Daily bump.
[official-gcc.git] / gcc / c-lex.c
blob5535eaeb42edb0a94da705a8d78e2b5b83170713
1 /* Mainly the interface between cpplib and the C front ends.
2 Copyright (C) 1987, 1988, 1989, 1992, 1994, 1995, 1996, 1997
3 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2007, 2008
4 Free Software Foundation, Inc.
6 This file is part of GCC.
8 GCC is free software; you can redistribute it and/or modify it under
9 the terms of the GNU General Public License as published by the Free
10 Software Foundation; either version 3, or (at your option) any later
11 version.
13 GCC is distributed in the hope that it will be useful, but WITHOUT ANY
14 WARRANTY; without even the implied warranty of MERCHANTABILITY or
15 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
16 for more details.
18 You should have received a copy of the GNU General Public License
19 along with GCC; see the file COPYING3. If not see
20 <http://www.gnu.org/licenses/>. */
22 #include "config.h"
23 #include "system.h"
24 #include "coretypes.h"
25 #include "tm.h"
27 #include "real.h"
28 #include "rtl.h"
29 #include "tree.h"
30 #include "input.h"
31 #include "output.h"
32 #include "c-tree.h"
33 #include "c-common.h"
34 #include "flags.h"
35 #include "timevar.h"
36 #include "cpplib.h"
37 #include "c-pragma.h"
38 #include "toplev.h"
39 #include "intl.h"
40 #include "tm_p.h"
41 #include "splay-tree.h"
42 #include "debug.h"
43 #include "target.h"
45 /* We may keep statistics about how long which files took to compile. */
46 static int header_time, body_time;
47 static splay_tree file_info_tree;
49 int pending_lang_change; /* If we need to switch languages - C++ only */
50 int c_header_level; /* depth in C headers - C++ only */
52 static tree interpret_integer (const cpp_token *, unsigned int);
53 static tree interpret_float (const cpp_token *, unsigned int);
54 static tree interpret_fixed (const cpp_token *, unsigned int);
55 static enum integer_type_kind narrowest_unsigned_type
56 (unsigned HOST_WIDE_INT, unsigned HOST_WIDE_INT, unsigned int);
57 static enum integer_type_kind narrowest_signed_type
58 (unsigned HOST_WIDE_INT, unsigned HOST_WIDE_INT, unsigned int);
59 static enum cpp_ttype lex_string (const cpp_token *, tree *, bool, bool);
60 static tree lex_charconst (const cpp_token *);
61 static void update_header_times (const char *);
62 static int dump_one_header (splay_tree_node, void *);
63 static void cb_line_change (cpp_reader *, const cpp_token *, int);
64 static void cb_ident (cpp_reader *, unsigned int, const cpp_string *);
65 static void cb_def_pragma (cpp_reader *, unsigned int);
66 static void cb_define (cpp_reader *, unsigned int, cpp_hashnode *);
67 static void cb_undef (cpp_reader *, unsigned int, cpp_hashnode *);
69 void
70 init_c_lex (void)
72 struct cpp_callbacks *cb;
73 struct c_fileinfo *toplevel;
75 /* The get_fileinfo data structure must be initialized before
76 cpp_read_main_file is called. */
77 toplevel = get_fileinfo ("<top level>");
78 if (flag_detailed_statistics)
80 header_time = 0;
81 body_time = get_run_time ();
82 toplevel->time = body_time;
85 cb = cpp_get_callbacks (parse_in);
87 cb->line_change = cb_line_change;
88 cb->ident = cb_ident;
89 cb->def_pragma = cb_def_pragma;
90 cb->valid_pch = c_common_valid_pch;
91 cb->read_pch = c_common_read_pch;
93 /* Set the debug callbacks if we can use them. */
94 if (debug_info_level == DINFO_LEVEL_VERBOSE
95 && (write_symbols == DWARF2_DEBUG
96 || write_symbols == VMS_AND_DWARF2_DEBUG))
98 cb->define = cb_define;
99 cb->undef = cb_undef;
103 struct c_fileinfo *
104 get_fileinfo (const char *name)
106 splay_tree_node n;
107 struct c_fileinfo *fi;
109 if (!file_info_tree)
110 file_info_tree = splay_tree_new ((splay_tree_compare_fn) strcmp,
112 (splay_tree_delete_value_fn) free);
