2004-02-24 Aldy Hernandez <aldyh@redhat.com>
[official-gcc.git] / gcc / cpphash.h
blob4e86bf37d5b320ae3e9d14f5c882d8fd63bcb9e6
1 /* Part of CPP library.
2 Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004
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's called cpphash.h for
21 historical reasons. */
23 #ifndef GCC_CPPHASH_H
24 #define GCC_CPPHASH_H
26 #include "hashtable.h"
28 #if defined HAVE_ICONV_H && defined HAVE_ICONV
29 #include <iconv.h>
30 #else
31 #define HAVE_ICONV 0
32 typedef int iconv_t; /* dummy */
33 #endif
35 struct directive; /* Deliberately incomplete. */
36 struct pending_option;
37 struct op;
38 struct _cpp_strbuf;
40 typedef bool (*convert_f) (iconv_t, const unsigned char *, size_t,
41 struct _cpp_strbuf *);
42 struct cset_converter
44 convert_f func;
45 iconv_t cd;
48 #ifndef HAVE_UCHAR
49 typedef unsigned char uchar;
50 #endif
51 #define U (const uchar *) /* Intended use: U"string" */
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_map *map \
69 = linemap_lookup (PFILE->line_table, PFILE->line); \
70 unsigned int line = SOURCE_LINE (map, PFILE->line) + 1; \
71 PFILE->line = linemap_line_start (PFILE->line_table, line, 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 /* Each macro definition is recorded in a cpp_macro structure.
94 Variadic macros cannot occur with traditional cpp. */
95 struct cpp_macro
97 /* Parameters, if any. */
98 cpp_hashnode **params;
100 /* Replacement tokens (ISO) or replacement text (traditional). See
101 comment at top of cpptrad.c for how traditional function-like
102 macros are encoded. */
103 union
105 cpp_token *tokens;
106 const uchar *text;
107 } exp;
109 /* Definition line number. */
110 fileline line;
112 /* Number of tokens in expansion, or bytes for traditional macros. */
113 unsigned int count;
115 /* Number of parameters. */
116 unsigned short paramc;
118 /* If a function-like macro. */
119 unsigned int fun_like : 1;
121 /* If a variadic macro. */
122 unsigned int variadic : 1;
124 /* If macro defined in system header. */
125 unsigned int syshdr : 1;
127 /* Nonzero if it has been expanded or had its existence tested. */
128 unsigned int used : 1;
131 #define _cpp_mark_macro_used(NODE) do { \
132 if ((NODE)->type == NT_MACRO && !((NODE)->flags & NODE_BUILTIN)) \
133 (NODE)->value.macro->used = 1; } while (0)
135 /* A generic memory buffer, and operations on it. */
136 typedef struct _cpp_buff _cpp_buff;
137 struct _cpp_buff
139 struct _cpp_buff *next;
140 unsigned char *base, *cur, *limit;
143 extern _cpp_buff *_cpp_get_buff (cpp_reader *, size_t);
144 extern void _cpp_release_buff (cpp_reader *, _cpp_buff *);
145 extern void _cpp_extend_buff (cpp_reader *, _cpp_buff **, size_t);
146 extern _cpp_buff *_cpp_append_extend_buff (cpp_reader *, _cpp_buff *, size_t);
147 extern void _cpp_free_buff (_cpp_buff *);
148 extern unsigned char *_cpp_aligned_alloc (cpp_reader *, size_t);
149 extern unsigned char *_cpp_unaligned_alloc (cpp_reader *, size_t);
151 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
152 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
153 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
155 /* #include types. */
156 enum include_type {IT_INCLUDE, IT_INCLUDE_NEXT, IT_IMPORT, IT_CMDLINE};
158 union utoken
160 const cpp_token *token;
161 const cpp_token **ptoken;
164 /* A "run" of tokens; part of a chain of runs. */
165 typedef struct tokenrun tokenrun;
166 struct tokenrun
168 tokenrun *next, *prev;
169 cpp_token *base, *limit;
172 /* Accessor macros for struct cpp_context. */
173 #define FIRST(c) ((c)->u.iso.first)
174 #define LAST(c) ((c)->u.iso.last)
175 #define CUR(c) ((c)->u.trad.cur)
176 #define RLIMIT(c) ((c)->u.trad.rlimit)
178 typedef struct cpp_context cpp_context;
179 struct cpp_context
181 /* Doubly-linked list. */
182 cpp_context *next, *prev;
184 union
186 /* For ISO macro expansion. Contexts other than the base context
187 are contiguous tokens. e.g. macro expansions, expanded
188 argument tokens. */
189 struct
191 union utoken first;
192 union utoken last;
193 } iso;
195 /* For traditional macro expansion. */
196 struct
198 const uchar *cur;
199 const uchar *rlimit;
200 } trad;
201 } u;
203 /* If non-NULL, a buffer used for storage related to this context.
