2 Copyright (C) 1986, 1987, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
3 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2007, 2009, 2010
4 Free Software Foundation, Inc.
5 Contributed by Per Bothner, 1994-95.
6 Based on CCCP program by Paul Rubin, June 1986
7 Adapted to ANSI C, Richard Stallman, Jan 1987
9 This program is free software; you can redistribute it and/or modify it
10 under the terms of the GNU General Public License as published by the
11 Free Software Foundation; either version 3, or (at your option) any
14 This program is distributed in the hope that it will be useful,
15 but WITHOUT ANY WARRANTY; without even the implied warranty of
16 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 GNU General Public License for more details.
19 You should have received a copy of the GNU General Public License
20 along with this program; see the file COPYING3. If not see
21 <http://www.gnu.org/licenses/>. */
28 #include "localedir.h"
30 static void init_library (void);
31 static void mark_named_operators (cpp_reader *);
32 static void read_original_filename (cpp_reader *);
33 static void read_original_directory (cpp_reader *);
34 static void post_options (cpp_reader *);
36 /* If we have designated initializers (GCC >2.7) these tables can be
37 initialized, constant data. Otherwise, they have to be filled in at
39 #if HAVE_DESIGNATED_INITIALIZERS
41 #define init_trigraph_map() /* Nothing. */
42 #define TRIGRAPH_MAP \
43 __extension__ const uchar _cpp_trigraph_map[UCHAR_MAX + 1] = {
46 #define s(p, v) [p] = v,
50 #define TRIGRAPH_MAP uchar _cpp_trigraph_map[UCHAR_MAX + 1] = { 0 }; \
51 static void init_trigraph_map (void) { \
52 unsigned char *x = _cpp_trigraph_map;
55 #define s(p, v) x[p] = v;
60 s('=', '#') s(')', ']') s('!', '|')
61 s('(', '[') s('\'', '^') s('>', '}')
62 s('/', '\\') s('<', '{') s('-', '~')
69 /* A set of booleans indicating what CPP features each source language
75 char extended_numbers;
76 char extended_identifiers;
78 char cplusplus_comments;
83 static const struct lang_flags lang_defaults[] =
84 { /* c99 c++ xnum xid std // digr ulit */
85 /* GNUC89 */ { 0, 0, 1, 0, 0, 1, 1, 0 },
86 /* GNUC99 */ { 1, 0, 1, 0, 0, 1, 1, 1 },
87 /* STDC89 */ { 0, 0, 0, 0, 1, 0, 0, 0 },
88 /* STDC94 */ { 0, 0, 0, 0, 1, 0, 1, 0 },
89 /* STDC99 */ { 1, 0, 1, 0, 1, 1, 1, 0 },
90 /* GNUCXX */ { 0, 1, 1, 0, 0, 1, 1, 0 },
91 /* CXX98 */ { 0, 1, 1, 0, 1, 1, 1, 0 },
92 /* GNUCXX0X */ { 1, 1, 1, 0, 0, 1, 1, 1 },
93 /* CXX0X */ { 1, 1, 1, 0, 1, 1, 1, 1 },
94 /* ASM */ { 0, 0, 1, 0, 0, 1, 0, 0 }
95 /* xid should be 1 for GNUC99, STDC99, GNUCXX, CXX98, GNUCXX0X, and
96 CXX0X when no longer experimental (when all uses of identifiers
97 in the compiler have been audited for correct handling of
98 extended identifiers). */
101 /* Sets internal flags correctly for a given language. */
103 cpp_set_lang (cpp_reader *pfile, enum c_lang lang)
105 const struct lang_flags *l = &lang_defaults[(int) lang];
107 CPP_OPTION (pfile, lang) = lang;
109 CPP_OPTION (pfile, c99) = l->c99;
