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 /* We may keep statistics about how long which files took to compile. */
61 static int header_time, body_time;
62 static splay_tree file_info_tree;
64 /* Cause the `yydebug' variable to be defined. */
67 /* File used for outputting assembler code. */
68 extern FILE *asm_out_file;
70 #undef WCHAR_TYPE_SIZE
71 #define WCHAR_TYPE_SIZE TYPE_PRECISION (wchar_type_node)
73 /* Number of bytes in a wide character. */
74 #define WCHAR_BYTES (WCHAR_TYPE_SIZE / BITS_PER_UNIT)
76 int indent_level; /* Number of { minus number of }. */
77 int pending_lang_change; /* If we need to switch languages - C++ only */
78 int c_header_level; /* depth in C headers - C++ only */
80 /* Nonzero tells yylex to ignore \ in string constants. */
81 static int ignore_escape_flag;
83 static void parse_float PARAMS ((PTR));
84 static tree lex_number PARAMS ((const char *, unsigned int));
85 static tree lex_string PARAMS ((const char *, unsigned int, int));
86 static tree lex_charconst PARAMS ((const cpp_token *));
87 static void update_header_times PARAMS ((const char *));
88 static int dump_one_header PARAMS ((splay_tree_node, void *));
89 static void cb_ident PARAMS ((cpp_reader *, const cpp_string *));
90 static void cb_file_change PARAMS ((cpp_reader *, const cpp_file_change *));
91 static void cb_def_pragma PARAMS ((cpp_reader *));
92 static void cb_define PARAMS ((cpp_reader *, cpp_hashnode *));
93 static void cb_undef PARAMS ((cpp_reader *, cpp_hashnode *));
99 struct cpp_callbacks *cb;
100 struct c_fileinfo *toplevel;
102 /* Set up filename timing. Must happen before cpp_start_read. */
103 file_info_tree = splay_tree_new ((splay_tree_compare_fn)strcmp,
105 (splay_tree_delete_value_fn)free);
106 toplevel = get_fileinfo ("<top level>");
107 if (flag_detailed_statistics)
110 body_time = get_run_time ();
111 toplevel->time = body_time;
114 #ifdef MULTIBYTE_CHARS
115 /* Change to the native locale for multibyte conversions. */
116 setlocale (LC_CTYPE, "");
117 GET_ENVIRONMENT (literal_codeset, "LANG");
120 cb = cpp_get_callbacks (parse_in);
122 cb->ident = cb_ident;
123 cb->file_change = cb_file_change;
124 cb->def_pragma = cb_def_pragma;
126 /* Set the debug callbacks if we can use them. */
127 if (debug_info_level == DINFO_LEVEL_VERBOSE
128 && (write_symbols == DWARF_DEBUG || write_symbols == DWARF2_DEBUG))
130 cb->define = cb_define;
131 cb->undef = cb_undef;
135 if (filename == 0 || !strcmp (filename, "-"))
136 filename = "stdin", cpp_filename = "";
138 cpp_filename = filename;
146 /* A thin wrapper around the real parser that initializes the
147 integrated preprocessor after debug output has been initialized. */
152 if (! cpp_start_read (parse_in, cpp_filename))
153 return 1; /* cpplib has emitted an error. */
163 struct c_fileinfo *fi;
165 n = splay_tree_lookup (file_info_tree, (splay_tree_key) name);
167 return (struct c_fileinfo *) n->value;
169 fi = (struct c_fileinfo *) xmalloc (sizeof (struct c_fileinfo));
171 fi->interface_only = 0;
172 fi->interface_unknown = 1;
173 splay_tree_insert (file_info_tree, (splay_tree_key) name,
174 (splay_tree_value) fi);
179 update_header_times (name)
182 /* Changing files again. This means currently collected time
183 is charged against header time, and body time starts back at 0. */
184 if (flag_detailed_statistics)
186 int this_time = get_run_time ();
187 struct c_fileinfo *file = get_fileinfo (name);
188 header_time += this_time - body_time;
189 file->time += this_time - body_time;
190 body_time = this_time;
195 dump_one_header (n, dummy)
197 void *dummy ATTRIBUTE_UNUSED;
199 print_time ((const char *) n->key,
200 ((struct c_fileinfo *) n->value)->time);
205 dump_time_statistics ()
207 struct c_fileinfo *file = get_fileinfo (input_filename);
208 int this_time = get_run_time ();
209 file->time += this_time - body_time;
211 fprintf (stderr, "\n******\n");
212 print_time ("header files (total)", header_time);
213 print_time ("main file (total)", this_time - body_time);
214 fprintf (stderr, "ratio = %g : 1\n",
215 (double)header_time / (double)(this_time - body_time));
216 fprintf (stderr, "\n******\n");
218 splay_tree_foreach (file_info_tree, dump_one_header, 0);
221 /* Not yet handled: #pragma, #define, #undef.