204 When the context is popped, the buffer is released. */
205 _cpp_buff *buff;
207 /* For a macro context, the macro node, otherwise NULL. */
208 cpp_hashnode *macro;
210 /* True if utoken element is token, else ptoken. */
211 bool direct_p;
214 struct lexer_state
216 /* Nonzero if first token on line is CPP_HASH. */
217 unsigned char in_directive;
219 /* Nonzero if in a directive that will handle padding tokens itself.
220 #include needs this to avoid problems with computed include and
221 spacing between tokens. */
222 unsigned char directive_wants_padding;
224 /* True if we are skipping a failed conditional group. */
225 unsigned char skipping;
227 /* Nonzero if in a directive that takes angle-bracketed headers. */
228 unsigned char angled_headers;
230 /* Nonzero if in a #if or #elif directive. */
231 unsigned char in_expression;
233 /* Nonzero to save comments. Turned off if discard_comments, and in
234 all directives apart from #define. */
235 unsigned char save_comments;
237 /* Nonzero if lexing __VA_ARGS__ is valid. */
238 unsigned char va_args_ok;
240 /* Nonzero if lexing poisoned identifiers is valid. */
241 unsigned char poisoned_ok;
243 /* Nonzero to prevent macro expansion. */
244 unsigned char prevent_expansion;
246 /* Nonzero when parsing arguments to a function-like macro. */
247 unsigned char parsing_args;
249 /* Nonzero to skip evaluating part of an expression. */
250 unsigned int skip_eval;
253 /* Special nodes - identifiers with predefined significance. */
254 struct spec_nodes
256 cpp_hashnode *n_defined; /* defined operator */
257 cpp_hashnode *n_true; /* C++ keyword true */
258 cpp_hashnode *n_false; /* C++ keyword false */
259 cpp_hashnode *n__VA_ARGS__; /* C99 vararg macros */
262 typedef struct _cpp_line_note _cpp_line_note;
263 struct _cpp_line_note
265 /* Location in the clean line the note refers to. */
266 const uchar *pos;
268 /* Type of note. The 9 'from' trigraph characters represent those
269 trigraphs, '\\' an escaped newline, ' ' an escaped newline with
270 intervening space, and anything else is invalid. */
271 unsigned int type;
274 /* Represents the contents of a file cpplib has read in. */
275 struct cpp_buffer
277 const uchar *cur; /* Current location. */
278 const uchar *line_base; /* Start of current physical line. */
279 const uchar *next_line; /* Start of to-be-cleaned logical line. */
281 const uchar *buf; /* Entire character buffer. */
282 const uchar *rlimit; /* Writable byte at end of file. */
284 _cpp_line_note *notes; /* Array of notes. */
285 unsigned int cur_note; /* Next note to process. */
286 unsigned int notes_used; /* Number of notes. */
287 unsigned int notes_cap; /* Size of allocated array. */
289 struct cpp_buffer *prev;
291 /* Pointer into the file table; non-NULL if this is a file buffer.
292 Used for include_next and to record control macros. */
293 struct _cpp_file *file;
295 /* Value of if_stack at start of this file.
296 Used to prohibit unmatched #endif (etc) in an include file. */
297 struct if_stack *if_stack;
299 /* True if we need to get the next clean line. */
300 bool need_line;
302 /* True if we have already warned about C++ comments in this file.
303 The warning happens only for C89 extended mode with -pedantic on,
304 or for -Wtraditional, and only once per file (otherwise it would
305 be far too noisy). */
306 unsigned int warned_cplusplus_comments : 1;
308 /* True if we don't process trigraphs and escaped newlines. True
309 for preprocessed input, command line directives, and _Pragma
310 buffers. */
311 unsigned int from_stage3 : 1;
313 /* At EOF, a buffer is automatically popped. If RETURN_AT_EOF is
314 true, a CPP_EOF token is then returned. Otherwise, the next
315 token from the enclosing buffer is returned. */
316 unsigned int return_at_eof : 1;
318 /* One for a system header, two for a C system header file that therefore
319 needs to be extern "C" protected in C++, and zero otherwise. */
320 unsigned char sysp;
322 /* The directory of the this buffer's file. Its NAME member is not
323 allocated, so we don't need to worry about freeing it. */
324 struct cpp_dir dir;
326 /* Descriptor for converting from the input character set to the
327 source character set. */
328 struct cset_converter input_cset_desc;
331 /* A cpp_reader encapsulates the "state" of a pre-processor run.
