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 tree lex_number PARAMS ((const char *, unsigned int));
88 static tree lex_string PARAMS ((const unsigned char *, unsigned int,
90 static tree lex_charconst PARAMS ((const cpp_token *));
91 static void update_header_times PARAMS ((const char *));
92 static int dump_one_header PARAMS ((splay_tree_node, void *));
93 static void cb_line_change PARAMS ((cpp_reader *, const cpp_token *, int));
94 static void cb_ident PARAMS ((cpp_reader *, unsigned int,
96 static void cb_file_change PARAMS ((cpp_reader *, const struct line_map *));
97 static void cb_def_pragma PARAMS ((cpp_reader *, unsigned int));
98 static void cb_define PARAMS ((cpp_reader *, unsigned int,
100 static void cb_undef PARAMS ((cpp_reader *, unsigned int,
104 init_c_lex (filename)
105 const char *filename;
107 struct cpp_callbacks *cb;
108 struct c_fileinfo *toplevel;
110 /* Set up filename timing. Must happen before cpp_read_main_file. */
111 file_info_tree = splay_tree_new ((splay_tree_compare_fn)strcmp,
113 (splay_tree_delete_value_fn)free);
114 toplevel = get_fileinfo ("<top level>");
115 if (flag_detailed_statistics)
118 body_time = get_run_time ();
119 toplevel->time = body_time;
122 #ifdef MULTIBYTE_CHARS
123 /* Change to the native locale for multibyte conversions. */
124 setlocale (LC_CTYPE, "");
125 GET_ENVIRONMENT (literal_codeset, "LANG");
128 cb = cpp_get_callbacks (parse_in);
130 cb->line_change = cb_line_change;
131 cb->ident = cb_ident;
132 cb->file_change = cb_file_change;
133 cb->def_pragma = cb_def_pragma;
135 /* Set the debug callbacks if we can use them. */
136 if (debug_info_level == DINFO_LEVEL_VERBOSE
137 && (write_symbols == DWARF_DEBUG || write_symbols == DWARF2_DEBUG
138 || write_symbols == VMS_AND_DWARF2_DEBUG))
140 cb->define = cb_define;
141 cb->undef = cb_undef;
147 if (filename == NULL || !strcmp (filename, "-"))
150 return cpp_read_main_file (parse_in, filename, ident_hash);
153 /* A thin wrapper around the real parser that initializes the
154 integrated preprocessor after debug output has been initialized.
155 Also, make sure the start_source_file debug hook gets called for
156 the primary source file. */
159 c_common_parse_file ()
161 (*debug_hooks->start_source_file) (lineno, input_filename);
162 cpp_finish_options (parse_in);
172 struct c_fileinfo *fi;
174 n = splay_tree_lookup (file_info_tree, (splay_tree_key) name);
176 return (struct c_fileinfo *) n->value;
178 fi = (struct c_fileinfo *) xmalloc (sizeof (struct c_fileinfo));
180 fi->interface_only = 0;
181 fi->interface_unknown = 1;
182 splay_tree_insert (file_info_tree, (splay_tree_key) name,
183 (splay_tree_value) fi);
188 update_header_times (name)
191 /* Changing files again. This means currently collected time
192 is charged against header time, and body time starts back at 0. */
193 if (flag_detailed_statistics)
195 int this_time = get_run_time ();
196 struct c_fileinfo *file = get_fileinfo (name);
197 header_time += this_time - body_time;
198 file->time += this_time - body_time;
199 body_time = this_time;
204 dump_one_header (n, dummy)
206 void *dummy ATTRIBUTE_UNUSED;
208 print_time ((const char *) n->key,
209 ((struct c_fileinfo *) n->value)->time);
214 dump_time_statistics ()
216 struct c_fileinfo *file = get_fileinfo (input_filename);
217 int this_time = get_run_time ();
218 file->time += this_time - body_time;
220 fprintf (stderr, "\n******\n");
221 print_time ("header files (total)", header_time);
222 print_time ("main file (total)", this_time - body_time);
223 fprintf (stderr, "ratio = %g : 1\n",
224 (double)header_time / (double)(this_time - body_time));
225 fprintf (stderr, "\n******\n");
227 splay_tree_foreach (file_info_tree, dump_one_header, 0);
230 /* Not yet handled: #pragma, #define, #undef.
