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 /* MULTIBYTE_CHARS support only works for native compilers.
44 ??? Ideally what we want is to model widechar support after
45 the current floating point support. */
47 #undef MULTIBYTE_CHARS
50 #ifdef MULTIBYTE_CHARS
53 #endif /* MULTIBYTE_CHARS */
54 #ifndef GET_ENVIRONMENT
55 #define GET_ENVIRONMENT(ENV_VALUE,ENV_NAME) ((ENV_VALUE) = getenv (ENV_NAME))
58 /* The current line map. */
59 static const struct line_map *map;
61 /* The line used to refresh the lineno global variable after each token. */
62 static unsigned int src_lineno;
64 /* We may keep statistics about how long which files took to compile. */
65 static int header_time, body_time;
66 static splay_tree file_info_tree;
68 /* Cause the `yydebug' variable to be defined. */
71 /* File used for outputting assembler code. */
72 extern FILE *asm_out_file;
74 #undef WCHAR_TYPE_SIZE
75 #define WCHAR_TYPE_SIZE TYPE_PRECISION (wchar_type_node)
77 /* Number of bytes in a wide character. */
78 #define WCHAR_BYTES (WCHAR_TYPE_SIZE / BITS_PER_UNIT)
80 int indent_level; /* Number of { minus number of }. */
81 int pending_lang_change; /* If we need to switch languages - C++ only */
82 int c_header_level; /* depth in C headers - C++ only */
84 /* Nonzero tells yylex to ignore \ in string constants. */
85 static int ignore_escape_flag;
87 static void parse_float PARAMS ((PTR));
88 static tree lex_number PARAMS ((const char *, unsigned int));
89 static tree lex_string PARAMS ((const unsigned char *, unsigned int,
91 static tree lex_charconst PARAMS ((const cpp_token *));
92 static void update_header_times PARAMS ((const char *));
93 static int dump_one_header PARAMS ((splay_tree_node, void *));
94 static void cb_line_change PARAMS ((cpp_reader *, const cpp_token *, int));
95 static void cb_ident PARAMS ((cpp_reader *, unsigned int,
97 static void cb_file_change PARAMS ((cpp_reader *, const struct line_map *));
98 static void cb_def_pragma PARAMS ((cpp_reader *, unsigned int));
99 static void cb_define PARAMS ((cpp_reader *, unsigned int,
101 static void cb_undef PARAMS ((cpp_reader *, unsigned int,
105 init_c_lex (filename)
106 const char *filename;
108 struct cpp_callbacks *cb;
109 struct c_fileinfo *toplevel;
111 /* Set up filename timing. Must happen before cpp_read_main_file. */
112 file_info_tree = splay_tree_new ((splay_tree_compare_fn)strcmp,
114 (splay_tree_delete_value_fn)free);
115 toplevel = get_fileinfo ("<top level>");
116 if (flag_detailed_statistics)
119 body_time = get_run_time ();
120 toplevel->time = body_time;
123 #ifdef MULTIBYTE_CHARS
124 /* Change to the native locale for multibyte conversions. */
125 setlocale (LC_CTYPE, "");
126 GET_ENVIRONMENT (literal_codeset, "LANG");
129 cb = cpp_get_callbacks (parse_in);
131 cb->line_change = cb_line_change;
132 cb->ident = cb_ident;
133 cb->file_change = cb_file_change;
134 cb->def_pragma = cb_def_pragma;
136 /* Set the debug callbacks if we can use them. */
137 if (debug_info_level == DINFO_LEVEL_VERBOSE
138 && (write_symbols == DWARF_DEBUG || write_symbols == DWARF2_DEBUG
139 || write_symbols == VMS_AND_DWARF2_DEBUG))
141 cb->define = cb_define;
142 cb->undef = cb_undef;
148 if (filename == NULL || !strcmp (filename, "-"))
151 return cpp_read_main_file (parse_in, filename, ident_hash);
154 /* A thin wrapper around the real parser that initializes the
155 integrated preprocessor after debug output has been initialized.
156 Also, make sure the start_source_file debug hook gets called for
157 the primary source file. */
160 c_common_parse_file ()
162 (*debug_hooks->start_source_file) (lineno, input_filename);
163 cpp_finish_options (parse_in);
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 /* Not yet handled: #pragma, #define, #undef.
