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 input filename as understood by CPP, where "" represents stdin. */
58 static const char *cpp_filename;
60 /* The current line map. */
61 static const struct line_map *map;
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 char *, unsigned int, int));
89 static tree lex_charconst PARAMS ((const cpp_token *));
90 static void update_header_times PARAMS ((const char *));
91 static int dump_one_header PARAMS ((splay_tree_node, void *));
92 static void cb_ident PARAMS ((cpp_reader *, unsigned int,
94 static void cb_file_change PARAMS ((cpp_reader *, const struct line_map *));
95 static void cb_def_pragma PARAMS ((cpp_reader *, unsigned int));
96 static void cb_define PARAMS ((cpp_reader *, unsigned int,
98 static void cb_undef PARAMS ((cpp_reader *, unsigned int,
102 init_c_lex (filename)
103 const char *filename;
105 struct cpp_callbacks *cb;
106 struct c_fileinfo *toplevel;
108 /* Set up filename timing. Must happen before cpp_start_read. */
109 file_info_tree = splay_tree_new ((splay_tree_compare_fn)strcmp,
111 (splay_tree_delete_value_fn)free);
112 toplevel = get_fileinfo ("<top level>");
113 if (flag_detailed_statistics)
116 body_time = get_run_time ();
117 toplevel->time = body_time;
120 #ifdef MULTIBYTE_CHARS
121 /* Change to the native locale for multibyte conversions. */
122 setlocale (LC_CTYPE, "");
123 GET_ENVIRONMENT (literal_codeset, "LANG");
126 cb = cpp_get_callbacks (parse_in);
128 cb->ident = cb_ident;
129 cb->file_change = cb_file_change;
130 cb->def_pragma = cb_def_pragma;
132 /* Set the debug callbacks if we can use them. */
133 if (debug_info_level == DINFO_LEVEL_VERBOSE
134 && (write_symbols == DWARF_DEBUG || write_symbols == DWARF2_DEBUG))
136 cb->define = cb_define;
137 cb->undef = cb_undef;
141 if (filename == 0 || !strcmp (filename, "-"))
142 filename = "stdin", cpp_filename = "";
144 cpp_filename = filename;
152 /* A thin wrapper around the real parser that initializes the
153 integrated preprocessor after debug output has been initialized. */
158 if (! cpp_start_read (parse_in, cpp_filename))
159 return 1; /* cpplib has emitted an error. */
169 struct c_fileinfo *fi;
171 n = splay_tree_lookup (file_info_tree, (splay_tree_key) name);
173 return (struct c_fileinfo *) n->value;
175 fi = (struct c_fileinfo *) xmalloc (sizeof (struct c_fileinfo));
177 fi->interface_only = 0;
178 fi->interface_unknown = 1;
179 splay_tree_insert (file_info_tree, (splay_tree_key) name,
180 (splay_tree_value) fi);
185 update_header_times (name)
188 /* Changing files again. This means currently collected time
189 is charged against header time, and body time starts back at 0. */
190 if (flag_detailed_statistics)
192 int this_time = get_run_time ();
193 struct c_fileinfo *file = get_fileinfo (name);
194 header_time += this_time - body_time;
195 file->time += this_time - body_time;
196 body_time = this_time;
201 dump_one_header (n, dummy)
203 void *dummy ATTRIBUTE_UNUSED;
205 print_time ((const char *) n->key,
206 ((struct c_fileinfo *) n->value)->time);
211 dump_time_statistics ()
213 struct c_fileinfo *file = get_fileinfo (input_filename);
214 int this_time = get_run_time ();
215 file->time += this_time - body_time;
217 fprintf (stderr, "\n******\n");
218 print_time ("header files (total)", header_time);
219 print_time ("main file (total)", this_time - body_time);
220 fprintf (stderr, "ratio = %g : 1\n",
221 (double)header_time / (double)(this_time - body_time));
222 fprintf (stderr, "\n******\n");
224 splay_tree_foreach (file_info_tree, dump_one_header, 0);
227 /* Not yet handled: #pragma, #define, #undef.