110 CPP_OPTION (pfile, cplusplus) = l->cplusplus;
111 CPP_OPTION (pfile, extended_numbers) = l->extended_numbers;
112 CPP_OPTION (pfile, extended_identifiers) = l->extended_identifiers;
113 CPP_OPTION (pfile, std) = l->std;
114 CPP_OPTION (pfile, trigraphs) = l->std;
115 CPP_OPTION (pfile, cplusplus_comments) = l->cplusplus_comments;
116 CPP_OPTION (pfile, digraphs) = l->digraphs;
117 CPP_OPTION (pfile, uliterals) = l->uliterals;
120 /* Initialize library global state. */
124 static int initialized = 0;
130 /* Set up the trigraph map. This doesn't need to do anything if
131 we were compiled with a compiler that supports C99 designated
133 init_trigraph_map ();
136 (void) bindtextdomain (PACKAGE, LOCALEDIR);
141 /* Initialize a cpp_reader structure. */
143 cpp_create_reader (enum c_lang lang, hash_table *table,
144 struct line_maps *line_table)
148 /* Initialize this instance of the library if it hasn't been already. */
151 pfile = XCNEW (cpp_reader);
153 cpp_set_lang (pfile, lang);
154 CPP_OPTION (pfile, warn_multichar) = 1;
155 CPP_OPTION (pfile, discard_comments) = 1;
156 CPP_OPTION (pfile, discard_comments_in_macro_exp) = 1;
157 CPP_OPTION (pfile, show_column) = 1;
158 CPP_OPTION (pfile, tabstop) = 8;
159 CPP_OPTION (pfile, operator_names) = 1;
160 CPP_OPTION (pfile, warn_trigraphs) = 2;
161 CPP_OPTION (pfile, warn_endif_labels) = 1;
162 CPP_OPTION (pfile, warn_deprecated) = 1;
163 CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
164 CPP_OPTION (pfile, dollars_in_ident) = 1;
165 CPP_OPTION (pfile, warn_dollars) = 1;
166 CPP_OPTION (pfile, warn_variadic_macros) = 1;
167 CPP_OPTION (pfile, warn_builtin_macro_redefined) = 1;
168 CPP_OPTION (pfile, warn_normalize) = normalized_C;
170 /* Default CPP arithmetic to something sensible for the host for the
171 benefit of dumb users like fix-header. */
172 CPP_OPTION (pfile, precision) = CHAR_BIT * sizeof (long);
173 CPP_OPTION (pfile, char_precision) = CHAR_BIT;
174 CPP_OPTION (pfile, wchar_precision) = CHAR_BIT * sizeof (int);
175 CPP_OPTION (pfile, int_precision) = CHAR_BIT * sizeof (int);
176 CPP_OPTION (pfile, unsigned_char) = 0;
177 CPP_OPTION (pfile, unsigned_wchar) = 1;
178 CPP_OPTION (pfile, bytes_big_endian) = 1; /* does not matter */
180 /* Default to no charset conversion. */
181 CPP_OPTION (pfile, narrow_charset) = _cpp_default_encoding ();
182 CPP_OPTION (pfile, wide_charset) = 0;
184 /* Default the input character set to UTF-8. */
185 CPP_OPTION (pfile, input_charset) = _cpp_default_encoding ();
187 /* A fake empty "directory" used as the starting point for files
188 looked up without a search path. Name cannot be '/' because we
189 don't want to prepend anything at all to filenames using it. All
190 other entries are correct zero-initialized. */
191 pfile->no_search_path.name = (char *) "";
193 /* Initialize the line map. */
194 pfile->line_table = line_table;
196 /* Initialize lexer state. */
197 pfile->state.save_comments = ! CPP_OPTION (pfile, discard_comments);
199 /* Set up static tokens. */
