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 indent_level; /* Number of { minus number of }. */
71 int pending_lang_change; /* If we need to switch languages - C++ only */
72 int c_header_level; /* depth in C headers - C++ only */
74 /* Nonzero tells yylex to ignore \ in string constants. */
75 static int ignore_escape_flag;
77 static tree lex_number PARAMS ((const char *, unsigned int));
78 static tree lex_string PARAMS ((const unsigned char *, unsigned int,
80 static tree lex_charconst PARAMS ((const cpp_token *));
81 static void update_header_times PARAMS ((const char *));
82 static int dump_one_header PARAMS ((splay_tree_node, void *));
83 static void cb_line_change PARAMS ((cpp_reader *, const cpp_token *, int));
84 static void cb_ident PARAMS ((cpp_reader *, unsigned int,
86 static void cb_file_change PARAMS ((cpp_reader *, const struct line_map *));
87 static void cb_def_pragma PARAMS ((cpp_reader *, unsigned int));
88 static void cb_define PARAMS ((cpp_reader *, unsigned int,
90 static void cb_undef PARAMS ((cpp_reader *, unsigned int,
97 struct cpp_callbacks *cb;
98 struct c_fileinfo *toplevel;
100 /* Set up filename timing. Must happen before cpp_read_main_file. */
101 file_info_tree = splay_tree_new ((splay_tree_compare_fn)strcmp,
103 (splay_tree_delete_value_fn)free);
104 toplevel = get_fileinfo ("<top level>");
105 if (flag_detailed_statistics)
108 body_time = get_run_time ();
109 toplevel->time = body_time;
112 #ifdef MULTIBYTE_CHARS
113 /* Change to the native locale for multibyte conversions. */
114 setlocale (LC_CTYPE, "");
115 GET_ENVIRONMENT (literal_codeset, "LANG");
118 cb = cpp_get_callbacks (parse_in);
120 cb->line_change = cb_line_change;
121 cb->ident = cb_ident;
122 cb->file_change = cb_file_change;
123 cb->def_pragma = cb_def_pragma;
125 /* Set the debug callbacks if we can use them. */
126 if (debug_info_level == DINFO_LEVEL_VERBOSE
127 && (write_symbols == DWARF_DEBUG || write_symbols == DWARF2_DEBUG
128 || write_symbols == VMS_AND_DWARF2_DEBUG))
130 cb->define = cb_define;
131 cb->undef = cb_undef;
137 if (filename == NULL || !strcmp (filename, "-"))
140 return cpp_read_main_file (parse_in, filename, ident_hash);
143 /* A thin wrapper around the real parser that initializes the
144 integrated preprocessor after debug output has been initialized.
145 Also, make sure the start_source_file debug hook gets called for
146 the primary source file. */
149 c_common_parse_file (set_yydebug)
150 int set_yydebug ATTRIBUTE_UNUSED;
153 yydebug = set_yydebug;
155 warning ("YYDEBUG not defined");
158 (*debug_hooks->start_source_file) (lineno, input_filename);
159 cpp_finish_options (parse_in);
162 free_parser_stacks ();
170 struct c_fileinfo *fi;
172 n = splay_tree_lookup (file_info_tree, (splay_tree_key) name);
174 return (struct c_fileinfo *) n->value;
176 fi = (struct c_fileinfo *) xmalloc (sizeof (struct c_fileinfo));
178 fi->interface_only = 0;
179 fi->interface_unknown = 1;
180 splay_tree_insert (file_info_tree, (splay_tree_key) name,
181 (splay_tree_value) fi);
186 update_header_times (name)
189 /* Changing files again. This means currently collected time
190 is charged against header time, and body time starts back at 0. */
191 if (flag_detailed_statistics)
193 int this_time = get_run_time ();
194 struct c_fileinfo *file = get_fileinfo (name);
195 header_time += this_time - body_time;
196 file->time += this_time - body_time;
197 body_time = this_time;
202 dump_one_header (n, dummy)
204 void *dummy ATTRIBUTE_UNUSED;
206 print_time ((const char *) n->key,
207 ((struct c_fileinfo *) n->value)->time);
212 dump_time_statistics ()
214 struct c_fileinfo *file = get_fileinfo (input_filename);
215 int this_time = get_run_time ();
216 file->time += this_time - body_time;
218 fprintf (stderr, "\n******\n");
219 print_time ("header files (total)", header_time);
220 print_time ("main file (total)", this_time - body_time);
221 fprintf (stderr, "ratio = %g : 1\n",
222 (double)header_time / (double)(this_time - body_time));
223 fprintf (stderr, "\n******\n");
225 splay_tree_foreach (file_info_tree, dump_one_header, 0);
228 /* Not yet handled: #pragma, #define, #undef.
