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
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 lex_number
PARAMS ((const char *, unsigned int));
77 static tree lex_string
PARAMS ((const unsigned char *, unsigned int,
79 static tree lex_charconst
PARAMS ((const cpp_token
*));
80 static void update_header_times
PARAMS ((const char *));
81 static int dump_one_header
PARAMS ((splay_tree_node
, void *));
82 static void cb_line_change
PARAMS ((cpp_reader
*, const cpp_token
*, int));
83 static void cb_ident
PARAMS ((cpp_reader
*, unsigned int,
85 static void cb_file_change
PARAMS ((cpp_reader
*, const struct line_map
*));
86 static void cb_def_pragma
PARAMS ((cpp_reader
*, unsigned int));
87 static void cb_define
PARAMS ((cpp_reader
*, unsigned int,
89 static void cb_undef
PARAMS ((cpp_reader
*, unsigned int,
96 struct cpp_callbacks
*cb
;
97 struct c_fileinfo
*toplevel
;
99 /* Set up filename timing. Must happen before cpp_read_main_file. */
100 file_info_tree
= splay_tree_new ((splay_tree_compare_fn
)strcmp
,
102 (splay_tree_delete_value_fn
)free
);
103 toplevel
= get_fileinfo ("<top level>");
104 if (flag_detailed_statistics
)
107 body_time
= get_run_time ();
108 toplevel
->time
= body_time
;
111 #ifdef MULTIBYTE_CHARS
112 /* Change to the native locale for multibyte conversions. */
113 setlocale (LC_CTYPE
, "");
114 GET_ENVIRONMENT (literal_codeset
, "LANG");
117 cb
= cpp_get_callbacks (parse_in
);
119 cb
->line_change
= cb_line_change
;
120 cb
->ident
= cb_ident
;
121 cb
->file_change
= cb_file_change
;
122 cb
->def_pragma
= cb_def_pragma
;
124 /* Set the debug callbacks if we can use them. */
125 if (debug_info_level
== DINFO_LEVEL_VERBOSE
126 && (write_symbols
== DWARF_DEBUG
|| write_symbols
== DWARF2_DEBUG
127 || write_symbols
== VMS_AND_DWARF2_DEBUG
))
129 cb
->define
= cb_define
;
130 cb
->undef
= cb_undef
;
136 if (filename
== NULL
|| !strcmp (filename
, "-"))
139 return cpp_read_main_file (parse_in
, filename
, ident_hash
);
142 /* A thin wrapper around the real parser that initializes the
143 integrated preprocessor after debug output has been initialized.
144 Also, make sure the start_source_file debug hook gets called for
145 the primary source file. */
148 c_common_parse_file (set_yydebug
)
149 int set_yydebug ATTRIBUTE_UNUSED
;
152 yydebug
= set_yydebug
;
154 warning ("YYDEBUG not defined");
157 (*debug_hooks
->start_source_file
) (lineno
, input_filename
);
158 cpp_finish_options (parse_in
);
161 free_parser_stacks ();
169 struct c_fileinfo
*fi
;
171 n
= splay_tree_lookup (file_info_tree
, (splay_tree_key
) name
);
173 return (struct c_fileinfo
*) n
->value
;
175 fi
= (struct c_fileinfo
*) xmalloc (sizeof (struct c_fileinfo
));
177 fi
->interface_only
= 0;
178 fi
->interface_unknown
= 1;
179 splay_tree_insert (file_info_tree
, (splay_tree_key
) name
,
180 (splay_tree_value
) fi
);
185 update_header_times (name
)
188 /* Changing files again. This means currently collected time
189 is charged against header time, and body time starts back at 0. */
190 if (flag_detailed_statistics
)
192 int this_time
= get_run_time ();
193 struct c_fileinfo
*file
= get_fileinfo (name
);
194 header_time
+= this_time
- body_time
;
195 file
->time
+= this_time
- body_time
;
196 body_time
= this_time
;
201 dump_one_header (n
, dummy
)
203 void *dummy ATTRIBUTE_UNUSED
;
205 print_time ((const char *) n
->key
,
206 ((struct c_fileinfo
*) n
->value
)->time
);
211 dump_time_statistics ()
213 struct c_fileinfo
*file
= get_fileinfo (input_filename
);
214 int this_time
