1 /* Part of CPP library.
2    Copyright (C) 1997-2016 Free Software Foundation, Inc.
3 
4 This program is free software; you can redistribute it and/or modify it
5 under the terms of the GNU General Public License as published by the
6 Free Software Foundation; either version 3, or (at your option) any
7 later version.
8 
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
12 GNU General Public License for more details.
13 
14 You should have received a copy of the GNU General Public License
15 along with this program; see the file COPYING3.  If not see
16 <http://www.gnu.org/licenses/>.  */
17 
18 /* This header defines all the internal data structures and functions
19    that need to be visible across files.  It should not be used outside
20    cpplib.  */
21 
22 #ifndef LIBCPP_INTERNAL_H
23 #define LIBCPP_INTERNAL_H
24 
25 #include "symtab.h"
26 #include "cpp-id-data.h"
27 
28 #if HAVE_ICONV
29 #include <iconv.h>
30 #else
31 #define HAVE_ICONV 0
32 typedef int iconv_t;  /* dummy */
33 #endif
34 
35 #ifdef __cplusplus
36 extern "C" {
37 #endif
38 
39 struct directive;		/* Deliberately incomplete.  */
40 struct pending_option;
41 struct op;
42 struct _cpp_strbuf;
43 
44 typedef bool (*convert_f) (iconv_t, const unsigned char *, size_t,
45 			   struct _cpp_strbuf *);
46 struct cset_converter
47 {
48   convert_f func;
49   iconv_t cd;
50   int width;
51 };
52 
53 #define BITS_PER_CPPCHAR_T (CHAR_BIT * sizeof (cppchar_t))
54 
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))))
61 
62 #define DIGIT_SEP(c) ((c) == '\'' && CPP_OPTION (pfile, digit_separators))
63 
64 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
65 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
66 #define CPP_BUF_COLUMN(BUF, CUR) ((CUR) - (BUF)->line_base)
67 #define CPP_BUF_COL(BUF) CPP_BUF_COLUMN(BUF, (BUF)->cur)
68 
69 #define CPP_INCREMENT_LINE(PFILE, COLS_HINT) do { \
70     const struct line_maps *line_table = PFILE->line_table; \
71     const struct line_map_ordinary *map = \
72       LINEMAPS_LAST_ORDINARY_MAP (line_table); \
73     linenum_type line = SOURCE_LINE (map, line_table->highest_line); \
74     linemap_line_start (PFILE->line_table, line + 1, COLS_HINT); \
75   } while (0)
76 
77 /* Maximum nesting of cpp_buffers.  We use a static limit, partly for
78    efficiency, and partly to limit runaway recursion.  */
79 #define CPP_STACK_MAX 200
80 
81 /* Host alignment handling.  */
82 struct dummy
83 {
84   char c;
85   union
86   {
87     double d;
88     int *p;
89   } u;
90 };
91 
92 #define DEFAULT_ALIGNMENT offsetof (struct dummy, u)
93 #define CPP_ALIGN2(size, align) (((size) + ((align) - 1)) & ~((align) - 1))
94 #define CPP_ALIGN(size) CPP_ALIGN2 (size, DEFAULT_ALIGNMENT)
95 
96 #define _cpp_mark_macro_used(NODE) do {					\
97   if ((NODE)->type == NT_MACRO && !((NODE)->flags & NODE_BUILTIN))	\
98     (NODE)->value.macro->used = 1; } while (0)
99 
100 /* A generic memory buffer, and operations on it.  */
101 typedef struct _cpp_buff _cpp_buff;
102 struct _cpp_buff
103 {
104   struct _cpp_buff *next;
105   unsigned char *base, *cur, *limit;
106 };
107 
108 extern _cpp_buff *_cpp_get_buff (cpp_reader *, size_t);
109 extern void _cpp_release_buff (cpp_reader *, _cpp_buff *);
110 extern void _cpp_extend_buff (cpp_reader *, _cpp_buff **, size_t);
111 extern _cpp_buff *_cpp_append_extend_buff (cpp_reader *, _cpp_buff *, size_t);
112 extern void _cpp_free_buff (_cpp_buff *);
113 extern unsigned char *_cpp_aligned_alloc (cpp_reader *, size_t);
114 extern unsigned char *_cpp_unaligned_alloc (cpp_reader *, size_t);
115 
116 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
117 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
118 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
119 
120 /* #include types.  */
121 enum include_type {IT_INCLUDE, IT_INCLUDE_NEXT, IT_IMPORT, IT_CMDLINE, IT_DEFAULT};
122 
123 union utoken
124 {
