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
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. */
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
;
75 char extended_identifiers
;
77 char cplusplus_comments
;
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. */
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. */
121 static int initialized
= 0;
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
130 init_trigraph_map ();
133 (void) bindtextdomain (PACKAGE
, LOCALEDIR
);
138 /* Initialize a cpp_reader structure. */
140 cpp_create_reader (enum c_lang lang
, hash_table
*table
,
141 struct line_maps
*line_table
)
145 /* Initialize this instance of the library if it hasn't been already. */
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
);
230 /* Free resources used by PFILE. Accessing PFILE after this function
231 returns leads to undefined behavior. Returns the error count. */
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
);
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;
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
)
269 if (run
!= &pfile
->base_run
)
273 for (context
= pfile
->base_context
.next
; context
; context
= contextn
)
275 contextn
= context
->next
;
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. */
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 B("__COUNTER__", BT_COUNTER
),
314 /* Keep builtins not used for -traditional-cpp at the end, and
315 update init_builtins() if any more are added. */
316 B("_Pragma", BT_PRAGMA
),
317 B("__STDC__", BT_STDC
),
320 static const struct builtin operator_array
[] =
322 B("and", CPP_AND_AND
),
323 B("and_eq", CPP_AND_EQ
),
324 B("bitand", CPP_AND
),
326 B("compl", CPP_COMPL
),
328 B("not_eq", CPP_NOT_EQ
),
330 B("or_eq", CPP_OR_EQ
),
332 B("xor_eq", CPP_XOR_EQ
)
336 /* Mark the C++ named operators in the hash table. */
338 mark_named_operators (cpp_reader
*pfile
)
340 const struct builtin
*b
;
342 for (b
= operator_array
;
343 b
< (operator_array
+ ARRAY_SIZE (operator_array
));
346 cpp_hashnode
*hp
= cpp_lookup (pfile
, b
->name
, b
->len
);
347 hp
->flags
|= NODE_OPERATOR
;
348 hp
->is_directive
= 0;
349 hp
->directive_index
= b
->value
;
353 /* Read the builtins table above and enter them, and language-specific
354 macros, into the hash table. HOSTED is true if this is a hosted
357 cpp_init_builtins (cpp_reader
*pfile
, int hosted
)
359 const struct builtin
*b
;
360 size_t n
= ARRAY_SIZE (builtin_array
);
362 if (CPP_OPTION (pfile
, traditional
))
364 else if (! CPP_OPTION (pfile
, stdc_0_in_system_headers
)
365 || CPP_OPTION (pfile
, std
))
368 _cpp_define_builtin (pfile
, "__STDC__ 1");
371 for (b
= builtin_array
; b
< builtin_array
+ n
; b
++)
373 cpp_hashnode
*hp
= cpp_lookup (pfile
, b
->name
, b
->len
);
375 hp
->flags
|= NODE_BUILTIN
| NODE_WARN
;
376 hp
->value
.builtin
= (enum builtin_type
) b
->value
;
379 if (CPP_OPTION (pfile
, cplusplus
))
380 _cpp_define_builtin (pfile
, "__cplusplus 1");
381 else if (CPP_OPTION (pfile
, lang
) == CLK_ASM
)
382 _cpp_define_builtin (pfile
, "__ASSEMBLER__ 1");
383 else if (CPP_OPTION (pfile
, lang
) == CLK_STDC94
)
384 _cpp_define_builtin (pfile
, "__STDC_VERSION__ 199409L");
385 else if (CPP_OPTION (pfile
, c99
))
386 _cpp_define_builtin (pfile
, "__STDC_VERSION__ 199901L");
389 _cpp_define_builtin (pfile
, "__STDC_HOSTED__ 1");
391 _cpp_define_builtin (pfile
