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 GNU CC.
7 GNU CC is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 2, or (at your option)
12 GNU CC is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with GNU CC; see the file COPYING. If not, write to
19 the Free Software Foundation, 59 Temple Place - Suite 330,
20 Boston, MA 02111-1307, USA. */
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 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 cpp_file_change *));
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, fc)
248 cpp_reader *pfile ATTRIBUTE_UNUSED;
249 const cpp_file_change *fc;
251 unsigned int from_line = SOURCE_LINE (fc->map - 1, fc->line - 1);
253 if (fc->reason == LC_ENTER)
255 /* Don't stack the main buffer on the input stack;
256 we already did in compile_file. */
257 if (MAIN_FILE_P (fc->map))
258 main_input_filename = fc->map->to_file;
262 push_srcloc (fc->map->to_file, 1);
263 input_file_stack->indent_level = indent_level;
264 (*debug_hooks->start_source_file) (lineno, fc->map->to_file);
265 #ifndef NO_IMPLICIT_EXTERN_C
268 else if (fc->externc)
271 ++pending_lang_change;
276 else if (fc->reason == LC_LEAVE)
278 /* Popping out of a file. */
279 if (input_file_stack->next)
281 #ifndef NO_IMPLICIT_EXTERN_C
282 if (c_header_level && --c_header_level == 0)
285 warning ("badly nested C headers from preprocessor");
286 --pending_lang_change;
290 if (indent_level != input_file_stack->indent_level)
292 warning_with_file_and_line
293 (input_filename, lineno,
294 "This file contains more '%c's than '%c's.",
295 indent_level > input_file_stack->indent_level ? '{' : '}',
296 indent_level > input_file_stack->indent_level ? '}' : '{');
300 (*debug_hooks->end_source_file) (from_line);
303 error ("leaving more files than we entered");
306 update_header_times (fc->map->to_file);
308 in_system_header = fc->sysp != 0;
309 input_filename = map->to_file;
310 lineno = SOURCE_LINE (map, fc->line);
313 extract_interface_info ();
317 cb_def_pragma (pfile, line)
321 /* Issue a warning message if we have been asked to do so. Ignore
322 unknown pragmas in system headers unless an explicit
323 -Wunknown-pragmas has been given. */
324 if (warn_unknown_pragmas > in_system_header)
326 const unsigned char *space, *name = 0;
329 cpp_get_token (pfile, &s);
330 space = cpp_token_as_text (pfile, &s);
331 cpp_get_token (pfile, &s);
332 if (s.type == CPP_NAME)
333 name = cpp_token_as_text (pfile, &s);
335 lineno = SOURCE_LINE (map, line);
337 warning ("ignoring #pragma %s %s", space, name);
339 warning ("ignoring #pragma %s", space);
343 /* #define callback for DWARF and DWARF2 debug info. */
345 cb_define (pfile, line, node)
350 (*debug_hooks->define) (SOURCE_LINE (map, line),
351 (const char *) cpp_macro_definition (pfile, node));
354 /* #undef callback for DWARF and DWARF2 debug info. */
356 cb_undef (pfile, line, node)
357 cpp_reader *pfile ATTRIBUTE_UNUSED;
361 (*debug_hooks->undef) (SOURCE_LINE (map, line),
362 (const char *) NODE_NAME (node));
366 /* Returns nonzero if C is a universal-character-name. Give an error if it
367 is not one which may appear in an identifier, as per [extendid].