114 n = splay_tree_lookup (file_info_tree, (splay_tree_key) name);
115 if (n)
116 return (struct c_fileinfo *) n->value;
118 fi = XNEW (struct c_fileinfo);
119 fi->time = 0;
120 fi->interface_only = 0;
121 fi->interface_unknown = 1;
122 splay_tree_insert (file_info_tree, (splay_tree_key) name,
123 (splay_tree_value) fi);
124 return fi;
127 static void
128 update_header_times (const char *name)
130 /* Changing files again. This means currently collected time
131 is charged against header time, and body time starts back at 0. */
132 if (flag_detailed_statistics)
134 int this_time = get_run_time ();
135 struct c_fileinfo *file = get_fileinfo (name);
136 header_time += this_time - body_time;
137 file->time += this_time - body_time;
138 body_time = this_time;
142 static int
143 dump_one_header (splay_tree_node n, void * ARG_UNUSED (dummy))
145 print_time ((const char *) n->key,
146 ((struct c_fileinfo *) n->value)->time);
147 return 0;
150 void
151 dump_time_statistics (void)
153 struct c_fileinfo *file = get_fileinfo (input_filename);
154 int this_time = get_run_time ();
155 file->time += this_time - body_time;
157 fprintf (stderr, "\n******\n");
158 print_time ("header files (total)", header_time);
159 print_time ("main file (total)", this_time - body_time);
160 fprintf (stderr, "ratio = %g : 1\n",
161 (double) header_time / (double) (this_time - body_time));
162 fprintf (stderr, "\n******\n");
164 splay_tree_foreach (file_info_tree, dump_one_header, 0);
167 static void
168 cb_ident (cpp_reader * ARG_UNUSED (pfile),
169 unsigned int ARG_UNUSED (line),
170 const cpp_string * ARG_UNUSED (str))
172 #ifdef ASM_OUTPUT_IDENT
173 if (!flag_no_ident)
175 /* Convert escapes in the string. */
176 cpp_string cstr = { 0, 0 };
177 if (cpp_interpret_string (pfile, str, 1, &cstr, CPP_STRING))
179 ASM_OUTPUT_IDENT (asm_out_file, (const char *) cstr.text);
180 free (CONST_CAST (unsigned char *, cstr.text));
183 #endif
186 /* Called at the start of every non-empty line. TOKEN is the first
187 lexed token on the line. Used for diagnostic line numbers. */
188 static void
189 cb_line_change (cpp_reader * ARG_UNUSED (pfile), const cpp_token *token,
190 int parsing_args)
192 if (token->type != CPP_EOF && !parsing_args)
193 input_location = token->src_loc;
196 void
197 fe_file_change (const struct line_map *new_map)
199 if (new_map == NULL)
200 return;
202 if (new_map->reason == LC_ENTER)
204 /* Don't stack the main buffer on the input stack;
205 we already did in compile_file. */
206 if (!MAIN_FILE_P (new_map))
208 unsigned int included_at = LAST_SOURCE_LINE_LOCATION (new_map - 1);
209 int line = 0;
210 if (included_at > BUILTINS_LOCATION)
211 line = SOURCE_LINE (new_map - 1, included_at);
213 input_location = new_map->start_location;
214 (*debug_hooks->start_source_file) (line, new_map->to_file);
215 #ifndef NO_IMPLICIT_EXTERN_C
216 if (c_header_level)
217 ++c_header_level;
218 else if (new_map->sysp == 2)
220 c_header_level = 1;
221 ++pending_lang_change;
223 #endif
226 else if (new_map->reason == LC_LEAVE)
228 #ifndef NO_IMPLICIT_EXTERN_C
229 if (c_header_level && --c_header_level == 0)
231 if (new_map->sysp == 2)
232 warning (0, "badly nested C headers from preprocessor");
233 --pending_lang_change;
235 #endif
236 input_location = new_map->start_location;
238 (*debug_hooks->end_source_file) (new_map->to_line);
241 update_header_times (new_map->to_file);
242 input_location = new_map->start_location;
245 static void
246 cb_def_pragma (cpp_reader *pfile, source_location loc)
248 /* Issue a warning message if we have been asked to do so. Ignore
249 unknown pragmas in system headers unless an explicit
250 -Wunknown-pragmas has been given. */
251 if (warn_unknown_pragmas > in_system_header)
253 const unsigned char *space, *name;
254 const cpp_token *s;