232 No need to deal with linemarkers under normal conditions. */
235 cb_ident (pfile, line, str)
236 cpp_reader *pfile ATTRIBUTE_UNUSED;
237 unsigned int line ATTRIBUTE_UNUSED;
238 const cpp_string *str ATTRIBUTE_UNUSED;
240 #ifdef ASM_OUTPUT_IDENT
243 /* Convert escapes in the string. */
244 tree value = lex_string (str->text, str->len, 0);
245 ASM_OUTPUT_IDENT (asm_out_file, TREE_STRING_POINTER (value));
250 /* Called at the start of every non-empty line. TOKEN is the first
251 lexed token on the line. Used for diagnostic line numbers. */
253 cb_line_change (pfile, token, parsing_args)
254 cpp_reader *pfile ATTRIBUTE_UNUSED;
255 const cpp_token *token;
256 int parsing_args ATTRIBUTE_UNUSED;
258 src_lineno = SOURCE_LINE (map, token->line);
262 cb_file_change (pfile, new_map)
263 cpp_reader *pfile ATTRIBUTE_UNUSED;
264 const struct line_map *new_map;
266 unsigned int to_line = SOURCE_LINE (new_map, new_map->to_line);
268 if (new_map->reason == LC_ENTER)
270 /* Don't stack the main buffer on the input stack;
271 we already did in compile_file. */
273 main_input_filename = new_map->to_file;
276 lineno = SOURCE_LINE (new_map - 1, new_map->from_line - 1);
277 push_srcloc (new_map->to_file, 1);
278 input_file_stack->indent_level = indent_level;
279 (*debug_hooks->start_source_file) (lineno, new_map->to_file);
280 #ifndef NO_IMPLICIT_EXTERN_C
283 else if (new_map->sysp == 2)
286 ++pending_lang_change;
291 else if (new_map->reason == LC_LEAVE)
293 #ifndef NO_IMPLICIT_EXTERN_C
294 if (c_header_level && --c_header_level == 0)
296 if (new_map->sysp == 2)
297 warning ("badly nested C headers from preprocessor");
298 --pending_lang_change;
302 if (indent_level != input_file_stack->indent_level)
304 warning_with_file_and_line
305 (input_filename, lineno,
306 "this file contains more '%c's than '%c's",
307 indent_level > input_file_stack->indent_level ? '{' : '}',
308 indent_level > input_file_stack->indent_level ? '}' : '{');
313 (*debug_hooks->end_source_file) (to_line);
316 update_header_times (new_map->to_file);
317 in_system_header = new_map->sysp != 0;
318 input_filename = new_map->to_file;
323 extract_interface_info ();
327 cb_def_pragma (pfile, line)
331 /* Issue a warning message if we have been asked to do so. Ignore
332 unknown pragmas in system headers unless an explicit
333 -Wunknown-pragmas has been given. */
334 if (warn_unknown_pragmas > in_system_header)
336 const unsigned char *space, *name = 0;
339 s = cpp_get_token (pfile);
340 space = cpp_token_as_text (pfile, s);
341 s = cpp_get_token (pfile);
342 if (s->type == CPP_NAME)
343 name = cpp_token_as_text (pfile, s);
345 lineno = SOURCE_LINE (map, line);
347 warning ("ignoring #pragma %s %s", space, name);
349 warning ("ignoring #pragma %s", space);
353 /* #define callback for DWARF and DWARF2 debug info. */
355 cb_define (pfile, line, node)
360 (*debug_hooks->define) (SOURCE_LINE (map, line),
361 (const char *) cpp_macro_definition (pfile, node));
364 /* #undef callback for DWARF and DWARF2 debug info. */
366 cb_undef (pfile, line, node)
367 cpp_reader *pfile ATTRIBUTE_UNUSED;
371 (*debug_hooks->undef) (SOURCE_LINE (map, line),
372 (const char *) NODE_NAME (node));
376 /* Returns nonzero if C is a universal-character-name. Give an error if it
377 is not one which may appear in an identifier, as per [extendid].