369 Note that extended character support in identifiers has not yet been
370 implemented. It is my personal opinion that this is not a desirable
371 feature. Portable code cannot count on support for more than the basic
372 identifier character set. */
385 /* None of the valid chars are outside the Basic Multilingual Plane (the
389 error ("universal-character-name '\\U%08x' not valid in identifier", c);
394 if ((c >= 0x00c0 && c <= 0x00d6)
395 || (c >= 0x00d8 && c <= 0x00f6)
396 || (c >= 0x00f8 && c <= 0x01f5)
397 || (c >= 0x01fa && c <= 0x0217)
398 || (c >= 0x0250 && c <= 0x02a8)
399 || (c >= 0x1e00 && c <= 0x1e9a)
400 || (c >= 0x1ea0 && c <= 0x1ef9))
405 || (c >= 0x0388 && c <= 0x038a)
407 || (c >= 0x038e && c <= 0x03a1)
408 || (c >= 0x03a3 && c <= 0x03ce)
409 || (c >= 0x03d0 && c <= 0x03d6)
414 || (c >= 0x03e2 && c <= 0x03f3)
415 || (c >= 0x1f00 && c <= 0x1f15)
416 || (c >= 0x1f18 && c <= 0x1f1d)
417 || (c >= 0x1f20 && c <= 0x1f45)
418 || (c >= 0x1f48 && c <= 0x1f4d)
419 || (c >= 0x1f50 && c <= 0x1f57)
423 || (c >= 0x1f5f && c <= 0x1f7d)
424 || (c >= 0x1f80 && c <= 0x1fb4)
425 || (c >= 0x1fb6 && c <= 0x1fbc)
426 || (c >= 0x1fc2 && c <= 0x1fc4)
427 || (c >= 0x1fc6 && c <= 0x1fcc)
428 || (c >= 0x1fd0 && c <= 0x1fd3)
429 || (c >= 0x1fd6 && c <= 0x1fdb)
430 || (c >= 0x1fe0 && c <= 0x1fec)
431 || (c >= 0x1ff2 && c <= 0x1ff4)
432 || (c >= 0x1ff6 && c <= 0x1ffc))
436 if ((c >= 0x0401 && c <= 0x040d)
437 || (c >= 0x040f && c <= 0x044f)
438 || (c >= 0x0451 && c <= 0x045c)
439 || (c >= 0x045e && c <= 0x0481)
440 || (c >= 0x0490 && c <= 0x04c4)
441 || (c >= 0x04c7 && c <= 0x04c8)
442 || (c >= 0x04cb && c <= 0x04cc)
443 || (c >= 0x04d0 && c <= 0x04eb)
444 || (c >= 0x04ee && c <= 0x04f5)
445 || (c >= 0x04f8 && c <= 0x04f9))
449 if ((c >= 0x0531 && c <= 0x0556)
450 || (c >= 0x0561 && c <= 0x0587))
454 if ((c >= 0x05d0 && c <= 0x05ea)
455 || (c >= 0x05f0 && c <= 0x05f4))
459 if ((c >= 0x0621 && c <= 0x063a)
460 || (c >= 0x0640 && c <= 0x0652)
461 || (c >= 0x0670 && c <= 0x06b7)
462 || (c >= 0x06ba && c <= 0x06be)
463 || (c >= 0x06c0 && c <= 0x06ce)
464 || (c >= 0x06e5 && c <= 0x06e7))
468 if ((c >= 0x0905 && c <= 0x0939)
469 || (c >= 0x0958 && c <= 0x0962))
473 if ((c >= 0x0985 && c <= 0x098c)
474 || (c >= 0x098f && c <= 0x0990)
475 || (c >= 0x0993 && c <= 0x09a8)
476 || (c >= 0x09aa && c <= 0x09b0)
478 || (c >= 0x09b6 && c <= 0x09b9)
479 || (c >= 0x09dc && c <= 0x09dd)
480 || (c >= 0x09df && c <= 0x09e1)
481 || (c >= 0x09f0 && c <= 0x09f1))
