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 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"
42 #include "splay-tree.h"
45 #ifdef MULTIBYTE_CHARS
48 #endif /* MULTIBYTE_CHARS */
50 /* The current line map. */
51 static const struct line_map
*map
;
53 /* The line used to refresh the lineno global variable after each token. */
54 static unsigned int src_lineno
;
56 /* We may keep statistics about how long which files took to compile. */
57 static int header_time
, body_time
;
58 static splay_tree file_info_tree
;
60 /* File used for outputting assembler code. */
61 extern FILE *asm_out_file
;
63 #undef WCHAR_TYPE_SIZE
64 #define WCHAR_TYPE_SIZE TYPE_PRECISION (wchar_type_node)
66 /* Number of bytes in a wide character. */
67 #define WCHAR_BYTES (WCHAR_TYPE_SIZE / BITS_PER_UNIT)
69 int pending_lang_change
; /* If we need to switch languages - C++ only */
70 int c_header_level
; /* depth in C headers - C++ only */
72 /* Nonzero tells yylex to ignore \ in string constants. */
73 static int ignore_escape_flag
;
75 static tree interpret_integer
PARAMS ((const cpp_token
*, unsigned int));
76 static tree interpret_float
PARAMS ((const cpp_token
*, unsigned int));
77 static enum integer_type_kind
78 narrowest_unsigned_type
PARAMS ((tree
, unsigned int));
79 static enum integer_type_kind
80 narrowest_signed_type
PARAMS ((tree
, unsigned int));
81 static tree lex_string
PARAMS ((const unsigned char *, unsigned int,
83 static tree lex_charconst
PARAMS ((const cpp_token
*));
84 static void update_header_times
PARAMS ((const char *));
85 static int dump_one_header
PARAMS ((splay_tree_node
, void *));
86 static void cb_line_change
PARAMS ((cpp_reader
*, const cpp_token
*, int));
87 static void cb_ident
PARAMS ((cpp_reader
*, unsigned int,
89 static void cb_def_pragma
PARAMS ((cpp_reader
*, unsigned int));
90 static void cb_define
PARAMS ((cpp_reader
*, unsigned int,
92 static void cb_undef
PARAMS ((cpp_reader
*, unsigned int,
98 struct cpp_callbacks
*cb
;
99 struct c_fileinfo
*toplevel
;
101 /* Set up filename timing. Must happen before cpp_read_main_file. */
102 file_info_tree
= splay_tree_new ((splay_tree_compare_fn
)strcmp
,
104 (splay_tree_delete_value_fn
)free
);
105 toplevel
= get_fileinfo ("<top level>");
106 if (flag_detailed_statistics
)
109 body_time
= get_run_time ();
110 toplevel
->time
= body_time
;
113 cb
= cpp_get_callbacks (parse_in
);
115 cb
->line_change
= cb_line_change
;
116 cb
->ident
= cb_ident
;
117 cb
->def_pragma
= cb_def_pragma
;
118 cb
->valid_pch
= c_common_valid_pch
;
119 cb
->read_pch
= c_common_read_pch
;
121 /* Set the debug callbacks if we can use them. */
122 if (debug_info_level
== DINFO_LEVEL_VERBOSE
123 && (write_symbols
== DWARF_DEBUG
|| write_symbols
== DWARF2_DEBUG
124 || write_symbols
== VMS_AND_DWARF2_DEBUG
))
126 cb
->define
= cb_define
;
127 cb
->undef
= cb_undef
;
136 struct c_fileinfo
*fi
;
138 n
= splay_tree_lookup (file_info_tree
, (splay_tree_key
) name
);
140 return (struct c_fileinfo
*) n
->value
;
142 fi
= (struct c_fileinfo
*) xmalloc (sizeof (struct c_fileinfo
));
144 fi
->interface_only
= 0;
145 fi
->interface_unknown
= 1;
146 splay_tree_insert (file_info_tree
, (splay_tree_key
) name
,
147 (splay_tree_value
) fi
);
152 update_header_times (name
)
155 /* Changing files again. This means currently collected time
156 is charged against header time, and body time starts back at 0. */
157 if (flag_detailed_statistics
)
159 int this_time
= get_run_time ();
160 struct c_fileinfo
*file
= get_fileinfo (name
);
161 header_time
+= this_time
- body_time
;
162 file
->time
+= this_time
- body_time
;
163 body_time