229 No need to deal with linemarkers under normal conditions. */
232 cb_ident (pfile, line, str)
233 cpp_reader *pfile ATTRIBUTE_UNUSED;
234 unsigned int line ATTRIBUTE_UNUSED;
235 const cpp_string *str ATTRIBUTE_UNUSED;
237 #ifdef ASM_OUTPUT_IDENT
240 /* Convert escapes in the string. */
241 tree value = lex_string (str->text, str->len, 0);
242 ASM_OUTPUT_IDENT (asm_out_file, TREE_STRING_POINTER (value));
247 /* Called at the start of every non-empty line. TOKEN is the first
248 lexed token on the line. Used for diagnostic line numbers. */
250 cb_line_change (pfile, token, parsing_args)
251 cpp_reader *pfile ATTRIBUTE_UNUSED;
252 const cpp_token *token;
253 int parsing_args ATTRIBUTE_UNUSED;
255 src_lineno = SOURCE_LINE (map, token->line);
259 cb_file_change (pfile, new_map)
260 cpp_reader *pfile ATTRIBUTE_UNUSED;
261 const struct line_map *new_map;
263 unsigned int to_line = SOURCE_LINE (new_map, new_map->to_line);
265 if (new_map->reason == LC_ENTER)
267 /* Don't stack the main buffer on the input stack;
268 we already did in compile_file. */
270 main_input_filename = new_map->to_file;
273 int included_at = SOURCE_LINE (new_map - 1, new_map->from_line - 1);
275 lineno = included_at;
276 push_srcloc (new_map->to_file, 1);
277 input_file_stack->indent_level = indent_level;
278 (*debug_hooks->start_source_file) (included_at, new_map->to_file);
279 #ifndef NO_IMPLICIT_EXTERN_C
282 else if (new_map->sysp == 2)
285 ++pending_lang_change;
290 else if (new_map->reason == LC_LEAVE)
292 #ifndef NO_IMPLICIT_EXTERN_C
293 if (c_header_level && --c_header_level == 0)
295 if (new_map->sysp == 2)
296 warning ("badly nested C headers from preprocessor");
297 --pending_lang_change;
301 if (indent_level != input_file_stack->indent_level)
303 warning_with_file_and_line
304 (input_filename, lineno,
305 "this file contains more '%c's than '%c's",
306 indent_level > input_file_stack->indent_level ? '{' : '}',
307 indent_level > input_file_stack->indent_level ? '}' : '{');
312 (*debug_hooks->end_source_file) (to_line);
315 update_header_times (new_map->to_file);
316 in_system_header = new_map->sysp != 0;
317 input_filename = new_map->to_file;
322 extract_interface_info ();
326 cb_def_pragma (pfile, line)
330 /* Issue a warning message if we have been asked to do so. Ignore
331 unknown pragmas in system headers unless an explicit
332 -Wunknown-pragmas has been given. */
333 if (warn_unknown_pragmas > in_system_header)
335 const unsigned char *space, *name = 0;
338 s = cpp_get_token (pfile);
339 space = cpp_token_as_text (pfile, s);
340 s = cpp_get_token (pfile);
341 if (s->type == CPP_NAME)
342 name = cpp_token_as_text (pfile, s);
344 lineno = SOURCE_LINE (map, line);
346 warning ("ignoring #pragma %s %s", space, name);
348 warning ("ignoring #pragma %s", space);
352 /* #define callback for DWARF and DWARF2 debug info. */
354 cb_define (pfile, line, node)
359 (*debug_hooks->define) (SOURCE_LINE (map, line),
360 (const char *) cpp_macro_definition (pfile, node));
363 /* #undef callback for DWARF and DWARF2 debug info. */
365 cb_undef (pfile, line, node)
366 cpp_reader *pfile ATTRIBUTE_UNUSED;
370 (*debug_hooks->undef) (SOURCE_LINE (map, line),
371 (const char *) NODE_NAME (node));
375 /* Returns nonzero if C is a universal-character-name. Give an error if it
376 is not one which may appear in an identifier, as per [extendid].