379 Note that extended character support in identifiers has not yet been
380 implemented. It is my personal opinion that this is not a desirable
381 feature. Portable code cannot count on support for more than the basic
382 identifier character set. */
395 /* None of the valid chars are outside the Basic Multilingual Plane (the
399 error ("universal-character-name '\\U%08x' not valid in identifier", c);
404 if ((c >= 0x00c0 && c <= 0x00d6)
405 || (c >= 0x00d8 && c <= 0x00f6)
406 || (c >= 0x00f8 && c <= 0x01f5)
407 || (c >= 0x01fa && c <= 0x0217)
408 || (c >= 0x0250 && c <= 0x02a8)
409 || (c >= 0x1e00 && c <= 0x1e9a)
410 || (c >= 0x1ea0 && c <= 0x1ef9))
415 || (c >= 0x0388 && c <= 0x038a)
417 || (c >= 0x038e && c <= 0x03a1)
418 || (c >= 0x03a3 && c <= 0x03ce)
419 || (c >= 0x03d0 && c <= 0x03d6)
424 || (c >= 0x03e2 && c <= 0x03f3)
425 || (c >= 0x1f00 && c <= 0x1f15)
426 || (c >= 0x1f18 && c <= 0x1f1d)
427 || (c >= 0x1f20 && c <= 0x1f45)
428 || (c >= 0x1f48 && c <= 0x1f4d)
429 || (c >= 0x1f50 && c <= 0x1f57)
433 || (c >= 0x1f5f && c <= 0x1f7d)
434 || (c >= 0x1f80 && c <= 0x1fb4)
435 || (c >= 0x1fb6 && c <= 0x1fbc)
436 || (c >= 0x1fc2 && c <= 0x1fc4)
437 || (c >= 0x1fc6 && c <= 0x1fcc)
438 || (c >= 0x1fd0 && c <= 0x1fd3)
439 || (c >= 0x1fd6 && c <= 0x1fdb)
440 || (c >= 0x1fe0 && c <= 0x1fec)
441 || (c >= 0x1ff2 && c <= 0x1ff4)
442 || (c >= 0x1ff6 && c <= 0x1ffc))
446 if ((c >= 0x0401 && c <= 0x040d)
447 || (c >= 0x040f && c <= 0x044f)
448 || (c >= 0x0451 && c <= 0x045c)
449 || (c >= 0x045e && c <= 0x0481)
450 || (c >= 0x0490 && c <= 0x04c4)
451 || (c >= 0x04c7 && c <= 0x04c8)
452 || (c >= 0x04cb && c <= 0x04cc)
453 || (c >= 0x04d0 && c <= 0x04eb)
454 || (c >= 0x04ee && c <= 0x04f5)
455 || (c >= 0x04f8 && c <= 0x04f9))
459 if ((c >= 0x0531 && c <= 0x0556)
460 || (c >= 0x0561 && c <= 0x0587))
464 if ((c >= 0x05d0 && c <= 0x05ea)
465 || (c >= 0x05f0 && c <= 0x05f4))
469 if ((c >= 0x0621 && c <= 0x063a)
470 || (c >= 0x0640 && c <= 0x0652)
471 || (c >= 0x0670 && c <= 0x06b7)
472 || (c >= 0x06ba && c <= 0x06be)
473 || (c >= 0x06c0 && c <= 0x06ce)
474 || (c >= 0x06e5 && c <= 0x06e7))
478 if ((c >= 0x0905 && c <= 0x0939)
479 || (c >= 0x0958 && c <= 0x0962))
483 if ((c >= 0x0985 && c <= 0x098c)
484 || (c >= 0x098f && c <= 0x0990)
485 || (c >= 0x0993 && c <= 0x09a8)
486 || (c >= 0x09aa && c <= 0x09b0)
488 || (c >= 0x09b6 && c <= 0x09b9)
489 || (c >= 0x09dc && c <= 0x09dd)
490 || (c >= 0x09df && c <= 0x09e1)
491 || (c >= 0x09f0 && c <= 0x09f1))
495 if ((c >= 0x0a05 && c <= 0x0a0a)
496 || (c >= 0x0a0f && c <= 0x0a10)
497 || (c >= 0x0a13 && c <= 0x0a28)
498 || (c >= 0x0a2a && c <= 0x0a30)
499 || (c >= 0x0a32 && c <= 0x0a33)
500 || (c >= 0x0a35 && c <= 0x0a36)
501 || (c >= 0x0a38 && c <= 0x0a39)
502 || (c >= 0x0a59 && c <= 0x0a5c)
507 if ((c >= 0x0a85 && c <= 0x0a8b)
509 || (c >= 0x0a8f && c <= 0x0a91)
510 || (c >= 0x0a93 && c <= 0x0aa8)
511 || (c >= 0x0aaa && c <= 0x0ab0)
512 || (c >= 0x0ab2 && c <= 0x0ab3)
513 || (c >= 0x0ab5 && c <= 0x0ab9)
518 if ((c >= 0x0b05 && c <= 0x0b0c)
519 || (c >= 0x0b0f && c <= 0x0b10)
520 || (c >= 0x0b13 && c <= 0x0b28)
521 || (c >= 0x0b2a && c <= 0x0b30)
522 || (c >= 0x0b32 && c <= 0x0b33)
523 || (c >= 0x0b36 && c <= 0x0b39)
524 || (c >= 0x0b5c && c <= 0x0b5d)
525 || (c >= 0x0b5f && c <= 0x0b61))
529 if ((c >= 0x0b85 && c <= 0x0b8a)
530 || (c >= 0x0b8e && c <= 0x0b90)
531 || (c >= 0x0b92 && c <= 0x0b95)
532 || (c >= 0x0b99 && c <= 0x0b9a)
534 || (c >= 0x0b9e && c <= 0x0b9f)
