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 /* MULTIBYTE_CHARS support only works for native compilers.
43 ??? Ideally what we want is to model widechar support after
44 the current floating point support. */
46 #undef MULTIBYTE_CHARS
49 #ifdef MULTIBYTE_CHARS
52 #endif /* MULTIBYTE_CHARS */
53 #ifndef GET_ENVIRONMENT
54 #define GET_ENVIRONMENT(ENV_VALUE,ENV_NAME) ((ENV_VALUE) = getenv (ENV_NAME))
57 /* The current line map. */
58 static const struct line_map
*map
;
60 /* The line used to refresh the lineno global variable after each token. */
61 static unsigned int src_lineno
;
63 /* We may keep statistics about how long which files took to compile. */
64 static int header_time
, body_time
;
65 static splay_tree file_info_tree
;
67 /* Cause the `yydebug' variable to be defined. */
70 /* File used for outputting assembler code. */
71 extern FILE *asm_out_file
;
73 #undef WCHAR_TYPE_SIZE
74 #define WCHAR_TYPE_SIZE TYPE_PRECISION (wchar_type_node)
76 /* Number of bytes in a wide character. */
77 #define WCHAR_BYTES (WCHAR_TYPE_SIZE / BITS_PER_UNIT)
79 int indent_level
; /* Number of { minus number of }. */
80 int pending_lang_change
; /* If we need to switch languages - C++ only */
81 int c_header_level
; /* depth in C headers - C++ only */
83 /* Nonzero tells yylex to ignore \ in string constants. */
84 static int ignore_escape_flag
;
86 static void parse_float
PARAMS ((PTR
));
87 static tree lex_number
PARAMS ((const char *, unsigned int));
88 static tree lex_string
PARAMS ((const unsigned char *, unsigned int,
90 static tree lex_charconst
PARAMS ((const cpp_token
*));
91 static void update_header_times
PARAMS ((const char *));
92 static int dump_one_header
PARAMS ((splay_tree_node
, void *));
93 static void cb_line_change
PARAMS ((cpp_reader
*, const cpp_token
*, int));
94 static void cb_ident
PARAMS ((cpp_reader
*, unsigned int,
96 static void cb_file_change
PARAMS ((cpp_reader
*, const struct line_map
*));
97 static void cb_def_pragma
PARAMS ((cpp_reader
*, unsigned int));
98 static void cb_define
PARAMS ((cpp_reader
*, unsigned int,
100 static void cb_undef
PARAMS ((cpp_reader
*, unsigned int,
104 init_c_lex (filename
)
105 const char *filename
;
107 struct cpp_callbacks
*cb
;
108 struct c_fileinfo
*toplevel
;
110 /* Set up filename timing. Must happen before cpp_read_main_file. */
111 file_info_tree
= splay_tree_new ((splay_tree_compare_fn
)strcmp
,
113 (splay_tree_delete_value_fn
)free
);
114 toplevel
= get_fileinfo ("<top level>");
115 if (flag_detailed_statistics
)
118 body_time
= get_run_time ();
119 toplevel
->time
= body_time
;
122 #ifdef MULTIBYTE_CHARS
123 /* Change to the native locale for multibyte conversions. */
124 setlocale (LC_CTYPE
, "");
125 GET_ENVIRONMENT (literal_codeset
, "LANG");
128 cb
= cpp_get_callbacks (parse_in
);
130 cb
->line_change
= cb_line_change
;
131 cb
->ident
= cb_ident
;
132 cb
->file_change
= cb_file_change
;
133 cb
->def_pragma
= cb_def_pragma
;
135 /* Set the debug callbacks if we can use them. */
136 if (debug_info_level
== DINFO_LEVEL_VERBOSE
137 && (write_symbols
== DWARF_DEBUG
|| write_symbols
== DWARF2_DEBUG
138 || write_symbols
== VMS_AND_DWARF2_DEBUG
))
140 cb
->define
= cb_define
;
141 cb
->undef
= cb_undef
;
147 if (filename
== NULL
|| !strcmp (filename
, "-"))
150 return cpp_read_main_file (parse_in
, filename
, ident_hash
);
153 /* A thin wrapper around the real parser that initializes the
154 integrated preprocessor after debug output has been initialized.
155 Also, make sure the start_source_file debug hook gets called for
156 the primary source file. */
161 (*debug_hooks
->start_source_file
) (lineno
, input_filename
);
162 cpp_finish_options (parse_in
);
172 struct c_fileinfo
*fi
;
174 n
= splay_tree_lookup (file_info_tree
, (splay_tree_key
) name
);
176 return (struct c_fileinfo
*) n
->value
;
178 fi
= (struct c_fileinfo
*) xmalloc (sizeof (struct c_fileinfo
));
180 fi
->interface_only
= 0;
181 fi
->interface_unknown
= 1;
182 splay_tree_insert (file_info_tree
, (splay_tree_key
) name
,
183 (splay_tree_value
) fi
);
188 update_header_times (name
)
191 /* Changing files again. This means currently collected time
192 is charged against header time, and body time starts back at 0. */
193 if (flag_detailed_statistics
)
195 int this_time
= get_run_time ();
196 struct c_fileinfo
*file
= get_fileinfo (name
);
197 header_time
+= this_time
- body_time
;
198 file
->time
+= this_time
- body_time
;
199 body_time
= this_time
;
204 dump_one_header (n
, dummy
)
206 void *dummy ATTRIBUTE_UNUSED
;
208 print_time ((const char *) n
->key
,
209 ((struct c_fileinfo
*) n
->value
)->time
);
214 dump_time_statistics ()
216 struct c_fileinfo
*file
= get_fileinfo (input_filename
);
217 int this_time
= get_run_time ();
218 file
->time
+= this_time
- body_time
;
220 fprintf (stderr
, "\n******\n");
221 print_time ("header files (total)", header_time
);
222 print_time ("main file (total)", this_time
- body_time
);
223 fprintf (stderr
, "ratio = %g : 1\n",
224 (double)header_time
/ (double)(this_time
- body_time
));
225 fprintf (stderr
, "\n******\n");
227 splay_tree_foreach (file_info_tree
, dump_one_header
, 0);
230 /* Not yet handled: #pragma, #define, #undef.
