libcpp/internal.h

   1 /* Part of CPP library.
   2    Copyright (C) 1997-2014 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 *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 in a __has_attribute__ statement.  */
 265   unsigned char in__has_attribute__;
 266
 267   /* Nonzero if prevent_expansion is true only because output is
 268      being discarded.  */
 269   unsigned char discarding_output;
 270
 271   /* Nonzero to skip evaluating part of an expression.  */
 272   unsigned int skip_eval;
 273
 274   /* Nonzero when handling a deferred pragma.  */
 275   unsigned char in_deferred_pragma;
 276
 277   /* Nonzero if the deferred pragma being handled allows macro expansion.  */
 278   unsigned char pragma_allow_expansion;
 279 };
 280
 281 /* Special nodes - identifiers with predefined significance.  */
 282 struct spec_nodes
 283 {
 284   cpp_hashnode *n_defined;              /* defined operator */
 285   cpp_hashnode *n_true;                 /* C++ keyword true */
 286   cpp_hashnode *n_false;                /* C++ keyword false */
 287   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 288   cpp_hashnode *n__has_include__;       /* __has_include__ operator */
 289   cpp_hashnode *n__has_include_next__;  /* __has_include_next__ operator */
 290   cpp_hashnode *n__has_attribute__;     /* __has_attribute__ operator */
 291 };
 292
 293 typedef struct _cpp_line_note _cpp_line_note;
 294 struct _cpp_line_note
 295 {
 296   /* Location in the clean line the note refers to.  */
 297   const unsigned char *pos;
 298
 299   /* Type of note.  The 9 'from' trigraph characters represent those
 300      trigraphs, '\\' an escaped newline, ' ' an escaped newline with
 301      intervening space, 0 represents a note that has already been handled,
 302      and anything else is invalid.  */
 303   unsigned int type;
 304 };
 305
 306 /* Represents the contents of a file cpplib has read in.  */
 307 struct cpp_buffer
 308 {
 309   const unsigned char *cur;        /* Current location.  */
 310   const unsigned char *line_base;  /* Start of current physical line.  */
 311   const unsigned char *next_line;  /* Start of to-be-cleaned logical line.  */
 312
 313   const unsigned char *buf;        /* Entire character buffer.  */
 314   const unsigned char *rlimit;     /* Writable byte at end of file.  */
 315   const unsigned char *to_free;    /* Pointer that should be freed when
 316                                       popping the buffer.  */
 317
 318   _cpp_line_note *notes;           /* Array of notes.  */
 319   unsigned int cur_note;           /* Next note to process.  */
 320   unsigned int notes_used;         /* Number of notes.  */
 321   unsigned int notes_cap;          /* Size of allocated array.  */
 322
 323   struct cpp_buffer *prev;
 324
 325   /* Pointer into the file table; non-NULL if this is a file buffer.
 326      Used for include_next and to record control macros.  */
 327   struct _cpp_file *file;
 328
 329   /* Saved value of __TIMESTAMP__ macro - date and time of last modification
 330      of the assotiated file.  */
 331   const unsigned char *timestamp;
 332
 333   /* Value of if_stack at start of this file.
 334      Used to prohibit unmatched #endif (etc) in an include file.  */
 335   struct if_stack *if_stack;
 336
 337   /* True if we need to get the next clean line.  */
 338   bool need_line;
 339
 340   /* True if we have already warned about C++ comments in this file.