231 No need to deal with linemarkers under normal conditions. */
234 cb_ident (pfile, line, str)
235 cpp_reader *pfile ATTRIBUTE_UNUSED;
236 unsigned int line ATTRIBUTE_UNUSED;
237 const cpp_string *str ATTRIBUTE_UNUSED;
239 #ifdef ASM_OUTPUT_IDENT
242 /* Convert escapes in the string. */
243 tree value = lex_string (str->text, str->len, 0);
244 ASM_OUTPUT_IDENT (asm_out_file, TREE_STRING_POINTER (value));
249 /* Called at the start of every non-empty line. TOKEN is the first
250 lexed token on the line. Used for diagnostic line numbers. */
252 cb_line_change (pfile, token, parsing_args)
253 cpp_reader *pfile ATTRIBUTE_UNUSED;
254 const cpp_token *token;
255 int parsing_args ATTRIBUTE_UNUSED;
257 src_lineno = SOURCE_LINE (map, token->line);
261 cb_file_change (pfile, new_map)
262 cpp_reader *pfile ATTRIBUTE_UNUSED;
263 const struct line_map *new_map;
265 unsigned int to_line = SOURCE_LINE (new_map, new_map->to_line);
267 if (new_map->reason == LC_ENTER)
269 /* Don't stack the main buffer on the input stack;
270 we already did in compile_file. */
272 main_input_filename = new_map->to_file;
275 int included_at = SOURCE_LINE (new_map - 1, new_map->from_line - 1);
277 lineno = included_at;
278 push_srcloc (new_map->to_file, 1);
279 input_file_stack->indent_level = indent_level;
280 (*debug_hooks->start_source_file) (included_at, new_map->to_file);
281 #ifndef NO_IMPLICIT_EXTERN_C
284 else if (new_map->sysp == 2)
287 ++pending_lang_change;
292 else if (new_map->reason == LC_LEAVE)
294 #ifndef NO_IMPLICIT_EXTERN_C
295 if (c_header_level && --c_header_level == 0)
297 if (new_map->sysp == 2)
298 warning ("badly nested C headers from preprocessor");
299 --pending_lang_change;
303 if (indent_level != input_file_stack->indent_level)
305 warning_with_file_and_line
306 (input_filename, lineno,
307 "this file contains more '%c's than '%c's",
308 indent_level > input_file_stack->indent_level ? '{' : '}',
309 indent_level > input_file_stack->indent_level ? '}' : '{');
314 (*debug_hooks->end_source_file) (to_line);
317 update_header_times (new_map->to_file);
318 in_system_header = new_map->sysp != 0;
319 input_filename = new_map->to_file;
324 extract_interface_info ();
328 cb_def_pragma (pfile, line)
332 /* Issue a warning message if we have been asked to do so. Ignore
333 unknown pragmas in system headers unless an explicit
334 -Wunknown-pragmas has been given. */
335 if (warn_unknown_pragmas > in_system_header)
337 const unsigned char *space, *name = 0;
340 s = cpp_get_token (pfile);
341 space = cpp_token_as_text (pfile, s);
342 s = cpp_get_token (pfile);
343 if (s->type == CPP_NAME)
344 name = cpp_token_as_text (pfile, s);
346 lineno = SOURCE_LINE (map, line);
348 warning ("ignoring #pragma %s %s", space, name);
350 warning ("ignoring #pragma %s", space);
354 /* #define callback for DWARF and DWARF2 debug info. */
356 cb_define (pfile, line, node)
361 (*debug_hooks->define) (SOURCE_LINE (map, line),
362 (const char *) cpp_macro_definition (pfile, node));
365 /* #undef callback for DWARF and DWARF2 debug info. */
367 cb_undef (pfile, line, node)
368 cpp_reader *pfile ATTRIBUTE_UNUSED;
372 (*debug_hooks->undef) (SOURCE_LINE (map, line),
373 (const char *) NODE_NAME (node));
377 /* Returns nonzero if C is a universal-character-name. Give an error if it
378 is not one which may appear in an identifier, as per [extendid].
