./:
[official-gcc.git] / libcpp / init.c
blobdb697e9ed34e4db3d5ed60e9e7a2967b489762a3
1 /* CPP Library.
2 Copyright (C) 1986, 1987, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
3 1999, 2000, 2001, 2002, 2003, 2004, 2005 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
11 later version.
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, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
22 #include "config.h"
23 #include "system.h"
24 #include "cpplib.h"
25 #include "internal.h"
26 #include "mkdeps.h"
27 #include "localedir.h"
29 static void init_library (void);
30 static void mark_named_operators (cpp_reader *);
31 static void read_original_filename (cpp_reader *);
32 static void read_original_directory (cpp_reader *);
33 static void post_options (cpp_reader *);
35 /* If we have designated initializers (GCC >2.7) these tables can be
36 initialized, constant data. Otherwise, they have to be filled in at
37 runtime. */
38 #if HAVE_DESIGNATED_INITIALIZERS
40 #define init_trigraph_map() /* Nothing. */
41 #define TRIGRAPH_MAP \
42 __extension__ const uchar _cpp_trigraph_map[UCHAR_MAX + 1] = {
44 #define END };
45 #define s(p, v) [p] = v,
47 #else
49 #define TRIGRAPH_MAP uchar _cpp_trigraph_map[UCHAR_MAX + 1] = { 0 }; \
50 static void init_trigraph_map (void) { \
51 unsigned char *x = _cpp_trigraph_map;
53 #define END }
54 #define s(p, v) x[p] = v;
56 #endif
58 TRIGRAPH_MAP
59 s('=', '#') s(')', ']') s('!', '|')
60 s('(', '[') s('\'', '^') s('>', '}')
61 s('/', '\\') s('<', '{') s('-', '~')
62 END
64 #undef s
65 #undef END
66 #undef TRIGRAPH_MAP
68 /* A set of booleans indicating what CPP features each source language
69 requires. */
70 struct lang_flags
72 char c99;
73 char cplusplus;
74 char extended_numbers;
75 char extended_identifiers;
76 char std;
77 char cplusplus_comments;
78 char digraphs;
81 static const struct lang_flags lang_defaults[] =
82 { /* c99 c++ xnum xid std // digr */
83 /* GNUC89 */ { 0, 0, 1, 0, 0, 1, 1 },
84 /* GNUC99 */ { 1, 0, 1, 0, 0, 1, 1 },
85 /* STDC89 */ { 0, 0, 0, 0, 1, 0, 0 },
86 /* STDC94 */ { 0, 0, 0, 0, 1, 0, 1 },
87 /* STDC99 */ { 1, 0, 1, 0, 1, 1, 1 },
88 /* GNUCXX */ { 0, 1, 1, 0, 0, 1, 1 },
89 /* CXX98 */ { 0, 1, 1, 0, 1, 1, 1 },
90 /* GNUCXX0X */ { 1, 1, 1, 0, 0, 1, 1 },
91 /* CXX0X */ { 1, 1, 1, 0, 1, 1, 1 },
92 /* ASM */ { 0, 0, 1, 0, 0, 1, 0 }
93 /* xid should be 1 for GNUC99, STDC99, GNUCXX, CXX98, GNUCXX0X, and
94 CXX0X when no longer experimental (when all uses of identifiers
95 in the compiler have been audited for correct handling of
96 extended identifiers). */
99 /* Sets internal flags correctly for a given language. */
100 void
101 cpp_set_lang (cpp_reader *pfile, enum c_lang lang)
103 const struct lang_flags *l = &lang_defaults[(int) lang];
105 CPP_OPTION (pfile, lang) = lang;
107 CPP_OPTION (pfile, c99) = l->c99;