 341      The warning happens only for C89 extended mode with -pedantic on,
 342      or for -Wtraditional, and only once per file (otherwise it would
 343      be far too noisy).  */
 344   unsigned int warned_cplusplus_comments : 1;
 345
 346   /* True if we don't process trigraphs and escaped newlines.  True
 347      for preprocessed input, command line directives, and _Pragma
 348      buffers.  */
 349   unsigned int from_stage3 : 1;
 350
 351   /* At EOF, a buffer is automatically popped.  If RETURN_AT_EOF is
 352      true, a CPP_EOF token is then returned.  Otherwise, the next
 353      token from the enclosing buffer is returned.  */
 354   unsigned int return_at_eof : 1;
 355
 356   /* One for a system header, two for a C system header file that therefore
 357      needs to be extern "C" protected in C++, and zero otherwise.  */
 358   unsigned char sysp;
 359
 360   /* The directory of the this buffer's file.  Its NAME member is not
 361      allocated, so we don't need to worry about freeing it.  */
 362   struct cpp_dir dir;
 363
 364   /* Descriptor for converting from the input character set to the
 365      source character set.  */
 366   struct cset_converter input_cset_desc;
 367 };
 368
 369 /* The list of saved macros by push_macro pragma.  */
 370 struct def_pragma_macro {
 371   /* Chain element to previous saved macro.  */
 372   struct def_pragma_macro *next;
 373   /* Name of the macro.  */
 374   char *name;
 375   /* The stored macro content.  */
 376   unsigned char *definition;
 377
 378   /* Definition line number.  */
 379   source_location line;
 380   /* If macro defined in system header.  */
 381   unsigned int syshdr   : 1;
 382   /* Nonzero if it has been expanded or had its existence tested.  */
 383   unsigned int used     : 1;
 384
 385   /* Mark if we save an undefined macro.  */
 386   unsigned int is_undef : 1;
 387 };
 388
 389 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 390    Applying cpp_get_token repeatedly yields a stream of pre-processor
 391    tokens.  Usually, there is only one cpp_reader object active.  */
 392 struct cpp_reader
 393 {
 394   /* Top of buffer stack.  */
 395   cpp_buffer *buffer;
 396
 397   /* Overlaid buffer (can be different after processing #include).  */
 398   cpp_buffer *overlaid_buffer;
 399
 400   /* Lexer state.  */
 401   struct lexer_state state;
 402
 403   /* Source line tracking.  */
 404   struct line_maps *line_table;
 405
 406   /* The line of the '#' of the current directive.  */
 407   source_location directive_line;
 408
 409   /* Memory buffers.  */
 410   _cpp_buff *a_buff;            /* Aligned permanent storage.  */
 411   _cpp_buff *u_buff;            /* Unaligned permanent storage.  */
 412   _cpp_buff *free_buffs;        /* Free buffer chain.  */
 413
 414   /* Context stack.  */
 415   struct cpp_context base_context;
 416   struct cpp_context *context;
 417
 418   /* If in_directive, the directive if known.  */
 419   const struct directive *directive;
 420
 421   /* Token generated while handling a directive, if any. */
 422   cpp_token directive_result;
 423
 424   /* When expanding a macro at top-level, this is the location of the
 425      macro invocation.  */
 426   source_location invocation_location;
 427
 428   /* Nonzero if we are about to expand a macro.  Note that if we are
 429      really expanding a macro, the function macro_of_context returns
 430      the macro being expanded and this flag is set to false.  Client
 431      code should use the function in_macro_expansion_p to know if we
 432      are either about to expand a macro, or are actually expanding
 433      one.  */
 434   bool about_to_expand_macro_p;
 435
 436   /* Search paths for include files.  */
 437   struct cpp_dir *quote_include;        /* "" */
 438   struct cpp_dir *bracket_include;      /* <> */
 439   struct cpp_dir no_search_path;        /* No path.  */
 440
 441   /* Chain of all hashed _cpp_file instances.  */
 442   struct _cpp_file *all_files;
 443
 444   struct _cpp_file *main_file;
 445
 446   /* File and directory hash table.  */
 447   struct htab *file_hash;
 448   struct htab *dir_hash;
 449   struct file_hash_entry_pool *file_hash_entries;
 450
 451   /* Negative path lookup hash table.  */
 452   struct htab *nonexistent_file_hash;
 453   struct obstack nonexistent_file_ob;
 454
 455   /* Nonzero means don't look for #include "foo" the source-file
 456      directory.  */
 457   bool quote_ignores_source_dir;
 458
 459   /* Nonzero if any file has contained #pragma once or #import has
 460      been used.  */
 461   bool seen_once_only;
 462
 463   /* Multiple include optimization.  */
 464   const cpp_hashnode *mi_cmacro;
 465   const cpp_hashnode *mi_ind_cmacro;
 466   bool mi_valid;
 467
 468   /* Lexing.  */
 469   cpp_token *cur_token;
 470   tokenrun base_run, *cur_run;
 471   unsigned int lookaheads;
 472
 473   /* Nonzero prevents the lexer from re-using the token runs.  */
 474   unsigned int keep_tokens;
 475
 476   /* Buffer to hold macro definition string.  */
 477   unsigned char *macro_buffer;
 478   unsigned int macro_buffer_len;
 479
 480   /* Descriptor for converting from the source character set to the
 481      execution character set.  */
 482   struct cset_converter narrow_cset_desc;
 483
 484   /* Descriptor for converting from the source character set to the
 485      UTF-8 execution character set.  */
 486   struct cset_converter utf8_cset_desc;
 487
 488   /* Descriptor for converting from the source character set to the
 489      UTF-16 execution character set.  */
 490   struct cset_converter char16_cset_desc;
 491
 492   /* Descriptor for converting from the source character set to the
 493      UTF-32 execution character set.  */
 494   struct cset_converter char32_cset_desc;
 495
 496   /* Descriptor for converting from the source character set to the
 497      wide execution character set.  */
 498   struct cset_converter wide_cset_desc;
 499
 500   /* Date and time text.  Calculated together if either is requested.  */
 501   const unsigned char *date;
 502   const unsigned char *time;
 503
 504   /* EOF token, and a token forcing paste avoidance.  */
 505   cpp_token avoid_paste;
 506   cpp_token eof;
 507
 508   /* Opaque handle to the dependencies of mkdeps.c.  */
 509   struct deps *deps;
 510
 511   /* Obstack holding all macro hash nodes.  This never shrinks.
