2005-04-26 Paul Brook <paul@codesourcery.com>
[official-gcc.git] / libcpp / internal.h
blobf0bd537d310cdbd0467ae7132d994f149170d7f2
1 /* Part of CPP library.
2 Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005
3 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 2, or (at your option) any
8 later version.
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; if not, write to the Free Software
17 Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
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
21 cpplib. */
23 #ifndef LIBCPP_INTERNAL_H
24 #define LIBCPP_INTERNAL_H
26 #include "symtab.h"
27 #include "cpp-id-data.h"
29 #ifndef HAVE_ICONV_H
30 #undef HAVE_ICONV
31 #endif
33 #if HAVE_ICONV
34 #include <iconv.h>
35 #else
36 #define HAVE_ICONV 0
37 typedef int iconv_t; /* dummy */
38 #endif
40 struct directive; /* Deliberately incomplete. */
41 struct pending_option;
42 struct op;
43 struct _cpp_strbuf;
45 typedef bool (*convert_f) (iconv_t, const unsigned char *, size_t,
46 struct _cpp_strbuf *);
47 struct cset_converter
49 convert_f func;
50 iconv_t cd;
53 #define BITS_PER_CPPCHAR_T (CHAR_BIT * sizeof (cppchar_t))
55 /* Test if a sign is valid within a preprocessing number. */
56 #define VALID_SIGN(c, prevc) \
57 (((c) == '+' || (c) == '-') && \
58 ((prevc) == 'e' || (prevc) == 'E' \
59 || (((prevc) == 'p' || (prevc) == 'P') \
60 && CPP_OPTION (pfile, extended_numbers))))
62 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
63 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
64 #define CPP_BUF_COLUMN(BUF, CUR) ((CUR) - (BUF)->line_base)
65 #define CPP_BUF_COL(BUF) CPP_BUF_COLUMN(BUF, (BUF)->cur)
67 #define CPP_INCREMENT_LINE(PFILE, COLS_HINT) do { \
68 const struct line_maps *line_table = PFILE->line_table; \
69 const struct line_map *map = &line_table->maps[line_table->used-1]; \
70 unsigned int line = SOURCE_LINE (map, line_table->highest_line); \
71 linemap_line_start (PFILE->line_table, line + 1, COLS_HINT); \
72 } while (0)
74 /* Maximum nesting of cpp_buffers. We use a static limit, partly for
75 efficiency, and partly to limit runaway recursion. */
76 #define CPP_STACK_MAX 200
78 /* Host alignment handling. */
79 struct dummy
81 char c;
82 union
84 double d;
85 int *p;
86 } u;
89 #define DEFAULT_ALIGNMENT offsetof (struct dummy, u)
90 #define CPP_ALIGN2(size, align) (((size) + ((align) - 1)) & ~((align) - 1))
91 #define CPP_ALIGN(size) CPP_ALIGN2 (size, DEFAULT_ALIGNMENT)
93 #define _cpp_mark_macro_used(NODE) do { \
94 if ((NODE)->type == NT_MACRO && !((NODE)->flags & NODE_BUILTIN)) \
95 (NODE)->value.macro->used = 1; } while (0)
97 /* A generic memory buffer, and operations on it. */
98 typedef struct _cpp_buff _cpp_buff;
99 struct _cpp_buff
101 struct _cpp_buff *next;
102 unsigned char *base, *cur, *limit;
105 extern _cpp_buff *_cpp_get_buff (cpp_reader *, size_t);
106 extern void _cpp_release_buff (cpp_reader *, _cpp_buff *);
107 extern void _cpp_extend_buff (cpp_reader *, _cpp_buff **, size_t);
108 extern _cpp_buff *_cpp_append_extend_buff (cpp_reader *, _cpp_buff *, size_t);
109 extern void _cpp_free_buff (_cpp_buff *);
110 extern unsigned char *_cpp_aligned_alloc (cpp_reader *, size_t);
111 extern unsigned char *_cpp_unaligned_alloc (cpp_reader *, size_t);
113 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
114 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
115 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
117 /* #include types. */
118 enum include_type {IT_INCLUDE, IT_INCLUDE_NEXT, IT_IMPORT, IT_CMDLINE};
120 union utoken
122 const cpp_token *token;
123 const cpp_token **ptoken;
126 /* A "run" of tokens; part of a chain of runs. */
127 typedef struct tokenrun tokenrun;
128 struct tokenrun
130 tokenrun *next, *prev;
131 cpp_token *base, *limit;
134 /* Accessor macros for struct cpp_context. */