, "__STDC_HOSTED__ 0");
393 if (CPP_OPTION (pfile
, objc
))
394 _cpp_define_builtin (pfile
, "__OBJC__ 1");
397 /* Sanity-checks are dependent on command-line options, so it is
398 called as a subroutine of cpp_read_main_file (). */
400 static void sanity_checks (cpp_reader
*);
401 static void sanity_checks (cpp_reader
*pfile
)
404 size_t max_precision
= 2 * CHAR_BIT
* sizeof (cpp_num_part
);
406 /* Sanity checks for assumptions about CPP arithmetic and target
407 type precisions made by cpplib. */
410 cpp_error (pfile
, CPP_DL_ICE
, "cppchar_t must be an unsigned type");
412 if (CPP_OPTION (pfile
, precision
) > max_precision
)
413 cpp_error (pfile
, CPP_DL_ICE
,
414 "preprocessor arithmetic has maximum precision of %lu bits;"
415 " target requires %lu bits",
416 (unsigned long) max_precision
,
417 (unsigned long) CPP_OPTION (pfile
, precision
));
419 if (CPP_OPTION (pfile
, precision
) < CPP_OPTION (pfile
, int_precision
))
420 cpp_error (pfile
, CPP_DL_ICE
,
421 "CPP arithmetic must be at least as precise as a target int");
423 if (CPP_OPTION (pfile
, char_precision
) < 8)
424 cpp_error (pfile
, CPP_DL_ICE
, "target char is less than 8 bits wide");
426 if (CPP_OPTION (pfile
, wchar_precision
) < CPP_OPTION (pfile
, char_precision
))
427 cpp_error (pfile
, CPP_DL_ICE
,
428 "target wchar_t is narrower than target char");
430 if (CPP_OPTION (pfile
, int_precision
) < CPP_OPTION (pfile
, char_precision
))
431 cpp_error (pfile
, CPP_DL_ICE
,
432 "target int is narrower than target char");
434 /* This is assumed in eval_token() and could be fixed if necessary. */
435 if (sizeof (cppchar_t
) > sizeof (cpp_num_part
))
436 cpp_error (pfile
, CPP_DL_ICE
,
437 "CPP half-integer narrower than CPP character");
439 if (CPP_OPTION (pfile
, wchar_precision
) > BITS_PER_CPPCHAR_T
)
440 cpp_error (pfile
, CPP_DL_ICE
,
441 "CPP on this host cannot handle wide character constants over"
442 " %lu bits, but the target requires %lu bits",
443 (unsigned long) BITS_PER_CPPCHAR_T
,
444 (unsigned long) CPP_OPTION (pfile
, wchar_precision
));
447 # define sanity_checks(PFILE)
450 /* This is called after options have been parsed, and partially
453 cpp_post_options (cpp_reader
*pfile
)
455 sanity_checks (pfile
);
457 post_options (pfile
);
459 /* Mark named operators before handling command line macros. */
460 if (CPP_OPTION (pfile
, cplusplus
) && CPP_OPTION (pfile
, operator_names
))
461 mark_named_operators (pfile
);
464 /* Setup for processing input from the file named FNAME, or stdin if
465 it is the empty string. Return the original filename
466 on success (e.g. foo.i->foo.c), or NULL on failure. */
468 cpp_read_main_file (cpp_reader
*pfile
, const char *fname
)
470 if (CPP_OPTION (pfile
, deps
.style
) != DEPS_NONE
)
473 pfile
->deps
= deps_init ();
475 /* Set the default target (if there is none already). */
476 deps_add_default_target (pfile
->deps
, fname
);
480 = _cpp_find_file (pfile
, fname
, &pfile
->no_search_path
, false, 0);
481 if (_cpp_find_failed (pfile
->main_file
))
484 _cpp_stack_file (pfile
, pfile
->main_file
, false);
486 /* For foo.i, read the original filename foo.c now, for the benefit
487 of the front ends. */
488 if (CPP_OPTION (pfile
, preprocessed
))
490 read_original_filename (pfile
);
491 fname
= pfile
->line_table
->maps
[pfile
->line_table
->used
-1].to_file
;
496 /* For preprocessed files, if the first tokens are of the form # NUM.
