1 /* Lexical analyzer for C and Objective C.
2 Copyright (C) 1987, 1988, 1989, 1992, 1994, 1995, 1996, 1997
3 1998, 1999, 2000 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
39 #include "splay-tree.h"
42 #ifdef MULTIBYTE_CHARS
45 #endif /* MULTIBYTE_CHARS */
46 #ifndef GET_ENVIRONMENT
47 #define GET_ENVIRONMENT(ENV_VALUE,ENV_NAME) ((ENV_VALUE) = getenv (ENV_NAME))
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 lex_number
PARAMS ((const char *, unsigned int));
76 static tree lex_string
PARAMS ((const unsigned char *, unsigned int,
78 static tree lex_charconst
PARAMS ((const cpp_token
*));
79 static void update_header_times
PARAMS ((const char *));
80 static int dump_one_header
PARAMS ((splay_tree_node
, void *));
81 static void cb_line_change
PARAMS ((cpp_reader
*, const cpp_token
*, int));
82 static void cb_ident
PARAMS ((cpp_reader
*, unsigned int,
84 static void cb_file_change
PARAMS ((cpp_reader
*, const struct line_map
*));
85 static void cb_def_pragma
PARAMS ((cpp_reader
*, unsigned int));
86 static void cb_define
PARAMS ((cpp_reader
*, unsigned int,
88 static void cb_undef
PARAMS ((cpp_reader
*, unsigned int,
95 struct cpp_callbacks
*cb
;
96 struct c_fileinfo
*toplevel
;
98 /* Set up filename timing. Must happen before cpp_read_main_file. */
99 file_info_tree
= splay_tree_new ((splay_tree_compare_fn
)strcmp
,
101 (splay_tree_delete_value_fn
)free
);
102 toplevel
= get_fileinfo ("<top level>");
103 if (flag_detailed_statistics
)
106 body_time
= get_run_time ();
107 toplevel
->time
= body_time
;
110 #ifdef MULTIBYTE_CHARS
111 /* Change to the native locale for multibyte conversions. */
112 setlocale (LC_CTYPE
, "");
113 GET_ENVIRONMENT (literal_codeset
, "LANG");
116 cb
= cpp_get_callbacks (parse_in
);
118 cb
->line_change
= cb_line_change
;
119 cb
->ident
= cb_ident
;
120 cb
->file_change
= cb_file_change
;
121 cb
->def_pragma
= cb_def_pragma
;
123 /* Set the debug callbacks if we can use them. */
124 if (debug_info_level
== DINFO_LEVEL_VERBOSE
125 && (write_symbols
== DWARF_DEBUG
|| write_symbols
== DWARF2_DEBUG
126 || write_symbols
== VMS_AND_DWARF2_DEBUG
))
128 cb
->define
= cb_define
;
129 cb
->undef
= cb_undef
;
135 if (filename
== NULL
|| !strcmp (filename
, "-"))
138 return cpp_read_main_file (parse_in
, filename
, ident_hash
);
141 /* A thin wrapper around the real parser that initializes the
142 integrated preprocessor after debug output has been initialized.
143 Also, make sure the start_source_file debug hook gets called for
144 the primary source file. */
147 c_common_parse_file (set_yydebug
)
148 int set_yydebug ATTRIBUTE_UNUSED
;
151 yydebug
= set_yydebug
;
153 warning ("YYDEBUG not defined");
156 (*debug_hooks
->start_source_file
) (lineno
, input_filename
);
157 cpp_finish_options (parse_in
);
160 free_parser_stacks ();
168 struct c_fileinfo
*fi
;
170 n
= splay_tree_lookup (file_info_tree
, (splay_tree_key
) name
);
172 return (struct c_fileinfo
*) n
->value
;
174 fi
= (struct c_fileinfo
*) xmalloc (sizeof (struct c_fileinfo
));
176 fi
->interface_only
= 0;
177 fi
->interface_unknown
= 1;
178 splay_tree_insert (file_info_tree
, (splay_tree_key
) name
,
179 (splay_tree_value
) fi
);
184 update_header_times (name
)
187 /* Changing files again. This means currently collected time
188 is charged against header time, and body time starts back at 0. */
189 if (flag_detailed_statistics
)
191 int this_time
= get_run_time ();
192 struct c_fileinfo
*file
= get_fileinfo (name
);
193 header_time
+= this_time
- body_time
;
194 file
->time
+= this_time
- body_time
;
195 body_time
= this_time
;
200 dump_one_header (n
, dummy
)
202 void *dummy ATTRIBUTE_UNUSED
;
204 print_time ((const char *) n
->key
,