231 No need to deal with linemarkers under normal conditions. */
234 cb_ident (pfile
, line
, str
)
235 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
236 unsigned int line ATTRIBUTE_UNUSED
;
237 const cpp_string
*str ATTRIBUTE_UNUSED
;
239 #ifdef ASM_OUTPUT_IDENT
242 /* Convert escapes in the string. */
243 tree value
= lex_string (str
->text
, str
->len
, 0);
244 ASM_OUTPUT_IDENT (asm_out_file
, TREE_STRING_POINTER (value
));
249 /* Called at the start of every non-empty line. TOKEN is the first
250 lexed token on the line. Used for diagnostic line numbers. */
252 cb_line_change (pfile
, token
, parsing_args
)
253 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
254 const cpp_token
*token
;
255 int parsing_args ATTRIBUTE_UNUSED
;
257 src_lineno
= SOURCE_LINE (map
, token
->line
);
261 cb_file_change (pfile
, new_map
)
262 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
263 const struct line_map
*new_map
;
265 unsigned int to_line
= SOURCE_LINE (new_map
, new_map
->to_line
);
267 if (new_map
->reason
== LC_ENTER
)
269 /* Don't stack the main buffer on the input stack;
270 we already did in compile_file. */
272 main_input_filename
= new_map
->to_file
;
275 int included_at
= SOURCE_LINE (new_map
- 1, new_map
->from_line
- 1);
277 lineno
= included_at
;
278 push_srcloc (new_map
->to_file
, 1);
279 input_file_stack
->indent_level
= indent_level
;
280 (*debug_hooks
->start_source_file
) (included_at
, new_map
->to_file
);
281 #ifndef NO_IMPLICIT_EXTERN_C
284 else if (new_map
->sysp
== 2)
287 ++pending_lang_change
;
292 else if (new_map
->reason
== LC_LEAVE
)
294 #ifndef NO_IMPLICIT_EXTERN_C
295 if (c_header_level
&& --c_header_level
== 0)
297 if (new_map
->sysp
== 2)
298 warning ("badly nested C headers from preprocessor");
299 --pending_lang_change
;
303 if (indent_level
!= input_file_stack
->indent_level
)
305 warning_with_file_and_line
306 (input_filename
, lineno
,
307 "this file contains more '%c's than '%c's",
308 indent_level
> input_file_stack
->indent_level
? '{' : '}',
309 indent_level
> input_file_stack
->indent_level
? '}' : '{');
314 (*debug_hooks
->end_source_file
) (to_line
);
317 update_header_times (new_map
->to_file
);
318 in_system_header
= new_map
->sysp
!= 0;
319 input_filename
= new_map
->to_file
;
324 extract_interface_info ();
328 cb_def_pragma (pfile
, line
)
332 /* Issue a warning message if we have been asked to do so. Ignore
333 unknown pragmas in system headers unless an explicit
334 -Wunknown-pragmas has been given. */
335 if (warn_unknown_pragmas
> in_system_header
)
337 const unsigned char *space
, *name
;
340 space
= name
= (const unsigned char *) "";
341 s
= cpp_get_token (pfile
);
342 if (s
->type
!= CPP_EOF
)
344 space
= cpp_token_as_text (pfile
, s
);
345 s
= cpp_get_token (pfile
);
346 if (s
->type
== CPP_NAME
)
347 name
= cpp_token_as_text (pfile
, s
);
350 lineno
= SOURCE_LINE (map
, line
);
351 warning ("ignoring #pragma %s %s", space
, name
);
355 /* #define callback for DWARF and DWARF2 debug info. */
357 cb_define (pfile
, line
, node
)
362 (*debug_hooks
->define
) (SOURCE_LINE (map
, line
),
363 (const char *) cpp_macro_definition (pfile
, node
));
366 /* #undef callback for DWARF and DWARF2 debug info. */
368 cb_undef (pfile
, line
, node
)
369 cpp_reader
*pfile ATTRIBUTE_UNUSED
;
373 (*debug_hooks
->undef
) (SOURCE_LINE (map
, line
),
374 (const char *) NODE_NAME (node
));
378 /* Returns nonzero if C is a universal-character-name. Give an error if it
379 is not one which may appear in an identifier, as per [extendid].