485 if ((c >= 0x0a05 && c <= 0x0a0a)
486 || (c >= 0x0a0f && c <= 0x0a10)
487 || (c >= 0x0a13 && c <= 0x0a28)
488 || (c >= 0x0a2a && c <= 0x0a30)
489 || (c >= 0x0a32 && c <= 0x0a33)
490 || (c >= 0x0a35 && c <= 0x0a36)
491 || (c >= 0x0a38 && c <= 0x0a39)
492 || (c >= 0x0a59 && c <= 0x0a5c)
497 if ((c >= 0x0a85 && c <= 0x0a8b)
499 || (c >= 0x0a8f && c <= 0x0a91)
500 || (c >= 0x0a93 && c <= 0x0aa8)
501 || (c >= 0x0aaa && c <= 0x0ab0)
502 || (c >= 0x0ab2 && c <= 0x0ab3)
503 || (c >= 0x0ab5 && c <= 0x0ab9)
508 if ((c >= 0x0b05 && c <= 0x0b0c)
509 || (c >= 0x0b0f && c <= 0x0b10)
510 || (c >= 0x0b13 && c <= 0x0b28)
511 || (c >= 0x0b2a && c <= 0x0b30)
512 || (c >= 0x0b32 && c <= 0x0b33)
513 || (c >= 0x0b36 && c <= 0x0b39)
514 || (c >= 0x0b5c && c <= 0x0b5d)
515 || (c >= 0x0b5f && c <= 0x0b61))
519 if ((c >= 0x0b85 && c <= 0x0b8a)
520 || (c >= 0x0b8e && c <= 0x0b90)
521 || (c >= 0x0b92 && c <= 0x0b95)
522 || (c >= 0x0b99 && c <= 0x0b9a)
524 || (c >= 0x0b9e && c <= 0x0b9f)
525 || (c >= 0x0ba3 && c <= 0x0ba4)
526 || (c >= 0x0ba8 && c <= 0x0baa)
527 || (c >= 0x0bae && c <= 0x0bb5)
528 || (c >= 0x0bb7 && c <= 0x0bb9))
532 if ((c >= 0x0c05 && c <= 0x0c0c)
533 || (c >= 0x0c0e && c <= 0x0c10)
534 || (c >= 0x0c12 && c <= 0x0c28)
535 || (c >= 0x0c2a && c <= 0x0c33)
536 || (c >= 0x0c35 && c <= 0x0c39)
537 || (c >= 0x0c60 && c <= 0x0c61))
541 if ((c >= 0x0c85 && c <= 0x0c8c)
542 || (c >= 0x0c8e && c <= 0x0c90)
543 || (c >= 0x0c92 && c <= 0x0ca8)
544 || (c >= 0x0caa && c <= 0x0cb3)
545 || (c >= 0x0cb5 && c <= 0x0cb9)
546 || (c >= 0x0ce0 && c <= 0x0ce1))
550 if ((c >= 0x0d05 && c <= 0x0d0c)
551 || (c >= 0x0d0e && c <= 0x0d10)
552 || (c >= 0x0d12 && c <= 0x0d28)
553 || (c >= 0x0d2a && c <= 0x0d39)
554 || (c >= 0x0d60 && c <= 0x0d61))
558 if ((c >= 0x0e01 && c <= 0x0e30)
559 || (c >= 0x0e32 && c <= 0x0e33)
560 || (c >= 0x0e40 && c <= 0x0e46)
561 || (c >= 0x0e4f && c <= 0x0e5b))
565 if ((c >= 0x0e81 && c <= 0x0e82)
571 || (c >= 0x0e94 && c <= 0x0e97)
572 || (c >= 0x0e99 && c <= 0x0e9f)
573 || (c >= 0x0ea1 && c <= 0x0ea3)
578 || (c >= 0x0ead && c <= 0x0eb0)
582 || (c >= 0x0ec0 && c <= 0x0ec4)
587 if ((c >= 0x10a0 && c <= 0x10c5)
588 || (c >= 0x10d0 && c <= 0x10f6))
592 if ((c >= 0x3041 && c <= 0x3094)
593 || (c >= 0x309b && c <= 0x309e))
597 if ((c >= 0x30a1 && c <= 0x30fe))
601 if ((c >= 0x3105 && c <= 0x312c))
605 if ((c >= 0x1100 && c <= 0x1159)
606 || (c >= 0x1161 && c <= 0x11a2)
607 || (c >= 0x11a8 && c <= 0x11f9))