= this_time
;
168 dump_one_header (n
, dummy
)
170 void *dummy ATTRIBUTE_UNUSED
;
172 print_time ((const char *) n
->key
,
173 ((struct c_fileinfo
*) n
->value
)->time
);
178 dump_time_statistics ()
180 struct c_fileinfo
*file
= get_fileinfo (input_filename
);
181 int this_time
= get_run_time ();
182 file
->time
+= this_time
- body_time
;
184 fprintf (stderr
, "\n******\n");
185 print_time ("header files (total)", header_time
);
186 print_time ("main file (total)", this_time
- body_time
);
187 fprintf (stderr
, "ratio = %g : 1\n",
188 (double)header_time
/ (double)(this_time
- body_time
));
189 fprintf (stderr
, "\n******\n");
191 splay_tree_foreach (file_info_tree
, dump_one_header
, 0);
195 cb_ident (pfile
, line
, str
)
196 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
197 unsigned int line ATTRIBUTE_UNUSED
;
198 const cpp_string
*str ATTRIBUTE_UNUSED
;
200 #ifdef ASM_OUTPUT_IDENT
203 /* Convert escapes in the string. */
204 tree value ATTRIBUTE_UNUSED
= lex_string (str
->text
, str
->len
, 0);
205 ASM_OUTPUT_IDENT (asm_out_file
, TREE_STRING_POINTER (value
));
210 /* Called at the start of every non-empty line. TOKEN is the first
211 lexed token on the line. Used for diagnostic line numbers. */
213 cb_line_change (pfile
, token
, parsing_args
)
214 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
215 const cpp_token
*token
;
216 int parsing_args ATTRIBUTE_UNUSED
;
218 src_lineno
= SOURCE_LINE (map
, token
->line
);
222 fe_file_change (new_map
)
223 const struct line_map
*new_map
;
225 unsigned int to_line
= SOURCE_LINE (new_map
, new_map
->to_line
);
227 if (new_map
->reason
== LC_ENTER
)
229 /* Don't stack the main buffer on the input stack;
230 we already did in compile_file. */
232 main_input_filename
= new_map
->to_file
;
235 int included_at
= SOURCE_LINE (new_map
- 1, new_map
->from_line
- 1);
237 lineno
= included_at
;
238 push_srcloc (new_map
->to_file
, 1);
239 (*debug_hooks
->start_source_file
) (included_at
, new_map
->to_file
);
240 #ifndef NO_IMPLICIT_EXTERN_C
243 else if (new_map
->sysp
== 2)
246 ++pending_lang_change
;
251 else if (new_map
->reason
== LC_LEAVE
)
253 #ifndef NO_IMPLICIT_EXTERN_C
254 if (c_header_level
&& --c_header_level
== 0)
256 if (new_map
->sysp
== 2)
257 warning ("badly nested C headers from preprocessor");
258 --pending_lang_change
;
263 (*debug_hooks
->end_source_file
) (to_line
);
266 update_header_times (new_map
->to_file
);
267 in_system_header
= new_map
->sysp
!= 0;
268 input_filename
= new_map
->to_file
;
273 extract_interface_info ();
277 cb_def_pragma (pfile
, line
)
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 const unsigned char *space
, *name
;
289 space
= name
= (const unsigned char *) "";
290 s
= cpp_get_token (pfile
);
291 if (s
->type
!= CPP_EOF
)
293 space
= cpp_token_as_text (pfile
, s
);
294 s
= cpp_get_token (pfile
);
295 if (s
->type
== CPP_NAME
)
296 name
= cpp_token_as_text (pfile
, s
);
299 lineno
= SOURCE_LINE (map
, line
);
300 warning ("ignoring #pragma %s %s", space
, name
);
304 /* #define callback for DWARF and DWARF2 debug info. */
306 cb_define (pfile
, line
, node
)
311 (*debug_hooks
->define
) (SOURCE_LINE (map
, line
),
312 (const char *) cpp_macro_definition (pfile
, node
));
315 /* #undef callback for DWARF and DWARF2 debug info. */
317 cb_undef (pfile
, line
, node
)
318 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
322 (*debug_hooks
->undef
) (SOURCE_LINE (map
, line
),
323 (const char *) NODE_NAME (node
));
327 /* Returns nonzero if C is a universal-character-name. Give an error if it
328 is not one which may appear in an identifier, as per [extendid].