= get_run_time ();
215 file
->time
+= this_time
- body_time
;
217 fprintf (stderr
, "\n******\n");
218 print_time ("header files (total)", header_time
);
219 print_time ("main file (total)", this_time
- body_time
);
220 fprintf (stderr
, "ratio = %g : 1\n",
221 (double)header_time
/ (double)(this_time
- body_time
));
222 fprintf (stderr
, "\n******\n");
224 splay_tree_foreach (file_info_tree
, dump_one_header
, 0);
228 cb_ident (pfile
, line
, str
)
229 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
230 unsigned int line ATTRIBUTE_UNUSED
;
231 const cpp_string
*str ATTRIBUTE_UNUSED
;
233 #ifdef ASM_OUTPUT_IDENT
236 /* Convert escapes in the string. */
237 tree value
= lex_string (str
->text
, str
->len
, 0);
238 ASM_OUTPUT_IDENT (asm_out_file
, TREE_STRING_POINTER (value
));
243 /* Called at the start of every non-empty line. TOKEN is the first
244 lexed token on the line. Used for diagnostic line numbers. */
246 cb_line_change (pfile
, token
, parsing_args
)
247 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
248 const cpp_token
*token
;
249 int parsing_args ATTRIBUTE_UNUSED
;
251 src_lineno
= SOURCE_LINE (map
, token
->line
);
255 cb_file_change (pfile
, new_map
)
256 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
257 const struct line_map
*new_map
;
259 unsigned int to_line
= SOURCE_LINE (new_map
, new_map
->to_line
);
261 if (new_map
->reason
== LC_ENTER
)
263 /* Don't stack the main buffer on the input stack;
264 we already did in compile_file. */
266 main_input_filename
= new_map
->to_file
;
269 int included_at
= SOURCE_LINE (new_map
- 1, new_map
->from_line
- 1);
271 lineno
= included_at
;
272 push_srcloc (new_map
->to_file
, 1);
273 (*debug_hooks
->start_source_file
) (included_at
, new_map
->to_file
);
274 #ifndef NO_IMPLICIT_EXTERN_C
277 else if (new_map
->sysp
== 2)
280 ++pending_lang_change
;
285 else if (new_map
->reason
== LC_LEAVE
)
287 #ifndef NO_IMPLICIT_EXTERN_C
288 if (c_header_level
&& --c_header_level
== 0)
290 if (new_map
->sysp
== 2)
291 warning ("badly nested C headers from preprocessor");
292 --pending_lang_change
;
297 (*debug_hooks
->end_source_file
) (to_line
);
300 update_header_times (new_map
->to_file
);
301 in_system_header
= new_map
->sysp
!= 0;
302 input_filename
= new_map
->to_file
;
307 extract_interface_info ();
311 cb_def_pragma (pfile
, line
)
315 /* Issue a warning message if we have been asked to do so. Ignore
316 unknown pragmas in system headers unless an explicit
317 -Wunknown-pragmas has been given. */
318 if (warn_unknown_pragmas
> in_system_header
)
320 const unsigned char *space
, *name
= 0;
323 s
= cpp_get_token (pfile
);
324 space
= cpp_token_as_text (pfile
, s
);
325 s
= cpp_get_token (pfile
);
326 if (s
->type
== CPP_NAME
)
327 name
= cpp_token_as_text (pfile
, s
);
329 lineno
= SOURCE_LINE (map
, line
);
331 warning ("ignoring #pragma %s %s", space
, name
);
333 warning ("ignoring #pragma %s", space
);
337 /* #define callback for DWARF and DWARF2 debug info. */
339 cb_define (pfile
, line
, node
)
344 (*debug_hooks
->define
) (SOURCE_LINE (map
, line
),
345 (const char *) cpp_macro_definition (pfile
, node
));
348 /* #undef callback for DWARF and DWARF2 debug info. */
350 cb_undef (pfile
, line
, node
)
351 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
355 (*debug_hooks
->undef
) (SOURCE_LINE (map
, line
),
356 (const char *) NODE_NAME (node
));
360 /* Returns nonzero if C is a universal-character-name. Give an error if it
361 is not one which may appear in an identifier, as per [extendid].