 512      See identifiers.c */
 513   struct obstack hash_ob;
 514
 515   /* Obstack holding buffer and conditional structures.  This is a
 516      real stack.  See directives.c.  */
 517   struct obstack buffer_ob;
 518
 519   /* Pragma table - dynamic, because a library user can add to the
 520      list of recognized pragmas.  */
 521   struct pragma_entry *pragmas;
 522
 523   /* Call backs to cpplib client.  */
 524   struct cpp_callbacks cb;
 525
 526   /* Identifier hash table.  */
 527   struct ht *hash_table;
 528
 529   /* Expression parser stack.  */
 530   struct op *op_stack, *op_limit;
 531
 532   /* User visible options.  */
 533   struct cpp_options opts;
 534
 535   /* Special nodes - identifiers with predefined significance to the
 536      preprocessor.  */
 537   struct spec_nodes spec_nodes;
 538
 539   /* Whether cpplib owns the hashtable.  */
 540   bool our_hashtable;
 541
 542   /* Traditional preprocessing output buffer (a logical line).  */
 543   struct
 544   {
 545     unsigned char *base;
 546     unsigned char *limit;
 547     unsigned char *cur;
 548     source_location first_line;
 549   } out;
 550
 551   /* Used for buffer overlays by traditional.c.  */
 552   const unsigned char *saved_cur, *saved_rlimit, *saved_line_base;
 553
 554   /* A saved list of the defined macros, for dependency checking
 555      of precompiled headers.  */
 556   struct cpp_savedstate *savedstate;
 557
 558   /* Next value of __COUNTER__ macro. */
 559   unsigned int counter;
 560
 561   /* Table of comments, when state.save_comments is true.  */
 562   cpp_comment_table comments;
 563
 564   /* List of saved macros by push_macro.  */
 565   struct def_pragma_macro *pushed_macros;
 566
 567   /* If non-null, the lexer will use this location for the next token
 568      instead of getting a location from the linemap.  */
 569   source_location *forced_token_location_p;
 570 };
 571
 572 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 573    If the definition of `numchar' looks odd to you, please look up the
 574    definition of a pp-number in the C standard [section 6.4.8 of C99].
 575
 576    In the unlikely event that characters other than \r and \n enter
 577    the set is_vspace, the macro handle_newline() in lex.c must be
 578    updated.  */
 579 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 580
 581 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 582 #define is_numchar(x)   ISIDNUM(x)
 583 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 584 #define is_numstart(x)  ISDIGIT(x)
 585 #define is_hspace(x)    ISBLANK(x)
 586 #define is_vspace(x)    IS_VSPACE(x)
 587 #define is_nvspace(x)   IS_NVSPACE(x)
 588 #define is_space(x)     IS_SPACE_OR_NUL(x)
 589
 590 /* This table is constant if it can be initialized at compile time,
 591    which is the case if cpp was compiled with GCC >=2.7, or another
 592    compiler that supports C99.  */
 593 #if HAVE_DESIGNATED_INITIALIZERS
 594 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 595 #else
 596 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 597 #endif
 598
 599 /* Macros.  */
 600
 601 static inline int cpp_in_system_header (cpp_reader *);
 602 static inline int
 603 cpp_in_system_header (cpp_reader *pfile)
 604 {
 605   return pfile->buffer ? pfile->buffer->sysp : 0;
 606 }
 607 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, cpp_pedantic)
 608 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, cpp_warn_traditional)
 609
 610 static inline int cpp_in_primary_file (cpp_reader *);
 611 static inline int
 612 cpp_in_primary_file (cpp_reader *pfile)
 613 {
 614   return pfile->line_table->depth == 1;
 615 }
 616
 617 /* In macro.c */
 618 extern void _cpp_free_definition (cpp_hashnode *);
 619 extern bool _cpp_create_definition (cpp_reader *, cpp_hashnode *);
 620 extern void _cpp_pop_context (cpp_reader *);
 621 extern void _cpp_push_text_context (cpp_reader *, cpp_hashnode *,
 622                                     const unsigned char *, size_t);
 623 extern bool _cpp_save_parameter (cpp_reader *, cpp_macro *, cpp_hashnode *,
 624                                  cpp_hashnode *);