330 Note that extended character support in identifiers has not yet been
331 implemented. It is my personal opinion that this is not a desirable
332 feature. Portable code cannot count on support for more than the basic
333 identifier character set. */
346 /* None of the valid chars are outside the Basic Multilingual Plane (the
350 error ("universal-character-name '\\U%08x' not valid in identifier", c
);
355 if ((c
>= 0x00c0 && c
<= 0x00d6)
356 || (c
>= 0x00d8 && c
<= 0x00f6)
357 || (c
>= 0x00f8 && c
<= 0x01f5)
358 || (c
>= 0x01fa && c
<= 0x0217)
359 || (c
>= 0x0250 && c
<= 0x02a8)
360 || (c
>= 0x1e00 && c
<= 0x1e9a)
361 || (c
>= 0x1ea0 && c
<= 0x1ef9))
366 || (c
>= 0x0388 && c
<= 0x038a)
368 || (c
>= 0x038e && c
<= 0x03a1)
369 || (c
>= 0x03a3 && c
<= 0x03ce)
370 || (c
>= 0x03d0 && c
<= 0x03d6)
375 || (c
>= 0x03e2 && c
<= 0x03f3)
376 || (c
>= 0x1f00 && c
<= 0x1f15)
377 || (c
>= 0x1f18 && c
<= 0x1f1d)
378 || (c
>= 0x1f20 && c
<= 0x1f45)
379 || (c
>= 0x1f48 && c
<= 0x1f4d)
380 || (c
>= 0x1f50 && c
<= 0x1f57)
384 || (c
>= 0x1f5f && c
<= 0x1f7d)
385 || (c
>= 0x1f80 && c
<= 0x1fb4)
386 || (c
>= 0x1fb6 && c
<= 0x1fbc)
387 || (c
>= 0x1fc2 && c
<= 0x1fc4)
388 || (c
>= 0x1fc6 && c
<= 0x1fcc)
389 || (c
>= 0x1fd0 && c
<= 0x1fd3)
390 || (c
>= 0x1fd6 && c
<= 0x1fdb)
391 || (c
>= 0x1fe0 && c
<= 0x1fec)
392 || (c
>= 0x1ff2 && c
<= 0x1ff4)
393 || (c
>= 0x1ff6 && c
<= 0x1ffc))
397 if ((c
>= 0x0401 && c
<= 0x040d)
398 || (c
>= 0x040f && c
<= 0x044f)
399 || (c
>= 0x0451 && c
<= 0x045c)
400 || (c
>= 0x045e && c
<= 0x0481)
401 || (c
>= 0x0490 && c
<= 0x04c4)
402 || (c
>= 0x04c7 && c
<= 0x04c8)
403 || (c
>= 0x04cb && c
<= 0x04cc)
404 || (c
>= 0x04d0 && c
<= 0x04eb)
405 || (c
>= 0x04ee && c
<= 0x04f5)
406 || (c
>= 0x04f8 && c
<= 0x04f9))
410 if ((c
>= 0x0531 && c
<= 0x0556)
411 || (c
>= 0x0561 && c
<= 0x0587))
415 if ((c
>= 0x05d0 && c
<= 0x05ea)
416 || (c
>= 0x05f0 && c
<= 0x05f4))
420 if ((c
>= 0x0621 && c
<= 0x063a)
421 || (c
>= 0x0640 && c
<= 0x0652)
422 || (c
>= 0x0670 && c
<= 0x06b7)
423 || (c
>= 0x06ba && c
<= 0x06be)
424 || (c
>= 0x06c0 && c
<= 0x06ce)
425 || (c
>= 0x06e5 && c
<= 0x06e7))
429 if ((c
>= 0x0905 && c
<= 0x0939)
430 || (c
>= 0x0958 && c
<= 0x0962))
434 if ((c
>= 0x0985 && c
<= 0x098c)
435 || (c
>= 0x098f && c
<= 0x0990)
436 || (c
>= 0x0993 && c
<= 0x09a8)
437 || (c
>= 0x09aa && c
<= 0x09b0)
439 || (c
>= 0x09b6 && c
<= 0x09b9)
440 || (c
>= 0x09dc && c
<= 0x09dd)
441 || (c
>= 0x09df && c
<= 0x09e1)
442 || (c
>= 0x09f0 && c
<= 0x09f1))
446 if ((c
>= 0x0a05 && c
<= 0x0a0a)
447 || (c
>= 0x0a0f && c
<= 0x0a10)
448 || (c
>= 0x0a13 && c
<= 0x0a28)
449 || (c
>= 0x0a2a && c
<= 0x0a30)
450 || (c
>= 0x0a32 && c
<= 0x0a33)
451 || (c
>= 0x0a35 && c
<= 0x0a36)
452 || (c
>= 0x0a38 && c
<= 0x0a39)
453 || (c
>= 0x0a59 && c
<= 0x0a5c)
458 if ((c
>= 0x0a85 && c
<= 0x0a8b)
460 || (c
>= 0x0a8f && c
<= 0x0a91)
461 || (c
>= 0x0a93 && c
<= 0x0aa8)
462 || (c
>= 0x0aaa && c
<= 0x0ab0)
463 || (c
>= 0x0ab2 && c
<= 0x0ab3)
464 || (c
>= 0x0ab5 && c
<= 0x0ab9)
469 if ((c