200 pfile->avoid_paste.type = CPP_PADDING;
201 pfile->avoid_paste.val.source = NULL;
202 pfile->eof.type = CPP_EOF;
203 pfile->eof.flags = 0;
205 /* Create a token buffer for the lexer. */
206 _cpp_init_tokenrun (&pfile->base_run, 250);
207 pfile->cur_run = &pfile->base_run;
208 pfile->cur_token = pfile->base_run.base;
210 /* Initialize the base context. */
211 pfile->context = &pfile->base_context;
212 pfile->base_context.macro = 0;
213 pfile->base_context.prev = pfile->base_context.next = 0;
215 /* Aligned and unaligned storage. */
216 pfile->a_buff = _cpp_get_buff (pfile, 0);
217 pfile->u_buff = _cpp_get_buff (pfile, 0);
219 /* Initialize table for push_macro/pop_macro. */
220 pfile->pushed_macros = 0;
222 /* The expression parser stack. */
223 _cpp_expand_op_stack (pfile);
225 /* Initialize the buffer obstack. */
226 _obstack_begin (&pfile->buffer_ob, 0, 0,
227 (void *(*) (long)) xmalloc,
228 (void (*) (void *)) free);
230 _cpp_init_files (pfile);
232 _cpp_init_hashtable (pfile, table);
237 /* Set the line_table entry in PFILE. This is called after reading a
238 PCH file, as the old line_table will be incorrect. */
240 cpp_set_line_map (cpp_reader *pfile, struct line_maps *line_table)
242 pfile->line_table = line_table;
245 /* Free resources used by PFILE. Accessing PFILE after this function
246 returns leads to undefined behavior. Returns the error count. */
248 cpp_destroy (cpp_reader *pfile)
250 cpp_context *context, *contextn;
251 struct def_pragma_macro *pmacro;
252 tokenrun *run, *runn;
255 free (pfile->op_stack);
257 while (CPP_BUFFER (pfile) != NULL)
258 _cpp_pop_buffer (pfile);
261 free (pfile->out.base);
263 if (pfile->macro_buffer)
265 free (pfile->macro_buffer);
266 pfile->macro_buffer = NULL;
267 pfile->macro_buffer_len = 0;
271 deps_free (pfile->deps);
272 obstack_free (&pfile->buffer_ob, 0);
274 _cpp_destroy_hashtable (pfile);
275 _cpp_cleanup_files (pfile);
276 _cpp_destroy_iconv (pfile);
278 _cpp_free_buff (pfile->a_buff);
279 _cpp_free_buff (pfile->u_buff);
280 _cpp_free_buff (pfile->free_buffs);
282 for (run = &pfile->base_run; run; run = runn)
286 if (run != &pfile->base_run)
290 for (context = pfile->base_context.next; context; context = contextn)
292 contextn = context->next;
296 if (pfile->comments.entries)
298 for (i = 0; i < pfile->comments.count; i++)
299 free (pfile->comments.entries[i].comment);
301 free (pfile->comments.entries);
303 if (pfile->pushed_macros)
307 pmacro = pfile->pushed_macros;
308 pfile->pushed_macros = pmacro->next;
312 while (pfile->pushed_macros);
318 /* This structure defines one built-in identifier. A node will be
319 entered in the hash table under the name NAME, with value VALUE.
321 There are two tables of these. builtin_array holds all the
322 "builtin" macros: these are handled by builtin_macro() in
323 macro.c. Builtin is somewhat of a misnomer -- the property of
324 interest is that these macros require special code to compute their
325 expansions. The value is a "builtin_type" enumerator.