108 CPP_OPTION (pfile, cplusplus) = l->cplusplus;
109 CPP_OPTION (pfile, extended_numbers) = l->extended_numbers;
110 CPP_OPTION (pfile, extended_identifiers) = l->extended_identifiers;
111 CPP_OPTION (pfile, std) = l->std;
112 CPP_OPTION (pfile, trigraphs) = l->std;
113 CPP_OPTION (pfile, cplusplus_comments) = l->cplusplus_comments;
114 CPP_OPTION (pfile, digraphs) = l->digraphs;
117 /* Initialize library global state. */
118 static void
119 init_library (void)
121 static int initialized = 0;
123 if (! initialized)
125 initialized = 1;
127 /* Set up the trigraph map. This doesn't need to do anything if
128 we were compiled with a compiler that supports C99 designated
129 initializers. */
130 init_trigraph_map ();
132 #ifdef ENABLE_NLS
133 (void) bindtextdomain (PACKAGE, LOCALEDIR);
134 #endif
138 /* Initialize a cpp_reader structure. */
139 cpp_reader *
140 cpp_create_reader (enum c_lang lang, hash_table *table,
141 struct line_maps *line_table)
143 cpp_reader *pfile;
145 /* Initialize this instance of the library if it hasn't been already. */
146 init_library ();
148 pfile = XCNEW (cpp_reader);
150 cpp_set_lang (pfile, lang);
151 CPP_OPTION (pfile, warn_multichar) = 1;
152 CPP_OPTION (pfile, discard_comments) = 1;
153 CPP_OPTION (pfile, discard_comments_in_macro_exp) = 1;
154 CPP_OPTION (pfile, show_column) = 1;
155 CPP_OPTION (pfile, tabstop) = 8;
156 CPP_OPTION (pfile, operator_names) = 1;
157 CPP_OPTION (pfile, warn_trigraphs) = 2;
158 CPP_OPTION (pfile, warn_endif_labels) = 1;
159 CPP_OPTION (pfile, warn_deprecated) = 1;
160 CPP_OPTION (pfile, warn_long_long) = !CPP_OPTION (pfile, c99);
161 CPP_OPTION (pfile, dollars_in_ident) = 1;
162 CPP_OPTION (pfile, warn_dollars) = 1;
163 CPP_OPTION (pfile, warn_variadic_macros) = 1;
164 CPP_OPTION (pfile, warn_normalize) = normalized_C;
166 /* Default CPP arithmetic to something sensible for the host for the
167 benefit of dumb users like fix-header. */
168 CPP_OPTION (pfile, precision) = CHAR_BIT * sizeof (long);
169 CPP_OPTION (pfile, char_precision) = CHAR_BIT;
170 CPP_OPTION (pfile, wchar_precision) = CHAR_BIT * sizeof (int);
171 CPP_OPTION (pfile, int_precision) = CHAR_BIT * sizeof (int);
172 CPP_OPTION (pfile, unsigned_char) = 0;
173 CPP_OPTION (pfile, unsigned_wchar) = 1;
174 CPP_OPTION (pfile, bytes_big_endian) = 1; /* does not matter */
176 /* Default to no charset conversion. */
177 CPP_OPTION (pfile, narrow_charset) = _cpp_default_encoding ();
178 CPP_OPTION (pfile, wide_charset) = 0;
180 /* Default the input character set to UTF-8. */
181 CPP_OPTION (pfile, input_charset) = _cpp_default_encoding ();
183 /* A fake empty "directory" used as the starting point for files
184 looked up without a search path. Name cannot be '/' because we
185 don't want to prepend anything at all to filenames using it. All