381 Note that extended character support in identifiers has not yet been
382 implemented. It is my personal opinion that this is not a desirable
383 feature. Portable code cannot count on support for more than the basic
384 identifier character set. */
397 /* None of the valid chars are outside the Basic Multilingual Plane (the
401 error ("universal-character-name '\\U%08x' not valid in identifier", c
);
406 if ((c
>= 0x00c0 && c
<= 0x00d6)
407 || (c
>= 0x00d8 && c
<= 0x00f6)
408 || (c
>= 0x00f8 && c
<= 0x01f5)
409 || (c
>= 0x01fa && c
<= 0x0217)
410 || (c
>= 0x0250 && c
<= 0x02a8)
411 || (c
>= 0x1e00 && c
<= 0x1e9a)
412 || (c
>= 0x1ea0 && c
<= 0x1ef9))
417 || (c
>= 0x0388 && c
<= 0x038a)
419 || (c
>= 0x038e && c
<= 0x03a1)
420 || (c
>= 0x03a3 && c
<= 0x03ce)
421 || (c
>= 0x03d0 && c
<= 0x03d6)
426 || (c
>= 0x03e2 && c
<= 0x03f3)
427 || (c
>= 0x1f00 && c
<= 0x1f15)
428 || (c
>= 0x1f18 && c
<= 0x1f1d)
429 || (c
>= 0x1f20 && c
<= 0x1f45)
430 || (c
>= 0x1f48 && c
<= 0x1f4d)
431 || (c
>= 0x1f50 && c
<= 0x1f57)
435 || (c
>= 0x1f5f && c
<= 0x1f7d)
436 || (c
>= 0x1f80 && c
<= 0x1fb4)
437 || (c
>= 0x1fb6 && c
<= 0x1fbc)
438 || (c
>= 0x1fc2 && c
<= 0x1fc4)
439 || (c
>= 0x1fc6 && c
<= 0x1fcc)
440 || (c
>= 0x1fd0 && c
<= 0x1fd3)
441 || (c
>= 0x1fd6 && c
<= 0x1fdb)
442 || (c
>= 0x1fe0 && c
<= 0x1fec)
443 || (c
>= 0x1ff2 && c
<= 0x1ff4)
444 || (c
>= 0x1ff6 && c
<= 0x1ffc))
448 if ((c
>= 0x0401 && c
<= 0x040d)
449 || (c
>= 0x040f && c
<= 0x044f)
450 || (c
>= 0x0451 && c
<= 0x045c)
451 || (c
>= 0x045e && c
<= 0x0481)
452 || (c
>= 0x0490 && c
<= 0x04c4)
453 || (c
>= 0x04c7 && c
<= 0x04c8)
454 || (c
>= 0x04cb && c
<= 0x04cc)
455 || (c
>= 0x04d0 && c
<= 0x04eb)
456 || (c
>= 0x04ee && c
<= 0x04f5)
457 || (c
>= 0x04f8 && c
<= 0x04f9))
461 if ((c
>= 0x0531 && c
<= 0x0556)
462 || (c
>= 0x0561 && c
<= 0x0587))
466 if ((c
>= 0x05d0 && c
<= 0x05ea)
467 || (c
>= 0x05f0 && c
<= 0x05f4))
471 if ((c
>= 0x0621 && c
<= 0x063a)
472 || (c
>= 0x0640 && c
<= 0x0652)
473 || (c
>= 0x0670 && c
<= 0x06b7)
474 || (c
>= 0x06ba && c
<= 0x06be)
475 || (c
>= 0x06c0 && c
<= 0x06ce)
476 || (c
>= 0x06e5 && c
<= 0x06e7))
480 if ((c
>= 0x0905 && c
<= 0x0939)
481 || (c
>= 0x0958 && c
<= 0x0962))
485 if ((c
>= 0x0985 && c
<= 0x098c)
486 || (c
>= 0x098f && c
<= 0x0990)
487 || (c
>= 0x0993 && c
<= 0x09a8)
488 || (c
>= 0x09aa && c
<= 0x09b0)
490 || (c
>= 0x09b6 && c
<= 0x09b9)
491 || (c
>= 0x09dc && c
<= 0x09dd)
492 || (c
>= 0x09df && c
<= 0x09e1)
493 || (c
>= 0x09f0 && c
<= 0x09f1))
497 if ((c
>= 0x0a05 && c
<= 0x0a0a)
498 || (c
>= 0x0a0f && c
<= 0x0a10)
499 || (c
>= 0x0a13 && c
<= 0x0a28)
500 || (c
>= 0x0a2a && c
<= 0x0a30)
501 || (c
>= 0x0a32 && c
<= 0x0a33)
502 || (c
>= 0x0a35 && c
<= 0x0a36)
503 || (c
>= 0x0a38 && c
<= 0x0a39)
504 || (c
>= 0x0a59 && c
<= 0x0a5c)
509 if ((c
>= 0x0a85 && c
<= 0x0a8b)
511 || (c
>= 0x0a8f && c
<= 0x0a91)
512 || (c
>= 0x0a93 && c
<= 0x0aa8)
513 || (c
>= 0x0aaa && c
<= 0x0ab0)
514 || (c
>= 0x0ab2 && c
<= 0x0ab3)
515 || (c
>= 0x0ab5 && c
<= 0x0ab9)
520 if ((c
>= 0x0b05 && c
<= 0x0b0c)
521 || (c
>= 0x0b0f && c
<= 0x0b10)
522 || (c
>= 0x0b13 && c
<= 0x0b28)
523 || (c
>= 0x0b2a && c
<= 0x0b30)
524 || (c
>= 0x0b32 && c
<= 0x0b33)
525 || (c
>= 0x0b36 && c
<= 0x0b39)
526 || (c
>= 0x0b5c && c
<= 0x0b5d)
527 || (c
>= 0x0b5f && c
<= 0x0b61))
531 if ((c
>= 0x0b85 && c
<= 0x0b8a)
532 || (c
>= 0x0b8e && c
<= 0x0b90)
533 || (c
>= 0x0b92 && c
<= 0x0b95)
534 || (c
>= 0x0b99 && c
<= 0x0b9a)
536 || (c
>= 0x0b9e && c
<= 0x0b9f)
537 || (c
>= 0x0ba3 && c