332 Applying cpp_get_token repeatedly yields a stream of pre-processor
333 tokens. Usually, there is only one cpp_reader object active. */
334 struct cpp_reader
336 /* Top of buffer stack. */
337 cpp_buffer *buffer;
339 /* Overlaid buffer (can be different after processing #include). */
340 cpp_buffer *overlaid_buffer;
342 /* Lexer state. */
343 struct lexer_state state;
345 /* Source line tracking. */
346 struct line_maps *line_table;
347 fileline line;
349 /* The line of the '#' of the current directive. */
350 fileline directive_line;
352 /* Memory buffers. */
353 _cpp_buff *a_buff; /* Aligned permanent storage. */
354 _cpp_buff *u_buff; /* Unaligned permanent storage. */
355 _cpp_buff *free_buffs; /* Free buffer chain. */
357 /* Context stack. */
358 struct cpp_context base_context;
359 struct cpp_context *context;
361 /* If in_directive, the directive if known. */
362 const struct directive *directive;
364 /* Search paths for include files. */
365 struct cpp_dir *quote_include; /* "" */
366 struct cpp_dir *bracket_include; /* <> */
367 struct cpp_dir no_search_path; /* No path. */
369 /* Chain of all hashed _cpp_file instances. */
370 struct _cpp_file *all_files;
372 struct _cpp_file *main_file;
374 /* File and directory hash table. */
375 struct htab *file_hash;
376 struct file_hash_entry *file_hash_entries;
377 unsigned int file_hash_entries_allocated, file_hash_entries_used;
379 /* Nonzero means don't look for #include "foo" the source-file
380 directory. */
381 bool quote_ignores_source_dir;
383 /* Nonzero if any file has contained #pragma once or #import has
384 been used. */
385 bool seen_once_only;
387 /* Multiple include optimization. */
388 const cpp_hashnode *mi_cmacro;
389 const cpp_hashnode *mi_ind_cmacro;
390 bool mi_valid;
392 /* Lexing. */
393 cpp_token *cur_token;
394 tokenrun base_run, *cur_run;
395 unsigned int lookaheads;
397 /* Nonzero prevents the lexer from re-using the token runs. */
398 unsigned int keep_tokens;
400 /* Error counter for exit code. */
401 unsigned int errors;
403 /* Buffer to hold macro definition string. */
404 unsigned char *macro_buffer;
405 unsigned int macro_buffer_len;
407 /* Descriptor for converting from the source character set to the
408 execution character set. */
409 struct cset_converter narrow_cset_desc;
411 /* Descriptor for converting from the source character set to the
412 wide execution character set. */
413 struct cset_converter wide_cset_desc;
415 /* Date and time text. Calculated together if either is requested. */
416 const uchar *date;
417 const uchar *time;
419 /* EOF token, and a token forcing paste avoidance. */
420 cpp_token avoid_paste;
421 cpp_token eof;
423 /* Opaque handle to the dependencies of mkdeps.c. */
424 struct deps *deps;
426 /* Obstack holding all macro hash nodes. This never shrinks.
427 See cpphash.c */
428 struct obstack hash_ob;
430 /* Obstack holding buffer and conditional structures. This is a
431 real stack. See cpplib.c. */
432 struct obstack buffer_ob;
434 /* Pragma table - dynamic, because a library user can add to the
435 list of recognized pragmas. */
436 struct pragma_entry *pragmas;
438 /* Call backs to cpplib client. */
439 struct cpp_callbacks cb;
441 /* Identifier hash table. */
442 struct ht *hash_table;
444 /* Expression parser stack. */
445 struct op *op_stack, *op_limit;
447 /* User visible options. */
448 struct cpp_options opts;
450 /* Special nodes - identifiers with predefined significance to the
451 preprocessor. */
452 struct spec_nodes spec_nodes;
454 /* Whether cpplib owns the hashtable. */
455 bool our_hashtable;
457 /* Traditional preprocessing output buffer (a logical line). */
458 struct
460 uchar *base;
461 uchar *limit;
462 uchar *cur;
463 fileline first_line;
464 } out;
466 /* Used for buffer overlays by cpptrad.c. */
467 const uchar *saved_cur, *saved_rlimit, *saved_line_base;
469 /* Used to save the original line number during traditional
470 preprocessing. */
471 unsigned int saved_line;
473 /* A saved list of the defined macros, for dependency checking
474 of precompiled headers. */
475 struct cpp_savedstate *savedstate;
478 /* Character classes. Based on the more primitive macros in safe-ctype.h.