378 Note that extended character support in identifiers has not yet been
379 implemented. It is my personal opinion that this is not a desirable
380 feature. Portable code cannot count on support for more than the basic
381 identifier character set. */
394 /* None of the valid chars are outside the Basic Multilingual Plane (the
398 error ("universal-character-name '\\U%08x' not valid in identifier", c);
403 if ((c >= 0x00c0 && c <= 0x00d6)
404 || (c >= 0x00d8 && c <= 0x00f6)
405 || (c >= 0x00f8 && c <= 0x01f5)
406 || (c >= 0x01fa && c <= 0x0217)
407 || (c >= 0x0250 && c <= 0x02a8)
408 || (c >= 0x1e00 && c <= 0x1e9a)
409 || (c >= 0x1ea0 && c <= 0x1ef9))
414 || (c >= 0x0388 && c <= 0x038a)
416 || (c >= 0x038e && c <= 0x03a1)
417 || (c >= 0x03a3 && c <= 0x03ce)
418 || (c >= 0x03d0 && c <= 0x03d6)
423 || (c >= 0x03e2 && c <= 0x03f3)
424 || (c >= 0x1f00 && c <= 0x1f15)
425 || (c >= 0x1f18 && c <= 0x1f1d)
426 || (c >= 0x1f20 && c <= 0x1f45)
427 || (c >= 0x1f48 && c <= 0x1f4d)
428 || (c >= 0x1f50 && c <= 0x1f57)
432 || (c >= 0x1f5f && c <= 0x1f7d)
433 || (c >= 0x1f80 && c <= 0x1fb4)
434 || (c >= 0x1fb6 && c <= 0x1fbc)
435 || (c >= 0x1fc2 && c <= 0x1fc4)
436 || (c >= 0x1fc6 && c <= 0x1fcc)
437 || (c >= 0x1fd0 && c <= 0x1fd3)
438 || (c >= 0x1fd6 && c <= 0x1fdb)
439 || (c >= 0x1fe0 && c <= 0x1fec)
440 || (c >= 0x1ff2 && c <= 0x1ff4)
441 || (c >= 0x1ff6 && c <= 0x1ffc))
445 if ((c >= 0x0401 && c <= 0x040d)
446 || (c >= 0x040f && c <= 0x044f)
447 || (c >= 0x0451 && c <= 0x045c)
448 || (c >= 0x045e && c <= 0x0481)
449 || (c >= 0x0490 && c <= 0x04c4)
450 || (c >= 0x04c7 && c <= 0x04c8)
451 || (c >= 0x04cb && c <= 0x04cc)
452 || (c >= 0x04d0 && c <= 0x04eb)
453 || (c >= 0x04ee && c <= 0x04f5)
454 || (c >= 0x04f8 && c <= 0x04f9))
458 if ((c >= 0x0531 && c <= 0x0556)
459 || (c >= 0x0561 && c <= 0x0587))
463 if ((c >= 0x05d0 && c <= 0x05ea)
464 || (c >= 0x05f0 && c <= 0x05f4))
468 if ((c >= 0x0621 && c <= 0x063a)
469 || (c >= 0x0640 && c <= 0x0652)
470 || (c >= 0x0670 && c <= 0x06b7)
471 || (c >= 0x06ba && c <= 0x06be)
472 || (c >= 0x06c0 && c <= 0x06ce)
473 || (c >= 0x06e5 && c <= 0x06e7))
477 if ((c >= 0x0905 && c <= 0x0939)
478 || (c >= 0x0958 && c <= 0x0962))
482 if ((c >= 0x0985 && c <= 0x098c)
483 || (c >= 0x098f && c <= 0x0990)
484 || (c >= 0x0993 && c <= 0x09a8)
485 || (c >= 0x09aa && c <= 0x09b0)
487 || (c >= 0x09b6 && c <= 0x09b9)
488 || (c >= 0x09dc && c <= 0x09dd)
489 || (c >= 0x09df && c <= 0x09e1)
490 || (c >= 0x09f0 && c <= 0x09f1))
494 if ((c >= 0x0a05 && c <= 0x0a0a)
495 || (c >= 0x0a0f && c <= 0x0a10)
496 || (c >= 0x0a13 && c <= 0x0a28)
497 || (c >= 0x0a2a && c <= 0x0a30)
498 || (c >= 0x0a32 && c <= 0x0a33)
499 || (c >= 0x0a35 && c <= 0x0a36)
500 || (c >= 0x0a38 && c <= 0x0a39)
501 || (c >= 0x0a59 && c <= 0x0a5c)
506 if ((c >= 0x0a85 && c <= 0x0a8b)
508 || (c >= 0x0a8f && c <= 0x0a91)
509 || (c >= 0x0a93 && c <= 0x0aa8)
510 || (c >= 0x0aaa && c <= 0x0ab0)
511 || (c >= 0x0ab2 && c <= 0x0ab3)
512 || (c >= 0x0ab5 && c <= 0x0ab9)
517 if ((c >= 0x0b05 && c <= 0x0b0c)
518 || (c >= 0x0b0f && c <= 0x0b10)
519 || (c >= 0x0b13 && c <= 0x0b28)