255 location_t fe_loc = loc;
257 space = name = (const unsigned char *) "";
258 s = cpp_get_token (pfile);
259 if (s->type != CPP_EOF)
261 space = cpp_token_as_text (pfile, s);
262 s = cpp_get_token (pfile);
263 if (s->type == CPP_NAME)
264 name = cpp_token_as_text (pfile, s);
267 warning_at (fe_loc, OPT_Wunknown_pragmas, "ignoring #pragma %s %s",
268 space, name);
272 /* #define callback for DWARF and DWARF2 debug info. */
273 static void
274 cb_define (cpp_reader *pfile, source_location loc, cpp_hashnode *node)
276 const struct line_map *map = linemap_lookup (line_table, loc);
277 (*debug_hooks->define) (SOURCE_LINE (map, loc),
278 (const char *) cpp_macro_definition (pfile, node));
281 /* #undef callback for DWARF and DWARF2 debug info. */
282 static void
283 cb_undef (cpp_reader * ARG_UNUSED (pfile), source_location loc,
284 cpp_hashnode *node)
286 const struct line_map *map = linemap_lookup (line_table, loc);
287 (*debug_hooks->undef) (SOURCE_LINE (map, loc),
288 (const char *) NODE_NAME (node));
291 /* Read a token and return its type. Fill *VALUE with its value, if
292 applicable. Fill *CPP_FLAGS with the token's flags, if it is
293 non-NULL. */
295 enum cpp_ttype
296 c_lex_with_flags (tree *value, location_t *loc, unsigned char *cpp_flags,
297 int lex_flags)
299 static bool no_more_pch;
300 const cpp_token *tok;
301 enum cpp_ttype type;
302 unsigned char add_flags = 0;
304 timevar_push (TV_CPP);
305 retry:
306 tok = cpp_get_token_with_location (parse_in, loc);
307 type = tok->type;
309 retry_after_at:
310 switch (type)
312 case CPP_PADDING:
313 goto retry;
315 case CPP_NAME:
316 *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node.node));
317 break;
319 case CPP_NUMBER:
321 unsigned int flags = cpp_classify_number (parse_in, tok);
323 switch (flags & CPP_N_CATEGORY)
325 case CPP_N_INVALID:
326 /* cpplib has issued an error. */
327 *value = error_mark_node;
328 errorcount++;
329 break;
331 case CPP_N_INTEGER:
332 /* C++ uses '0' to mark virtual functions as pure.
333 Set PURE_ZERO to pass this information to the C++ parser. */
334 if (tok->val.str.len == 1 && *tok->val.str.text == '0')
335 add_flags = PURE_ZERO;
336 *value = interpret_integer (tok, flags);
337 break;
339 case CPP_N_FLOATING:
340 *value = interpret_float (tok, flags);
341 break;
343 default:
344 gcc_unreachable ();
347 break;
349 case CPP_ATSIGN:
350 /* An @ may give the next token special significance in Objective-C. */
351 if (c_dialect_objc ())
353 location_t atloc = *loc;
354 location_t newloc;
356 retry_at:
357 tok = cpp_get_token_with_location (parse_in, &newloc);
358 type = tok->type;
359 switch (type)
361 case CPP_PADDING:
362 goto retry_at;
364 case CPP_STRING:
365 case CPP_WSTRING:
366 case CPP_STRING16:
367 case CPP_STRING32:
368 case CPP_UTF8STRING:
369 type = lex_string (tok, value, true, true);
370 break;
372 case CPP_NAME:
373 *value = HT_IDENT_TO_GCC_IDENT (HT_NODE (tok->val.node.node));
374 if (objc_is_reserved_word (*value))
376 type = CPP_AT_NAME;
377 break;
379 /* FALLTHROUGH */
381 default:
382 /* ... or not. */
383 error_at (atloc, "stray %<@%> in program");
384 *loc = newloc;
385 goto retry_after_at;
387 break;
390 /* FALLTHROUGH */
391 case CPP_HASH:
392 case CPP_PASTE:
394 unsigned char name[8];
396 *cpp_spell_token (parse_in, tok, name, true) = 0;
398 error ("stray %qs in program", name);
401 goto retry;
403 case CPP_OTHER:
405 cppchar_t c = tok->val.str.text[0];
407 if (c == '"' || c == '\'')
408 error ("missing terminating %c character", (int) c);
409 else if (ISGRAPH (c))
410 error ("stray %qc in program", (int) c);
411 else
412 error ("stray %<\\%o%> in program", (int) c);
414 goto retry;
416 case CPP_CHAR:
417 case CPP_WCHAR:
418 case CPP_CHAR16:
419 case CPP_CHAR32:
420 *value = lex_charconst (tok);
421 break;