 625 extern bool _cpp_arguments_ok (cpp_reader *, cpp_macro *, const cpp_hashnode *,
 626                                unsigned int);
 627 extern const unsigned char *_cpp_builtin_macro_text (cpp_reader *,
 628                                                      cpp_hashnode *);
 629 extern int _cpp_warn_if_unused_macro (cpp_reader *, cpp_hashnode *, void *);
 630 extern void _cpp_push_token_context (cpp_reader *, cpp_hashnode *,
 631                                      const cpp_token *, unsigned int);
 632 extern void _cpp_backup_tokens_direct (cpp_reader *, unsigned int);
 633
 634 /* In identifiers.c */
 635 extern void _cpp_init_hashtable (cpp_reader *, cpp_hash_table *);
 636 extern void _cpp_destroy_hashtable (cpp_reader *);
 637
 638 /* In files.c */
 639 typedef struct _cpp_file _cpp_file;
 640 extern _cpp_file *_cpp_find_file (cpp_reader *, const char *, cpp_dir *,
 641                                   bool, int, bool);
 642 extern bool _cpp_find_failed (_cpp_file *);
 643 extern void _cpp_mark_file_once_only (cpp_reader *, struct _cpp_file *);
 644 extern void _cpp_fake_include (cpp_reader *, const char *);
 645 extern bool _cpp_stack_file (cpp_reader *, _cpp_file*, bool);
 646 extern bool _cpp_stack_include (cpp_reader *, const char *, int,
 647                                 enum include_type);
 648 extern int _cpp_compare_file_date (cpp_reader *, const char *, int);
 649 extern void _cpp_report_missing_guards (cpp_reader *);
 650 extern void _cpp_init_files (cpp_reader *);
 651 extern void _cpp_cleanup_files (cpp_reader *);
 652 extern void _cpp_pop_file_buffer (cpp_reader *, struct _cpp_file *,
 653                                   const unsigned char *);
 654 extern bool _cpp_save_file_entries (cpp_reader *pfile, FILE *f);
 655 extern bool _cpp_read_file_entries (cpp_reader *, FILE *);
 656 extern const char *_cpp_get_file_name (_cpp_file *);
 657 extern struct stat *_cpp_get_file_stat (_cpp_file *);
 658 extern bool _cpp_has_header (cpp_reader *, const char *, int,
 659                              enum include_type);
 660
 661 /* In expr.c */
 662 extern bool _cpp_parse_expr (cpp_reader *, bool);
 663 extern struct op *_cpp_expand_op_stack (cpp_reader *);
 664
 665 /* In lex.c */
 666 extern void _cpp_process_line_notes (cpp_reader *, int);
 667 extern void _cpp_clean_line (cpp_reader *);
 668 extern bool _cpp_get_fresh_line (cpp_reader *);
 669 extern bool _cpp_skip_block_comment (cpp_reader *);
 670 extern cpp_token *_cpp_temp_token (cpp_reader *);
 671 extern const cpp_token *_cpp_lex_token (cpp_reader *);
 672 extern cpp_token *_cpp_lex_direct (cpp_reader *);
 673 extern unsigned char *_cpp_spell_ident_ucns (unsigned char *, cpp_hashnode *);
 674 extern int _cpp_equiv_tokens (const cpp_token *, const cpp_token *);
 675 extern void _cpp_init_tokenrun (tokenrun *, unsigned int);
 676 extern cpp_hashnode *_cpp_lex_identifier (cpp_reader *, const char *);
 677 extern int _cpp_remaining_tokens_num_in_context (cpp_context *);
 678 extern void _cpp_init_lexer (void);
 679
 680 /* In init.c.  */
 681 extern void _cpp_maybe_push_include_file (cpp_reader *);
 682 extern const char *cpp_named_operator2name (enum cpp_ttype type);
 683
 684 /* In directives.c */
 685 extern int _cpp_test_assertion (cpp_reader *, unsigned int *);
 686 extern int _cpp_handle_directive (cpp_reader *, int);
 687 extern void _cpp_define_builtin (cpp_reader *, const char *);
 688 extern char ** _cpp_save_pragma_names (cpp_reader *);
 689 extern void _cpp_restore_pragma_names (cpp_reader *, char **);
 690 extern int _cpp_do__Pragma (cpp_reader *);
 691 extern void _cpp_init_directives (cpp_reader *);
 692 extern void _cpp_init_internal_pragmas (cpp_reader *);
 693 extern void _cpp_do_file_change (cpp_reader *, enum lc_reason, const char *,
 694                                  linenum_type, unsigned int);
 695 extern void _cpp_pop_buffer (cpp_reader *);
 696 extern char *_cpp_bracket_include (cpp_reader *);
 697
 698 /* In directives.c */
 699 struct _cpp_dir_only_callbacks
 700 {
 701   /* Called to print a block of lines. */
 702   void (*print_lines) (int, const void *, size_t);