520 || (c >= 0x0b2a && c <= 0x0b30)
521 || (c >= 0x0b32 && c <= 0x0b33)
522 || (c >= 0x0b36 && c <= 0x0b39)
523 || (c >= 0x0b5c && c <= 0x0b5d)
524 || (c >= 0x0b5f && c <= 0x0b61))
528 if ((c >= 0x0b85 && c <= 0x0b8a)
529 || (c >= 0x0b8e && c <= 0x0b90)
530 || (c >= 0x0b92 && c <= 0x0b95)
531 || (c >= 0x0b99 && c <= 0x0b9a)
533 || (c >= 0x0b9e && c <= 0x0b9f)
534 || (c >= 0x0ba3 && c <= 0x0ba4)
535 || (c >= 0x0ba8 && c <= 0x0baa)
536 || (c >= 0x0bae && c <= 0x0bb5)
537 || (c >= 0x0bb7 && c <= 0x0bb9))
541 if ((c >= 0x0c05 && c <= 0x0c0c)
542 || (c >= 0x0c0e && c <= 0x0c10)
543 || (c >= 0x0c12 && c <= 0x0c28)
544 || (c >= 0x0c2a && c <= 0x0c33)
545 || (c >= 0x0c35 && c <= 0x0c39)
546 || (c >= 0x0c60 && c <= 0x0c61))
550 if ((c >= 0x0c85 && c <= 0x0c8c)
551 || (c >= 0x0c8e && c <= 0x0c90)
552 || (c >= 0x0c92 && c <= 0x0ca8)
553 || (c >= 0x0caa && c <= 0x0cb3)
554 || (c >= 0x0cb5 && c <= 0x0cb9)
555 || (c >= 0x0ce0 && c <= 0x0ce1))
559 if ((c >= 0x0d05 && c <= 0x0d0c)
560 || (c >= 0x0d0e && c <= 0x0d10)
561 || (c >= 0x0d12 && c <= 0x0d28)
562 || (c >= 0x0d2a && c <= 0x0d39)
563 || (c >= 0x0d60 && c <= 0x0d61))
567 if ((c >= 0x0e01 && c <= 0x0e30)
568 || (c >= 0x0e32 && c <= 0x0e33)
569 || (c >= 0x0e40 && c <= 0x0e46)
570 || (c >= 0x0e4f && c <= 0x0e5b))
574 if ((c >= 0x0e81 && c <= 0x0e82)
580 || (c >= 0x0e94 && c <= 0x0e97)
581 || (c >= 0x0e99 && c <= 0x0e9f)
582 || (c >= 0x0ea1 && c <= 0x0ea3)
587 || (c >= 0x0ead && c <= 0x0eb0)
591 || (c >= 0x0ec0 && c <= 0x0ec4)
596 if ((c >= 0x10a0 && c <= 0x10c5)
597 || (c >= 0x10d0 && c <= 0x10f6))
601 if ((c >= 0x3041 && c <= 0x3094)
602 || (c >= 0x309b && c <= 0x309e))
606 if ((c >= 0x30a1 && c <= 0x30fe))
610 if ((c >= 0x3105 && c <= 0x312c))
614 if ((c >= 0x1100 && c <= 0x1159)
615 || (c >= 0x1161 && c <= 0x11a2)
616 || (c >= 0x11a8 && c <= 0x11f9))
619 /* CJK Unified Ideographs */
620 if ((c >= 0xf900 && c <= 0xfa2d)
621 || (c >= 0xfb1f && c <= 0xfb36)
622 || (c >= 0xfb38 && c <= 0xfb3c)
624 || (c >= 0xfb40 && c <= 0xfb41)
625 || (c >= 0xfb42 && c <= 0xfb44)
626 || (c >= 0xfb46 && c <= 0xfbb1)
627 || (c >= 0xfbd3 && c <= 0xfd3f)
628 || (c >= 0xfd50 && c <= 0xfd8f)
629 || (c >= 0xfd92 && c <= 0xfdc7)
630 || (c >= 0xfdf0 && c <= 0xfdfb)
631 || (c >= 0xfe70 && c <= 0xfe72)
633 || (c >= 0xfe76 && c <= 0xfefc)
634 || (c >= 0xff21 && c <= 0xff3a)
635 || (c >= 0xff41 && c <= 0xff5a)
636 || (c >= 0xff66 && c <= 0xffbe)
637 || (c >= 0xffc2 && c <= 0xffc7)
638 || (c >= 0xffca && c <= 0xffcf)
639 || (c >= 0xffd2 && c <= 0xffd7)
640 || (c >= 0xffda && c <= 0xffdc)
641 || (c >= 0x4e00 && c <= 0x9fa5))
644 error ("universal-character-name '\\u%04x' not valid in identifier", c);
649 /* Add the UTF-8 representation of C to the token_buffer. */
652 utf8_extend_token (c)
662 else if (c <= 0x000007ff)
663 shift = 6, mask = 0xc0;
664 else if (c <= 0x0000ffff)
665 shift = 12, mask = 0xe0;
666 else if (c <= 0x001fffff)
667 shift = 18, mask = 0xf0;
668 else if (c <= 0x03ffffff)
669 shift = 24, mask = 0xf8;
671 shift = 30, mask = 0xfc;
673 extend_token (mask | (c >> shift));