363 Note that extended character support in identifiers has not yet been
364 implemented. It is my personal opinion that this is not a desirable
365 feature. Portable code cannot count on support for more than the basic
366 identifier character set. */
379 /* None of the valid chars are outside the Basic Multilingual Plane (the
383 error ("universal-character-name '\\U%08x' not valid in identifier", c
);
388 if ((c
>= 0x00c0 && c
<= 0x00d6)
389 || (c
>= 0x00d8 && c
<= 0x00f6)
390 || (c
>= 0x00f8 && c
<= 0x01f5)
391 || (c
>= 0x01fa && c
<= 0x0217)
392 || (c
>= 0x0250 && c
<= 0x02a8)
393 || (c
>= 0x1e00 && c
<= 0x1e9a)
394 || (c
>= 0x1ea0 && c
<= 0x1ef9))
399 || (c
>= 0x0388 && c
<= 0x038a)
401 || (c
>= 0x038e && c
<= 0x03a1)
402 || (c
>= 0x03a3 && c
<= 0x03ce)
403 || (c
>= 0x03d0 && c
<= 0x03d6)
408 || (c
>= 0x03e2 && c
<= 0x03f3)
409 || (c
>= 0x1f00 && c
<= 0x1f15)
410 || (c
>= 0x1f18 && c
<= 0x1f1d)
411 || (c
>= 0x1f20 && c
<= 0x1f45)
412 || (c
>= 0x1f48 && c
<= 0x1f4d)
413 || (c
>= 0x1f50 && c
<= 0x1f57)
417 || (c
>= 0x1f5f && c
<= 0x1f7d)
418 || (c
>= 0x1f80 && c
<= 0x1fb4)
419 || (c
>= 0x1fb6 && c
<= 0x1fbc)
420 || (c
>= 0x1fc2 && c
<= 0x1fc4)
421 || (c
>= 0x1fc6 && c
<= 0x1fcc)
422 || (c
>= 0x1fd0 && c
<= 0x1fd3)
423 || (c
>= 0x1fd6 && c
<= 0x1fdb)
424 || (c
>= 0x1fe0 && c
<= 0x1fec)
425 || (c
>= 0x1ff2 && c
<= 0x1ff4)
426 || (c
>= 0x1ff6 && c
<= 0x1ffc))
430 if ((c
>= 0x0401 && c
<= 0x040d)
431 || (c
>= 0x040f && c
<= 0x044f)
432 || (c
>= 0x0451 && c
<= 0x045c)
433 || (c
>= 0x045e && c
<= 0x0481)
434 || (c
>= 0x0490 && c
<= 0x04c4)
435 || (c
>= 0x04c7 && c
<= 0x04c8)
436 || (c
>= 0x04cb && c
<= 0x04cc)
437 || (c
>= 0x04d0 && c
<= 0x04eb)
438 || (c
>= 0x04ee && c
<= 0x04f5)
439 || (c
>= 0x04f8 && c
<= 0x04f9))
443 if ((c
>= 0x0531 && c
<= 0x0556)
444 || (c
>= 0x0561 && c
<= 0x0587))
448 if ((c
>= 0x05d0 && c
<= 0x05ea)
449 || (c
>= 0x05f0 && c
<= 0x05f4))
453 if ((c
>= 0x0621 && c
<= 0x063a)
454 || (c
>= 0x0640 && c
<= 0x0652)
455 || (c
>= 0x0670 && c
<= 0x06b7)
456 || (c
>= 0x06ba && c
<= 0x06be)
457 || (c
>= 0x06c0 && c
<= 0x06ce)
458 || (c
>= 0x06e5 && c
<= 0x06e7))
462 if ((c
>= 0x0905 && c
<= 0x0939)
463 || (c
>= 0x0958 && c
<= 0x0962))
467 if ((c
>= 0x0985 && c
<= 0x098c)
468 || (c
>= 0x098f && c
<= 0x0990)
469 || (c
>= 0x0993 && c
<= 0x09a8)
470 || (c
>= 0x09aa && c
<= 0x09b0)
472 || (c
>= 0x09b6 && c
<= 0x09b9)
473 || (c
>= 0x09dc && c
<= 0x09dd)
474 || (c
>= 0x09df && c
<= 0x09e1)
475 || (c
>= 0x09f0 && c
<= 0x09f1))
479 if ((c
>= 0x0a05 && c
<= 0x0a0a)
480 || (c
>= 0x0a0f && c
<= 0x0a10)
481 || (c
>= 0x0a13 && c
<= 0x0a28)
482 || (c
>= 0x0a2a && c
<= 0x0a30)
483 || (c
>= 0x0a32 && c
<= 0x0a33)
484 || (c
>= 0x0a35 && c
<= 0x0a36)
485 || (c
>= 0x0a38 && c
<= 0x0a39)
486 || (c
>= 0x0a59 && c
<= 0x0a5c)
491 if ((c
>= 0x0a85 && c
<= 0x0a8b)
493 || (c
>= 0x0a8f && c
<= 0x0a91)
494 || (c
>= 0x0a93 && c
<= 0x0aa8)
495 || (c
>= 0x0aaa && c
<= 0x0ab0)
496 || (c
>= 0x0ab2 && c
<= 0x0ab3)
497 || (c
>= 0x0ab5 && c
<= 0x0ab9)
502 if ((c
>= 0x0b05 && c
<= 0x0b0c)
503 || (c
>= 0x0b0f && c
<= 0x0b10)
504 || (c