380 Note that extended character support in identifiers has not yet been
381 implemented. It is my personal opinion that this is not a desirable
382 feature. Portable code cannot count on support for more than the basic
383 identifier character set. */
396 /* None of the valid chars are outside the Basic Multilingual Plane (the
400 error ("universal-character-name '\\U%08x' not valid in identifier", c);
405 if ((c >= 0x00c0 && c <= 0x00d6)
406 || (c >= 0x00d8 && c <= 0x00f6)
407 || (c >= 0x00f8 && c <= 0x01f5)
408 || (c >= 0x01fa && c <= 0x0217)
409 || (c >= 0x0250 && c <= 0x02a8)
410 || (c >= 0x1e00 && c <= 0x1e9a)
411 || (c >= 0x1ea0 && c <= 0x1ef9))
416 || (c >= 0x0388 && c <= 0x038a)
418 || (c >= 0x038e && c <= 0x03a1)
419 || (c >= 0x03a3 && c <= 0x03ce)
420 || (c >= 0x03d0 && c <= 0x03d6)
425 || (c >= 0x03e2 && c <= 0x03f3)
426 || (c >= 0x1f00 && c <= 0x1f15)
427 || (c >= 0x1f18 && c <= 0x1f1d)
428 || (c >= 0x1f20 && c <= 0x1f45)
429 || (c >= 0x1f48 && c <= 0x1f4d)
430 || (c >= 0x1f50 && c <= 0x1f57)
434 || (c >= 0x1f5f && c <= 0x1f7d)
435 || (c >= 0x1f80 && c <= 0x1fb4)
436 || (c >= 0x1fb6 && c <= 0x1fbc)
437 || (c >= 0x1fc2 && c <= 0x1fc4)
438 || (c >= 0x1fc6 && c <= 0x1fcc)
439 || (c >= 0x1fd0 && c <= 0x1fd3)
440 || (c >= 0x1fd6 && c <= 0x1fdb)
441 || (c >= 0x1fe0 && c <= 0x1fec)
442 || (c >= 0x1ff2 && c <= 0x1ff4)
443 || (c >= 0x1ff6 && c <= 0x1ffc))
447 if ((c >= 0x0401 && c <= 0x040d)
448 || (c >= 0x040f && c <= 0x044f)
449 || (c >= 0x0451 && c <= 0x045c)
450 || (c >= 0x045e && c <= 0x0481)
451 || (c >= 0x0490 && c <= 0x04c4)
452 || (c >= 0x04c7 && c <= 0x04c8)
453 || (c >= 0x04cb && c <= 0x04cc)
454 || (c >= 0x04d0 && c <= 0x04eb)
455 || (c >= 0x04ee && c <= 0x04f5)
456 || (c >= 0x04f8 && c <= 0x04f9))
460 if ((c >= 0x0531 && c <= 0x0556)
461 || (c >= 0x0561 && c <= 0x0587))
465 if ((c >= 0x05d0 && c <= 0x05ea)
466 || (c >= 0x05f0 && c <= 0x05f4))
470 if ((c >= 0x0621 && c <= 0x063a)
471 || (c >= 0x0640 && c <= 0x0652)
472 || (c >= 0x0670 && c <= 0x06b7)
473 || (c >= 0x06ba && c <= 0x06be)
474 || (c >= 0x06c0 && c <= 0x06ce)
475 || (c >= 0x06e5 && c <= 0x06e7))
479 if ((c >= 0x0905 && c <= 0x0939)
480 || (c >= 0x0958 && c <= 0x0962))
484 if ((c >= 0x0985 && c <= 0x098c)
485 || (c >= 0x098f && c <= 0x0990)
486 || (c >= 0x0993 && c <= 0x09a8)
487 || (c >= 0x09aa && c <= 0x09b0)
489 || (c >= 0x09b6 && c <= 0x09b9)
490 || (c >= 0x09dc && c <= 0x09dd)
491 || (c >= 0x09df && c <= 0x09e1)
492 || (c >= 0x09f0 && c <= 0x09f1))
496 if ((c >= 0x0a05 && c <= 0x0a0a)
497 || (c >= 0x0a0f && c <= 0x0a10)
498 || (c >= 0x0a13 && c <= 0x0a28)
499 || (c >= 0x0a2a && c <= 0x0a30)
500 || (c >= 0x0a32 && c <= 0x0a33)
501 || (c >= 0x0a35 && c <= 0x0a36)
502 || (c >= 0x0a38 && c <= 0x0a39)
503 || (c >= 0x0a59 && c <= 0x0a5c)
508 if ((c >= 0x0a85 && c <= 0x0a8b)
510 || (c >= 0x0a8f && c <= 0x0a91)
511 || (c >= 0x0a93 && c <= 0x0aa8)
512 || (c >= 0x0aaa && c <= 0x0ab0)
513 || (c >= 0x0ab2 && c <= 0x0ab3)
514 || (c >= 0x0ab5 && c <= 0x0ab9)
519 if ((c >= 0x0b05 && c <= 0x0b0c)
520 || (c >= 0x0b0f && c <= 0x0b10)
521 || (c >= 0x0b13 && c <= 0x0b28)