<= 0x0ba4)
538 || (c
>= 0x0ba8 && c
<= 0x0baa)
539 || (c
>= 0x0bae && c
<= 0x0bb5)
540 || (c
>= 0x0bb7 && c
<= 0x0bb9))
544 if ((c
>= 0x0c05 && c
<= 0x0c0c)
545 || (c
>= 0x0c0e && c
<= 0x0c10)
546 || (c
>= 0x0c12 && c
<= 0x0c28)
547 || (c
>= 0x0c2a && c
<= 0x0c33)
548 || (c
>= 0x0c35 && c
<= 0x0c39)
549 || (c
>= 0x0c60 && c
<= 0x0c61))
553 if ((c
>= 0x0c85 && c
<= 0x0c8c)
554 || (c
>= 0x0c8e && c
<= 0x0c90)
555 || (c
>= 0x0c92 && c
<= 0x0ca8)
556 || (c
>= 0x0caa && c
<= 0x0cb3)
557 || (c
>= 0x0cb5 && c
<= 0x0cb9)
558 || (c
>= 0x0ce0 && c
<= 0x0ce1))
562 if ((c
>= 0x0d05 && c
<= 0x0d0c)
563 || (c
>= 0x0d0e && c
<= 0x0d10)
564 || (c
>= 0x0d12 && c
<= 0x0d28)
565 || (c
>= 0x0d2a && c
<= 0x0d39)
566 || (c
>= 0x0d60 && c
<= 0x0d61))
570 if ((c
>= 0x0e01 && c
<= 0x0e30)
571 || (c
>= 0x0e32 && c
<= 0x0e33)
572 || (c
>= 0x0e40 && c
<= 0x0e46)
573 || (c
>= 0x0e4f && c
<= 0x0e5b))
577 if ((c
>= 0x0e81 && c
<= 0x0e82)
583 || (c
>= 0x0e94 && c
<= 0x0e97)
584 || (c
>= 0x0e99 && c
<= 0x0e9f)
585 || (c
>= 0x0ea1 && c
<= 0x0ea3)
590 || (c
>= 0x0ead && c
<= 0x0eb0)
594 || (c
>= 0x0ec0 && c
<= 0x0ec4)
599 if ((c
>= 0x10a0 && c
<= 0x10c5)
600 || (c
>= 0x10d0 && c
<= 0x10f6))
604 if ((c
>= 0x3041 && c
<= 0x3094)
605 || (c
>= 0x309b && c
<= 0x309e))
609 if ((c
>= 0x30a1 && c
<= 0x30fe))
613 if ((c
>= 0x3105 && c
<= 0x312c))
617 if ((c
>= 0x1100 && c
<= 0x1159)
618 || (c
>= 0x1161 && c
<= 0x11a2)
619 || (c
>= 0x11a8 && c
<= 0x11f9))
622 /* CJK Unified Ideographs */
623 if ((c
>= 0xf900 && c
<= 0xfa2d)
624 || (c
>= 0xfb1f && c
<= 0xfb36)
625 || (c
>= 0xfb38 && c
<= 0xfb3c)
627 || (c
>= 0xfb40 && c
<= 0xfb41)
628 || (c
>= 0xfb42 && c
<= 0xfb44)
629 || (c
>= 0xfb46 && c
<= 0xfbb1)
630 || (c
>= 0xfbd3 && c
<= 0xfd3f)
631 || (c
>= 0xfd50 && c
<= 0xfd8f)
632 || (c
>= 0xfd92 && c
<= 0xfdc7)
633 || (c
>= 0xfdf0 && c
<= 0xfdfb)
634 || (c
>= 0xfe70 && c
<= 0xfe72)
636 || (c
>= 0xfe76 && c
<= 0xfefc)
637 || (c
>= 0xff21 && c
<= 0xff3a)
638 || (c
>= 0xff41 && c
<= 0xff5a)
639 || (c
>= 0xff66 && c
<= 0xffbe)
640 || (c
>= 0xffc2 && c
<= 0xffc7)
641 || (c
>= 0xffca && c
<= 0xffcf)
642 || (c
>= 0xffd2 && c
<= 0xffd7)
643 || (c
>= 0xffda && c
<= 0xffdc)
644 || (c
>= 0x4e00 && c
<= 0x9fa5))
647 error ("universal-character-name '\\u%04x' not valid in identifier", c
);
652 /* Add the UTF-8 representation of C to the token_buffer. */
655 utf8_extend_token (c
)
665 else if (c
<= 0x000007ff)
666 shift
= 6, mask
= 0xc0;
667 else if (c
<= 0x0000ffff)
668 shift
= 12, mask
= 0xe0;
669 else if (c
<= 0x001fffff)
670 shift
= 18, mask
= 0xf0;
671 else if (c
<= 0x03ffffff)
672 shift
= 24, mask
= 0xf8;
674 shift
= 30, mask
= 0xfc;
676 extend_token (mask
| (c
>> shift
));
680 extend_token ((unsigned char) (0x80 | (c
>> shift
)));
689 tree
*const node_var
;
690 const char unsigned_flag
;
691 const char long_flag
;
692 const char long_long_flag
;
695 struct try_type type_sequence
[] =
697 { &integer_type_node
, 0, 0, 0},
698 { &unsigned_type_node
, 1, 0, 0},
699 { &long_integer_type_node
, 0, 1, 0},
700 { &long_unsigned_type_node
, 1, 1, 0},
701 { &long_long_integer_type_node
, 0, 1, 1},
702 { &long_long_unsigned_type_node
, 1, 1, 1}
714 int conversion_errno
;
715 REAL_VALUE_TYPE value
;
723 struct pf_args
* args
= (struct pf_args
*) data
;
724 const char *typename
;
726 args
->conversion_errno
= 0;
727 args
->type
= double_type_node
;
730 /* The second argument, machine_mode, of REAL_VALUE_ATOF
731 tells the desired precision of the binary result
732 of decimal-to-binary conversion. */