>= 0x0b13 && c
<= 0x0b28)
505 || (c
>= 0x0b2a && c
<= 0x0b30)
506 || (c
>= 0x0b32 && c
<= 0x0b33)
507 || (c
>= 0x0b36 && c
<= 0x0b39)
508 || (c
>= 0x0b5c && c
<= 0x0b5d)
509 || (c
>= 0x0b5f && c
<= 0x0b61))
513 if ((c
>= 0x0b85 && c
<= 0x0b8a)
514 || (c
>= 0x0b8e && c
<= 0x0b90)
515 || (c
>= 0x0b92 && c
<= 0x0b95)
516 || (c
>= 0x0b99 && c
<= 0x0b9a)
518 || (c
>= 0x0b9e && c
<= 0x0b9f)
519 || (c
>= 0x0ba3 && c
<= 0x0ba4)
520 || (c
>= 0x0ba8 && c
<= 0x0baa)
521 || (c
>= 0x0bae && c
<= 0x0bb5)
522 || (c
>= 0x0bb7 && c
<= 0x0bb9))
526 if ((c
>= 0x0c05 && c
<= 0x0c0c)
527 || (c
>= 0x0c0e && c
<= 0x0c10)
528 || (c
>= 0x0c12 && c
<= 0x0c28)
529 || (c
>= 0x0c2a && c
<= 0x0c33)
530 || (c
>= 0x0c35 && c
<= 0x0c39)
531 || (c
>= 0x0c60 && c
<= 0x0c61))
535 if ((c
>= 0x0c85 && c
<= 0x0c8c)
536 || (c
>= 0x0c8e && c
<= 0x0c90)
537 || (c
>= 0x0c92 && c
<= 0x0ca8)
538 || (c
>= 0x0caa && c
<= 0x0cb3)
539 || (c
>= 0x0cb5 && c
<= 0x0cb9)
540 || (c
>= 0x0ce0 && c
<= 0x0ce1))
544 if ((c
>= 0x0d05 && c
<= 0x0d0c)
545 || (c
>= 0x0d0e && c
<= 0x0d10)
546 || (c
>= 0x0d12 && c
<= 0x0d28)
547 || (c
>= 0x0d2a && c
<= 0x0d39)
548 || (c
>= 0x0d60 && c
<= 0x0d61))
552 if ((c
>= 0x0e01 && c
<= 0x0e30)
553 || (c
>= 0x0e32 && c
<= 0x0e33)
554 || (c
>= 0x0e40 && c
<= 0x0e46)
555 || (c
>= 0x0e4f && c
<= 0x0e5b))
559 if ((c
>= 0x0e81 && c
<= 0x0e82)
565 || (c
>= 0x0e94 && c
<= 0x0e97)
566 || (c
>= 0x0e99 && c
<= 0x0e9f)
567 || (c
>= 0x0ea1 && c
<= 0x0ea3)
572 || (c
>= 0x0ead && c
<= 0x0eb0)
576 || (c
>= 0x0ec0 && c
<= 0x0ec4)
581 if ((c
>= 0x10a0 && c
<= 0x10c5)
582 || (c
>= 0x10d0 && c
<= 0x10f6))
586 if ((c
>= 0x3041 && c
<= 0x3094)
587 || (c
>= 0x309b && c
<= 0x309e))
591 if ((c
>= 0x30a1 && c
<= 0x30fe))
595 if ((c
>= 0x3105 && c
<= 0x312c))
599 if ((c
>= 0x1100 && c
<= 0x1159)
600 || (c
>= 0x1161 && c
<= 0x11a2)
601 || (c
>= 0x11a8 && c
<= 0x11f9))
604 /* CJK Unified Ideographs */
605 if ((c
>= 0xf900 && c
<= 0xfa2d)
606 || (c
>= 0xfb1f && c
<= 0xfb36)
607 || (c
>= 0xfb38 && c
<= 0xfb3c)
609 || (c
>= 0xfb40 && c
<= 0xfb41)
610 || (c
>= 0xfb42 && c
<= 0xfb44)
611 || (c
>= 0xfb46 && c
<= 0xfbb1)
612 || (c
>= 0xfbd3 && c
<= 0xfd3f)
613 || (c
>= 0xfd50 && c
<= 0xfd8f)
614 || (c
>= 0xfd92 && c
<= 0xfdc7)
615 || (c
>= 0xfdf0 && c
<= 0xfdfb)
616 || (c
>= 0xfe70 && c
<= 0xfe72)
618 || (c
>= 0xfe76 && c
<= 0xfefc)
619 || (c
>= 0xff21 && c
<= 0xff3a)
620 || (c
>= 0xff41 && c
<= 0xff5a)
621 || (c
>= 0xff66 && c
<= 0xffbe)
622 || (c
>= 0xffc2 && c
<= 0xffc7)
623 || (c
>= 0xffca && c
<= 0xffcf)
624 || (c
>= 0xffd2 && c
<= 0xffd7)
625 || (c
>= 0xffda && c
<= 0xffdc)
626 || (c
>= 0x4e00 && c
<= 0x9fa5))
629 error ("universal-character-name '\\u%04x' not valid in identifier", c
);
634 /* Add the UTF-8 representation of C to the token_buffer. */
637 utf8_extend_token (c
)
647 else if (c
<= 0x000007ff)
648 shift
= 6, mask
= 0xc0;
649 else if (c
<= 0x0000ffff)
650 shift
= 12, mask
= 0xe0;
651 else if (c
<= 0x001fffff)
652 shift
= 18, mask
= 0xf0;
653 else if (c
<= 0x03ffffff)
654 shift
= 24, mask
= 0xf8;
656 shift
= 30, mask
= 0xfc;
658 extend_token (mask
| (c