205 ((struct c_fileinfo
*) n
->value
)->time
);
210 dump_time_statistics ()
212 struct c_fileinfo
*file
= get_fileinfo (input_filename
);
213 int this_time
= get_run_time ();
214 file
->time
+= this_time
- body_time
;
216 fprintf (stderr
, "\n******\n");
217 print_time ("header files (total)", header_time
);
218 print_time ("main file (total)", this_time
- body_time
);
219 fprintf (stderr
, "ratio = %g : 1\n",
220 (double)header_time
/ (double)(this_time
- body_time
));
221 fprintf (stderr
, "\n******\n");
223 splay_tree_foreach (file_info_tree
, dump_one_header
, 0);
227 cb_ident (pfile
, line
, str
)
228 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
229 unsigned int line ATTRIBUTE_UNUSED
;
230 const cpp_string
*str ATTRIBUTE_UNUSED
;
232 #ifdef ASM_OUTPUT_IDENT
235 /* Convert escapes in the string. */
236 tree value
= lex_string (str
->text
, str
->len
, 0);
237 ASM_OUTPUT_IDENT (asm_out_file
, TREE_STRING_POINTER (value
));
242 /* Called at the start of every non-empty line. TOKEN is the first
243 lexed token on the line. Used for diagnostic line numbers. */
245 cb_line_change (pfile
, token
, parsing_args
)
246 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
247 const cpp_token
*token
;
248 int parsing_args ATTRIBUTE_UNUSED
;
250 src_lineno
= SOURCE_LINE (map
, token
->line
);
254 cb_file_change (pfile
, new_map
)
255 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
256 const struct line_map
*new_map
;
258 unsigned int to_line
= SOURCE_LINE (new_map
, new_map
->to_line
);
260 if (new_map
->reason
== LC_ENTER
)
262 /* Don't stack the main buffer on the input stack;
263 we already did in compile_file. */
265 main_input_filename
= new_map
->to_file
;
268 int included_at
= SOURCE_LINE (new_map
- 1, new_map
->from_line
- 1);
270 lineno
= included_at
;
271 push_srcloc (new_map
->to_file
, 1);
272 (*debug_hooks
->start_source_file
) (included_at
, new_map
->to_file
);
273 #ifndef NO_IMPLICIT_EXTERN_C
276 else if (new_map
->sysp
== 2)
279 ++pending_lang_change
;
284 else if (new_map
->reason
== LC_LEAVE
)
286 #ifndef NO_IMPLICIT_EXTERN_C
287 if (c_header_level
&& --c_header_level
== 0)
289 if (new_map
->sysp
== 2)
290 warning ("badly nested C headers from preprocessor");
291 --pending_lang_change
;
296 (*debug_hooks
->end_source_file
) (to_line
);
299 update_header_times (new_map
->to_file
);
300 in_system_header
= new_map
->sysp
!= 0;
301 input_filename
= new_map
->to_file
;
306 extract_interface_info ();
310 cb_def_pragma (pfile
, line
)
314 /* Issue a warning message if we have been asked to do so. Ignore
315 unknown pragmas in system headers unless an explicit
316 -Wunknown-pragmas has been given. */
317 if (warn_unknown_pragmas
> in_system_header
)
319 const unsigned char *space
, *name
= 0;
322 s
= cpp_get_token (pfile
);
323 space
= cpp_token_as_text (pfile
, s
);
324 s
= cpp_get_token (pfile
);
325 if (s
->type
== CPP_NAME
)
326 name
= cpp_token_as_text (pfile
, s
);
328 lineno
= SOURCE_LINE (map
, line
);
330 warning ("ignoring #pragma %s %s", space
, name
);
332 warning ("ignoring #pragma %s", space
);
336 /* #define callback for DWARF and DWARF2 debug info. */
338 cb_define (pfile
, line
, node
)
343 (*debug_hooks
->define
) (SOURCE_LINE (map
, line
),
344 (const char *) cpp_macro_definition (pfile
, node
));
347 /* #undef callback for DWARF and DWARF2 debug info. */
349 cb_undef (pfile
, line
, node
)
350 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
354 (*debug_hooks
->undef
) (SOURCE_LINE (map
, line
),
355 (const char *) NODE_NAME (node
));
359 /* Returns nonzero if C is a universal-character-name. Give an error if it
360 is not one which may appear in an identifier, as per [extendid].
