1 /* Part of CPP library.
2 Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2007,
3 2008, 2009 Free Software Foundation, Inc.
5 This program is free software; you can redistribute it and/or modify it
6 under the terms of the GNU General Public License as published by the
7 Free Software Foundation; either version 3, or (at your option) any
10 This program is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 GNU General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with this program; see the file COPYING3. If not see
17 <http://www.gnu.org/licenses/>. */
19 /* This header defines all the internal data structures and functions
20 that need to be visible across files. It should not be used outside
23 #ifndef LIBCPP_INTERNAL_H
24 #define LIBCPP_INTERNAL_H
27 #include "cpp-id-data.h"
33 typedef int iconv_t
; /* dummy */
40 struct directive
; /* Deliberately incomplete. */
41 struct pending_option
;
45 typedef bool (*convert_f
) (iconv_t
, const unsigned char *, size_t,
46 struct _cpp_strbuf
*);
54 #define BITS_PER_CPPCHAR_T (CHAR_BIT * sizeof (cppchar_t))
56 /* Test if a sign is valid within a preprocessing number. */
57 #define VALID_SIGN(c, prevc) \
58 (((c) == '+' || (c) == '-') && \
59 ((prevc) == 'e' || (prevc) == 'E' \
60 || (((prevc) == 'p' || (prevc) == 'P') \
61 && CPP_OPTION (pfile, extended_numbers))))
63 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
64 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
65 #define CPP_BUF_COLUMN(BUF, CUR) ((CUR) - (BUF)->line_base)
66 #define CPP_BUF_COL(BUF) CPP_BUF_COLUMN(BUF, (BUF)->cur)
68 #define CPP_INCREMENT_LINE(PFILE, COLS_HINT) do { \
69 const struct line_maps *line_table = PFILE->line_table; \
70 const struct line_map *map = &line_table->maps[line_table->used-1]; \
71 linenum_type line = SOURCE_LINE (map, line_table->highest_line); \
72 linemap_line_start (PFILE->line_table, line + 1, COLS_HINT); \
75 /* Maximum nesting of cpp_buffers. We use a static limit, partly for
76 efficiency, and partly to limit runaway recursion. */
77 #define CPP_STACK_MAX 200
79 /* Host alignment handling. */
90 #define DEFAULT_ALIGNMENT offsetof (struct dummy, u)
91 #define CPP_ALIGN2(size, align) (((size) + ((align) - 1)) & ~((align) - 1))
92 #define CPP_ALIGN(size) CPP_ALIGN2 (size, DEFAULT_ALIGNMENT)
94 #define _cpp_mark_macro_used(NODE) do { \
95 if ((NODE)->type == NT_MACRO && !((NODE)->flags & NODE_BUILTIN)) \
96 (NODE)->value.macro->used = 1; } while (0)
98 /* A generic memory buffer, and operations on it. */
99 typedef struct _cpp_buff _cpp_buff
;
102 struct _cpp_buff
*next
;
103 unsigned char *base
, *cur
, *limit
;
106 extern _cpp_buff
*_cpp_get_buff (cpp_reader
*, size_t);
107 extern void _cpp_release_buff (cpp_reader
*, _cpp_buff
*);
108 extern void _cpp_extend_buff (cpp_reader
*, _cpp_buff
**, size_t);
109 extern _cpp_buff
*_cpp_append_extend_buff (cpp_reader
*, _cpp_buff
*, size_t);
110 extern void _cpp_free_buff (_cpp_buff
*);
111 extern unsigned char *_cpp_aligned_alloc (cpp_reader
*, size_t);
112 extern unsigned char *_cpp_unaligned_alloc (cpp_reader
*, size_t);
114 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
115 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
116 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
118 /* #include types. */
119 enum include_type
{IT_INCLUDE
, IT_INCLUDE_NEXT
, IT_IMPORT
, IT_CMDLINE
};
123 const cpp_token
*token
;
124 const cpp_token
**ptoken
;
127 /* A "run" of tokens; part of a chain of runs. */
128 typedef struct tokenrun tokenrun
;
131 tokenrun
*next
, *prev
;
132 cpp_token
*base
, *limit
;
135 /* Accessor macros for struct cpp_context. */
136 #define FIRST(c) ((c)->u.iso.first)
137 #define LAST(c) ((c)->u.iso.last)
138 #define CUR(c) ((c)->u.trad.cur)
139 #define RLIMIT(c) ((c)->u.trad.rlimit)
141 typedef struct cpp_context cpp_context
;
144 /* Doubly-linked list. */
145 cpp_context
*next
, *prev
;
149 /* For ISO macro expansion. Contexts other than the base context
150 are contiguous tokens. e.g. macro expansions, expanded
158 /* For traditional macro expansion. */
161 const unsigned char *cur
;
162 const unsigned char *rlimit
;
166 /* If non-NULL, a buffer used for storage related to this context.