737 error ("both 'f' and 'l' suffixes on floating constant");
739 args
->type
= float_type_node
;
742 else if (args
->lflag
)
744 args
->type
= long_double_type_node
;
745 typename
= "long double";
747 else if (flag_single_precision_constant
)
749 args
->type
= float_type_node
;
754 if (args
->base
== 16)
755 args
->value
= REAL_VALUE_HTOF (args
->str
, TYPE_MODE (args
->type
));
757 args
->value
= REAL_VALUE_ATOF (args
->str
, TYPE_MODE (args
->type
));
759 args
->conversion_errno
= errno
;
760 /* A diagnostic is required here by some ISO C testsuites.
761 This is not pedwarn, because some people don't want
762 an error for this. */
763 if (REAL_VALUE_ISINF (args
->value
) && pedantic
)
764 warning ("floating point number exceeds range of '%s'", typename
);
771 const cpp_token
*tok
;
774 timevar_push (TV_CPP
);
776 tok
= cpp_get_token (parse_in
);
777 while (tok
->type
== CPP_PADDING
);
778 timevar_pop (TV_CPP
);
780 /* The C++ front end does horrible things with the current line
781 number. To ensure an accurate line number, we must reset it
782 every time we return a token. */
788 case CPP_OPEN_BRACE
: indent_level
++; break;
789 case CPP_CLOSE_BRACE
: indent_level
--; break;
791 /* Issue this error here, where we can get at tok->val.c. */
793 if (ISGRAPH (tok
->val
.c
))
794 error ("stray '%c' in program", tok
->val
.c
);
796 error ("stray '\\%o' in program", tok
->val
.c
);
800 *value
= HT_IDENT_TO_GCC_IDENT (HT_NODE (tok
->val
.node
));
804 *value
= lex_number ((const char *)tok
->val
.str
.text
, tok
->val
.str
.len
);
809 *value
= lex_charconst (tok
);
814 *value
= lex_string (tok
->val
.str
.text
, tok
->val
.str
.len
,
815 tok
->type
== CPP_WSTRING
);
818 /* These tokens should not be visible outside cpplib. */
819 case CPP_HEADER_NAME
:
830 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
833 lex_number (str
, len
)
839 int largest_digit
= 0;
845 enum anon1
{ NOT_FLOAT
= 0, AFTER_POINT
, AFTER_EXPON
} floatflag
= NOT_FLOAT
;
847 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
848 The code below which fills the parts array assumes that a host
849 int is at least twice as wide as a host char, and that
850 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
851 Two HOST_WIDE_INTs is the largest int literal we can store.
852 In order to detect overflow below, the number of parts (TOTAL_PARTS)
853 must be exactly the number of parts needed to hold the bits
854 of two HOST_WIDE_INTs. */
855 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
856 unsigned int parts
[TOTAL_PARTS
];
858 /* Optimize for most frequent case. */
862 return integer_zero_node
;
863 else if (*str
== '1')
864 return integer_one_node
;
866 return build_int_2 (*str
- '0', 0);
869 for (count
= 0; count
< TOTAL_PARTS
; count
++)
872 /* len is known to be >1 at this point. */
875 if (len
> 2 && str
[0] == '0' && (str
[1] == 'x' || str
[1] == 'X'))
880 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
881 else if (str
[0] == '0' && ISDIGIT (str
[1]))
893 if (floatflag
== AFTER_POINT
)
894 ERROR ("too many decimal points in floating constant");
895 else if (floatflag
== AFTER_EXPON
)
896 ERROR ("decimal point in exponent - impossible!");
898 floatflag
= AFTER_POINT
;
904 /* Possible future extension: silently ignore _ in numbers,
905 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
906 but somewhat easier to read. Ada has this? */
907 ERROR ("underscore in number");
911 /* It is not a decimal point.