535 || (c >= 0x0ba3 && c <= 0x0ba4)
536 || (c >= 0x0ba8 && c <= 0x0baa)
537 || (c >= 0x0bae && c <= 0x0bb5)
538 || (c >= 0x0bb7 && c <= 0x0bb9))
542 if ((c >= 0x0c05 && c <= 0x0c0c)
543 || (c >= 0x0c0e && c <= 0x0c10)
544 || (c >= 0x0c12 && c <= 0x0c28)
545 || (c >= 0x0c2a && c <= 0x0c33)
546 || (c >= 0x0c35 && c <= 0x0c39)
547 || (c >= 0x0c60 && c <= 0x0c61))
551 if ((c >= 0x0c85 && c <= 0x0c8c)
552 || (c >= 0x0c8e && c <= 0x0c90)
553 || (c >= 0x0c92 && c <= 0x0ca8)
554 || (c >= 0x0caa && c <= 0x0cb3)
555 || (c >= 0x0cb5 && c <= 0x0cb9)
556 || (c >= 0x0ce0 && c <= 0x0ce1))
560 if ((c >= 0x0d05 && c <= 0x0d0c)
561 || (c >= 0x0d0e && c <= 0x0d10)
562 || (c >= 0x0d12 && c <= 0x0d28)
563 || (c >= 0x0d2a && c <= 0x0d39)
564 || (c >= 0x0d60 && c <= 0x0d61))
568 if ((c >= 0x0e01 && c <= 0x0e30)
569 || (c >= 0x0e32 && c <= 0x0e33)
570 || (c >= 0x0e40 && c <= 0x0e46)
571 || (c >= 0x0e4f && c <= 0x0e5b))
575 if ((c >= 0x0e81 && c <= 0x0e82)
581 || (c >= 0x0e94 && c <= 0x0e97)
582 || (c >= 0x0e99 && c <= 0x0e9f)
583 || (c >= 0x0ea1 && c <= 0x0ea3)
588 || (c >= 0x0ead && c <= 0x0eb0)
592 || (c >= 0x0ec0 && c <= 0x0ec4)
597 if ((c >= 0x10a0 && c <= 0x10c5)
598 || (c >= 0x10d0 && c <= 0x10f6))
602 if ((c >= 0x3041 && c <= 0x3094)
603 || (c >= 0x309b && c <= 0x309e))
607 if ((c >= 0x30a1 && c <= 0x30fe))
611 if ((c >= 0x3105 && c <= 0x312c))
615 if ((c >= 0x1100 && c <= 0x1159)
616 || (c >= 0x1161 && c <= 0x11a2)
617 || (c >= 0x11a8 && c <= 0x11f9))
620 /* CJK Unified Ideographs */
621 if ((c >= 0xf900 && c <= 0xfa2d)
622 || (c >= 0xfb1f && c <= 0xfb36)
623 || (c >= 0xfb38 && c <= 0xfb3c)
625 || (c >= 0xfb40 && c <= 0xfb41)
626 || (c >= 0xfb42 && c <= 0xfb44)
627 || (c >= 0xfb46 && c <= 0xfbb1)
628 || (c >= 0xfbd3 && c <= 0xfd3f)
629 || (c >= 0xfd50 && c <= 0xfd8f)
630 || (c >= 0xfd92 && c <= 0xfdc7)
631 || (c >= 0xfdf0 && c <= 0xfdfb)
632 || (c >= 0xfe70 && c <= 0xfe72)
634 || (c >= 0xfe76 && c <= 0xfefc)
635 || (c >= 0xff21 && c <= 0xff3a)
636 || (c >= 0xff41 && c <= 0xff5a)
637 || (c >= 0xff66 && c <= 0xffbe)
638 || (c >= 0xffc2 && c <= 0xffc7)
639 || (c >= 0xffca && c <= 0xffcf)
640 || (c >= 0xffd2 && c <= 0xffd7)
641 || (c >= 0xffda && c <= 0xffdc)
642 || (c >= 0x4e00 && c <= 0x9fa5))
645 error ("universal-character-name '\\u%04x' not valid in identifier", c);
650 /* Add the UTF-8 representation of C to the token_buffer. */
653 utf8_extend_token (c)
663 else if (c <= 0x000007ff)
664 shift = 6, mask = 0xc0;
665 else if (c <= 0x0000ffff)
666 shift = 12, mask = 0xe0;
667 else if (c <= 0x001fffff)
668 shift = 18, mask = 0xf0;
669 else if (c <= 0x03ffffff)
670 shift = 24, mask = 0xf8;
672 shift = 30, mask = 0xfc;
674 extend_token (mask | (c >> shift));
678 extend_token ((unsigned char) (0x80 | (c >> shift)));
687 tree *const node_var;
688 const char unsigned_flag;
689 const char long_flag;
690 const char long_long_flag;
693 struct try_type type_sequence[] =
695 { &integer_type_node, 0, 0, 0},
696 { &unsigned_type_node, 1, 0, 0},
697 { &long_integer_type_node, 0, 1, 0},
698 { &long_unsigned_type_node, 1, 1, 0},
699 { &long_long_integer_type_node, 0, 1, 1},
700 { &long_long_unsigned_type_node, 1, 1, 1}
712 int conversion_errno;
713 REAL_VALUE_TYPE value;
721 struct pf_args * args = (struct pf_args *) data;
722 const char *typename;
724 args->conversion_errno = 0;
725 args->type = double_type_node;
728 /* The second argument, machine_mode, of REAL_VALUE_ATOF
729 tells the desired precision of the binary result