610 /* CJK Unified Ideographs */
611 if ((c >= 0xf900 && c <= 0xfa2d)
612 || (c >= 0xfb1f && c <= 0xfb36)
613 || (c >= 0xfb38 && c <= 0xfb3c)
615 || (c >= 0xfb40 && c <= 0xfb41)
616 || (c >= 0xfb42 && c <= 0xfb44)
617 || (c >= 0xfb46 && c <= 0xfbb1)
618 || (c >= 0xfbd3 && c <= 0xfd3f)
619 || (c >= 0xfd50 && c <= 0xfd8f)
620 || (c >= 0xfd92 && c <= 0xfdc7)
621 || (c >= 0xfdf0 && c <= 0xfdfb)
622 || (c >= 0xfe70 && c <= 0xfe72)
624 || (c >= 0xfe76 && c <= 0xfefc)
625 || (c >= 0xff21 && c <= 0xff3a)
626 || (c >= 0xff41 && c <= 0xff5a)
627 || (c >= 0xff66 && c <= 0xffbe)
628 || (c >= 0xffc2 && c <= 0xffc7)
629 || (c >= 0xffca && c <= 0xffcf)
630 || (c >= 0xffd2 && c <= 0xffd7)
631 || (c >= 0xffda && c <= 0xffdc)
632 || (c >= 0x4e00 && c <= 0x9fa5))
635 error ("universal-character-name '\\u%04x' not valid in identifier", c);
640 /* Add the UTF-8 representation of C to the token_buffer. */
643 utf8_extend_token (c)
653 else if (c <= 0x000007ff)
654 shift = 6, mask = 0xc0;
655 else if (c <= 0x0000ffff)
656 shift = 12, mask = 0xe0;
657 else if (c <= 0x001fffff)
658 shift = 18, mask = 0xf0;
659 else if (c <= 0x03ffffff)
660 shift = 24, mask = 0xf8;
662 shift = 30, mask = 0xfc;
664 extend_token (mask | (c >> shift));
668 extend_token ((unsigned char) (0x80 | (c >> shift)));
683 struct try_type type_sequence[] =
685 { &integer_type_node, 0, 0, 0},
686 { &unsigned_type_node, 1, 0, 0},
687 { &long_integer_type_node, 0, 1, 0},
688 { &long_unsigned_type_node, 1, 1, 0},
689 { &long_long_integer_type_node, 0, 1, 1},
690 { &long_long_unsigned_type_node, 1, 1, 1}
702 int conversion_errno;
703 REAL_VALUE_TYPE value;
711 struct pf_args * args = (struct pf_args *) data;
712 const char *typename;
714 args->conversion_errno = 0;
715 args->type = double_type_node;
718 /* The second argument, machine_mode, of REAL_VALUE_ATOF
719 tells the desired precision of the binary result
720 of decimal-to-binary conversion. */
725 error ("both 'f' and 'l' suffixes on floating constant");
727 args->type = float_type_node;
730 else if (args->lflag)
732 args->type = long_double_type_node;
733 typename = "long double";
735 else if (flag_single_precision_constant)
737 args->type = float_type_node;
742 if (args->base == 16)
743 args->value = REAL_VALUE_HTOF (args->str, TYPE_MODE (args->type));
745 args->value = REAL_VALUE_ATOF (args->str, TYPE_MODE (args->type));
747 args->conversion_errno = errno;
748 /* A diagnostic is required here by some ISO C testsuites.
