2 Copyright (C) 1986, 1987, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
3 1999, 2000, 2001, 2002, 2003 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 no charset conversion. */
161 CPP_OPTION (pfile
, narrow_charset
) = 0;
162 CPP_OPTION (pfile
, wide_charset
) = 0;
164 /* A fake empty "directory" used as the starting point for files
165 looked up without a search path. Name cannot be '/' because we
166 don't want to prepend anything at all to filenames using it. All
167 other entries are correct zero-initialized. */
168 pfile
->no_search_path
.name
= (char *) "";
170 /* Initialize the line map. Start at logical line 1, so we can use
171 a line number of zero for special states. */
172 linemap_init (&pfile
->line_maps
);
175 /* Initialize lexer state. */
176 pfile
->state
.save_comments
= ! CPP_OPTION (pfile
, discard_comments
);
178 /* Set up static tokens. */
179 pfile
->avoid_paste
.type
= CPP_PADDING
;
180 pfile
->avoid_paste
.val
.source
= NULL
;
181 pfile
->eof
.type
= CPP_EOF
;
182 pfile
->eof
.flags
= 0;
184 /* Create a token buffer for the lexer. */
185 _cpp_init_tokenrun (&pfile
->base_run
, 250);
186 pfile
->cur_run
= &pfile
->base_run
;
187 pfile
->cur_token
= pfile
->base_run
.base
;
189 /* Initialize the base context. */
190 pfile
->context
= &pfile
->base_context
;
191 pfile
->base_context
.macro
= 0;
192 pfile
->base_context
.prev
= pfile
->base_context
.next
= 0;
194 /* Aligned and unaligned storage. */
195 pfile
->a_buff
= _cpp_get_buff (pfile
, 0);
196 pfile
->u_buff
= _cpp_get_buff (pfile
, 0);
198 /* The expression parser stack. */
199 _cpp_expand_op_stack (pfile
);
201 /* Initialize the buffer obstack. */
202 _obstack_begin (&pfile
->buffer_ob
, 0, 0,
203 (void *(*) (long)) xmalloc
,
204 (void (*) (void *)) free
);
206 _cpp_init_files (pfile
);
208 _cpp_init_hashtable (pfile
, table
);
213 /* Free resources used by PFILE. Accessing PFILE after this function
214 returns leads to undefined behavior. Returns the error count. */
216 cpp_destroy (cpp_reader
*pfile
)
218 cpp_context
*context
, *contextn
;
219 tokenrun
*run
, *runn
;
221 free (pfile
->op_stack
);
223 while (CPP_BUFFER (pfile
) != NULL
)
224 _cpp_pop_buffer (pfile
);
227 free (pfile
->out
.base
);
229 if (pfile
->macro_buffer
)
231 free (pfile
->macro_buffer
);
232 pfile
->macro_buffer
= NULL
;
233 pfile
->macro_buffer_len
= 0;
237 deps_free (pfile
->deps
);
238 obstack_free (&pfile
->buffer_ob
, 0);
240 _cpp_destroy_hashtable (pfile
);
241 _cpp_cleanup_files (pfile
);
242 _cpp_destroy_iconv (pfile
);
244 _cpp_free_buff (pfile
->a_buff
);
245 _cpp_free_buff (pfile
->u_buff
);
246 _cpp_free_buff (pfile
->free_buffs
);
248 for (run
= &pfile
->base_run
; run
; run
= runn
)
252 if (run
!= &pfile
->base_run
)
256 for (context
= pfile
->base_context
.next
; context
; context
= contextn
)
258 contextn
= context
->next
;
262 linemap_free (&pfile
->line_maps
);
266 /* This structure defines one built-in identifier. A node will be
267 entered in the hash table under the name NAME, with value VALUE.
269 There are two tables of these. builtin_array holds all the
270 "builtin" macros: these are handled by builtin_macro() in
271 cppmacro.c. Builtin is somewhat of a misnomer -- the property of
272 interest is that these macros require special code to compute their
273 expansions. The value is a "builtin_type" enumerator.