522 || (c >= 0x0b2a && c <= 0x0b30)
523 || (c >= 0x0b32 && c <= 0x0b33)
524 || (c >= 0x0b36 && c <= 0x0b39)
525 || (c >= 0x0b5c && c <= 0x0b5d)
526 || (c >= 0x0b5f && c <= 0x0b61))
530 if ((c >= 0x0b85 && c <= 0x0b8a)
531 || (c >= 0x0b8e && c <= 0x0b90)
532 || (c >= 0x0b92 && c <= 0x0b95)
533 || (c >= 0x0b99 && c <= 0x0b9a)
535 || (c >= 0x0b9e && c <= 0x0b9f)
536 || (c >= 0x0ba3 && c <= 0x0ba4)
537 || (c >= 0x0ba8 && c <= 0x0baa)
538 || (c >= 0x0bae && c <= 0x0bb5)
539 || (c >= 0x0bb7 && c <= 0x0bb9))
543 if ((c >= 0x0c05 && c <= 0x0c0c)
544 || (c >= 0x0c0e && c <= 0x0c10)
545 || (c >= 0x0c12 && c <= 0x0c28)
546 || (c >= 0x0c2a && c <= 0x0c33)
547 || (c >= 0x0c35 && c <= 0x0c39)
548 || (c >= 0x0c60 && c <= 0x0c61))
552 if ((c >= 0x0c85 && c <= 0x0c8c)
553 || (c >= 0x0c8e && c <= 0x0c90)
554 || (c >= 0x0c92 && c <= 0x0ca8)
555 || (c >= 0x0caa && c <= 0x0cb3)
556 || (c >= 0x0cb5 && c <= 0x0cb9)
557 || (c >= 0x0ce0 && c <= 0x0ce1))
561 if ((c >= 0x0d05 && c <= 0x0d0c)
562 || (c >= 0x0d0e && c <= 0x0d10)
563 || (c >= 0x0d12 && c <= 0x0d28)
564 || (c >= 0x0d2a && c <= 0x0d39)
565 || (c >= 0x0d60 && c <= 0x0d61))
569 if ((c >= 0x0e01 && c <= 0x0e30)
570 || (c >= 0x0e32 && c <= 0x0e33)
571 || (c >= 0x0e40 && c <= 0x0e46)
572 || (c >= 0x0e4f && c <= 0x0e5b))
576 if ((c >= 0x0e81 && c <= 0x0e82)
582 || (c >= 0x0e94 && c <= 0x0e97)
583 || (c >= 0x0e99 && c <= 0x0e9f)
584 || (c >= 0x0ea1 && c <= 0x0ea3)
589 || (c >= 0x0ead && c <= 0x0eb0)
593 || (c >= 0x0ec0 && c <= 0x0ec4)
598 if ((c >= 0x10a0 && c <= 0x10c5)
599 || (c >= 0x10d0 && c <= 0x10f6))
603 if ((c >= 0x3041 && c <= 0x3094)
604 || (c >= 0x309b && c <= 0x309e))
608 if ((c >= 0x30a1 && c <= 0x30fe))
612 if ((c >= 0x3105 && c <= 0x312c))
616 if ((c >= 0x1100 && c <= 0x1159)
617 || (c >= 0x1161 && c <= 0x11a2)
618 || (c >= 0x11a8 && c <= 0x11f9))
621 /* CJK Unified Ideographs */
622 if ((c >= 0xf900 && c <= 0xfa2d)
623 || (c >= 0xfb1f && c <= 0xfb36)
624 || (c >= 0xfb38 && c <= 0xfb3c)
626 || (c >= 0xfb40 && c <= 0xfb41)
627 || (c >= 0xfb42 && c <= 0xfb44)
628 || (c >= 0xfb46 && c <= 0xfbb1)
629 || (c >= 0xfbd3 && c <= 0xfd3f)
630 || (c >= 0xfd50 && c <= 0xfd8f)
631 || (c >= 0xfd92 && c <= 0xfdc7)
632 || (c >= 0xfdf0 && c <= 0xfdfb)
633 || (c >= 0xfe70 && c <= 0xfe72)
635 || (c >= 0xfe76 && c <= 0xfefc)
636 || (c >= 0xff21 && c <= 0xff3a)
637 || (c >= 0xff41 && c <= 0xff5a)
638 || (c >= 0xff66 && c <= 0xffbe)
639 || (c >= 0xffc2 && c <= 0xffc7)
640 || (c >= 0xffca && c <= 0xffcf)
641 || (c >= 0xffd2 && c <= 0xffd7)
642 || (c >= 0xffda && c <= 0xffdc)
643 || (c >= 0x4e00 && c <= 0x9fa5))
646 error ("universal-character-name '\\u%04x' not valid in identifier", c);
651 /* Add the UTF-8 representation of C to the token_buffer. */
654 utf8_extend_token (c)
664 else if (c <= 0x000007ff)
665 shift = 6, mask = 0xc0;
666 else if (c <= 0x0000ffff)
667 shift = 12, mask = 0xe0;
668 else if (c <= 0x001fffff)
669 shift = 18, mask = 0xf0;
670 else if (c <= 0x03ffffff)
671 shift = 24, mask = 0xf8;
673 shift = 30, mask = 0xfc;
675 extend_token (mask | (c >> shift));