228 No need to deal with linemarkers under normal conditions. */
231 cb_ident (pfile, line, str)
232 cpp_reader *pfile ATTRIBUTE_UNUSED;
233 unsigned int line ATTRIBUTE_UNUSED;
234 const cpp_string *str ATTRIBUTE_UNUSED;
236 #ifdef ASM_OUTPUT_IDENT
239 /* Convert escapes in the string. */
240 tree value = lex_string ((const char *)str->text, str->len, 0);
241 ASM_OUTPUT_IDENT (asm_out_file, TREE_STRING_POINTER (value));
247 cb_file_change (pfile, new_map)
248 cpp_reader *pfile ATTRIBUTE_UNUSED;
249 const struct line_map *new_map;
251 unsigned int to_line = SOURCE_LINE (new_map, new_map->to_line);
253 if (new_map->reason == LC_ENTER)
255 /* Don't stack the main buffer on the input stack;
256 we already did in compile_file. */
258 main_input_filename = new_map->to_file;
261 lineno = SOURCE_LINE (new_map - 1, new_map->from_line - 1);
262 push_srcloc (new_map->to_file, 1);
263 input_file_stack->indent_level = indent_level;
264 (*debug_hooks->start_source_file) (lineno, new_map->to_file);
265 #ifndef NO_IMPLICIT_EXTERN_C
268 else if (new_map->sysp == 2)
271 ++pending_lang_change;
276 else if (new_map->reason == LC_LEAVE)
278 #ifndef NO_IMPLICIT_EXTERN_C
279 if (c_header_level && --c_header_level == 0)
281 if (new_map->sysp == 2)
282 warning ("badly nested C headers from preprocessor");
283 --pending_lang_change;
287 if (indent_level != input_file_stack->indent_level)
289 warning_with_file_and_line
290 (input_filename, lineno,
291 "This file contains more '%c's than '%c's.",
292 indent_level > input_file_stack->indent_level ? '{' : '}',
293 indent_level > input_file_stack->indent_level ? '}' : '{');
298 (*debug_hooks->end_source_file) (to_line);
301 update_header_times (new_map->to_file);
302 in_system_header = new_map->sysp != 0;
303 input_filename = new_map->to_file;
308 extract_interface_info ();
312 cb_def_pragma (pfile, line)
316 /* Issue a warning message if we have been asked to do so. Ignore
317 unknown pragmas in system headers unless an explicit
318 -Wunknown-pragmas has been given. */
319 if (warn_unknown_pragmas > in_system_header)
321 const unsigned char *space, *name = 0;
324 cpp_get_token (pfile, &s);
325 space = cpp_token_as_text (pfile, &s);
326 cpp_get_token (pfile, &s);
327 if (s.type == CPP_NAME)
328 name = cpp_token_as_text (pfile, &s);
330 lineno = SOURCE_LINE (map, line);
332 warning ("ignoring #pragma %s %s", space, name);
334 warning ("ignoring #pragma %s", space);
338 /* #define callback for DWARF and DWARF2 debug info. */
340 cb_define (pfile, line, node)
345 (*debug_hooks->define) (SOURCE_LINE (map, line),
346 (const char *) cpp_macro_definition (pfile, node));
349 /* #undef callback for DWARF and DWARF2 debug info. */
351 cb_undef (pfile, line, node)
352 cpp_reader *pfile ATTRIBUTE_UNUSED;
356 (*debug_hooks->undef) (SOURCE_LINE (map, line),
357 (const char *) NODE_NAME (node));
361 /* Returns nonzero if C is a universal-character-name. Give an error if it
362 is not one which may appear in an identifier, as per [extendid].
364 Note that extended character support in identifiers has not yet been
365 implemented. It is my personal opinion that this is not a desirable
366 feature. Portable code cannot count on support for more than the basic
367 identifier character set. */
380 /* None of the valid chars are outside the Basic Multilingual Plane (the
