c)
541 || (c
>= 0x0c8e && c
<= 0x0c90)
542 || (c
>= 0x0c92 && c
<= 0x0ca8)
543 || (c
>= 0x0caa && c
<= 0x0cb3)
544 || (c
>= 0x0cb5 && c
<= 0x0cb9)
545 || (c
>= 0x0ce0 && c
<= 0x0ce1))
549 if ((c
>= 0x0d05 && c
<= 0x0d0c)
550 || (c
>= 0x0d0e && c
<= 0x0d10)
551 || (c
>= 0x0d12 && c
<= 0x0d28)
552 || (c
>= 0x0d2a && c
<= 0x0d39)
553 || (c
>= 0x0d60 && c
<= 0x0d61))
557 if ((c
>= 0x0e01 && c
<= 0x0e30)
558 || (c
>= 0x0e32 && c
<= 0x0e33)
559 || (c
>= 0x0e40 && c
<= 0x0e46)
560 || (c
>= 0x0e4f && c
<= 0x0e5b))
564 if ((c
>= 0x0e81 && c
<= 0x0e82)
570 || (c
>= 0x0e94 && c
<= 0x0e97)
571 || (c
>= 0x0e99 && c
<= 0x0e9f)
572 || (c
>= 0x0ea1 && c
<= 0x0ea3)
577 || (c
>= 0x0ead && c
<= 0x0eb0)
581 || (c
>= 0x0ec0 && c
<= 0x0ec4)
586 if ((c
>= 0x10a0 && c
<= 0x10c5)
587 || (c
>= 0x10d0 && c
<= 0x10f6))
591 if ((c
>= 0x3041 && c
<= 0x3094)
592 || (c
>= 0x309b && c
<= 0x309e))
596 if ((c
>= 0x30a1 && c
<= 0x30fe))
600 if ((c
>= 0x3105 && c
<= 0x312c))
604 if ((c
>= 0x1100 && c
<= 0x1159)
605 || (c
>= 0x1161 && c
<= 0x11a2)
606 || (c
>= 0x11a8 && c
<= 0x11f9))
609 /* CJK Unified Ideographs */
610 if ((c
>= 0xf900 && c
<= 0xfa2d)
611 || (c
>= 0xfb1f && c
<= 0xfb36)
612 || (c
>= 0xfb38 && c
<= 0xfb3c)
614 || (c
>= 0xfb40 && c
<= 0xfb41)
615 || (c
>= 0xfb42 && c
<= 0xfb44)
616 || (c
>= 0xfb46 && c
<= 0xfbb1)
617 || (c
>= 0xfbd3 && c
<= 0xfd3f)
618 || (c
>= 0xfd50 && c
<= 0xfd8f)
619 || (c
>= 0xfd92 && c
<= 0xfdc7)
620 || (c
>= 0xfdf0 && c
<= 0xfdfb)
621 || (c
>= 0xfe70 && c
<= 0xfe72)
623 || (c
>= 0xfe76 && c
<= 0xfefc)
624 || (c
>= 0xff21 && c
<= 0xff3a)
625 || (c
>= 0xff41 && c
<= 0xff5a)
626 || (c
>= 0xff66 && c
<= 0xffbe)
627 || (c
>= 0xffc2 && c
<= 0xffc7)
628 || (c
>= 0xffca && c
<= 0xffcf)
629 || (c
>= 0xffd2 && c
<= 0xffd7)
630 || (c
>= 0xffda && c
<= 0xffdc)
631 || (c
>= 0x4e00 && c
<= 0x9fa5))
634 error ("universal-character-name '\\u%04x' not valid in identifier", c
);
639 /* Add the UTF-8 representation of C to the token_buffer. */
642 utf8_extend_token (c
)
652 else if (c
<= 0x000007ff)
653 shift
= 6, mask
= 0xc0;
654 else if (c
<= 0x0000ffff)
655 shift
= 12, mask
= 0xe0;
656 else if (c
<= 0x001fffff)
657 shift
= 18, mask
= 0xf0;
658 else if (c
<= 0x03ffffff)
659 shift
= 24, mask
= 0xf8;
661 shift
= 30, mask
= 0xfc;
663 extend_token (mask
| (c
>> shift
));
667 extend_token ((unsigned char) (0x80 | (c
>> shift
)));
677 const cpp_token
*tok
;
680 timevar_push (TV_CPP
);
682 tok
= cpp_get_token (parse_in
);
683 while (tok
->type
== CPP_PADDING
);
684 timevar_pop (TV_CPP
);
686 /* The C++ front end does horrible things with the current line
687 number. To ensure an accurate line number, we must reset it
688 every time we return a token. */
694 /* Issue this error here, where we can get at tok->val.c. */
696 if (ISGRAPH (tok
->val
.c
))
697 error ("stray '%c' in program", tok
->val
.c
);
699 error ("stray '\\%o' in program", tok
->val
.c
);
703 *value
= HT_IDENT_TO_GCC_IDENT (HT_NODE (tok
->val
.node
));
708 unsigned int flags
= cpp_classify_number (parse_in
, tok
);
710 switch (flags
& CPP_N_CATEGORY
)
713 /* cpplib has issued an error. */