167 When the context is popped, the buffer is released. */
170 /* For a macro context, the macro node, otherwise NULL. */
173 /* True if utoken element is token, else ptoken. */
179 /* Nonzero if first token on line is CPP_HASH. */
180 unsigned char in_directive
;
182 /* Nonzero if in a directive that will handle padding tokens itself.
183 #include needs this to avoid problems with computed include and
184 spacing between tokens. */
185 unsigned char directive_wants_padding
;
187 /* True if we are skipping a failed conditional group. */
188 unsigned char skipping
;
190 /* Nonzero if in a directive that takes angle-bracketed headers. */
191 unsigned char angled_headers
;
193 /* Nonzero if in a #if or #elif directive. */
194 unsigned char in_expression
;
196 /* Nonzero to save comments. Turned off if discard_comments, and in
197 all directives apart from #define. */
198 unsigned char save_comments
;
200 /* Nonzero if lexing __VA_ARGS__ is valid. */
201 unsigned char va_args_ok
;
203 /* Nonzero if lexing poisoned identifiers is valid. */
204 unsigned char poisoned_ok
;
206 /* Nonzero to prevent macro expansion. */
207 unsigned char prevent_expansion
;
209 /* Nonzero when parsing arguments to a function-like macro. */
210 unsigned char parsing_args
;
212 /* Nonzero if prevent_expansion is true only because output is
214 unsigned char discarding_output
;
216 /* Nonzero to skip evaluating part of an expression. */
217 unsigned int skip_eval
;
219 /* Nonzero when handling a deferred pragma. */
220 unsigned char in_deferred_pragma
;
222 /* Nonzero if the deferred pragma being handled allows macro expansion. */
223 unsigned char pragma_allow_expansion
;
226 /* Special nodes - identifiers with predefined significance. */
229 cpp_hashnode
*n_defined
; /* defined operator */
230 cpp_hashnode
*n_true
; /* C++ keyword true */
231 cpp_hashnode
*n_false
; /* C++ keyword false */
232 cpp_hashnode
*n__VA_ARGS__
; /* C99 vararg macros */
235 typedef struct _cpp_line_note _cpp_line_note
;
236 struct _cpp_line_note
238 /* Location in the clean line the note refers to. */
239 const unsigned char *pos
;
241 /* Type of note. The 9 'from' trigraph characters represent those
242 trigraphs, '\\' an escaped newline, ' ' an escaped newline with
243 intervening space, and anything else is invalid. */
247 /* Represents the contents of a file cpplib has read in. */
250 const unsigned char *cur
; /* Current location. */
251 const unsigned char *line_base
; /* Start of current physical line. */
252 const unsigned char *next_line
; /* Start of to-be-cleaned logical line. */
254 const unsigned char *buf
; /* Entire character buffer. */
255 const unsigned char *rlimit
; /* Writable byte at end of file. */
257 _cpp_line_note
*notes
; /* Array of notes. */
258 unsigned int cur_note
; /* Next note to process. */
259 unsigned int notes_used
; /* Number of notes. */
260 unsigned int notes_cap
; /* Size of allocated array. */
262 struct cpp_buffer
*prev
;
264 /* Pointer into the file table; non-NULL if this is a file buffer.
265 Used for include_next and to record control macros. */
266 struct _cpp_file
*file
;
268 /* Saved value of __TIMESTAMP__ macro - date and time of last modification
269 of the assotiated file. */
270 const unsigned char *timestamp
;
272 /* Value of if_stack at start of this file.