135 #define FIRST(c) ((c)->u.iso.first)
136 #define LAST(c) ((c)->u.iso.last)
137 #define CUR(c) ((c)->u.trad.cur)
138 #define RLIMIT(c) ((c)->u.trad.rlimit)
140 typedef struct cpp_context cpp_context;
141 struct cpp_context
143 /* Doubly-linked list. */
144 cpp_context *next, *prev;
146 union
148 /* For ISO macro expansion. Contexts other than the base context
149 are contiguous tokens. e.g. macro expansions, expanded
150 argument tokens. */
151 struct
153 union utoken first;
154 union utoken last;
155 } iso;
157 /* For traditional macro expansion. */
158 struct
160 const unsigned char *cur;
161 const unsigned char *rlimit;
162 } trad;
163 } u;
165 /* If non-NULL, a buffer used for storage related to this context.
166 When the context is popped, the buffer is released. */
167 _cpp_buff *buff;
169 /* For a macro context, the macro node, otherwise NULL. */
170 cpp_hashnode *macro;
172 /* True if utoken element is token, else ptoken. */
173 bool direct_p;
176 struct lexer_state
178 /* Nonzero if first token on line is CPP_HASH. */
179 unsigned char in_directive;
181 /* Nonzero if in a directive that will handle padding tokens itself.
182 #include needs this to avoid problems with computed include and
183 spacing between tokens. */
184 unsigned char directive_wants_padding;
186 /* True if we are skipping a failed conditional group. */
187 unsigned char skipping;
189 /* Nonzero if in a directive that takes angle-bracketed headers. */
190 unsigned char angled_headers;
192 /* Nonzero if in a #if or #elif directive. */
193 unsigned char in_expression;
195 /* Nonzero to save comments. Turned off if discard_comments, and in
196 all directives apart from #define. */
197 unsigned char save_comments;
199 /* Nonzero if lexing __VA_ARGS__ is valid. */
200 unsigned char va_args_ok;
202 /* Nonzero if lexing poisoned identifiers is valid. */
203 unsigned char poisoned_ok;
205 /* Nonzero to prevent macro expansion. */
206 unsigned char prevent_expansion;
208 /* Nonzero when handling a deferred pragma. */
209 unsigned char in_deferred_pragma;
211 /* Nonzero when parsing arguments to a function-like macro. */
212 unsigned char parsing_args;
214 /* Nonzero if prevent_expansion is true only because output is
215 being discarded. */
216 unsigned char discarding_output;
218 /* Nonzero to skip evaluating part of an expression. */
219 unsigned int skip_eval;
222 /* Special nodes - identifiers with predefined significance. */
223 struct spec_nodes
225 cpp_hashnode *n_defined; /* defined operator */
226 cpp_hashnode *n_true; /* C++ keyword true */
227 cpp_hashnode *n_false; /* C++ keyword false */
228 cpp_hashnode *n__VA_ARGS__; /* C99 vararg macros */
231 typedef struct _cpp_line_note _cpp_line_note;
232 struct _cpp_line_note
234 /* Location in the clean line the note refers to. */
235 const unsigned char *pos;
237 /* Type of note. The 9 'from' trigraph characters represent those
238 trigraphs, '\\' an escaped newline, ' ' an escaped newline with
239 intervening space, and anything else is invalid. */
240 unsigned int type;
243 /* Represents the contents of a file cpplib has read in. */
244 struct cpp_buffer
246 const unsigned char *cur; /* Current location. */
247 const unsigned char *line_base; /* Start of current physical line. */
248 const unsigned char *next_line; /* Start of to-be-cleaned logical line. */
250 const unsigned char *buf; /* Entire character buffer. */
251 const unsigned char *rlimit; /* Writable byte at end of file. */
253 _cpp_line_note *notes; /* Array of notes. */
254 unsigned int cur_note; /* Next note to process. */
255 unsigned int notes_used; /* Number of notes. */
256 unsigned int notes_cap; /* Size of allocated array. */
258 struct cpp_buffer *prev;
260 /* Pointer into the file table; non-NULL if this is a file buffer.
