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. */
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
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] = {
45 #define s(p, v) [p] = v,
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;
54 #define s(p, v) x[p] = v;
59 s('=', '#') s(')', ']') s('!', '|')
60 s('(', '[') s('\'', '^') s('>', '}')
61 s('/', '\\') s('<', '{') s('-', '~')
68 /* A set of booleans indicating what CPP features each source language
74 char extended_numbers
;
76 char cplusplus_comments
;
80 static const struct lang_flags lang_defaults
[] =
81 { /* c99 c++ xnum std // digr */
82 /* GNUC89 */ { 0, 0, 1, 0, 1, 1 },
83 /* GNUC99 */ { 1, 0, 1, 0, 1, 1 },
84 /* STDC89 */ { 0, 0, 0, 1, 0, 0 },
85 /* STDC94 */ { 0, 0, 0, 1, 0, 1 },
86 /* STDC99 */ { 1, 0, 1, 1, 1, 1 },
87 /* GNUCXX */ { 0, 1, 1, 0, 1, 1 },
88 /* CXX98 */ { 0, 1, 1, 1, 1, 1 },
89 /* ASM */ { 0, 0, 1, 0, 1, 0 }
92 /* Sets internal flags correctly for a given language. */
94 cpp_set_lang (cpp_reader
*pfile
, enum c_lang lang
)
96 const struct lang_flags
*l
= &lang_defaults
[(int) lang
];
98 CPP_OPTION (pfile
, lang
) = lang
;
100 CPP_OPTION (pfile
, c99
) = l
->c99
;
101 CPP_OPTION (pfile
, cplusplus
) = l
->cplusplus
;
102 CPP_OPTION (pfile
, extended_numbers
) = l
->extended_numbers
;
103 CPP_OPTION (pfile
, std
) = l
->std
;
104 CPP_OPTION (pfile
, trigraphs
) = l
->std
;
105 CPP_OPTION (pfile
, cplusplus_comments
) = l
->cplusplus_comments
;
106 CPP_OPTION (pfile
, digraphs
) = l
->digraphs
;
109 /* Initialize library global state. */
113 static int initialized
= 0;
119 /* Set up the trigraph map. This doesn't need to do anything if
120 we were compiled with a compiler that supports C99 designated
122 init_trigraph_map ();
125 (void) bindtextdomain (PACKAGE
, LOCALEDIR
);
130 /* Initialize a cpp_reader structure. */
132 cpp_create_reader (enum c_lang lang
, hash_table
*table
,
133 struct line_maps
*line_table
)
137 /* Initialize this instance of the library if it hasn't been already. */
140 pfile
= xcalloc (1, sizeof (cpp_reader
));
142 cpp_set_lang (pfile
, lang
);
143 CPP_OPTION (pfile
, warn_multichar
) = 1;
144 CPP_OPTION (pfile
, discard_comments
) = 1;
145 CPP_OPTION (pfile
, discard_comments_in_macro_exp
) = 1;
146 CPP_OPTION (pfile
, show_column
) = 1;
147 CPP_OPTION (pfile
, tabstop
) = 8;
148 CPP_OPTION (pfile
, operator_names
) = 1;
149 CPP_OPTION (pfile
, warn_trigraphs
) = 2;
150 CPP_OPTION (pfile
, warn_endif_labels
) = 1;
151 CPP_OPTION (pfile
, warn_deprecated
) = 1;
152 CPP_OPTION (pfile
, warn_long_long
) = !CPP_OPTION (pfile
, c99
);
153 CPP_OPTION (pfile
, dollars_in_ident
) = 1;
154 CPP_OPTION (pfile
, warn_dollars
) = 1;
155 CPP_OPTION (pfile
, warn_variadic_macros
) = 1;
157 /* Default CPP arithmetic to something sensible for the host for the
158 benefit of dumb users like fix-header. */
159 CPP_OPTION (pfile
, precision
) = CHAR_BIT
* sizeof (long);
160 CPP_OPTION (pfile
, char_precision
) = CHAR_BIT
;
161 CPP_OPTION (pfile
, wchar_precision
) = CHAR_BIT
* sizeof (int);
162 CPP_OPTION (pfile
, int_precision
) = CHAR_BIT
* sizeof (int);
163 CPP_OPTION (pfile
, unsigned_char
) = 0;
164 CPP_OPTION (pfile
, unsigned_wchar
) = 1;
165 CPP_OPTION (pfile
, bytes_big_endian