479 If the definition of `numchar' looks odd to you, please look up the
480 definition of a pp-number in the C standard [section 6.4.8 of C99].
482 In the unlikely event that characters other than \r and \n enter
483 the set is_vspace, the macro handle_newline() in cpplex.c must be
484 updated. */
485 #define _dollar_ok(x) ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
487 #define is_idchar(x) (ISIDNUM(x) || _dollar_ok(x))
488 #define is_numchar(x) ISIDNUM(x)
489 #define is_idstart(x) (ISIDST(x) || _dollar_ok(x))
490 #define is_numstart(x) ISDIGIT(x)
491 #define is_hspace(x) ISBLANK(x)
492 #define is_vspace(x) IS_VSPACE(x)
493 #define is_nvspace(x) IS_NVSPACE(x)
494 #define is_space(x) IS_SPACE_OR_NUL(x)
496 /* This table is constant if it can be initialized at compile time,
497 which is the case if cpp was compiled with GCC >=2.7, or another
498 compiler that supports C99. */
499 #if HAVE_DESIGNATED_INITIALIZERS
500 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
501 #else
502 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
503 #endif
505 /* Macros. */
507 static inline int cpp_in_system_header (cpp_reader *);
508 static inline int
509 cpp_in_system_header (cpp_reader *pfile)
511 return pfile->buffer ? pfile->buffer->sysp : 0;
513 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
514 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
516 /* In cpperror.c */
517 extern int _cpp_begin_message (cpp_reader *, int,
518 source_location, unsigned int);
520 /* In cppmacro.c */
521 extern void _cpp_free_definition (cpp_hashnode *);
522 extern bool _cpp_create_definition (cpp_reader *, cpp_hashnode *);
523 extern void _cpp_pop_context (cpp_reader *);
524 extern void _cpp_push_text_context (cpp_reader *, cpp_hashnode *,
525 const uchar *, size_t);
526 extern bool _cpp_save_parameter (cpp_reader *, cpp_macro *, cpp_hashnode *);
527 extern bool _cpp_arguments_ok (cpp_reader *, cpp_macro *, const cpp_hashnode *,
528 unsigned int);
529 extern const uchar *_cpp_builtin_macro_text (cpp_reader *, cpp_hashnode *);
530 int _cpp_warn_if_unused_macro (cpp_reader *, cpp_hashnode *, void *);
531 /* In cpphash.c */
532 extern void _cpp_init_hashtable (cpp_reader *, hash_table *);
533 extern void _cpp_destroy_hashtable (cpp_reader *);
535 /* In cppfiles.c */
536 typedef struct _cpp_file _cpp_file;
537 extern _cpp_file *_cpp_find_file (cpp_reader *, const char *fname,
538 cpp_dir *start_dir, bool fake);
539 extern bool _cpp_find_failed (_cpp_file *);
540 extern void _cpp_mark_file_once_only (cpp_reader *, struct _cpp_file *);
541 extern void _cpp_fake_include (cpp_reader *, const char *);
542 extern bool _cpp_stack_file (cpp_reader *, _cpp_file*, bool);
543 extern bool _cpp_stack_include (cpp_reader *, const char *, int,
544 enum include_type);
545 extern int _cpp_compare_file_date (cpp_reader *, const char *, int);
546 extern void _cpp_report_missing_guards (cpp_reader *);
547 extern void _cpp_init_files (cpp_reader *);
548 extern void _cpp_cleanup_files (cpp_reader *);
549 extern void _cpp_pop_file_buffer (cpp_reader *, struct _cpp_file *);
550 extern bool _cpp_save_file_entries (cpp_reader *pfile, FILE *f);
551 extern bool _cpp_read_file_entries (cpp_reader *, FILE *);
553 /* In cppexp.c */
554 extern bool _cpp_parse_expr (cpp_reader *);
555 extern struct op *_cpp_expand_op_stack (cpp_reader *);
557 /* In cpplex.c */
558 extern void _cpp_process_line_notes (cpp_reader *, int);
559 extern void _cpp_clean_line (cpp_reader *);
560 extern bool _cpp_get_fresh_line (cpp_reader *);
