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"
44 #ifdef MULTIBYTE_CHARS
47 #endif /* MULTIBYTE_CHARS */
48 #ifndef GET_ENVIRONMENT
49 #define GET_ENVIRONMENT(ENV_VALUE,ENV_NAME) ((ENV_VALUE) = getenv (ENV_NAME))
52 /* The current line map. */
53 static const struct line_map *map;
55 /* The line used to refresh the lineno global variable after each token. */
56 static unsigned int src_lineno;
58 /* We may keep statistics about how long which files took to compile. */
59 static int header_time, body_time;
60 static splay_tree file_info_tree;
62 /* File used for outputting assembler code. */
63 extern FILE *asm_out_file;
65 #undef WCHAR_TYPE_SIZE
66 #define WCHAR_TYPE_SIZE TYPE_PRECISION (wchar_type_node)
68 /* Number of bytes in a wide character. */
69 #define WCHAR_BYTES (WCHAR_TYPE_SIZE / BITS_PER_UNIT)
71 int indent_level; /* Number of { minus number of }. */
72 int pending_lang_change; /* If we need to switch languages - C++ only */
73 int c_header_level; /* depth in C headers - C++ only */
75 /* Nonzero tells yylex to ignore \ in string constants. */
76 static int ignore_escape_flag;
78 static tree lex_number PARAMS ((const char *, unsigned int));
79 static tree lex_string PARAMS ((const unsigned char *, unsigned int,
81 static tree lex_charconst PARAMS ((const cpp_token *));
82 static void update_header_times PARAMS ((const char *));
83 static int dump_one_header PARAMS ((splay_tree_node, void *));
84 static void cb_register_builtins PARAMS ((cpp_reader *));
85 static void cb_line_change PARAMS ((cpp_reader *, const cpp_token *, int));
86 static void cb_ident PARAMS ((cpp_reader *, unsigned int,
88 static void cb_file_change PARAMS ((cpp_reader *, const struct line_map *));
89 static void cb_def_pragma PARAMS ((cpp_reader *, unsigned int));
90 static void cb_define PARAMS ((cpp_reader *, unsigned int,
92 static void cb_undef PARAMS ((cpp_reader *, unsigned int,
99 struct cpp_callbacks *cb;
100 struct c_fileinfo *toplevel;
102 /* Set up filename timing. Must happen before cpp_read_main_file. */
103 file_info_tree = splay_tree_new ((splay_tree_compare_fn)strcmp,
105 (splay_tree_delete_value_fn)free);
106 toplevel = get_fileinfo ("<top level>");
107 if (flag_detailed_statistics)
110 body_time = get_run_time ();
111 toplevel->time = body_time;
114 #ifdef MULTIBYTE_CHARS
115 /* Change to the native locale for multibyte conversions. */
116 setlocale (LC_CTYPE, "");
117 GET_ENVIRONMENT (literal_codeset, "LANG");
120 cb = cpp_get_callbacks (parse_in);
122 cb->line_change = cb_line_change;
123 cb->ident = cb_ident;
124 cb->file_change = cb_file_change;
125 cb->def_pragma = cb_def_pragma;
126 cb->register_builtins = cb_register_builtins;
128 /* Set the debug callbacks if we can use them. */
129 if (debug_info_level == DINFO_LEVEL_VERBOSE
130 && (write_symbols == DWARF_DEBUG || write_symbols == DWARF2_DEBUG
131 || write_symbols == VMS_AND_DWARF2_DEBUG))
133 cb->define = cb_define;
134 cb->undef = cb_undef;
140 if (filename == NULL || !strcmp (filename, "-"))
143 return cpp_read_main_file (parse_in, filename, ident_hash);
146 /* A thin wrapper around the real parser that initializes the
147 integrated preprocessor after debug output has been initialized.
