2 Copyright (C) 1986, 1987, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
3 1999, 2000, 2001, 2002, 2003, 2004 Free Software Foundation, Inc.
4 Contributed by Per Bothner, 1994-95.
5 Based on CCCP program by Paul Rubin, June 1986
6 Adapted to ANSI C, Richard Stallman, Jan 1987
8 This program is free software; you can redistribute it and/or modify it
9 under the terms of the GNU General Public License as published by the
10 Free Software Foundation; either version 2, or (at your option) any
13 This program is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with this program; if not, write to the Free Software
20 Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
28 static void init_library (void);
29 static void mark_named_operators (cpp_reader *);
30 static void read_original_filename (cpp_reader *);
31 static void read_original_directory (cpp_reader *);
32 static void post_options (cpp_reader *);
34 /* If we have designated initializers (GCC >2.7) these tables can be
35 initialized, constant data. Otherwise, they have to be filled in at
37 #if HAVE_DESIGNATED_INITIALIZERS
39 #define init_trigraph_map() /* Nothing. */
40 #define TRIGRAPH_MAP \
41 __extension__ const uchar _cpp_trigraph_map[UCHAR_MAX + 1] = {
44 #define s(p, v) [p] = v,
48 #define TRIGRAPH_MAP uchar _cpp_trigraph_map[UCHAR_MAX + 1] = { 0 }; \
49 static void init_trigraph_map (void) { \
50 unsigned char *x = _cpp_trigraph_map;
53 #define s(p, v) x[p] = v;
58 s('=', '#') s(')', ']') s('!', '|')
59 s('(', '[') s('\'', '^') s('>', '}')
60 s('/', '\\') s('<', '{') s('-', '~')
67 /* A set of booleans indicating what CPP features each source language
73 char extended_numbers;
75 char cplusplus_comments;
79 static const struct lang_flags lang_defaults[] =
80 { /* c99 c++ xnum std // digr */
81 /* GNUC89 */ { 0, 0, 1, 0, 1, 1 },
82 /* GNUC99 */ { 1, 0, 1, 0, 1, 1 },
83 /* STDC89 */ { 0, 0, 0, 1, 0, 0 },
84 /* STDC94 */ { 0, 0, 0, 1, 0, 1 },
85 /* STDC99 */ { 1, 0, 1, 1, 1, 1 },
86 /* GNUCXX */ { 0, 1, 1, 0, 1, 1 },
87 /* CXX98 */ { 0, 1, 1, 1, 1, 1 },
88 /* ASM */ { 0, 0, 1, 0, 1, 0 }
91 /* Sets internal flags correctly for a given language. */
93 cpp_set_lang (cpp_reader *pfile, enum c_lang lang)
95 const struct lang_flags *l = &lang_defaults[(int) lang];
97 CPP_OPTION (pfile, lang) = lang;
99 CPP_OPTION (pfile, c99) = l->c99;
100 CPP_OPTION (pfile, cplusplus) = l->cplusplus;
101 CPP_OPTION (pfile, extended_numbers) = l->extended_numbers;
102 CPP_OPTION (pfile, std) = l->std;
103 CPP_OPTION (pfile, trigraphs) = l->std;
104 CPP_OPTION (pfile, cplusplus_comments) = l->cplusplus_comments;
105 CPP_OPTION (pfile, digraphs) = l->digraphs;
108 /* Initialize library global state. */
112 static int initialized = 0;
118 /* Set up the trigraph map. This doesn't need to do anything if
119 we were compiled with a compiler that supports C99 designated
121 init_trigraph_map ();
125 /* Initialize a cpp_reader structure. */
127 cpp_create_reader (enum c_lang lang, hash_table *table)
131 /* Initialize this instance of the library if it hasn't been already. */
134 pfile = xcalloc (1, sizeof (cpp_reader));
136 cpp_set_lang (pfile, lang);
137 CPP_OPTION (pfile, warn_multichar) = 1;
138 CPP_OPTION (pfile, discard_comments) = 1;
139 CPP_OPTION (pfile, discard_comments_in_macro_exp) = 1;
140 CPP_OPTION (pfile, show_column) = 1;
141 CPP_OPTION (pfile, tabstop) = 8;
142 CPP_OPTION (pfile, operator_names) = 1;
143 CPP_OPTION (pfile, warn_trigraphs) = 2;
144 CPP_OPTION (pfile, warn_endif_labels) = 1;
145 CPP_OPTION (pfile, warn_deprecated) = 1;
146 CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
