libcpp/internal.h

   1 /* Part of CPP library.
   2    Copyright (C) 1997-2021 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 "cpplib.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   const char* from;
  52   const char* to;
  53 };
  54
  55 #define BITS_PER_CPPCHAR_T (CHAR_BIT * sizeof (cppchar_t))
  56
  57 /* Test if a sign is valid within a preprocessing number.  */
  58 #define VALID_SIGN(c, prevc) \
  59   (((c) == '+' || (c) == '-') && \
  60    ((prevc) == 'e' || (prevc) == 'E' \
  61     || (((prevc) == 'p' || (prevc) == 'P') \
  62         && CPP_OPTION (pfile, extended_numbers))))
  63
  64 #define DIGIT_SEP(c) ((c) == '\'' && CPP_OPTION (pfile, digit_separators))
  65
  66 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
  67 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
  68 #define CPP_BUF_COLUMN(BUF, CUR) ((CUR) - (BUF)->line_base)
  69 #define CPP_BUF_COL(BUF) CPP_BUF_COLUMN(BUF, (BUF)->cur)
  70
  71 #define CPP_INCREMENT_LINE(PFILE, COLS_HINT) do { \
  72     const class line_maps *line_table = PFILE->line_table; \
  73     const struct line_map_ordinary *map = \
  74       LINEMAPS_LAST_ORDINARY_MAP (line_table); \
  75     linenum_type line = SOURCE_LINE (map, line_table->highest_line); \
  76     linemap_line_start (PFILE->line_table, line + 1, COLS_HINT); \
  77   } while (0)
  78
  79 /* Host alignment handling.  */
  80 struct dummy
  81 {
  82   char c;
  83   union
  84   {
  85     double d;
  86     int *p;
  87   } u;
  88 };
  89
  90 #define DEFAULT_ALIGNMENT offsetof (struct dummy, u)
  91 #define CPP_ALIGN2(size, align) (((size) + ((align) - 1)) & ~((align) - 1))
  92 #define CPP_ALIGN(size) CPP_ALIGN2 (size, DEFAULT_ALIGNMENT)
  93
  94 #define _cpp_mark_macro_used(NODE)                                      \
  95   (cpp_user_macro_p (NODE) ? (NODE)->value.macro->used = 1 : 0)
  96
  97 /* A generic memory buffer, and operations on it.  */
  98 typedef struct _cpp_buff _cpp_buff;
  99 struct _cpp_buff
 100 {
 101   struct _cpp_buff *next;
 102   unsigned char *base, *cur, *limit;
 103 };
 104
 105 extern _cpp_buff *_cpp_get_buff (cpp_reader *, size_t);
 106 extern void _cpp_release_buff (cpp_reader *, _cpp_buff *);
 107 extern void _cpp_extend_buff (cpp_reader *, _cpp_buff **, size_t);
 108 extern _cpp_buff *_cpp_append_extend_buff (cpp_reader *, _cpp_buff *, size_t);
 109 extern void _cpp_free_buff (_cpp_buff *);
 110 extern unsigned char *_cpp_aligned_alloc (cpp_reader *, size_t);
 111 extern unsigned char *_cpp_unaligned_alloc (cpp_reader *, size_t);
 112
 113 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
 114 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
 115 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
 116
 117 /* #include types.  */
 118 enum include_type
 119   {
 120    /* Directive-based including mechanisms.  */
 121    IT_INCLUDE,  /* #include */
 122    IT_INCLUDE_NEXT,  /* #include_next */
 123    IT_IMPORT,   /* #import  */
 124
 125    /* Non-directive including mechanisms.  */
 126    IT_CMDLINE,  /* -include */
 127    IT_DEFAULT,  /* forced header  */
 128    IT_MAIN,     /* main, start on line 1 */
 129    IT_PRE_MAIN,  /* main, but there will be a preamble before line
 130                     1 */
 131
 132    IT_DIRECTIVE_HWM = IT_IMPORT + 1,  /* Directives below this.  */
 133    IT_HEADER_HWM = IT_DEFAULT + 1     /* Header files below this.  */