148 Also, make sure the start_source_file debug hook gets called for
149 the primary source file. */
152 c_common_parse_file (set_yydebug)
153 int set_yydebug ATTRIBUTE_UNUSED;
156 yydebug = set_yydebug;
158 warning ("YYDEBUG not defined");
161 (*debug_hooks->start_source_file) (lineno, input_filename);
162 cpp_finish_options (parse_in);
165 free_parser_stacks ();
173 struct c_fileinfo *fi;
175 n = splay_tree_lookup (file_info_tree, (splay_tree_key) name);
177 return (struct c_fileinfo *) n->value;
179 fi = (struct c_fileinfo *) xmalloc (sizeof (struct c_fileinfo));
181 fi->interface_only = 0;
182 fi->interface_unknown = 1;
183 splay_tree_insert (file_info_tree, (splay_tree_key) name,
184 (splay_tree_value) fi);
189 update_header_times (name)
192 /* Changing files again. This means currently collected time
193 is charged against header time, and body time starts back at 0. */
194 if (flag_detailed_statistics)
196 int this_time = get_run_time ();
197 struct c_fileinfo *file = get_fileinfo (name);
198 header_time += this_time - body_time;
199 file->time += this_time - body_time;
200 body_time = this_time;
205 dump_one_header (n, dummy)
207 void *dummy ATTRIBUTE_UNUSED;
209 print_time ((const char *) n->key,
210 ((struct c_fileinfo *) n->value)->time);
215 dump_time_statistics ()
217 struct c_fileinfo *file = get_fileinfo (input_filename);
218 int this_time = get_run_time ();
219 file->time += this_time - body_time;
221 fprintf (stderr, "\n******\n");
222 print_time ("header files (total)", header_time);
223 print_time ("main file (total)", this_time - body_time);
224 fprintf (stderr, "ratio = %g : 1\n",
225 (double)header_time / (double)(this_time - body_time));
226 fprintf (stderr, "\n******\n");
228 splay_tree_foreach (file_info_tree, dump_one_header, 0);
231 /* Register preprocessor built-ins. */
233 cb_register_builtins (pfile)
236 if (c_language == clk_cplusplus)
238 if (SUPPORTS_ONE_ONLY)
239 cpp_define (pfile, "__GXX_WEAK__");
241 cpp_define (pfile, "__GXX_WEAK__=0");
244 (*targetm.register_cpp_builtins) (pfile);
247 /* Not yet handled: #pragma, #define, #undef.
248 No need to deal with linemarkers under normal conditions. */
251 cb_ident (pfile, line, str)
252 cpp_reader *pfile ATTRIBUTE_UNUSED;
253 unsigned int line ATTRIBUTE_UNUSED;
254 const cpp_string *str ATTRIBUTE_UNUSED;
256 #ifdef ASM_OUTPUT_IDENT
259 /* Convert escapes in the string. */
260 tree value = lex_string (str->text, str->len, 0);
261 ASM_OUTPUT_IDENT (asm_out_file, TREE_STRING_POINTER (value));
266 /* Called at the start of every non-empty line. TOKEN is the first
267 lexed token on the line. Used for diagnostic line numbers. */
269 cb_line_change (pfile, token, parsing_args)
270 cpp_reader *pfile ATTRIBUTE_UNUSED;
271 const cpp_token *token;
272 int parsing_args ATTRIBUTE_UNUSED;
274 src_lineno = SOURCE_LINE (map, token->line);
278 cb_file_change (pfile, new_map)
279 cpp_reader *pfile ATTRIBUTE_UNUSED;
280 const struct line_map *new_map;
282 unsigned int to_line = SOURCE_LINE (new_map, new_map->to_line);
284 if (new_map->reason == LC_ENTER)
286 /* Don't stack the main buffer on the input stack;
287 we already did in compile_file. */
289 main_input_filename = new_map->to_file;
292 int included_at = SOURCE_LINE (new_map - 1, new_map->from_line - 1);
294 lineno = included_at;
295 push_srcloc (new_map->to_file, 1);
296 input_file_stack->indent_level = indent_level;
297 (*debug_hooks->start_source_file) (included_at, new_map->to_file);
298 #ifndef NO_IMPLICIT_EXTERN_C
301 else if (new_map->sysp == 2)
304 ++pending_lang_change;
309 else if (new_map->reason == LC_LEAVE)
311 #ifndef NO_IMPLICIT_EXTERN_C
312 if (c_header_level && --c_header_level == 0)
314 if (new_map->sysp == 2)
315 warning ("badly nested C headers from preprocessor");
316 --pending_lang_change;
320 if (indent_level != input_file_stack->indent_level)
322 warning_with_file_and_line
323 (input_filename, lineno,
324 "this file contains more '%c's than '%c's",
325 indent_level > input_file_stack->indent_level ? '{' : '}',
326 indent_level > input_file_stack->indent_level ? '}' : '{');
331 (*debug_hooks->end_source_file) (to_line);
334 update_header_times (new_map->to_file);
335 in_system_header = new_map->sysp != 0;
336 input_filename = new_map->to_file;
341 extract_interface_info ();
345 cb_def_pragma (pfile, line)
349 /* Issue a warning message if we have been asked to do so. Ignore
350 unknown pragmas in system headers unless an explicit
351 -Wunknown-pragmas has been given. */
352 if (warn_unknown_pragmas > in_system_header)
354 const unsigned char *space, *name = 0;
357 s = cpp_get_token (pfile);
358 space = cpp_token_as_text (pfile, s);
359 s = cpp_get_token (pfile);
360 if (s->type == CPP_NAME)
361 name = cpp_token_as_text (pfile, s);
363 lineno = SOURCE_LINE (map, line);
365 warning ("ignoring #pragma %s %s", space, name);
367 warning ("ignoring #pragma %s", space);
371 /* #define callback for DWARF and DWARF2 debug info. */
373 cb_define (pfile, line, node)
378 (*debug_hooks->define) (SOURCE_LINE (map, line),
379 (const char *) cpp_macro_definition (pfile, node));
382 /* #undef callback for DWARF and DWARF2 debug info. */
384 cb_undef (pfile, line, node)
385 cpp_reader *pfile ATTRIBUTE_UNUSED;
389 (*debug_hooks->undef) (SOURCE_LINE (map, line),
390 (const char *) NODE_NAME (node));
394 /* Returns nonzero if C is a universal-character-name. Give an error if it
395 is not one which may appear in an identifier, as per [extendid].