147 CPP_OPTION (pfile, dollars_in_ident) = 1;
148 CPP_OPTION (pfile, warn_dollars) = 1;
150 /* Default CPP arithmetic to something sensible for the host for the
151 benefit of dumb users like fix-header. */
152 CPP_OPTION (pfile, precision) = CHAR_BIT * sizeof (long);
153 CPP_OPTION (pfile, char_precision) = CHAR_BIT;
154 CPP_OPTION (pfile, wchar_precision) = CHAR_BIT * sizeof (int);
155 CPP_OPTION (pfile, int_precision) = CHAR_BIT * sizeof (int);
156 CPP_OPTION (pfile, unsigned_char) = 0;
157 CPP_OPTION (pfile, unsigned_wchar) = 1;
158 CPP_OPTION (pfile, bytes_big_endian) = 1; /* does not matter */
160 /* Default to locale/UTF-8. */
161 CPP_OPTION (pfile, narrow_charset) = _cpp_default_encoding ();
162 CPP_OPTION (pfile, wide_charset) = 0;
163 CPP_OPTION (pfile, input_charset) = _cpp_default_encoding ();
166 CPP_OPTION (pfile, preproc_asm) = 1;
167 CPP_OPTION (pfile, pedantic_parse_number) = 0;
168 CPP_OPTION (pfile, obj_ext) = NULL;
170 /* A fake empty "directory" used as the starting point for files
171 looked up without a search path. Name cannot be '/' because we
172 don't want to prepend anything at all to filenames using it. All
173 other entries are correct zero-initialized. */
174 pfile->no_search_path.name = (char *) "";
176 /* Initialize the line map. Start at logical line 1, so we can use
177 a line number of zero for special states. */
178 linemap_init (&pfile->line_maps);
181 /* Initialize lexer state. */
182 pfile->state.save_comments = ! CPP_OPTION (pfile, discard_comments);
184 /* Set up static tokens. */
185 pfile->avoid_paste.type = CPP_PADDING;
186 pfile->avoid_paste.val.source = NULL;
187 pfile->eof.type = CPP_EOF;
188 pfile->eof.flags = 0;
190 /* Create a token buffer for the lexer. */
191 _cpp_init_tokenrun (&pfile->base_run, 250);
192 pfile->cur_run = &pfile->base_run;
193 pfile->cur_token = pfile->base_run.base;
195 /* Initialize the base context. */
196 pfile->context = &pfile->base_context;
197 pfile->base_context.macro = 0;
198 pfile->base_context.prev = pfile->base_context.next = 0;
200 /* Aligned and unaligned storage. */
201 pfile->a_buff = _cpp_get_buff (pfile, 0);
202 pfile->u_buff = _cpp_get_buff (pfile, 0);
204 /* The expression parser stack. */
205 _cpp_expand_op_stack (pfile);
207 /* Initialize the buffer obstack. */
208 _obstack_begin (&pfile->buffer_ob, 0, 0,
209 (void *(*) (long)) xmalloc,
210 (void (*) (void *)) free);
212 _cpp_init_files (pfile);
214 _cpp_init_hashtable (pfile, table);
219 /* Free resources used by PFILE. Accessing PFILE after this function
220 returns leads to undefined behavior. Returns the error count. */
222 cpp_destroy (cpp_reader *pfile)
224 cpp_context *context, *contextn;
225 tokenrun *run, *runn;
227 free (pfile->op_stack);
229 while (CPP_BUFFER (pfile) != NULL)
230 _cpp_pop_buffer (pfile);
233 free (pfile->out.base);
235 if (pfile->macro_buffer)
237 free (pfile->macro_buffer);
238 pfile->macro_buffer = NULL;
239 pfile->macro_buffer_len = 0;
243 deps_free (pfile->deps);
244 obstack_free (&pfile->buffer_ob, 0);
246 _cpp_destroy_hashtable (pfile);
247 _cpp_cleanup_files (pfile);
248 _cpp_destroy_iconv (pfile);
250 _cpp_free_buff (pfile->a_buff);
251 _cpp_free_buff (pfile->u_buff);
252 _cpp_free_buff (pfile->free_buffs);
254 for (run = &pfile->base_run; run; run = runn)
258 if (run != &pfile->base_run)
262 for (context = pfile->base_context.next; context; context = contextn)
264 contextn = context->next;
268 linemap_free (&pfile->line_maps);
272 /* This structure defines one built-in identifier. A node will be
273 entered in the hash table under the name NAME, with value VALUE.