125   const cpp_token *token;
126   const cpp_token **ptoken;
127 };
128 
129 /* A "run" of tokens; part of a chain of runs.  */
130 typedef struct tokenrun tokenrun;
131 struct tokenrun
132 {
133   tokenrun *next, *prev;
134   cpp_token *base, *limit;
135 };
136 
137 /* Accessor macros for struct cpp_context.  */
138 #define FIRST(c) ((c)->u.iso.first)
139 #define LAST(c) ((c)->u.iso.last)
140 #define CUR(c) ((c)->u.trad.cur)
141 #define RLIMIT(c) ((c)->u.trad.rlimit)
142 
143 /* This describes some additional data that is added to the macro
144    token context of type cpp_context, when -ftrack-macro-expansion is
145    on.  */
146 typedef struct
147 {
148   /* The node of the macro we are referring to.  */
149   cpp_hashnode *macro_node;
150   /* This buffer contains an array of virtual locations.  The virtual
151      location at index 0 is the virtual location of the token at index
152      0 in the current instance of cpp_context; similarly for all the
153      other virtual locations.  */
154   source_location *virt_locs;
155   /* This is a pointer to the current virtual location.  This is used
156      to iterate over the virtual locations while we iterate over the
157      tokens they belong to.  */
158   source_location *cur_virt_loc;
159 } macro_context;
160 
161 /* The kind of tokens carried by a cpp_context.  */
162 enum context_tokens_kind {
163   /* This is the value of cpp_context::tokens_kind if u.iso.first
164      contains an instance of cpp_token **.  */
165   TOKENS_KIND_INDIRECT,
166   /* This is the value of cpp_context::tokens_kind if u.iso.first
167      contains an instance of cpp_token *.  */
168   TOKENS_KIND_DIRECT,
169   /* This is the value of cpp_context::tokens_kind when the token
170      context contains tokens resulting from macro expansion.  In that
171      case struct cpp_context::macro points to an instance of struct
172      macro_context.  This is used only when the
173      -ftrack-macro-expansion flag is on.  */
174   TOKENS_KIND_EXTENDED
175 };
176 
177 typedef struct cpp_context cpp_context;
178 struct cpp_context
179 {
180   /* Doubly-linked list.  */
181   cpp_context *next, *prev;
182 
183   union
184   {
185     /* For ISO macro expansion.  Contexts other than the base context
186        are contiguous tokens.  e.g. macro expansions, expanded
187        argument tokens.  */
188     struct
189     {
190       union utoken first;
191       union utoken last;
192     } iso;
193 
194     /* For traditional macro expansion.  */
195     struct
196     {
197       const unsigned char *cur;
198       const unsigned char *rlimit;
199     } trad;
200   } u;
201 
202   /* If non-NULL, a buffer used for storage related to this context.
203      When the context is popped, the buffer is released.  */
204   _cpp_buff *buff;
205 
206   /* If tokens_kind is TOKEN_KIND_EXTENDED, then (as we thus are in a
207      macro context) this is a pointer to an instance of macro_context.
208      Otherwise if tokens_kind is *not* TOKEN_KIND_EXTENDED, then, if
209      we are in a macro context, this is a pointer to an instance of
210      cpp_hashnode, representing the name of the macro this context is
211      for.  If we are not in a macro context, then this is just NULL.
212      Note that when tokens_kind is TOKEN_KIND_EXTENDED, the memory
213      used by the instance of macro_context pointed to by this member
214      is de-allocated upon de-allocation of the instance of struct
215      cpp_context.  */
216   union
217   {
218     macro_context *mc;
219     cpp_hashnode *macro;
220   } c;
221 
222   /* This determines the type of tokens held by this context.  */
223   enum context_tokens_kind tokens_kind;
224 };
225 
226 struct lexer_state
227 {
228   /* Nonzero if first token on line is CPP_HASH.  */
229   unsigned char in_directive;
230 
231   /* Nonzero if in a directive that will handle padding tokens itself.