>> shift
));
662 extend_token ((unsigned char) (0x80 | (c
>> shift
)));
671 tree
*const node_var
;
672 const char unsigned_flag
;
673 const char long_flag
;
674 const char long_long_flag
;
677 struct try_type type_sequence
[] =
679 { &integer_type_node
, 0, 0, 0},
680 { &unsigned_type_node
, 1, 0, 0},
681 { &long_integer_type_node
, 0, 1, 0},
682 { &long_unsigned_type_node
, 1, 1, 0},
683 { &long_long_integer_type_node
, 0, 1, 1},
684 { &long_long_unsigned_type_node
, 1, 1, 1}
692 const cpp_token
*tok
;
695 timevar_push (TV_CPP
);
697 tok
= cpp_get_token (parse_in
);
698 while (tok
->type
== CPP_PADDING
);
699 timevar_pop (TV_CPP
);
701 /* The C++ front end does horrible things with the current line
702 number. To ensure an accurate line number, we must reset it
703 every time we return a token. */
709 /* Issue this error here, where we can get at tok->val.c. */
711 if (ISGRAPH (tok
->val
.c
))
712 error ("stray '%c' in program", tok
->val
.c
);
714 error ("stray '\\%o' in program", tok
->val
.c
);
718 *value
= HT_IDENT_TO_GCC_IDENT (HT_NODE (tok
->val
.node
));
722 *value
= lex_number ((const char *)tok
->val
.str
.text
, tok
->val
.str
.len
);
727 *value
= lex_charconst (tok
);
732 *value
= lex_string (tok
->val
.str
.text
, tok
->val
.str
.len
,
733 tok
->type
== CPP_WSTRING
);
736 /* These tokens should not be visible outside cpplib. */
737 case CPP_HEADER_NAME
:
748 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
751 lex_number (str
, len
)
757 int largest_digit
= 0;
763 enum anon1
{ NOT_FLOAT
= 0, AFTER_POINT
, AFTER_EXPON
} floatflag
= NOT_FLOAT
;
765 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
766 The code below which fills the parts array assumes that a host
767 int is at least twice as wide as a host char, and that
768 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
769 Two HOST_WIDE_INTs is the largest int literal we can store.
770 In order to detect overflow below, the number of parts (TOTAL_PARTS)
771 must be exactly the number of parts needed to hold the bits
772 of two HOST_WIDE_INTs. */
773 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
774 unsigned int parts
[TOTAL_PARTS
];
776 /* Optimize for most frequent case. */
780 return integer_zero_node
;
781 else if (*str
== '1')
782 return integer_one_node
;
784 return build_int_2 (*str
- '0', 0);
787 for (count
= 0; count
< TOTAL_PARTS
; count
++)
790 /* len is known to be >1 at this point. */
793 if (len
> 2 && str
[0] == '0' && (str
[1] == 'x' || str
[1] == 'X'))
798 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
799 else if (str
[0] == '0' && ISDIGIT (str
[1]))
811 if (floatflag
== AFTER_POINT
)
812 ERROR ("too many decimal points in floating constant");
813 else if (floatflag
== AFTER_EXPON
)
814 ERROR ("decimal point in exponent - impossible!");
816 floatflag
= AFTER_POINT
;
822 /* Possible future extension: silently ignore _ in numbers,
823 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
824 but somewhat easier to read. Ada has this? */
825 ERROR ("underscore in number");
829 /* It is not a decimal point.