397 Note that extended character support in identifiers has not yet been
398 implemented. It is my personal opinion that this is not a desirable
399 feature. Portable code cannot count on support for more than the basic
400 identifier character set. */
413 /* None of the valid chars are outside the Basic Multilingual Plane (the
417 error ("universal-character-name '\\U%08x' not valid in identifier", c);
422 if ((c >= 0x00c0 && c <= 0x00d6)
423 || (c >= 0x00d8 && c <= 0x00f6)
424 || (c >= 0x00f8 && c <= 0x01f5)
425 || (c >= 0x01fa && c <= 0x0217)
426 || (c >= 0x0250 && c <= 0x02a8)
427 || (c >= 0x1e00 && c <= 0x1e9a)
428 || (c >= 0x1ea0 && c <= 0x1ef9))
433 || (c >= 0x0388 && c <= 0x038a)
435 || (c >= 0x038e && c <= 0x03a1)
436 || (c >= 0x03a3 && c <= 0x03ce)
437 || (c >= 0x03d0 && c <= 0x03d6)
442 || (c >= 0x03e2 && c <= 0x03f3)
443 || (c >= 0x1f00 && c <= 0x1f15)
444 || (c >= 0x1f18 && c <= 0x1f1d)
445 || (c >= 0x1f20 && c <= 0x1f45)
446 || (c >= 0x1f48 && c <= 0x1f4d)
447 || (c >= 0x1f50 && c <= 0x1f57)
451 || (c >= 0x1f5f && c <= 0x1f7d)
452 || (c >= 0x1f80 && c <= 0x1fb4)
453 || (c >= 0x1fb6 && c <= 0x1fbc)
454 || (c >= 0x1fc2 && c <= 0x1fc4)
455 || (c >= 0x1fc6 && c <= 0x1fcc)
456 || (c >= 0x1fd0 && c <= 0x1fd3)
457 || (c >= 0x1fd6 && c <= 0x1fdb)
458 || (c >= 0x1fe0 && c <= 0x1fec)
459 || (c >= 0x1ff2 && c <= 0x1ff4)
460 || (c >= 0x1ff6 && c <= 0x1ffc))
464 if ((c >= 0x0401 && c <= 0x040d)
465 || (c >= 0x040f && c <= 0x044f)
466 || (c >= 0x0451 && c <= 0x045c)
467 || (c >= 0x045e && c <= 0x0481)
468 || (c >= 0x0490 && c <= 0x04c4)
469 || (c >= 0x04c7 && c <= 0x04c8)
470 || (c >= 0x04cb && c <= 0x04cc)
471 || (c >= 0x04d0 && c <= 0x04eb)
472 || (c >= 0x04ee && c <= 0x04f5)
473 || (c >= 0x04f8 && c <= 0x04f9))
477 if ((c >= 0x0531 && c <= 0x0556)
478 || (c >= 0x0561 && c <= 0x0587))
482 if ((c >= 0x05d0 && c <= 0x05ea)
483 || (c >= 0x05f0 && c <= 0x05f4))
487 if ((c >= 0x0621 && c <= 0x063a)
488 || (c >= 0x0640 && c <= 0x0652)
489 || (c >= 0x0670 && c <= 0x06b7)
490 || (c >= 0x06ba && c <= 0x06be)
491 || (c >= 0x06c0 && c <= 0x06ce)
492 || (c >= 0x06e5 && c <= 0x06e7))
496 if ((c >= 0x0905 && c <= 0x0939)
497 || (c >= 0x0958 && c <= 0x0962))
501 if ((c >= 0x0985 && c <= 0x098c)
502 || (c >= 0x098f && c <= 0x0990)
503 || (c >= 0x0993 && c <= 0x09a8)
504 || (c >= 0x09aa && c <= 0x09b0)
506 || (c >= 0x09b6 && c <= 0x09b9)
507 || (c >= 0x09dc && c <= 0x09dd)
508 || (c >= 0x09df && c <= 0x09e1)
509 || (c >= 0x09f0 && c <= 0x09f1))
513 if ((c >= 0x0a05 && c <= 0x0a0a)
514 || (c >= 0x0a0f && c <= 0x0a10)
515 || (c >= 0x0a13 && c <= 0x0a28)
516 || (c >= 0x0a2a && c <= 0x0a30)
517 || (c >= 0x0a32 && c <= 0x0a33)
518 || (c >= 0x0a35 && c <= 0x0a36)
519 || (c >= 0x0a38 && c <= 0x0a39)
520 || (c >= 0x0a59 && c <= 0x0a5c)
525 if ((c >= 0x0a85 && c <= 0x0a8b)
527 || (c >= 0x0a8f && c <= 0x0a91)
528 || (c >= 0x0a93 && c <= 0x0aa8)
529 || (c >= 0x0aaa && c <= 0x0ab0)
530 || (c >= 0x0ab2 && c <= 0x0ab3)
531 || (c >= 0x0ab5 && c <= 0x0ab9)
536 if ((c >= 0x0b05 && c <= 0x0b0c)
537 || (c >= 0x0b0f && c <= 0x0b10)
538 || (c >= 0x0b13 && c <= 0x0b28)