561 extern bool _cpp_skip_block_comment (cpp_reader *);
562 extern cpp_token *_cpp_temp_token (cpp_reader *);
563 extern const cpp_token *_cpp_lex_token (cpp_reader *);
564 extern cpp_token *_cpp_lex_direct (cpp_reader *);
565 extern int _cpp_equiv_tokens (const cpp_token *, const cpp_token *);
566 extern void _cpp_init_tokenrun (tokenrun *, unsigned int);
568 /* In cppinit.c. */
569 extern void _cpp_maybe_push_include_file (cpp_reader *);
571 /* In cpplib.c */
572 extern int _cpp_test_assertion (cpp_reader *, unsigned int *);
573 extern int _cpp_handle_directive (cpp_reader *, int);
574 extern void _cpp_define_builtin (cpp_reader *, const char *);
575 extern char ** _cpp_save_pragma_names (cpp_reader *);
576 extern void _cpp_restore_pragma_names (cpp_reader *, char **);
577 extern void _cpp_do__Pragma (cpp_reader *);
578 extern void _cpp_init_directives (cpp_reader *);
579 extern void _cpp_init_internal_pragmas (cpp_reader *);
580 extern void _cpp_do_file_change (cpp_reader *, enum lc_reason, const char *,
581 unsigned int, unsigned int);
582 extern void _cpp_pop_buffer (cpp_reader *);
584 /* In cpptrad.c. */
585 extern bool _cpp_scan_out_logical_line (cpp_reader *, cpp_macro *);
586 extern bool _cpp_read_logical_line_trad (cpp_reader *);
587 extern void _cpp_overlay_buffer (cpp_reader *pfile, const uchar *, size_t);
588 extern void _cpp_remove_overlay (cpp_reader *);
589 extern bool _cpp_create_trad_definition (cpp_reader *, cpp_macro *);
590 extern bool _cpp_expansions_different_trad (const cpp_macro *,
591 const cpp_macro *);
592 extern uchar *_cpp_copy_replacement_text (const cpp_macro *, uchar *);
593 extern size_t _cpp_replacement_text_len (const cpp_macro *);
595 /* In cppcharset.c. */
596 extern cppchar_t _cpp_valid_ucn (cpp_reader *, const uchar **,
597 const uchar *, int);
598 extern void _cpp_destroy_iconv (cpp_reader *);
599 extern uchar *_cpp_convert_input (cpp_reader *, const char *, uchar *,
600 size_t, size_t, off_t *);
601 extern const char *_cpp_default_encoding (void);
603 /* Utility routines and macros. */
604 #define DSC(str) (const uchar *)str, sizeof str - 1
605 #define xnew(T) (T *) xmalloc (sizeof(T))
606 #define xcnew(T) (T *) xcalloc (1, sizeof(T))
607 #define xnewvec(T, N) (T *) xmalloc (sizeof(T) * (N))
608 #define xcnewvec(T, N) (T *) xcalloc (N, sizeof(T))
609 #define xobnew(O, T) (T *) obstack_alloc (O, sizeof(T))
611 /* These are inline functions instead of macros so we can get type
612 checking. */
613 static inline int ustrcmp (const uchar *, const uchar *);
614 static inline int ustrncmp (const uchar *, const uchar *, size_t);
615 static inline size_t ustrlen (const uchar *);
616 static inline uchar *uxstrdup (const uchar *);
617 static inline uchar *ustrchr (const uchar *, int);
618 static inline int ufputs (const uchar *, FILE *);
620 static inline int
621 ustrcmp (const uchar *s1, const uchar *s2)
623 return strcmp ((const char *)s1, (const char *)s2);
626 static inline int
627 ustrncmp (const uchar *s1, const uchar *s2, size_t n)
629 return strncmp ((const char *)s1, (const char *)s2, n);
632 static inline size_t
633 ustrlen (const uchar *s1)
635 return strlen ((const char *)s1);
638 static inline uchar *
639 uxstrdup (const uchar *s1)
641 return (uchar *) xstrdup ((const char *)s1);
644 static inline uchar *
645 ustrchr (const uchar *s1, int c)
647 return (uchar *) strchr ((const char *)s1, c);
650 static inline int
651 ufputs (const uchar *s, FILE *f)
653 return fputs ((const char *)s, f);
656 #endif /* ! GCC_CPPHASH_H */