362 Note that extended character support in identifiers has not yet been
363 implemented. It is my personal opinion that this is not a desirable
364 feature. Portable code cannot count on support for more than the basic
365 identifier character set. */
378 /* None of the valid chars are outside the Basic Multilingual Plane (the
382 error ("universal-character-name '\\U%08x' not valid in identifier", c
);
387 if ((c
>= 0x00c0 && c
<= 0x00d6)
388 || (c
>= 0x00d8 && c
<= 0x00f6)
389 || (c
>= 0x00f8 && c
<= 0x01f5)
390 || (c
>= 0x01fa && c
<= 0x0217)
391 || (c
>= 0x0250 && c
<= 0x02a8)
392 || (c
>= 0x1e00 && c
<= 0x1e9a)
393 || (c
>= 0x1ea0 && c
<= 0x1ef9))
398 || (c
>= 0x0388 && c
<= 0x038a)
400 || (c
>= 0x038e && c
<= 0x03a1)
401 || (c
>= 0x03a3 && c
<= 0x03ce)
402 || (c
>= 0x03d0 && c
<= 0x03d6)
407 || (c
>= 0x03e2 && c
<= 0x03f3)
408 || (c
>= 0x1f00 && c
<= 0x1f15)
409 || (c
>= 0x1f18 && c
<= 0x1f1d)
410 || (c
>= 0x1f20 && c
<= 0x1f45)
411 || (c
>= 0x1f48 && c
<= 0x1f4d)
412 || (c
>= 0x1f50 && c
<= 0x1f57)
416 || (c
>= 0x1f5f && c
<= 0x1f7d)
417 || (c
>= 0x1f80 && c
<= 0x1fb4)
418 || (c
>= 0x1fb6 && c
<= 0x1fbc)
419 || (c
>= 0x1fc2 && c
<= 0x1fc4)
420 || (c
>= 0x1fc6 && c
<= 0x1fcc)
421 || (c
>= 0x1fd0 && c
<= 0x1fd3)
422 || (c
>= 0x1fd6 && c
<= 0x1fdb)
423 || (c
>= 0x1fe0 && c
<= 0x1fec)
424 || (c
>= 0x1ff2 && c
<= 0x1ff4)
425 || (c
>= 0x1ff6 && c
<= 0x1ffc))
429 if ((c
>= 0x0401 && c
<= 0x040d)
430 || (c
>= 0x040f && c
<= 0x044f)
431 || (c
>= 0x0451 && c
<= 0x045c)
432 || (c
>= 0x045e && c
<= 0x0481)
433 || (c
>= 0x0490 && c
<= 0x04c4)
434 || (c
>= 0x04c7 && c
<= 0x04c8)
435 || (c
>= 0x04cb && c
<= 0x04cc)
436 || (c
>= 0x04d0 && c
<= 0x04eb)
437 || (c
>= 0x04ee && c
<= 0x04f5)
438 || (c
>= 0x04f8 && c
<= 0x04f9))
442 if ((c
>= 0x0531 && c
<= 0x0556)
443 || (c
>= 0x0561 && c
<= 0x0587))
447 if ((c
>= 0x05d0 && c
<= 0x05ea)
448 || (c
>= 0x05f0 && c
<= 0x05f4))
452 if ((c
>= 0x0621 && c
<= 0x063a)
453 || (c
>= 0x0640 && c
<= 0x0652)
454 || (c
>= 0x0670 && c
<= 0x06b7)
455 || (c
>= 0x06ba && c
<= 0x06be)
456 || (c
>= 0x06c0 && c
<= 0x06ce)
457 || (c
>= 0x06e5 && c
<= 0x06e7))
461 if ((c
>= 0x0905 && c
<= 0x0939)
462 || (c
>= 0x0958 && c
<= 0x0962))
466 if ((c
>= 0x0985 && c
<= 0x098c)
467 || (c
>= 0x098f && c
<= 0x0990)
468 || (c
>= 0x0993 && c
<= 0x09a8)
469 || (c
>= 0x09aa && c
<= 0x09b0)
471 || (c
>= 0x09b6 && c
<= 0x09b9)
472 || (c
>= 0x09dc && c
<= 0x09dd)
473 || (c
>= 0x09df && c
<= 0x09e1)
474 || (c
>= 0x09f0 && c
<= 0x09f1))
478 if ((c
>= 0x0a05 && c
<= 0x0a0a)
479 || (c
>= 0x0a0f && c
<= 0x0a10)
480 || (c
>= 0x0a13 && c
<= 0x0a28)
481 || (c
>= 0x0a2a && c
<= 0x0a30)
482 || (c
>= 0x0a32 && c
<= 0x0a33)
483 || (c
>= 0x0a35 && c
<= 0x0a36)
484 || (c
>= 0x0a38 && c
<= 0x0a39)
485 || (c
>= 0x0a59 && c
<= 0x0a5c)
490 if ((c
>= 0x0a85 && c
<= 0x0a8b)
492 || (c
>= 0x0a8f && c
<= 0x0a91)
493 || (c
>= 0x0a93 && c
<= 0x0aa8)
494 || (c
>= 0x0aaa && c
<= 0x0ab0)
495 || (c
>= 0x0ab2 && c
<= 0x0ab3)
496 || (c
>= 0x0ab5 && c
<= 0x0ab9)
501 if ((c
>= 0x0b05 && c
<= 0x0b0c)
502 || (c
>= 0x0b0f && c
<= 0x0b10)
503 || (c