275 operator_array holds the C++ named operators. These are keywords
276 which act as aliases for punctuators. In C++, they cannot be
277 altered through #define, and #if recognizes them as operators. In
278 C, these are not entered into the hash table at all (but see
279 <iso646.h>). The value is a token-type enumerator. */
284 unsigned short value
;
287 #define B(n, t) { DSC(n), t }
288 static const struct builtin builtin_array
[] =
290 B("__TIME__", BT_TIME
),
291 B("__DATE__", BT_DATE
),
292 B("__FILE__", BT_FILE
),
293 B("__BASE_FILE__", BT_BASE_FILE
),
294 B("__LINE__", BT_SPECLINE
),
295 B("__INCLUDE_LEVEL__", BT_INCLUDE_LEVEL
),
296 /* Keep builtins not used for -traditional-cpp at the end, and
297 update init_builtins() if any more are added. */
298 B("_Pragma", BT_PRAGMA
),
299 B("__STDC__", BT_STDC
),
302 static const struct builtin operator_array
[] =
304 B("and", CPP_AND_AND
),
305 B("and_eq", CPP_AND_EQ
),
306 B("bitand", CPP_AND
),
308 B("compl", CPP_COMPL
),
310 B("not_eq", CPP_NOT_EQ
),
312 B("or_eq", CPP_OR_EQ
),
314 B("xor_eq", CPP_XOR_EQ
)
318 /* Mark the C++ named operators in the hash table. */
320 mark_named_operators (cpp_reader
*pfile
)
322 const struct builtin
*b
;
324 for (b
= operator_array
;
325 b
< (operator_array
+ ARRAY_SIZE (operator_array
));
328 cpp_hashnode
*hp
= cpp_lookup (pfile
, b
->name
, b
->len
);
329 hp
->flags
|= NODE_OPERATOR
;
330 hp
->is_directive
= 0;
331 hp
->directive_index
= b
->value
;
335 /* Read the builtins table above and enter them, and language-specific
336 macros, into the hash table. HOSTED is true if this is a hosted
339 cpp_init_builtins (cpp_reader
*pfile
, int hosted
)
341 const struct builtin
*b
;
342 size_t n
= ARRAY_SIZE (builtin_array
);
344 if (CPP_OPTION (pfile
, traditional
))
347 for(b
= builtin_array
; b
< builtin_array
+ n
; b
++)
349 cpp_hashnode
*hp
= cpp_lookup (pfile
, b
->name
, b
->len
);
351 hp
->flags
|= NODE_BUILTIN
| NODE_WARN
;
352 hp
->value
.builtin
= b
->value
;
355 if (CPP_OPTION (pfile
, cplusplus
))
356 _cpp_define_builtin (pfile
, "__cplusplus 1");
357 else if (CPP_OPTION (pfile
, lang
) == CLK_ASM
)
358 _cpp_define_builtin (pfile
, "__ASSEMBLER__ 1");
359 else if (CPP_OPTION (pfile
, lang
) == CLK_STDC94
)
360 _cpp_define_builtin (pfile
, "__STDC_VERSION__ 199409L");
361 else if (CPP_OPTION (pfile
, c99
))
362 _cpp_define_builtin (pfile
, "__STDC_VERSION__ 199901L");
365 _cpp_define_builtin (pfile
, "__STDC_HOSTED__ 1");
367 _cpp_define_builtin (pfile
, "__STDC_HOSTED__ 0");
369 if (CPP_OPTION (pfile
, objc
))
370 _cpp_define_builtin (pfile
, "__OBJC__ 1");
373 /* Sanity-checks are dependent on command-line options, so it is
374 called as a subroutine of cpp_read_main_file (). */
376 static void sanity_checks (cpp_reader
*);
377 static void sanity_checks (cpp_reader
*pfile
)
380 size_t max_precision
= 2 * CHAR_BIT
* sizeof (cpp_num_part
);
382 /* Sanity checks for assumptions about CPP arithmetic and target
383 type precisions made by cpplib. */
386 cpp_error (pfile
, DL_ICE
, "cppchar_t must be an unsigned type");
388 if (CPP_OPTION (pfile
, precision
) > max_precision
)
389 cpp_error (pfile
, DL_ICE
,
390 "preprocessor arithmetic has maximum precision of %lu bits;"
391 " target requires %lu bits",
392 (unsigned long) max_precision
,
393 (unsigned long) CPP_OPTION (pfile
, precision
));
395 if (CPP_OPTION (pfile
, precision
) < CPP_OPTION (pfile
, int_precision
))
396 cpp_error (pfile