384 error ("universal-character-name '\\U%08x' not valid in identifier", c);
389 if ((c >= 0x00c0 && c <= 0x00d6)
390 || (c >= 0x00d8 && c <= 0x00f6)
391 || (c >= 0x00f8 && c <= 0x01f5)
392 || (c >= 0x01fa && c <= 0x0217)
393 || (c >= 0x0250 && c <= 0x02a8)
394 || (c >= 0x1e00 && c <= 0x1e9a)
395 || (c >= 0x1ea0 && c <= 0x1ef9))
400 || (c >= 0x0388 && c <= 0x038a)
402 || (c >= 0x038e && c <= 0x03a1)
403 || (c >= 0x03a3 && c <= 0x03ce)
404 || (c >= 0x03d0 && c <= 0x03d6)
409 || (c >= 0x03e2 && c <= 0x03f3)
410 || (c >= 0x1f00 && c <= 0x1f15)
411 || (c >= 0x1f18 && c <= 0x1f1d)
412 || (c >= 0x1f20 && c <= 0x1f45)
413 || (c >= 0x1f48 && c <= 0x1f4d)
414 || (c >= 0x1f50 && c <= 0x1f57)
418 || (c >= 0x1f5f && c <= 0x1f7d)
419 || (c >= 0x1f80 && c <= 0x1fb4)
420 || (c >= 0x1fb6 && c <= 0x1fbc)
421 || (c >= 0x1fc2 && c <= 0x1fc4)
422 || (c >= 0x1fc6 && c <= 0x1fcc)
423 || (c >= 0x1fd0 && c <= 0x1fd3)
424 || (c >= 0x1fd6 && c <= 0x1fdb)
425 || (c >= 0x1fe0 && c <= 0x1fec)
426 || (c >= 0x1ff2 && c <= 0x1ff4)
427 || (c >= 0x1ff6 && c <= 0x1ffc))
431 if ((c >= 0x0401 && c <= 0x040d)
432 || (c >= 0x040f && c <= 0x044f)
433 || (c >= 0x0451 && c <= 0x045c)
434 || (c >= 0x045e && c <= 0x0481)
435 || (c >= 0x0490 && c <= 0x04c4)
436 || (c >= 0x04c7 && c <= 0x04c8)
437 || (c >= 0x04cb && c <= 0x04cc)
438 || (c >= 0x04d0 && c <= 0x04eb)
439 || (c >= 0x04ee && c <= 0x04f5)
440 || (c >= 0x04f8 && c <= 0x04f9))
444 if ((c >= 0x0531 && c <= 0x0556)
445 || (c >= 0x0561 && c <= 0x0587))
449 if ((c >= 0x05d0 && c <= 0x05ea)
450 || (c >= 0x05f0 && c <= 0x05f4))
454 if ((c >= 0x0621 && c <= 0x063a)
455 || (c >= 0x0640 && c <= 0x0652)
456 || (c >= 0x0670 && c <= 0x06b7)
457 || (c >= 0x06ba && c <= 0x06be)
458 || (c >= 0x06c0 && c <= 0x06ce)
459 || (c >= 0x06e5 && c <= 0x06e7))
463 if ((c >= 0x0905 && c <= 0x0939)
464 || (c >= 0x0958 && c <= 0x0962))
468 if ((c >= 0x0985 && c <= 0x098c)
469 || (c >= 0x098f && c <= 0x0990)
470 || (c >= 0x0993 && c <= 0x09a8)
471 || (c >= 0x09aa && c <= 0x09b0)
473 || (c >= 0x09b6 && c <= 0x09b9)
474 || (c >= 0x09dc && c <= 0x09dd)
475 || (c >= 0x09df && c <= 0x09e1)
476 || (c >= 0x09f0 && c <= 0x09f1))
480 if ((c >= 0x0a05 && c <= 0x0a0a)
481 || (c >= 0x0a0f && c <= 0x0a10)
482 || (c >= 0x0a13 && c <= 0x0a28)
483 || (c >= 0x0a2a && c <= 0x0a30)
484 || (c >= 0x0a32 && c <= 0x0a33)
485 || (c >= 0x0a35 && c <= 0x0a36)
486 || (c >= 0x0a38 && c <= 0x0a39)
487 || (c >= 0x0a59 && c <= 0x0a5c)
492 if ((c >= 0x0a85 && c <= 0x0a8b)
494 || (c >= 0x0a8f && c <= 0x0a91)
495 || (c >= 0x0a93 && c <= 0x0aa8)
496 || (c >= 0x0aaa && c <= 0x0ab0)
497 || (c >= 0x0ab2 && c <= 0x0ab3)
498 || (c >= 0x0ab5 && c <= 0x0ab9)
503 if ((c >= 0x0b05 && c <= 0x0b0c)
504 || (c >= 0x0b0f && c <= 0x0b10)
505 || (c >= 0x0b13 && c <= 0x0b28)
506 || (c >= 0x0b2a && c <= 0x0b30)
507 || (c >= 0x0b32 && c <= 0x0b33)
508 || (c >= 0x0b36 && c <= 0x0b39)
509 || (c >= 0x0b5c && c <= 0x0b5d)
510 || (c >= 0x0b5f && c <= 0x0b61))
514 if ((c >= 0x0b85 && c <= 0x0b8a)
515 || (c >= 0x0b8e && c <= 0x0b90)
516 || (c >= 0x0b92 && c <= 0x0b95)
517 || (c >= 0x0b99 && c <= 0x0b9a)
519 || (c >= 0x0b9e && c <= 0x0b9f)
520 || (c >= 0x0ba3 && c <= 0x0ba4)
521 || (c >= 0x0ba8 && c <= 0x0baa)
522 || (c >= 0x0bae && c <= 0x0bb5)
523 || (c >= 0x0bb7 && c <= 0x0bb9))