273 Used to prohibit unmatched #endif (etc) in an include file. */
274 struct if_stack
*if_stack
;
276 /* True if we need to get the next clean line. */
279 /* True if we have already warned about C++ comments in this file.
280 The warning happens only for C89 extended mode with -pedantic on,
281 or for -Wtraditional, and only once per file (otherwise it would
282 be far too noisy). */
283 unsigned int warned_cplusplus_comments
: 1;
285 /* True if we don't process trigraphs and escaped newlines. True
286 for preprocessed input, command line directives, and _Pragma
288 unsigned int from_stage3
: 1;
290 /* At EOF, a buffer is automatically popped. If RETURN_AT_EOF is
291 true, a CPP_EOF token is then returned. Otherwise, the next
292 token from the enclosing buffer is returned. */
293 unsigned int return_at_eof
: 1;
295 /* One for a system header, two for a C system header file that therefore
296 needs to be extern "C" protected in C++, and zero otherwise. */
299 /* The directory of the this buffer's file. Its NAME member is not
300 allocated, so we don't need to worry about freeing it. */
303 /* Descriptor for converting from the input character set to the
304 source character set. */
305 struct cset_converter input_cset_desc
;
308 /* A cpp_reader encapsulates the "state" of a pre-processor run.
309 Applying cpp_get_token repeatedly yields a stream of pre-processor
310 tokens. Usually, there is only one cpp_reader object active. */
313 /* Top of buffer stack. */
316 /* Overlaid buffer (can be different after processing #include). */
317 cpp_buffer
*overlaid_buffer
;
320 struct lexer_state state
;
322 /* Source line tracking. */
323 struct line_maps
*line_table
;
325 /* The line of the '#' of the current directive. */
326 source_location directive_line
;
328 /* Memory buffers. */
329 _cpp_buff
*a_buff
; /* Aligned permanent storage. */
330 _cpp_buff
*u_buff
; /* Unaligned permanent storage. */
331 _cpp_buff
*free_buffs
; /* Free buffer chain. */
334 struct cpp_context base_context
;
335 struct cpp_context
*context
;
337 /* If in_directive, the directive if known. */
338 const struct directive
*directive
;
340 /* Token generated while handling a directive, if any. */
341 cpp_token directive_result
;
343 /* When expanding a macro at top-level, this is the location of the
345 source_location invocation_location
;
347 /* True if this call to cpp_get_token should consider setting
348 invocation_location. */
349 bool set_invocation_location
;
351 /* Search paths for include files. */
352 struct cpp_dir
*quote_include
; /* "" */
353 struct cpp_dir
*bracket_include
; /* <> */
354 struct cpp_dir no_search_path
; /* No path. */
356 /* Chain of all hashed _cpp_file instances. */
357 struct _cpp_file
*all_files
;
359 struct _cpp_file
*main_file
;
361 /* File and directory hash table. */
362 struct htab
*file_hash
;
363 struct htab
*dir_hash
;
364 struct file_hash_entry_pool
*file_hash_entries
;
366 /* Negative path lookup hash table. */
367 struct htab
*nonexistent_file_hash
;
368 struct obstack nonexistent_file_ob
;
370 /* Nonzero means don't look for #include "foo" the source-file
372 bool quote_ignores_source_dir
;
374 /* Nonzero if any file has contained #pragma once or #import has
378 /* Multiple include optimization. */
379 const cpp_hashnode
*mi_cmacro
;
380 const cpp_hashnode
*mi_ind_cmacro
;
384 cpp_token
*cur_token
;
385 tokenrun base_run
, *cur_run
;
386 unsigned int lookaheads
;
388 /* Nonzero prevents the lexer from re-using the token runs. */
389 unsigned int keep_tokens
;
391 /* Buffer to hold macro definition string. */
392 unsigned char *macro_buffer
;
393 unsigned int macro_buffer_len
;
395 /* Descriptor for converting from the source character set to the
396 execution character set. */
397 struct cset_converter narrow_cset_desc
;
399 /* Descriptor for converting from the source character set to the
400 UTF-16 execution character set. */
401 struct cset_converter char16_cset_desc
;
403 /* Descriptor for converting from the source character set to the
404 UTF-32 execution character set. */
405 struct cset_converter char32_cset_desc
;
407 /* Descriptor for converting from the source character set to the
408 wide execution character set. */
409 struct cset_converter wide_cset_desc
;
411 /* Date and time text. Calculated together if either is requested. */
412 const unsigned char *date
;
413 const unsigned char *time
;
415 /* EOF token, and a token forcing paste avoidance. */
416 cpp_token avoid_paste
;
419 /* Opaque handle to the dependencies of mkdeps.c. */
422 /* Obstack holding all macro hash nodes. This never shrinks.