261 Used for include_next and to record control macros. */
262 struct _cpp_file *file;
264 /* Value of if_stack at start of this file.
265 Used to prohibit unmatched #endif (etc) in an include file. */
266 struct if_stack *if_stack;
268 /* True if we need to get the next clean line. */
269 bool need_line;
271 /* True if we have already warned about C++ comments in this file.
272 The warning happens only for C89 extended mode with -pedantic on,
273 or for -Wtraditional, and only once per file (otherwise it would
274 be far too noisy). */
275 unsigned int warned_cplusplus_comments : 1;
277 /* True if we don't process trigraphs and escaped newlines. True
278 for preprocessed input, command line directives, and _Pragma
279 buffers. */
280 unsigned int from_stage3 : 1;
282 /* At EOF, a buffer is automatically popped. If RETURN_AT_EOF is
283 true, a CPP_EOF token is then returned. Otherwise, the next
284 token from the enclosing buffer is returned. */
285 unsigned int return_at_eof : 1;
287 /* One for a system header, two for a C system header file that therefore
288 needs to be extern "C" protected in C++, and zero otherwise. */
289 unsigned char sysp;
291 /* The directory of the this buffer's file. Its NAME member is not
292 allocated, so we don't need to worry about freeing it. */
293 struct cpp_dir dir;
295 /* Descriptor for converting from the input character set to the
296 source character set. */
297 struct cset_converter input_cset_desc;
300 /* A cpp_reader encapsulates the "state" of a pre-processor run.
301 Applying cpp_get_token repeatedly yields a stream of pre-processor
302 tokens. Usually, there is only one cpp_reader object active. */
303 struct cpp_reader
305 /* Top of buffer stack. */
306 cpp_buffer *buffer;
308 /* Overlaid buffer (can be different after processing #include). */
309 cpp_buffer *overlaid_buffer;
311 /* Lexer state. */
312 struct lexer_state state;
314 /* Source line tracking. */
315 struct line_maps *line_table;
317 /* The line of the '#' of the current directive. */
318 source_location directive_line;
320 /* Memory buffers. */
321 _cpp_buff *a_buff; /* Aligned permanent storage. */
322 _cpp_buff *u_buff; /* Unaligned permanent storage. */
323 _cpp_buff *free_buffs; /* Free buffer chain. */
325 /* Context stack. */
326 struct cpp_context base_context;
327 struct cpp_context *context;
329 /* If in_directive, the directive if known. */
330 const struct directive *directive;
332 /* Token generated while handling a directive, if any. */
333 cpp_token directive_result;
335 /* Search paths for include files. */
336 struct cpp_dir *quote_include; /* "" */
337 struct cpp_dir *bracket_include; /* <> */
338 struct cpp_dir no_search_path; /* No path. */
340 /* Chain of all hashed _cpp_file instances. */
341 struct _cpp_file *all_files;
343 struct _cpp_file *main_file;
345 /* File and directory hash table. */
346 struct htab *file_hash;
347 struct htab *dir_hash;
348 struct file_hash_entry *file_hash_entries;
349 unsigned int file_hash_entries_allocated, file_hash_entries_used;
351 /* Nonzero means don't look for #include "foo" the source-file
352 directory. */
353 bool quote_ignores_source_dir;
355 /* Nonzero if any file has contained #pragma once or #import has
356 been used. */
357 bool seen_once_only;
359 /* Multiple include optimization. */
360 const cpp_hashnode *mi_cmacro;
361 const cpp_hashnode *mi_ind_cmacro;
362 bool mi_valid;
364 /* Lexing. */
365 cpp_token *cur_token;
366 tokenrun base_run, *cur_run;
367 unsigned int lookaheads;
369 /* Nonzero prevents the lexer from re-using the token runs. */
370 unsigned int keep_tokens;
372 /* Error counter for exit code. */
373 unsigned int errors;
375 /* Buffer to hold macro definition string. */