232      #include needs this to avoid problems with computed include and
233      spacing between tokens.  */
234   unsigned char directive_wants_padding;
235 
236   /* True if we are skipping a failed conditional group.  */
237   unsigned char skipping;
238 
239   /* Nonzero if in a directive that takes angle-bracketed headers.  */
240   unsigned char angled_headers;
241 
242   /* Nonzero if in a #if or #elif directive.  */
243   unsigned char in_expression;
244 
245   /* Nonzero to save comments.  Turned off if discard_comments, and in
246      all directives apart from #define.  */
247   unsigned char save_comments;
248 
249   /* Nonzero if lexing __VA_ARGS__ is valid.  */
250   unsigned char va_args_ok;
251 
252   /* Nonzero if lexing poisoned identifiers is valid.  */
253   unsigned char poisoned_ok;
254 
255   /* Nonzero to prevent macro expansion.  */
256   unsigned char prevent_expansion;
257 
258   /* Nonzero when parsing arguments to a function-like macro.  */
259   unsigned char parsing_args;
260 
261   /* Nonzero if in a __has_include__ or __has_include_next__ statement.  */
262   unsigned char in__has_include__;
263 
264   /* Nonzero if prevent_expansion is true only because output is
265      being discarded.  */
266   unsigned char discarding_output;
267 
268   /* Nonzero to skip evaluating part of an expression.  */
269   unsigned int skip_eval;
270 
271   /* Nonzero when handling a deferred pragma.  */
272   unsigned char in_deferred_pragma;
273 
274   /* Nonzero if the deferred pragma being handled allows macro expansion.  */
275   unsigned char pragma_allow_expansion;
276 };
277 
278 /* Special nodes - identifiers with predefined significance.  */
279 struct spec_nodes
280 {
281   cpp_hashnode *n_defined;		/* defined operator */
282   cpp_hashnode *n_true;			/* C++ keyword true */
283   cpp_hashnode *n_false;		/* C++ keyword false */
284   cpp_hashnode *n__VA_ARGS__;		/* C99 vararg macros */
285   cpp_hashnode *n__has_include__;	/* __has_include__ operator */
286   cpp_hashnode *n__has_include_next__;	/* __has_include_next__ operator */
287 };
288 
289 typedef struct _cpp_line_note _cpp_line_note;
290 struct _cpp_line_note
291 {
292   /* Location in the clean line the note refers to.  */
293   const unsigned char *pos;
294 
295   /* Type of note.  The 9 'from' trigraph characters represent those
296      trigraphs, '\\' an escaped newline, ' ' an escaped newline with
297      intervening space, 0 represents a note that has already been handled,
298      and anything else is invalid.  */
299   unsigned int type;
300 };
301 
302 /* Represents the contents of a file cpplib has read in.  */
303 struct cpp_buffer
304 {
305   const unsigned char *cur;        /* Current location.  */
306   const unsigned char *line_base;  /* Start of current physical line.  */
307   const unsigned char *next_line;  /* Start of to-be-cleaned logical line.  */
308 
309   const unsigned char *buf;        /* Entire character buffer.  */
310   const unsigned char *rlimit;     /* Writable byte at end of file.  */
311   const unsigned char *to_free;	   /* Pointer that should be freed when
312 				      popping the buffer.  */
313 
314   _cpp_line_note *notes;           /* Array of notes.  */
315   unsigned int cur_note;           /* Next note to process.  */
316   unsigned int notes_used;         /* Number of notes.  */
317   unsigned int notes_cap;          /* Size of allocated array.  */
318 
319   struct cpp_buffer *prev;
320 
321   /* Pointer into the file table; non-NULL if this is a file buffer.
322      Used for include_next and to record control macros.  */
323   struct _cpp_file *file;
324 
325   /* Saved value of __TIMESTAMP__ macro - date and time of last modification
326      of the assotiated file.  */
327   const unsigned char *timestamp;
328 
329   /* Value of if_stack at start of this file.
330      Used to prohibit unmatched #endif (etc) in an include file.  */
331   struct if_stack *if_stack;
332 
333   /* True if we need to get the next clean line.  */
334   bool need_line;
335 
336   /* True if we have already warned about C++ comments in this file.