527 if ((c >= 0x0c05 && c <= 0x0c0c)
528 || (c >= 0x0c0e && c <= 0x0c10)
529 || (c >= 0x0c12 && c <= 0x0c28)
530 || (c >= 0x0c2a && c <= 0x0c33)
531 || (c >= 0x0c35 && c <= 0x0c39)
532 || (c >= 0x0c60 && c <= 0x0c61))
536 if ((c >= 0x0c85 && c <= 0x0c8c)
537 || (c >= 0x0c8e && c <= 0x0c90)
538 || (c >= 0x0c92 && c <= 0x0ca8)
539 || (c >= 0x0caa && c <= 0x0cb3)
540 || (c >= 0x0cb5 && c <= 0x0cb9)
541 || (c >= 0x0ce0 && c <= 0x0ce1))
545 if ((c >= 0x0d05 && c <= 0x0d0c)
546 || (c >= 0x0d0e && c <= 0x0d10)
547 || (c >= 0x0d12 && c <= 0x0d28)
548 || (c >= 0x0d2a && c <= 0x0d39)
549 || (c >= 0x0d60 && c <= 0x0d61))
553 if ((c >= 0x0e01 && c <= 0x0e30)
554 || (c >= 0x0e32 && c <= 0x0e33)
555 || (c >= 0x0e40 && c <= 0x0e46)
556 || (c >= 0x0e4f && c <= 0x0e5b))
560 if ((c >= 0x0e81 && c <= 0x0e82)
566 || (c >= 0x0e94 && c <= 0x0e97)
567 || (c >= 0x0e99 && c <= 0x0e9f)
568 || (c >= 0x0ea1 && c <= 0x0ea3)
573 || (c >= 0x0ead && c <= 0x0eb0)
577 || (c >= 0x0ec0 && c <= 0x0ec4)
582 if ((c >= 0x10a0 && c <= 0x10c5)
583 || (c >= 0x10d0 && c <= 0x10f6))
587 if ((c >= 0x3041 && c <= 0x3094)
588 || (c >= 0x309b && c <= 0x309e))
592 if ((c >= 0x30a1 && c <= 0x30fe))
596 if ((c >= 0x3105 && c <= 0x312c))
600 if ((c >= 0x1100 && c <= 0x1159)
601 || (c >= 0x1161 && c <= 0x11a2)
602 || (c >= 0x11a8 && c <= 0x11f9))
605 /* CJK Unified Ideographs */
606 if ((c >= 0xf900 && c <= 0xfa2d)
607 || (c >= 0xfb1f && c <= 0xfb36)
608 || (c >= 0xfb38 && c <= 0xfb3c)
610 || (c >= 0xfb40 && c <= 0xfb41)
611 || (c >= 0xfb42 && c <= 0xfb44)
612 || (c >= 0xfb46 && c <= 0xfbb1)
613 || (c >= 0xfbd3 && c <= 0xfd3f)
614 || (c >= 0xfd50 && c <= 0xfd8f)
615 || (c >= 0xfd92 && c <= 0xfdc7)
616 || (c >= 0xfdf0 && c <= 0xfdfb)
617 || (c >= 0xfe70 && c <= 0xfe72)
619 || (c >= 0xfe76 && c <= 0xfefc)
620 || (c >= 0xff21 && c <= 0xff3a)
621 || (c >= 0xff41 && c <= 0xff5a)
622 || (c >= 0xff66 && c <= 0xffbe)
623 || (c >= 0xffc2 && c <= 0xffc7)
624 || (c >= 0xffca && c <= 0xffcf)
625 || (c >= 0xffd2 && c <= 0xffd7)
626 || (c >= 0xffda && c <= 0xffdc)
627 || (c >= 0x4e00 && c <= 0x9fa5))
630 error ("universal-character-name '\\u%04x' not valid in identifier", c);
635 /* Add the UTF-8 representation of C to the token_buffer. */
638 utf8_extend_token (c)
648 else if (c <= 0x000007ff)
649 shift = 6, mask = 0xc0;
650 else if (c <= 0x0000ffff)
651 shift = 12, mask = 0xe0;
652 else if (c <= 0x001fffff)
653 shift = 18, mask = 0xf0;
654 else if (c <= 0x03ffffff)
655 shift = 24, mask = 0xf8;
657 shift = 30, mask = 0xfc;
659 extend_token (mask | (c >> shift));
663 extend_token ((unsigned char) (0x80 | (c >> shift)));
678 struct try_type type_sequence[] =
680 { &integer_type_node, 0, 0, 0},
681 { &unsigned_type_node, 1, 0, 0},
682 { &long_integer_type_node, 0, 1, 0},
683 { &long_unsigned_type_node, 1, 1, 0},
684 { &long_long_integer_type_node, 0, 1, 1},
685 { &long_long_unsigned_type_node, 1, 1, 1}
697 int conversion_errno;
698 REAL_VALUE_TYPE value;
706 struct pf_args * args = (struct pf_args *) data;
707 const char *typename;
709 args->conversion_errno = 0;
710 args->type = double_type_node;
713 /* The second argument, machine_mode, of REAL_VALUE_ATOF
714 tells the desired precision of the binary result
715 of decimal-to-binary conversion. */
720 error ("both 'f' and 'l' suffixes on floating constant");
722 args->type = float_type_node;