222 No need to deal with linemarkers under normal conditions. */
225 cb_ident (pfile, str)
226 cpp_reader *pfile ATTRIBUTE_UNUSED;
227 const cpp_string *str ATTRIBUTE_UNUSED;
229 #ifdef ASM_OUTPUT_IDENT
232 /* Convert escapes in the string. */
233 tree value = lex_string ((const char *)str->text, str->len, 0);
234 ASM_OUTPUT_IDENT (asm_out_file, TREE_STRING_POINTER (value));
240 cb_file_change (pfile, fc)
241 cpp_reader *pfile ATTRIBUTE_UNUSED;
242 const cpp_file_change *fc;
244 if (fc->reason == FC_ENTER)
246 /* Don't stack the main buffer on the input stack;
247 we already did in compile_file. */
248 if (fc->from.filename)
250 lineno = fc->from.lineno;
251 push_srcloc (fc->to.filename, 1);
252 input_file_stack->indent_level = indent_level;
253 (*debug_hooks->start_source_file) (fc->from.lineno, fc->to.filename);
254 #ifndef NO_IMPLICIT_EXTERN_C
257 else if (fc->externc)
260 ++pending_lang_change;
265 main_input_filename = fc->to.filename;
267 else if (fc->reason == FC_LEAVE)
269 /* Popping out of a file. */
270 if (input_file_stack->next)
272 #ifndef NO_IMPLICIT_EXTERN_C
273 if (c_header_level && --c_header_level == 0)
276 warning ("badly nested C headers from preprocessor");
277 --pending_lang_change;
281 if (indent_level != input_file_stack->indent_level)
283 warning_with_file_and_line
284 (input_filename, lineno,
285 "This file contains more '%c's than '%c's.",
286 indent_level > input_file_stack->indent_level ? '{' : '}',
287 indent_level > input_file_stack->indent_level ? '}' : '{');
291 (*debug_hooks->end_source_file) (input_file_stack->line);
294 error ("leaving more files than we entered");
297 update_header_times (fc->to.filename);
298 in_system_header = fc->sysp != 0;
299 input_filename = fc->to.filename;
300 lineno = fc->to.lineno; /* Do we need this? */
303 extract_interface_info ();
307 cb_def_pragma (pfile)
310 /* Issue a warning message if we have been asked to do so. Ignore
311 unknown pragmas in system headers unless an explicit
312 -Wunknown-pragmas has been given. */
313 if (warn_unknown_pragmas > in_system_header)
315 const unsigned char *space, *name = 0;
318 cpp_get_token (pfile, &s);
319 space = cpp_token_as_text (pfile, &s);
320 cpp_get_token (pfile, &s);
321 if (s.type == CPP_NAME)
322 name = cpp_token_as_text (pfile, &s);
324 lineno = cpp_get_line (parse_in)->line;
326 warning ("ignoring #pragma %s %s", space, name);
328 warning ("ignoring #pragma %s", space);
332 /* #define callback for DWARF and DWARF2 debug info. */
334 cb_define (pfile, node)
338 (*debug_hooks->define) (cpp_get_line (pfile)->line,
339 (const char *) cpp_macro_definition (pfile, node));
342 /* #undef callback for DWARF and DWARF2 debug info. */
344 cb_undef (pfile, node)
348 (*debug_hooks->undef) (cpp_get_line (pfile)->line,
349 (const char *) NODE_NAME (node));
353 /* Returns nonzero if C is a universal-character-name. Give an error if it
354 is not one which may appear in an identifier, as per [extendid].