>= 0x0b13 && c
<= 0x0b28)
504 || (c
>= 0x0b2a && c
<= 0x0b30)
505 || (c
>= 0x0b32 && c
<= 0x0b33)
506 || (c
>= 0x0b36 && c
<= 0x0b39)
507 || (c
>= 0x0b5c && c
<= 0x0b5d)
508 || (c
>= 0x0b5f && c
<= 0x0b61))
512 if ((c
>= 0x0b85 && c
<= 0x0b8a)
513 || (c
>= 0x0b8e && c
<= 0x0b90)
514 || (c
>= 0x0b92 && c
<= 0x0b95)
515 || (c
>= 0x0b99 && c
<= 0x0b9a)
517 || (c
>= 0x0b9e && c
<= 0x0b9f)
518 || (c
>= 0x0ba3 && c
<= 0x0ba4)
519 || (c
>= 0x0ba8 && c
<= 0x0baa)
520 || (c
>= 0x0bae && c
<= 0x0bb5)
521 || (c
>= 0x0bb7 && c
<= 0x0bb9))
525 if ((c
>= 0x0c05 && c
<= 0x0c0c)
526 || (c
>= 0x0c0e && c
<= 0x0c10)
527 || (c
>= 0x0c12 && c
<= 0x0c28)
528 || (c
>= 0x0c2a && c
<= 0x0c33)
529 || (c
>= 0x0c35 && c
<= 0x0c39)
530 || (c
>= 0x0c60 && c
<= 0x0c61))
534 if ((c
>= 0x0c85 && c
<= 0x0c8c)
535 || (c
>= 0x0c8e && c
<= 0x0c90)
536 || (c
>= 0x0c92 && c
<= 0x0ca8)
537 || (c
>= 0x0caa && c
<= 0x0cb3)
538 || (c
>= 0x0cb5 && c
<= 0x0cb9)
539 || (c
>= 0x0ce0 && c
<= 0x0ce1))
543 if ((c
>= 0x0d05 && c
<= 0x0d0c)
544 || (c
>= 0x0d0e && c
<= 0x0d10)
545 || (c
>= 0x0d12 && c
<= 0x0d28)
546 || (c
>= 0x0d2a && c
<= 0x0d39)
547 || (c
>= 0x0d60 && c
<= 0x0d61))
551 if ((c
>= 0x0e01 && c
<= 0x0e30)
552 || (c
>= 0x0e32 && c
<= 0x0e33)
553 || (c
>= 0x0e40 && c
<= 0x0e46)
554 || (c
>= 0x0e4f && c
<= 0x0e5b))
558 if ((c
>= 0x0e81 && c
<= 0x0e82)
564 || (c
>= 0x0e94 && c
<= 0x0e97)
565 || (c
>= 0x0e99 && c
<= 0x0e9f)
566 || (c
>= 0x0ea1 && c
<= 0x0ea3)
571 || (c
>= 0x0ead && c
<= 0x0eb0)
575 || (c
>= 0x0ec0 && c
<= 0x0ec4)
580 if ((c
>= 0x10a0 && c
<= 0x10c5)
581 || (c
>= 0x10d0 && c
<= 0x10f6))
585 if ((c
>= 0x3041 && c
<= 0x3094)
586 || (c
>= 0x309b && c
<= 0x309e))
590 if ((c
>= 0x30a1 && c
<= 0x30fe))
594 if ((c
>= 0x3105 && c
<= 0x312c))
598 if ((c
>= 0x1100 && c
<= 0x1159)
599 || (c
>= 0x1161 && c
<= 0x11a2)
600 || (c
>= 0x11a8 && c
<= 0x11f9))
603 /* CJK Unified Ideographs */
604 if ((c
>= 0xf900 && c
<= 0xfa2d)
605 || (c
>= 0xfb1f && c
<= 0xfb36)
606 || (c
>= 0xfb38 && c
<= 0xfb3c)
608 || (c
>= 0xfb40 && c
<= 0xfb41)
609 || (c
>= 0xfb42 && c
<= 0xfb44)
610 || (c
>= 0xfb46 && c
<= 0xfbb1)
611 || (c
>= 0xfbd3 && c
<= 0xfd3f)
612 || (c
>= 0xfd50 && c
<= 0xfd8f)
613 || (c
>= 0xfd92 && c
<= 0xfdc7)
614 || (c
>= 0xfdf0 && c
<= 0xfdfb)
615 || (c
>= 0xfe70 && c
<= 0xfe72)
617 || (c
>= 0xfe76 && c
<= 0xfefc)
618 || (c
>= 0xff21 && c
<= 0xff3a)
619 || (c
>= 0xff41 && c
<= 0xff5a)
620 || (c
>= 0xff66 && c
<= 0xffbe)
621 || (c
>= 0xffc2 && c
<= 0xffc7)
622 || (c
>= 0xffca && c
<= 0xffcf)
623 || (c
>= 0xffd2 && c
<= 0xffd7)
624 || (c
>= 0xffda && c
<= 0xffdc)
625 || (c
>= 0x4e00 && c
<= 0x9fa5))
628 error ("universal-character-name '\\u%04x' not valid in identifier", c
);
633 /* Add the UTF-8 representation of C to the token_buffer. */
636 utf8_extend_token (c
)
646 else if (c
<= 0x000007ff)
647 shift
= 6, mask
= 0xc0;
648 else if (c
<= 0x0000ffff)
649 shift
= 12, mask
= 0xe0;
650 else if (c
<= 0x001fffff)
651 shift
= 18, mask
= 0xf0;
652 else if (c
<= 0x03ffffff)
653 shift
= 24, mask
= 0xf8;
655 shift
= 30, mask
= 0xfc;
657 extend_token (mask
| (c