539 || (c >= 0x0b2a && c <= 0x0b30)
540 || (c >= 0x0b32 && c <= 0x0b33)
541 || (c >= 0x0b36 && c <= 0x0b39)
542 || (c >= 0x0b5c && c <= 0x0b5d)
543 || (c >= 0x0b5f && c <= 0x0b61))
547 if ((c >= 0x0b85 && c <= 0x0b8a)
548 || (c >= 0x0b8e && c <= 0x0b90)
549 || (c >= 0x0b92 && c <= 0x0b95)
550 || (c >= 0x0b99 && c <= 0x0b9a)
552 || (c >= 0x0b9e && c <= 0x0b9f)
553 || (c >= 0x0ba3 && c <= 0x0ba4)
554 || (c >= 0x0ba8 && c <= 0x0baa)
555 || (c >= 0x0bae && c <= 0x0bb5)
556 || (c >= 0x0bb7 && c <= 0x0bb9))
560 if ((c >= 0x0c05 && c <= 0x0c0c)
561 || (c >= 0x0c0e && c <= 0x0c10)
562 || (c >= 0x0c12 && c <= 0x0c28)
563 || (c >= 0x0c2a && c <= 0x0c33)
564 || (c >= 0x0c35 && c <= 0x0c39)
565 || (c >= 0x0c60 && c <= 0x0c61))
569 if ((c >= 0x0c85 && c <= 0x0c8c)
570 || (c >= 0x0c8e && c <= 0x0c90)
571 || (c >= 0x0c92 && c <= 0x0ca8)
572 || (c >= 0x0caa && c <= 0x0cb3)
573 || (c >= 0x0cb5 && c <= 0x0cb9)
574 || (c >= 0x0ce0 && c <= 0x0ce1))
578 if ((c >= 0x0d05 && c <= 0x0d0c)
579 || (c >= 0x0d0e && c <= 0x0d10)
580 || (c >= 0x0d12 && c <= 0x0d28)
581 || (c >= 0x0d2a && c <= 0x0d39)
582 || (c >= 0x0d60 && c <= 0x0d61))
586 if ((c >= 0x0e01 && c <= 0x0e30)
587 || (c >= 0x0e32 && c <= 0x0e33)
588 || (c >= 0x0e40 && c <= 0x0e46)
589 || (c >= 0x0e4f && c <= 0x0e5b))
593 if ((c >= 0x0e81 && c <= 0x0e82)
599 || (c >= 0x0e94 && c <= 0x0e97)
600 || (c >= 0x0e99 && c <= 0x0e9f)
601 || (c >= 0x0ea1 && c <= 0x0ea3)
606 || (c >= 0x0ead && c <= 0x0eb0)
610 || (c >= 0x0ec0 && c <= 0x0ec4)
615 if ((c >= 0x10a0 && c <= 0x10c5)
616 || (c >= 0x10d0 && c <= 0x10f6))
620 if ((c >= 0x3041 && c <= 0x3094)
621 || (c >= 0x309b && c <= 0x309e))
625 if ((c >= 0x30a1 && c <= 0x30fe))
629 if ((c >= 0x3105 && c <= 0x312c))
633 if ((c >= 0x1100 && c <= 0x1159)
634 || (c >= 0x1161 && c <= 0x11a2)
635 || (c >= 0x11a8 && c <= 0x11f9))
638 /* CJK Unified Ideographs */
639 if ((c >= 0xf900 && c <= 0xfa2d)
640 || (c >= 0xfb1f && c <= 0xfb36)
641 || (c >= 0xfb38 && c <= 0xfb3c)
643 || (c >= 0xfb40 && c <= 0xfb41)
644 || (c >= 0xfb42 && c <= 0xfb44)
645 || (c >= 0xfb46 && c <= 0xfbb1)
646 || (c >= 0xfbd3 && c <= 0xfd3f)
647 || (c >= 0xfd50 && c <= 0xfd8f)
648 || (c >= 0xfd92 && c <= 0xfdc7)
649 || (c >= 0xfdf0 && c <= 0xfdfb)
650 || (c >= 0xfe70 && c <= 0xfe72)
652 || (c >= 0xfe76 && c <= 0xfefc)
653 || (c >= 0xff21 && c <= 0xff3a)
654 || (c >= 0xff41 && c <= 0xff5a)
655 || (c >= 0xff66 && c <= 0xffbe)
656 || (c >= 0xffc2 && c <= 0xffc7)
657 || (c >= 0xffca && c <= 0xffcf)
658 || (c >= 0xffd2 && c <= 0xffd7)
659 || (c >= 0xffda && c <= 0xffdc)
660 || (c >= 0x4e00 && c <= 0x9fa5))
663 error ("universal-character-name '\\u%04x' not valid in identifier", c);
668 /* Add the UTF-8 representation of C to the token_buffer. */
671 utf8_extend_token (c)
681 else if (c <= 0x000007ff)
682 shift = 6, mask = 0xc0;
683 else if (c <= 0x0000ffff)
684 shift = 12, mask = 0xe0;
685 else if (c <= 0x001fffff)
686 shift = 18, mask = 0xf0;
687 else if (c <= 0x03ffffff)
688 shift = 24, mask = 0xf8;
690 shift = 30, mask = 0xfc;
692 extend_token (mask | (c >> shift));