356 Note that extended character support in identifiers has not yet been
357 implemented. It is my personal opinion that this is not a desirable
358 feature. Portable code cannot count on support for more than the basic
359 identifier character set. */
372 /* None of the valid chars are outside the Basic Multilingual Plane (the
376 error ("universal-character-name '\\U%08x' not valid in identifier", c);
381 if ((c >= 0x00c0 && c <= 0x00d6)
382 || (c >= 0x00d8 && c <= 0x00f6)
383 || (c >= 0x00f8 && c <= 0x01f5)
384 || (c >= 0x01fa && c <= 0x0217)
385 || (c >= 0x0250 && c <= 0x02a8)
386 || (c >= 0x1e00 && c <= 0x1e9a)
387 || (c >= 0x1ea0 && c <= 0x1ef9))
392 || (c >= 0x0388 && c <= 0x038a)
394 || (c >= 0x038e && c <= 0x03a1)
395 || (c >= 0x03a3 && c <= 0x03ce)
396 || (c >= 0x03d0 && c <= 0x03d6)
401 || (c >= 0x03e2 && c <= 0x03f3)
402 || (c >= 0x1f00 && c <= 0x1f15)
403 || (c >= 0x1f18 && c <= 0x1f1d)
404 || (c >= 0x1f20 && c <= 0x1f45)
405 || (c >= 0x1f48 && c <= 0x1f4d)
406 || (c >= 0x1f50 && c <= 0x1f57)
410 || (c >= 0x1f5f && c <= 0x1f7d)
411 || (c >= 0x1f80 && c <= 0x1fb4)
412 || (c >= 0x1fb6 && c <= 0x1fbc)
413 || (c >= 0x1fc2 && c <= 0x1fc4)
414 || (c >= 0x1fc6 && c <= 0x1fcc)
415 || (c >= 0x1fd0 && c <= 0x1fd3)
416 || (c >= 0x1fd6 && c <= 0x1fdb)
417 || (c >= 0x1fe0 && c <= 0x1fec)
418 || (c >= 0x1ff2 && c <= 0x1ff4)
419 || (c >= 0x1ff6 && c <= 0x1ffc))
423 if ((c >= 0x0401 && c <= 0x040d)
424 || (c >= 0x040f && c <= 0x044f)
425 || (c >= 0x0451 && c <= 0x045c)
426 || (c >= 0x045e && c <= 0x0481)
427 || (c >= 0x0490 && c <= 0x04c4)
428 || (c >= 0x04c7 && c <= 0x04c8)
429 || (c >= 0x04cb && c <= 0x04cc)
430 || (c >= 0x04d0 && c <= 0x04eb)
431 || (c >= 0x04ee && c <= 0x04f5)
432 || (c >= 0x04f8 && c <= 0x04f9))
436 if ((c >= 0x0531 && c <= 0x0556)
437 || (c >= 0x0561 && c <= 0x0587))
441 if ((c >= 0x05d0 && c <= 0x05ea)
442 || (c >= 0x05f0 && c <= 0x05f4))
446 if ((c >= 0x0621 && c <= 0x063a)
447 || (c >= 0x0640 && c <= 0x0652)
448 || (c >= 0x0670 && c <= 0x06b7)
449 || (c >= 0x06ba && c <= 0x06be)
450 || (c >= 0x06c0 && c <= 0x06ce)
451 || (c >= 0x06e5 && c <= 0x06e7))
455 if ((c >= 0x0905 && c <= 0x0939)
456 || (c >= 0x0958 && c <= 0x0962))
460 if ((c >= 0x0985 && c <= 0x098c)
461 || (c >= 0x098f && c <= 0x0990)
462 || (c >= 0x0993 && c <= 0x09a8)
463 || (c >= 0x09aa && c <= 0x09b0)
465 || (c >= 0x09b6 && c <= 0x09b9)
466 || (c >= 0x09dc && c <= 0x09dd)
467 || (c >= 0x09df && c <= 0x09e1)
468 || (c >= 0x09f0 && c <= 0x09f1))
472 if ((c >= 0x0a05 && c <= 0x0a0a)
473 || (c >= 0x0a0f && c <= 0x0a10)
474 || (c >= 0x0a13 && c <= 0x0a28)
475 || (c >= 0x0a2a && c <= 0x0a30)
476 || (c >= 0x0a32 && c <= 0x0a33)
477 || (c >= 0x0a35 && c <= 0x0a36)
478 || (c >= 0x0a38 && c <= 0x0a39)
479 || (c >= 0x0a59 && c <= 0x0a5c)
484 if ((c >= 0x0a85 && c <= 0x0a8b)
486 || (c >= 0x0a8f && c <= 0x0a91)
487 || (c >= 0x0a93 && c <= 0x0aa8)
488 || (c >= 0x0aaa && c <= 0x0ab0)
489 || (c >= 0x0ab2 && c <= 0x0ab3)
490 || (c >= 0x0ab5 && c <= 0x0ab9)
495 if ((c >= 0x0b05 && c <= 0x0b0c)
496 || (c >= 0x0b0f && c <= 0x0b10)
497 || (c >= 0x0b13 && c <= 0x0b28)
498 || (c >= 0x0b2a && c <= 0x0b30)
499 || (c >= 0x0b32 && c <= 0x0b33)
500 || (c >= 0x0b36 && c <= 0x0b39)
501 || (c >= 0x0b5c && c <= 0x0b5d)
502 || (c >= 0x0b5f && c <= 0x0b61))
506 if ((c >= 0x0b85 && c <= 0x0b8a)
507 || (c >= 0x0b8e && c <= 0x0b90)
508 || (c >= 0x0b92 && c <= 0x0b95)
509 || (c >= 0x0b99 && c <= 0x0b9a)