 134   };
 135
 136 union utoken
 137 {
 138   const cpp_token *token;
 139   const cpp_token **ptoken;
 140 };
 141
 142 /* A "run" of tokens; part of a chain of runs.  */
 143 typedef struct tokenrun tokenrun;
 144 struct tokenrun
 145 {
 146   tokenrun *next, *prev;
 147   cpp_token *base, *limit;
 148 };
 149
 150 /* Accessor macros for struct cpp_context.  */
 151 #define FIRST(c) ((c)->u.iso.first)
 152 #define LAST(c) ((c)->u.iso.last)
 153 #define CUR(c) ((c)->u.trad.cur)
 154 #define RLIMIT(c) ((c)->u.trad.rlimit)
 155
 156 /* This describes some additional data that is added to the macro
 157    token context of type cpp_context, when -ftrack-macro-expansion is
 158    on.  */
 159 typedef struct
 160 {
 161   /* The node of the macro we are referring to.  */
 162   cpp_hashnode *macro_node;
 163   /* This buffer contains an array of virtual locations.  The virtual
 164      location at index 0 is the virtual location of the token at index
 165      0 in the current instance of cpp_context; similarly for all the
 166      other virtual locations.  */
 167   location_t *virt_locs;
 168   /* This is a pointer to the current virtual location.  This is used
 169      to iterate over the virtual locations while we iterate over the
 170      tokens they belong to.  */
 171   location_t *cur_virt_loc;
 172 } macro_context;
 173
 174 /* The kind of tokens carried by a cpp_context.  */
 175 enum context_tokens_kind {
 176   /* This is the value of cpp_context::tokens_kind if u.iso.first
 177      contains an instance of cpp_token **.  */
 178   TOKENS_KIND_INDIRECT,
 179   /* This is the value of cpp_context::tokens_kind if u.iso.first
 180      contains an instance of cpp_token *.  */
 181   TOKENS_KIND_DIRECT,
 182   /* This is the value of cpp_context::tokens_kind when the token
 183      context contains tokens resulting from macro expansion.  In that
 184      case struct cpp_context::macro points to an instance of struct
 185      macro_context.  This is used only when the
 186      -ftrack-macro-expansion flag is on.  */
 187   TOKENS_KIND_EXTENDED
 188 };
 189
 190 typedef struct cpp_context cpp_context;
 191 struct cpp_context
 192 {
 193   /* Doubly-linked list.  */
 194   cpp_context *next, *prev;
 195
 196   union
 197   {
 198     /* For ISO macro expansion.  Contexts other than the base context
 199        are contiguous tokens.  e.g. macro expansions, expanded
 200        argument tokens.  */
 201     struct
 202     {
 203       union utoken first;
 204       union utoken last;
 205     } iso;
 206
 207     /* For traditional macro expansion.  */
 208     struct
 209     {
 210       const unsigned char *cur;
 211       const unsigned char *rlimit;
 212     } trad;
 213   } u;
 214
 215   /* If non-NULL, a buffer used for storage related to this context.
 216      When the context is popped, the buffer is released.  */
 217   _cpp_buff *buff;
 218
 219   /* If tokens_kind is TOKEN_KIND_EXTENDED, then (as we thus are in a
 220      macro context) this is a pointer to an instance of macro_context.
 221      Otherwise if tokens_kind is *not* TOKEN_KIND_EXTENDED, then, if
 222      we are in a macro context, this is a pointer to an instance of
 223      cpp_hashnode, representing the name of the macro this context is
 224      for.  If we are not in a macro context, then this is just NULL.