749 This is not pedwarn, because some people don't want
750 an error for this. */
751 if (REAL_VALUE_ISINF (args->value) && pedantic)
752 warning ("floating point number exceeds range of '%s'", typename);
763 timevar_push (TV_CPP);
764 cpp_get_token (parse_in, &tok);
765 timevar_pop (TV_CPP);
767 /* The C++ front end does horrible things with the current line
768 number. To ensure an accurate line number, we must reset it
769 every time we return a token. */
770 lineno = SOURCE_LINE (map, cpp_get_line (parse_in)->line);
776 case CPP_OPEN_BRACE: indent_level++; break;
777 case CPP_CLOSE_BRACE: indent_level--; break;
779 /* Issue this error here, where we can get at tok.val.c. */
781 if (ISGRAPH (tok.val.c))
782 error ("stray '%c' in program", tok.val.c);
784 error ("stray '\\%o' in program", tok.val.c);
788 *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok.val.node));
792 *value = lex_number ((const char *)tok.val.str.text, tok.val.str.len);
797 *value = lex_charconst (&tok);
802 *value = lex_string ((const char *)tok.val.str.text,
803 tok.val.str.len, tok.type == CPP_WSTRING);
806 /* These tokens should not be visible outside cpplib. */
807 case CPP_HEADER_NAME:
818 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
821 lex_number (str, len)
827 int largest_digit = 0;
833 enum anon1 { NOT_FLOAT = 0, AFTER_POINT, AFTER_EXPON } floatflag = NOT_FLOAT;
835 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
836 The code below which fills the parts array assumes that a host
837 int is at least twice as wide as a host char, and that
838 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
839 Two HOST_WIDE_INTs is the largest int literal we can store.
840 In order to detect overflow below, the number of parts (TOTAL_PARTS)
841 must be exactly the number of parts needed to hold the bits
842 of two HOST_WIDE_INTs. */
843 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
844 unsigned int parts[TOTAL_PARTS];
846 /* Optimize for most frequent case. */
850 return integer_zero_node;
851 else if (*str == '1')
852 return integer_one_node;
854 return build_int_2 (*str - '0', 0);
857 for (count = 0; count < TOTAL_PARTS; count++)
860 /* len is known to be >1 at this point. */
863 if (len > 2 && str[0] == '0' && (str[1] == 'x' || str[1] == 'X'))
868 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
869 else if (str[0] == '0' && ISDIGIT (str[1]))
881 if (floatflag == AFTER_POINT)
882 ERROR ("too many decimal points in floating constant");
883 else if (floatflag == AFTER_EXPON)
884 ERROR ("decimal point in exponent - impossible!");
886 floatflag = AFTER_POINT;
892 /* Possible future extension: silently ignore _ in numbers,
893 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
894 but somewhat easier to read. Ada has this? */
895 ERROR ("underscore in number");
899 /* It is not a decimal point.
900 It should be a digit (perhaps a hex digit). */
906 else if (base <= 10 && (c == 'e' || c == 'E'))
909 floatflag = AFTER_EXPON;
912 else if (base == 16 && (c == 'p' || c == 'P'))
914 floatflag = AFTER_EXPON;
915 break; /* start of exponent */
917 else if (base == 16 && c >= 'a' && c <= 'f')
921 else if (base == 16 && c >= 'A' && c <= 'F')
928 break; /* start of suffix */
931 if (n >= largest_digit)
935 for (count = 0; count < TOTAL_PARTS; count++)
937 parts[count] *= base;
941 += (parts[count-1] >> HOST_BITS_PER_CHAR);
943 &= (1 << HOST_BITS_PER_CHAR) - 1;
949 /* If the highest-order part overflows (gets larger than
950 a host char will hold) then the whole number has
951 overflowed. Record this and truncate the highest-order
953 if (parts[TOTAL_PARTS - 1] >> HOST_BITS_PER_CHAR)
956 parts[TOTAL_PARTS - 1] &= (1 << HOST_BITS_PER_CHAR) - 1;
960 while (p < str + len);
962 /* This can happen on input like `int i = 0x;' */
964 ERROR ("numeric constant with no digits");
966 if (largest_digit >= base)
967 ERROR ("numeric constant contains digits beyond the radix");
969 if (floatflag != NOT_FLOAT)
972 int imag, fflag, lflag, conversion_errno;
973 REAL_VALUE_TYPE real;