677 extend_token ((unsigned char) (0x80 | (c >> shift)));
686 tree *const node_var;
687 const char unsigned_flag;
688 const char long_flag;
689 const char long_long_flag;
692 struct try_type type_sequence[] =
694 { &integer_type_node, 0, 0, 0},
695 { &unsigned_type_node, 1, 0, 0},
696 { &long_integer_type_node, 0, 1, 0},
697 { &long_unsigned_type_node, 1, 1, 0},
698 { &long_long_integer_type_node, 0, 1, 1},
699 { &long_long_unsigned_type_node, 1, 1, 1}
707 const cpp_token *tok;
710 timevar_push (TV_CPP);
712 tok = cpp_get_token (parse_in);
713 while (tok->type == CPP_PADDING);
714 timevar_pop (TV_CPP);
716 /* The C++ front end does horrible things with the current line
717 number. To ensure an accurate line number, we must reset it
718 every time we return a token. */
724 case CPP_OPEN_BRACE: indent_level++; break;
725 case CPP_CLOSE_BRACE: indent_level--; break;
727 /* Issue this error here, where we can get at tok->val.c. */
729 if (ISGRAPH (tok->val.c))
730 error ("stray '%c' in program", tok->val.c);
732 error ("stray '\\%o' in program", tok->val.c);
736 *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node));
740 *value = lex_number ((const char *)tok->val.str.text, tok->val.str.len);
745 *value = lex_charconst (tok);
750 *value = lex_string (tok->val.str.text, tok->val.str.len,
751 tok->type == CPP_WSTRING);
754 /* These tokens should not be visible outside cpplib. */
755 case CPP_HEADER_NAME:
766 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
769 lex_number (str, len)
775 int largest_digit = 0;
781 enum anon1 { NOT_FLOAT = 0, AFTER_POINT, AFTER_EXPON } floatflag = NOT_FLOAT;
783 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
784 The code below which fills the parts array assumes that a host
785 int is at least twice as wide as a host char, and that
786 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
787 Two HOST_WIDE_INTs is the largest int literal we can store.
788 In order to detect overflow below, the number of parts (TOTAL_PARTS)
789 must be exactly the number of parts needed to hold the bits
790 of two HOST_WIDE_INTs. */
791 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
792 unsigned int parts[TOTAL_PARTS];
794 /* Optimize for most frequent case. */
798 return integer_zero_node;
799 else if (*str == '1')
800 return integer_one_node;
802 return build_int_2 (*str - '0', 0);
805 for (count = 0; count < TOTAL_PARTS; count++)
808 /* len is known to be >1 at this point. */
811 if (len > 2 && str[0] == '0' && (str[1] == 'x' || str[1] == 'X'))
816 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
817 else if (str[0] == '0' && ISDIGIT (str[1]))
829 if (floatflag == AFTER_POINT)
830 ERROR ("too many decimal points in floating constant");
831 else if (floatflag == AFTER_EXPON)
832 ERROR ("decimal point in exponent - impossible!");
834 floatflag = AFTER_POINT;
840 /* Possible future extension: silently ignore _ in numbers,
841 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
842 but somewhat easier to read. Ada has this? */
843 ERROR ("underscore in number");
847 /* It is not a decimal point.