 703   void (*maybe_print_line) (source_location);
 704 };
 705
 706 extern void _cpp_preprocess_dir_only (cpp_reader *,
 707                                       const struct _cpp_dir_only_callbacks *);
 708
 709 /* In traditional.c.  */
 710 extern bool _cpp_scan_out_logical_line (cpp_reader *, cpp_macro *);
 711 extern bool _cpp_read_logical_line_trad (cpp_reader *);
 712 extern void _cpp_overlay_buffer (cpp_reader *pfile, const unsigned char *,
 713                                  size_t);
 714 extern void _cpp_remove_overlay (cpp_reader *);
 715 extern bool _cpp_create_trad_definition (cpp_reader *, cpp_macro *);
 716 extern bool _cpp_expansions_different_trad (const cpp_macro *,
 717                                             const cpp_macro *);
 718 extern unsigned char *_cpp_copy_replacement_text (const cpp_macro *,
 719                                                   unsigned char *);
 720 extern size_t _cpp_replacement_text_len (const cpp_macro *);
 721
 722 /* In charset.c.  */
 723
 724 /* The normalization state at this point in the sequence.
 725    It starts initialized to all zeros, and at the end
 726    'level' is the normalization level of the sequence.  */
 727
 728 struct normalize_state
 729 {
 730   /* The previous starter character.  */
 731   cppchar_t previous;
 732   /* The combining class of the previous character (whether or not a
 733      starter).  */
 734   unsigned char prev_class;
 735   /* The lowest normalization level so far.  */
 736   enum cpp_normalize_level level;
 737 };
 738 #define INITIAL_NORMALIZE_STATE { 0, 0, normalized_KC }
 739 #define NORMALIZE_STATE_RESULT(st) ((st)->level)
 740
 741 /* We saw a character C that matches ISIDNUM(), update a
 742    normalize_state appropriately.  */
 743 #define NORMALIZE_STATE_UPDATE_IDNUM(st, c)     \
 744   ((st)->previous = (c), (st)->prev_class = 0)
 745
 746 extern cppchar_t _cpp_valid_ucn (cpp_reader *, const unsigned char **,
 747                                  const unsigned char *, int,
 748                                  struct normalize_state *state);
 749 extern void _cpp_destroy_iconv (cpp_reader *);
 750 extern unsigned char *_cpp_convert_input (cpp_reader *, const char *,
 751                                           unsigned char *, size_t, size_t,
 752                                           const unsigned char **, off_t *);
 753 extern const char *_cpp_default_encoding (void);
 754 extern cpp_hashnode * _cpp_interpret_identifier (cpp_reader *pfile,
 755                                                  const unsigned char *id,
 756                                                  size_t len);
 757
 758 /* Utility routines and macros.  */
 759 #define DSC(str) (const unsigned char *)str, sizeof str - 1
 760
 761 /* These are inline functions instead of macros so we can get type
 762    checking.  */
 763 static inline int ustrcmp (const unsigned char *, const unsigned char *);
 764 static inline int ustrncmp (const unsigned char *, const unsigned char *,
 765                             size_t);
 766 static inline size_t ustrlen (const unsigned char *);
 767 static inline const unsigned char *uxstrdup (const unsigned char *);
 768 static inline const unsigned char *ustrchr (const unsigned char *, int);
 769 static inline int ufputs (const unsigned char *, FILE *);
 770
 771 /* Use a const char for the second parameter since it is usually a literal.  */
 772 static inline int ustrcspn (const unsigned char *, const char *);
 773
 774 static inline int
 775 ustrcmp (const unsigned char *s1, const unsigned char *s2)
 776 {
 777   return strcmp ((const char *)s1, (const char *)s2);
 778 }
 779
 780 static inline int
 781 ustrncmp (const unsigned char *s1, const unsigned char *s2, size_t n)
 782 {
 783   return strncmp ((const char *)s1, (const char *)s2, n);
 784 }
 785
 786 static inline int
 787 ustrcspn (const unsigned char *s1, const char *s2)
 788 {
 789   return strcspn ((const char *)s1, s2);
 790 }
 791
 792 static inline size_t
 793 ustrlen (const unsigned char *s1)
 794 {
 795   return strlen ((const char *)s1);
 796 }
 797
 798 static inline const unsigned char *