186 other entries are correct zero-initialized. */
187 pfile->no_search_path.name = (char *) "";
189 /* Initialize the line map. */
190 pfile->line_table = line_table;
192 /* Initialize lexer state. */
193 pfile->state.save_comments = ! CPP_OPTION (pfile, discard_comments);
195 /* Set up static tokens. */
196 pfile->avoid_paste.type = CPP_PADDING;
197 pfile->avoid_paste.val.source = NULL;
198 pfile->eof.type = CPP_EOF;
199 pfile->eof.flags = 0;
201 /* Create a token buffer for the lexer. */
202 _cpp_init_tokenrun (&pfile->base_run, 250);
203 pfile->cur_run = &pfile->base_run;
204 pfile->cur_token = pfile->base_run.base;
206 /* Initialize the base context. */
207 pfile->context = &pfile->base_context;
208 pfile->base_context.macro = 0;
209 pfile->base_context.prev = pfile->base_context.next = 0;
211 /* Aligned and unaligned storage. */
212 pfile->a_buff = _cpp_get_buff (pfile, 0);
213 pfile->u_buff = _cpp_get_buff (pfile, 0);
215 /* The expression parser stack. */
216 _cpp_expand_op_stack (pfile);
218 /* Initialize the buffer obstack. */
219 _obstack_begin (&pfile->buffer_ob, 0, 0,
220 (void *(*) (long)) xmalloc,
221 (void (*) (void *)) free);
223 _cpp_init_files (pfile);
225 _cpp_init_hashtable (pfile, table);
227 return pfile;
230 /* Free resources used by PFILE. Accessing PFILE after this function
231 returns leads to undefined behavior. Returns the error count. */
232 void
233 cpp_destroy (cpp_reader *pfile)
235 cpp_context *context, *contextn;
236 tokenrun *run, *runn;
238 free (pfile->op_stack);
240 while (CPP_BUFFER (pfile) != NULL)
241 _cpp_pop_buffer (pfile);
243 if (pfile->out.base)
244 free (pfile->out.base);
246 if (pfile->macro_buffer)
248 free (pfile->macro_buffer);
249 pfile->macro_buffer = NULL;
250 pfile->macro_buffer_len = 0;
253 if (pfile->deps)
254 deps_free (pfile->deps);
255 obstack_free (&pfile->buffer_ob, 0);
257 _cpp_destroy_hashtable (pfile);
258 _cpp_cleanup_files (pfile);
259 _cpp_destroy_iconv (pfile);
261 _cpp_free_buff (pfile->a_buff);
262 _cpp_free_buff (pfile->u_buff);
263 _cpp_free_buff (pfile->free_buffs);
265 for (run = &pfile->base_run; run; run = runn)
267 runn = run->next;
268 free (run->base);
269 if (run != &pfile->base_run)
270 free (run);
273 for (context = pfile->base_context.next; context; context = contextn)
275 contextn = context->next;
276 free (context);
279 free (pfile);
282 /* This structure defines one built-in identifier. A node will be
283 entered in the hash table under the name NAME, with value VALUE.
285 There are two tables of these. builtin_array holds all the
286 "builtin" macros: these are handled by builtin_macro() in
287 macro.c. Builtin is somewhat of a misnomer -- the property of
288 interest is that these macros require special code to compute their
289 expansions. The value is a "builtin_type" enumerator.