912 It should be a digit (perhaps a hex digit). */
915 || (base
== 16 && ISXDIGIT (c
)))
919 else if (base
<= 10 && (c
== 'e' || c
== 'E'))
922 floatflag
= AFTER_EXPON
;
925 else if (base
== 16 && (c
== 'p' || c
== 'P'))
927 floatflag
= AFTER_EXPON
;
928 break; /* start of exponent */
933 break; /* start of suffix */
936 if (n
>= largest_digit
)
940 for (count
= 0; count
< TOTAL_PARTS
; count
++)
942 parts
[count
] *= base
;
946 += (parts
[count
-1] >> HOST_BITS_PER_CHAR
);
948 &= (1 << HOST_BITS_PER_CHAR
) - 1;
954 /* If the highest-order part overflows (gets larger than
955 a host char will hold) then the whole number has
956 overflowed. Record this and truncate the highest-order
958 if (parts
[TOTAL_PARTS
- 1] >> HOST_BITS_PER_CHAR
)
961 parts
[TOTAL_PARTS
- 1] &= (1 << HOST_BITS_PER_CHAR
) - 1;
965 while (p
< str
+ len
);
967 /* This can happen on input like `int i = 0x;' */
969 ERROR ("numeric constant with no digits");
971 if (largest_digit
>= base
)
972 ERROR ("numeric constant contains digits beyond the radix");
974 if (floatflag
!= NOT_FLOAT
)
977 int imag
, fflag
, lflag
, conversion_errno
;
978 REAL_VALUE_TYPE real
;
982 if (base
== 16 && pedantic
&& !flag_isoc99
)
983 pedwarn ("floating constant may not be in radix 16");
985 if (base
== 16 && floatflag
!= AFTER_EXPON
)
986 ERROR ("hexadecimal floating constant has no exponent");
988 /* Read explicit exponent if any, and put it in tokenbuf. */
989 if ((base
== 10 && ((c
== 'e') || (c
== 'E')))
990 || (base
== 16 && (c
== 'p' || c
== 'P')))
994 if (p
< str
+ len
&& (c
== '+' || c
== '-'))
996 /* Exponent is decimal, even if string is a hex float. */
998 ERROR ("floating constant exponent has no digits");
999 while (p
< str
+ len
&& ISDIGIT (c
))
1005 /* Copy the float constant now; we don't want any suffixes in the
1006 string passed to parse_float. */
1007 copy
= alloca (p
- str
+ 1);
1008 memcpy (copy
, str
, p
- str
);
1009 copy
[p
- str
] = '\0';
1011 /* Now parse suffixes. */
1012 fflag
= lflag
= imag
= 0;
1013 while (p
< str
+ len
)
1018 ERROR ("more than one 'f' suffix on floating constant");
1019 else if (warn_traditional
&& !in_system_header
1020 && ! cpp_sys_macro_p (parse_in
))
1021 warning ("traditional C rejects the 'f' suffix");
1028 ERROR ("more than one 'l' suffix on floating constant");
1029 else if (warn_traditional
&& !in_system_header
1030 && ! cpp_sys_macro_p (parse_in
))
1031 warning ("traditional C rejects the 'l' suffix");
1039 ERROR ("more than one 'i' or 'j' suffix on floating constant");
1041 pedwarn ("ISO C forbids imaginary numeric constants");
1046 ERROR ("invalid suffix on floating constant");
1049 /* Setup input for parse_float() */
1055 /* Convert string to a double, checking for overflow. */
1056 if (do_float_handler (parse_float
, (PTR
) &args
))
1058 /* Receive output from parse_float() */
1062 /* We got an exception from parse_float() */
1063 ERROR ("floating constant out of range");
1065 /* Receive output from parse_float() */
1066 conversion_errno
= args
.conversion_errno
;
1070 /* ERANGE is also reported for underflow,
1071 so test the value to distinguish overflow from that. */
1072 if (conversion_errno
== ERANGE
&& !flag_traditional
&& pedantic
1073 && (REAL_VALUES_LESS (dconst1
, real
)
1074 || REAL_VALUES_LESS (real
, dconstm1
)))
1075 warning ("floating point number exceeds range of 'double'");
1078 /* Create a node with determined type and value. */
1080 value
= build_complex (NULL_TREE
, convert (type
, integer_zero_node
),
1081 build_real (type
, real
));
1083 value
= build_real (type
, real
);
1087 tree trad_type
, ansi_type
, type
;
1088 HOST_WIDE_INT high
, low
;
1089 int spec_unsigned
= 0;
1091 int spec_long_long
= 0;
1096 trad_type
= ansi_type
= type
= NULL_TREE
;
1097 while (p
< str
+ len
)
1104 error ("two 'u' suffixes on integer constant");
1105 else if (warn_traditional
&& !in_system_header
1106 && ! cpp_sys_macro_p (parse_in
))
1107 warning ("traditional C rejects the 'u' suffix");
1118 error ("three 'l' suffixes on integer constant");
1120 error ("'lul' is not a valid integer suffix");
1121 else if (c
!= spec_long
)
1122 error ("'Ll' and 'lL' are not valid integer suffixes");
1123 else if (pedantic
&& ! flag_isoc99
1124 && ! in_system_header
&& warn_long_long
)
1125 pedwarn ("ISO C89 forbids long long integer constants");
1131 case 'i': case 'I': case 'j': case 'J':
1133 error ("more than one 'i' or 'j' suffix on integer constant");
1135 pedwarn ("ISO C forbids imaginary numeric constants");
1140 ERROR ("invalid suffix on integer constant");
1144 /* If the literal overflowed, pedwarn about it now. */
1148 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT
* 2);
1151 /* This is simplified by the fact that our constant
1152 is always positive. */
1156 for (i
= 0; i
< HOST_BITS_PER_WIDE_INT
/ HOST_BITS_PER_CHAR
; i
++)
1158 high
|= ((HOST_WIDE_INT
) parts
[i
+ (HOST_BITS_PER_WIDE_INT
1159 / HOST_BITS_PER_CHAR
)]
1160 << (i
* HOST_BITS_PER_CHAR
));
1161 low
|= (HOST_WIDE_INT
) parts
[i
] << (i
* HOST_BITS_PER_CHAR
);
1164 value
= build_int_2 (low
, high
);
1165 TREE_TYPE (value
) = long_long_unsigned_type_node
;
1167 /* If warn_traditional, calculate both the ISO type and the
1168 traditional type, then see if they disagree.