717 *value
= interpret_integer (tok
, flags
);
721 *value
= interpret_float (tok
, flags
);
732 *value
= lex_charconst (tok
);
737 *value
= lex_string (tok
->val
.str
.text
, tok
->val
.str
.len
,
738 tok
->type
== CPP_WSTRING
);
741 /* These tokens should not be visible outside cpplib. */
742 case CPP_HEADER_NAME
:
753 /* Returns the narrowest C-visible unsigned type, starting with the
754 minimum specified by FLAGS, that can fit VALUE, or itk_none if
756 static enum integer_type_kind
757 narrowest_unsigned_type (value
, flags
)
761 enum integer_type_kind itk
;
763 if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
)
764 itk
= itk_unsigned_int
;
765 else if ((flags
& CPP_N_WIDTH
) == CPP_N_MEDIUM
)
766 itk
= itk_unsigned_long
;
768 itk
= itk_unsigned_long_long
;
770 /* int_fits_type_p must think the type of its first argument is
771 wider than its second argument, or it won't do the proper check. */
772 TREE_TYPE (value
) = widest_unsigned_literal_type_node
;
774 for (; itk
< itk_none
; itk
+= 2 /* skip unsigned types */)
775 if (int_fits_type_p (value
, integer_types
[itk
]))
781 /* Ditto, but narrowest signed type. */
782 static enum integer_type_kind
783 narrowest_signed_type (value
, flags
)
787 enum integer_type_kind itk
;
789 if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
)
791 else if ((flags
& CPP_N_WIDTH
) == CPP_N_MEDIUM
)
796 /* int_fits_type_p must think the type of its first argument is
797 wider than its second argument, or it won't do the proper check. */
798 TREE_TYPE (value
) = widest_unsigned_literal_type_node
;
800 for (; itk
< itk_none
; itk
+= 2 /* skip signed types */)
801 if (int_fits_type_p (value
, integer_types
[itk
]))
807 /* Interpret TOKEN, an integer with FLAGS as classified by cpplib. */
809 interpret_integer (token
, flags
)
810 const cpp_token
*token
;
814 enum integer_type_kind itk
;
816 cpp_options
*options
= cpp_get_options (parse_in
);
818 integer
= cpp_interpret_integer (parse_in
, token
, flags
);
819 integer
= cpp_num_sign_extend (integer
, options
->precision
);
820 value
= build_int_2_wide (integer
.low
, integer
.high
);
822 /* The type of a constant with a U suffix is straightforward. */
823 if (flags
& CPP_N_UNSIGNED
)
824 itk
= narrowest_unsigned_type (value
, flags
);
827 /* The type of a potentially-signed integer constant varies
828 depending on the base it's in, the standard in use, and the
830 enum integer_type_kind itk_u
= narrowest_unsigned_type (value
, flags
);
831 enum integer_type_kind itk_s
= narrowest_signed_type (value
, flags
);
833 /* In both C89 and C99, octal and hex constants may be signed or
834 unsigned, whichever fits tighter. We do not warn about this
835 choice differing from the traditional choice, as the constant
836 is probably a bit pattern and either way will work. */
837 if ((flags
& CPP_N_RADIX
) != CPP_N_DECIMAL
)
838 itk
= MIN (itk_u
, itk_s
);
841 /* In C99, decimal constants are always signed.
842 In C89, decimal constants that don't fit in long have
843 undefined behaviour; we try to make them unsigned long.