848 It should be a digit (perhaps a hex digit). */
851 || (base == 16 && ISXDIGIT (c)))
855 else if (base <= 10 && (c == 'e' || c == 'E'))
858 floatflag = AFTER_EXPON;
861 else if (base == 16 && (c == 'p' || c == 'P'))
863 floatflag = AFTER_EXPON;
864 break; /* start of exponent */
869 break; /* start of suffix */
872 if (n >= largest_digit)
876 for (count = 0; count < TOTAL_PARTS; count++)
878 parts[count] *= base;
882 += (parts[count-1] >> HOST_BITS_PER_CHAR);
884 &= (1 << HOST_BITS_PER_CHAR) - 1;
890 /* If the highest-order part overflows (gets larger than
891 a host char will hold) then the whole number has
892 overflowed. Record this and truncate the highest-order
894 if (parts[TOTAL_PARTS - 1] >> HOST_BITS_PER_CHAR)
897 parts[TOTAL_PARTS - 1] &= (1 << HOST_BITS_PER_CHAR) - 1;
901 while (p < str + len);
903 /* This can happen on input like `int i = 0x;' */
905 ERROR ("numeric constant with no digits");
907 if (largest_digit >= base)
908 ERROR ("numeric constant contains digits beyond the radix");
910 if (floatflag != NOT_FLOAT)
913 const char *typename;
914 int imag, fflag, lflag;
915 REAL_VALUE_TYPE real;
918 if (base == 16 && floatflag != AFTER_EXPON)
919 ERROR ("hexadecimal floating constant has no exponent");
921 /* Read explicit exponent if any, and put it in tokenbuf. */
922 if ((base == 10 && ((c == 'e') || (c == 'E')))
923 || (base == 16 && (c == 'p' || c == 'P')))
927 if (p < str + len && (c == '+' || c == '-'))
929 /* Exponent is decimal, even if string is a hex float. */
931 ERROR ("floating constant exponent has no digits");
932 while (p < str + len && ISDIGIT (c))
938 /* Copy the float constant now; we don't want any suffixes in the
939 string passed to parse_float. */
940 copy = alloca (p - str + 1);
941 memcpy (copy, str, p - str);
942 copy[p - str] = '\0';
944 /* Now parse suffixes. */
945 fflag = lflag = imag = 0;
946 while (p < str + len)
951 ERROR ("more than one 'f' suffix on floating constant");
952 else if (warn_traditional && !in_system_header
953 && ! cpp_sys_macro_p (parse_in))
954 warning ("traditional C rejects the 'f' suffix");
961 ERROR ("more than one 'l' suffix on floating constant");
962 else if (warn_traditional && !in_system_header
963 && ! cpp_sys_macro_p (parse_in))
964 warning ("traditional C rejects the 'l' suffix");
972 ERROR ("more than one 'i' or 'j' suffix on floating constant");
974 pedwarn ("ISO C forbids imaginary numeric constants");
979 ERROR ("invalid suffix on floating constant");
982 type = double_type_node;
988 ERROR ("both 'f' and 'l' suffixes on floating constant");
990 type = float_type_node;
995 type = long_double_type_node;
996 typename = "long double";
998 else if (flag_single_precision_constant)
1000 type = float_type_node;
1004 /* Warn about this only after we know we're not issuing an error. */
1005 if (base == 16 && pedantic && !flag_isoc99)
1006 pedwarn ("hexadecimal floating constants are only valid in C99");
1008 /* The second argument, machine_mode, of REAL_VALUE_ATOF
1009 tells the desired precision of the binary result
1010 of decimal-to-binary conversion. */
1012 real = REAL_VALUE_HTOF (copy, TYPE_MODE (type));
1014 real = REAL_VALUE_ATOF (copy, TYPE_MODE (type));
1016 /* A diagnostic is required here by some ISO C testsuites.