>> shift
));
661 extend_token ((unsigned char) (0x80 | (c
>> shift
)));
670 tree
*const node_var
;
671 const char unsigned_flag
;
672 const char long_flag
;
673 const char long_long_flag
;
676 struct try_type type_sequence
[] =
678 { &integer_type_node
, 0, 0, 0},
679 { &unsigned_type_node
, 1, 0, 0},
680 { &long_integer_type_node
, 0, 1, 0},
681 { &long_unsigned_type_node
, 1, 1, 0},
682 { &long_long_integer_type_node
, 0, 1, 1},
683 { &long_long_unsigned_type_node
, 1, 1, 1}
691 const cpp_token
*tok
;
694 timevar_push (TV_CPP
);
696 tok
= cpp_get_token (parse_in
);
697 while (tok
->type
== CPP_PADDING
);
698 timevar_pop (TV_CPP
);
700 /* The C++ front end does horrible things with the current line
701 number. To ensure an accurate line number, we must reset it
702 every time we return a token. */
708 /* Issue this error here, where we can get at tok->val.c. */
710 if (ISGRAPH (tok
->val
.c
))
711 error ("stray '%c' in program", tok
->val
.c
);
713 error ("stray '\\%o' in program", tok
->val
.c
);
717 *value
= HT_IDENT_TO_GCC_IDENT (HT_NODE (tok
->val
.node
));
721 *value
= lex_number ((const char *)tok
->val
.str
.text
, tok
->val
.str
.len
);
726 *value
= lex_charconst (tok
);
731 *value
= lex_string (tok
->val
.str
.text
, tok
->val
.str
.len
,
732 tok
->type
== CPP_WSTRING
);
735 /* These tokens should not be visible outside cpplib. */
736 case CPP_HEADER_NAME
:
747 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
750 lex_number (str
, len
)
756 int largest_digit
= 0;
762 enum anon1
{ NOT_FLOAT
= 0, AFTER_POINT
, AFTER_EXPON
} floatflag
= NOT_FLOAT
;
764 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
765 The code below which fills the parts array assumes that a host
766 int is at least twice as wide as a host char, and that
767 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
768 Two HOST_WIDE_INTs is the largest int literal we can store.
769 In order to detect overflow below, the number of parts (TOTAL_PARTS)
770 must be exactly the number of parts needed to hold the bits
771 of two HOST_WIDE_INTs. */
772 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
773 unsigned int parts
[TOTAL_PARTS
];
775 /* Optimize for most frequent case. */
779 return integer_zero_node
;
780 else if (*str
== '1')
781 return integer_one_node
;
783 return build_int_2 (*str
- '0', 0);
786 for (count
= 0; count
< TOTAL_PARTS
; count
++)
789 /* len is known to be >1 at this point. */
792 if (len
> 2 && str
[0] == '0' && (str
[1] == 'x' || str
[1] == 'X'))
797 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
798 else if (str
[0] == '0' && ISDIGIT (str
[1]))
810 if (floatflag
== AFTER_POINT
)
811 ERROR ("too many decimal points in floating constant");
812 else if (floatflag
== AFTER_EXPON
)
813 ERROR ("decimal point in exponent - impossible!");
815 floatflag
= AFTER_POINT
;
821 /* Possible future extension: silently ignore _ in numbers,
822 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
823 but somewhat easier to read. Ada has this? */
824 ERROR ("underscore in number");
828 /* It is not a decimal point.