291 operator_array holds the C++ named operators. These are keywords
292 which act as aliases for punctuators. In C++, they cannot be
293 altered through #define, and #if recognizes them as operators. In
294 C, these are not entered into the hash table at all (but see
295 <iso646.h>). The value is a token-type enumerator. */
296 struct builtin
298 const uchar *name;
299 unsigned short len;
300 unsigned short value;
303 #define B(n, t) { DSC(n), t }
304 static const struct builtin builtin_array[] =
306 B("__TIMESTAMP__", BT_TIMESTAMP),
307 B("__TIME__", BT_TIME),
308 B("__DATE__", BT_DATE),
309 B("__FILE__", BT_FILE),
310 B("__BASE_FILE__", BT_BASE_FILE),
311 B("__LINE__", BT_SPECLINE),
312 B("__INCLUDE_LEVEL__", BT_INCLUDE_LEVEL),
313 /* Keep builtins not used for -traditional-cpp at the end, and
314 update init_builtins() if any more are added. */
315 B("_Pragma", BT_PRAGMA),
316 B("__STDC__", BT_STDC),
319 static const struct builtin operator_array[] =
321 B("and", CPP_AND_AND),
322 B("and_eq", CPP_AND_EQ),
323 B("bitand", CPP_AND),
324 B("bitor", CPP_OR),
325 B("compl", CPP_COMPL),
326 B("not", CPP_NOT),
327 B("not_eq", CPP_NOT_EQ),
328 B("or", CPP_OR_OR),
329 B("or_eq", CPP_OR_EQ),
330 B("xor", CPP_XOR),
331 B("xor_eq", CPP_XOR_EQ)
333 #undef B
335 /* Mark the C++ named operators in the hash table. */
336 static void
337 mark_named_operators (cpp_reader *pfile)
339 const struct builtin *b;
341 for (b = operator_array;
342 b < (operator_array + ARRAY_SIZE (operator_array));
343 b++)
345 cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
346 hp->flags |= NODE_OPERATOR;
347 hp->is_directive = 0;
348 hp->directive_index = b->value;
352 /* Read the builtins table above and enter them, and language-specific
353 macros, into the hash table. HOSTED is true if this is a hosted
354 environment. */
355 void
356 cpp_init_builtins (cpp_reader *pfile, int hosted)
358 const struct builtin *b;
359 size_t n = ARRAY_SIZE (builtin_array);
361 if (CPP_OPTION (pfile, traditional))
362 n -= 2;
363 else if (! CPP_OPTION (pfile, stdc_0_in_system_headers)
364 || CPP_OPTION (pfile, std))
366 n--;
367 _cpp_define_builtin (pfile, "__STDC__ 1");
370 for (b = builtin_array; b < builtin_array + n; b++)
372 cpp_hashnode *hp = cpp_lookup (pfile, b->name, b->len);
373 hp->type = NT_MACRO;
374 hp->flags |= NODE_BUILTIN | NODE_WARN;
375 hp->value.builtin = (enum builtin_type) b->value;
378 if (CPP_OPTION (pfile, cplusplus))
379 _cpp_define_builtin (pfile, "__cplusplus 1");
380 else if (CPP_OPTION (pfile, lang) == CLK_ASM)
381 _cpp_define_builtin (pfile, "__ASSEMBLER__ 1");
382 else if (CPP_OPTION (pfile, lang) == CLK_STDC94)
383 _cpp_define_builtin (pfile, "__STDC_VERSION__ 199409L");
384 else if (CPP_OPTION (pfile, c99))
385 _cpp_define_builtin (pfile, "__STDC_VERSION__ 199901L");
387 if (hosted)
388 _cpp_define_builtin (pfile, "__STDC_HOSTED__ 1");
389 else
390 _cpp_define_builtin (pfile, "__STDC_HOSTED__ 0");
392 if (CPP_OPTION (pfile, objc))
393 _cpp_define_builtin (pfile, "__OBJC__ 1");
396 /* Sanity-checks are dependent on command-line options, so it is
397 called as a subroutine of cpp_read_main_file (). */
398 #if ENABLE_CHECKING
399 static void sanity_checks (cpp_reader *);