>= 0x0b05 && c
<= 0x0b0c)
470 || (c
>= 0x0b0f && c
<= 0x0b10)
471 || (c
>= 0x0b13 && c
<= 0x0b28)
472 || (c
>= 0x0b2a && c
<= 0x0b30)
473 || (c
>= 0x0b32 && c
<= 0x0b33)
474 || (c
>= 0x0b36 && c
<= 0x0b39)
475 || (c
>= 0x0b5c && c
<= 0x0b5d)
476 || (c
>= 0x0b5f && c
<= 0x0b61))
480 if ((c
>= 0x0b85 && c
<= 0x0b8a)
481 || (c
>= 0x0b8e && c
<= 0x0b90)
482 || (c
>= 0x0b92 && c
<= 0x0b95)
483 || (c
>= 0x0b99 && c
<= 0x0b9a)
485 || (c
>= 0x0b9e && c
<= 0x0b9f)
486 || (c
>= 0x0ba3 && c
<= 0x0ba4)
487 || (c
>= 0x0ba8 && c
<= 0x0baa)
488 || (c
>= 0x0bae && c
<= 0x0bb5)
489 || (c
>= 0x0bb7 && c
<= 0x0bb9))
493 if ((c
>= 0x0c05 && c
<= 0x0c0c)
494 || (c
>= 0x0c0e && c
<= 0x0c10)
495 || (c
>= 0x0c12 && c
<= 0x0c28)
496 || (c
>= 0x0c2a && c
<= 0x0c33)
497 || (c
>= 0x0c35 && c
<= 0x0c39)
498 || (c
>= 0x0c60 && c
<= 0x0c61))
502 if ((c
>= 0x0c85 && c
<= 0x0c8c)
503 || (c
>= 0x0c8e && c
<= 0x0c90)
504 || (c
>= 0x0c92 && c
<= 0x0ca8)
505 || (c
>= 0x0caa && c
<= 0x0cb3)
506 || (c
>= 0x0cb5 && c
<= 0x0cb9)
507 || (c
>= 0x0ce0 && c
<= 0x0ce1))
511 if ((c
>= 0x0d05 && c
<= 0x0d0c)
512 || (c
>= 0x0d0e && c
<= 0x0d10)
513 || (c
>= 0x0d12 && c
<= 0x0d28)
514 || (c
>= 0x0d2a && c
<= 0x0d39)
515 || (c
>= 0x0d60 && c
<= 0x0d61))
519 if ((c
>= 0x0e01 && c
<= 0x0e30)
520 || (c
>= 0x0e32 && c
<= 0x0e33)
521 || (c
>= 0x0e40 && c
<= 0x0e46)
522 || (c
>= 0x0e4f && c
<= 0x0e5b))
526 if ((c
>= 0x0e81 && c
<= 0x0e82)
532 || (c
>= 0x0e94 && c
<= 0x0e97)
533 || (c
>= 0x0e99 && c
<= 0x0e9f)
534 || (c
>= 0x0ea1 && c
<= 0x0ea3)
539 || (c
>= 0x0ead && c
<= 0x0eb0)
543 || (c
>= 0x0ec0 && c
<= 0x0ec4)
548 if ((c
>= 0x10a0 && c
<= 0x10c5)
549 || (c
>= 0x10d0 && c
<= 0x10f6))
553 if ((c
>= 0x3041 && c
<= 0x3094)
554 || (c
>= 0x309b && c
<= 0x309e))
558 if ((c
>= 0x30a1 && c
<= 0x30fe))
562 if ((c
>= 0x3105 && c
<= 0x312c))
566 if ((c
>= 0x1100 && c
<= 0x1159)
567 || (c
>= 0x1161 && c
<= 0x11a2)
568 || (c
>= 0x11a8 && c
<= 0x11f9))
571 /* CJK Unified Ideographs */
572 if ((c
>= 0xf900 && c
<= 0xfa2d)
573 || (c
>= 0xfb1f && c
<= 0xfb36)
574 || (c
>= 0xfb38 && c
<= 0xfb3c)
576 || (c
>= 0xfb40 && c
<= 0xfb41)
577 || (c
>= 0xfb42 && c
<= 0xfb44)
578 || (c
>= 0xfb46 && c
<= 0xfbb1)
579 || (c
>= 0xfbd3 && c
<= 0xfd3f)
580 || (c
>= 0xfd50 && c
<= 0xfd8f)
581 || (c
>= 0xfd92 && c
<= 0xfdc7)
582 || (c
>= 0xfdf0 && c
<= 0xfdfb)
583 || (c
>= 0xfe70 && c
<= 0xfe72)
585 || (c
>= 0xfe76 && c
<= 0xfefc)
586 || (c
>= 0xff21 && c
<= 0xff3a)
587 || (c
>= 0xff41 && c
<= 0xff5a)
588 || (c
>= 0xff66 && c
<= 0xffbe)
589 || (c
>= 0xffc2 && c
<= 0xffc7)
590 || (c
>= 0xffca && c
<= 0xffcf)
591 || (c
>= 0xffd2 && c
<= 0xffd7)
592 || (c
>= 0xffda && c
<= 0xffdc)
593 || (c
>= 0x4e00 && c
<= 0x9fa5))
596 error ("universal-character-name '\\u%04x' not valid in identifier", c
);
601 /* Add the UTF-8 representation of C to the token_buffer. */
604 utf8_extend_token (c
)
614 else if (c
<= 0x000007ff)
615 shift
= 6, mask
= 0xc0;
616 else if (c
<= 0x0000ffff)
617 shift
= 12, mask
= 0xe0;
618 else if (c
<= 0x001fffff)
619 shift