423 case CPP_STRING:
424 case CPP_WSTRING:
425 case CPP_STRING16:
426 case CPP_STRING32:
427 case CPP_UTF8STRING:
428 if ((lex_flags & C_LEX_STRING_NO_JOIN) == 0)
430 type = lex_string (tok, value, false,
431 (lex_flags & C_LEX_STRING_NO_TRANSLATE) == 0);
432 break;
434 *value = build_string (tok->val.str.len, (const char *) tok->val.str.text);
435 break;
437 case CPP_PRAGMA:
438 *value = build_int_cst (NULL, tok->val.pragma);
439 break;
441 /* These tokens should not be visible outside cpplib. */
442 case CPP_HEADER_NAME:
443 case CPP_COMMENT:
444 case CPP_MACRO_ARG:
445 gcc_unreachable ();
447 default:
448 *value = NULL_TREE;
449 break;
452 if (cpp_flags)
453 *cpp_flags = tok->flags | add_flags;
455 if (!no_more_pch)
457 no_more_pch = true;
458 c_common_no_more_pch ();
461 timevar_pop (TV_CPP);
463 return type;
466 /* Returns the narrowest C-visible unsigned type, starting with the
467 minimum specified by FLAGS, that can fit HIGH:LOW, or itk_none if
468 there isn't one. */
470 static enum integer_type_kind
471 narrowest_unsigned_type (unsigned HOST_WIDE_INT low,
472 unsigned HOST_WIDE_INT high,
473 unsigned int flags)
475 int itk;
477 if ((flags & CPP_N_WIDTH) == CPP_N_SMALL)
478 itk = itk_unsigned_int;
479 else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM)
480 itk = itk_unsigned_long;
481 else
482 itk = itk_unsigned_long_long;
484 for (; itk < itk_none; itk += 2 /* skip unsigned types */)
486 tree upper = TYPE_MAX_VALUE (integer_types[itk]);
488 if ((unsigned HOST_WIDE_INT) TREE_INT_CST_HIGH (upper) > high
489 || ((unsigned HOST_WIDE_INT) TREE_INT_CST_HIGH (upper) == high
490 && TREE_INT_CST_LOW (upper) >= low))
491 return (enum integer_type_kind) itk;
494 return itk_none;
497 /* Ditto, but narrowest signed type. */
498 static enum integer_type_kind
499 narrowest_signed_type (unsigned HOST_WIDE_INT low,
500 unsigned HOST_WIDE_INT high, unsigned int flags)
502 int itk;
504 if ((flags & CPP_N_WIDTH) == CPP_N_SMALL)
505 itk = itk_int;
506 else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM)
507 itk = itk_long;
508 else
509 itk = itk_long_long;
512 for (; itk < itk_none; itk += 2 /* skip signed types */)
514 tree upper = TYPE_MAX_VALUE (integer_types[itk]);
516 if ((unsigned HOST_WIDE_INT) TREE_INT_CST_HIGH (upper) > high
517 || ((unsigned HOST_WIDE_INT) TREE_INT_CST_HIGH (upper) == high
518 && TREE_INT_CST_LOW (upper) >= low))
519 return (enum integer_type_kind) itk;
522 return itk_none;
525 /* Interpret TOKEN, an integer with FLAGS as classified by cpplib. */
526 static tree
527 interpret_integer (const cpp_token *token, unsigned int flags)
529 tree value, type;
530 enum integer_type_kind itk;
531 cpp_num integer;
532 cpp_options *options = cpp_get_options (parse_in);
534 integer = cpp_interpret_integer (parse_in, token, flags);
535 integer = cpp_num_sign_extend (integer, options->precision);
537 /* The type of a constant with a U suffix is straightforward. */
538 if (flags & CPP_N_UNSIGNED)
539 itk = narrowest_unsigned_type (integer.low, integer.high, flags);
540 else
542 /* The type of a potentially-signed integer constant varies
543 depending on the base it's in, the standard in use, and the
544 length suffixes. */
545 enum integer_type_kind itk_u
546 = narrowest_unsigned_type (integer.low, integer.high, flags);
547 enum integer_type_kind itk_s
548 = narrowest_signed_type (integer.low, integer.high, flags);
550 /* In both C89 and C99, octal and hex constants may be signed or
551 unsigned, whichever fits tighter. We do not warn about this
552 choice differing from the traditional choice, as the constant
553 is probably a bit pattern and either way will work. */
554 if ((flags & CPP_N_RADIX) != CPP_N_DECIMAL)
555 itk = MIN (itk_u, itk_s);
556 else
558 /* In C99, decimal constants are always signed.