497 handle the directive so we know the original file name. This will
498 generate file_change callbacks, which the front ends must handle
499 appropriately given their state of initialization. */
501 read_original_filename (cpp_reader
*pfile
)
503 const cpp_token
*token
, *token1
;
505 /* Lex ahead; if the first tokens are of the form # NUM, then
506 process the directive, otherwise back up. */
507 token
= _cpp_lex_direct (pfile
);
508 if (token
->type
== CPP_HASH
)
510 pfile
->state
.in_directive
= 1;
511 token1
= _cpp_lex_direct (pfile
);
512 _cpp_backup_tokens (pfile
, 1);
513 pfile
->state
.in_directive
= 0;
515 /* If it's a #line directive, handle it. */
516 if (token1
->type
== CPP_NUMBER
)
518 _cpp_handle_directive (pfile
, token
->flags
& PREV_WHITE
);
519 read_original_directory (pfile
);
524 /* Backup as if nothing happened. */
525 _cpp_backup_tokens (pfile
, 1);
528 /* For preprocessed files, if the tokens following the first filename
529 line is of the form # <line> "/path/name//", handle the
530 directive so we know the original current directory. */
532 read_original_directory (cpp_reader
*pfile
)
534 const cpp_token
*hash
, *token
;
536 /* Lex ahead; if the first tokens are of the form # NUM, then
537 process the directive, otherwise back up. */
538 hash
= _cpp_lex_direct (pfile
);
539 if (hash
->type
!= CPP_HASH
)
541 _cpp_backup_tokens (pfile
, 1);
545 token
= _cpp_lex_direct (pfile
);
547 if (token
->type
!= CPP_NUMBER
)
549 _cpp_backup_tokens (pfile
, 2);
553 token
= _cpp_lex_direct (pfile
);
555 if (token
->type
!= CPP_STRING
556 || ! (token
->val
.str
.len
>= 5
557 && token
->val
.str
.text
[token
->val
.str
.len
-2] == '/'
558 && token
->val
.str
.text
[token
->val
.str
.len
-3] == '/'))
560 _cpp_backup_tokens (pfile
, 3);
564 if (pfile
->cb
.dir_change
)
566 char *debugdir
= (char *) alloca (token
->val
.str
.len
- 3);
568 memcpy (debugdir
, (const char *) token
->val
.str
.text
+ 1,
569 token
->val
.str
.len
- 4);
570 debugdir
[token
->val
.str
.len
- 4] = '\0';
572 pfile
->cb
.dir_change (pfile
, debugdir
);
576 /* This is called at the end of preprocessing. It pops the last
577 buffer and writes dependency output, and returns the number of
580 Maybe it should also reset state, such that you could call
581 cpp_start_read with a new filename to restart processing. */
583 cpp_finish (cpp_reader
*pfile
, FILE *deps_stream
)
585 /* Warn about unused macros before popping the final buffer. */
586 if (CPP_OPTION (pfile
, warn_unused_macros
))
587 cpp_forall_identifiers (pfile
, _cpp_warn_if_unused_macro
, NULL
);
589 /* lex.c leaves the final buffer on the stack. This it so that
590 it returns an unending stream of CPP_EOFs to the client. If we
591 popped the buffer, we'd dereference a NULL buffer pointer and
592 segfault. It's nice to allow the client to do worry-free excess
593 cpp_get_token calls. */
594 while (pfile
->buffer
)
595 _cpp_pop_buffer (pfile
);
597 /* Don't write the deps file if there are errors. */
598 if (CPP_OPTION (pfile
, deps
.style
) != DEPS_NONE
599 && deps_stream
&& pfile
->errors
== 0)
601 deps_write (pfile
->deps
, deps_stream
, 72);
603 if (CPP_OPTION (pfile
, deps
.phony_targets
))
604 deps_phony_targets (pfile
->deps
, deps_stream
);
607 /* Report on headers that could use multiple include guards. */
608 if (CPP_OPTION (pfile
, print_include_names
))
609 _cpp_report_missing_guards (pfile
);
611 return pfile
->errors
;
615 post_options (cpp_reader
*pfile
)
617 /* -Wtraditional is not useful in C++ mode. */
618 if (CPP_OPTION (pfile
, cplusplus
))
619 CPP_OPTION (pfile
, warn_traditional
) = 0;
621 /* Permanently disable macro expansion if we are rescanning
622 preprocessed text. Read preprocesed source in ISO mode. */
623 if (CPP_OPTION (pfile
, preprocessed
))
625 pfile
->state
.prevent_expansion
= 1;
626 CPP_OPTION (pfile
, traditional
) = 0;
629 if (CPP_OPTION (pfile
, warn_trigraphs
) == 2)
630 CPP_OPTION (pfile
, warn_trigraphs
) = !CPP_OPTION (pfile
, trigraphs
);
632 if (CPP_OPTION (pfile
, traditional
))
634 CPP_OPTION (pfile
, cplusplus_comments
) = 0;
636 /* Traditional CPP does not accurately track column information. */
637 CPP_OPTION (pfile
, show_column
) = 0;
638 CPP_OPTION (pfile
, trigraphs
) = 0;
639 CPP_OPTION (pfile
, warn_trigraphs
) = 0;