511 || (c >= 0x0b9e && c <= 0x0b9f)
512 || (c >= 0x0ba3 && c <= 0x0ba4)
513 || (c >= 0x0ba8 && c <= 0x0baa)
514 || (c >= 0x0bae && c <= 0x0bb5)
515 || (c >= 0x0bb7 && c <= 0x0bb9))
519 if ((c >= 0x0c05 && c <= 0x0c0c)
520 || (c >= 0x0c0e && c <= 0x0c10)
521 || (c >= 0x0c12 && c <= 0x0c28)
522 || (c >= 0x0c2a && c <= 0x0c33)
523 || (c >= 0x0c35 && c <= 0x0c39)
524 || (c >= 0x0c60 && c <= 0x0c61))
528 if ((c >= 0x0c85 && c <= 0x0c8c)
529 || (c >= 0x0c8e && c <= 0x0c90)
530 || (c >= 0x0c92 && c <= 0x0ca8)
531 || (c >= 0x0caa && c <= 0x0cb3)
532 || (c >= 0x0cb5 && c <= 0x0cb9)
533 || (c >= 0x0ce0 && c <= 0x0ce1))
537 if ((c >= 0x0d05 && c <= 0x0d0c)
538 || (c >= 0x0d0e && c <= 0x0d10)
539 || (c >= 0x0d12 && c <= 0x0d28)
540 || (c >= 0x0d2a && c <= 0x0d39)
541 || (c >= 0x0d60 && c <= 0x0d61))
545 if ((c >= 0x0e01 && c <= 0x0e30)
546 || (c >= 0x0e32 && c <= 0x0e33)
547 || (c >= 0x0e40 && c <= 0x0e46)
548 || (c >= 0x0e4f && c <= 0x0e5b))
552 if ((c >= 0x0e81 && c <= 0x0e82)
558 || (c >= 0x0e94 && c <= 0x0e97)
559 || (c >= 0x0e99 && c <= 0x0e9f)
560 || (c >= 0x0ea1 && c <= 0x0ea3)
565 || (c >= 0x0ead && c <= 0x0eb0)
569 || (c >= 0x0ec0 && c <= 0x0ec4)
574 if ((c >= 0x10a0 && c <= 0x10c5)
575 || (c >= 0x10d0 && c <= 0x10f6))
579 if ((c >= 0x3041 && c <= 0x3094)
580 || (c >= 0x309b && c <= 0x309e))
584 if ((c >= 0x30a1 && c <= 0x30fe))
588 if ((c >= 0x3105 && c <= 0x312c))
592 if ((c >= 0x1100 && c <= 0x1159)
593 || (c >= 0x1161 && c <= 0x11a2)
594 || (c >= 0x11a8 && c <= 0x11f9))
597 /* CJK Unified Ideographs */
598 if ((c >= 0xf900 && c <= 0xfa2d)
599 || (c >= 0xfb1f && c <= 0xfb36)
600 || (c >= 0xfb38 && c <= 0xfb3c)
602 || (c >= 0xfb40 && c <= 0xfb41)
603 || (c >= 0xfb42 && c <= 0xfb44)
604 || (c >= 0xfb46 && c <= 0xfbb1)
605 || (c >= 0xfbd3 && c <= 0xfd3f)
606 || (c >= 0xfd50 && c <= 0xfd8f)
607 || (c >= 0xfd92 && c <= 0xfdc7)
608 || (c >= 0xfdf0 && c <= 0xfdfb)
609 || (c >= 0xfe70 && c <= 0xfe72)
611 || (c >= 0xfe76 && c <= 0xfefc)
612 || (c >= 0xff21 && c <= 0xff3a)
613 || (c >= 0xff41 && c <= 0xff5a)
614 || (c >= 0xff66 && c <= 0xffbe)
615 || (c >= 0xffc2 && c <= 0xffc7)
616 || (c >= 0xffca && c <= 0xffcf)
617 || (c >= 0xffd2 && c <= 0xffd7)
618 || (c >= 0xffda && c <= 0xffdc)
619 || (c >= 0x4e00 && c <= 0x9fa5))
622 error ("universal-character-name '\\u%04x' not valid in identifier", c);
627 /* Add the UTF-8 representation of C to the token_buffer. */
630 utf8_extend_token (c)
640 else if (c <= 0x000007ff)
641 shift = 6, mask = 0xc0;
642 else if (c <= 0x0000ffff)
643 shift = 12, mask = 0xe0;
644 else if (c <= 0x001fffff)
645 shift = 18, mask = 0xf0;
646 else if (c <= 0x03ffffff)
647 shift = 24, mask = 0xf8;
649 shift = 30, mask = 0xfc;
651 extend_token (mask | (c >> shift));
655 extend_token ((unsigned char) (0x80 | (c >> shift)));
670 struct try_type type_sequence[] =
672 { &integer_type_node, 0, 0, 0},
673 { &unsigned_type_node, 1, 0, 0},
674 { &long_integer_type_node, 0, 1, 0},
675 { &long_unsigned_type_node, 1, 1, 0},
676 { &long_long_integer_type_node, 0, 1, 1},
677 { &long_long_unsigned_type_node, 1, 1, 1}
689 int conversion_errno;
690 REAL_VALUE_TYPE value;
698 struct pf_args * args = (struct pf_args *) data;
699 const char *typename;
701 args->conversion_errno = 0;
702 args->type = double_type_node;
705 /* The second argument, machine_mode, of REAL_VALUE_ATOF
706 tells the desired precision of the binary result
707 of decimal-to-binary conversion. */