= 18, mask
= 0xf0;
620 else if (c
<= 0x03ffffff)
621 shift
= 24, mask
= 0xf8;
623 shift
= 30, mask
= 0xfc;
625 extend_token (mask
| (c
>> shift
));
629 extend_token ((unsigned char) (0x80 | (c
>> shift
)));
639 const cpp_token
*tok
;
642 timevar_push (TV_CPP
);
644 tok
= cpp_get_token (parse_in
);
645 while (tok
->type
== CPP_PADDING
);
646 timevar_pop (TV_CPP
);
648 /* The C++ front end does horrible things with the current line
649 number. To ensure an accurate line number, we must reset it
650 every time we return a token. */
656 /* Issue this error here, where we can get at tok->val.c. */
658 if (ISGRAPH (tok
->val
.c
))
659 error ("stray '%c' in program", tok
->val
.c
);
661 error ("stray '\\%o' in program", tok
->val
.c
);
665 *value
= HT_IDENT_TO_GCC_IDENT (HT_NODE (tok
->val
.node
));
670 unsigned int flags
= cpp_classify_number (parse_in
, tok
);
672 switch (flags
& CPP_N_CATEGORY
)
675 /* cpplib has issued an error. */
676 *value
= error_mark_node
;
680 *value
= interpret_integer (tok
, flags
);
684 *value
= interpret_float (tok
, flags
);
695 *value
= lex_charconst (tok
);
700 *value
= lex_string (tok
->val
.str
.text
, tok
->val
.str
.len
,
701 tok
->type
== CPP_WSTRING
);
704 /* These tokens should not be visible outside cpplib. */
705 case CPP_HEADER_NAME
:
716 /* Returns the narrowest C-visible unsigned type, starting with the
717 minimum specified by FLAGS, that can fit VALUE, or itk_none if
719 static enum integer_type_kind
720 narrowest_unsigned_type (value
, flags
)
724 enum integer_type_kind itk
;
726 if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
)
727 itk
= itk_unsigned_int
;
728 else if ((flags
& CPP_N_WIDTH
) == CPP_N_MEDIUM
)
729 itk
= itk_unsigned_long
;
731 itk
= itk_unsigned_long_long
;
733 /* int_fits_type_p must think the type of its first argument is
734 wider than its second argument, or it won't do the proper check. */
735 TREE_TYPE (value
) = widest_unsigned_literal_type_node
;
737 for (; itk
< itk_none
; itk
+= 2 /* skip unsigned types */)
738 if (int_fits_type_p (value
, integer_types
[itk
]))
744 /* Ditto, but narrowest signed type. */
745 static enum integer_type_kind
746 narrowest_signed_type (value
, flags
)
750 enum integer_type_kind itk
;
752 if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
)
754 else if ((flags
& CPP_N_WIDTH
) == CPP_N_MEDIUM
)
759 /* int_fits_type_p must think the type of its first argument is
760 wider than its second argument, or it won't do the proper check. */
761 TREE_TYPE (value
) = widest_unsigned_literal_type_node
;
763 for (; itk
< itk_none
; itk
+= 2 /* skip signed types */)
764 if (int_fits_type_p (value
, integer_types
[itk
]))
770 /* Interpret TOKEN, an integer with FLAGS as classified by cpplib. */
772 interpret_integer (token
, flags
)
773 const cpp_token
*token
;
777 enum integer_type_kind itk
;
779 cpp_options
*options
= cpp_get_options (parse_in
);
781 integer
= cpp_interpret_integer (parse_in
, token
, flags
);
782 integer
= cpp_num_sign_extend (integer