830 It should be a digit (perhaps a hex digit). */
833 || (base
== 16 && ISXDIGIT (c
)))
837 else if (base
<= 10 && (c
== 'e' || c
== 'E'))
840 floatflag
= AFTER_EXPON
;
843 else if (base
== 16 && (c
== 'p' || c
== 'P'))
845 floatflag
= AFTER_EXPON
;
846 break; /* start of exponent */
851 break; /* start of suffix */
854 if (n
>= largest_digit
)
858 for (count
= 0; count
< TOTAL_PARTS
; count
++)
860 parts
[count
] *= base
;
864 += (parts
[count
-1] >> HOST_BITS_PER_CHAR
);
866 &= (1 << HOST_BITS_PER_CHAR
) - 1;
872 /* If the highest-order part overflows (gets larger than
873 a host char will hold) then the whole number has
874 overflowed. Record this and truncate the highest-order
876 if (parts
[TOTAL_PARTS
- 1] >> HOST_BITS_PER_CHAR
)
879 parts
[TOTAL_PARTS
- 1] &= (1 << HOST_BITS_PER_CHAR
) - 1;
883 while (p
< str
+ len
);
885 /* This can happen on input like `int i = 0x;' */
887 ERROR ("numeric constant with no digits");
889 if (largest_digit
>= base
)
890 ERROR ("numeric constant contains digits beyond the radix");
892 if (floatflag
!= NOT_FLOAT
)
895 const char *typename
;
896 int imag
, fflag
, lflag
;
897 REAL_VALUE_TYPE real
;
900 if (base
== 16 && floatflag
!= AFTER_EXPON
)
901 ERROR ("hexadecimal floating constant has no exponent");
903 /* Read explicit exponent if any, and put it in tokenbuf. */
904 if ((base
== 10 && ((c
== 'e') || (c
== 'E')))
905 || (base
== 16 && (c
== 'p' || c
== 'P')))
909 if (p
< str
+ len
&& (c
== '+' || c
== '-'))
911 /* Exponent is decimal, even if string is a hex float. */
913 ERROR ("floating constant exponent has no digits");
914 while (p
< str
+ len
&& ISDIGIT (c
))
920 /* Copy the float constant now; we don't want any suffixes in the
921 string passed to parse_float. */
922 copy
= alloca (p
- str
+ 1);
923 memcpy (copy
, str
, p
- str
);
924 copy
[p
- str
] = '\0';
926 /* Now parse suffixes. */
927 fflag
= lflag
= imag
= 0;
928 while (p
< str
+ len
)
933 ERROR ("more than one 'f' suffix on floating constant");
934 else if (warn_traditional
&& !in_system_header
935 && ! cpp_sys_macro_p (parse_in
))
936 warning ("traditional C rejects the 'f' suffix");
943 ERROR ("more than one 'l' suffix on floating constant");
944 else if (warn_traditional
&& !in_system_header
945 && ! cpp_sys_macro_p (parse_in
))
946 warning ("traditional C rejects the 'l' suffix");
954 ERROR ("more than one 'i' or 'j' suffix on floating constant");
956 pedwarn ("ISO C forbids imaginary numeric constants");
961 ERROR ("invalid suffix on floating constant");
964 type
= double_type_node
;
970 ERROR ("both 'f' and 'l' suffixes on floating constant");
972 type
= float_type_node
;
977 type
= long_double_type_node
;
978 typename
= "long double";
980 else if (flag_single_precision_constant
)
982 type
= float_type_node
;
986 /* Warn about this only after we know we're not issuing an error. */
987 if (base
== 16 && pedantic
&& !flag_isoc99
)
988 pedwarn ("hexadecimal floating constants are only valid in C99");
990 /* The second argument, machine_mode, of REAL_VALUE_ATOF
991 tells the desired precision of the binary result
992 of decimal-to-binary conversion. */
994 real
= REAL_VALUE_HTOF (copy
, TYPE_MODE (type
));
996 real
= REAL_VALUE_ATOF (copy
, TYPE_MODE (type
));
998 /* A diagnostic is required here by some ISO C testsuites.