725 else if (args->lflag)
727 args->type = long_double_type_node;
728 typename = "long double";
730 else if (flag_single_precision_constant)
732 args->type = float_type_node;
737 if (args->base == 16)
738 args->value = REAL_VALUE_HTOF (args->str, TYPE_MODE (args->type));
740 args->value = REAL_VALUE_ATOF (args->str, TYPE_MODE (args->type));
742 args->conversion_errno = errno;
743 /* A diagnostic is required here by some ISO C testsuites.
744 This is not pedwarn, because some people don't want
745 an error for this. */
746 if (REAL_VALUE_ISINF (args->value) && pedantic)
747 warning ("floating point number exceeds range of '%s'", typename);
758 timevar_push (TV_CPP);
759 cpp_get_token (parse_in, &tok);
760 timevar_pop (TV_CPP);
762 /* The C++ front end does horrible things with the current line
763 number. To ensure an accurate line number, we must reset it
764 every time we return a token. */
765 lineno = SOURCE_LINE (map, cpp_get_line (parse_in)->line);
771 case CPP_OPEN_BRACE: indent_level++; break;
772 case CPP_CLOSE_BRACE: indent_level--; break;
774 /* Issue this error here, where we can get at tok.val.c. */
776 if (ISGRAPH (tok.val.c))
777 error ("stray '%c' in program", tok.val.c);
779 error ("stray '\\%o' in program", tok.val.c);
783 *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok.val.node));
787 *value = lex_number ((const char *)tok.val.str.text, tok.val.str.len);
792 *value = lex_charconst (&tok);
797 *value = lex_string ((const char *)tok.val.str.text,
798 tok.val.str.len, tok.type == CPP_WSTRING);
801 /* These tokens should not be visible outside cpplib. */
802 case CPP_HEADER_NAME:
813 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
816 lex_number (str, len)
822 int largest_digit = 0;
828 enum anon1 { NOT_FLOAT = 0, AFTER_POINT, AFTER_EXPON } floatflag = NOT_FLOAT;
830 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
831 The code below which fills the parts array assumes that a host
832 int is at least twice as wide as a host char, and that
833 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
834 Two HOST_WIDE_INTs is the largest int literal we can store.
835 In order to detect overflow below, the number of parts (TOTAL_PARTS)
836 must be exactly the number of parts needed to hold the bits
837 of two HOST_WIDE_INTs. */
838 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
839 unsigned int parts[TOTAL_PARTS];
841 /* Optimize for most frequent case. */
845 return integer_zero_node;
846 else if (*str == '1')
847 return integer_one_node;
849 return build_int_2 (*str - '0', 0);
852 for (count = 0; count < TOTAL_PARTS; count++)
855 /* len is known to be >1 at this point. */
858 if (len > 2 && str[0] == '0' && (str[1] == 'x' || str[1] == 'X'))
863 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
864 else if (str[0] == '0' && ISDIGIT (str[1]))
876 if (floatflag == AFTER_POINT)
877 ERROR ("too many decimal points in floating constant");
878 else if (floatflag == AFTER_EXPON)
879 ERROR ("decimal point in exponent - impossible!");
881 floatflag = AFTER_POINT;
887 /* Possible future extension: silently ignore _ in numbers,
888 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
889 but somewhat easier to read. Ada has this? */
890 ERROR ("underscore in number");
894 /* It is not a decimal point.