696 extend_token ((unsigned char) (0x80 | (c >> shift)));
705 tree *const node_var;
706 const char unsigned_flag;
707 const char long_flag;
708 const char long_long_flag;
711 struct try_type type_sequence[] =
713 { &integer_type_node, 0, 0, 0},
714 { &unsigned_type_node, 1, 0, 0},
715 { &long_integer_type_node, 0, 1, 0},
716 { &long_unsigned_type_node, 1, 1, 0},
717 { &long_long_integer_type_node, 0, 1, 1},
718 { &long_long_unsigned_type_node, 1, 1, 1}
726 const cpp_token *tok;
729 timevar_push (TV_CPP);
731 tok = cpp_get_token (parse_in);
732 while (tok->type == CPP_PADDING);
733 timevar_pop (TV_CPP);
735 /* The C++ front end does horrible things with the current line
736 number. To ensure an accurate line number, we must reset it
737 every time we return a token. */
743 case CPP_OPEN_BRACE: indent_level++; break;
744 case CPP_CLOSE_BRACE: indent_level--; break;
746 /* Issue this error here, where we can get at tok->val.c. */
748 if (ISGRAPH (tok->val.c))
749 error ("stray '%c' in program", tok->val.c);
751 error ("stray '\\%o' in program", tok->val.c);
755 *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node));
759 *value = lex_number ((const char *)tok->val.str.text, tok->val.str.len);
764 *value = lex_charconst (tok);
769 *value = lex_string (tok->val.str.text, tok->val.str.len,
770 tok->type == CPP_WSTRING);
773 /* These tokens should not be visible outside cpplib. */
774 case CPP_HEADER_NAME:
785 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
788 lex_number (str, len)
794 int largest_digit = 0;
800 enum anon1 { NOT_FLOAT = 0, AFTER_POINT, AFTER_EXPON } floatflag = NOT_FLOAT;
802 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
803 The code below which fills the parts array assumes that a host
804 int is at least twice as wide as a host char, and that
805 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
806 Two HOST_WIDE_INTs is the largest int literal we can store.
807 In order to detect overflow below, the number of parts (TOTAL_PARTS)
808 must be exactly the number of parts needed to hold the bits
809 of two HOST_WIDE_INTs. */
810 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
811 unsigned int parts[TOTAL_PARTS];
813 /* Optimize for most frequent case. */
817 return integer_zero_node;
818 else if (*str == '1')
819 return integer_one_node;
821 return build_int_2 (*str - '0', 0);
824 for (count = 0; count < TOTAL_PARTS; count++)
827 /* len is known to be >1 at this point. */
830 if (len > 2 && str[0] == '0' && (str[1] == 'x' || str[1] == 'X'))
835 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
836 else if (str[0] == '0' && ISDIGIT (str[1]))
848 if (floatflag == AFTER_POINT)
849 ERROR ("too many decimal points in floating constant");
850 else if (floatflag == AFTER_EXPON)
851 ERROR ("decimal point in exponent - impossible!");
853 floatflag = AFTER_POINT;
859 /* Possible future extension: silently ignore _ in numbers,
860 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
861 but somewhat easier to read. Ada has this? */
862 ERROR ("underscore in number");
866 /* It is not a decimal point.