, DL_ICE
,
397 "CPP arithmetic must be at least as precise as a target int");
399 if (CPP_OPTION (pfile
, char_precision
) < 8)
400 cpp_error (pfile
, DL_ICE
, "target char is less than 8 bits wide");
402 if (CPP_OPTION (pfile
, wchar_precision
) < CPP_OPTION (pfile
, char_precision
))
403 cpp_error (pfile
, DL_ICE
,
404 "target wchar_t is narrower than target char");
406 if (CPP_OPTION (pfile
, int_precision
) < CPP_OPTION (pfile
, char_precision
))
407 cpp_error (pfile
, DL_ICE
,
408 "target int is narrower than target char");
410 /* This is assumed in eval_token() and could be fixed if necessary. */
411 if (sizeof (cppchar_t
) > sizeof (cpp_num_part
))
412 cpp_error (pfile
, DL_ICE
, "CPP half-integer narrower than CPP character");
414 if (CPP_OPTION (pfile
, wchar_precision
) > BITS_PER_CPPCHAR_T
)
415 cpp_error (pfile
, DL_ICE
,
416 "CPP on this host cannot handle wide character constants over"
417 " %lu bits, but the target requires %lu bits",
418 (unsigned long) BITS_PER_CPPCHAR_T
,
419 (unsigned long) CPP_OPTION (pfile
, wchar_precision
));
422 # define sanity_checks(PFILE)
425 /* Add a dependency target. Can be called any number of times before
426 cpp_read_main_file(). If no targets have been added before
427 cpp_read_main_file(), then the default target is used. */
429 cpp_add_dependency_target (cpp_reader
*pfile
, const char *target
, int quote
)
432 pfile
->deps
= deps_init ();
434 deps_add_target (pfile
->deps
, target
, quote
);
437 /* This is called after options have been parsed, and partially
440 cpp_post_options (cpp_reader
*pfile
)
442 sanity_checks (pfile
);
444 post_options (pfile
);
446 /* Mark named operators before handling command line macros. */
447 if (CPP_OPTION (pfile
, cplusplus
) && CPP_OPTION (pfile
, operator_names
))
448 mark_named_operators (pfile
);
451 /* Setup for processing input from the file named FNAME,
452 or stdin if it is the empty string. Return the original filename
453 on success (e.g. foo.i->foo.c), or NULL on failure. */
455 cpp_read_main_file (cpp_reader
*pfile
, const char *fname
)
457 if (CPP_OPTION (pfile
, deps
.style
) != DEPS_NONE
)
460 pfile
->deps
= deps_init ();
462 /* Set the default target (if there is none already). */
463 deps_add_default_target (pfile
->deps
, fname
);
466 if (!_cpp_stack_file (pfile
, fname
))
469 /* Set this here so the client can change the option if it wishes,
470 and after stacking the main file so we don't trace the main
472 pfile
->line_maps
.trace_includes
= CPP_OPTION (pfile
, print_include_names
);
474 /* For foo.i, read the original filename foo.c now, for the benefit
475 of the front ends. */
476 if (CPP_OPTION (pfile
, preprocessed
))
477 read_original_filename (pfile
);
479 if (CPP_OPTION (pfile
, working_directory
))
481 const char *name
= pfile
->map
->to_file
;
482 const char *dir
= getpwd ();
483 char *dir_with_slashes
= alloca (strlen (dir
) + 3);
485 memcpy (dir_with_slashes
, dir
, strlen (dir
));
486 memcpy (dir_with_slashes
+ strlen (dir
), "//", 3);
488 if (pfile
->cb
.dir_change
)
489 pfile
->cb
.dir_change (pfile
, dir
);
490 /* Emit file renames that will be recognized by
491 read_directory_filename, since dir_change doesn't output
493 _cpp_do_file_change (pfile
, LC_RENAME
, dir_with_slashes
, 1, 0);
494 _cpp_do_file_change (pfile
, LC_RENAME
, name
, 1, 0);
497 return pfile
->map
->to_file
;
500 /* For preprocessed files, if the first tokens are of the form # NUM.