829 It should be a digit (perhaps a hex digit). */
832 || (base
== 16 && ISXDIGIT (c
)))
836 else if (base
<= 10 && (c
== 'e' || c
== 'E'))
839 floatflag
= AFTER_EXPON
;
842 else if (base
== 16 && (c
== 'p' || c
== 'P'))
844 floatflag
= AFTER_EXPON
;
845 break; /* start of exponent */
850 break; /* start of suffix */
853 if (n
>= largest_digit
)
857 for (count
= 0; count
< TOTAL_PARTS
; count
++)
859 parts
[count
] *= base
;
863 += (parts
[count
-1] >> HOST_BITS_PER_CHAR
);
865 &= (1 << HOST_BITS_PER_CHAR
) - 1;
871 /* If the highest-order part overflows (gets larger than
872 a host char will hold) then the whole number has
873 overflowed. Record this and truncate the highest-order
875 if (parts
[TOTAL_PARTS
- 1] >> HOST_BITS_PER_CHAR
)
878 parts
[TOTAL_PARTS
- 1] &= (1 << HOST_BITS_PER_CHAR
) - 1;
882 while (p
< str
+ len
);
884 /* This can happen on input like `int i = 0x;' */
886 ERROR ("numeric constant with no digits");
888 if (largest_digit
>= base
)
889 ERROR ("numeric constant contains digits beyond the radix");
891 if (floatflag
!= NOT_FLOAT
)
894 const char *typename
;
895 int imag
, fflag
, lflag
;
896 REAL_VALUE_TYPE real
;
899 if (base
== 16 && floatflag
!= AFTER_EXPON
)
900 ERROR ("hexadecimal floating constant has no exponent");
902 /* Read explicit exponent if any, and put it in tokenbuf. */
903 if ((base
== 10 && ((c
== 'e') || (c
== 'E')))
904 || (base
== 16 && (c
== 'p' || c
== 'P')))
908 if (p
< str
+ len
&& (c
== '+' || c
== '-'))
910 /* Exponent is decimal, even if string is a hex float. */
912 ERROR ("floating constant exponent has no digits");
913 while (p
< str
+ len
&& ISDIGIT (c
))
919 /* Copy the float constant now; we don't want any suffixes in the
920 string passed to parse_float. */
921 copy
= alloca (p
- str
+ 1);
922 memcpy (copy
, str
, p
- str
);
923 copy
[p
- str
] = '\0';
925 /* Now parse suffixes. */
926 fflag
= lflag
= imag
= 0;
927 while (p
< str
+ len
)
932 ERROR ("more than one 'f' suffix on floating constant");
933 else if (warn_traditional
&& !in_system_header
934 && ! cpp_sys_macro_p (parse_in
))
935 warning ("traditional C rejects the 'f' suffix");
942 ERROR ("more than one 'l' suffix on floating constant");
943 else if (warn_traditional
&& !in_system_header
944 && ! cpp_sys_macro_p (parse_in
))
945 warning ("traditional C rejects the 'l' suffix");
953 ERROR ("more than one 'i' or 'j' suffix on floating constant");
955 pedwarn ("ISO C forbids imaginary numeric constants");
960 ERROR ("invalid suffix on floating constant");
963 type
= double_type_node
;
969 ERROR ("both 'f' and 'l' suffixes on floating constant");
971 type
= float_type_node
;
976 type
= long_double_type_node
;
977 typename
= "long double";
979 else if (flag_single_precision_constant
)
981 type
= float_type_node
;
985 /* Warn about this only after we know we're not issuing an error. */
986 if (base
== 16 && pedantic
&& !flag_isoc99
)
987 pedwarn ("hexadecimal floating constants are only valid in C99");
989 /* The second argument, machine_mode, of REAL_VALUE_ATOF
990 tells the desired precision of the binary result
991 of decimal-to-binary conversion. */
993 real
= REAL_VALUE_HTOF (copy
, TYPE_MODE (type
));
995 real
= REAL_VALUE_ATOF (copy
, TYPE_MODE (type
));
997 /* A diagnostic is required here by some ISO C testsuites.