) = 1; /* does not matter */
167 /* Default to no charset conversion. */
168 CPP_OPTION (pfile
, narrow_charset
) = _cpp_default_encoding ();
169 CPP_OPTION (pfile
, wide_charset
) = 0;
171 /* Default the input character set to UTF-8. */
172 CPP_OPTION (pfile
, input_charset
) = _cpp_default_encoding ();
174 /* A fake empty "directory" used as the starting point for files
175 looked up without a search path. Name cannot be '/' because we
176 don't want to prepend anything at all to filenames using it. All
177 other entries are correct zero-initialized. */
178 pfile
->no_search_path
.name
= (char *) "";
180 /* Initialize the line map. */
181 pfile
->line_table
= line_table
;
183 /* Initialize lexer state. */
184 pfile
->state
.save_comments
= ! CPP_OPTION (pfile
, discard_comments
);
186 /* Set up static tokens. */
187 pfile
->avoid_paste
.type
= CPP_PADDING
;
188 pfile
->avoid_paste
.val
.source
= NULL
;
189 pfile
->eof
.type
= CPP_EOF
;
190 pfile
->eof
.flags
= 0;
192 /* Create a token buffer for the lexer. */
193 _cpp_init_tokenrun (&pfile
->base_run
, 250);
194 pfile
->cur_run
= &pfile
->base_run
;
195 pfile
->cur_token
= pfile
->base_run
.base
;
197 /* Initialize the base context. */
198 pfile
->context
= &pfile
->base_context
;
199 pfile
->base_context
.macro
= 0;
200 pfile
->base_context
.prev
= pfile
->base_context
.next
= 0;
202 /* Aligned and unaligned storage. */
203 pfile
->a_buff
= _cpp_get_buff (pfile
, 0);
204 pfile
->u_buff
= _cpp_get_buff (pfile
, 0);
206 /* The expression parser stack. */
207 _cpp_expand_op_stack (pfile
);
209 /* Initialize the buffer obstack. */
210 _obstack_begin (&pfile
->buffer_ob
, 0, 0,
211 (void *(*) (long)) xmalloc
,
212 (void (*) (void *)) free
);
214 _cpp_init_files (pfile
);
216 _cpp_init_hashtable (pfile
, table
);
221 /* Free resources used by PFILE. Accessing PFILE after this function
222 returns leads to undefined behavior. Returns the error count. */
224 cpp_destroy (cpp_reader
*pfile
)
226 cpp_context
*context
, *contextn
;
227 tokenrun
*run
, *runn
;
229 free (pfile
->op_stack
);
231 while (CPP_BUFFER (pfile
) != NULL
)
232 _cpp_pop_buffer (pfile
);
235 free (pfile
->out
.base
);
237 if (pfile
->macro_buffer
)
239 free (pfile
->macro_buffer
);
240 pfile
->macro_buffer
= NULL
;
241 pfile
->macro_buffer_len
= 0;
245 deps_free (pfile
->deps
);
246 obstack_free (&pfile
->buffer_ob
, 0);
248 _cpp_destroy_hashtable (pfile
);
249 _cpp_cleanup_files (pfile
);
250 _cpp_destroy_iconv (pfile
);
252 _cpp_free_buff (pfile
->a_buff
);
253 _cpp_free_buff (pfile
->u_buff
);
254 _cpp_free_buff (pfile
->free_buffs
);
256 for (run
= &pfile
->base_run
; run
; run
= runn
)
260 if (run
!= &pfile
->base_run
)
264 for (context
= pfile
->base_context
.next
; context
; context
= contextn
)
266 contextn
= context
->next
;
273 /* This structure defines one built-in identifier. A node will be
274 entered in the hash table under the name NAME, with value VALUE.
276 There are two tables of these. builtin_array holds all the
277 "builtin" macros: these are handled by builtin_macro() in
278 macro.c. Builtin is somewhat of a misnomer -- the property of
279 interest is that these macros require special code to compute their
280 expansions. The value is a "builtin_type" enumerator.