977 if (base == 16 && pedantic && !flag_isoc99)
978 pedwarn ("floating constant may not be in radix 16");
980 if (base == 16 && floatflag != AFTER_EXPON)
981 ERROR ("hexadecimal floating constant has no exponent");
983 /* Read explicit exponent if any, and put it in tokenbuf. */
984 if ((base == 10 && ((c == 'e') || (c == 'E')))
985 || (base == 16 && (c == 'p' || c == 'P')))
989 if (p < str + len && (c == '+' || c == '-'))
991 /* Exponent is decimal, even if string is a hex float. */
993 ERROR ("floating constant exponent has no digits");
994 while (p < str + len && ISDIGIT (c))
1000 /* Copy the float constant now; we don't want any suffixes in the
1001 string passed to parse_float. */
1002 copy = alloca (p - str + 1);
1003 memcpy (copy, str, p - str);
1004 copy[p - str] = '\0';
1006 /* Now parse suffixes. */
1007 fflag = lflag = imag = 0;
1008 while (p < str + len)
1013 ERROR ("more than one 'f' suffix on floating constant");
1014 else if (warn_traditional && !in_system_header
1015 && ! cpp_sys_macro_p (parse_in))
1016 warning ("traditional C rejects the 'f' suffix");
1023 ERROR ("more than one 'l' suffix on floating constant");
1024 else if (warn_traditional && !in_system_header
1025 && ! cpp_sys_macro_p (parse_in))
1026 warning ("traditional C rejects the 'l' suffix");
1034 ERROR ("more than one 'i' or 'j' suffix on floating constant");
1036 pedwarn ("ISO C forbids imaginary numeric constants");
1041 ERROR ("invalid suffix on floating constant");
1044 /* Setup input for parse_float() */
1050 /* Convert string to a double, checking for overflow. */
1051 if (do_float_handler (parse_float, (PTR) &args))
1053 /* Receive output from parse_float() */
1057 /* We got an exception from parse_float() */
1058 ERROR ("floating constant out of range");
1060 /* Receive output from parse_float() */
1061 conversion_errno = args.conversion_errno;
1065 /* ERANGE is also reported for underflow,
1066 so test the value to distinguish overflow from that. */
1067 if (conversion_errno == ERANGE && !flag_traditional && pedantic
1068 && (REAL_VALUES_LESS (dconst1, real)
1069 || REAL_VALUES_LESS (real, dconstm1)))
1070 warning ("floating point number exceeds range of 'double'");
1073 /* Create a node with determined type and value. */
1075 value = build_complex (NULL_TREE, convert (type, integer_zero_node),
1076 build_real (type, real));
1078 value = build_real (type, real);
1082 tree trad_type, ansi_type, type;
1083 HOST_WIDE_INT high, low;
1084 int spec_unsigned = 0;
1086 int spec_long_long = 0;
1091 trad_type = ansi_type = type = NULL_TREE;
1092 while (p < str + len)
1099 error ("two 'u' suffixes on integer constant");
1100 else if (warn_traditional && !in_system_header
1101 && ! cpp_sys_macro_p (parse_in))
1102 warning ("traditional C rejects the 'u' suffix");
1113 error ("three 'l' suffixes on integer constant");
1115 error ("'lul' is not a valid integer suffix");
1116 else if (c != spec_long)
1117 error ("'Ll' and 'lL' are not valid integer suffixes");
1118 else if (pedantic && ! flag_isoc99
1119 && ! in_system_header && warn_long_long)
1120 pedwarn ("ISO C89 forbids long long integer constants");
1126 case 'i': case 'I': case 'j': case 'J':
1128 error ("more than one 'i' or 'j' suffix on integer constant");
1130 pedwarn ("ISO C forbids imaginary numeric constants");
1135 ERROR ("invalid suffix on integer constant");
1139 /* If the literal overflowed, pedwarn about it now. */
1143 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT * 2);
1146 /* This is simplified by the fact that our constant
1147 is always positive. */
1151 for (i = 0; i < HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR; i++)
1153 high |= ((HOST_WIDE_INT) parts[i + (HOST_BITS_PER_WIDE_INT
1154 / HOST_BITS_PER_CHAR)]
1155 << (i * HOST_BITS_PER_CHAR));
1156 low |= (HOST_WIDE_INT) parts[i] << (i * HOST_BITS_PER_CHAR);
1159 value = build_int_2 (low, high);
1160 TREE_TYPE (value) = long_long_unsigned_type_node;
1162 /* If warn_traditional, calculate both the ISO type and the
1163 traditional type, then see if they disagree.
