2 Copyright (C) 1986, 1987, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
3 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2007, 2009
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/>. */
29 #include "localedir.h"
32 static void init_library (void);
33 static void mark_named_operators (cpp_reader *);
34 static void read_original_filename (cpp_reader *);
35 static void read_original_directory (cpp_reader *);
36 static void post_options (cpp_reader *);
38 /* If we have designated initializers (GCC >2.7) these tables can be
39 initialized, constant data. Otherwise, they have to be filled in at
41 #if HAVE_DESIGNATED_INITIALIZERS
43 #define init_trigraph_map() /* Nothing. */
44 #define TRIGRAPH_MAP \
45 __extension__ const uchar _cpp_trigraph_map[UCHAR_MAX + 1] = {
48 #define s(p, v) [p] = v,
52 #define TRIGRAPH_MAP uchar _cpp_trigraph_map[UCHAR_MAX + 1] = { 0 }; \
53 static void init_trigraph_map (void) { \
54 unsigned char *x = _cpp_trigraph_map;
57 #define s(p, v) x[p] = v;
62 s('=', '#') s(')', ']') s('!', '|')
63 s('(', '[') s('\'', '^') s('>', '}')
64 s('/', '\\') s('<', '{') s('-', '~')
71 /* A set of booleans indicating what CPP features each source language
77 char extended_numbers;
78 char extended_identifiers;
80 char cplusplus_comments;
85 static const struct lang_flags lang_defaults[] =
86 { /* c99 c++ xnum xid std // digr ulit */
87 /* GNUC89 */ { 0, 0, 1, 0, 0, 1, 1, 0 },
88 /* GNUC99 */ { 1, 0, 1, 0, 0, 1, 1, 1 },
89 /* STDC89 */ { 0, 0, 0, 0, 1, 0, 0, 0 },
90 /* STDC94 */ { 0, 0, 0, 0, 1, 0, 1, 0 },
91 /* STDC99 */ { 1, 0, 1, 0, 1, 1, 1, 0 },
92 /* GNUCXX */ { 0, 1, 1, 0, 0, 1, 1, 0 },
93 /* CXX98 */ { 0, 1, 1, 0, 1, 1, 1, 0 },
94 /* GNUCXX0X */ { 1, 1, 1, 0, 0, 1, 1, 1 },
95 /* CXX0X */ { 1, 1, 1, 0, 1, 1, 1, 1 },
96 /* ASM */ { 0, 0, 1, 0, 0, 1, 0, 0 }
97 /* xid should be 1 for GNUC99, STDC99, GNUCXX, CXX98, GNUCXX0X, and
98 CXX0X when no longer experimental (when all uses of identifiers
99 in the compiler have been audited for correct handling of
100 extended identifiers). */
103 /* Sets internal flags correctly for a given language. */
105 cpp_set_lang (cpp_reader *pfile, enum c_lang lang)
107 const struct lang_flags *l = &lang_defaults[(int) lang];
109 CPP_OPTION (pfile, lang) = lang;
111 CPP_OPTION (pfile, c99) = l->c99;
112 CPP_OPTION (pfile, cplusplus) = l->cplusplus;
113 CPP_OPTION (pfile, extended_numbers) = l->extended_numbers;
114 CPP_OPTION (pfile, extended_identifiers) = l->extended_identifiers;
115 CPP_OPTION (pfile, std) = l->std;
116 CPP_OPTION (pfile, trigraphs) = l->std;
117 CPP_OPTION (pfile, cplusplus_comments) = l->cplusplus_comments;
118 CPP_OPTION (pfile, digraphs) = l->digraphs;
119 CPP_OPTION (pfile, uliterals) = l->uliterals;
122 /* Initialize library global state. */
126 static int initialized = 0;
132 /* Set up the trigraph map. This doesn't need to do anything if
133 we were compiled with a compiler that supports C99 designated
135 init_trigraph_map ();
138 (void) bindtextdomain (PACKAGE, LOCALEDIR);
143 /* Initialize a cpp_reader structure. */
145 cpp_create_reader (enum c_lang lang, hash_table *table,
146 struct line_maps *line_table)
150 /* Initialize this instance of the library if it hasn't been already. */
153 pfile = XCNEW (cpp_reader);
155 cpp_set_lang (pfile, lang);
156 CPP_OPTION (pfile, warn_multichar) = 1;
157 CPP_OPTION (pfile, discard_comments) = 1;
158 CPP_OPTION (pfile, discard_comments_in_macro_exp) = 1;
159 CPP_OPTION (pfile, show_column) = 1;
160 CPP_OPTION (pfile, tabstop) = 8;
161 CPP_OPTION (pfile, operator_names) = 1;
162 CPP_OPTION (pfile, warn_trigraphs) = 2;
163 CPP_OPTION (pfile, warn_endif_labels) = 1;
164 CPP_OPTION (pfile, warn_deprecated) = 1;
165 CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
166 CPP_OPTION (pfile, dollars_in_ident) = 1;
167 CPP_OPTION (pfile, warn_dollars) = 1;