400 static void sanity_checks (cpp_reader *pfile)
402 cppchar_t test = 0;
403 size_t max_precision = 2 * CHAR_BIT * sizeof (cpp_num_part);
405 /* Sanity checks for assumptions about CPP arithmetic and target
406 type precisions made by cpplib. */
407 test--;
408 if (test < 1)
409 cpp_error (pfile, CPP_DL_ICE, "cppchar_t must be an unsigned type");
411 if (CPP_OPTION (pfile, precision) > max_precision)
412 cpp_error (pfile, CPP_DL_ICE,
413 "preprocessor arithmetic has maximum precision of %lu bits;"
414 " target requires %lu bits",
415 (unsigned long) max_precision,
416 (unsigned long) CPP_OPTION (pfile, precision));
418 if (CPP_OPTION (pfile, precision) < CPP_OPTION (pfile, int_precision))
419 cpp_error (pfile, CPP_DL_ICE,
420 "CPP arithmetic must be at least as precise as a target int");
422 if (CPP_OPTION (pfile, char_precision) < 8)
423 cpp_error (pfile, CPP_DL_ICE, "target char is less than 8 bits wide");
425 if (CPP_OPTION (pfile, wchar_precision) < CPP_OPTION (pfile, char_precision))
426 cpp_error (pfile, CPP_DL_ICE,
427 "target wchar_t is narrower than target char");
429 if (CPP_OPTION (pfile, int_precision) < CPP_OPTION (pfile, char_precision))
430 cpp_error (pfile, CPP_DL_ICE,
431 "target int is narrower than target char");
433 /* This is assumed in eval_token() and could be fixed if necessary. */
434 if (sizeof (cppchar_t) > sizeof (cpp_num_part))
435 cpp_error (pfile, CPP_DL_ICE,
436 "CPP half-integer narrower than CPP character");
438 if (CPP_OPTION (pfile, wchar_precision) > BITS_PER_CPPCHAR_T)
439 cpp_error (pfile, CPP_DL_ICE,
440 "CPP on this host cannot handle wide character constants over"
441 " %lu bits, but the target requires %lu bits",
442 (unsigned long) BITS_PER_CPPCHAR_T,
443 (unsigned long) CPP_OPTION (pfile, wchar_precision));
445 #else
446 # define sanity_checks(PFILE)
447 #endif
449 /* This is called after options have been parsed, and partially
450 processed. */
451 void
452 cpp_post_options (cpp_reader *pfile)
454 sanity_checks (pfile);
456 post_options (pfile);
458 /* Mark named operators before handling command line macros. */
459 if (CPP_OPTION (pfile, cplusplus) && CPP_OPTION (pfile, operator_names))
460 mark_named_operators (pfile);
463 /* Setup for processing input from the file named FNAME, or stdin if
464 it is the empty string. Return the original filename
465 on success (e.g. foo.i->foo.c), or NULL on failure. */
466 const char *
467 cpp_read_main_file (cpp_reader *pfile, const char *fname)
469 if (CPP_OPTION (pfile, deps.style) != DEPS_NONE)
471 if (!pfile->deps)
472 pfile->deps = deps_init ();
474 /* Set the default target (if there is none already). */
475 deps_add_default_target (pfile->deps, fname);
478 pfile->main_file
479 = _cpp_find_file (pfile, fname, &pfile->no_search_path, false, 0);
480 if (_cpp_find_failed (pfile->main_file))
481 return NULL;
483 _cpp_stack_file (pfile, pfile->main_file, false);
485 /* For foo.i, read the original filename foo.c now, for the benefit
486 of the front ends. */
487 if (CPP_OPTION (pfile, preprocessed))
489 read_original_filename (pfile);
490 fname = pfile->line_table->maps[pfile->line_table->used-1].to_file;
492 return fname;
495 /* For preprocessed files, if the first tokens are of the form # NUM.