376 unsigned char *macro_buffer;
377 unsigned int macro_buffer_len;
379 /* Descriptor for converting from the source character set to the
380 execution character set. */
381 struct cset_converter narrow_cset_desc;
383 /* Descriptor for converting from the source character set to the
384 wide execution character set. */
385 struct cset_converter wide_cset_desc;
387 /* Date and time text. Calculated together if either is requested. */
388 const unsigned char *date;
389 const unsigned char *time;
391 /* EOF token, and a token forcing paste avoidance. */
392 cpp_token avoid_paste;
393 cpp_token eof;
395 /* Opaque handle to the dependencies of mkdeps.c. */
396 struct deps *deps;
398 /* Obstack holding all macro hash nodes. This never shrinks.
399 See identifiers.c */
400 struct obstack hash_ob;
402 /* Obstack holding buffer and conditional structures. This is a
403 real stack. See directives.c. */
404 struct obstack buffer_ob;
406 /* Pragma table - dynamic, because a library user can add to the
407 list of recognized pragmas. */
408 struct pragma_entry *pragmas;
410 /* Call backs to cpplib client. */
411 struct cpp_callbacks cb;
413 /* Identifier hash table. */
414 struct ht *hash_table;
416 /* Expression parser stack. */
417 struct op *op_stack, *op_limit;
419 /* User visible options. */
420 struct cpp_options opts;
422 /* Special nodes - identifiers with predefined significance to the
423 preprocessor. */
424 struct spec_nodes spec_nodes;
426 /* Whether cpplib owns the hashtable. */
427 bool our_hashtable;
429 /* Traditional preprocessing output buffer (a logical line). */
430 struct
432 unsigned char *base;
433 unsigned char *limit;
434 unsigned char *cur;
435 source_location first_line;
436 } out;
438 /* Used for buffer overlays by traditional.c. */
439 const unsigned char *saved_cur, *saved_rlimit, *saved_line_base;
441 /* A saved list of the defined macros, for dependency checking
442 of precompiled headers. */
443 struct cpp_savedstate *savedstate;
446 /* Character classes. Based on the more primitive macros in safe-ctype.h.
447 If the definition of `numchar' looks odd to you, please look up the
448 definition of a pp-number in the C standard [section 6.4.8 of C99].
450 In the unlikely event that characters other than \r and \n enter
451 the set is_vspace, the macro handle_newline() in lex.c must be
452 updated. */
453 #define _dollar_ok(x) ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
455 #define is_idchar(x) (ISIDNUM(x) || _dollar_ok(x))
456 #define is_numchar(x) ISIDNUM(x)
457 #define is_idstart(x) (ISIDST(x) || _dollar_ok(x))
458 #define is_numstart(x) ISDIGIT(x)
459 #define is_hspace(x) ISBLANK(x)
460 #define is_vspace(x) IS_VSPACE(x)
461 #define is_nvspace(x) IS_NVSPACE(x)
462 #define is_space(x) IS_SPACE_OR_NUL(x)
464 /* This table is constant if it can be initialized at compile time,
465 which is the case if cpp was compiled with GCC >=2.7, or another
466 compiler that supports C99. */
467 #if HAVE_DESIGNATED_INITIALIZERS
468 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
469 #else
470 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
471 #endif
473 /* Macros. */
475 static inline int cpp_in_system_header (cpp_reader *);
476 static inline int
477 cpp_in_system_header (cpp_reader *pfile)
479 return pfile->buffer ? pfile->buffer->sysp : 0;
481 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
482 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
484 /* In errors.c */
485 extern int _cpp_begin_message (cpp_reader *, int,
486 source_location, unsigned int);
488 /* In macro.c */
489 extern void _cpp_free_definition (cpp_hashnode *);
490 extern bool _cpp_create_definition (cpp_reader *, cpp_hashnode *);