1169 Otherwise, calculate only the type for the dialect in use. */
1170 if (warn_traditional
|| flag_traditional
)
1172 /* Calculate the traditional type. */
1173 /* Traditionally, any constant is signed; but if unsigned is
1174 specified explicitly, obey that. Use the smallest size
1175 with the right number of bits, except for one special
1176 case with decimal constants. */
1177 if (! spec_long
&& base
!= 10
1178 && int_fits_type_p (value
, unsigned_type_node
))
1179 trad_type
= spec_unsigned
? unsigned_type_node
: integer_type_node
;
1180 /* A decimal constant must be long if it does not fit in
1181 type int. I think this is independent of whether the
1182 constant is signed. */
1183 else if (! spec_long
&& base
== 10
1184 && int_fits_type_p (value
, integer_type_node
))
1185 trad_type
= spec_unsigned
? unsigned_type_node
: integer_type_node
;
1186 else if (! spec_long_long
)
1187 trad_type
= (spec_unsigned
1188 ? long_unsigned_type_node
1189 : long_integer_type_node
);
1190 else if (int_fits_type_p (value
,
1192 ? long_long_unsigned_type_node
1193 : long_long_integer_type_node
))
1194 trad_type
= (spec_unsigned
1195 ? long_long_unsigned_type_node
1196 : long_long_integer_type_node
);
1198 trad_type
= (spec_unsigned
1199 ? widest_unsigned_literal_type_node
1200 : widest_integer_literal_type_node
);
1202 if (warn_traditional
|| ! flag_traditional
)
1204 /* Calculate the ISO type. */
1205 if (! spec_long
&& ! spec_unsigned
1206 && int_fits_type_p (value
, integer_type_node
))
1207 ansi_type
= integer_type_node
;
1208 else if (! spec_long
&& (base
!= 10 || spec_unsigned
)
1209 && int_fits_type_p (value
, unsigned_type_node
))
1210 ansi_type
= unsigned_type_node
;
1211 else if (! spec_unsigned
&& !spec_long_long
1212 && int_fits_type_p (value
, long_integer_type_node
))
1213 ansi_type
= long_integer_type_node
;
1214 else if (! spec_long_long
1215 && int_fits_type_p (value
, long_unsigned_type_node
))
1216 ansi_type
= long_unsigned_type_node
;
1217 else if (! spec_unsigned
1218 && int_fits_type_p (value
, long_long_integer_type_node
))
1219 ansi_type
= long_long_integer_type_node
;
1220 else if (int_fits_type_p (value
, long_long_unsigned_type_node
))
1221 ansi_type
= long_long_unsigned_type_node
;
1222 else if (! spec_unsigned
1223 && int_fits_type_p (value
, widest_integer_literal_type_node
))
1224 ansi_type
= widest_integer_literal_type_node
;
1226 ansi_type
= widest_unsigned_literal_type_node
;
1229 type
= flag_traditional
? trad_type
: ansi_type
;
1231 /* We assume that constants specified in a non-decimal
1232 base are bit patterns, and that the programmer really
1233 meant what they wrote. */
1234 if (warn_traditional
&& !in_system_header
1235 && base
== 10 && trad_type
!= ansi_type
)
1237 if (TYPE_PRECISION (trad_type
) != TYPE_PRECISION (ansi_type
))
1238 warning ("width of integer constant changes with -traditional");
1239 else if (TREE_UNSIGNED (trad_type
) != TREE_UNSIGNED (ansi_type
))
1240 warning ("integer constant is unsigned in ISO C, signed with -traditional");
1242 warning ("width of integer constant may change on other systems with -traditional");
1245 if (pedantic
&& !flag_traditional
&& (flag_isoc99
|| !spec_long_long
)
1248 ? TYPE_PRECISION (long_long_integer_type_node
)
1249 : TYPE_PRECISION (long_integer_type_node
)) < TYPE_PRECISION (type
)))
1252 pedwarn ("integer constant larger than the maximum value of %s",
1254 ? (TREE_UNSIGNED (type
)
1255 ? _("an unsigned long long int")
1256 : _("a long long int"))
1257 : _("an unsigned long int")));
1260 if (base
== 10 && ! spec_unsigned
&& TREE_UNSIGNED (type
))
1261 warning ("decimal constant is so large that it is unsigned");
1265 if (TYPE_PRECISION (type
)
1266 <= TYPE_PRECISION (integer_type_node
))
1267 value
= build_complex (NULL_TREE
, integer_zero_node
,
1268 convert (integer_type_node
, value
));
1270 ERROR ("complex integer constant is too wide for 'complex int'");
1272 else if (flag_traditional
&& !int_fits_type_p (value
, type
))
1273 /* The traditional constant 0x80000000 is signed
1274 but doesn't fit in the range of int.