, options
->precision
);
783 value
= build_int_2_wide (integer
.low
, integer
.high
);
785 /* The type of a constant with a U suffix is straightforward. */
786 if (flags
& CPP_N_UNSIGNED
)
787 itk
= narrowest_unsigned_type (value
, flags
);
790 /* The type of a potentially-signed integer constant varies
791 depending on the base it's in, the standard in use, and the
793 enum integer_type_kind itk_u
= narrowest_unsigned_type (value
, flags
);
794 enum integer_type_kind itk_s
= narrowest_signed_type (value
, flags
);
796 /* In both C89 and C99, octal and hex constants may be signed or
797 unsigned, whichever fits tighter. We do not warn about this
798 choice differing from the traditional choice, as the constant
799 is probably a bit pattern and either way will work. */
800 if ((flags
& CPP_N_RADIX
) != CPP_N_DECIMAL
)
801 itk
= MIN (itk_u
, itk_s
);
804 /* In C99, decimal constants are always signed.
805 In C89, decimal constants that don't fit in long have
806 undefined behavior; we try to make them unsigned long.
807 In GCC's extended C89, that last is true of decimal
808 constants that don't fit in long long, too. */
811 if (itk_s
> itk_u
&& itk_s
> itk_long
)
815 if (itk_u
< itk_unsigned_long
)
816 itk_u
= itk_unsigned_long
;
818 warning ("this decimal constant is unsigned only in ISO C90");
820 else if (warn_traditional
)
821 warning ("this decimal constant would be unsigned in ISO C90");
827 /* cpplib has already issued a warning for overflow. */
828 type
= ((flags
& CPP_N_UNSIGNED
)
829 ? widest_unsigned_literal_type_node
830 : widest_integer_literal_type_node
);
832 type
= integer_types
[itk
];
834 if (itk
> itk_unsigned_long
835 && (flags
& CPP_N_WIDTH
) != CPP_N_LARGE
836 && ! in_system_header
&& ! flag_isoc99
)
837 pedwarn ("integer constant is too large for \"%s\" type",
838 (flags
& CPP_N_UNSIGNED
) ? "unsigned long" : "long");
840 TREE_TYPE (value
) = type
;
842 /* Convert imaginary to a complex type. */
843 if (flags
& CPP_N_IMAGINARY
)
844 value
= build_complex (NULL_TREE
, convert (type
, integer_zero_node
), value
);
849 /* Interpret TOKEN, a floating point number with FLAGS as classified
852 interpret_float (token
, flags
)
853 const cpp_token
*token
;
858 REAL_VALUE_TYPE real
;
861 const char *typename
;
863 /* FIXME: make %T work in error/warning, then we don't need typename. */
864 if ((flags
& CPP_N_WIDTH
) == CPP_N_LARGE
)
866 type
= long_double_type_node
;
867 typename
= "long double";
869 else if ((flags
& CPP_N_WIDTH
) == CPP_N_SMALL
870 || flag_single_precision_constant
)
872 type
= float_type_node
;
877 type
= double_type_node
;
881 /* Copy the constant to a nul-terminated buffer. If the constant
882 has any suffixes, cut them off; REAL_VALUE_ATOF/ REAL_VALUE_HTOF
883 can't handle them. */
884 copylen
= token
->val
.str
.len
;
885 if ((flags
& CPP_N_WIDTH
) != CPP_N_MEDIUM
)
886 /* Must be an F or L suffix. */
888 if (flags
& CPP_N_IMAGINARY
)
892 copy
= alloca (copylen
+ 1);
893 memcpy (copy
, token
->val
.str
.text