337      The warning happens only for C89 extended mode with -pedantic on,
338      or for -Wtraditional, and only once per file (otherwise it would
339      be far too noisy).  */
340   unsigned int warned_cplusplus_comments : 1;
341 
342   /* True if we don't process trigraphs and escaped newlines.  True
343      for preprocessed input, command line directives, and _Pragma
344      buffers.  */
345   unsigned int from_stage3 : 1;
346 
347   /* At EOF, a buffer is automatically popped.  If RETURN_AT_EOF is
348      true, a CPP_EOF token is then returned.  Otherwise, the next
349      token from the enclosing buffer is returned.  */
350   unsigned int return_at_eof : 1;
351 
352   /* One for a system header, two for a C system header file that therefore
353      needs to be extern "C" protected in C++, and zero otherwise.  */
354   unsigned char sysp;
355 
356   /* The directory of the this buffer's file.  Its NAME member is not
357      allocated, so we don't need to worry about freeing it.  */
358   struct cpp_dir dir;
359 
360   /* Descriptor for converting from the input character set to the
361      source character set.  */
362   struct cset_converter input_cset_desc;
363 };
364 
365 /* The list of saved macros by push_macro pragma.  */
366 struct def_pragma_macro {
367   /* Chain element to previous saved macro.  */
368   struct def_pragma_macro *next;
369   /* Name of the macro.  */
370   char *name;
371   /* The stored macro content.  */
372   unsigned char *definition;
373 
374   /* Definition line number.  */
375   source_location line;
376   /* If macro defined in system header.  */
377   unsigned int syshdr   : 1;
378   /* Nonzero if it has been expanded or had its existence tested.  */
379   unsigned int used     : 1;
380 
381   /* Mark if we save an undefined macro.  */
382   unsigned int is_undef : 1;
383 };
384 
385 /* A cpp_reader encapsulates the "state" of a pre-processor run.
386    Applying cpp_get_token repeatedly yields a stream of pre-processor
387    tokens.  Usually, there is only one cpp_reader object active.  */
388 struct cpp_reader
389 {
390   /* Top of buffer stack.  */
391   cpp_buffer *buffer;
392 
393   /* Overlaid buffer (can be different after processing #include).  */
394   cpp_buffer *overlaid_buffer;
395 
396   /* Lexer state.  */
397   struct lexer_state state;
398 
399   /* Source line tracking.  */
400   struct line_maps *line_table;
401 
402   /* The line of the '#' of the current directive.  */
403   source_location directive_line;
404 
405   /* Memory buffers.  */
406   _cpp_buff *a_buff;		/* Aligned permanent storage.  */
407   _cpp_buff *u_buff;		/* Unaligned permanent storage.  */
408   _cpp_buff *free_buffs;	/* Free buffer chain.  */
409 
410   /* Context stack.  */
411   struct cpp_context base_context;
412   struct cpp_context *context;
413 
414   /* If in_directive, the directive if known.  */
415   const struct directive *directive;
416 
417   /* Token generated while handling a directive, if any. */
418   cpp_token directive_result;
419 
420   /* When expanding a macro at top-level, this is the location of the
421      macro invocation.  */
422   source_location invocation_location;
423 
424   /* This is the node representing the macro being expanded at
425      top-level.  The value of this data member is valid iff
426      in_macro_expansion_p() returns TRUE.  */
427   cpp_hashnode *top_most_macro_node;
428 
429   /* Nonzero if we are about to expand a macro.  Note that if we are
430      really expanding a macro, the function macro_of_context returns
431      the macro being expanded and this flag is set to false.  Client
432      code should use the function in_macro_expansion_p to know if we
433      are either about to expand a macro, or are actually expanding
434      one.  */
435   bool about_to_expand_macro_p;
436 
437   /* Search paths for include files.  */
438   struct cpp_dir *quote_include;	/* "" */
439   struct cpp_dir *bracket_include;	/* <> */
440   struct cpp_dir no_search_path;	/* No path.  */
441 
442   /* Chain of all hashed _cpp_file instances.  */
443   struct _cpp_file *all_files;
444 
445   struct _cpp_file *main_file;
446 
447   /* File and directory hash table.  */
448   struct htab *file_hash;
449   struct htab *dir_hash;
450   struct file_hash_entry_pool *file_hash_entries;
451 
452   /* Negative path lookup hash table.  */
453   struct htab *nonexistent_file_hash;
454   struct obstack nonexistent_file_ob;
455 
456   /* Nonzero means don't look for #include "foo" the source-file
457      directory.  */
458   bool quote_ignores_source_dir;
459 
460   /* Nonzero if any file has contained #pragma once or #import has
461      been used.  */
462   bool seen_once_only;
463 
464   /* Multiple include optimization.  */
465   const cpp_hashnode *mi_cmacro;
466   const cpp_hashnode *mi_ind_cmacro;
467   bool mi_valid;
468 
469   /* Lexing.  */
470   cpp_token *cur_token;
471   tokenrun base_run, *cur_run;
472   unsigned int lookaheads;
473 
474   /* Nonzero prevents the lexer from re-using the token runs.  */
475   unsigned int keep_tokens;
476 
477   /* Buffer to hold macro definition string.  */
478   unsigned char *macro_buffer;
479   unsigned int macro_buffer_len;
480 
481   /* Descriptor for converting from the source character set to the
482      execution character set.  */
483   struct cset_converter narrow_cset_desc;
484 
485   /* Descriptor for converting from the source character set to the
486      UTF-8 execution character set.  */
487   struct cset_converter utf8_cset_desc;
488 
489   /* Descriptor for converting from the source character set to the
490      UTF-16 execution character set.  */
491   struct cset_converter char16_cset_desc;
492 
493   /* Descriptor for converting from the source character set to the
494      UTF-32 execution character set.  */
495   struct cset_converter char32_cset_desc;
496 
497   /* Descriptor for converting from the source character set to the
498      wide execution character set.  */
499   struct cset_converter wide_cset_desc;
500 
501   /* Date and time text.  Calculated together if either is requested.  */
502   const unsigned char *date;
503   const unsigned char *time;
504 
505   /* EOF token, and a token forcing paste avoidance.  */
506   cpp_token avoid_paste;
507   cpp_token eof;
508 
509   /* Opaque handle to the dependencies of mkdeps.c.  */
510   struct deps *deps;
511 
512   /* Obstack holding all macro hash nodes.  This never shrinks.