327 operator_array holds the C++ named operators. These are keywords
328 which act as aliases for punctuators. In C++, they cannot be
329 altered through #define, and #if recognizes them as operators. In
330 C, these are not entered into the hash table at all (but see
331 <iso646.h>). The value is a token-type enumerator. */
334 const uchar *const name;
335 const unsigned short len;
336 const unsigned short value;
337 const bool always_warn_if_redefined;
340 #define B(n, t, f) { DSC(n), t, f }
341 static const struct builtin_macro builtin_array[] =
343 B("__TIMESTAMP__", BT_TIMESTAMP, false),
344 B("__TIME__", BT_TIME, false),
345 B("__DATE__", BT_DATE, false),
346 B("__FILE__", BT_FILE, false),
347 B("__BASE_FILE__", BT_BASE_FILE, false),
348 B("__LINE__", BT_SPECLINE, true),
349 B("__INCLUDE_LEVEL__", BT_INCLUDE_LEVEL, true),
350 B("__COUNTER__", BT_COUNTER, true),
351 /* Keep builtins not used for -traditional-cpp at the end, and
352 update init_builtins() if any more are added. */
353 B("_Pragma", BT_PRAGMA, true),
354 B("__STDC__", BT_STDC, true),
358 struct builtin_operator
360 const uchar *const name;
361 const unsigned short len;
362 const unsigned short value;
365 #define B(n, t) { DSC(n), t }
366 static const struct builtin_operator operator_array[] =
368 B("and", CPP_AND_AND),
369 B("and_eq", CPP_AND_EQ),
370 B("bitand", CPP_AND),
372 B("compl", CPP_COMPL),
374 B("not_eq", CPP_NOT_EQ),
376 B("or_eq", CPP_OR_EQ),
378 B("xor_eq", CPP_XOR_EQ)
382 /* Mark the C++ named operators in the hash table. */
384 mark_named_operators (cpp_reader *pfile)
386 const struct builtin_operator *b;
388 for (b = operator_array;
389 b < (operator_array + ARRAY_SIZE (operator_array));
392 cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
393 hp->flags |= NODE_OPERATOR;
394 hp->is_directive = 0;
395 hp->directive_index = b->value;
400 cpp_init_special_builtins (cpp_reader *pfile)
402 const struct builtin_macro *b;
403 size_t n = ARRAY_SIZE (builtin_array);
405 if (CPP_OPTION (pfile, traditional))
407 else if (! CPP_OPTION (pfile, stdc_0_in_system_headers)
408 || CPP_OPTION (pfile, std))
411 for (b = builtin_array; b < builtin_array + n; b++)
413 cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
415 hp->flags |= NODE_BUILTIN;
416 if (b->always_warn_if_redefined
417 || CPP_OPTION (pfile, warn_builtin_macro_redefined))
418 hp->flags |= NODE_WARN;
419 hp->value.builtin = (enum builtin_type) b->value;
423 /* Read the builtins table above and enter them, and language-specific
424 macros, into the hash table. HOSTED is true if this is a hosted
427 cpp_init_builtins (cpp_reader *pfile, int hosted)
429 cpp_init_special_builtins (pfile);
431 if (!CPP_OPTION (pfile, traditional)
432 && (! CPP_OPTION (pfile, stdc_0_in_system_headers)
433 || CPP_OPTION (pfile, std)))
434 _cpp_define_builtin (pfile, "__STDC__ 1");
436 if (CPP_OPTION (pfile, cplusplus))
437 _cpp_define_builtin (pfile, "__cplusplus 1");
438 else if (CPP_OPTION (pfile, lang) == CLK_ASM)
439 _cpp_define_builtin (pfile, "__ASSEMBLER__ 1");
440 else if (CPP_OPTION (pfile, lang) == CLK_STDC94)
441 _cpp_define_builtin (pfile, "__STDC_VERSION__ 199409L");
442 else if (CPP_OPTION (pfile, c99))
443 _cpp_define_builtin (pfile, "__STDC_VERSION__ 199901L");
446 _cpp_define_builtin (pfile, "__STDC_HOSTED__ 1");