998 This is not pedwarn, because some people don't want
999 an error for this. */
1000 if (REAL_VALUE_ISINF (real
) && pedantic
)
1001 warning ("floating point number exceeds range of 'double'");
1003 /* Create a node with determined type and value. */
1005 value
= build_complex (NULL_TREE
, convert (type
, integer_zero_node
),
1006 build_real (type
, real
));
1008 value
= build_real (type
, real
);
1012 tree trad_type
, type
;
1013 HOST_WIDE_INT high
, low
;
1014 int spec_unsigned
= 0;
1016 int spec_long_long
= 0;
1021 trad_type
= type
= NULL_TREE
;
1022 while (p
< str
+ len
)
1029 error ("two 'u' suffixes on integer constant");
1030 else if (warn_traditional
&& !in_system_header
1031 && ! cpp_sys_macro_p (parse_in
))
1032 warning ("traditional C rejects the 'u' suffix");
1043 error ("three 'l' suffixes on integer constant");
1045 error ("'lul' is not a valid integer suffix");
1046 else if (c
!= spec_long
)
1047 error ("'Ll' and 'lL' are not valid integer suffixes");
1048 else if (pedantic
&& ! flag_isoc99
1049 && ! in_system_header
&& warn_long_long
)
1050 pedwarn ("ISO C89 forbids long long integer constants");
1056 case 'i': case 'I': case 'j': case 'J':
1058 error ("more than one 'i' or 'j' suffix on integer constant");
1060 pedwarn ("ISO C forbids imaginary numeric constants");
1065 ERROR ("invalid suffix on integer constant");
1069 /* If the literal overflowed, pedwarn about it now. */
1073 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT
* 2);
1076 /* This is simplified by the fact that our constant
1077 is always positive. */
1081 for (i
= 0; i
< HOST_BITS_PER_WIDE_INT
/ HOST_BITS_PER_CHAR
; i
++)
1083 high
|= ((HOST_WIDE_INT
) parts
[i
+ (HOST_BITS_PER_WIDE_INT
1084 / HOST_BITS_PER_CHAR
)]
1085 << (i
* HOST_BITS_PER_CHAR
));
1086 low
|= (HOST_WIDE_INT
) parts
[i
] << (i
* HOST_BITS_PER_CHAR
);
1089 value
= build_int_2 (low
, high
);
1090 TREE_TYPE (value
) = long_long_unsigned_type_node
;
1092 /* If warn_traditional, calculate both the ISO type and the
1093 traditional type, then see if they disagree. */
1094 if (warn_traditional
)
1096 /* Traditionally, any constant is signed; but if unsigned is
1097 specified explicitly, obey that. Use the smallest size
1098 with the right number of bits, except for one special
1099 case with decimal constants. */
1100 if (! spec_long
&& base
!= 10
1101 && int_fits_type_p (value
, unsigned_type_node
))
1102 trad_type
= spec_unsigned
? unsigned_type_node
: integer_type_node
;
1103 /* A decimal constant must be long if it does not fit in
1104 type int. I think this is independent of whether the
1105 constant is signed. */
1106 else if (! spec_long
&& base
== 10
1107 && int_fits_type_p (value
, integer_type_node
))
1108 trad_type
= spec_unsigned
? unsigned_type_node
: integer_type_node
;
1109 else if (! spec_long_long
)
1110 trad_type
= (spec_unsigned
1111 ? long_unsigned_type_node
1112 : long_integer_type_node
);
1113 else if (int_fits_type_p (value
,
1115 ? long_long_unsigned_type_node
1116 : long_long_integer_type_node
))
1117 trad_type
= (spec_unsigned
1118 ? long_long_unsigned_type_node
1119 : long_long_integer_type_node
);
1121 trad_type
= (spec_unsigned
1122 ? widest_unsigned_literal_type_node
1123 : widest_integer_literal_type_node
);
1126 /* Calculate the ISO type. */
1127 if (! spec_long
&& ! spec_unsigned
1128 && int_fits_type_p (value
, integer_type_node
))
1129 type
= integer_type_node
;
1130 else if (! spec_long
&& (base
!= 10 || spec_unsigned
)
1131 && int_fits_type_p (value
, unsigned_type_node
))
1132 type
= unsigned_type_node
;
1133 else if (! spec_unsigned
&& !spec_long_long
1134 && int_fits_type_p (value
, long_integer_type_node
))
1135 type
= long_integer_type_node
;
1136 else if (! spec_long_long
1137 && int_fits_type_p (value
, long_unsigned_type_node
))
1138 type
= long_unsigned_type_node
;
1139 else if (! spec_unsigned
1140 && int_fits_type_p (value
, long_long_integer_type_node
))
1141 type
= long_long_integer_type_node
;
1142 else if (int_fits_type_p (value
, long_long_unsigned_type_node
))
1143 type
= long_long_unsigned_type_node
;
1144 else if (! spec_unsigned
1145 && int_fits_type_p (value
, widest_integer_literal_type_node
))
1146 type
= widest_integer_literal_type_node
;