513      See identifiers.c */
514   struct obstack hash_ob;
515 
516   /* Obstack holding buffer and conditional structures.  This is a
517      real stack.  See directives.c.  */
518   struct obstack buffer_ob;
519 
520   /* Pragma table - dynamic, because a library user can add to the
521      list of recognized pragmas.  */
522   struct pragma_entry *pragmas;
523 
524   /* Call backs to cpplib client.  */
525   struct cpp_callbacks cb;
526 
527   /* Identifier hash table.  */
528   struct ht *hash_table;
529 
530   /* Expression parser stack.  */
531   struct op *op_stack, *op_limit;
532 
533   /* User visible options.  */
534   struct cpp_options opts;
535 
536   /* Special nodes - identifiers with predefined significance to the
537      preprocessor.  */
538   struct spec_nodes spec_nodes;
539 
540   /* Whether cpplib owns the hashtable.  */
541   bool our_hashtable;
542 
543   /* Traditional preprocessing output buffer (a logical line).  */
544   struct
545   {
546     unsigned char *base;
547     unsigned char *limit;
548     unsigned char *cur;
549     source_location first_line;
550   } out;
551 
552   /* Used for buffer overlays by traditional.c.  */
553   const unsigned char *saved_cur, *saved_rlimit, *saved_line_base;
554 
555   /* A saved list of the defined macros, for dependency checking
556      of precompiled headers.  */
557   struct cpp_savedstate *savedstate;
558 
559   /* Next value of __COUNTER__ macro. */
560   unsigned int counter;
561 
562   /* Table of comments, when state.save_comments is true.  */
563   cpp_comment_table comments;
564 
565   /* List of saved macros by push_macro.  */
566   struct def_pragma_macro *pushed_macros;
567 
568   /* If non-null, the lexer will use this location for the next token
569      instead of getting a location from the linemap.  */
570   source_location *forced_token_location_p;
571 };
572 
573 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
574    If the definition of `numchar' looks odd to you, please look up the
575    definition of a pp-number in the C standard [section 6.4.8 of C99].
576 
577    In the unlikely event that characters other than \r and \n enter
578    the set is_vspace, the macro handle_newline() in lex.c must be
579    updated.  */
580 #define _dollar_ok(x)	((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
581 
582 #define is_idchar(x)	(ISIDNUM(x) || _dollar_ok(x))
583 #define is_numchar(x)	ISIDNUM(x)
584 #define is_idstart(x)	(ISIDST(x) || _dollar_ok(x))
585 #define is_numstart(x)	ISDIGIT(x)
586 #define is_hspace(x)	ISBLANK(x)
587 #define is_vspace(x)	IS_VSPACE(x)
588 #define is_nvspace(x)	IS_NVSPACE(x)
589 #define is_space(x)	IS_SPACE_OR_NUL(x)
590 
591 /* This table is constant if it can be initialized at compile time,
592    which is the case if cpp was compiled with GCC >=2.7, or another
593    compiler that supports C99.  */
594 #if HAVE_DESIGNATED_INITIALIZERS
595 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
596 #else
597 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
598 #endif
599 
600 /* Macros.  */
601 
602 static inline int cpp_in_system_header (cpp_reader *);
603 static inline int
cpp_in_system_header(cpp_reader * pfile)604 cpp_in_system_header (cpp_reader *pfile)
605 {
606   return pfile->buffer ? pfile->buffer->sysp : 0;
607 }
608 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, cpp_pedantic)
609 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, cpp_warn_traditional)
610 
611 static inline int cpp_in_primary_file (cpp_reader *);
612 static inline int
cpp_in_primary_file(cpp_reader * pfile)613 cpp_in_primary_file (cpp_reader *pfile)
614 {
615   return pfile->line_table->depth == 1;
616 }
617 
618 /* In macro.c */
619 extern void _cpp_free_definition (cpp_hashnode *);
620 extern bool _cpp_create_definition (cpp_reader *, cpp_hashnode *);
621 extern void _cpp_pop_context (cpp_reader *);
622 extern void _cpp_push_text_context (cpp_reader *, cpp_hashnode *,