282 operator_array holds the C++ named operators. These are keywords
283 which act as aliases for punctuators. In C++, they cannot be
284 altered through #define, and #if recognizes them as operators. In
285 C, these are not entered into the hash table at all (but see
286 <iso646.h>). The value is a token-type enumerator. */
291 unsigned short value
;
294 #define B(n, t) { DSC(n), t }
295 static const struct builtin builtin_array
[] =
297 B("__TIME__", BT_TIME
),
298 B("__DATE__", BT_DATE
),
299 B("__FILE__", BT_FILE
),
300 B("__BASE_FILE__", BT_BASE_FILE
),
301 B("__LINE__", BT_SPECLINE
),
302 B("__INCLUDE_LEVEL__", BT_INCLUDE_LEVEL
),
303 /* Keep builtins not used for -traditional-cpp at the end, and
304 update init_builtins() if any more are added. */
305 B("_Pragma", BT_PRAGMA
),
306 B("__STDC__", BT_STDC
),
309 static const struct builtin operator_array
[] =
311 B("and", CPP_AND_AND
),
312 B("and_eq", CPP_AND_EQ
),
313 B("bitand", CPP_AND
),
315 B("compl", CPP_COMPL
),
317 B("not_eq", CPP_NOT_EQ
),
319 B("or_eq", CPP_OR_EQ
),
321 B("xor_eq", CPP_XOR_EQ
)
325 /* Mark the C++ named operators in the hash table. */
327 mark_named_operators (cpp_reader
*pfile
)
329 const struct builtin
*b
;
331 for (b
= operator_array
;
332 b
< (operator_array
+ ARRAY_SIZE (operator_array
));
335 cpp_hashnode
*hp
= cpp_lookup (pfile
, b
->name
, b
->len
);
336 hp
->flags
|= NODE_OPERATOR
;
337 hp
->is_directive
= 0;
338 hp
->directive_index
= b
->value
;
342 /* Read the builtins table above and enter them, and language-specific
343 macros, into the hash table. HOSTED is true if this is a hosted
346 cpp_init_builtins (cpp_reader
*pfile
, int hosted
)
348 const struct builtin
*b
;
349 size_t n
= ARRAY_SIZE (builtin_array
);
351 if (CPP_OPTION (pfile
, traditional
))
354 for(b
= builtin_array
; b
< builtin_array
+ n
; b
++)
356 cpp_hashnode
*hp
= cpp_lookup (pfile
, b
->name
, b
->len
);
358 hp
->flags
|= NODE_BUILTIN
| NODE_WARN
;
359 hp
->value
.builtin
= b
->value
;
362 if (CPP_OPTION (pfile
, cplusplus
))
363 _cpp_define_builtin (pfile
, "__cplusplus 1");
364 else if (CPP_OPTION (pfile
, lang
) == CLK_ASM
)
365 _cpp_define_builtin (pfile
, "__ASSEMBLER__ 1");
366 else if (CPP_OPTION (pfile
, lang
) == CLK_STDC94
)
367 _cpp_define_builtin (pfile
, "__STDC_VERSION__ 199409L");
368 else if (CPP_OPTION (pfile
, c99
))
369 _cpp_define_builtin (pfile
, "__STDC_VERSION__ 199901L");
372 _cpp_define_builtin (pfile
, "__STDC_HOSTED__ 1");
374 _cpp_define_builtin (pfile
, "__STDC_HOSTED__ 0");
376 if (CPP_OPTION (pfile
, objc
))
377 _cpp_define_builtin (pfile
, "__OBJC__ 1");
380 /* Sanity-checks are dependent on command-line options, so it is
381 called as a subroutine of cpp_read_main_file (). */
383 static void sanity_checks (cpp_reader
*);
384 static void sanity_checks (cpp_reader
*pfile
)
387 size_t max_precision
= 2 * CHAR_BIT
* sizeof (cpp_num_part
);
389 /* Sanity checks for assumptions about CPP arithmetic and target