275 There are two tables of these. builtin_array holds all the
276 "builtin" macros: these are handled by builtin_macro() in
277 cppmacro.c. Builtin is somewhat of a misnomer -- the property of
278 interest is that these macros require special code to compute their
279 expansions. The value is a "builtin_type" enumerator.
281 operator_array holds the C++ named operators. These are keywords
282 which act as aliases for punctuators. In C++, they cannot be
283 altered through #define, and #if recognizes them as operators. In
284 C, these are not entered into the hash table at all (but see
285 <iso646.h>). The value is a token-type enumerator. */
290 unsigned short value;
293 #define B(n, t) { DSC(n), t }
294 static const struct builtin builtin_array[] =
296 B("__TIME__", BT_TIME),
297 B("__DATE__", BT_DATE),
298 B("__FILE__", BT_FILE),
299 B("__BASE_FILE__", BT_BASE_FILE),
300 B("__LINE__", BT_SPECLINE),
301 B("__INCLUDE_LEVEL__", BT_INCLUDE_LEVEL),
302 /* Keep builtins not used for -traditional-cpp at the end, and
303 update init_builtins() if any more are added. */
304 B("_Pragma", BT_PRAGMA),
305 B("__STDC__", BT_STDC),
308 static const struct builtin operator_array[] =
310 B("and", CPP_AND_AND),
311 B("and_eq", CPP_AND_EQ),
312 B("bitand", CPP_AND),
314 B("compl", CPP_COMPL),
316 B("not_eq", CPP_NOT_EQ),
318 B("or_eq", CPP_OR_EQ),
320 B("xor_eq", CPP_XOR_EQ)
324 /* Mark the C++ named operators in the hash table. */
326 mark_named_operators (cpp_reader *pfile)
328 const struct builtin *b;
330 for (b = operator_array;
331 b < (operator_array + ARRAY_SIZE (operator_array));
334 cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
335 hp->flags |= NODE_OPERATOR;
336 hp->is_directive = 0;
337 hp->directive_index = b->value;
341 /* Read the builtins table above and enter them, and language-specific
342 macros, into the hash table. HOSTED is true if this is a hosted
345 cpp_init_builtins (cpp_reader *pfile, int hosted)
347 const struct builtin *b;
348 size_t n = ARRAY_SIZE (builtin_array);
350 if (CPP_OPTION (pfile, traditional))
353 for(b = builtin_array; b < builtin_array + n; b++)
355 cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
357 hp->flags |= NODE_BUILTIN | NODE_WARN;
358 hp->value.builtin = b->value;
361 if (CPP_OPTION (pfile, cplusplus))
362 _cpp_define_builtin (pfile, "__cplusplus 1");
363 else if (CPP_OPTION (pfile, lang) == CLK_ASM)
364 _cpp_define_builtin (pfile, "__ASSEMBLER__ 1");
365 else if (CPP_OPTION (pfile, lang) == CLK_STDC94)
366 _cpp_define_builtin (pfile, "__STDC_VERSION__ 199409L");
367 else if (CPP_OPTION (pfile, c99))
368 _cpp_define_builtin (pfile, "__STDC_VERSION__ 199901L");
371 _cpp_define_builtin (pfile, "__STDC_HOSTED__ 1");
373 _cpp_define_builtin (pfile, "__STDC_HOSTED__ 0");
375 if (CPP_OPTION (pfile, objc))
376 _cpp_define_builtin (pfile, "__OBJC__ 1");
379 /* Sanity-checks are dependent on command-line options, so it is
380 called as a subroutine of cpp_read_main_file (). */
382 static void sanity_checks (cpp_reader *);
383 static void sanity_checks (cpp_reader *pfile)
386 size_t max_precision = 2 * CHAR_BIT * sizeof (cpp_num_part);
388 /* Sanity checks for assumptions about CPP arithmetic and target