168 CPP_OPTION (pfile, warn_variadic_macros) = 1;
169 CPP_OPTION (pfile, warn_builtin_macro_redefined) = 1;
170 CPP_OPTION (pfile, warn_normalize) = normalized_C;
172 /* Default CPP arithmetic to something sensible for the host for the
173 benefit of dumb users like fix-header. */
174 CPP_OPTION (pfile, precision) = CHAR_BIT * sizeof (long);
175 CPP_OPTION (pfile, char_precision) = CHAR_BIT;
176 CPP_OPTION (pfile, wchar_precision) = CHAR_BIT * sizeof (int);
177 CPP_OPTION (pfile, int_precision) = CHAR_BIT * sizeof (int);
178 CPP_OPTION (pfile, unsigned_char) = 0;
179 CPP_OPTION (pfile, unsigned_wchar) = 1;
180 CPP_OPTION (pfile, bytes_big_endian) = 1; /* does not matter */
182 /* Default to no charset conversion. */
183 CPP_OPTION (pfile, narrow_charset) = _cpp_default_encoding ();
184 CPP_OPTION (pfile, wide_charset) = 0;
186 /* Default the input character set to UTF-8. */
187 CPP_OPTION (pfile, input_charset) = _cpp_default_encoding ();
189 /* A fake empty "directory" used as the starting point for files
190 looked up without a search path. Name cannot be '/' because we
191 don't want to prepend anything at all to filenames using it. All
192 other entries are correct zero-initialized. */
193 pfile->no_search_path.name = (char *) "";
195 /* Initialize the line map. */
196 pfile->line_table = line_table;
198 /* Initialize lexer state. */
199 pfile->state.save_comments = ! CPP_OPTION (pfile, discard_comments);
201 /* Set up static tokens. */
202 pfile->avoid_paste.type = CPP_PADDING;
203 pfile->avoid_paste.val.source = NULL;
204 pfile->eof.type = CPP_EOF;
205 pfile->eof.flags = 0;
207 /* Create a token buffer for the lexer. */
208 _cpp_init_tokenrun (&pfile->base_run, 250);
209 pfile->cur_run = &pfile->base_run;
210 pfile->cur_token = pfile->base_run.base;
212 /* Initialize the base context. */
213 pfile->context = &pfile->base_context;
214 pfile->base_context.macro = 0;
215 pfile->base_context.prev = pfile->base_context.next = 0;
217 /* Aligned and unaligned storage. */
218 pfile->a_buff = _cpp_get_buff (pfile, 0);
219 pfile->u_buff = _cpp_get_buff (pfile, 0);
221 /* The expression parser stack. */
222 _cpp_expand_op_stack (pfile);
224 /* Initialize the buffer obstack. */
225 _obstack_begin (&pfile->buffer_ob, 0, 0,
226 (void *(*) (long)) xmalloc,
227 (void (*) (void *)) free);
229 _cpp_init_files (pfile);
231 _cpp_init_hashtable (pfile, table);
236 /* Set the line_table entry in PFILE. This is called after reading a
237 PCH file, as the old line_table will be incorrect. */
239 cpp_set_line_map (cpp_reader *pfile, struct line_maps *line_table)
241 pfile->line_table = line_table;
244 /* Free resources used by PFILE. Accessing PFILE after this function
245 returns leads to undefined behavior. Returns the error count. */
247 cpp_destroy (cpp_reader *pfile)
249 cpp_context *context, *contextn;
250 tokenrun *run, *runn;
253 free (pfile->op_stack);
255 while (CPP_BUFFER (pfile) != NULL)
256 _cpp_pop_buffer (pfile);
259 free (pfile->out.base);
261 if (pfile->macro_buffer)
263 free (pfile->macro_buffer);
264 pfile->macro_buffer = NULL;
265 pfile->macro_buffer_len = 0;
269 deps_free (pfile->deps);
270 obstack_free (&pfile->buffer_ob, 0);
272 _cpp_destroy_hashtable (pfile);
273 _cpp_cleanup_files (pfile);
274 _cpp_destroy_iconv (pfile);
276 _cpp_free_buff (pfile->a_buff);
277 _cpp_free_buff (pfile->u_buff);
278 _cpp_free_buff (pfile->free_buffs);
280 for (run = &pfile->base_run; run; run = runn)
284 if (run != &pfile->base_run)
288 for (context = pfile->base_context.next; context; context = contextn)
290 contextn = context->next;
294 if (pfile->comments.entries)
296 for (i = 0; i < pfile->comments.count; i++)
297 free (pfile->comments.entries[i].comment);
299 free (pfile->comments.entries);
305 /* This structure defines one built-in identifier. A node will be
306 entered in the hash table under the name NAME, with value VALUE.