867 It should be a digit (perhaps a hex digit). */
870 || (base == 16 && ISXDIGIT (c)))
874 else if (base <= 10 && (c == 'e' || c == 'E'))
877 floatflag = AFTER_EXPON;
880 else if (base == 16 && (c == 'p' || c == 'P'))
882 floatflag = AFTER_EXPON;
883 break; /* start of exponent */
888 break; /* start of suffix */
891 if (n >= largest_digit)
895 for (count = 0; count < TOTAL_PARTS; count++)
897 parts[count] *= base;
901 += (parts[count-1] >> HOST_BITS_PER_CHAR);
903 &= (1 << HOST_BITS_PER_CHAR) - 1;
909 /* If the highest-order part overflows (gets larger than
910 a host char will hold) then the whole number has
911 overflowed. Record this and truncate the highest-order
913 if (parts[TOTAL_PARTS - 1] >> HOST_BITS_PER_CHAR)
916 parts[TOTAL_PARTS - 1] &= (1 << HOST_BITS_PER_CHAR) - 1;
920 while (p < str + len);
922 /* This can happen on input like `int i = 0x;' */
924 ERROR ("numeric constant with no digits");
926 if (largest_digit >= base)
927 ERROR ("numeric constant contains digits beyond the radix");
929 if (floatflag != NOT_FLOAT)
932 const char *typename;
933 int imag, fflag, lflag;
934 REAL_VALUE_TYPE real;
937 if (base == 16 && floatflag != AFTER_EXPON)
938 ERROR ("hexadecimal floating constant has no exponent");
940 /* Read explicit exponent if any, and put it in tokenbuf. */
941 if ((base == 10 && ((c == 'e') || (c == 'E')))
942 || (base == 16 && (c == 'p' || c == 'P')))
946 if (p < str + len && (c == '+' || c == '-'))
948 /* Exponent is decimal, even if string is a hex float. */
950 ERROR ("floating constant exponent has no digits");
951 while (p < str + len && ISDIGIT (c))
957 /* Copy the float constant now; we don't want any suffixes in the
958 string passed to parse_float. */
959 copy = alloca (p - str + 1);
960 memcpy (copy, str, p - str);
961 copy[p - str] = '\0';
963 /* Now parse suffixes. */
964 fflag = lflag = imag = 0;
965 while (p < str + len)
970 ERROR ("more than one 'f' suffix on floating constant");
971 else if (warn_traditional && !in_system_header
972 && ! cpp_sys_macro_p (parse_in))
973 warning ("traditional C rejects the 'f' suffix");
980 ERROR ("more than one 'l' suffix on floating constant");
981 else if (warn_traditional && !in_system_header
982 && ! cpp_sys_macro_p (parse_in))
983 warning ("traditional C rejects the 'l' suffix");
991 ERROR ("more than one 'i' or 'j' suffix on floating constant");
993 pedwarn ("ISO C forbids imaginary numeric constants");
998 ERROR ("invalid suffix on floating constant");
1001 type = double_type_node;
1002 typename = "double";
1007 ERROR ("both 'f' and 'l' suffixes on floating constant");
1009 type = float_type_node;
1014 type = long_double_type_node;
1015 typename = "long double";
1017 else if (flag_single_precision_constant)
1019 type = float_type_node;
1023 /* Warn about this only after we know we're not issuing an error. */
1024 if (base == 16 && pedantic && !flag_isoc99)
1025 pedwarn ("hexadecimal floating constants are only valid in C99");
1027 /* The second argument, machine_mode, of REAL_VALUE_ATOF
1028 tells the desired precision of the binary result
1029 of decimal-to-binary conversion. */
1031 real = REAL_VALUE_HTOF (copy, TYPE_MODE (type));
1033 real = REAL_VALUE_ATOF (copy, TYPE_MODE (type));
1035 /* A diagnostic is required here by some ISO C testsuites.