679 extend_token ((unsigned char) (0x80 | (c >> shift)));
688 tree *const node_var;
689 const char unsigned_flag;
690 const char long_flag;
691 const char long_long_flag;
694 struct try_type type_sequence[] =
696 { &integer_type_node, 0, 0, 0},
697 { &unsigned_type_node, 1, 0, 0},
698 { &long_integer_type_node, 0, 1, 0},
699 { &long_unsigned_type_node, 1, 1, 0},
700 { &long_long_integer_type_node, 0, 1, 1},
701 { &long_long_unsigned_type_node, 1, 1, 1}
709 const cpp_token *tok;
712 timevar_push (TV_CPP);
714 tok = cpp_get_token (parse_in);
715 while (tok->type == CPP_PADDING);
716 timevar_pop (TV_CPP);
718 /* The C++ front end does horrible things with the current line
719 number. To ensure an accurate line number, we must reset it
720 every time we return a token. */
726 case CPP_OPEN_BRACE: indent_level++; break;
727 case CPP_CLOSE_BRACE: indent_level--; break;
729 /* Issue this error here, where we can get at tok->val.c. */
731 if (ISGRAPH (tok->val.c))
732 error ("stray '%c' in program", tok->val.c);
734 error ("stray '\\%o' in program", tok->val.c);
738 *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node));
742 *value = lex_number ((const char *)tok->val.str.text, tok->val.str.len);
747 *value = lex_charconst (tok);
752 *value = lex_string (tok->val.str.text, tok->val.str.len,
753 tok->type == CPP_WSTRING);
756 /* These tokens should not be visible outside cpplib. */
757 case CPP_HEADER_NAME:
768 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
771 lex_number (str, len)
777 int largest_digit = 0;
783 enum anon1 { NOT_FLOAT = 0, AFTER_POINT, AFTER_EXPON } floatflag = NOT_FLOAT;
785 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
786 The code below which fills the parts array assumes that a host
787 int is at least twice as wide as a host char, and that
788 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
789 Two HOST_WIDE_INTs is the largest int literal we can store.
790 In order to detect overflow below, the number of parts (TOTAL_PARTS)
791 must be exactly the number of parts needed to hold the bits
792 of two HOST_WIDE_INTs. */
793 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
794 unsigned int parts[TOTAL_PARTS];
796 /* Optimize for most frequent case. */
800 return integer_zero_node;
801 else if (*str == '1')
802 return integer_one_node;
804 return build_int_2 (*str - '0', 0);
807 for (count = 0; count < TOTAL_PARTS; count++)
810 /* len is known to be >1 at this point. */
813 if (len > 2 && str[0] == '0' && (str[1] == 'x' || str[1] == 'X'))
818 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
819 else if (str[0] == '0' && ISDIGIT (str[1]))
831 if (floatflag == AFTER_POINT)
832 ERROR ("too many decimal points in floating constant");
833 else if (floatflag == AFTER_EXPON)
834 ERROR ("decimal point in exponent - impossible!");
836 floatflag = AFTER_POINT;
842 /* Possible future extension: silently ignore _ in numbers,
843 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
844 but somewhat easier to read. Ada has this? */
845 ERROR ("underscore in number");
849 /* It is not a decimal point.