424 struct obstack hash_ob
;
426 /* Obstack holding buffer and conditional structures. This is a
427 real stack. See directives.c. */
428 struct obstack buffer_ob
;
430 /* Pragma table - dynamic, because a library user can add to the
431 list of recognized pragmas. */
432 struct pragma_entry
*pragmas
;
434 /* Call backs to cpplib client. */
435 struct cpp_callbacks cb
;
437 /* Identifier hash table. */
438 struct ht
*hash_table
;
440 /* Expression parser stack. */
441 struct op
*op_stack
, *op_limit
;
443 /* User visible options. */
444 struct cpp_options opts
;
446 /* Special nodes - identifiers with predefined significance to the
448 struct spec_nodes spec_nodes
;
450 /* Whether cpplib owns the hashtable. */
453 /* Traditional preprocessing output buffer (a logical line). */
457 unsigned char *limit
;
459 source_location first_line
;
462 /* Used for buffer overlays by traditional.c. */
463 const unsigned char *saved_cur
, *saved_rlimit
, *saved_line_base
;
465 /* A saved list of the defined macros, for dependency checking
466 of precompiled headers. */
467 struct cpp_savedstate
*savedstate
;
469 /* Next value of __COUNTER__ macro. */
470 unsigned int counter
;
472 /* Table of comments, when state.save_comments is true. */
473 cpp_comment_table comments
;
476 /* Character classes. Based on the more primitive macros in safe-ctype.h.
477 If the definition of `numchar' looks odd to you, please look up the
478 definition of a pp-number in the C standard [section 6.4.8 of C99].
480 In the unlikely event that characters other than \r and \n enter
481 the set is_vspace, the macro handle_newline() in lex.c must be
483 #define _dollar_ok(x) ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
485 #define is_idchar(x) (ISIDNUM(x) || _dollar_ok(x))
486 #define is_numchar(x) ISIDNUM(x)
487 #define is_idstart(x) (ISIDST(x) || _dollar_ok(x))
488 #define is_numstart(x) ISDIGIT(x)
489 #define is_hspace(x) ISBLANK(x)
490 #define is_vspace(x) IS_VSPACE(x)
491 #define is_nvspace(x) IS_NVSPACE(x)
492 #define is_space(x) IS_SPACE_OR_NUL(x)
494 /* This table is constant if it can be initialized at compile time,
495 which is the case if cpp was compiled with GCC >=2.7, or another
496 compiler that supports C99. */
497 #if HAVE_DESIGNATED_INITIALIZERS
498 extern const unsigned char _cpp_trigraph_map
[UCHAR_MAX
+ 1];
500 extern unsigned char _cpp_trigraph_map
[UCHAR_MAX
+ 1];
505 static inline int cpp_in_system_header (cpp_reader
*);
507 cpp_in_system_header (cpp_reader
*pfile
)
509 return pfile
->buffer
? pfile
->buffer
->sysp
: 0;
511 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
512 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
514 static inline int cpp_in_primary_file (cpp_reader
*);
516 cpp_in_primary_file (cpp_reader
*pfile
)
518 return pfile
->line_table
->depth
== 1;
522 extern void _cpp_free_definition (cpp_hashnode
*);
523 extern bool _cpp_create_definition (cpp_reader
*, cpp_hashnode
*);
524 extern void _cpp_pop_context (cpp_reader
*);
525 extern void _cpp_push_text_context (cpp_reader
*, cpp_hashnode
*,
526 const unsigned char *, size_t);
527 extern bool _cpp_save_parameter (cpp_reader
*, cpp_macro
*, cpp_hashnode
*);
528 extern bool _cpp_arguments_ok (cpp_reader
*, cpp_macro
*, const cpp_hashnode
*,
530 extern const unsigned char *_cpp_builtin_macro_text (cpp_reader
*,
532 extern int _cpp_warn_if_unused_macro (cpp_reader
*, cpp_hashnode
*, void *);
533 extern void _cpp_push_token_context (cpp_reader
*, cpp_hashnode
*,
534 const cpp_token
*, unsigned int);