 225      Note that when tokens_kind is TOKEN_KIND_EXTENDED, the memory
 226      used by the instance of macro_context pointed to by this member
 227      is de-allocated upon de-allocation of the instance of struct
 228      cpp_context.  */
 229   union
 230   {
 231     macro_context *mc;
 232     cpp_hashnode *macro;
 233   } c;
 234
 235   /* This determines the type of tokens held by this context.  */
 236   enum context_tokens_kind tokens_kind;
 237 };
 238
 239 struct lexer_state
 240 {
 241   /* 1 if we're handling a directive.  2 if it's an include-like
 242      directive.  */
 243   unsigned char in_directive;
 244
 245   /* Nonzero if in a directive that will handle padding tokens itself.
 246      #include needs this to avoid problems with computed include and
 247      spacing between tokens.  */
 248   unsigned char directive_wants_padding;
 249
 250   /* True if we are skipping a failed conditional group.  */
 251   unsigned char skipping;
 252
 253   /* Nonzero if in a directive that takes angle-bracketed headers.  */
 254   unsigned char angled_headers;
 255
 256   /* Nonzero if in a #if or #elif directive.  */
 257   unsigned char in_expression;
 258
 259   /* Nonzero to save comments.  Turned off if discard_comments, and in
 260      all directives apart from #define.  */
 261   unsigned char save_comments;
 262
 263   /* Nonzero if lexing __VA_ARGS__ and __VA_OPT__ are valid.  */
 264   unsigned char va_args_ok;
 265
 266   /* Nonzero if lexing poisoned identifiers is valid.  */
 267   unsigned char poisoned_ok;
 268
 269   /* Nonzero to prevent macro expansion.  */
 270   unsigned char prevent_expansion;
 271
 272   /* Nonzero when parsing arguments to a function-like macro.  */
 273   unsigned char parsing_args;
 274
 275   /* Nonzero if prevent_expansion is true only because output is
 276      being discarded.  */
 277   unsigned char discarding_output;
 278
 279   /* Nonzero to skip evaluating part of an expression.  */
 280   unsigned int skip_eval;
 281
 282   /* Nonzero when tokenizing a deferred pragma.  */
 283   unsigned char in_deferred_pragma;
 284
 285   /* Count to token that is a header-name.  */
 286   unsigned char directive_file_token;
 287
 288   /* Nonzero if the deferred pragma being handled allows macro expansion.  */
 289   unsigned char pragma_allow_expansion;
 290 };
 291
 292 /* Special nodes - identifiers with predefined significance.  */
 293 struct spec_nodes
 294 {
 295   cpp_hashnode *n_defined;              /* defined operator */
 296   cpp_hashnode *n_true;                 /* C++ keyword true */
 297   cpp_hashnode *n_false;                /* C++ keyword false */
 298   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 299   cpp_hashnode *n__VA_OPT__;            /* C++ vararg macros */
 300
 301   enum {M_EXPORT, M_MODULE, M_IMPORT, M__IMPORT, M_HWM};
 302
 303   /* C++20 modules, only set when module_directives is in effect.
 304      incoming variants [0], outgoing ones [1] */
 305   cpp_hashnode *n_modules[M_HWM][2];
 306 };
 307
 308 typedef struct _cpp_line_note _cpp_line_note;
 309 struct _cpp_line_note
 310 {
 311   /* Location in the clean line the note refers to.  */
 312   const unsigned char *pos;
 313
 314   /* Type of note.  The 9 'from' trigraph characters represent those
 315      trigraphs, '\\' an escaped newline, ' ' an escaped newline with
 316      intervening space, 0 represents a note that has already been handled,
 317      and anything else is invalid.  */
 318   unsigned int type;
 319 };
 320
 321 /* Represents the contents of a file cpplib has read in.  */
 322 struct cpp_buffer
 323 {
 324   const unsigned char *cur;        /* Current location.  */
 325   const unsigned char *line_base;  /* Start of current physical line.  */
 326   const unsigned char *next_line;  /* Start of to-be-cleaned logical line.  */
 327
 328   const unsigned char *buf;        /* Entire character buffer.  */
 329   const unsigned char *rlimit;     /* Writable byte at end of file.  */
 330   const unsigned char *to_free;    /* Pointer that should be freed when
 331                                       popping the buffer.  */
 332
 333   _cpp_line_note *notes;           /* Array of notes.  */
 334   unsigned int cur_note;           /* Next note to process.  */
 335   unsigned int notes_used;         /* Number of notes.  */
 336   unsigned int notes_cap;          /* Size of allocated array.  */
 337
 338   struct cpp_buffer *prev;
 339
 340   /* Pointer into the file table; non-NULL if this is a file buffer.