448 _cpp_define_builtin (pfile, "__STDC_HOSTED__ 0");
450 if (CPP_OPTION (pfile, objc))
451 _cpp_define_builtin (pfile, "__OBJC__ 1");
454 /* Sanity-checks are dependent on command-line options, so it is
455 called as a subroutine of cpp_read_main_file (). */
457 static void sanity_checks (cpp_reader *);
458 static void sanity_checks (cpp_reader *pfile)
461 size_t max_precision = 2 * CHAR_BIT * sizeof (cpp_num_part);
463 /* Sanity checks for assumptions about CPP arithmetic and target
464 type precisions made by cpplib. */
467 cpp_error (pfile, CPP_DL_ICE, "cppchar_t must be an unsigned type");
469 if (CPP_OPTION (pfile, precision) > max_precision)
470 cpp_error (pfile, CPP_DL_ICE,
471 "preprocessor arithmetic has maximum precision of %lu bits;"
472 " target requires %lu bits",
473 (unsigned long) max_precision,
474 (unsigned long) CPP_OPTION (pfile, precision));
476 if (CPP_OPTION (pfile, precision) < CPP_OPTION (pfile, int_precision))
477 cpp_error (pfile, CPP_DL_ICE,
478 "CPP arithmetic must be at least as precise as a target int");
480 if (CPP_OPTION (pfile, char_precision) < 8)
481 cpp_error (pfile, CPP_DL_ICE, "target char is less than 8 bits wide");
483 if (CPP_OPTION (pfile, wchar_precision) < CPP_OPTION (pfile, char_precision))
484 cpp_error (pfile, CPP_DL_ICE,
485 "target wchar_t is narrower than target char");
487 if (CPP_OPTION (pfile, int_precision) < CPP_OPTION (pfile, char_precision))
488 cpp_error (pfile, CPP_DL_ICE,
489 "target int is narrower than target char");
491 /* This is assumed in eval_token() and could be fixed if necessary. */
492 if (sizeof (cppchar_t) > sizeof (cpp_num_part))
493 cpp_error (pfile, CPP_DL_ICE,
494 "CPP half-integer narrower than CPP character");
496 if (CPP_OPTION (pfile, wchar_precision) > BITS_PER_CPPCHAR_T)
497 cpp_error (pfile, CPP_DL_ICE,
498 "CPP on this host cannot handle wide character constants over"
499 " %lu bits, but the target requires %lu bits",
500 (unsigned long) BITS_PER_CPPCHAR_T,
501 (unsigned long) CPP_OPTION (pfile, wchar_precision));
504 # define sanity_checks(PFILE)
507 /* This is called after options have been parsed, and partially
510 cpp_post_options (cpp_reader *pfile)
512 sanity_checks (pfile);
514 post_options (pfile);
516 /* Mark named operators before handling command line macros. */
517 if (CPP_OPTION (pfile, cplusplus) && CPP_OPTION (pfile, operator_names))
518 mark_named_operators (pfile);
521 /* Setup for processing input from the file named FNAME, or stdin if
522 it is the empty string. Return the original filename
523 on success (e.g. foo.i->foo.c), or NULL on failure. */
525 cpp_read_main_file (cpp_reader *pfile, const char *fname)
527 if (CPP_OPTION (pfile, deps.style) != DEPS_NONE)
530 pfile->deps = deps_init ();
532 /* Set the default target (if there is none already). */
533 deps_add_default_target (pfile->deps, fname);
537 = _cpp_find_file (pfile, fname, &pfile->no_search_path, false, 0);
538 if (_cpp_find_failed (pfile->main_file))
541 _cpp_stack_file (pfile, pfile->main_file, false);
543 /* For foo.i, read the original filename foo.c now, for the benefit
544 of the front ends. */
545 if (CPP_OPTION (pfile, preprocessed))
547 read_original_filename (pfile);
548 fname = pfile->line_table->maps[pfile->line_table->used-1].to_file;
553 /* For preprocessed files, if the first tokens are of the form # NUM.