501 handle the directive so we know the original file name. This will
502 generate file_change callbacks, which the front ends must handle
503 appropriately given their state of initialization. */
505 read_original_filename (cpp_reader
*pfile
)
507 const cpp_token
*token
, *token1
;
509 /* Lex ahead; if the first tokens are of the form # NUM, then
510 process the directive, otherwise back up. */
511 token
= _cpp_lex_direct (pfile
);
512 if (token
->type
== CPP_HASH
)
514 token1
= _cpp_lex_direct (pfile
);
515 _cpp_backup_tokens (pfile
, 1);
517 /* If it's a #line directive, handle it. */
518 if (token1
->type
== CPP_NUMBER
)
520 _cpp_handle_directive (pfile
, token
->flags
& PREV_WHITE
);
521 read_original_directory (pfile
);
526 /* Backup as if nothing happened. */
527 _cpp_backup_tokens (pfile
, 1);
530 /* For preprocessed files, if the tokens following the first filename
531 line is of the form # <line> "/path/name//", handle the
532 directive so we know the original current directory. */
534 read_original_directory (cpp_reader
*pfile
)
536 const cpp_token
*hash
, *token
;
538 /* Lex ahead; if the first tokens are of the form # NUM, then
539 process the directive, otherwise back up. */
540 hash
= _cpp_lex_direct (pfile
);
541 if (hash
->type
!= CPP_HASH
)
543 _cpp_backup_tokens (pfile
, 1);
547 token
= _cpp_lex_direct (pfile
);
549 if (token
->type
!= CPP_NUMBER
)
551 _cpp_backup_tokens (pfile
, 2);
555 token
= _cpp_lex_direct (pfile
);
557 if (token
->type
!= CPP_STRING
558 || ! (token
->val
.str
.len
>= 5
559 && token
->val
.str
.text
[token
->val
.str
.len
-2] == '/'
560 && token
->val
.str
.text
[token
->val
.str
.len
-3] == '/'))
562 _cpp_backup_tokens (pfile
, 3);
566 if (pfile
->cb
.dir_change
)
568 char *debugdir
= alloca (token
->val
.str
.len
- 3);
570 memcpy (debugdir
, (const char *) token
->val
.str
.text
+ 1,
571 token
->val
.str
.len
- 4);
572 debugdir
[token
->val
.str
.len
- 4] = '\0';
574 pfile
->cb
.dir_change (pfile
, debugdir
);
577 /* We want to process the fake line changes as regular changes, to
579 _cpp_backup_tokens (pfile
, 3);
581 CPP_OPTION (pfile
, working_directory
) = false;
584 /* This is called at the end of preprocessing. It pops the last
585 buffer and writes dependency output, and returns the number of
588 Maybe it should also reset state, such that you could call
589 cpp_start_read with a new filename to restart processing. */
591 cpp_finish (cpp_reader
*pfile
, FILE *deps_stream
)
593 /* Warn about unused macros before popping the final buffer. */
594 if (CPP_OPTION (pfile
, warn_unused_macros
))
595 cpp_forall_identifiers (pfile
, _cpp_warn_if_unused_macro
, NULL
);
597 /* cpplex.c leaves the final buffer on the stack. This it so that
598 it returns an unending stream of CPP_EOFs to the client. If we
599 popped the buffer, we'd dereference a NULL buffer pointer and
600 segfault. It's nice to allow the client to do worry-free excess
601 cpp_get_token calls. */
602 while (pfile
->buffer
)
603 _cpp_pop_buffer (pfile
);
605 /* Don't write the deps file if there are errors. */
606 if (CPP_OPTION (pfile
, deps
.style
) != DEPS_NONE
607 && deps_stream
&& pfile
->errors
== 0)
609 deps_write (pfile
->deps
, deps_stream
, 72);
611 if (CPP_OPTION (pfile
, deps
.phony_targets
))
612 deps_phony_targets (pfile
->deps
, deps_stream
);
615 /* Report on headers that could use multiple include guards. */
616 if (CPP_OPTION (pfile
, print_include_names
))
617 _cpp_report_missing_guards (pfile
);
619 return pfile
->errors
;
623 post_options (cpp_reader
*pfile
)
625 /* -Wtraditional is not useful in C++ mode. */
626 if (CPP_OPTION (pfile
, cplusplus
))
627 CPP_OPTION (pfile
, warn_traditional
) = 0;
629 /* Permanently disable macro expansion if we are rescanning
630 preprocessed text. Read preprocesed source in ISO mode. */
631 if (CPP_OPTION (pfile
, preprocessed
))
633 pfile
->state
.prevent_expansion
= 1;
634 CPP_OPTION (pfile
, traditional
) = 0;
637 if (CPP_OPTION (pfile
, warn_trigraphs
) == 2)
638 CPP_OPTION (pfile
, warn_trigraphs
) = !CPP_OPTION (pfile
, trigraphs
);
640 if (CPP_OPTION (pfile
, traditional
))
642 /* Traditional CPP does not accurately track column information. */
643 CPP_OPTION (pfile
, show_column
) = 0;
644 CPP_OPTION (pfile
, trigraphs
) = 0;
645 CPP_OPTION (pfile
, warn_trigraphs
) = 0;