850 It should be a digit (perhaps a hex digit). */
853 || (base == 16 && ISXDIGIT (c)))
857 else if (base <= 10 && (c == 'e' || c == 'E'))
860 floatflag = AFTER_EXPON;
863 else if (base == 16 && (c == 'p' || c == 'P'))
865 floatflag = AFTER_EXPON;
866 break; /* start of exponent */
871 break; /* start of suffix */
874 if (n >= largest_digit)
878 for (count = 0; count < TOTAL_PARTS; count++)
880 parts[count] *= base;
884 += (parts[count-1] >> HOST_BITS_PER_CHAR);
886 &= (1 << HOST_BITS_PER_CHAR) - 1;
892 /* If the highest-order part overflows (gets larger than
893 a host char will hold) then the whole number has
894 overflowed. Record this and truncate the highest-order
896 if (parts[TOTAL_PARTS - 1] >> HOST_BITS_PER_CHAR)
899 parts[TOTAL_PARTS - 1] &= (1 << HOST_BITS_PER_CHAR) - 1;
903 while (p < str + len);
905 /* This can happen on input like `int i = 0x;' */
907 ERROR ("numeric constant with no digits");
909 if (largest_digit >= base)
910 ERROR ("numeric constant contains digits beyond the radix");
912 if (floatflag != NOT_FLOAT)
915 const char *typename;
916 int imag, fflag, lflag;
917 REAL_VALUE_TYPE real;
920 if (base == 16 && floatflag != AFTER_EXPON)
921 ERROR ("hexadecimal floating constant has no exponent");
923 /* Read explicit exponent if any, and put it in tokenbuf. */
924 if ((base == 10 && ((c == 'e') || (c == 'E')))
925 || (base == 16 && (c == 'p' || c == 'P')))
929 if (p < str + len && (c == '+' || c == '-'))
931 /* Exponent is decimal, even if string is a hex float. */
933 ERROR ("floating constant exponent has no digits");
934 while (p < str + len && ISDIGIT (c))
940 /* Copy the float constant now; we don't want any suffixes in the
941 string passed to parse_float. */
942 copy = alloca (p - str + 1);
943 memcpy (copy, str, p - str);
944 copy[p - str] = '\0';
946 /* Now parse suffixes. */
947 fflag = lflag = imag = 0;
948 while (p < str + len)
953 ERROR ("more than one 'f' suffix on floating constant");
954 else if (warn_traditional && !in_system_header
955 && ! cpp_sys_macro_p (parse_in))
956 warning ("traditional C rejects the 'f' suffix");
963 ERROR ("more than one 'l' suffix on floating constant");
964 else if (warn_traditional && !in_system_header
965 && ! cpp_sys_macro_p (parse_in))
966 warning ("traditional C rejects the 'l' suffix");
974 ERROR ("more than one 'i' or 'j' suffix on floating constant");
976 pedwarn ("ISO C forbids imaginary numeric constants");
981 ERROR ("invalid suffix on floating constant");
984 type = double_type_node;
990 ERROR ("both 'f' and 'l' suffixes on floating constant");
992 type = float_type_node;
997 type = long_double_type_node;
998 typename = "long double";
1000 else if (flag_single_precision_constant)
1002 type = float_type_node;
1006 /* Warn about this only after we know we're not issuing an error. */
1007 if (base == 16 && pedantic && !flag_isoc99)
1008 pedwarn ("hexadecimal floating constants are only valid in C99");
1010 /* The second argument, machine_mode, of REAL_VALUE_ATOF
1011 tells the desired precision of the binary result
1012 of decimal-to-binary conversion. */
1014 real = REAL_VALUE_HTOF (copy, TYPE_MODE (type));
1016 real = REAL_VALUE_ATOF (copy, TYPE_MODE (type));
1018 /* A diagnostic is required here by some ISO C testsuites.