389 type precisions made by cpplib. */
392 cpp_error (pfile, CPP_DL_ICE, "cppchar_t must be an unsigned type");
394 if (CPP_OPTION (pfile, precision) > max_precision)
395 cpp_error (pfile, CPP_DL_ICE,
396 "preprocessor arithmetic has maximum precision of %lu bits;"
397 " target requires %lu bits",
398 (unsigned long) max_precision,
399 (unsigned long) CPP_OPTION (pfile, precision));
401 if (CPP_OPTION (pfile, precision) < CPP_OPTION (pfile, int_precision))
402 cpp_error (pfile, CPP_DL_ICE,
403 "CPP arithmetic must be at least as precise as a target int");
405 if (CPP_OPTION (pfile, char_precision) < 8)
406 cpp_error (pfile, CPP_DL_ICE, "target char is less than 8 bits wide");
408 if (CPP_OPTION (pfile, wchar_precision) < CPP_OPTION (pfile, char_precision))
409 cpp_error (pfile, CPP_DL_ICE,
410 "target wchar_t is narrower than target char");
412 if (CPP_OPTION (pfile, int_precision) < CPP_OPTION (pfile, char_precision))
413 cpp_error (pfile, CPP_DL_ICE,
414 "target int is narrower than target char");
416 /* This is assumed in eval_token() and could be fixed if necessary. */
417 if (sizeof (cppchar_t) > sizeof (cpp_num_part))
418 cpp_error (pfile, CPP_DL_ICE,
419 "CPP half-integer narrower than CPP character");
421 if (CPP_OPTION (pfile, wchar_precision) > BITS_PER_CPPCHAR_T)
422 cpp_error (pfile, CPP_DL_ICE,
423 "CPP on this host cannot handle wide character constants over"
424 " %lu bits, but the target requires %lu bits",
425 (unsigned long) BITS_PER_CPPCHAR_T,
426 (unsigned long) CPP_OPTION (pfile, wchar_precision));
429 # define sanity_checks(PFILE)
432 /* Add a dependency target. Can be called any number of times before
433 cpp_read_main_file(). If no targets have been added before
434 cpp_read_main_file(), then the default target is used. */
436 cpp_add_dependency_target (cpp_reader *pfile, const char *target, int quote)
439 pfile->deps = deps_init ();
441 deps_add_target (pfile->deps, target, quote);
444 /* This is called after options have been parsed, and partially
447 cpp_post_options (cpp_reader *pfile)
449 sanity_checks (pfile);
451 post_options (pfile);
453 /* Mark named operators before handling command line macros. */
454 if (CPP_OPTION (pfile, cplusplus) && CPP_OPTION (pfile, operator_names))
455 mark_named_operators (pfile);
458 /* Setup for processing input from the file named FNAME, or stdin if
459 it is the empty string. Return the original filename
460 on success (e.g. foo.i->foo.c), or NULL on failure. */
462 cpp_read_main_file (cpp_reader *pfile, const char *fname)
464 if (CPP_OPTION (pfile, deps.style) != DEPS_NONE)
467 pfile->deps = deps_init ();
469 /* Set the default target (if there is none already). */
470 deps_add_default_target (pfile, fname);
474 = _cpp_find_file (pfile, fname, &pfile->no_search_path, false, 0);
475 if (_cpp_find_failed (pfile->main_file))
478 _cpp_stack_file (pfile, pfile->main_file, false);
480 /* For foo.i, read the original filename foo.c now, for the benefit
481 of the front ends. */
482 if (CPP_OPTION (pfile, preprocessed))
484 read_original_filename (pfile);
487 fname = pfile->map->to_file;
492 /* For preprocessed files, if the first tokens are of the form # NUM.