712 error ("both 'f' and 'l' suffixes on floating constant");
714 args->type = float_type_node;
717 else if (args->lflag)
719 args->type = long_double_type_node;
720 typename = "long double";
722 else if (flag_single_precision_constant)
724 args->type = float_type_node;
729 if (args->base == 16)
730 args->value = REAL_VALUE_HTOF (args->str, TYPE_MODE (args->type));
732 args->value = REAL_VALUE_ATOF (args->str, TYPE_MODE (args->type));
734 args->conversion_errno = errno;
735 /* A diagnostic is required here by some ISO C testsuites.
736 This is not pedwarn, because some people don't want
737 an error for this. */
738 if (REAL_VALUE_ISINF (args->value) && pedantic)
739 warning ("floating point number exceeds range of '%s'", typename);
750 timevar_push (TV_CPP);
751 cpp_get_token (parse_in, &tok);
752 timevar_pop (TV_CPP);
754 /* The C++ front end does horrible things with the current line
755 number. To ensure an accurate line number, we must reset it
756 every time we return a token. */
757 lineno = cpp_get_line (parse_in)->line;
763 case CPP_OPEN_BRACE: indent_level++; break;
764 case CPP_CLOSE_BRACE: indent_level--; break;
766 /* Issue this error here, where we can get at tok.val.c. */
768 if (ISGRAPH (tok.val.c))
769 error ("stray '%c' in program", tok.val.c);
771 error ("stray '\\%o' in program", tok.val.c);
775 *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok.val.node));
779 *value = lex_number ((const char *)tok.val.str.text, tok.val.str.len);
784 *value = lex_charconst (&tok);
789 *value = lex_string ((const char *)tok.val.str.text,
790 tok.val.str.len, tok.type == CPP_WSTRING);
793 /* These tokens should not be visible outside cpplib. */
794 case CPP_HEADER_NAME:
805 #define ERROR(msgid) do { error(msgid); goto syntax_error; } while(0)
808 lex_number (str, len)
814 int largest_digit = 0;
820 enum anon1 { NOT_FLOAT = 0, AFTER_POINT, AFTER_EXPON } floatflag = NOT_FLOAT;
822 /* We actually store only HOST_BITS_PER_CHAR bits in each part.
823 The code below which fills the parts array assumes that a host
824 int is at least twice as wide as a host char, and that
825 HOST_BITS_PER_WIDE_INT is an even multiple of HOST_BITS_PER_CHAR.
826 Two HOST_WIDE_INTs is the largest int literal we can store.
827 In order to detect overflow below, the number of parts (TOTAL_PARTS)
828 must be exactly the number of parts needed to hold the bits
829 of two HOST_WIDE_INTs. */
830 #define TOTAL_PARTS ((HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR) * 2)
831 unsigned int parts[TOTAL_PARTS];
833 /* Optimize for most frequent case. */
837 return integer_zero_node;
838 else if (*str == '1')
839 return integer_one_node;
841 return build_int_2 (*str - '0', 0);
844 for (count = 0; count < TOTAL_PARTS; count++)
847 /* len is known to be >1 at this point. */
850 if (len > 2 && str[0] == '0' && (str[1] == 'x' || str[1] == 'X'))
855 /* The ISDIGIT check is so we are not confused by a suffix on 0. */
856 else if (str[0] == '0' && ISDIGIT (str[1]))
868 if (floatflag == AFTER_POINT)
869 ERROR ("too many decimal points in floating constant");
870 else if (floatflag == AFTER_EXPON)
871 ERROR ("decimal point in exponent - impossible!");
873 floatflag = AFTER_POINT;
879 /* Possible future extension: silently ignore _ in numbers,
880 permitting cosmetic grouping - e.g. 0x8000_0000 == 0x80000000
881 but somewhat easier to read. Ada has this? */
882 ERROR ("underscore in number");
886 /* It is not a decimal point.