1019 This is not pedwarn, because some people don't want
1020 an error for this. */
1021 if (REAL_VALUE_ISINF (real) && pedantic)
1022 warning ("floating point number exceeds range of 'double'");
1024 /* Create a node with determined type and value. */
1026 value = build_complex (NULL_TREE, convert (type, integer_zero_node),
1027 build_real (type, real));
1029 value = build_real (type, real);
1033 tree trad_type, type;
1034 HOST_WIDE_INT high, low;
1035 int spec_unsigned = 0;
1037 int spec_long_long = 0;
1042 trad_type = type = NULL_TREE;
1043 while (p < str + len)
1050 error ("two 'u' suffixes on integer constant");
1051 else if (warn_traditional && !in_system_header
1052 && ! cpp_sys_macro_p (parse_in))
1053 warning ("traditional C rejects the 'u' suffix");
1064 error ("three 'l' suffixes on integer constant");
1066 error ("'lul' is not a valid integer suffix");
1067 else if (c != spec_long)
1068 error ("'Ll' and 'lL' are not valid integer suffixes");
1069 else if (pedantic && ! flag_isoc99
1070 && ! in_system_header && warn_long_long)
1071 pedwarn ("ISO C89 forbids long long integer constants");
1077 case 'i': case 'I': case 'j': case 'J':
1079 error ("more than one 'i' or 'j' suffix on integer constant");
1081 pedwarn ("ISO C forbids imaginary numeric constants");
1086 ERROR ("invalid suffix on integer constant");
1090 /* If the literal overflowed, pedwarn about it now. */
1094 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT * 2);
1097 /* This is simplified by the fact that our constant
1098 is always positive. */
1102 for (i = 0; i < HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR; i++)
1104 high |= ((HOST_WIDE_INT) parts[i + (HOST_BITS_PER_WIDE_INT
1105 / HOST_BITS_PER_CHAR)]
1106 << (i * HOST_BITS_PER_CHAR));
1107 low |= (HOST_WIDE_INT) parts[i] << (i * HOST_BITS_PER_CHAR);
1110 value = build_int_2 (low, high);
1111 TREE_TYPE (value) = long_long_unsigned_type_node;
1113 /* If warn_traditional, calculate both the ISO type and the
1114 traditional type, then see if they disagree. */
1115 if (warn_traditional)
1117 /* Traditionally, any constant is signed; but if unsigned is
1118 specified explicitly, obey that. Use the smallest size
1119 with the right number of bits, except for one special
1120 case with decimal constants. */
1121 if (! spec_long && base != 10
1122 && int_fits_type_p (value, unsigned_type_node))
1123 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1124 /* A decimal constant must be long if it does not fit in
1125 type int. I think this is independent of whether the
1126 constant is signed. */
1127 else if (! spec_long && base == 10
1128 && int_fits_type_p (value, integer_type_node))
1129 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1130 else if (! spec_long_long)
1131 trad_type = (spec_unsigned
1132 ? long_unsigned_type_node
1133 : long_integer_type_node);
1134 else if (int_fits_type_p (value,
1136 ? long_long_unsigned_type_node
1137 : long_long_integer_type_node))
1138 trad_type = (spec_unsigned
1139 ? long_long_unsigned_type_node
1140 : long_long_integer_type_node);
1142 trad_type = (spec_unsigned
1143 ? widest_unsigned_literal_type_node
1144 : widest_integer_literal_type_node);
1147 /* Calculate the ISO type. */
1148 if (! spec_long && ! spec_unsigned
1149 && int_fits_type_p (value, integer_type_node))
1150 type = integer_type_node;
1151 else if (! spec_long && (base != 10 || spec_unsigned)
1152 && int_fits_type_p (value, unsigned_type_node))
1153 type = unsigned_type_node;
1154 else if (! spec_unsigned && !spec_long_long
1155 && int_fits_type_p (value, long_integer_type_node))
1156 type = long_integer_type_node;
1157 else if (! spec_long_long
1158 && int_fits_type_p (value, long_unsigned_type_node))
1159 type = long_unsigned_type_node;
1160 else if (! spec_unsigned
1161 && int_fits_type_p (value, long_long_integer_type_node))
1162 type = long_long_integer_type_node;
1163 else if (int_fits_type_p (value, long_long_unsigned_type_node))
1164 type = long_long_unsigned_type_node;
1165 else if (! spec_unsigned