623 				    const unsigned char *, size_t);
624 extern bool _cpp_save_parameter (cpp_reader *, cpp_macro *, cpp_hashnode *,
625 				 cpp_hashnode *);
626 extern bool _cpp_arguments_ok (cpp_reader *, cpp_macro *, const cpp_hashnode *,
627 			       unsigned int);
628 extern const unsigned char *_cpp_builtin_macro_text (cpp_reader *,
629 						     cpp_hashnode *,
630 						     source_location = 0);
631 extern int _cpp_warn_if_unused_macro (cpp_reader *, cpp_hashnode *, void *);
632 extern void _cpp_push_token_context (cpp_reader *, cpp_hashnode *,
633 				     const cpp_token *, unsigned int);
634 extern void _cpp_backup_tokens_direct (cpp_reader *, unsigned int);
635 
636 /* In identifiers.c */
637 extern void _cpp_init_hashtable (cpp_reader *, cpp_hash_table *);
638 extern void _cpp_destroy_hashtable (cpp_reader *);
639 
640 /* In files.c */
641 typedef struct _cpp_file _cpp_file;
642 extern _cpp_file *_cpp_find_file (cpp_reader *, const char *, cpp_dir *,
643 				  bool, int, bool);
644 extern bool _cpp_find_failed (_cpp_file *);
645 extern void _cpp_mark_file_once_only (cpp_reader *, struct _cpp_file *);
646 extern void _cpp_fake_include (cpp_reader *, const char *);
647 extern bool _cpp_stack_file (cpp_reader *, _cpp_file*, bool);
648 extern bool _cpp_stack_include (cpp_reader *, const char *, int,
649 				enum include_type);
650 extern int _cpp_compare_file_date (cpp_reader *, const char *, int);
651 extern void _cpp_report_missing_guards (cpp_reader *);
652 extern void _cpp_init_files (cpp_reader *);
653 extern void _cpp_cleanup_files (cpp_reader *);
654 extern void _cpp_pop_file_buffer (cpp_reader *, struct _cpp_file *,
655 				  const unsigned char *);
656 extern bool _cpp_save_file_entries (cpp_reader *pfile, FILE *f);
657 extern bool _cpp_read_file_entries (cpp_reader *, FILE *);
658 extern const char *_cpp_get_file_name (_cpp_file *);
659 extern struct stat *_cpp_get_file_stat (_cpp_file *);
660 extern bool _cpp_has_header (cpp_reader *, const char *, int,
661 			     enum include_type);
662 
663 /* In expr.c */
664 extern bool _cpp_parse_expr (cpp_reader *, bool);
665 extern struct op *_cpp_expand_op_stack (cpp_reader *);
666 
667 /* In lex.c */
668 extern void _cpp_process_line_notes (cpp_reader *, int);
669 extern void _cpp_clean_line (cpp_reader *);
670 extern bool _cpp_get_fresh_line (cpp_reader *);
671 extern bool _cpp_skip_block_comment (cpp_reader *);
672 extern cpp_token *_cpp_temp_token (cpp_reader *);
673 extern const cpp_token *_cpp_lex_token (cpp_reader *);
674 extern cpp_token *_cpp_lex_direct (cpp_reader *);
675 extern unsigned char *_cpp_spell_ident_ucns (unsigned char *, cpp_hashnode *);
676 extern int _cpp_equiv_tokens (const cpp_token *, const cpp_token *);
677 extern void _cpp_init_tokenrun (tokenrun *, unsigned int);
678 extern cpp_hashnode *_cpp_lex_identifier (cpp_reader *, const char *);
679 extern int _cpp_remaining_tokens_num_in_context (cpp_context *);
680 extern void _cpp_init_lexer (void);
681 
682 /* In init.c.  */
683 extern void _cpp_maybe_push_include_file (cpp_reader *);
684 extern const char *cpp_named_operator2name (enum cpp_ttype type);
685 
686 /* In directives.c */
687 extern int _cpp_test_assertion (cpp_reader *, unsigned int *);
688 extern int _cpp_handle_directive (cpp_reader *, int);
689 extern void _cpp_define_builtin (cpp_reader *, const char *);
690 extern char ** _cpp_save_pragma_names (cpp_reader *);
691 extern void _cpp_restore_pragma_names (cpp_reader *, char **);
692 extern int _cpp_do__Pragma (cpp_reader *, source_location);
693 extern void _cpp_init_directives (cpp_reader *);
694 extern void _cpp_init_internal_pragmas (cpp_reader *);
695 extern void _cpp_do_file_change (cpp_reader *, enum lc_reason, const char *,
696 				 linenum_type, unsigned int);
697 extern void _cpp_pop_buffer (cpp_reader *);
698 extern char *_cpp_bracket_include (cpp_reader *);
699 
700 /* In directives.c */
701 struct _cpp_dir_only_callbacks
702 {
703   /* Called to print a block of lines. */