1275 This will change it to -0x80000000, which does fit. */
1277 TREE_TYPE (value
) = unsigned_type (type
);
1278 value
= convert (type
, value
);
1279 TREE_OVERFLOW (value
) = TREE_CONSTANT_OVERFLOW (value
) = 0;
1282 TREE_TYPE (value
) = type
;
1284 /* If it's still an integer (not a complex), and it doesn't
1285 fit in the type we choose for it, then pedwarn. */
1288 && TREE_CODE (TREE_TYPE (value
)) == INTEGER_TYPE
1289 && ! int_fits_type_p (value
, TREE_TYPE (value
)))
1290 pedwarn ("integer constant is larger than the maximum value for its type");
1294 error ("missing white space after number '%.*s'", (int) (p
- str
), str
);
1299 return integer_zero_node
;
1303 lex_string (str
, len
, wide
)
1304 const unsigned char *str
;
1309 char *buf
= alloca ((len
+ 1) * (wide
? WCHAR_BYTES
: 1));
1311 const unsigned char *p
= str
, *limit
= str
+ len
;
1313 unsigned width
= wide
? WCHAR_TYPE_SIZE
1314 : TYPE_PRECISION (char_type_node
);
1316 #ifdef MULTIBYTE_CHARS
1317 /* Reset multibyte conversion state. */
1318 (void) local_mbtowc (NULL
, NULL
, 0);
1323 #ifdef MULTIBYTE_CHARS
1327 char_len
= local_mbtowc (&wc
, (const char *) p
, limit
- p
);
1330 warning ("ignoring invalid multibyte character");
1343 if (c
== '\\' && !ignore_escape_flag
)
1347 if (width
< HOST_BITS_PER_INT
)
1348 mask
= ((unsigned int) 1 << width
) - 1;
1351 c
= cpp_parse_escape (parse_in
, &p
, limit
,
1352 mask
, flag_traditional
);
1355 /* Add this single character into the buffer either as a wchar_t
1356 or as a single byte. */
1359 unsigned charwidth
= TYPE_PRECISION (char_type_node
);
1360 unsigned bytemask
= (1 << charwidth
) - 1;
1363 for (byte
= 0; byte
< WCHAR_BYTES
; ++byte
)
1366 if (byte
>= (int) sizeof (c
))
1369 n
= (c
>> (byte
* charwidth
)) & bytemask
;
1370 if (BYTES_BIG_ENDIAN
)
1371 q
[WCHAR_BYTES
- byte
- 1] = n
;
1383 /* Terminate the string value, either with a single byte zero
1384 or with a wide zero. */
1388 memset (q
, 0, WCHAR_BYTES
);
1396 value
= build_string (q
- buf
, buf
);
1399 TREE_TYPE (value
) = wchar_array_type_node
;
1401 TREE_TYPE (value
) = char_array_type_node
;
1405 /* Converts a (possibly wide) character constant token into a tree. */
1407 lex_charconst (token
)
1408 const cpp_token
*token
;
1410 HOST_WIDE_INT result
;
1412 unsigned int chars_seen
;
1414 result
= cpp_interpret_charconst (parse_in
, token
, warn_multichar
,
1415 flag_traditional
, &chars_seen
);
1416 if (token
->type
== CPP_WCHAR
)
1418 value
= build_int_2 (result
, 0);
1419 type
= wchar_type_node
;
1424 value
= build_int_2 (result
, -1);
1426 value
= build_int_2 (result
, 0);
1428 /* In C, a character constant has type 'int'.
1429 In C++ 'char', but multi-char charconsts have type 'int'. */
1430 if (c_language
== clk_cplusplus
&& chars_seen
<= 1)
1431 type
= char_type_node
;
1433 type
= integer_type_node
;
1436 /* cpp_interpret_charconst issues a warning if the constant
1437 overflows, but if the number fits in HOST_WIDE_INT anyway, it
1438 will return it un-truncated, which may cause problems down the
1439 line. So set the type to widest_integer_literal_type, call
1440 convert to truncate it to the proper type, then clear
1441 TREE_OVERFLOW so we don't get a second warning.
1443 FIXME: cpplib's assessment of overflow may not be accurate on a
1444 platform where the final type can change at (compiler's) runtime. */
1446 TREE_TYPE (value
) = widest_integer_literal_type_node
;
1447 value
= convert (type
, value
);
1448 TREE_OVERFLOW (value
) = 0;