496 handle the directive so we know the original file name. This will
497 generate file_change callbacks, which the front ends must handle
498 appropriately given their state of initialization. */
499 static void
500 read_original_filename (cpp_reader *pfile)
502 const cpp_token *token, *token1;
504 /* Lex ahead; if the first tokens are of the form # NUM, then
505 process the directive, otherwise back up. */
506 token = _cpp_lex_direct (pfile);
507 if (token->type == CPP_HASH)
509 pfile->state.in_directive = 1;
510 token1 = _cpp_lex_direct (pfile);
511 _cpp_backup_tokens (pfile, 1);
512 pfile->state.in_directive = 0;
514 /* If it's a #line directive, handle it. */
515 if (token1->type == CPP_NUMBER)
517 _cpp_handle_directive (pfile, token->flags & PREV_WHITE);
518 read_original_directory (pfile);
519 return;
523 /* Backup as if nothing happened. */
524 _cpp_backup_tokens (pfile, 1);
527 /* For preprocessed files, if the tokens following the first filename
528 line is of the form # <line> "/path/name//", handle the
529 directive so we know the original current directory. */
530 static void
531 read_original_directory (cpp_reader *pfile)
533 const cpp_token *hash, *token;
535 /* Lex ahead; if the first tokens are of the form # NUM, then
536 process the directive, otherwise back up. */
537 hash = _cpp_lex_direct (pfile);
538 if (hash->type != CPP_HASH)
540 _cpp_backup_tokens (pfile, 1);
541 return;
544 token = _cpp_lex_direct (pfile);
546 if (token->type != CPP_NUMBER)
548 _cpp_backup_tokens (pfile, 2);
549 return;
552 token = _cpp_lex_direct (pfile);
554 if (token->type != CPP_STRING
555 || ! (token->val.str.len >= 5
556 && token->val.str.text[token->val.str.len-2] == '/'
557 && token->val.str.text[token->val.str.len-3] == '/'))
559 _cpp_backup_tokens (pfile, 3);
560 return;
563 if (pfile->cb.dir_change)
565 char *debugdir = (char *) alloca (token->val.str.len - 3);
567 memcpy (debugdir, (const char *) token->val.str.text + 1,
568 token->val.str.len - 4);
569 debugdir[token->val.str.len - 4] = '\0';
571 pfile->cb.dir_change (pfile, debugdir);
575 /* This is called at the end of preprocessing. It pops the last
576 buffer and writes dependency output, and returns the number of
577 errors.
579 Maybe it should also reset state, such that you could call
580 cpp_start_read with a new filename to restart processing. */
582 cpp_finish (cpp_reader *pfile, FILE *deps_stream)
584 /* Warn about unused macros before popping the final buffer. */
585 if (CPP_OPTION (pfile, warn_unused_macros))
586 cpp_forall_identifiers (pfile, _cpp_warn_if_unused_macro, NULL);
588 /* lex.c leaves the final buffer on the stack. This it so that
589 it returns an unending stream of CPP_EOFs to the client. If we
590 popped the buffer, we'd dereference a NULL buffer pointer and
591 segfault. It's nice to allow the client to do worry-free excess
592 cpp_get_token calls. */
593 while (pfile->buffer)
594 _cpp_pop_buffer (pfile);
596 /* Don't write the deps file if there are errors. */
597 if (CPP_OPTION (pfile, deps.style) != DEPS_NONE
598 && deps_stream && pfile->errors == 0)
600 deps_write (pfile->deps, deps_stream, 72);
602 if (CPP_OPTION (pfile, deps.phony_targets))
603 deps_phony_targets (pfile->deps, deps_stream);
606 /* Report on headers that could use multiple include guards. */
607 if (CPP_OPTION (pfile, print_include_names))
608 _cpp_report_missing_guards (pfile);
610 return pfile->errors;
613 static void
614 post_options (cpp_reader *pfile)
616 /* -Wtraditional is not useful in C++ mode. */
617 if (CPP_OPTION (pfile, cplusplus))
618 CPP_OPTION (pfile, warn_traditional) = 0;
620 /* Permanently disable macro expansion if we are rescanning
621 preprocessed text. Read preprocesed source in ISO mode. */
622 if (CPP_OPTION (pfile, preprocessed))
624 pfile->state.prevent_expansion = 1;
625 CPP_OPTION (pfile, traditional) = 0;
628 if (CPP_OPTION (pfile, warn_trigraphs) == 2)
629 CPP_OPTION (pfile, warn_trigraphs) = !CPP_OPTION (pfile, trigraphs);
631 if (CPP_OPTION (pfile, traditional))
633 CPP_OPTION (pfile, cplusplus_comments) = 0;
635 /* Traditional CPP does not accurately track column information. */
636 CPP_OPTION (pfile, show_column) = 0;
637 CPP_OPTION (pfile, trigraphs) = 0;
638 CPP_OPTION (pfile, warn_trigraphs) = 0;