390 type precisions made by cpplib. */
393 cpp_error (pfile
, CPP_DL_ICE
, "cppchar_t must be an unsigned type");
395 if (CPP_OPTION (pfile
, precision
) > max_precision
)
396 cpp_error (pfile
, CPP_DL_ICE
,
397 "preprocessor arithmetic has maximum precision of %lu bits;"
398 " target requires %lu bits",
399 (unsigned long) max_precision
,
400 (unsigned long) CPP_OPTION (pfile
, precision
));
402 if (CPP_OPTION (pfile
, precision
) < CPP_OPTION (pfile
, int_precision
))
403 cpp_error (pfile
, CPP_DL_ICE
,
404 "CPP arithmetic must be at least as precise as a target int");
406 if (CPP_OPTION (pfile
, char_precision
) < 8)
407 cpp_error (pfile
, CPP_DL_ICE
, "target char is less than 8 bits wide");
409 if (CPP_OPTION (pfile
, wchar_precision
) < CPP_OPTION (pfile
, char_precision
))
410 cpp_error (pfile
, CPP_DL_ICE
,
411 "target wchar_t is narrower than target char");
413 if (CPP_OPTION (pfile
, int_precision
) < CPP_OPTION (pfile
, char_precision
))
414 cpp_error (pfile
, CPP_DL_ICE
,
415 "target int is narrower than target char");
417 /* This is assumed in eval_token() and could be fixed if necessary. */
418 if (sizeof (cppchar_t
) > sizeof (cpp_num_part
))
419 cpp_error (pfile
, CPP_DL_ICE
,
420 "CPP half-integer narrower than CPP character");
422 if (CPP_OPTION (pfile
, wchar_precision
) > BITS_PER_CPPCHAR_T
)
423 cpp_error (pfile
, CPP_DL_ICE
,
424 "CPP on this host cannot handle wide character constants over"
425 " %lu bits, but the target requires %lu bits",
426 (unsigned long) BITS_PER_CPPCHAR_T
,
427 (unsigned long) CPP_OPTION (pfile
, wchar_precision
));
430 # define sanity_checks(PFILE)
433 /* This is called after options have been parsed, and partially
436 cpp_post_options (cpp_reader
*pfile
)
438 sanity_checks (pfile
);
440 post_options (pfile
);
442 /* Mark named operators before handling command line macros. */
443 if (CPP_OPTION (pfile
, cplusplus
) && CPP_OPTION (pfile
, operator_names
))
444 mark_named_operators (pfile
);
447 /* Setup for processing input from the file named FNAME, or stdin if
448 it is the empty string. Return the original filename
449 on success (e.g. foo.i->foo.c), or NULL on failure. */
451 cpp_read_main_file (cpp_reader
*pfile
, const char *fname
)
453 if (CPP_OPTION (pfile
, deps
.style
) != DEPS_NONE
)
456 pfile
->deps
= deps_init ();
458 /* Set the default target (if there is none already). */
459 deps_add_default_target (pfile
->deps
, fname
);
463 = _cpp_find_file (pfile
, fname
, &pfile
->no_search_path
, false);
464 if (_cpp_find_failed (pfile
->main_file
))
467 _cpp_stack_file (pfile
, pfile
->main_file
, false);
469 /* For foo.i, read the original filename foo.c now, for the benefit
470 of the front ends. */
471 if (CPP_OPTION (pfile
, preprocessed
))
473 read_original_filename (pfile
);
474 fname
= pfile
->line_table
->maps
[pfile
->line_table
->used
-1].to_file
;
479 /* For preprocessed files, if the first tokens are of the form # NUM.