704   void (*print_lines) (int, const void *, size_t);
705   void (*maybe_print_line) (source_location);
706 };
707 
708 extern void _cpp_preprocess_dir_only (cpp_reader *,
709 				      const struct _cpp_dir_only_callbacks *);
710 
711 /* In traditional.c.  */
712 extern bool _cpp_scan_out_logical_line (cpp_reader *, cpp_macro *, bool);
713 extern bool _cpp_read_logical_line_trad (cpp_reader *);
714 extern void _cpp_overlay_buffer (cpp_reader *pfile, const unsigned char *,
715 				 size_t);
716 extern void _cpp_remove_overlay (cpp_reader *);
717 extern bool _cpp_create_trad_definition (cpp_reader *, cpp_macro *);
718 extern bool _cpp_expansions_different_trad (const cpp_macro *,
719 					    const cpp_macro *);
720 extern unsigned char *_cpp_copy_replacement_text (const cpp_macro *,
721 						  unsigned char *);
722 extern size_t _cpp_replacement_text_len (const cpp_macro *);
723 
724 /* In charset.c.  */
725 
726 /* The normalization state at this point in the sequence.
727    It starts initialized to all zeros, and at the end
728    'level' is the normalization level of the sequence.  */
729 
730 struct normalize_state
731 {
732   /* The previous starter character.  */
733   cppchar_t previous;
734   /* The combining class of the previous character (whether or not a
735      starter).  */
736   unsigned char prev_class;
737   /* The lowest normalization level so far.  */
738   enum cpp_normalize_level level;
739 };
740 #define INITIAL_NORMALIZE_STATE { 0, 0, normalized_KC }
741 #define NORMALIZE_STATE_RESULT(st) ((st)->level)
742 
743 /* We saw a character C that matches ISIDNUM(), update a
744    normalize_state appropriately.  */
745 #define NORMALIZE_STATE_UPDATE_IDNUM(st, c)	\
746   ((st)->previous = (c), (st)->prev_class = 0)
747 
748 extern bool _cpp_valid_ucn (cpp_reader *, const unsigned char **,
749 			    const unsigned char *, int,
750 			    struct normalize_state *state,
751 			    cppchar_t *);
752 extern void _cpp_destroy_iconv (cpp_reader *);
753 extern unsigned char *_cpp_convert_input (cpp_reader *, const char *,
754 					  unsigned char *, size_t, size_t,
755 					  const unsigned char **, off_t *);
756 extern const char *_cpp_default_encoding (void);
757 extern cpp_hashnode * _cpp_interpret_identifier (cpp_reader *pfile,
758 						 const unsigned char *id,
759 						 size_t len);
760 
761 /* Utility routines and macros.  */
762 #define DSC(str) (const unsigned char *)str, sizeof str - 1
763 
764 /* These are inline functions instead of macros so we can get type
765    checking.  */
766 static inline int ustrcmp (const unsigned char *, const unsigned char *);
767 static inline int ustrncmp (const unsigned char *, const unsigned char *,
768 			    size_t);
769 static inline size_t ustrlen (const unsigned char *);
770 static inline const unsigned char *uxstrdup (const unsigned char *);
771 static inline const unsigned char *ustrchr (const unsigned char *, int);
772 static inline int ufputs (const unsigned char *, FILE *);
773 
774 /* Use a const char for the second parameter since it is usually a literal.  */
775 static inline int ustrcspn (const unsigned char *, const char *);
776 
777 static inline int
ustrcmp(const unsigned char * s1,const unsigned char * s2)778 ustrcmp (const unsigned char *s1, const unsigned char *s2)
779 {
780   return strcmp ((const char *)s1, (const char *)s2);
781 }
782 
783 static inline int
ustrncmp(const unsigned char * s1,const unsigned char * s2,size_t n)784 ustrncmp (const unsigned char *s1, const unsigned char *s2, size_t n)
785 {
786   return strncmp ((const char *)s1, (const char *)s2, n);
787 }
788 
789 static inline int
ustrcspn(const unsigned char * s1,const char * s2)790 ustrcspn (const unsigned char *s1, const char *s2)
791 {
792   return strcspn ((const char *)s1, s2);
793 }
794 
795 static inline size_t
ustrlen(const unsigned char * s1)796 ustrlen (const unsigned char *s1)
797 {
798   return strlen ((const char *)s1);
799 }