895 It should be a digit (perhaps a hex digit). */
901 else if (base <= 10 && (c == 'e' || c == 'E'))
904 floatflag = AFTER_EXPON;
907 else if (base == 16 && (c == 'p' || c == 'P'))
909 floatflag = AFTER_EXPON;
910 break; /* start of exponent */
912 else if (base == 16 && c >= 'a' && c <= 'f')
916 else if (base == 16 && c >= 'A' && c <= 'F')
923 break; /* start of suffix */
926 if (n >= largest_digit)
930 for (count = 0; count < TOTAL_PARTS; count++)
932 parts[count] *= base;
936 += (parts[count-1] >> HOST_BITS_PER_CHAR);
938 &= (1 << HOST_BITS_PER_CHAR) - 1;
944 /* If the highest-order part overflows (gets larger than
945 a host char will hold) then the whole number has
946 overflowed. Record this and truncate the highest-order
948 if (parts[TOTAL_PARTS - 1] >> HOST_BITS_PER_CHAR)
951 parts[TOTAL_PARTS - 1] &= (1 << HOST_BITS_PER_CHAR) - 1;
955 while (p < str + len);
957 /* This can happen on input like `int i = 0x;' */
959 ERROR ("numeric constant with no digits");
961 if (largest_digit >= base)
962 ERROR ("numeric constant contains digits beyond the radix");
964 if (floatflag != NOT_FLOAT)
967 int imag, fflag, lflag, conversion_errno;
968 REAL_VALUE_TYPE real;
972 if (base == 16 && pedantic && !flag_isoc99)
973 pedwarn ("floating constant may not be in radix 16");
975 if (base == 16 && floatflag != AFTER_EXPON)
976 ERROR ("hexadecimal floating constant has no exponent");
978 /* Read explicit exponent if any, and put it in tokenbuf. */
979 if ((base == 10 && ((c == 'e') || (c == 'E')))
980 || (base == 16 && (c == 'p' || c == 'P')))
984 if (p < str + len && (c == '+' || c == '-'))
986 /* Exponent is decimal, even if string is a hex float. */
988 ERROR ("floating constant exponent has no digits");
989 while (p < str + len && ISDIGIT (c))
995 /* Copy the float constant now; we don't want any suffixes in the
996 string passed to parse_float. */
997 copy = alloca (p - str + 1);
998 memcpy (copy, str, p - str);
999 copy[p - str] = '\0';
1001 /* Now parse suffixes. */
1002 fflag = lflag = imag = 0;
1003 while (p < str + len)
1008 ERROR ("more than one 'f' suffix on floating constant");
1009 else if (warn_traditional && !in_system_header
1010 && ! cpp_sys_macro_p (parse_in))
1011 warning ("traditional C rejects the 'f' suffix");
1018 ERROR ("more than one 'l' 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 'l' suffix");
1029 ERROR ("more than one 'i' or 'j' suffix on floating constant");
1031 pedwarn ("ISO C forbids imaginary numeric constants");
1036 ERROR ("invalid suffix on floating constant");
1039 /* Setup input for parse_float() */
1045 /* Convert string to a double, checking for overflow. */
1046 if (do_float_handler (parse_float, (PTR) &args))
1048 /* Receive output from parse_float() */
1052 /* We got an exception from parse_float() */
1053 ERROR ("floating constant out of range");
1055 /* Receive output from parse_float() */
1056 conversion_errno = args.conversion_errno;
1060 /* ERANGE is also reported for underflow,
1061 so test the value to distinguish overflow from that. */
1062 if (conversion_errno == ERANGE && !flag_traditional && pedantic
1063 && (REAL_VALUES_LESS (dconst1, real)
1064 || REAL_VALUES_LESS (real, dconstm1)))
1065 warning ("floating point number exceeds range of 'double'");
1068 /* Create a node with determined type and value. */
1070 value = build_complex (NULL_TREE, convert (type, integer_zero_node),
1071 build_real (type, real));
1073 value = build_real (type, real);
1077 tree trad_type, ansi_type, type;
1078 HOST_WIDE_INT high, low;
1079 int spec_unsigned = 0;
1081 int spec_long_long = 0;
1086 trad_type = ansi_type = type = NULL_TREE;
1087 while (p < str + len)
1094 error ("two 'u' suffixes on integer constant");
1095 else if (warn_traditional && !in_system_header
1096 && ! cpp_sys_macro_p (parse_in))
1097 warning ("traditional C rejects the 'u' suffix");
1108 error ("three 'l' suffixes on integer constant");
1110 error ("'lul' is not a valid integer suffix");
1111 else if (c != spec_long)
1112 error ("'Ll' and 'lL' are not valid integer suffixes");
1113 else if (pedantic && ! flag_isoc99
1114 && ! in_system_header && warn_long_long)
1115 pedwarn ("ISO C89 forbids long long integer constants");
1121 case 'i': case 'I': case 'j': case 'J':
1123 error ("more than one 'i' or 'j' suffix on integer constant");
1125 pedwarn ("ISO C forbids imaginary numeric constants");
1130 ERROR ("invalid suffix on integer constant");
1134 /* If the literal overflowed, pedwarn about it now. */
1138 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT * 2);
1141 /* This is simplified by the fact that our constant
1142 is always positive. */
1146 for (i = 0; i < HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR; i++)
1148 high |= ((HOST_WIDE_INT) parts[i + (HOST_BITS_PER_WIDE_INT
1149 / HOST_BITS_PER_CHAR)]
1150 << (i * HOST_BITS_PER_CHAR));
1151 low |= (HOST_WIDE_INT) parts[i] << (i * HOST_BITS_PER_CHAR);
1154 value = build_int_2 (low, high);
1155 TREE_TYPE (value) = long_long_unsigned_type_node;
1157 /* If warn_traditional, calculate both the ISO type and the
1158 traditional type, then see if they disagree.