493 handle the directive so we know the original file name. This will
494 generate file_change callbacks, which the front ends must handle
495 appropriately given their state of initialization. */
497 read_original_filename (cpp_reader *pfile)
499 const cpp_token *token, *token1;
501 /* Lex ahead; if the first tokens are of the form # NUM, then
502 process the directive, otherwise back up. */
503 token = _cpp_lex_direct (pfile);
504 if (token->type == CPP_HASH)
506 pfile->state.in_directive = 1;
507 token1 = _cpp_lex_direct (pfile);
508 _cpp_backup_tokens (pfile, 1);
509 pfile->state.in_directive = 0;
511 /* If it's a #line directive, handle it. */
512 if (token1->type == CPP_NUMBER)
514 _cpp_handle_directive (pfile, token->flags & PREV_WHITE);
515 read_original_directory (pfile);
520 /* Backup as if nothing happened. */
521 _cpp_backup_tokens (pfile, 1);
524 /* For preprocessed files, if the tokens following the first filename
525 line is of the form # <line> "/path/name//", handle the
526 directive so we know the original current directory. */
528 read_original_directory (cpp_reader *pfile)
530 const cpp_token *hash, *token;
532 /* Lex ahead; if the first tokens are of the form # NUM, then
533 process the directive, otherwise back up. */
534 hash = _cpp_lex_direct (pfile);
535 if (hash->type != CPP_HASH)
537 _cpp_backup_tokens (pfile, 1);
541 token = _cpp_lex_direct (pfile);
543 if (token->type != CPP_NUMBER)
545 _cpp_backup_tokens (pfile, 2);
549 token = _cpp_lex_direct (pfile);
551 if (token->type != CPP_STRING
552 || ! (token->val.str.len >= 5
553 && token->val.str.text[token->val.str.len-2] == '/'
554 && token->val.str.text[token->val.str.len-3] == '/'))
556 _cpp_backup_tokens (pfile, 3);
560 if (pfile->cb.dir_change)
562 char *debugdir = alloca (token->val.str.len - 3);
564 memcpy (debugdir, (const char *) token->val.str.text + 1,
565 token->val.str.len - 4);
566 debugdir[token->val.str.len - 4] = '\0';
568 pfile->cb.dir_change (pfile, debugdir);
572 /* This is called at the end of preprocessing. It pops the last
573 buffer and writes dependency output, and returns the number of
576 Maybe it should also reset state, such that you could call
577 cpp_start_read with a new filename to restart processing. */
579 cpp_finish (cpp_reader *pfile, FILE *deps_stream)
581 /* Warn about unused macros before popping the final buffer. */
582 if (CPP_OPTION (pfile, warn_unused_macros))
583 cpp_forall_identifiers (pfile, _cpp_warn_if_unused_macro, NULL);
585 /* cpplex.c leaves the final buffer on the stack. This it so that
586 it returns an unending stream of CPP_EOFs to the client. If we
587 popped the buffer, we'd dereference a NULL buffer pointer and
588 segfault. It's nice to allow the client to do worry-free excess
589 cpp_get_token calls. */
590 while (pfile->buffer)
591 _cpp_pop_buffer (pfile);
593 /* Don't write the deps file if there are errors. */
594 if (CPP_OPTION (pfile, deps.style) != DEPS_NONE
595 && deps_stream && pfile->errors == 0)
597 deps_write (pfile->deps, deps_stream, 72);
599 if (CPP_OPTION (pfile, deps.phony_targets))
600 deps_phony_targets (pfile->deps, deps_stream);
603 /* Report on headers that could use multiple include guards. */
604 if (CPP_OPTION (pfile, print_include_names))
605 _cpp_report_missing_guards (pfile);
607 return pfile->errors;
611 post_options (cpp_reader *pfile)
613 /* -Wtraditional is not useful in C++ mode. */
614 if (CPP_OPTION (pfile, cplusplus))
615 CPP_OPTION (pfile, warn_traditional) = 0;
617 /* Permanently disable macro expansion if we are rescanning
618 preprocessed text. Read preprocesed source in ISO mode. */
619 if (CPP_OPTION (pfile, preprocessed))
621 pfile->state.prevent_expansion = 1;
622 CPP_OPTION (pfile, traditional) = 0;
625 if (CPP_OPTION (pfile, warn_trigraphs) == 2)
626 CPP_OPTION (pfile, warn_trigraphs) = !CPP_OPTION (pfile, trigraphs);
628 if (CPP_OPTION (pfile, traditional))
630 CPP_OPTION (pfile, cplusplus_comments) = 0;
632 /* Traditional CPP does not accurately track column information. */
633 CPP_OPTION (pfile, show_column) = 0;
634 CPP_OPTION (pfile, trigraphs) = 0;
635 CPP_OPTION (pfile, warn_trigraphs) = 0;