308 There are two tables of these. builtin_array holds all the
309 "builtin" macros: these are handled by builtin_macro() in
310 macro.c. Builtin is somewhat of a misnomer -- the property of
311 interest is that these macros require special code to compute their
312 expansions. The value is a "builtin_type" enumerator.
314 operator_array holds the C++ named operators. These are keywords
315 which act as aliases for punctuators. In C++, they cannot be
316 altered through #define, and #if recognizes them as operators. In
317 C, these are not entered into the hash table at all (but see
318 <iso646.h>). The value is a token-type enumerator. */
321 const uchar *const name;
322 const unsigned short len;
323 const unsigned short value;
324 const bool always_warn_if_redefined;
327 #define B(n, t, f) { DSC(n), t, f }
328 static const struct builtin_macro builtin_array[] =
330 B("__TIMESTAMP__", BT_TIMESTAMP, false),
331 B("__TIME__", BT_TIME, false),
332 B("__DATE__", BT_DATE, false),
333 B("__FILE__", BT_FILE, false),
334 B("__BASE_FILE__", BT_BASE_FILE, false),
335 B("__LINE__", BT_SPECLINE, true),
336 B("__INCLUDE_LEVEL__", BT_INCLUDE_LEVEL, true),
337 B("__COUNTER__", BT_COUNTER, true),
338 /* Keep builtins not used for -traditional-cpp at the end, and
339 update init_builtins() if any more are added. */
340 B("_Pragma", BT_PRAGMA, true),
341 B("__STDC__", BT_STDC, true),
345 struct builtin_operator
347 const uchar *const name;
348 const unsigned short len;
349 const unsigned short value;
352 #define B(n, t) { DSC(n), t }
353 static const struct builtin_operator operator_array[] =
355 B("and", CPP_AND_AND),
356 B("and_eq", CPP_AND_EQ),
357 B("bitand", CPP_AND),
359 B("compl", CPP_COMPL),
361 B("not_eq", CPP_NOT_EQ),
363 B("or_eq", CPP_OR_EQ),
365 B("xor_eq", CPP_XOR_EQ)
369 /* Mark the C++ named operators in the hash table. */
371 mark_named_operators (cpp_reader *pfile)
373 const struct builtin_operator *b;
375 for (b = operator_array;
376 b < (operator_array + ARRAY_SIZE (operator_array));
379 cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
380 hp->flags |= NODE_OPERATOR;
381 hp->is_directive = 0;
382 hp->directive_index = b->value;
387 cpp_init_special_builtins (cpp_reader *pfile)
389 const struct builtin_macro *b;
390 size_t n = ARRAY_SIZE (builtin_array);
392 if (CPP_OPTION (pfile, traditional))
394 else if (! CPP_OPTION (pfile, stdc_0_in_system_headers)
395 || CPP_OPTION (pfile, std))
398 for (b = builtin_array; b < builtin_array + n; b++)
400 cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
402 hp->flags |= NODE_BUILTIN;
403 if (b->always_warn_if_redefined
404 || CPP_OPTION (pfile, warn_builtin_macro_redefined))
405 hp->flags |= NODE_WARN;
406 hp->value.builtin = (enum builtin_type) b->value;
410 /* Read the builtins table above and enter them, and language-specific
411 macros, into the hash table. HOSTED is true if this is a hosted
414 cpp_init_builtins (cpp_reader *pfile, int hosted)
416 cpp_init_special_builtins (pfile);
418 if (!CPP_OPTION (pfile, traditional)
419 && (! CPP_OPTION (pfile, stdc_0_in_system_headers)
420 || CPP_OPTION (pfile, std)))
421 _cpp_define_builtin (pfile, "__STDC__ 1");
423 if (CPP_OPTION (pfile, cplusplus))
424 _cpp_define_builtin (pfile, "__cplusplus 1");
425 else if (CPP_OPTION (pfile, lang) == CLK_ASM)
426 _cpp_define_builtin (pfile, "__ASSEMBLER__ 1");
427 else if (CPP_OPTION (pfile, lang) == CLK_STDC94)
428 _cpp_define_builtin (pfile, "__STDC_VERSION__ 199409L");