535 extern void _cpp_backup_tokens_direct (cpp_reader
*, unsigned int);
537 /* In identifiers.c */
538 extern void _cpp_init_hashtable (cpp_reader
*, hash_table
*);
539 extern void _cpp_destroy_hashtable (cpp_reader
*);
542 typedef struct _cpp_file _cpp_file
;
543 extern _cpp_file
*_cpp_find_file (cpp_reader
*, const char *, cpp_dir
*,
545 extern bool _cpp_find_failed (_cpp_file
*);
546 extern void _cpp_mark_file_once_only (cpp_reader
*, struct _cpp_file
*);
547 extern void _cpp_fake_include (cpp_reader
*, const char *);
548 extern bool _cpp_stack_file (cpp_reader
*, _cpp_file
*, bool);
549 extern bool _cpp_stack_include (cpp_reader
*, const char *, int,
551 extern int _cpp_compare_file_date (cpp_reader
*, const char *, int);
552 extern void _cpp_report_missing_guards (cpp_reader
*);
553 extern void _cpp_init_files (cpp_reader
*);
554 extern void _cpp_cleanup_files (cpp_reader
*);
555 extern void _cpp_pop_file_buffer (cpp_reader
*, struct _cpp_file
*);
556 extern bool _cpp_save_file_entries (cpp_reader
*pfile
, FILE *f
);
557 extern bool _cpp_read_file_entries (cpp_reader
*, FILE *);
558 extern struct stat
*_cpp_get_file_stat (_cpp_file
*);
561 extern bool _cpp_parse_expr (cpp_reader
*, bool);
562 extern struct op
*_cpp_expand_op_stack (cpp_reader
*);
565 extern void _cpp_process_line_notes (cpp_reader
*, int);
566 extern void _cpp_clean_line (cpp_reader
*);
567 extern bool _cpp_get_fresh_line (cpp_reader
*);
568 extern bool _cpp_skip_block_comment (cpp_reader
*);
569 extern cpp_token
*_cpp_temp_token (cpp_reader
*);
570 extern const cpp_token
*_cpp_lex_token (cpp_reader
*);
571 extern cpp_token
*_cpp_lex_direct (cpp_reader
*);
572 extern int _cpp_equiv_tokens (const cpp_token
*, const cpp_token
*);
573 extern void _cpp_init_tokenrun (tokenrun
*, unsigned int);
576 extern void _cpp_maybe_push_include_file (cpp_reader
*);
577 extern const char *cpp_named_operator2name (enum cpp_ttype type
);
579 /* In directives.c */
580 extern int _cpp_test_assertion (cpp_reader
*, unsigned int *);
581 extern int _cpp_handle_directive (cpp_reader
*, int);
582 extern void _cpp_define_builtin (cpp_reader
*, const char *);
583 extern char ** _cpp_save_pragma_names (cpp_reader
*);
584 extern void _cpp_restore_pragma_names (cpp_reader
*, char **);
585 extern int _cpp_do__Pragma (cpp_reader
*);
586 extern void _cpp_init_directives (cpp_reader
*);
587 extern void _cpp_init_internal_pragmas (cpp_reader
*);
588 extern void _cpp_do_file_change (cpp_reader
*, enum lc_reason
, const char *,
589 linenum_type
, unsigned int);
590 extern void _cpp_pop_buffer (cpp_reader
*);
592 /* In directives.c */
593 struct _cpp_dir_only_callbacks
595 /* Called to print a block of lines. */
596 void (*print_lines
) (int, const void *, size_t);
597 void (*maybe_print_line
) (source_location
);
600 extern void _cpp_preprocess_dir_only (cpp_reader
*,
601 const struct _cpp_dir_only_callbacks
*);
603 /* In traditional.c. */
604 extern bool _cpp_scan_out_logical_line (cpp_reader
*, cpp_macro
*);
605 extern bool _cpp_read_logical_line_trad (cpp_reader
*);
606 extern void _cpp_overlay_buffer (cpp_reader
*pfile
, const unsigned char *,
608 extern void _cpp_remove_overlay (cpp_reader
*);
609 extern bool _cpp_create_trad_definition (cpp_reader
*, cpp_macro
*);
610 extern bool _cpp_expansions_different_trad (const cpp_macro
*,
612 extern unsigned char *_cpp_copy_replacement_text (const cpp_macro
*,
614 extern size_t _cpp_replacement_text_len (const cpp_macro
*);
618 /* The normalization state at this point in the sequence.