1017 This is not pedwarn, because some people don't want
1018 an error for this. */
1019 if (REAL_VALUE_ISINF (real) && pedantic)
1020 warning ("floating point number exceeds range of 'double'");
1022 /* Create a node with determined type and value. */
1024 value = build_complex (NULL_TREE, convert (type, integer_zero_node),
1025 build_real (type, real));
1027 value = build_real (type, real);
1031 tree trad_type, type;
1032 HOST_WIDE_INT high, low;
1033 int spec_unsigned = 0;
1035 int spec_long_long = 0;
1040 trad_type = type = NULL_TREE;
1041 while (p < str + len)
1048 error ("two 'u' suffixes on integer constant");
1049 else if (warn_traditional && !in_system_header
1050 && ! cpp_sys_macro_p (parse_in))
1051 warning ("traditional C rejects the 'u' suffix");
1062 error ("three 'l' suffixes on integer constant");
1064 error ("'lul' is not a valid integer suffix");
1065 else if (c != spec_long)
1066 error ("'Ll' and 'lL' are not valid integer suffixes");
1067 else if (pedantic && ! flag_isoc99
1068 && ! in_system_header && warn_long_long)
1069 pedwarn ("ISO C89 forbids long long integer constants");
1075 case 'i': case 'I': case 'j': case 'J':
1077 error ("more than one 'i' or 'j' suffix on integer constant");
1079 pedwarn ("ISO C forbids imaginary numeric constants");
1084 ERROR ("invalid suffix on integer constant");
1088 /* If the literal overflowed, pedwarn about it now. */
1092 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT * 2);
1095 /* This is simplified by the fact that our constant
1096 is always positive. */
1100 for (i = 0; i < HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR; i++)
1102 high |= ((HOST_WIDE_INT) parts[i + (HOST_BITS_PER_WIDE_INT
1103 / HOST_BITS_PER_CHAR)]
1104 << (i * HOST_BITS_PER_CHAR));
1105 low |= (HOST_WIDE_INT) parts[i] << (i * HOST_BITS_PER_CHAR);
1108 value = build_int_2 (low, high);
1109 TREE_TYPE (value) = long_long_unsigned_type_node;
1111 /* If warn_traditional, calculate both the ISO type and the
1112 traditional type, then see if they disagree. */
1113 if (warn_traditional)
1115 /* Traditionally, any constant is signed; but if unsigned is
1116 specified explicitly, obey that. Use the smallest size
1117 with the right number of bits, except for one special
1118 case with decimal constants. */
1119 if (! spec_long && base != 10
1120 && int_fits_type_p (value, unsigned_type_node))
1121 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1122 /* A decimal constant must be long if it does not fit in
1123 type int. I think this is independent of whether the
1124 constant is signed. */
1125 else if (! spec_long && base == 10
1126 && int_fits_type_p (value, integer_type_node))
1127 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1128 else if (! spec_long_long)
1129 trad_type = (spec_unsigned
1130 ? long_unsigned_type_node
1131 : long_integer_type_node);
1132 else if (int_fits_type_p (value,
1134 ? long_long_unsigned_type_node
1135 : long_long_integer_type_node))
1136 trad_type = (spec_unsigned
1137 ? long_long_unsigned_type_node
1138 : long_long_integer_type_node);
1140 trad_type = (spec_unsigned
1141 ? widest_unsigned_literal_type_node
1142 : widest_integer_literal_type_node);
1145 /* Calculate the ISO type. */
1146 if (! spec_long && ! spec_unsigned
1147 && int_fits_type_p (value, integer_type_node))
1148 type = integer_type_node;
1149 else if (! spec_long && (base != 10 || spec_unsigned)
1150 && int_fits_type_p (value, unsigned_type_node))
1151 type = unsigned_type_node;
1152 else if (! spec_unsigned && !spec_long_long
1153 && int_fits_type_p (value, long_integer_type_node))
1154 type = long_integer_type_node;
1155 else if (! spec_long_long
1156 && int_fits_type_p (value, long_unsigned_type_node))
1157 type = long_unsigned_type_node;
1158 else if (! spec_unsigned
1159 && int_fits_type_p (value, long_long_integer_type_node))
1160 type = long_long_integer_type_node;
1161 else if (int_fits_type_p (value, long_long_unsigned_type_node))
1162 type = long_long_unsigned_type_node;
1163 else if (! spec_unsigned
1164 && int_fits_type_p (value, widest_integer_literal_type_node))
1165 type = widest_integer_literal_type_node;
1167 type = widest_unsigned_literal_type_node;
1169 /* We assume that constants specified in a non-decimal
1170 base are bit patterns, and that the programmer really
1171 meant what they wrote. */
1172 if (warn_traditional && !in_system_header
1173 && base == 10 && trad_type != type)