1164 Otherwise, calculate only the type for the dialect in use. */
1165 if (warn_traditional || flag_traditional)
1167 /* Calculate the traditional type. */
1168 /* Traditionally, any constant is signed; but if unsigned is
1169 specified explicitly, obey that. Use the smallest size
1170 with the right number of bits, except for one special
1171 case with decimal constants. */
1172 if (! spec_long && base != 10
1173 && int_fits_type_p (value, unsigned_type_node))
1174 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1175 /* A decimal constant must be long if it does not fit in
1176 type int. I think this is independent of whether the
1177 constant is signed. */
1178 else if (! spec_long && base == 10
1179 && int_fits_type_p (value, integer_type_node))
1180 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1181 else if (! spec_long_long)
1182 trad_type = (spec_unsigned
1183 ? long_unsigned_type_node
1184 : long_integer_type_node);
1185 else if (int_fits_type_p (value,
1187 ? long_long_unsigned_type_node
1188 : long_long_integer_type_node))
1189 trad_type = (spec_unsigned
1190 ? long_long_unsigned_type_node
1191 : long_long_integer_type_node);
1193 trad_type = (spec_unsigned
1194 ? widest_unsigned_literal_type_node
1195 : widest_integer_literal_type_node);
1197 if (warn_traditional || ! flag_traditional)
1199 /* Calculate the ISO type. */
1200 if (! spec_long && ! spec_unsigned
1201 && int_fits_type_p (value, integer_type_node))
1202 ansi_type = integer_type_node;
1203 else if (! spec_long && (base != 10 || spec_unsigned)
1204 && int_fits_type_p (value, unsigned_type_node))
1205 ansi_type = unsigned_type_node;
1206 else if (! spec_unsigned && !spec_long_long
1207 && int_fits_type_p (value, long_integer_type_node))
1208 ansi_type = long_integer_type_node;
1209 else if (! spec_long_long
1210 && int_fits_type_p (value, long_unsigned_type_node))
1211 ansi_type = long_unsigned_type_node;
1212 else if (! spec_unsigned
1213 && int_fits_type_p (value, long_long_integer_type_node))
1214 ansi_type = long_long_integer_type_node;
1215 else if (int_fits_type_p (value, long_long_unsigned_type_node))
1216 ansi_type = long_long_unsigned_type_node;
1217 else if (! spec_unsigned
1218 && int_fits_type_p (value, widest_integer_literal_type_node))
1219 ansi_type = widest_integer_literal_type_node;
1221 ansi_type = widest_unsigned_literal_type_node;
1224 type = flag_traditional ? trad_type : ansi_type;
1226 /* We assume that constants specified in a non-decimal
1227 base are bit patterns, and that the programmer really
1228 meant what they wrote. */
1229 if (warn_traditional && !in_system_header
1230 && base == 10 && trad_type != ansi_type)
1232 if (TYPE_PRECISION (trad_type) != TYPE_PRECISION (ansi_type))
1233 warning ("width of integer constant changes with -traditional");
1234 else if (TREE_UNSIGNED (trad_type) != TREE_UNSIGNED (ansi_type))
1235 warning ("integer constant is unsigned in ISO C, signed with -traditional");
1237 warning ("width of integer constant may change on other systems with -traditional");
1240 if (pedantic && !flag_traditional && (flag_isoc99 || !spec_long_long)
1243 ? TYPE_PRECISION (long_long_integer_type_node)
1244 : TYPE_PRECISION (long_integer_type_node)) < TYPE_PRECISION (type)))
1247 pedwarn ("integer constant larger than the maximum value of %s",
1249 ? (TREE_UNSIGNED (type)
1250 ? "an unsigned long long int"
1251 : "a long long int")
1252 : "an unsigned long int"));
1255 if (base == 10 && ! spec_unsigned && TREE_UNSIGNED (type))
1256 warning ("decimal constant is so large that it is unsigned");
1260 if (TYPE_PRECISION (type)
1261 <= TYPE_PRECISION (integer_type_node))
1262 value = build_complex (NULL_TREE, integer_zero_node,
1263 convert (integer_type_node, value));
1265 ERROR ("complex integer constant is too wide for 'complex int'");
1267 else if (flag_traditional && !int_fits_type_p (value, type))
1268 /* The traditional constant 0x80000000 is signed
1269 but doesn't fit in the range of int.