730 of decimal-to-binary conversion. */
735 error ("both 'f' and 'l' suffixes on floating constant");
737 args->type = float_type_node;
740 else if (args->lflag)
742 args->type = long_double_type_node;
743 typename = "long double";
745 else if (flag_single_precision_constant)
747 args->type = float_type_node;
752 if (args->base == 16)
753 args->value = REAL_VALUE_HTOF (args->str, TYPE_MODE (args->type));
755 args->value = REAL_VALUE_ATOF (args->str, TYPE_MODE (args->type));
757 args->conversion_errno = errno;
758 /* A diagnostic is required here by some ISO C testsuites.
759 This is not pedwarn, because some people don't want
760 an error for this. */
761 if (REAL_VALUE_ISINF (args->value) && pedantic)
762 warning ("floating point number exceeds range of '%s'", typename);
769 const cpp_token *tok;
772 timevar_push (TV_CPP);
774 tok = cpp_get_token (parse_in);
775 while (tok->type == CPP_PADDING);
776 timevar_pop (TV_CPP);
778 /* The C++ front end does horrible things with the current line
779 number. To ensure an accurate line number, we must reset it
780 every time we return a token. */
786 case CPP_OPEN_BRACE: indent_level++; break;
787 case CPP_CLOSE_BRACE: indent_level--; break;
789 /* Issue this error here, where we can get at tok->val.c. */
791 if (ISGRAPH (tok->val.c))
792 error ("stray '%c' in program", tok->val.c);
794 error ("stray '\\%o' in program", tok->val.c);
798 *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node));
802 *value = lex_number ((const char *)tok->val.str.text, tok->val.str.len);
807 *value = lex_charconst (tok);
812 *value = lex_string (tok->val.str.text, tok->val.str.len,
813 tok->type == CPP_WSTRING);
816 /* These tokens should not be visible outside cpplib. */
817 case CPP_HEADER_NAME:
828 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
831 lex_number (str, len)
837 int largest_digit = 0;
843 enum anon1 { NOT_FLOAT = 0, AFTER_POINT, AFTER_EXPON } floatflag = NOT_FLOAT;
845 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
846 The code below which fills the parts array assumes that a host
847 int is at least twice as wide as a host char, and that
848 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
849 Two HOST_WIDE_INTs is the largest int literal we can store.
850 In order to detect overflow below, the number of parts (TOTAL_PARTS)
851 must be exactly the number of parts needed to hold the bits
852 of two HOST_WIDE_INTs. */
853 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
854 unsigned int parts[TOTAL_PARTS];
856 /* Optimize for most frequent case. */
860 return integer_zero_node;
861 else if (*str == '1')
862 return integer_one_node;
864 return build_int_2 (*str - '0', 0);
867 for (count = 0; count < TOTAL_PARTS; count++)
870 /* len is known to be >1 at this point. */
873 if (len > 2 && str[0] == '0' && (str[1] == 'x' || str[1] == 'X'))
878 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
879 else if (str[0] == '0' && ISDIGIT (str[1]))
891 if (floatflag == AFTER_POINT)
892 ERROR ("too many decimal points in floating constant");
893 else if (floatflag == AFTER_EXPON)
894 ERROR ("decimal point in exponent - impossible!");
896 floatflag = AFTER_POINT;
902 /* Possible future extension: silently ignore _ in numbers,
903 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
904 but somewhat easier to read. Ada has this? */
905 ERROR ("underscore in number");
909 /* It is not a decimal point.