, copylen
);
894 copy
[copylen
] = '\0';
896 real_from_string (&real
, copy
);
897 real_convert (&real
, TYPE_MODE (type
), &real
);
899 /* A diagnostic is required for "soft" overflow by some ISO C
900 testsuites. This is not pedwarn, because some people don't want
902 ??? That's a dubious reason... is this a mandatory diagnostic or
903 isn't it? -- zw, 2001-08-21. */
904 if (REAL_VALUE_ISINF (real
) && pedantic
)
905 warning ("floating constant exceeds range of \"%s\"", typename
);
907 /* Create a node with determined type and value. */
908 value
= build_real (type
, real
);
909 if (flags
& CPP_N_IMAGINARY
)
910 value
= build_complex (NULL_TREE
, convert (type
, integer_zero_node
), value
);
916 lex_string (str
, len
, wide
)
917 const unsigned char *str
;
922 char *buf
= alloca ((len
+ 1) * (wide
? WCHAR_BYTES
: 1));
924 const unsigned char *p
= str
, *limit
= str
+ len
;
927 #ifdef MULTIBYTE_CHARS
928 /* Reset multibyte conversion state. */
929 (void) local_mbtowc (NULL
, NULL
, 0);
934 #ifdef MULTIBYTE_CHARS
938 char_len
= local_mbtowc (&wc
, (const char *) p
, limit
- p
);
941 warning ("ignoring invalid multibyte character");
954 if (c
== '\\' && !ignore_escape_flag
)
955 c
= cpp_parse_escape (parse_in
, &p
, limit
, wide
);
957 /* Add this single character into the buffer either as a wchar_t,
958 a multibyte sequence, or as a single byte. */
961 unsigned charwidth
= TYPE_PRECISION (char_type_node
);
962 unsigned bytemask
= (1 << charwidth
) - 1;
965 for (byte
= 0; byte
< WCHAR_BYTES
; ++byte
)
968 if (byte
>= (int) sizeof (c
))
971 n
= (c
>> (byte
* charwidth
)) & bytemask
;
972 if (BYTES_BIG_ENDIAN
)
973 q
[WCHAR_BYTES
- byte
- 1] = n
;
979 #ifdef MULTIBYTE_CHARS
980 else if (char_len
> 1)
982 /* We're dealing with a multibyte character. */
983 for ( ; char_len
>0; --char_len
)
985 *q
++ = *(p
- char_len
);
995 /* Terminate the string value, either with a single byte zero
996 or with a wide zero. */
1000 memset (q
, 0, WCHAR_BYTES
);
1008 value
= build_string (q
- buf
, buf
);
1011 TREE_TYPE (value
) = wchar_array_type_node
;
1013 TREE_TYPE (value
) = char_array_type_node
;
1017 /* Converts a (possibly wide) character constant token into a tree. */
1019 lex_charconst (token
)
1020 const cpp_token
*token
;
1024 unsigned int chars_seen
;
1027 result
= cpp_interpret_charconst (parse_in
, token
,
1028 &chars_seen
, &unsignedp
);
1030 /* Cast to cppchar_signed_t to get correct sign-extension of RESULT
1031 before possibly widening to HOST_WIDE_INT for build_int_2. */
1032 if (unsignedp
|| (cppchar_signed_t
) result
>= 0)
1033 value
= build_int_2 (result
, 0);
1035 value
= build_int_2 ((cppchar_signed_t
) result
, -1);
1037 if (token
->type
== CPP_WCHAR
)
1038 type
= wchar_type_node
;
1039 /* In C, a character constant has type 'int'.
1040 In C++ 'char', but multi-char charconsts have type 'int'. */
1041 else if ((c_language
== clk_c
) || chars_seen
> 1)
1042 type
= integer_type_node
;
1044 type
= char_type_node
;
1046 TREE_TYPE (value
) = type
;