559 In C89, decimal constants that don't fit in long have
560 undefined behavior; we try to make them unsigned long.
561 In GCC's extended C89, that last is true of decimal
562 constants that don't fit in long long, too. */
564 itk = itk_s;
565 if (itk_s > itk_u && itk_s > itk_long)
567 if (!flag_isoc99)
569 if (itk_u < itk_unsigned_long)
570 itk_u = itk_unsigned_long;
571 itk = itk_u;
572 warning (0, "this decimal constant is unsigned only in ISO C90");
574 else
575 warning (OPT_Wtraditional,
576 "this decimal constant would be unsigned in ISO C90");
581 if (itk == itk_none)
582 /* cpplib has already issued a warning for overflow. */
583 type = ((flags & CPP_N_UNSIGNED)
584 ? widest_unsigned_literal_type_node
585 : widest_integer_literal_type_node);
586 else
588 type = integer_types[itk];
589 if (itk > itk_unsigned_long
590 && (flags & CPP_N_WIDTH) != CPP_N_LARGE)
591 emit_diagnostic
592 ((c_dialect_cxx () ? cxx_dialect == cxx98 : !flag_isoc99)
593 ? DK_PEDWARN : DK_WARNING,
594 input_location, OPT_Wlong_long,
595 (flags & CPP_N_UNSIGNED)
596 ? "integer constant is too large for %<unsigned long%> type"
597 : "integer constant is too large for %<long%> type");
600 value = build_int_cst_wide (type, integer.low, integer.high);
602 /* Convert imaginary to a complex type. */
603 if (flags & CPP_N_IMAGINARY)
604 value = build_complex (NULL_TREE, build_int_cst (type, 0), value);
606 return value;
609 /* Interpret TOKEN, a floating point number with FLAGS as classified
610 by cpplib. */
611 static tree
612 interpret_float (const cpp_token *token, unsigned int flags)
614 tree type;
615 tree const_type;
616 tree value;
617 REAL_VALUE_TYPE real;
618 REAL_VALUE_TYPE real_trunc;
619 char *copy;
620 size_t copylen;
622 /* Default (no suffix) depends on whether the FLOAT_CONST_DECIMAL64
623 pragma has been used and is either double or _Decimal64. Types
624 that are not allowed with decimal float default to double. */
625 if (flags & CPP_N_DEFAULT)
627 flags ^= CPP_N_DEFAULT;
628 flags |= CPP_N_MEDIUM;
630 if (((flags & CPP_N_HEX) == 0) && ((flags & CPP_N_IMAGINARY) == 0))
632 warning (OPT_Wunsuffixed_float_constants,
633 "unsuffixed float constant");
634 if (float_const_decimal64_p ())
635 flags |= CPP_N_DFLOAT;
639 /* Decode _Fract and _Accum. */
640 if (flags & CPP_N_FRACT || flags & CPP_N_ACCUM)
641 return interpret_fixed (token, flags);
643 /* Decode type based on width and properties. */
644 if (flags & CPP_N_DFLOAT)
645 if ((flags & CPP_N_WIDTH) == CPP_N_LARGE)
646 type = dfloat128_type_node;
647 else if ((flags & CPP_N_WIDTH) == CPP_N_SMALL)
648 type = dfloat32_type_node;
649 else
650 type = dfloat64_type_node;
651 else
652 if (flags & CPP_N_WIDTH_MD)
654 char suffix;
655 enum machine_mode mode;
657 if ((flags & CPP_N_WIDTH_MD) == CPP_N_MD_W)
658 suffix = 'w';
659 else
660 suffix = 'q';