491 extern void _cpp_pop_context (cpp_reader *);
492 extern void _cpp_push_text_context (cpp_reader *, cpp_hashnode *,
493 const unsigned char *, size_t);
494 extern bool _cpp_save_parameter (cpp_reader *, cpp_macro *, cpp_hashnode *);
495 extern bool _cpp_arguments_ok (cpp_reader *, cpp_macro *, const cpp_hashnode *,
496 unsigned int);
497 extern const unsigned char *_cpp_builtin_macro_text (cpp_reader *,
498 cpp_hashnode *);
499 int _cpp_warn_if_unused_macro (cpp_reader *, cpp_hashnode *, void *);
500 /* In identifiers.c */
501 extern void _cpp_init_hashtable (cpp_reader *, hash_table *);
502 extern void _cpp_destroy_hashtable (cpp_reader *);
504 /* In files.c */
505 typedef struct _cpp_file _cpp_file;
506 extern _cpp_file *_cpp_find_file (cpp_reader *, const char *fname,
507 cpp_dir *start_dir, bool fake);
508 extern bool _cpp_find_failed (_cpp_file *);
509 extern void _cpp_mark_file_once_only (cpp_reader *, struct _cpp_file *);
510 extern void _cpp_fake_include (cpp_reader *, const char *);
511 extern bool _cpp_stack_file (cpp_reader *, _cpp_file*, bool);
512 extern bool _cpp_stack_include (cpp_reader *, const char *, int,
513 enum include_type);
514 extern int _cpp_compare_file_date (cpp_reader *, const char *, int);
515 extern void _cpp_report_missing_guards (cpp_reader *);
516 extern void _cpp_init_files (cpp_reader *);
517 extern void _cpp_cleanup_files (cpp_reader *);
518 extern void _cpp_pop_file_buffer (cpp_reader *, struct _cpp_file *);
519 extern bool _cpp_save_file_entries (cpp_reader *pfile, FILE *f);
520 extern bool _cpp_read_file_entries (cpp_reader *, FILE *);
522 /* In expr.c */
523 extern bool _cpp_parse_expr (cpp_reader *);
524 extern struct op *_cpp_expand_op_stack (cpp_reader *);
526 /* In lex.c */
527 extern void _cpp_process_line_notes (cpp_reader *, int);
528 extern void _cpp_clean_line (cpp_reader *);
529 extern bool _cpp_get_fresh_line (cpp_reader *);
530 extern bool _cpp_skip_block_comment (cpp_reader *);
531 extern cpp_token *_cpp_temp_token (cpp_reader *);
532 extern const cpp_token *_cpp_lex_token (cpp_reader *);
533 extern cpp_token *_cpp_lex_direct (cpp_reader *);
534 extern int _cpp_equiv_tokens (const cpp_token *, const cpp_token *);
535 extern void _cpp_init_tokenrun (tokenrun *, unsigned int);
537 /* In init.c. */
538 extern void _cpp_maybe_push_include_file (cpp_reader *);
540 /* In directives.c */
541 extern int _cpp_test_assertion (cpp_reader *, unsigned int *);
542 extern int _cpp_handle_directive (cpp_reader *, int);
543 extern void _cpp_define_builtin (cpp_reader *, const char *);
544 extern char ** _cpp_save_pragma_names (cpp_reader *);
545 extern void _cpp_restore_pragma_names (cpp_reader *, char **);
546 extern void _cpp_do__Pragma (cpp_reader *);
547 extern void _cpp_init_directives (cpp_reader *);
548 extern void _cpp_init_internal_pragmas (cpp_reader *);
549 extern void _cpp_do_file_change (cpp_reader *, enum lc_reason, const char *,
550 unsigned int, unsigned int);
551 extern void _cpp_pop_buffer (cpp_reader *);
553 /* In traditional.c. */
554 extern bool _cpp_scan_out_logical_line (cpp_reader *, cpp_macro *);
555 extern bool _cpp_read_logical_line_trad (cpp_reader *);
556 extern void _cpp_overlay_buffer (cpp_reader *pfile, const unsigned char *,
557 size_t);
558 extern void _cpp_remove_overlay (cpp_reader *);
559 extern bool _cpp_create_trad_definition (cpp_reader *, cpp_macro *);
560 extern bool _cpp_expansions_different_trad (const cpp_macro *,
561 const cpp_macro *);
562 extern unsigned char *_cpp_copy_replacement_text (const cpp_macro *,
563 unsigned char *);
564 extern size_t _cpp_replacement_text_len (const cpp_macro *);
566 /* In charset.c. */
568 /* The normalization state at this point in the sequence.