910 It should be a digit (perhaps a hex digit). */
913 || (base == 16 && ISXDIGIT (c)))
917 else if (base <= 10 && (c == 'e' || c == 'E'))
920 floatflag = AFTER_EXPON;
923 else if (base == 16 && (c == 'p' || c == 'P'))
925 floatflag = AFTER_EXPON;
926 break; /* start of exponent */
931 break; /* start of suffix */
934 if (n >= largest_digit)
938 for (count = 0; count < TOTAL_PARTS; count++)
940 parts[count] *= base;
944 += (parts[count-1] >> HOST_BITS_PER_CHAR);
946 &= (1 << HOST_BITS_PER_CHAR) - 1;
952 /* If the highest-order part overflows (gets larger than
953 a host char will hold) then the whole number has
954 overflowed. Record this and truncate the highest-order
956 if (parts[TOTAL_PARTS - 1] >> HOST_BITS_PER_CHAR)
959 parts[TOTAL_PARTS - 1] &= (1 << HOST_BITS_PER_CHAR) - 1;
963 while (p < str + len);
965 /* This can happen on input like `int i = 0x;' */
967 ERROR ("numeric constant with no digits");
969 if (largest_digit >= base)
970 ERROR ("numeric constant contains digits beyond the radix");
972 if (floatflag != NOT_FLOAT)
975 int imag, fflag, lflag, conversion_errno;
976 REAL_VALUE_TYPE real;
980 if (base == 16 && pedantic && !flag_isoc99)
981 pedwarn ("floating constant may not be in radix 16");
983 if (base == 16 && floatflag != AFTER_EXPON)
984 ERROR ("hexadecimal floating constant has no exponent");
986 /* Read explicit exponent if any, and put it in tokenbuf. */
987 if ((base == 10 && ((c == 'e') || (c == 'E')))
988 || (base == 16 && (c == 'p' || c == 'P')))
992 if (p < str + len && (c == '+' || c == '-'))
994 /* Exponent is decimal, even if string is a hex float. */
996 ERROR ("floating constant exponent has no digits");
997 while (p < str + len && ISDIGIT (c))
1003 /* Copy the float constant now; we don't want any suffixes in the
1004 string passed to parse_float. */
1005 copy = alloca (p - str + 1);
1006 memcpy (copy, str, p - str);
1007 copy[p - str] = '\0';
1009 /* Now parse suffixes. */
1010 fflag = lflag = imag = 0;
1011 while (p < str + len)
1016 ERROR ("more than one 'f' suffix on floating constant");
1017 else if (warn_traditional && !in_system_header
1018 && ! cpp_sys_macro_p (parse_in))
1019 warning ("traditional C rejects the 'f' suffix");
1026 ERROR ("more than one 'l' suffix on floating constant");
1027 else if (warn_traditional && !in_system_header
1028 && ! cpp_sys_macro_p (parse_in))
1029 warning ("traditional C rejects the 'l' suffix");
1037 ERROR ("more than one 'i' or 'j' suffix on floating constant");
1039 pedwarn ("ISO C forbids imaginary numeric constants");
1044 ERROR ("invalid suffix on floating constant");
1047 /* Setup input for parse_float() */
1053 /* Convert string to a double, checking for overflow. */
1054 if (do_float_handler (parse_float, (PTR) &args))
1056 /* Receive output from parse_float() */
1060 /* We got an exception from parse_float() */
1061 ERROR ("floating constant out of range");
1063 /* Receive output from parse_float() */
1064 conversion_errno = args.conversion_errno;
1068 /* ERANGE is also reported for underflow,
1069 so test the value to distinguish overflow from that. */
1070 if (conversion_errno == ERANGE && pedantic
1071 && (REAL_VALUES_LESS (dconst1, real)
1072 || REAL_VALUES_LESS (real, dconstm1)))
1073 warning ("floating point number exceeds range of 'double'");
1076 /* Create a node with determined type and value. */
1078 value = build_complex (NULL_TREE, convert (type, integer_zero_node),
1079 build_real (type, real));
1081 value = build_real (type, real);
1085 tree trad_type, type;
1086 HOST_WIDE_INT high, low;
1087 int spec_unsigned = 0;
1089 int spec_long_long = 0;
1094 trad_type = type = NULL_TREE;
1095 while (p < str + len)
1102 error ("two 'u' suffixes on integer constant");