800 
801 static inline const unsigned char *
uxstrdup(const unsigned char * s1)802 uxstrdup (const unsigned char *s1)
803 {
804   return (const unsigned char *) xstrdup ((const char *)s1);
805 }
806 
807 static inline const unsigned char *
ustrchr(const unsigned char * s1,int c)808 ustrchr (const unsigned char *s1, int c)
809 {
810   return (const unsigned char *) strchr ((const char *)s1, c);
811 }
812 
813 static inline int
ufputs(const unsigned char * s,FILE * f)814 ufputs (const unsigned char *s, FILE *f)
815 {
816   return fputs ((const char *)s, f);
817 }
818 
819   /* In line-map.c.  */
820 
821 /* Create a macro map.  A macro map encodes source locations of tokens
822    that are part of a macro replacement-list, at a macro expansion
823    point. See the extensive comments of struct line_map and struct
824    line_map_macro, in line-map.h.
825 
826    This map shall be created when the macro is expanded. The map
827    encodes the source location of the expansion point of the macro as
828    well as the "original" source location of each token that is part
829    of the macro replacement-list. If a macro is defined but never
830    expanded, it has no macro map.  SET is the set of maps the macro
831    map should be part of.  MACRO_NODE is the macro which the new macro
832    map should encode source locations for.  EXPANSION is the location
833    of the expansion point of MACRO. For function-like macros
834    invocations, it's best to make it point to the closing parenthesis
835    of the macro, rather than the the location of the first character
836    of the macro.  NUM_TOKENS is the number of tokens that are part of
837    the replacement-list of MACRO.  */
838 const line_map_macro *linemap_enter_macro (struct line_maps *,
839 					   struct cpp_hashnode*,
840 					   source_location,
841 					   unsigned int);
842 
843 /* Create and return a virtual location for a token that is part of a
844    macro expansion-list at a macro expansion point.  See the comment
845    inside struct line_map_macro to see what an expansion-list exactly
846    is.
847 
848    A call to this function must come after a call to
849    linemap_enter_macro.
850 
851    MAP is the map into which the source location is created.  TOKEN_NO
852    is the index of the token in the macro replacement-list, starting
853    at number 0.
854 
855    ORIG_LOC is the location of the token outside of this macro
856    expansion.  If the token comes originally from the macro
857    definition, it is the locus in the macro definition; otherwise it
858    is a location in the context of the caller of this macro expansion
859    (which is a virtual location or a source location if the caller is
860    itself a macro expansion or not).
861 
862    MACRO_DEFINITION_LOC is the location in the macro definition,
863    either of the token itself or of a macro parameter that it
864    replaces.  */
865 source_location linemap_add_macro_token (const line_map_macro *,
866 					 unsigned int,
867 					 source_location,
868 					 source_location);
869 
870 /* Return the source line number corresponding to source location
871    LOCATION.  SET is the line map set LOCATION comes from.  If
872    LOCATION is the location of token that is part of the
873    expansion-list of a macro expansion return the line number of the
874    macro expansion point.  */
875 int linemap_get_expansion_line (struct line_maps *,
876 				source_location);
877 
878 /* Return the path of the file corresponding to source code location
879    LOCATION.
880 
881    If LOCATION is the location of a token that is part of the
882    replacement-list of a macro expansion return the file path of the
883    macro expansion point.
884 
885    SET is the line map set LOCATION comes from.  */
886 const char* linemap_get_expansion_filename (struct line_maps *,
887 					    source_location);
888 
889 #ifdef __cplusplus
890 }
891 #endif
892 
893 #endif /* ! LIBCPP_INTERNAL_H */
894