569 It starts initialized to all zeros, and at the end
570 'level' is the normalization level of the sequence. */
572 struct normalize_state
574 /* The previous character. */
575 cppchar_t previous;
576 /* The combining class of the previous character. */
577 unsigned char prev_class;
578 /* The lowest normalization level so far. */
579 enum cpp_normalize_level level;
581 #define INITIAL_NORMALIZE_STATE { 0, 0, normalized_KC }
582 #define NORMALIZE_STATE_RESULT(st) ((st)->level)
584 /* We saw a character that matches ISIDNUM(), update a
585 normalize_state appropriately. */
586 #define NORMALIZE_STATE_UPDATE_IDNUM(st) \
587 ((st)->previous = 0, (st)->prev_class = 0)
589 extern cppchar_t _cpp_valid_ucn (cpp_reader *, const unsigned char **,
590 const unsigned char *, int,
591 struct normalize_state *state);
592 extern void _cpp_destroy_iconv (cpp_reader *);
593 extern unsigned char *_cpp_convert_input (cpp_reader *, const char *,
594 unsigned char *, size_t, size_t,
595 off_t *);
596 extern const char *_cpp_default_encoding (void);
597 extern cpp_hashnode * _cpp_interpret_identifier (cpp_reader *pfile,
598 const unsigned char *id,
599 size_t len);
601 /* Utility routines and macros. */
602 #define DSC(str) (const unsigned char *)str, sizeof str - 1
604 /* These are inline functions instead of macros so we can get type
605 checking. */
606 static inline int ustrcmp (const unsigned char *, const unsigned char *);
607 static inline int ustrncmp (const unsigned char *, const unsigned char *,
608 size_t);
609 static inline size_t ustrlen (const unsigned char *);
610 static inline unsigned char *uxstrdup (const unsigned char *);
611 static inline unsigned char *ustrchr (const unsigned char *, int);
612 static inline int ufputs (const unsigned char *, FILE *);
614 /* Use a const char for the second parameter since it is usually a literal. */
615 static inline int ustrcspn (const unsigned char *, const char *);
617 static inline int
618 ustrcmp (const unsigned char *s1, const unsigned char *s2)
620 return strcmp ((const char *)s1, (const char *)s2);
623 static inline int
624 ustrncmp (const unsigned char *s1, const unsigned char *s2, size_t n)
626 return strncmp ((const char *)s1, (const char *)s2, n);
629 static inline int
630 ustrcspn (const unsigned char *s1, const char *s2)
632 return strcspn ((const char *)s1, s2);
635 static inline size_t
636 ustrlen (const unsigned char *s1)
638 return strlen ((const char *)s1);
641 static inline unsigned char *
642 uxstrdup (const unsigned char *s1)
644 return (unsigned char *) xstrdup ((const char *)s1);
647 static inline unsigned char *
648 ustrchr (const unsigned char *s1, int c)
650 return (unsigned char *) strchr ((const char *)s1, c);
653 static inline int
654 ufputs (const unsigned char *s, FILE *f)
656 return fputs ((const char *)s, f);
659 #endif /* ! LIBCPP_INTERNAL_H */