1103 else if (warn_traditional && !in_system_header
1104 && ! cpp_sys_macro_p (parse_in))
1105 warning ("traditional C rejects the 'u' suffix");
1116 error ("three 'l' suffixes on integer constant");
1118 error ("'lul' is not a valid integer suffix");
1119 else if (c != spec_long)
1120 error ("'Ll' and 'lL' are not valid integer suffixes");
1121 else if (pedantic && ! flag_isoc99
1122 && ! in_system_header && warn_long_long)
1123 pedwarn ("ISO C89 forbids long long integer constants");
1129 case 'i': case 'I': case 'j': case 'J':
1131 error ("more than one 'i' or 'j' suffix on integer constant");
1133 pedwarn ("ISO C forbids imaginary numeric constants");
1138 ERROR ("invalid suffix on integer constant");
1142 /* If the literal overflowed, pedwarn about it now. */
1146 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT * 2);
1149 /* This is simplified by the fact that our constant
1150 is always positive. */
1154 for (i = 0; i < HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR; i++)
1156 high |= ((HOST_WIDE_INT) parts[i + (HOST_BITS_PER_WIDE_INT
1157 / HOST_BITS_PER_CHAR)]
1158 << (i * HOST_BITS_PER_CHAR));
1159 low |= (HOST_WIDE_INT) parts[i] << (i * HOST_BITS_PER_CHAR);
1162 value = build_int_2 (low, high);
1163 TREE_TYPE (value) = long_long_unsigned_type_node;
1165 /* If warn_traditional, calculate both the ISO type and the
1166 traditional type, then see if they disagree. */
1167 if (warn_traditional)
1169 /* Traditionally, any constant is signed; but if unsigned is
1170 specified explicitly, obey that. Use the smallest size
1171 with the right number of bits, except for one special
1172 case with decimal constants. */
1173 if (! spec_long && base != 10
1174 && int_fits_type_p (value, unsigned_type_node))
1175 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1176 /* A decimal constant must be long if it does not fit in
1177 type int. I think this is independent of whether the
1178 constant is signed. */
1179 else if (! spec_long && base == 10
1180 && int_fits_type_p (value, integer_type_node))
1181 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1182 else if (! spec_long_long)
1183 trad_type = (spec_unsigned
1184 ? long_unsigned_type_node
1185 : long_integer_type_node);
1186 else if (int_fits_type_p (value,
1188 ? long_long_unsigned_type_node
1189 : long_long_integer_type_node))
1190 trad_type = (spec_unsigned
1191 ? long_long_unsigned_type_node
1192 : long_long_integer_type_node);
1194 trad_type = (spec_unsigned
1195 ? widest_unsigned_literal_type_node
1196 : widest_integer_literal_type_node);
1199 /* Calculate the ISO type. */
1200 if (! spec_long && ! spec_unsigned
1201 && int_fits_type_p (value, integer_type_node))
1202 type = integer_type_node;
1203 else if (! spec_long && (base != 10 || spec_unsigned)
1204 && int_fits_type_p (value, unsigned_type_node))
1205 type = unsigned_type_node;
1206 else if (! spec_unsigned && !spec_long_long
1207 && int_fits_type_p (value, long_integer_type_node))
1208 type = long_integer_type_node;
1209 else if (! spec_long_long
1210 && int_fits_type_p (value, long_unsigned_type_node))
1211 type = long_unsigned_type_node;
1212 else if (! spec_unsigned
1213 && int_fits_type_p (value, long_long_integer_type_node))
1214 type = long_long_integer_type_node;
1215 else if (int_fits_type_p (value, long_long_unsigned_type_node))
1216 type = long_long_unsigned_type_node;
1217 else if (! spec_unsigned
1218 && int_fits_type_p (value, widest_integer_literal_type_node))
1219 type = widest_integer_literal_type_node;
1221 type = widest_unsigned_literal_type_node;
1223 /* We assume that constants specified in a non-decimal
1224 base are bit patterns, and that the programmer really
1225 meant what they wrote. */