1159 Otherwise, calculate only the type for the dialect in use. */
1160 if (warn_traditional || flag_traditional)
1162 /* Calculate the traditional type. */
1163 /* Traditionally, any constant is signed; but if unsigned is
1164 specified explicitly, obey that. Use the smallest size
1165 with the right number of bits, except for one special
1166 case with decimal constants. */
1167 if (! spec_long && base != 10
1168 && int_fits_type_p (value, unsigned_type_node))
1169 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1170 /* A decimal constant must be long if it does not fit in
1171 type int. I think this is independent of whether the
1172 constant is signed. */
1173 else if (! spec_long && base == 10
1174 && int_fits_type_p (value, integer_type_node))
1175 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1176 else if (! spec_long_long)
1177 trad_type = (spec_unsigned
1178 ? long_unsigned_type_node
1179 : long_integer_type_node);
1180 else if (int_fits_type_p (value,
1182 ? long_long_unsigned_type_node
1183 : long_long_integer_type_node))
1184 trad_type = (spec_unsigned
1185 ? long_long_unsigned_type_node
1186 : long_long_integer_type_node);
1188 trad_type = (spec_unsigned
1189 ? widest_unsigned_literal_type_node
1190 : widest_integer_literal_type_node);
1192 if (warn_traditional || ! flag_traditional)
1194 /* Calculate the ISO type. */
1195 if (! spec_long && ! spec_unsigned
1196 && int_fits_type_p (value, integer_type_node))
1197 ansi_type = integer_type_node;
1198 else if (! spec_long && (base != 10 || spec_unsigned)
1199 && int_fits_type_p (value, unsigned_type_node))
1200 ansi_type = unsigned_type_node;
1201 else if (! spec_unsigned && !spec_long_long
1202 && int_fits_type_p (value, long_integer_type_node))
1203 ansi_type = long_integer_type_node;
1204 else if (! spec_long_long
1205 && int_fits_type_p (value, long_unsigned_type_node))
1206 ansi_type = long_unsigned_type_node;
1207 else if (! spec_unsigned
1208 && int_fits_type_p (value, long_long_integer_type_node))
1209 ansi_type = long_long_integer_type_node;
1210 else if (int_fits_type_p (value, long_long_unsigned_type_node))
1211 ansi_type = long_long_unsigned_type_node;
1212 else if (! spec_unsigned
1213 && int_fits_type_p (value, widest_integer_literal_type_node))
1214 ansi_type = widest_integer_literal_type_node;
1216 ansi_type = widest_unsigned_literal_type_node;
1219 type = flag_traditional ? trad_type : ansi_type;
1221 /* We assume that constants specified in a non-decimal
1222 base are bit patterns, and that the programmer really
1223 meant what they wrote. */
1224 if (warn_traditional && !in_system_header
1225 && base == 10 && trad_type != ansi_type)
1227 if (TYPE_PRECISION (trad_type) != TYPE_PRECISION (ansi_type))
1228 warning ("width of integer constant changes with -traditional");
1229 else if (TREE_UNSIGNED (trad_type) != TREE_UNSIGNED (ansi_type))
1230 warning ("integer constant is unsigned in ISO C, signed with -traditional");
1232 warning ("width of integer constant may change on other systems with -traditional");
1235 if (pedantic && !flag_traditional && (flag_isoc99 || !spec_long_long)
1238 ? TYPE_PRECISION (long_long_integer_type_node)
1239 : TYPE_PRECISION (long_integer_type_node)) < TYPE_PRECISION (type)))
1242 pedwarn ("integer constant larger than the maximum value of %s",
1244 ? (TREE_UNSIGNED (type)
1245 ? "an unsigned long long int"
1246 : "a long long int")
1247 : "an unsigned long int"));
1250 if (base == 10 && ! spec_unsigned && TREE_UNSIGNED (type))
1251 warning ("decimal constant is so large that it is unsigned");
1255 if (TYPE_PRECISION (type)
1256 <= TYPE_PRECISION (integer_type_node))
1257 value = build_complex (NULL_TREE, integer_zero_node,
1258 convert (integer_type_node, value));
1260 ERROR ("complex integer constant is too wide for 'complex int'");
1262 else if (flag_traditional && !int_fits_type_p (value, type))
1263 /* The traditional constant 0x80000000 is signed
1264 but doesn't fit in the range of int.