999 This is not pedwarn, because some people don't want
1000 an error for this. */
1001 if (REAL_VALUE_ISINF (real
) && pedantic
)
1002 warning ("floating point number exceeds range of 'double'");
1004 /* Create a node with determined type and value. */
1006 value
= build_complex (NULL_TREE
, convert (type
, integer_zero_node
),
1007 build_real (type
, real
));
1009 value
= build_real (type
, real
);
1013 tree trad_type
, type
;
1014 HOST_WIDE_INT high
, low
;
1015 int spec_unsigned
= 0;
1017 int spec_long_long
= 0;
1022 trad_type
= type
= NULL_TREE
;
1023 while (p
< str
+ len
)
1030 error ("two 'u' suffixes on integer constant");
1031 else if (warn_traditional
&& !in_system_header
1032 && ! cpp_sys_macro_p (parse_in
))
1033 warning ("traditional C rejects the 'u' suffix");
1044 error ("three 'l' suffixes on integer constant");
1046 error ("'lul' is not a valid integer suffix");
1047 else if (c
!= spec_long
)
1048 error ("'Ll' and 'lL' are not valid integer suffixes");
1049 else if (pedantic
&& ! flag_isoc99
1050 && ! in_system_header
&& warn_long_long
)
1051 pedwarn ("ISO C89 forbids long long integer constants");
1057 case 'i': case 'I': case 'j': case 'J':
1059 error ("more than one 'i' or 'j' suffix on integer constant");
1061 pedwarn ("ISO C forbids imaginary numeric constants");
1066 ERROR ("invalid suffix on integer constant");
1070 /* If the literal overflowed, pedwarn about it now. */
1074 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT
* 2);
1077 /* This is simplified by the fact that our constant
1078 is always positive. */
1082 for (i
= 0; i
< HOST_BITS_PER_WIDE_INT
/ HOST_BITS_PER_CHAR
; i
++)
1084 high
|= ((HOST_WIDE_INT
) parts
[i
+ (HOST_BITS_PER_WIDE_INT
1085 / HOST_BITS_PER_CHAR
)]
1086 << (i
* HOST_BITS_PER_CHAR
));
1087 low
|= (HOST_WIDE_INT
) parts
[i
] << (i
* HOST_BITS_PER_CHAR
);
1090 value
= build_int_2 (low
, high
);
1091 TREE_TYPE (value
) = long_long_unsigned_type_node
;
1093 /* If warn_traditional, calculate both the ISO type and the
1094 traditional type, then see if they disagree. */
1095 if (warn_traditional
)
1097 /* Traditionally, any constant is signed; but if unsigned is
1098 specified explicitly, obey that. Use the smallest size
1099 with the right number of bits, except for one special
1100 case with decimal constants. */
1101 if (! spec_long
&& base
!= 10
1102 && int_fits_type_p (value
, unsigned_type_node
))
1103 trad_type
= spec_unsigned
? unsigned_type_node
: integer_type_node
;
1104 /* A decimal constant must be long if it does not fit in
1105 type int. I think this is independent of whether the
1106 constant is signed. */
1107 else if (! spec_long
&& base
== 10
1108 && int_fits_type_p (value
, integer_type_node
))
1109 trad_type
= spec_unsigned
? unsigned_type_node
: integer_type_node
;
1110 else if (! spec_long_long
)
1111 trad_type
= (spec_unsigned
1112 ? long_unsigned_type_node
1113 : long_integer_type_node
);
1114 else if (int_fits_type_p (value
,
1116 ? long_long_unsigned_type_node
1117 : long_long_integer_type_node
))
1118 trad_type
= (spec_unsigned
1119 ? long_long_unsigned_type_node
1120 : long_long_integer_type_node
);
1122 trad_type
= (spec_unsigned
1123 ? widest_unsigned_literal_type_node
1124 : widest_integer_literal_type_node
);
1127 /* Calculate the ISO type. */
1128 if (! spec_long
&& ! spec_unsigned
1129 && int_fits_type_p (value
, integer_type_node
))
1130 type
= integer_type_node
;
1131 else if (! spec_long
&& (base
!= 10 || spec_unsigned
)
1132 && int_fits_type_p (value
, unsigned_type_node
))
1133 type
= unsigned_type_node
;
1134 else if (! spec_unsigned
&& !spec_long_long
1135 && int_fits_type_p (value
, long_integer_type_node
))
1136 type
= long_integer_type_node
;
1137 else if (! spec_long_long
1138 && int_fits_type_p (value
, long_unsigned_type_node
))
1139 type
= long_unsigned_type_node
;
1140 else if (! spec_unsigned
1141 && int_fits_type_p (value
, long_long_integer_type_node
))
1142 type
= long_long_integer_type_node
;
1143 else if (int_fits_type_p (value
, long_long_unsigned_type_node
))
1144 type
= long_long_unsigned_type_node
;
1145 else if (! spec_unsigned
1146 && int_fits_type_p (value
, widest_integer_literal_type_node
))
1147 type
= widest_integer_literal_type_node
;