662 mode = targetm.c.mode_for_suffix (suffix);
663 if (mode == VOIDmode)
665 error ("unsupported non-standard suffix on floating constant");
666 errorcount++;
668 return error_mark_node;
670 else
671 pedwarn (input_location, OPT_pedantic, "non-standard suffix on floating constant");
673 type = c_common_type_for_mode (mode, 0);
674 gcc_assert (type);
676 else if ((flags & CPP_N_WIDTH) == CPP_N_LARGE)
677 type = long_double_type_node;
678 else if ((flags & CPP_N_WIDTH) == CPP_N_SMALL
679 || flag_single_precision_constant)
680 type = float_type_node;
681 else
682 type = double_type_node;
684 const_type = excess_precision_type (type);
685 if (!const_type)
686 const_type = type;
688 /* Copy the constant to a nul-terminated buffer. If the constant
689 has any suffixes, cut them off; REAL_VALUE_ATOF/ REAL_VALUE_HTOF
690 can't handle them. */
691 copylen = token->val.str.len;
692 if (flags & CPP_N_DFLOAT)
693 copylen -= 2;
694 else
696 if ((flags & CPP_N_WIDTH) != CPP_N_MEDIUM)
697 /* Must be an F or L or machine defined suffix. */
698 copylen--;
699 if (flags & CPP_N_IMAGINARY)
700 /* I or J suffix. */
701 copylen--;
704 copy = (char *) alloca (copylen + 1);
705 memcpy (copy, token->val.str.text, copylen);
706 copy[copylen] = '\0';
708 real_from_string3 (&real, copy, TYPE_MODE (const_type));
709 if (const_type != type)
710 /* Diagnosing if the result of converting the value with excess
711 precision to the semantic type would overflow (with associated
712 double rounding) is more appropriate than diagnosing if the
713 result of converting the string directly to the semantic type
714 would overflow. */
715 real_convert (&real_trunc, TYPE_MODE (type), &real);
717 /* Both C and C++ require a diagnostic for a floating constant
718 outside the range of representable values of its type. Since we
719 have __builtin_inf* to produce an infinity, this is now a
720 mandatory pedwarn if the target does not support infinities. */
721 if (REAL_VALUE_ISINF (real)
722 || (const_type != type && REAL_VALUE_ISINF (real_trunc)))
724 if (!MODE_HAS_INFINITIES (TYPE_MODE (type)))
725 pedwarn (input_location, 0, "floating constant exceeds range of %qT", type);
726 else
727 warning (OPT_Woverflow, "floating constant exceeds range of %qT", type);
729 /* We also give a warning if the value underflows. */
730 else if (REAL_VALUES_EQUAL (real, dconst0)
731 || (const_type != type && REAL_VALUES_EQUAL (real_trunc, dconst0)))
733 REAL_VALUE_TYPE realvoidmode;
734 int overflow = real_from_string (&realvoidmode, copy);
735 if (overflow < 0 || !REAL_VALUES_EQUAL (realvoidmode, dconst0))
736 warning (OPT_Woverflow, "floating constant truncated to zero");
739 /* Create a node with determined type and value. */
740 value = build_real (const_type, real);
741 if (flags & CPP_N_IMAGINARY)
742 value = build_complex (NULL_TREE, convert (const_type, integer_zero_node),