 341      Used for include_next and to record control macros.  */
 342   struct _cpp_file *file;
 343
 344   /* Saved value of __TIMESTAMP__ macro - date and time of last modification
 345      of the assotiated file.  */
 346   const unsigned char *timestamp;
 347
 348   /* Value of if_stack at start of this file.
 349      Used to prohibit unmatched #endif (etc) in an include file.  */
 350   struct if_stack *if_stack;
 351
 352   /* True if we need to get the next clean line.  */
 353   bool need_line : 1;
 354
 355   /* True if we have already warned about C++ comments in this file.
 356      The warning happens only for C89 extended mode with -pedantic on,
 357      or for -Wtraditional, and only once per file (otherwise it would
 358      be far too noisy).  */
 359   bool warned_cplusplus_comments : 1;
 360
 361   /* True if we don't process trigraphs and escaped newlines.  True
 362      for preprocessed input, command line directives, and _Pragma
 363      buffers.  */
 364   bool from_stage3 : 1;
 365
 366   /* At EOF, a buffer is automatically popped.  If RETURN_AT_EOF is
 367      true, a CPP_EOF token is then returned.  Otherwise, the next
 368      token from the enclosing buffer is returned.  */
 369   bool return_at_eof : 1;
 370
 371   /* One for a system header, two for a C system header file that therefore
 372      needs to be extern "C" protected in C++, and zero otherwise.  */
 373   unsigned char sysp;
 374
 375   /* The directory of the this buffer's file.  Its NAME member is not
 376      allocated, so we don't need to worry about freeing it.  */
 377   struct cpp_dir dir;
 378
 379   /* Descriptor for converting from the input character set to the
 380      source character set.  */
 381   struct cset_converter input_cset_desc;
 382 };
 383
 384 /* The list of saved macros by push_macro pragma.  */
 385 struct def_pragma_macro {
 386   /* Chain element to previous saved macro.  */
 387   struct def_pragma_macro *next;
 388   /* Name of the macro.  */
 389   char *name;
 390   /* The stored macro content.  */
 391   unsigned char *definition;
 392
 393   /* Definition line number.  */
 394   location_t line;
 395   /* If macro defined in system header.  */
 396   unsigned int syshdr   : 1;
 397   /* Nonzero if it has been expanded or had its existence tested.  */
 398   unsigned int used     : 1;
 399
 400   /* Mark if we save an undefined macro.  */
 401   unsigned int is_undef : 1;
 402   /* Nonzero if it was a builtin macro.  */
 403   unsigned int is_builtin : 1;
 404 };
 405
 406 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 407    Applying cpp_get_token repeatedly yields a stream of pre-processor
 408    tokens.  Usually, there is only one cpp_reader object active.  */
 409 struct cpp_reader
 410 {
 411   /* Top of buffer stack.  */
 412   cpp_buffer *buffer;
 413
 414   /* Overlaid buffer (can be different after processing #include).  */
 415   cpp_buffer *overlaid_buffer;
 416
 417   /* Lexer state.  */
 418   struct lexer_state state;
 419
 420   /* Source line tracking.  */
 421   class line_maps *line_table;
 422
 423   /* The line of the '#' of the current directive.  */
 424   location_t directive_line;
 425
 426   /* Memory buffers.  */
 427   _cpp_buff *a_buff;            /* Aligned permanent storage.  */
 428   _cpp_buff *u_buff;            /* Unaligned permanent storage.  */
 429   _cpp_buff *free_buffs;        /* Free buffer chain.  */
 430
 431   /* Context stack.  */
 432   struct cpp_context base_context;
 433   struct cpp_context *context;
 434
 435   /* If in_directive, the directive if known.  */
 436   const struct directive *directive;
 437
 438   /* Token generated while handling a directive, if any. */