429 else if (CPP_OPTION (pfile, c99))
430 _cpp_define_builtin (pfile, "__STDC_VERSION__ 199901L");
433 _cpp_define_builtin (pfile, "__STDC_HOSTED__ 1");
435 _cpp_define_builtin (pfile, "__STDC_HOSTED__ 0");
437 if (CPP_OPTION (pfile, objc))
438 _cpp_define_builtin (pfile, "__OBJC__ 1");
441 /* Sanity-checks are dependent on command-line options, so it is
442 called as a subroutine of cpp_read_main_file (). */
444 static void sanity_checks (cpp_reader *);
445 static void sanity_checks (cpp_reader *pfile)
448 size_t max_precision = 2 * CHAR_BIT * sizeof (cpp_num_part);
450 /* Sanity checks for assumptions about CPP arithmetic and target
451 type precisions made by cpplib. */
454 cpp_error (pfile, CPP_DL_ICE, "cppchar_t must be an unsigned type");
456 if (CPP_OPTION (pfile, precision) > max_precision)
457 cpp_error (pfile, CPP_DL_ICE,
458 "preprocessor arithmetic has maximum precision of %lu bits;"
459 " target requires %lu bits",
460 (unsigned long) max_precision,
461 (unsigned long) CPP_OPTION (pfile, precision));
463 if (CPP_OPTION (pfile, precision) < CPP_OPTION (pfile, int_precision))
464 cpp_error (pfile, CPP_DL_ICE,
465 "CPP arithmetic must be at least as precise as a target int");
467 if (CPP_OPTION (pfile, char_precision) < 8)
468 cpp_error (pfile, CPP_DL_ICE, "target char is less than 8 bits wide");
470 if (CPP_OPTION (pfile, wchar_precision) < CPP_OPTION (pfile, char_precision))
471 cpp_error (pfile, CPP_DL_ICE,
472 "target wchar_t is narrower than target char");
474 if (CPP_OPTION (pfile, int_precision) < CPP_OPTION (pfile, char_precision))
475 cpp_error (pfile, CPP_DL_ICE,
476 "target int is narrower than target char");
478 /* This is assumed in eval_token() and could be fixed if necessary. */
479 if (sizeof (cppchar_t) > sizeof (cpp_num_part))
480 cpp_error (pfile, CPP_DL_ICE,
481 "CPP half-integer narrower than CPP character");
483 if (CPP_OPTION (pfile, wchar_precision) > BITS_PER_CPPCHAR_T)
484 cpp_error (pfile, CPP_DL_ICE,
485 "CPP on this host cannot handle wide character constants over"
486 " %lu bits, but the target requires %lu bits",
487 (unsigned long) BITS_PER_CPPCHAR_T,
488 (unsigned long) CPP_OPTION (pfile, wchar_precision));
491 # define sanity_checks(PFILE)
494 /* This is called after options have been parsed, and partially
497 cpp_post_options (cpp_reader *pfile)
499 sanity_checks (pfile);
501 post_options (pfile);
503 /* Mark named operators before handling command line macros. */
504 if (CPP_OPTION (pfile, cplusplus) && CPP_OPTION (pfile, operator_names))
505 mark_named_operators (pfile);
508 /* Setup for processing input from the file named FNAME, or stdin if
509 it is the empty string. Return the original filename
510 on success (e.g. foo.i->foo.c), or NULL on failure. */
512 cpp_read_main_file (cpp_reader *pfile, const char *fname)
514 if (CPP_OPTION (pfile, deps.style) != DEPS_NONE)
517 pfile->deps = deps_init ();
519 /* Set the default target (if there is none already). */
520 deps_add_default_target (pfile, fname);
524 = _cpp_find_file (pfile, fname, &pfile->no_search_path, false, 0);
525 if (_cpp_find_failed (pfile->main_file))
528 _cpp_stack_file (pfile, pfile->main_file, false);
530 /* For foo.i, read the original filename foo.c now, for the benefit
531 of the front ends. */
532 if (CPP_OPTION (pfile, preprocessed))
534 read_original_filename (pfile);
535 fname = pfile->line_table->maps[pfile->line_table->used-1].to_file;
540 /* For preprocessed files, if the first tokens are of the form # NUM.