1036 This is not pedwarn, because some people don't want
1037 an error for this. */
1038 if (REAL_VALUE_ISINF (real) && pedantic)
1039 warning ("floating point number exceeds range of 'double'");
1041 /* Create a node with determined type and value. */
1043 value = build_complex (NULL_TREE, convert (type, integer_zero_node),
1044 build_real (type, real));
1046 value = build_real (type, real);
1050 tree trad_type, type;
1051 HOST_WIDE_INT high, low;
1052 int spec_unsigned = 0;
1054 int spec_long_long = 0;
1059 trad_type = type = NULL_TREE;
1060 while (p < str + len)
1067 error ("two 'u' suffixes on integer constant");
1068 else if (warn_traditional && !in_system_header
1069 && ! cpp_sys_macro_p (parse_in))
1070 warning ("traditional C rejects the 'u' suffix");
1081 error ("three 'l' suffixes on integer constant");
1083 error ("'lul' is not a valid integer suffix");
1084 else if (c != spec_long)
1085 error ("'Ll' and 'lL' are not valid integer suffixes");
1086 else if (pedantic && ! flag_isoc99
1087 && ! in_system_header && warn_long_long)
1088 pedwarn ("ISO C89 forbids long long integer constants");
1094 case 'i': case 'I': case 'j': case 'J':
1096 error ("more than one 'i' or 'j' suffix on integer constant");
1098 pedwarn ("ISO C forbids imaginary numeric constants");
1103 ERROR ("invalid suffix on integer constant");
1107 /* If the literal overflowed, pedwarn about it now. */
1111 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT * 2);
1114 /* This is simplified by the fact that our constant
1115 is always positive. */
1119 for (i = 0; i < HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR; i++)
1121 high |= ((HOST_WIDE_INT) parts[i + (HOST_BITS_PER_WIDE_INT
1122 / HOST_BITS_PER_CHAR)]
1123 << (i * HOST_BITS_PER_CHAR));
1124 low |= (HOST_WIDE_INT) parts[i] << (i * HOST_BITS_PER_CHAR);
1127 value = build_int_2 (low, high);
1128 TREE_TYPE (value) = long_long_unsigned_type_node;
1130 /* If warn_traditional, calculate both the ISO type and the
1131 traditional type, then see if they disagree. */
1132 if (warn_traditional)
1134 /* Traditionally, any constant is signed; but if unsigned is
1135 specified explicitly, obey that. Use the smallest size
1136 with the right number of bits, except for one special
1137 case with decimal constants. */
1138 if (! spec_long && base != 10
1139 && int_fits_type_p (value, unsigned_type_node))
1140 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1141 /* A decimal constant must be long if it does not fit in
1142 type int. I think this is independent of whether the
1143 constant is signed. */
1144 else if (! spec_long && base == 10
1145 && int_fits_type_p (value, integer_type_node))
1146 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1147 else if (! spec_long_long)
1148 trad_type = (spec_unsigned
1149 ? long_unsigned_type_node
1150 : long_integer_type_node);
1151 else if (int_fits_type_p (value,
1153 ? long_long_unsigned_type_node
1154 : long_long_integer_type_node))
1155 trad_type = (spec_unsigned
1156 ? long_long_unsigned_type_node
1157 : long_long_integer_type_node);
1159 trad_type = (spec_unsigned
1160 ? widest_unsigned_literal_type_node
1161 : widest_integer_literal_type_node);
1164 /* Calculate the ISO type. */
1165 if (! spec_long && ! spec_unsigned
1166 && int_fits_type_p (value, integer_type_node))
1167 type = integer_type_node;
1168 else if (! spec_long && (base != 10 || spec_unsigned)
1169 && int_fits_type_p (value, unsigned_type_node))
1170 type = unsigned_type_node;
1171 else if (! spec_unsigned && !spec_long_long
1172 && int_fits_type_p (value, long_integer_type_node))
1173 type = long_integer_type_node;
1174 else if (! spec_long_long
1175 && int_fits_type_p (value, long_unsigned_type_node))
1176 type = long_unsigned_type_node;
1177 else if (! spec_unsigned
1178 && int_fits_type_p (value, long_long_integer_type_node))
1179 type = long_long_integer_type_node;
1180 else if (int_fits_type_p (value, long_long_unsigned_type_node))
1181 type = long_long_unsigned_type_node;
1182 else if (! spec_unsigned
1183 && int_fits_type_p (value, widest_integer_literal_type_node))
1184 type = widest_integer_literal_type_node;