887 It should be a digit (perhaps a hex digit). */
893 else if (base <= 10 && (c == 'e' || c == 'E'))
896 floatflag = AFTER_EXPON;
899 else if (base == 16 && (c == 'p' || c == 'P'))
901 floatflag = AFTER_EXPON;
902 break; /* start of exponent */
904 else if (base == 16 && c >= 'a' && c <= 'f')
908 else if (base == 16 && c >= 'A' && c <= 'F')
915 break; /* start of suffix */
918 if (n >= largest_digit)
922 for (count = 0; count < TOTAL_PARTS; count++)
924 parts[count] *= base;
928 += (parts[count-1] >> HOST_BITS_PER_CHAR);
930 &= (1 << HOST_BITS_PER_CHAR) - 1;
936 /* If the highest-order part overflows (gets larger than
937 a host char will hold) then the whole number has
938 overflowed. Record this and truncate the highest-order
940 if (parts[TOTAL_PARTS - 1] >> HOST_BITS_PER_CHAR)
943 parts[TOTAL_PARTS - 1] &= (1 << HOST_BITS_PER_CHAR) - 1;
947 while (p < str + len);
949 /* This can happen on input like `int i = 0x;' */
951 ERROR ("numeric constant with no digits");
953 if (largest_digit >= base)
954 ERROR ("numeric constant contains digits beyond the radix");
956 if (floatflag != NOT_FLOAT)
959 int imag, fflag, lflag, conversion_errno;
960 REAL_VALUE_TYPE real;
964 if (base == 16 && pedantic && !flag_isoc99)
965 pedwarn ("floating constant may not be in radix 16");
967 if (base == 16 && floatflag != AFTER_EXPON)
968 ERROR ("hexadecimal floating constant has no exponent");
970 /* Read explicit exponent if any, and put it in tokenbuf. */
971 if ((base == 10 && ((c == 'e') || (c == 'E')))
972 || (base == 16 && (c == 'p' || c == 'P')))
976 if (p < str + len && (c == '+' || c == '-'))
978 /* Exponent is decimal, even if string is a hex float. */
980 ERROR ("floating constant exponent has no digits");
981 while (p < str + len && ISDIGIT (c))
987 /* Copy the float constant now; we don't want any suffixes in the
988 string passed to parse_float. */
989 copy = alloca (p - str + 1);
990 memcpy (copy, str, p - str);
991 copy[p - str] = '\0';
993 /* Now parse suffixes. */
994 fflag = lflag = imag = 0;
995 while (p < str + len)
1000 ERROR ("more than one 'f' suffix on floating constant");
1001 else if (warn_traditional && !in_system_header
1002 && ! cpp_sys_macro_p (parse_in))
1003 warning ("traditional C rejects the 'f' suffix");
1010 ERROR ("more than one 'l' suffix on floating constant");
1011 else if (warn_traditional && !in_system_header
1012 && ! cpp_sys_macro_p (parse_in))
1013 warning ("traditional C rejects the 'l' suffix");
1021 ERROR ("more than one 'i' or 'j' suffix on floating constant");
1023 pedwarn ("ISO C forbids imaginary numeric constants");
1028 ERROR ("invalid suffix on floating constant");
1031 /* Setup input for parse_float() */
1037 /* Convert string to a double, checking for overflow. */
1038 if (do_float_handler (parse_float, (PTR) &args))
1040 /* Receive output from parse_float() */
1044 /* We got an exception from parse_float() */
1045 ERROR ("floating constant out of range");
1047 /* Receive output from parse_float() */
1048 conversion_errno = args.conversion_errno;
1052 /* ERANGE is also reported for underflow,
1053 so test the value to distinguish overflow from that. */
1054 if (conversion_errno == ERANGE && !flag_traditional && pedantic
1055 && (REAL_VALUES_LESS (dconst1, real)
1056 || REAL_VALUES_LESS (real, dconstm1)))
1057 warning ("floating point number exceeds range of 'double'");
1060 /* Create a node with determined type and value. */
1062 value = build_complex (NULL_TREE, convert (type, integer_zero_node),
1063 build_real (type, real));
1065 value = build_real (type, real);
1069 tree trad_type, ansi_type, type;
1070 HOST_WIDE_INT high, low;
1071 int spec_unsigned = 0;
1073 int spec_long_long = 0;
1078 trad_type = ansi_type = type = NULL_TREE;
1079 while (p < str + len)
1086 error ("two 'u' suffixes on integer constant");
1087 else if (warn_traditional && !in_system_header
1088 && ! cpp_sys_macro_p (parse_in))
1089 warning ("traditional C rejects the 'u' suffix");
1100 error ("three 'l' suffixes on integer constant");
1102 error ("'lul' is not a valid integer suffix");
1103 else if (c != spec_long)
1104 error ("'Ll' and 'lL' are not valid integer suffixes");
1105 else if (pedantic && ! flag_isoc99
1106 && ! in_system_header && warn_long_long)
1107 pedwarn ("ISO C89 forbids long long integer constants");
1113 case 'i': case 'I': case 'j': case 'J':
1115 error ("more than one 'i' or 'j' suffix on integer constant");
1117 pedwarn ("ISO C forbids imaginary numeric constants");
1122 ERROR ("invalid suffix on integer constant");
1126 /* If the literal overflowed, pedwarn about it now. */