1270 This will change it to -0x80000000, which does fit. */
1272 TREE_TYPE (value) = unsigned_type (type);
1273 value = convert (type, value);
1274 TREE_OVERFLOW (value) = TREE_CONSTANT_OVERFLOW (value) = 0;
1277 TREE_TYPE (value) = type;
1279 /* If it's still an integer (not a complex), and it doesn't
1280 fit in the type we choose for it, then pedwarn. */
1283 && TREE_CODE (TREE_TYPE (value)) == INTEGER_TYPE
1284 && ! int_fits_type_p (value, TREE_TYPE (value)))
1285 pedwarn ("integer constant is larger than the maximum value for its type");
1289 error ("missing white space after number '%.*s'", (int) (p - str), str);
1294 return integer_zero_node;
1298 lex_string (str, len, wide)
1304 char *buf = alloca ((len + 1) * (wide ? WCHAR_BYTES : 1));
1306 const char *p = str, *limit = str + len;
1308 unsigned width = wide ? WCHAR_TYPE_SIZE
1309 : TYPE_PRECISION (char_type_node);
1311 #ifdef MULTIBYTE_CHARS
1312 /* Reset multibyte conversion state. */
1313 (void) local_mbtowc (NULL, NULL, 0);
1318 #ifdef MULTIBYTE_CHARS
1322 char_len = local_mbtowc (&wc, p, limit - p);
1325 warning ("Ignoring invalid multibyte character");
1338 if (c == '\\' && !ignore_escape_flag)
1342 if (width < HOST_BITS_PER_INT)
1343 mask = ((unsigned int) 1 << width) - 1;
1346 c = cpp_parse_escape (parse_in, (const unsigned char **) &p,
1347 (const unsigned char *) limit,
1348 mask, flag_traditional);
1351 /* Add this single character into the buffer either as a wchar_t
1352 or as a single byte. */
1355 unsigned charwidth = TYPE_PRECISION (char_type_node);
1356 unsigned bytemask = (1 << charwidth) - 1;
1359 for (byte = 0; byte < WCHAR_BYTES; ++byte)
1362 if (byte >= (int) sizeof (c))
1365 n = (c >> (byte * charwidth)) & bytemask;
1366 if (BYTES_BIG_ENDIAN)
1367 q[WCHAR_BYTES - byte - 1] = n;
1379 /* Terminate the string value, either with a single byte zero
1380 or with a wide zero. */
1384 memset (q, 0, WCHAR_BYTES);
1392 value = build_string (q - buf, buf);
1395 TREE_TYPE (value) = wchar_array_type_node;
1397 TREE_TYPE (value) = char_array_type_node;
1401 /* Converts a (possibly wide) character constant token into a tree. */
1403 lex_charconst (token)
1404 const cpp_token *token;
1406 HOST_WIDE_INT result;
1408 unsigned int chars_seen;
1410 result = cpp_interpret_charconst (parse_in, token, warn_multichar,
1411 flag_traditional, &chars_seen);
1412 if (token->type == CPP_WCHAR)
1414 value = build_int_2 (result, 0);
1415 TREE_TYPE (value) = wchar_type_node;
1420 value = build_int_2 (result, -1);
1422 value = build_int_2 (result, 0);
1424 /* In C, a character constant has type 'int'.
1425 In C++ 'char', but multi-char charconsts have type 'int'. */
1426 if (c_language == clk_cplusplus && chars_seen <= 1)
1427 TREE_TYPE (value) = char_type_node;
1429 TREE_TYPE (value) = integer_type_node;