1186 type = widest_unsigned_literal_type_node;
1188 /* We assume that constants specified in a non-decimal
1189 base are bit patterns, and that the programmer really
1190 meant what they wrote. */
1191 if (warn_traditional && !in_system_header
1192 && base == 10 && trad_type != type)
1194 if (TYPE_PRECISION (trad_type) != TYPE_PRECISION (type))
1195 warning ("width of integer constant is different in traditional C");
1196 else if (TREE_UNSIGNED (trad_type) != TREE_UNSIGNED (type))
1197 warning ("integer constant is unsigned in ISO C, signed in traditional C");
1199 warning ("width of integer constant may change on other systems in traditional C");
1202 if (pedantic && (flag_isoc99 || !spec_long_long)
1205 ? TYPE_PRECISION (long_long_integer_type_node)
1206 : TYPE_PRECISION (long_integer_type_node)) < TYPE_PRECISION (type)))
1209 pedwarn ("integer constant larger than the maximum value of %s",
1211 ? (TREE_UNSIGNED (type)
1212 ? _("an unsigned long long int")
1213 : _("a long long int"))
1214 : _("an unsigned long int")));
1217 if (base == 10 && ! spec_unsigned && TREE_UNSIGNED (type))
1218 warning ("decimal constant is so large that it is unsigned");
1222 if (TYPE_PRECISION (type)
1223 <= TYPE_PRECISION (integer_type_node))
1224 value = build_complex (NULL_TREE, integer_zero_node,
1225 convert (integer_type_node, value));
1227 ERROR ("complex integer constant is too wide for 'complex int'");
1230 TREE_TYPE (value) = type;
1232 /* If it's still an integer (not a complex), and it doesn't
1233 fit in the type we choose for it, then pedwarn. */
1236 && TREE_CODE (TREE_TYPE (value)) == INTEGER_TYPE
1237 && ! int_fits_type_p (value, TREE_TYPE (value)))
1238 pedwarn ("integer constant is larger than the maximum value for its type");
1242 error ("missing white space after number '%.*s'", (int) (p - str), str);
1247 return integer_zero_node;
1251 lex_string (str, len, wide)
1252 const unsigned char *str;
1257 char *buf = alloca ((len + 1) * (wide ? WCHAR_BYTES : 1));
1259 const unsigned char *p = str, *limit = str + len;
1262 #ifdef MULTIBYTE_CHARS
1263 /* Reset multibyte conversion state. */
1264 (void) local_mbtowc (NULL, NULL, 0);
1269 #ifdef MULTIBYTE_CHARS
1273 char_len = local_mbtowc (&wc, (const char *) p, limit - p);
1276 warning ("ignoring invalid multibyte character");
1289 if (c == '\\' && !ignore_escape_flag)
1290 c = cpp_parse_escape (parse_in, &p, limit, wide);
1292 /* Add this single character into the buffer either as a wchar_t,
1293 a multibyte sequence, or as a single byte. */
1296 unsigned charwidth = TYPE_PRECISION (char_type_node);
1297 unsigned bytemask = (1 << charwidth) - 1;
1300 for (byte = 0; byte < WCHAR_BYTES; ++byte)
1303 if (byte >= (int) sizeof (c))
1306 n = (c >> (byte * charwidth)) & bytemask;
1307 if (BYTES_BIG_ENDIAN)
1308 q[WCHAR_BYTES - byte - 1] = n;
1314 #ifdef MULTIBYTE_CHARS
1315 else if (char_len > 1)
1317 /* We're dealing with a multibyte character. */
1318 for ( ; char_len >0; --char_len)
1320 *q++ = *(p - char_len);
1330 /* Terminate the string value, either with a single byte zero
1331 or with a wide zero. */
1335 memset (q, 0, WCHAR_BYTES);
1343 value = build_string (q - buf, buf);
1346 TREE_TYPE (value) = wchar_array_type_node;
1348 TREE_TYPE (value) = char_array_type_node;
1352 /* Converts a (possibly wide) character constant token into a tree. */
1354 lex_charconst (token)
1355 const cpp_token *token;
1359 unsigned int chars_seen;
1362 result = cpp_interpret_charconst (parse_in, token, warn_multichar,
1363 &chars_seen, &unsignedp);
1365 /* Cast to cppchar_signed_t to get correct sign-extension of RESULT
1366 before possibly widening to HOST_WIDE_INT for build_int_2. */
1367 if (unsignedp || (cppchar_signed_t) result >= 0)
1368 value = build_int_2 (result, 0);
1370 value = build_int_2 ((cppchar_signed_t) result, -1);
1372 if (token->type == CPP_WCHAR)
1373 type = wchar_type_node;
1374 /* In C, a character constant has type 'int'.
1375 In C++ 'char', but multi-char charconsts have type 'int'. */
1376 else if ((c_language == clk_c || c_language == clk_objective_c)
1378 type = integer_type_node;
1380 type = char_type_node;
1382 TREE_TYPE (value) = type;