1166 && int_fits_type_p (value, widest_integer_literal_type_node))
1167 type = widest_integer_literal_type_node;
1169 type = widest_unsigned_literal_type_node;
1171 /* We assume that constants specified in a non-decimal
1172 base are bit patterns, and that the programmer really
1173 meant what they wrote. */
1174 if (warn_traditional && !in_system_header
1175 && base == 10 && trad_type != type)
1177 if (TYPE_PRECISION (trad_type) != TYPE_PRECISION (type))
1178 warning ("width of integer constant is different in traditional C");
1179 else if (TREE_UNSIGNED (trad_type) != TREE_UNSIGNED (type))
1180 warning ("integer constant is unsigned in ISO C, signed in traditional C");
1182 warning ("width of integer constant may change on other systems in traditional C");
1185 if (pedantic && (flag_isoc99 || !spec_long_long)
1188 ? TYPE_PRECISION (long_long_integer_type_node)
1189 : TYPE_PRECISION (long_integer_type_node)) < TYPE_PRECISION (type)))
1192 pedwarn ("integer constant larger than the maximum value of %s",
1194 ? (TREE_UNSIGNED (type)
1195 ? _("an unsigned long long int")
1196 : _("a long long int"))
1197 : _("an unsigned long int")));
1200 if (base == 10 && ! spec_unsigned && TREE_UNSIGNED (type))
1201 warning ("decimal constant is so large that it is unsigned");
1205 if (TYPE_PRECISION (type)
1206 <= TYPE_PRECISION (integer_type_node))
1207 value = build_complex (NULL_TREE, integer_zero_node,
1208 convert (integer_type_node, value));
1210 ERROR ("complex integer constant is too wide for 'complex int'");
1213 TREE_TYPE (value) = type;
1215 /* If it's still an integer (not a complex), and it doesn't
1216 fit in the type we choose for it, then pedwarn. */
1219 && TREE_CODE (TREE_TYPE (value)) == INTEGER_TYPE
1220 && ! int_fits_type_p (value, TREE_TYPE (value)))
1221 pedwarn ("integer constant is larger than the maximum value for its type");
1225 error ("missing white space after number '%.*s'", (int) (p - str), str);
1230 return integer_zero_node;
1234 lex_string (str, len, wide)
1235 const unsigned char *str;
1240 char *buf = alloca ((len + 1) * (wide ? WCHAR_BYTES : 1));
1242 const unsigned char *p = str, *limit = str + len;
1244 unsigned width = wide ? WCHAR_TYPE_SIZE
1245 : TYPE_PRECISION (char_type_node);
1247 #ifdef MULTIBYTE_CHARS
1248 /* Reset multibyte conversion state. */
1249 (void) local_mbtowc (NULL, NULL, 0);
1254 #ifdef MULTIBYTE_CHARS
1258 char_len = local_mbtowc (&wc, (const char *) p, limit - p);
1261 warning ("ignoring invalid multibyte character");
1274 if (c == '\\' && !ignore_escape_flag)
1278 if (width < HOST_BITS_PER_INT)
1279 mask = ((unsigned int) 1 << width) - 1;
1282 c = cpp_parse_escape (parse_in, &p, limit, mask);
1285 /* Add this single character into the buffer either as a wchar_t
1286 or as a single byte. */
1289 unsigned charwidth = TYPE_PRECISION (char_type_node);
1290 unsigned bytemask = (1 << charwidth) - 1;
1293 for (byte = 0; byte < WCHAR_BYTES; ++byte)
1296 if (byte >= (int) sizeof (c))
1299 n = (c >> (byte * charwidth)) & bytemask;
1300 if (BYTES_BIG_ENDIAN)
1301 q[WCHAR_BYTES - byte - 1] = n;
1313 /* Terminate the string value, either with a single byte zero
1314 or with a wide zero. */
1318 memset (q, 0, WCHAR_BYTES);
1326 value = build_string (q - buf, buf);
1329 TREE_TYPE (value) = wchar_array_type_node;
1331 TREE_TYPE (value) = char_array_type_node;
1335 /* Converts a (possibly wide) character constant token into a tree. */
1337 lex_charconst (token)
1338 const cpp_token *token;
1340 HOST_WIDE_INT result;
1342 unsigned int chars_seen;
1344 result = cpp_interpret_charconst (parse_in, token, warn_multichar,
1346 if (token->type == CPP_WCHAR)
1348 value = build_int_2 (result, 0);
1349 TREE_TYPE (value) = wchar_type_node;
1354 value = build_int_2 (result, -1);
1356 value = build_int_2 (result, 0);
1358 /* In C, a character constant has type 'int'.
1359 In C++ 'char', but multi-char charconsts have type 'int'. */
1360 if (c_language == clk_cplusplus && chars_seen <= 1)
1361 TREE_TYPE (value) = char_type_node;
1363 TREE_TYPE (value) = integer_type_node;