619 It starts initialized to all zeros, and at the end
620 'level' is the normalization level of the sequence. */
622 struct normalize_state
624 /* The previous character. */
626 /* The combining class of the previous character. */
627 unsigned char prev_class
;
628 /* The lowest normalization level so far. */
629 enum cpp_normalize_level level
;
631 #define INITIAL_NORMALIZE_STATE { 0, 0, normalized_KC }
632 #define NORMALIZE_STATE_RESULT(st) ((st)->level)
634 /* We saw a character that matches ISIDNUM(), update a
635 normalize_state appropriately. */
636 #define NORMALIZE_STATE_UPDATE_IDNUM(st) \
637 ((st)->previous = 0, (st)->prev_class = 0)
639 extern cppchar_t
_cpp_valid_ucn (cpp_reader
*, const unsigned char **,
640 const unsigned char *, int,
641 struct normalize_state
*state
);
642 extern void _cpp_destroy_iconv (cpp_reader
*);
643 extern unsigned char *_cpp_convert_input (cpp_reader
*, const char *,
644 unsigned char *, size_t, size_t,
645 const unsigned char **, off_t
*);
646 extern const char *_cpp_default_encoding (void);
647 extern cpp_hashnode
* _cpp_interpret_identifier (cpp_reader
*pfile
,
648 const unsigned char *id
,
651 /* Utility routines and macros. */
652 #define DSC(str) (const unsigned char *)str, sizeof str - 1
654 /* These are inline functions instead of macros so we can get type
656 static inline int ustrcmp (const unsigned char *, const unsigned char *);
657 static inline int ustrncmp (const unsigned char *, const unsigned char *,
659 static inline size_t ustrlen (const unsigned char *);
660 static inline unsigned char *uxstrdup (const unsigned char *);
661 static inline unsigned char *ustrchr (const unsigned char *, int);
662 static inline int ufputs (const unsigned char *, FILE *);
664 /* Use a const char for the second parameter since it is usually a literal. */
665 static inline int ustrcspn (const unsigned char *, const char *);
668 ustrcmp (const unsigned char *s1
, const unsigned char *s2
)
670 return strcmp ((const char *)s1
, (const char *)s2
);
674 ustrncmp (const unsigned char *s1
, const unsigned char *s2
, size_t n
)
676 return strncmp ((const char *)s1
, (const char *)s2
, n
);
680 ustrcspn (const unsigned char *s1
, const char *s2
)
682 return strcspn ((const char *)s1
, s2
);
686 ustrlen (const unsigned char *s1
)
688 return strlen ((const char *)s1
);
691 static inline unsigned char *
692 uxstrdup (const unsigned char *s1
)
694 return (unsigned char *) xstrdup ((const char *)s1
);
697 static inline unsigned char *
698 ustrchr (const unsigned char *s1
, int c
)
700 return (unsigned char *) strchr ((const char *)s1
, c
);
704 ufputs (const unsigned char *s
, FILE *f
)
706 return fputs ((const char *)s
, f
);
713 #endif /* ! LIBCPP_INTERNAL_H */