1148 type
= widest_unsigned_literal_type_node
;
1150 /* We assume that constants specified in a non-decimal
1151 base are bit patterns, and that the programmer really
1152 meant what they wrote. */
1153 if (warn_traditional
&& !in_system_header
1154 && base
== 10 && trad_type
!= type
)
1156 if (TYPE_PRECISION (trad_type
) != TYPE_PRECISION (type
))
1157 warning ("width of integer constant is different in traditional C");
1158 else if (TREE_UNSIGNED (trad_type
) != TREE_UNSIGNED (type
))
1159 warning ("integer constant is unsigned in ISO C, signed in traditional C");
1161 warning ("width of integer constant may change on other systems in traditional C");
1164 if (pedantic
&& (flag_isoc99
|| !spec_long_long
)
1167 ? TYPE_PRECISION (long_long_integer_type_node
)
1168 : TYPE_PRECISION (long_integer_type_node
)) < TYPE_PRECISION (type
)))
1171 pedwarn ("integer constant larger than the maximum value of %s",
1173 ? (TREE_UNSIGNED (type
)
1174 ? _("an unsigned long long int")
1175 : _("a long long int"))
1176 : _("an unsigned long int")));
1179 if (base
== 10 && ! spec_unsigned
&& TREE_UNSIGNED (type
))
1180 warning ("decimal constant is so large that it is unsigned");
1184 if (TYPE_PRECISION (type
)
1185 <= TYPE_PRECISION (integer_type_node
))
1186 value
= build_complex (NULL_TREE
, integer_zero_node
,
1187 convert (integer_type_node
, value
));
1189 ERROR ("complex integer constant is too wide for 'complex int'");
1192 TREE_TYPE (value
) = type
;
1194 /* If it's still an integer (not a complex), and it doesn't
1195 fit in the type we choose for it, then pedwarn. */
1198 && TREE_CODE (TREE_TYPE (value
)) == INTEGER_TYPE
1199 && ! int_fits_type_p (value
, TREE_TYPE (value
)))
1200 pedwarn ("integer constant is larger than the maximum value for its type");
1204 error ("missing white space after number '%.*s'", (int) (p
- str
), str
);
1209 return integer_zero_node
;
1213 lex_string (str
, len
, wide
)
1214 const unsigned char *str
;
1219 char *buf
= alloca ((len
+ 1) * (wide
? WCHAR_BYTES
: 1));
1221 const unsigned char *p
= str
, *limit
= str
+ len
;
1224 #ifdef MULTIBYTE_CHARS
1225 /* Reset multibyte conversion state. */
1226 (void) local_mbtowc (NULL
, NULL
, 0);
1231 #ifdef MULTIBYTE_CHARS
1235 char_len
= local_mbtowc (&wc
, (const char *) p
, limit
- p
);
1238 warning ("ignoring invalid multibyte character");
1251 if (c
== '\\' && !ignore_escape_flag
)
1252 c
= cpp_parse_escape (parse_in
, &p
, limit
, wide
);
1254 /* Add this single character into the buffer either as a wchar_t,
1255 a multibyte sequence, or as a single byte. */
1258 unsigned charwidth
= TYPE_PRECISION (char_type_node
);
1259 unsigned bytemask
= (1 << charwidth
) - 1;
1262 for (byte
= 0; byte
< WCHAR_BYTES
; ++byte
)
1265 if (byte
>= (int) sizeof (c
))
1268 n
= (c
>> (byte
* charwidth
)) & bytemask
;
1269 if (BYTES_BIG_ENDIAN
)
1270 q
[WCHAR_BYTES
- byte
- 1] = n
;
1276 #ifdef MULTIBYTE_CHARS
1277 else if (char_len
> 1)
1279 /* We're dealing with a multibyte character. */
1280 for ( ; char_len
>0; --char_len
)
1282 *q
++ = *(p
- char_len
);
1292 /* Terminate the string value, either with a single byte zero
1293 or with a wide zero. */
1297 memset (q
, 0, WCHAR_BYTES
);
1305 value
= build_string (q
- buf
, buf
);
1308 TREE_TYPE (value
) = wchar_array_type_node
;
1310 TREE_TYPE (value
) = char_array_type_node
;
1314 /* Converts a (possibly wide) character constant token into a tree. */
1316 lex_charconst (token
)
1317 const cpp_token
*token
;
1321 unsigned int chars_seen
;
1324 result
= cpp_interpret_charconst (parse_in
, token
,
1325 &chars_seen
, &unsignedp
);
1327 /* Cast to cppchar_signed_t to get correct sign-extension of RESULT
1328 before possibly widening to HOST_WIDE_INT for build_int_2. */
1329 if (unsignedp
|| (cppchar_signed_t
) result
>= 0)
1330 value
= build_int_2 (result
, 0);
1332 value
= build_int_2 ((cppchar_signed_t
) result
, -1);
1334 if (token
->type
== CPP_WCHAR
)
1335 type
= wchar_type_node
;
1336 /* In C, a character constant has type 'int'.
1337 In C++ 'char', but multi-char charconsts have type 'int'. */
1338 else if ((c_language
== clk_c
|| c_language
== clk_objective_c
)
1340 type
= integer_type_node
;
1342 type
= char_type_node
;
1344 TREE_TYPE (value
) = type
;