1130 pedwarn ("integer constant is too large for this configuration of the compiler - truncated to %d bits", HOST_BITS_PER_WIDE_INT * 2);
1133 /* This is simplified by the fact that our constant
1134 is always positive. */
1138 for (i = 0; i < HOST_BITS_PER_WIDE_INT / HOST_BITS_PER_CHAR; i++)
1140 high |= ((HOST_WIDE_INT) parts[i + (HOST_BITS_PER_WIDE_INT
1141 / HOST_BITS_PER_CHAR)]
1142 << (i * HOST_BITS_PER_CHAR));
1143 low |= (HOST_WIDE_INT) parts[i] << (i * HOST_BITS_PER_CHAR);
1146 value = build_int_2 (low, high);
1147 TREE_TYPE (value) = long_long_unsigned_type_node;
1149 /* If warn_traditional, calculate both the ISO type and the
1150 traditional type, then see if they disagree.
1151 Otherwise, calculate only the type for the dialect in use. */
1152 if (warn_traditional || flag_traditional)
1154 /* Calculate the traditional type. */
1155 /* Traditionally, any constant is signed; but if unsigned is
1156 specified explicitly, obey that. Use the smallest size
1157 with the right number of bits, except for one special
1158 case with decimal constants. */
1159 if (! spec_long && base != 10
1160 && int_fits_type_p (value, unsigned_type_node))
1161 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1162 /* A decimal constant must be long if it does not fit in
1163 type int. I think this is independent of whether the
1164 constant is signed. */
1165 else if (! spec_long && base == 10
1166 && int_fits_type_p (value, integer_type_node))
1167 trad_type = spec_unsigned ? unsigned_type_node : integer_type_node;
1168 else if (! spec_long_long)
1169 trad_type = (spec_unsigned
1170 ? long_unsigned_type_node
1171 : long_integer_type_node);
1172 else if (int_fits_type_p (value,
1174 ? long_long_unsigned_type_node
1175 : long_long_integer_type_node))
1176 trad_type = (spec_unsigned
1177 ? long_long_unsigned_type_node
1178 : long_long_integer_type_node);
1180 trad_type = (spec_unsigned
1181 ? widest_unsigned_literal_type_node
1182 : widest_integer_literal_type_node);
1184 if (warn_traditional || ! flag_traditional)
1186 /* Calculate the ISO type. */
1187 if (! spec_long && ! spec_unsigned
1188 && int_fits_type_p (value, integer_type_node))
1189 ansi_type = integer_type_node;
1190 else if (! spec_long && (base != 10 || spec_unsigned)
1191 && int_fits_type_p (value, unsigned_type_node))
1192 ansi_type = unsigned_type_node;
1193 else if (! spec_unsigned && !spec_long_long
1194 && int_fits_type_p (value, long_integer_type_node))
1195 ansi_type = long_integer_type_node;
1196 else if (! spec_long_long
1197 && int_fits_type_p (value, long_unsigned_type_node))
1198 ansi_type = long_unsigned_type_node;
1199 else if (! spec_unsigned
1200 && int_fits_type_p (value, long_long_integer_type_node))
1201 ansi_type = long_long_integer_type_node;
1202 else if (int_fits_type_p (value, long_long_unsigned_type_node))
1203 ansi_type = long_long_unsigned_type_node;
1204 else if (! spec_unsigned
1205 && int_fits_type_p (value, widest_integer_literal_type_node))
1206 ansi_type = widest_integer_literal_type_node;
1208 ansi_type = widest_unsigned_literal_type_node;
1211 type = flag_traditional ? trad_type : ansi_type;
1213 /* We assume that constants specified in a non-decimal
1214 base are bit patterns, and that the programmer really
1215 meant what they wrote. */
1216 if (warn_traditional && !in_system_header
1217 && base == 10 && trad_type != ansi_type)
1219 if (TYPE_PRECISION (trad_type) != TYPE_PRECISION (ansi_type))
1220 warning ("width of integer constant changes with -traditional");
1221 else if (TREE_UNSIGNED (trad_type) != TREE_UNSIGNED (ansi_type))
1222 warning ("integer constant is unsigned in ISO C, signed with -traditional");
1224 warning ("width of integer constant may change on other systems with -traditional");
1227 if (pedantic && !flag_traditional && (flag_isoc99 || !spec_long_long)
1230 ? TYPE_PRECISION (long_long_integer_type_node)
1231 : TYPE_PRECISION (long_integer_type_node)) < TYPE_PRECISION (type)))
1234 pedwarn ("integer constant larger than the maximum value of %s",
1236 ? (TREE_UNSIGNED (type)
1237 ? "an unsigned long long int"
1238 : "a long long int")
1239 : "an unsigned long int"));
1242 if (base == 10 && ! spec_unsigned && TREE_UNSIGNED (type))
1243 warning ("decimal constant is so large that it is unsigned");
1247 if (TYPE_PRECISION (type)
1248 <= TYPE_PRECISION (integer_type_node))
1249 value = build_complex (NULL_TREE, integer_zero_node,
1250 convert (integer_type_node, value));
1252 ERROR ("complex integer constant is too wide for 'complex int'");
1254 else if (flag_traditional && !int_fits_type_p (value, type))
1255 /* The traditional constant 0x80000000 is signed
1256 but doesn't fit in the range of int.