1226 if (warn_traditional && !in_system_header
1227 && base == 10 && trad_type != type)
1229 if (TYPE_PRECISION (trad_type) != TYPE_PRECISION (type))
1230 warning ("width of integer constant is different in traditional C");
1231 else if (TREE_UNSIGNED (trad_type) != TREE_UNSIGNED (type))
1232 warning ("integer constant is unsigned in ISO C, signed in traditional C");
1234 warning ("width of integer constant may change on other systems in traditional C");
1237 if (pedantic && (flag_isoc99 || !spec_long_long)
1240 ? TYPE_PRECISION (long_long_integer_type_node)
1241 : TYPE_PRECISION (long_integer_type_node)) < TYPE_PRECISION (type)))
1244 pedwarn ("integer constant larger than the maximum value of %s",
1246 ? (TREE_UNSIGNED (type)
1247 ? _("an unsigned long long int")
1248 : _("a long long int"))
1249 : _("an unsigned long int")));
1252 if (base == 10 && ! spec_unsigned && TREE_UNSIGNED (type))
1253 warning ("decimal constant is so large that it is unsigned");
1257 if (TYPE_PRECISION (type)
1258 <= TYPE_PRECISION (integer_type_node))
1259 value = build_complex (NULL_TREE, integer_zero_node,
1260 convert (integer_type_node, value));
1262 ERROR ("complex integer constant is too wide for 'complex int'");
1265 TREE_TYPE (value) = type;
1267 /* If it's still an integer (not a complex), and it doesn't
1268 fit in the type we choose for it, then pedwarn. */
1271 && TREE_CODE (TREE_TYPE (value)) == INTEGER_TYPE
1272 && ! int_fits_type_p (value, TREE_TYPE (value)))
1273 pedwarn ("integer constant is larger than the maximum value for its type");
1277 error ("missing white space after number '%.*s'", (int) (p - str), str);
1282 return integer_zero_node;
1286 lex_string (str, len, wide)
1287 const unsigned char *str;
1292 char *buf = alloca ((len + 1) * (wide ? WCHAR_BYTES : 1));
1294 const unsigned char *p = str, *limit = str + len;
1296 unsigned width = wide ? WCHAR_TYPE_SIZE
1297 : TYPE_PRECISION (char_type_node);
1299 #ifdef MULTIBYTE_CHARS
1300 /* Reset multibyte conversion state. */
1301 (void) local_mbtowc (NULL, NULL, 0);
1306 #ifdef MULTIBYTE_CHARS
1310 char_len = local_mbtowc (&wc, (const char *) p, limit - p);
1313 warning ("ignoring invalid multibyte character");
1326 if (c == '\\' && !ignore_escape_flag)
1330 if (width < HOST_BITS_PER_INT)
1331 mask = ((unsigned int) 1 << width) - 1;
1334 c = cpp_parse_escape (parse_in, &p, limit, mask);
1337 /* Add this single character into the buffer either as a wchar_t
1338 or as a single byte. */
1341 unsigned charwidth = TYPE_PRECISION (char_type_node);
1342 unsigned bytemask = (1 << charwidth) - 1;
1345 for (byte = 0; byte < WCHAR_BYTES; ++byte)
1348 if (byte >= (int) sizeof (c))
1351 n = (c >> (byte * charwidth)) & bytemask;
1352 if (BYTES_BIG_ENDIAN)
1353 q[WCHAR_BYTES - byte - 1] = n;
1365 /* Terminate the string value, either with a single byte zero
1366 or with a wide zero. */
1370 memset (q, 0, WCHAR_BYTES);
1378 value = build_string (q - buf, buf);
1381 TREE_TYPE (value) = wchar_array_type_node;
1383 TREE_TYPE (value) = char_array_type_node;
1387 /* Converts a (possibly wide) character constant token into a tree. */
1389 lex_charconst (token)
1390 const cpp_token *token;
1392 HOST_WIDE_INT result;
1394 unsigned int chars_seen;
1396 result = cpp_interpret_charconst (parse_in, token, warn_multichar,
1398 if (token->type == CPP_WCHAR)
1400 value = build_int_2 (result, 0);
1401 TREE_TYPE (value) = wchar_type_node;
1406 value = build_int_2 (result, -1);
1408 value = build_int_2 (result, 0);
1410 /* In C, a character constant has type 'int'.
1411 In C++ 'char', but multi-char charconsts have type 'int'. */
1412 if (c_language == clk_cplusplus && chars_seen <= 1)
1413 TREE_TYPE (value) = char_type_node;
1415 TREE_TYPE (value) = integer_type_node;