541 handle the directive so we know the original file name. This will
542 generate file_change callbacks, which the front ends must handle
543 appropriately given their state of initialization. */
545 read_original_filename (cpp_reader *pfile)
547 const cpp_token *token, *token1;
549 /* Lex ahead; if the first tokens are of the form # NUM, then
550 process the directive, otherwise back up. */
551 token = _cpp_lex_direct (pfile);
552 if (token->type == CPP_HASH)
554 pfile->state.in_directive = 1;
555 token1 = _cpp_lex_direct (pfile);
556 _cpp_backup_tokens (pfile, 1);
557 pfile->state.in_directive = 0;
559 /* If it's a #line directive, handle it. */
560 if (token1->type == CPP_NUMBER)
562 _cpp_handle_directive (pfile, token->flags & PREV_WHITE);
563 read_original_directory (pfile);
568 /* Backup as if nothing happened. */
569 _cpp_backup_tokens (pfile, 1);
572 /* For preprocessed files, if the tokens following the first filename
573 line is of the form # <line> "/path/name//", handle the
574 directive so we know the original current directory. */
576 read_original_directory (cpp_reader *pfile)
578 const cpp_token *hash, *token;
580 /* Lex ahead; if the first tokens are of the form # NUM, then
581 process the directive, otherwise back up. */
582 hash = _cpp_lex_direct (pfile);
583 if (hash->type != CPP_HASH)
585 _cpp_backup_tokens (pfile, 1);
589 token = _cpp_lex_direct (pfile);
591 if (token->type != CPP_NUMBER)
593 _cpp_backup_tokens (pfile, 2);
597 token = _cpp_lex_direct (pfile);
599 if (token->type != CPP_STRING
600 || ! (token->val.str.len >= 5
601 && token->val.str.text[token->val.str.len-2] == '/'
602 && token->val.str.text[token->val.str.len-3] == '/'))
604 _cpp_backup_tokens (pfile, 3);
608 if (pfile->cb.dir_change)
610 char *debugdir = (char *) alloca (token->val.str.len - 3);
612 memcpy (debugdir, (const char *) token->val.str.text + 1,
613 token->val.str.len - 4);
614 debugdir[token->val.str.len - 4] = '\0';
616 pfile->cb.dir_change (pfile, debugdir);
620 /* This is called at the end of preprocessing. It pops the last
621 buffer and writes dependency output, and returns the number of
624 Maybe it should also reset state, such that you could call
625 cpp_start_read with a new filename to restart processing. */
627 cpp_finish (cpp_reader *pfile, FILE *deps_stream)
629 /* Warn about unused macros before popping the final buffer. */
630 if (CPP_OPTION (pfile, warn_unused_macros))
631 cpp_forall_identifiers (pfile, _cpp_warn_if_unused_macro, NULL);
633 /* lex.c leaves the final buffer on the stack. This it so that
634 it returns an unending stream of CPP_EOFs to the client. If we
635 popped the buffer, we'd dereference a NULL buffer pointer and
636 segfault. It's nice to allow the client to do worry-free excess
637 cpp_get_token calls. */
638 while (pfile->buffer)
639 _cpp_pop_buffer (pfile);
641 /* Don't write the deps file if there are errors. */
642 if (CPP_OPTION (pfile, deps.style) != DEPS_NONE
643 && deps_stream && pfile->errors == 0)
645 deps_write (pfile->deps, deps_stream, 72);
647 if (CPP_OPTION (pfile, deps.phony_targets))
648 deps_phony_targets (pfile->deps, deps_stream);
651 /* Report on headers that could use multiple include guards. */
652 if (CPP_OPTION (pfile, print_include_names))
653 _cpp_report_missing_guards (pfile);
655 return pfile->errors;
659 post_options (cpp_reader *pfile)
661 /* -Wtraditional is not useful in C++ mode. */
662 if (CPP_OPTION (pfile, cplusplus))
663 CPP_OPTION (pfile, warn_traditional) = 0;
665 /* Permanently disable macro expansion if we are rescanning
666 preprocessed text. Read preprocesed source in ISO mode. */
667 if (CPP_OPTION (pfile, preprocessed))
669 if (!CPP_OPTION (pfile, directives_only))
670 pfile->state.prevent_expansion = 1;
671 CPP_OPTION (pfile, traditional) = 0;
674 if (CPP_OPTION (pfile, warn_trigraphs) == 2)
675 CPP_OPTION (pfile, warn_trigraphs) = !CPP_OPTION (pfile, trigraphs);
677 if (CPP_OPTION (pfile, traditional))
679 CPP_OPTION (pfile, cplusplus_comments) = 0;
681 /* Traditional CPP does not accurately track column information. */
682 CPP_OPTION (pfile, show_column) = 0;
683 CPP_OPTION (pfile, trigraphs) = 0;
684 CPP_OPTION (pfile, warn_trigraphs) = 0;