743 value);
745 if (type != const_type)
746 value = build1 (EXCESS_PRECISION_EXPR, type, value);
748 return value;
751 /* Interpret TOKEN, a fixed-point number with FLAGS as classified
752 by cpplib. */
754 static tree
755 interpret_fixed (const cpp_token *token, unsigned int flags)
757 tree type;
758 tree value;
759 FIXED_VALUE_TYPE fixed;
760 char *copy;
761 size_t copylen;
763 copylen = token->val.str.len;
765 if (flags & CPP_N_FRACT) /* _Fract. */
767 if (flags & CPP_N_UNSIGNED) /* Unsigned _Fract. */
769 if ((flags & CPP_N_WIDTH) == CPP_N_LARGE)
771 type = unsigned_long_long_fract_type_node;
772 copylen -= 4;
774 else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM)
776 type = unsigned_long_fract_type_node;
777 copylen -= 3;
779 else if ((flags & CPP_N_WIDTH) == CPP_N_SMALL)
781 type = unsigned_short_fract_type_node;
782 copylen -= 3;
784 else
786 type = unsigned_fract_type_node;
787 copylen -= 2;
790 else /* Signed _Fract. */
792 if ((flags & CPP_N_WIDTH) == CPP_N_LARGE)
794 type = long_long_fract_type_node;
795 copylen -= 3;
797 else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM)
799 type = long_fract_type_node;
800 copylen -= 2;
802 else if ((flags & CPP_N_WIDTH) == CPP_N_SMALL)
804 type = short_fract_type_node;
805 copylen -= 2;
807 else
809 type = fract_type_node;
810 copylen --;
814 else /* _Accum. */
816 if (flags & CPP_N_UNSIGNED) /* Unsigned _Accum. */
818 if ((flags & CPP_N_WIDTH) == CPP_N_LARGE)
820 type = unsigned_long_long_accum_type_node;
821 copylen -= 4;
823 else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM)
825 type = unsigned_long_accum_type_node;
826 copylen -= 3;
828 else if ((flags & CPP_N_WIDTH) == CPP_N_SMALL)
830 type = unsigned_short_accum_type_node;
831 copylen -= 3;
833 else
835 type = unsigned_accum_type_node;
836 copylen -= 2;
839 else /* Signed _Accum. */
841 if ((flags & CPP_N_WIDTH) == CPP_N_LARGE)
843 type = long_long_accum_type_node;
844 copylen -= 3;
846 else if ((flags & CPP_N_WIDTH) == CPP_N_MEDIUM)
848 type = long_accum_type_node;
849 copylen -= 2;
851 else if ((flags & CPP_N_WIDTH) == CPP_N_SMALL)
853 type = short_accum_type_node;
854 copylen -= 2;
856 else
858 type = accum_type_node;
859 copylen --;
864 copy = (char *) alloca (copylen + 1);
865 memcpy (copy, token->val.str.text, copylen);
866 copy[copylen] = '\0';
868 fixed_from_string (&fixed, copy, TYPE_MODE (type));
870 /* Create a node with determined type and value. */
871 value = build_fixed (type, fixed);
873 return value;
876 /* Convert a series of STRING, WSTRING, STRING16, STRING32 and/or
877 UTF8STRING tokens into a tree, performing string constant
878 concatenation. TOK is the first of these. VALP is the location
879 to write the string into. OBJC_STRING indicates whether an '@' token
880 preceded the incoming token.