844 In GCC's extended C89, that last is true of decimal
845 constants that don't fit in long long, too. */
848 if (itk_s
> itk_u
&& itk_s
> itk_long
)
852 if (itk_u
< itk_unsigned_long
)
853 itk_u
= itk_unsigned_long
;
855 warning ("this decimal constant is unsigned only in ISO C89");
857 else if (warn_traditional
)
858 warning ("this decimal constant would be unsigned in ISO C89");
864 /* cpplib has already issued a warning for overflow. */
865 type
= ((flags
& CPP_N_UNSIGNED
)
866 ? widest_unsigned_literal_type_node
867 : widest_integer_literal_type_node
);
869 type
= integer_types
[itk
];
871 if (itk
> itk_unsigned_long
872 && (flags
& CPP_N_WIDTH
) != CPP_N_LARGE
873 && ! in_system_header
&& ! flag_isoc99
)
874 pedwarn ("integer constant is too large for \"%s\" type",
875 (flags
& CPP_N_UNSIGNED
) ? "unsigned long" : "long");
877 TREE_TYPE (value
) = type
;
879 /* Convert imaginary to a complex type. */
880 if (flags
& CPP_N_IMAGINARY
)
881 value
= build_complex (NULL_TREE
, convert (type
, integer_zero_node
), value
);
886 /* Interpret TOKEN, a floating point number with FLAGS as classified
889 interpret_float (token
, flags
)
890 const cpp_token
*token
;
895 REAL_VALUE_TYPE real
;
898 const char *typename
;
900 /* FIXME: make %T work in error/warning, then we don't need typename. */
901 if ((flags
& CPP_N_WIDTH
) == CPP_N_LARGE
)
903 type
= long_double_type_node
;
904 typename
= "long double";
906 else if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
907 || flag_single_precision_constant
)
909 type
= float_type_node
;
914 type
= double_type_node
;
918 /* Copy the constant to a nul-terminated buffer. If the constant
919 has any suffixes, cut them off; REAL_VALUE_ATOF/ REAL_VALUE_HTOF
920 can't handle them. */
921 copylen
= token
->val
.str
.len
;
922 if ((flags
& CPP_N_WIDTH
) != CPP_N_MEDIUM
)
923 /* Must be an F or L suffix. */
925 if (flags
& CPP_N_IMAGINARY
)
929 copy
= alloca (copylen
+ 1);
930 memcpy (copy
, token
->val
.str
.text
, copylen
);
931 copy
[copylen
] = '\0';
933 /* The second argument, machine_mode, of REAL_VALUE_ATOF tells the
934 desired precision of the binary result of decimal-to-binary
936 if (flags
& CPP_N_HEX
)
937 real
= REAL_VALUE_HTOF (copy
, TYPE_MODE (type
));
939 real
= REAL_VALUE_ATOF (copy
, TYPE_MODE (type
));
941 /* A diagnostic is required for "soft" overflow by some ISO C
942 testsuites. This is not pedwarn, because some people don't want
944 ??? That's a dubious reason... is this a mandatory diagnostic or
945 isn't it? -- zw, 2001-08-21. */
946 if (REAL_VALUE_ISINF (real
) && pedantic
)
947 warning ("floating constant exceeds range of \"%s\"", typename
);
949 /* Create a node with determined type and value. */
950 value
= build_real (type
, real
);
951 if (flags
& CPP_N_IMAGINARY
)
952 value
= build_complex (NULL_TREE
, convert (type
, integer_zero_node
), value
);
958 lex_string (str
, len
, wide
)
959 const unsigned char *str
;
964 char *buf
= alloca ((len
+ 1) * (wide
? WCHAR_BYTES
: 1));
966 const unsigned char *p
= str
, *limit
= str
+ len
;
969 #ifdef MULTIBYTE_CHARS
970 /* Reset multibyte conversion state. */
971 (void) local_mbtowc (NULL
, NULL
, 0);
976 #ifdef MULTIBYTE_CHARS
980 char_len
= local_mbtowc (&wc
, (const char *) p
, limit
- p
);
983 warning ("ignoring invalid multibyte character");
996 if (c
== '\\' && !ignore_escape_flag
)
997 c
= cpp_parse_escape (parse_in
, &p
, limit
, wide
);
999 /* Add this single character into the buffer either as a wchar_t,
1000 a multibyte sequence, or as a single byte. */
1003 unsigned charwidth
= TYPE_PRECISION (char_type_node
);
1004 unsigned bytemask
= (1 << charwidth
) - 1;
1007 for (byte
= 0; byte
< WCHAR_BYTES
; ++byte
)
1010 if (byte
>= (int) sizeof (c
))
1013 n
= (c
>> (byte
* charwidth
)) & bytemask
;
1014 if (BYTES_BIG_ENDIAN
)
1015 q
[WCHAR_BYTES
- byte
- 1] = n
;
1021 #ifdef MULTIBYTE_CHARS
1022 else if (char_len
> 1)
1024 /* We're dealing with a multibyte character. */
1025 for ( ; char_len
>0; --char_len
)
1027 *q
++ = *(p
- char_len
);
1037 /* Terminate the string value, either with a single byte zero
1038 or with a wide zero. */
1042 memset (q
, 0, WCHAR_BYTES
);
1050 value
= build_string (q
- buf
, buf
);
1053 TREE_TYPE (value
) = wchar_array_type_node
;
1055 TREE_TYPE (value
) = char_array_type_node
;
1059 /* Converts a (possibly wide) character constant token into a tree. */
1061 lex_charconst (token
)
1062 const cpp_token
*token
;
1066 unsigned int chars_seen
;
1069 result
= cpp_interpret_charconst (parse_in
, token
,
1070 &chars_seen
, &unsignedp
);
1072 /* Cast to cppchar_signed_t to get correct sign-extension of RESULT
1073 before possibly widening to HOST_WIDE_INT for build_int_2. */
1074 if (unsignedp
|| (cppchar_signed_t
) result
>= 0)
1075 value
= build_int_2 (result
, 0);
1077 value
= build_int_2 ((cppchar_signed_t
) result
, -1);
1079 if (token
->type
== CPP_WCHAR
)
1080 type
= wchar_type_node
;
1081 /* In C, a character constant has type 'int'.
1082 In C++ 'char', but multi-char charconsts have type 'int'. */
1083 else if ((c_language
== clk_c
|| c_language
== clk_objective_c
)
1085 type
= integer_type_node
;
1087 type
= char_type_node
;
1089 TREE_TYPE (value
) = type
;