554 handle the directive so we know the original file name. This will
555 generate file_change callbacks, which the front ends must handle
556 appropriately given their state of initialization. */
558 read_original_filename (cpp_reader *pfile)
560 const cpp_token *token, *token1;
562 /* Lex ahead; if the first tokens are of the form # NUM, then
563 process the directive, otherwise back up. */
564 token = _cpp_lex_direct (pfile);
565 if (token->type == CPP_HASH)
567 pfile->state.in_directive = 1;
568 token1 = _cpp_lex_direct (pfile);
569 _cpp_backup_tokens (pfile, 1);
570 pfile->state.in_directive = 0;
572 /* If it's a #line directive, handle it. */
573 if (token1->type == CPP_NUMBER
574 && _cpp_handle_directive (pfile, token->flags & PREV_WHITE))
576 read_original_directory (pfile);
581 /* Backup as if nothing happened. */
582 _cpp_backup_tokens (pfile, 1);
585 /* For preprocessed files, if the tokens following the first filename
586 line is of the form # <line> "/path/name//", handle the
587 directive so we know the original current directory. */
589 read_original_directory (cpp_reader *pfile)
591 const cpp_token *hash, *token;
593 /* Lex ahead; if the first tokens are of the form # NUM, then
594 process the directive, otherwise back up. */
595 hash = _cpp_lex_direct (pfile);
596 if (hash->type != CPP_HASH)
598 _cpp_backup_tokens (pfile, 1);
602 token = _cpp_lex_direct (pfile);
604 if (token->type != CPP_NUMBER)
606 _cpp_backup_tokens (pfile, 2);
610 token = _cpp_lex_direct (pfile);
612 if (token->type != CPP_STRING
613 || ! (token->val.str.len >= 5
614 && token->val.str.text[token->val.str.len-2] == '/'
615 && token->val.str.text[token->val.str.len-3] == '/'))
617 _cpp_backup_tokens (pfile, 3);
621 if (pfile->cb.dir_change)
623 char *debugdir = (char *) alloca (token->val.str.len - 3);
625 memcpy (debugdir, (const char *) token->val.str.text + 1,
626 token->val.str.len - 4);
627 debugdir[token->val.str.len - 4] = '\0';
629 pfile->cb.dir_change (pfile, debugdir);
633 /* This is called at the end of preprocessing. It pops the last
634 buffer and writes dependency output, and returns the number of
637 Maybe it should also reset state, such that you could call
638 cpp_start_read with a new filename to restart processing. */
640 cpp_finish (cpp_reader *pfile, FILE *deps_stream)
642 /* Warn about unused macros before popping the final buffer. */
643 if (CPP_OPTION (pfile, warn_unused_macros))
644 cpp_forall_identifiers (pfile, _cpp_warn_if_unused_macro, NULL);
646 /* lex.c leaves the final buffer on the stack. This it so that
647 it returns an unending stream of CPP_EOFs to the client. If we
648 popped the buffer, we'd dereference a NULL buffer pointer and
649 segfault. It's nice to allow the client to do worry-free excess
650 cpp_get_token calls. */
651 while (pfile->buffer)
652 _cpp_pop_buffer (pfile);
654 /* Don't write the deps file if there are errors. */
655 if (CPP_OPTION (pfile, deps.style) != DEPS_NONE
656 && deps_stream && pfile->errors == 0)
658 deps_write (pfile->deps, deps_stream, 72);
660 if (CPP_OPTION (pfile, deps.phony_targets))
661 deps_phony_targets (pfile->deps, deps_stream);
664 /* Report on headers that could use multiple include guards. */
665 if (CPP_OPTION (pfile, print_include_names))
666 _cpp_report_missing_guards (pfile);
668 return pfile->errors;
672 post_options (cpp_reader *pfile)
674 /* -Wtraditional is not useful in C++ mode. */
675 if (CPP_OPTION (pfile, cplusplus))
676 CPP_OPTION (pfile, warn_traditional) = 0;
678 /* Permanently disable macro expansion if we are rescanning
679 preprocessed text. Read preprocesed source in ISO mode. */
680 if (CPP_OPTION (pfile, preprocessed))
682 if (!CPP_OPTION (pfile, directives_only))
683 pfile->state.prevent_expansion = 1;
684 CPP_OPTION (pfile, traditional) = 0;
687 if (CPP_OPTION (pfile, warn_trigraphs) == 2)
688 CPP_OPTION (pfile, warn_trigraphs) = !CPP_OPTION (pfile, trigraphs);
690 if (CPP_OPTION (pfile, traditional))
692 CPP_OPTION (pfile, cplusplus_comments) = 0;
694 /* Traditional CPP does not accurately track column information. */
695 CPP_OPTION (pfile, show_column) = 0;
696 CPP_OPTION (pfile, trigraphs) = 0;
697 CPP_OPTION (pfile, warn_trigraphs) = 0;