881 Returns the CPP token type of the result (CPP_STRING, CPP_WSTRING,
882 CPP_STRING32, CPP_STRING16, CPP_UTF8STRING, or CPP_OBJC_STRING).
884 This is unfortunately more work than it should be. If any of the
885 strings in the series has an L prefix, the result is a wide string
886 (6.4.5p4). Whether or not the result is a wide string affects the
887 meaning of octal and hexadecimal escapes (6.4.4.4p6,9). But escape
888 sequences do not continue across the boundary between two strings in
889 a series (6.4.5p7), so we must not lose the boundaries. Therefore
890 cpp_interpret_string takes a vector of cpp_string structures, which
891 we must arrange to provide. */
893 static enum cpp_ttype
894 lex_string (const cpp_token *tok, tree *valp, bool objc_string, bool translate)
896 tree value;
897 size_t concats = 0;
898 struct obstack str_ob;
899 cpp_string istr;
900 enum cpp_ttype type = tok->type;
902 /* Try to avoid the overhead of creating and destroying an obstack
903 for the common case of just one string. */
904 cpp_string str = tok->val.str;
905 cpp_string *strs = &str;
907 retry:
908 tok = cpp_get_token (parse_in);
909 switch (tok->type)
911 case CPP_PADDING:
912 goto retry;
913 case CPP_ATSIGN:
914 if (c_dialect_objc ())
916 objc_string = true;
917 goto retry;
919 /* FALLTHROUGH */
921 default:
922 break;
924 case CPP_WSTRING:
925 case CPP_STRING16:
926 case CPP_STRING32:
927 case CPP_UTF8STRING:
928 if (type != tok->type)
930 if (type == CPP_STRING)
931 type = tok->type;
932 else
933 error ("unsupported non-standard concatenation of string literals");
936 case CPP_STRING:
937 if (!concats)
939 gcc_obstack_init (&str_ob);
940 obstack_grow (&str_ob, &str, sizeof (cpp_string));
943 concats++;
944 obstack_grow (&str_ob, &tok->val.str, sizeof (cpp_string));
945 goto retry;
948 /* We have read one more token than we want. */
949 _cpp_backup_tokens (parse_in, 1);
950 if (concats)
951 strs = XOBFINISH (&str_ob, cpp_string *);
953 if (concats && !objc_string && !in_system_header)
954 warning (OPT_Wtraditional,
955 "traditional C rejects string constant concatenation");
957 if ((translate
958 ? cpp_interpret_string : cpp_interpret_string_notranslate)
959 (parse_in, strs, concats + 1, &istr, type))
961 value = build_string (istr.len, (const char *) istr.text);
962 free (CONST_CAST (unsigned char *, istr.text));
964 else
966 /* Callers cannot generally handle error_mark_node in this context,
967 so return the empty string instead. cpp_interpret_string has
968 issued an error. */
969 switch (type)
971 default:
972 case CPP_STRING:
973 case CPP_UTF8STRING:
974 value = build_string (1, "");
975 break;
976 case CPP_STRING16:
977 value = build_string (TYPE_PRECISION (char16_type_node)
978 / TYPE_PRECISION (char_type_node),
979 "\0"); /* char16_t is 16 bits */
980 break;
981 case CPP_STRING32:
982 value = build_string (TYPE_PRECISION (char32_type_node)
983 / TYPE_PRECISION (char_type_node),
984 "\0\0\0"); /* char32_t is 32 bits */
985 break;
986 case CPP_WSTRING:
987 value = build_string (TYPE_PRECISION (wchar_type_node)
988 / TYPE_PRECISION (char_type_node),
989 "\0\0\0"); /* widest supported wchar_t
990 is 32 bits */
991 break;
995 switch (type)
997 default:
998 case CPP_STRING:
999 case CPP_UTF8STRING:
1000 TREE_TYPE (value) = char_array_type_node;
1001 break;
1002 case CPP_STRING16:
1003 TREE_TYPE (value) = char16_array_type_node;
1004 break;
1005 case CPP_STRING32:
1006 TREE_TYPE (value) = char32_array_type_node;
1007 break;
1008 case CPP_WSTRING:
1009 TREE_TYPE (value) = wchar_array_type_node;
1011 *valp = fix_string_type (value);
1013 if (concats)
1014 obstack_free (&str_ob, 0);
1016 return objc_string ? CPP_OBJC_STRING : type;
1019 /* Converts a (possibly wide) character constant token into a tree. */
1020 static tree
1021 lex_charconst (const cpp_token *token)
1023 cppchar_t result;
1024 tree type, value;
1025 unsigned int chars_seen;
1026 int unsignedp = 0;
1028 result = cpp_interpret_charconst (parse_in, token,
1029 &chars_seen, &unsignedp);
1031 if (token->type == CPP_WCHAR)
1032 type = wchar_type_node;
1033 else if (token->type == CPP_CHAR32)
1034 type = char32_type_node;
1035 else if (token->type == CPP_CHAR16)
1036 type = char16_type_node;
1037 /* In C, a character constant has type 'int'.
1038 In C++ 'char', but multi-char charconsts have type 'int'. */
1039 else if (!c_dialect_cxx () || chars_seen > 1)
1040 type = integer_type_node;
1041 else
1042 type = char_type_node;
1044 /* Cast to cppchar_signed_t to get correct sign-extension of RESULT
1045 before possibly widening to HOST_WIDE_INT for build_int_cst. */
1046 if (unsignedp || (cppchar_signed_t) result >= 0)
1047 value = build_int_cst_wide (type, result, 0);
1048 else
1049 value = build_int_cst_wide (type, (cppchar_signed_t) result, -1);
1051 return value;