1265 This will change it to -0x80000000, which does fit. */
1267 TREE_TYPE (value) = unsigned_type (type);
1268 value = convert (type, value);
1269 TREE_OVERFLOW (value) = TREE_CONSTANT_OVERFLOW (value) = 0;
1272 TREE_TYPE (value) = type;
1274 /* If it's still an integer (not a complex), and it doesn't
1275 fit in the type we choose for it, then pedwarn. */
1278 && TREE_CODE (TREE_TYPE (value)) == INTEGER_TYPE
1279 && ! int_fits_type_p (value, TREE_TYPE (value)))
1280 pedwarn ("integer constant is larger than the maximum value for its type");
1284 error ("missing white space after number '%.*s'", (int) (p - str), str);
1289 return integer_zero_node;
1293 lex_string (str, len, wide)
1299 char *buf = alloca ((len + 1) * (wide ? WCHAR_BYTES : 1));
1301 const char *p = str, *limit = str + len;
1303 unsigned width = wide ? WCHAR_TYPE_SIZE
1304 : TYPE_PRECISION (char_type_node);
1306 #ifdef MULTIBYTE_CHARS
1307 /* Reset multibyte conversion state. */
1308 (void) local_mbtowc (NULL, NULL, 0);
1313 #ifdef MULTIBYTE_CHARS
1317 char_len = local_mbtowc (&wc, p, limit - p);
1320 warning ("Ignoring invalid multibyte character");
1333 if (c == '\\' && !ignore_escape_flag)
1337 if (width < HOST_BITS_PER_INT)
1338 mask = ((unsigned int) 1 << width) - 1;
1341 c = cpp_parse_escape (parse_in, (const unsigned char **) &p,
1342 (const unsigned char *) limit,
1343 mask, flag_traditional);
1346 /* Add this single character into the buffer either as a wchar_t
1347 or as a single byte. */
1350 unsigned charwidth = TYPE_PRECISION (char_type_node);
1351 unsigned bytemask = (1 << charwidth) - 1;
1354 for (byte = 0; byte < WCHAR_BYTES; ++byte)
1357 if (byte >= (int) sizeof (c))
1360 n = (c >> (byte * charwidth)) & bytemask;
1361 if (BYTES_BIG_ENDIAN)
1362 q[WCHAR_BYTES - byte - 1] = n;
1374 /* Terminate the string value, either with a single byte zero
1375 or with a wide zero. */
1379 memset (q, 0, WCHAR_BYTES);
1387 value = build_string (q - buf, buf);
1390 TREE_TYPE (value) = wchar_array_type_node;
1392 TREE_TYPE (value) = char_array_type_node;
1396 /* Converts a (possibly wide) character constant token into a tree. */
1398 lex_charconst (token)
1399 const cpp_token *token;
1401 HOST_WIDE_INT result;
1403 unsigned int chars_seen;
1405 result = cpp_interpret_charconst (parse_in, token, warn_multichar,
1406 flag_traditional, &chars_seen);
1407 if (token->type == CPP_WCHAR)
1409 value = build_int_2 (result, 0);
1410 TREE_TYPE (value) = wchar_type_node;
1415 value = build_int_2 (result, -1);
1417 value = build_int_2 (result, 0);
1419 /* In C, a character constant has type 'int'.
1420 In C++ 'char', but multi-char charconsts have type 'int'. */
1421 if (c_language == clk_cplusplus && chars_seen <= 1)
1422 TREE_TYPE (value) = char_type_node;
1424 TREE_TYPE (value) = integer_type_node;