480 handle the directive so we know the original file name. This will
481 generate file_change callbacks, which the front ends must handle
482 appropriately given their state of initialization. */
484 read_original_filename (cpp_reader
*pfile
)
486 const cpp_token
*token
, *token1
;
488 /* Lex ahead; if the first tokens are of the form # NUM, then
489 process the directive, otherwise back up. */
490 token
= _cpp_lex_direct (pfile
);
491 if (token
->type
== CPP_HASH
)
493 token1
= _cpp_lex_direct (pfile
);
494 _cpp_backup_tokens (pfile
, 1);
496 /* If it's a #line directive, handle it. */
497 if (token1
->type
== CPP_NUMBER
)
499 _cpp_handle_directive (pfile
, token
->flags
& PREV_WHITE
);
500 read_original_directory (pfile
);
505 /* Backup as if nothing happened. */
506 _cpp_backup_tokens (pfile
, 1);
509 /* For preprocessed files, if the tokens following the first filename
510 line is of the form # <line> "/path/name//", handle the
511 directive so we know the original current directory. */
513 read_original_directory (cpp_reader
*pfile
)
515 const cpp_token
*hash
, *token
;
517 /* Lex ahead; if the first tokens are of the form # NUM, then
518 process the directive, otherwise back up. */
519 hash
= _cpp_lex_direct (pfile
);
520 if (hash
->type
!= CPP_HASH
)
522 _cpp_backup_tokens (pfile
, 1);
526 token
= _cpp_lex_direct (pfile
);
528 if (token
->type
!= CPP_NUMBER
)
530 _cpp_backup_tokens (pfile
, 2);
534 token
= _cpp_lex_direct (pfile
);
536 if (token
->type
!= CPP_STRING
537 || ! (token
->val
.str
.len
>= 5
538 && token
->val
.str
.text
[token
->val
.str
.len
-2] == '/'
539 && token
->val
.str
.text
[token
->val
.str
.len
-3] == '/'))
541 _cpp_backup_tokens (pfile
, 3);
545 if (pfile
->cb
.dir_change
)
547 char *debugdir
= alloca (token
->val
.str
.len
- 3);
549 memcpy (debugdir
, (const char *) token
->val
.str
.text
+ 1,
550 token
->val
.str
.len
- 4);
551 debugdir
[token
->val
.str
.len
- 4] = '\0';
553 pfile
->cb
.dir_change (pfile
, debugdir
);
557 /* This is called at the end of preprocessing. It pops the last
558 buffer and writes dependency output, and returns the number of
561 Maybe it should also reset state, such that you could call
562 cpp_start_read with a new filename to restart processing. */
564 cpp_finish (cpp_reader
*pfile
, FILE *deps_stream
)
566 /* Warn about unused macros before popping the final buffer. */
567 if (CPP_OPTION (pfile
, warn_unused_macros
))
568 cpp_forall_identifiers (pfile
, _cpp_warn_if_unused_macro
, NULL
);
570 /* lex.c leaves the final buffer on the stack. This it so that
571 it returns an unending stream of CPP_EOFs to the client. If we
572 popped the buffer, we'd dereference a NULL buffer pointer and
573 segfault. It's nice to allow the client to do worry-free excess
574 cpp_get_token calls. */
575 while (pfile
->buffer
)
576 _cpp_pop_buffer (pfile
);
578 /* Don't write the deps file if there are errors. */
579 if (CPP_OPTION (pfile
, deps
.style
) != DEPS_NONE
580 && deps_stream
&& pfile
->errors
== 0)
582 deps_write (pfile
->deps
, deps_stream
, 72);
584 if (CPP_OPTION (pfile
, deps
.phony_targets
))
585 deps_phony_targets (pfile
->deps
, deps_stream
);
588 /* Report on headers that could use multiple include guards. */
589 if (CPP_OPTION (pfile
, print_include_names
))
590 _cpp_report_missing_guards (pfile
);
592 return pfile
->errors
;
596 post_options (cpp_reader
*pfile
)
598 /* -Wtraditional is not useful in C++ mode. */
599 if (CPP_OPTION (pfile
, cplusplus
))
600 CPP_OPTION (pfile
, warn_traditional
) = 0;
602 /* Permanently disable macro expansion if we are rescanning
603 preprocessed text. Read preprocesed source in ISO mode. */
604 if (CPP_OPTION (pfile
, preprocessed
))
606 pfile
->state
.prevent_expansion
= 1;
607 CPP_OPTION (pfile
, traditional
) = 0;
610 if (CPP_OPTION (pfile
, warn_trigraphs
) == 2)
611 CPP_OPTION (pfile
, warn_trigraphs
) = !CPP_OPTION (pfile
, trigraphs
);
613 if (CPP_OPTION (pfile
, traditional
))
615 CPP_OPTION (pfile
, cplusplus_comments
) = 0;
617 /* Traditional CPP does not accurately track column information. */
618 CPP_OPTION (pfile
, show_column
) = 0;
619 CPP_OPTION (pfile
, trigraphs
) = 0;
620 CPP_OPTION (pfile
, warn_trigraphs
) = 0;