1149 type
= widest_unsigned_literal_type_node
;
1151 /* We assume that constants specified in a non-decimal
1152 base are bit patterns, and that the programmer really
1153 meant what they wrote. */
1154 if (warn_traditional
&& !in_system_header
1155 && base
== 10 && trad_type
!= type
)
1157 if (TYPE_PRECISION (trad_type
) != TYPE_PRECISION (type
))
1158 warning ("width of integer constant is different in traditional C");
1159 else if (TREE_UNSIGNED (trad_type
) != TREE_UNSIGNED (type
))
1160 warning ("integer constant is unsigned in ISO C, signed in traditional C");
1162 warning ("width of integer constant may change on other systems in traditional C");
1165 if (pedantic
&& (flag_isoc99
|| !spec_long_long
)
1168 ? TYPE_PRECISION (long_long_integer_type_node
)
1169 : TYPE_PRECISION (long_integer_type_node
)) < TYPE_PRECISION (type
)))
1172 pedwarn ("integer constant larger than the maximum value of %s",
1174 ? (TREE_UNSIGNED (type
)
1175 ? _("an unsigned long long int")
1176 : _("a long long int"))
1177 : _("an unsigned long int")));
1180 if (base
== 10 && ! spec_unsigned
&& TREE_UNSIGNED (type
))
1181 warning ("decimal constant is so large that it is unsigned");
1185 if (TYPE_PRECISION (type
)
1186 <= TYPE_PRECISION (integer_type_node
))
1187 value
= build_complex (NULL_TREE
, integer_zero_node
,
1188 convert (integer_type_node
, value
));
1190 ERROR ("complex integer constant is too wide for 'complex int'");
1193 TREE_TYPE (value
) = type
;
1195 /* If it's still an integer (not a complex), and it doesn't
1196 fit in the type we choose for it, then pedwarn. */
1199 && TREE_CODE (TREE_TYPE (value
)) == INTEGER_TYPE
1200 && ! int_fits_type_p (value
, TREE_TYPE (value
)))
1201 pedwarn ("integer constant is larger than the maximum value for its type");
1205 error ("missing white space after number '%.*s'", (int) (p
- str
), str
);
1210 return integer_zero_node
;
1214 lex_string (str
, len
, wide
)
1215 const unsigned char *str
;
1220 char *buf
= alloca ((len
+ 1) * (wide
? WCHAR_BYTES
: 1));
1222 const unsigned char *p
= str
, *limit
= str
+ len
;
1225 #ifdef MULTIBYTE_CHARS
1226 /* Reset multibyte conversion state. */
1227 (void) local_mbtowc (NULL
, NULL
, 0);
1232 #ifdef MULTIBYTE_CHARS
1236 char_len
= local_mbtowc (&wc
, (const char *) p
, limit
- p
);
1239 warning ("ignoring invalid multibyte character");
1252 if (c
== '\\' && !ignore_escape_flag
)
1253 c
= cpp_parse_escape (parse_in
, &p
, limit
, wide
);
1255 /* Add this single character into the buffer either as a wchar_t,
1256 a multibyte sequence, or as a single byte. */
1259 unsigned charwidth
= TYPE_PRECISION (char_type_node
);
1260 unsigned bytemask
= (1 << charwidth
) - 1;
1263 for (byte
= 0; byte
< WCHAR_BYTES
; ++byte
)
1266 if (byte
>= (int) sizeof (c
))
1269 n
= (c
>> (byte
* charwidth
)) & bytemask
;
1270 if (BYTES_BIG_ENDIAN
)
1271 q
[WCHAR_BYTES
- byte
- 1] = n
;
1277 #ifdef MULTIBYTE_CHARS
1278 else if (char_len
> 1)
1280 /* We're dealing with a multibyte character. */
1281 for ( ; char_len
>0; --char_len
)
1283 *q
++ = *(p
- char_len
);
1293 /* Terminate the string value, either with a single byte zero
1294 or with a wide zero. */
1298 memset (q
, 0, WCHAR_BYTES
);
1306 value
= build_string (q
- buf
, buf
);
1309 TREE_TYPE (value
) = wchar_array_type_node
;
1311 TREE_TYPE (value
) = char_array_type_node
;
1315 /* Converts a (possibly wide) character constant token into a tree. */
1317 lex_charconst (token
)
1318 const cpp_token
*token
;
1322 unsigned int chars_seen
;
1325 result
= cpp_interpret_charconst (parse_in
, token
,
1326 &chars_seen
, &unsignedp
);
1328 /* Cast to cppchar_signed_t to get correct sign-extension of RESULT
1329 before possibly widening to HOST_WIDE_INT for build_int_2. */
1330 if (unsignedp
|| (cppchar_signed_t
) result
>= 0)
1331 value
= build_int_2 (result
, 0);
1333 value
= build_int_2 ((cppchar_signed_t
) result
, -1);
1335 if (token
->type
== CPP_WCHAR
)
1336 type
= wchar_type_node
;
1337 /* In C, a character constant has type 'int'.
1338 In C++ 'char', but multi-char charconsts have type 'int'. */
1339 else if ((c_language
== clk_c
|| c_language
== clk_objective_c
)
1341 type
= integer_type_node
;
1343 type
= char_type_node
;
1345 TREE_TYPE (value
) = type
;