1257 This will change it to -0x80000000, which does fit. */
1259 TREE_TYPE (value) = unsigned_type (type);
1260 value = convert (type, value);
1261 TREE_OVERFLOW (value) = TREE_CONSTANT_OVERFLOW (value) = 0;
1264 TREE_TYPE (value) = type;
1266 /* If it's still an integer (not a complex), and it doesn't
1267 fit in the type we choose for it, then pedwarn. */
1270 && TREE_CODE (TREE_TYPE (value)) == INTEGER_TYPE
1271 && ! int_fits_type_p (value, TREE_TYPE (value)))
1272 pedwarn ("integer constant is larger than the maximum value for its type");
1276 error ("missing white space after number '%.*s'", (int) (p - str), str);
1281 return integer_zero_node;
1285 lex_string (str, len, wide)
1291 char *buf = alloca ((len + 1) * (wide ? WCHAR_BYTES : 1));
1293 const char *p = str, *limit = str + len;
1295 unsigned width = wide ? WCHAR_TYPE_SIZE
1296 : TYPE_PRECISION (char_type_node);
1298 #ifdef MULTIBYTE_CHARS
1299 /* Reset multibyte conversion state. */
1300 (void) local_mbtowc (NULL, NULL, 0);
1305 #ifdef MULTIBYTE_CHARS
1309 char_len = local_mbtowc (&wc, p, limit - p);
1312 warning ("Ignoring invalid multibyte character");
1325 if (c == '\\' && !ignore_escape_flag)
1329 if (width < HOST_BITS_PER_INT)
1330 mask = ((unsigned int) 1 << width) - 1;
1333 c = cpp_parse_escape (parse_in, (const unsigned char **) &p,
1334 (const unsigned char *) limit,
1335 mask, flag_traditional);
1338 /* Add this single character into the buffer either as a wchar_t
1339 or as a single byte. */
1342 unsigned charwidth = TYPE_PRECISION (char_type_node);
1343 unsigned bytemask = (1 << charwidth) - 1;
1346 for (byte = 0; byte < WCHAR_BYTES; ++byte)
1349 if (byte >= (int) sizeof (c))
1352 n = (c >> (byte * charwidth)) & bytemask;
1353 if (BYTES_BIG_ENDIAN)
1354 q[WCHAR_BYTES - byte - 1] = n;
1366 /* Terminate the string value, either with a single byte zero
1367 or with a wide zero. */
1371 memset (q, 0, WCHAR_BYTES);
1379 value = build_string (q - buf, buf);
1382 TREE_TYPE (value) = wchar_array_type_node;
1384 TREE_TYPE (value) = char_array_type_node;
1388 /* Converts a (possibly wide) character constant token into a tree. */
1390 lex_charconst (token)
1391 const cpp_token *token;
1393 HOST_WIDE_INT result;
1395 unsigned int chars_seen;
1397 result = cpp_interpret_charconst (parse_in, token, warn_multichar,
1398 flag_traditional, &chars_seen);
1399 if (token->type == CPP_WCHAR)
1401 value = build_int_2 (result, 0);
1402 TREE_TYPE (value) = wchar_type_node;
1407 value = build_int_2 (result, -1);
1409 value = build_int_2 (result, 0);
1411 /* In C, a character constant has type 'int'.
1412 In C++ 'char', but multi-char charconsts have type 'int'. */
1413 if (c_language == clk_cplusplus && chars_seen <= 1)
1414 TREE_TYPE (value) = char_type_node;
1416 TREE_TYPE (value) = integer_type_node;