 799 uxstrdup (const unsigned char *s1)
 800 {
 801   return (const unsigned char *) xstrdup ((const char *)s1);
 802 }
 803
 804 static inline const unsigned char *
 805 ustrchr (const unsigned char *s1, int c)
 806 {
 807   return (const unsigned char *) strchr ((const char *)s1, c);
 808 }
 809
 810 static inline int
 811 ufputs (const unsigned char *s, FILE *f)
 812 {
 813   return fputs ((const char *)s, f);
 814 }
 815
 816   /* In line-map.c.  */
 817
 818 /* Create a macro map.  A macro map encodes source locations of tokens
 819    that are part of a macro replacement-list, at a macro expansion
 820    point. See the extensive comments of struct line_map and struct
 821    line_map_macro, in line-map.h.
 822
 823    This map shall be created when the macro is expanded. The map
 824    encodes the source location of the expansion point of the macro as
 825    well as the "original" source location of each token that is part
 826    of the macro replacement-list. If a macro is defined but never
 827    expanded, it has no macro map.  SET is the set of maps the macro
 828    map should be part of.  MACRO_NODE is the macro which the new macro
 829    map should encode source locations for.  EXPANSION is the location
 830    of the expansion point of MACRO. For function-like macros
 831    invocations, it's best to make it point to the closing parenthesis
 832    of the macro, rather than the the location of the first character
 833    of the macro.  NUM_TOKENS is the number of tokens that are part of
 834    the replacement-list of MACRO.  */
 835 const struct line_map *linemap_enter_macro (struct line_maps *,
 836                                             struct cpp_hashnode*,
 837                                             source_location,
 838                                             unsigned int);
 839
 840 /* Create and return a virtual location for a token that is part of a
 841    macro expansion-list at a macro expansion point.  See the comment
 842    inside struct line_map_macro to see what an expansion-list exactly
 843    is.
 844
 845    A call to this function must come after a call to
 846    linemap_enter_macro.
 847
 848    MAP is the map into which the source location is created.  TOKEN_NO
 849    is the index of the token in the macro replacement-list, starting
 850    at number 0.
 851
 852    ORIG_LOC is the location of the token outside of this macro
 853    expansion.  If the token comes originally from the macro
 854    definition, it is the locus in the macro definition; otherwise it
 855    is a location in the context of the caller of this macro expansion
 856    (which is a virtual location or a source location if the caller is
 857    itself a macro expansion or not).
 858
 859    MACRO_DEFINITION_LOC is the location in the macro definition,
 860    either of the token itself or of a macro parameter that it
 861    replaces.  */
 862 source_location linemap_add_macro_token (const struct line_map *,
 863                                          unsigned int,
 864                                          source_location,
 865                                          source_location);
 866
 867 /* Return the source line number corresponding to source location
 868    LOCATION.  SET is the line map set LOCATION comes from.  If
 869    LOCATION is the location of token that is part of the
 870    expansion-list of a macro expansion return the line number of the
 871    macro expansion point.  */
 872 int linemap_get_expansion_line (struct line_maps *,
 873                                 source_location);
 874
 875 /* Return the path of the file corresponding to source code location
 876    LOCATION.
 877
 878    If LOCATION is the location of a token that is part of the
 879    replacement-list of a macro expansion return the file path of the
 880    macro expansion point.
 881
 882    SET is the line map set LOCATION comes from.  */
 883 const char* linemap_get_expansion_filename (struct line_maps *,
 884                                             source_location);
 885
 886 #ifdef __cplusplus
 887 }
 888 #endif
 889
 890 #endif /* ! LIBCPP_INTERNAL_H */