1175 if (TYPE_PRECISION (trad_type) != TYPE_PRECISION (type))
1176 warning ("width of integer constant is different in traditional C");
1177 else if (TREE_UNSIGNED (trad_type) != TREE_UNSIGNED (type))
1178 warning ("integer constant is unsigned in ISO C, signed in traditional C");
1180 warning ("width of integer constant may change on other systems in traditional C");
1183 if (pedantic && (flag_isoc99 || !spec_long_long)
1186 ? TYPE_PRECISION (long_long_integer_type_node)
1187 : TYPE_PRECISION (long_integer_type_node)) < TYPE_PRECISION (type)))
1190 pedwarn ("integer constant larger than the maximum value of %s",
1192 ? (TREE_UNSIGNED (type)
1193 ? _("an unsigned long long int")
1194 : _("a long long int"))
1195 : _("an unsigned long int")));
1198 if (base == 10 && ! spec_unsigned && TREE_UNSIGNED (type))
1199 warning ("decimal constant is so large that it is unsigned");
1203 if (TYPE_PRECISION (type)
1204 <= TYPE_PRECISION (integer_type_node))
1205 value = build_complex (NULL_TREE, integer_zero_node,
1206 convert (integer_type_node, value));
1208 ERROR ("complex integer constant is too wide for 'complex int'");
1211 TREE_TYPE (value) = type;
1213 /* If it's still an integer (not a complex), and it doesn't
1214 fit in the type we choose for it, then pedwarn. */
1217 && TREE_CODE (TREE_TYPE (value)) == INTEGER_TYPE
1218 && ! int_fits_type_p (value, TREE_TYPE (value)))
1219 pedwarn ("integer constant is larger than the maximum value for its type");
1223 error ("missing white space after number '%.*s'", (int) (p - str), str);
1228 return integer_zero_node;
1232 lex_string (str, len, wide)
1233 const unsigned char *str;
1238 char *buf = alloca ((len + 1) * (wide ? WCHAR_BYTES : 1));
1240 const unsigned char *p = str, *limit = str + len;
1242 unsigned width = wide ? WCHAR_TYPE_SIZE
1243 : TYPE_PRECISION (char_type_node);
1245 #ifdef MULTIBYTE_CHARS
1246 /* Reset multibyte conversion state. */
1247 (void) local_mbtowc (NULL, NULL, 0);
1252 #ifdef MULTIBYTE_CHARS
1256 char_len = local_mbtowc (&wc, (const char *) p, limit - p);
1259 warning ("ignoring invalid multibyte character");
1272 if (c == '\\' && !ignore_escape_flag)
1276 if (width < HOST_BITS_PER_INT)
1277 mask = ((unsigned int) 1 << width) - 1;
1280 c = cpp_parse_escape (parse_in, &p, limit, mask);
1283 /* Add this single character into the buffer either as a wchar_t,
1284 a multibyte sequence, or as a single byte. */
1287 unsigned charwidth = TYPE_PRECISION (char_type_node);
1288 unsigned bytemask = (1 << charwidth) - 1;
1291 for (byte = 0; byte < WCHAR_BYTES; ++byte)
1294 if (byte >= (int) sizeof (c))
1297 n = (c >> (byte * charwidth)) & bytemask;
1298 if (BYTES_BIG_ENDIAN)
1299 q[WCHAR_BYTES - byte - 1] = n;
1305 #ifdef MULTIBYTE_CHARS
1306 else if (char_len > 1)
1308 /* We're dealing with a multibyte character. */
1309 for ( ; char_len >0; --char_len)
1311 *q++ = *(p - char_len);
1321 /* Terminate the string value, either with a single byte zero
1322 or with a wide zero. */
1326 memset (q, 0, WCHAR_BYTES);
1334 value = build_string (q - buf, buf);
1337 TREE_TYPE (value) = wchar_array_type_node;
1339 TREE_TYPE (value) = char_array_type_node;
1343 /* Converts a (possibly wide) character constant token into a tree. */
1345 lex_charconst (token)
1346 const cpp_token *token;
1348 HOST_WIDE_INT result;
1350 unsigned int chars_seen;
1352 result = cpp_interpret_charconst (parse_in, token, warn_multichar,
1354 if (token->type == CPP_WCHAR)
1356 value = build_int_2 (result, 0);
1357 type = wchar_type_node;
1362 value = build_int_2 (result, -1);
1364 value = build_int_2 (result, 0);
1366 /* In C, a character constant has type 'int'.
1367 In C++ 'char', but multi-char charconsts have type 'int'. */
1368 if (c_language == clk_cplusplus && chars_seen <= 1)
1369 type = char_type_node;
1371 type = integer_type_node;
1374 /* cpp_interpret_charconst issues a warning if the constant
1375 overflows, but if the number fits in HOST_WIDE_INT anyway, it
1376 will return it un-truncated, which may cause problems down the
1377 line. So set the type to widest_integer_literal_type, call
1378 convert to truncate it to the proper type, then clear
1379 TREE_OVERFLOW so we don't get a second warning.
1381 FIXME: cpplib's assessment of overflow may not be accurate on a
1382 platform where the final type can change at (compiler's) runtime. */
1384 TREE_TYPE (value) = widest_integer_literal_type_node;
1385 value = convert (type, value);
1386 TREE_OVERFLOW (value) = 0;