 439   cpp_token directive_result;
 440
 441   /* When expanding a macro at top-level, this is the location of the
 442      macro invocation.  */
 443   location_t invocation_location;
 444
 445   /* This is the node representing the macro being expanded at
 446      top-level.  The value of this data member is valid iff
 447      cpp_in_macro_expansion_p() returns TRUE.  */
 448   cpp_hashnode *top_most_macro_node;
 449
 450   /* Nonzero if we are about to expand a macro.  Note that if we are
 451      really expanding a macro, the function macro_of_context returns
 452      the macro being expanded and this flag is set to false.  Client
 453      code should use the function cpp_in_macro_expansion_p to know if we
 454      are either about to expand a macro, or are actually expanding
 455      one.  */
 456   bool about_to_expand_macro_p;
 457
 458   /* Search paths for include files.  */
 459   struct cpp_dir *quote_include;        /* "" */
 460   struct cpp_dir *bracket_include;      /* <> */
 461   struct cpp_dir no_search_path;        /* No path.  */
 462
 463   /* Chain of all hashed _cpp_file instances.  */
 464   struct _cpp_file *all_files;
 465
 466   struct _cpp_file *main_file;
 467
 468   /* File and directory hash table.  */
 469   struct htab *file_hash;
 470   struct htab *dir_hash;
 471   struct file_hash_entry_pool *file_hash_entries;
 472
 473   /* Negative path lookup hash table.  */
 474   struct htab *nonexistent_file_hash;
 475   struct obstack nonexistent_file_ob;
 476
 477   /* Nonzero means don't look for #include "foo" the source-file
 478      directory.  */
 479   bool quote_ignores_source_dir;
 480
 481   /* Nonzero if any file has contained #pragma once or #import has
 482      been used.  */
 483   bool seen_once_only;
 484
 485   /* Multiple include optimization.  */
 486   const cpp_hashnode *mi_cmacro;
 487   const cpp_hashnode *mi_ind_cmacro;
 488   bool mi_valid;
 489
 490   /* Lexing.  */
 491   cpp_token *cur_token;
 492   tokenrun base_run, *cur_run;
 493   unsigned int lookaheads;
 494
 495   /* Nonzero prevents the lexer from re-using the token runs.  */
 496   unsigned int keep_tokens;
 497
 498   /* Buffer to hold macro definition string.  */
 499   unsigned char *macro_buffer;
 500   unsigned int macro_buffer_len;
 501
 502   /* Descriptor for converting from the source character set to the
 503      execution character set.  */
 504   struct cset_converter narrow_cset_desc;
 505
 506   /* Descriptor for converting from the source character set to the
 507      UTF-8 execution character set.  */
 508   struct cset_converter utf8_cset_desc;
 509
 510   /* Descriptor for converting from the source character set to the
 511      UTF-16 execution character set.  */
 512   struct cset_converter char16_cset_desc;
 513
 514   /* Descriptor for converting from the source character set to the
 515      UTF-32 execution character set.  */
 516   struct cset_converter char32_cset_desc;
 517
 518   /* Descriptor for converting from the source character set to the
 519      wide execution character set.  */
 520   struct cset_converter wide_cset_desc;
 521
 522   /* Date and time text.  Calculated together if either is requested.  */
 523   const unsigned char *date;
 524   const unsigned char *time;
 525
 526   /* Time stamp, set idempotently lazily.  */
 527   time_t time_stamp;
 528   int time_stamp_kind; /* Or errno.  */
 529
 530   /* A token forcing paste avoidance, and one demarking macro arguments.  */
 531   cpp_token avoid_paste;
 532   cpp_token endarg;
 533
 534   /* Opaque handle to the dependencies of mkdeps.c.  */
 535   class mkdeps *deps;
 536
 537   /* Obstack holding all macro hash nodes.  This never shrinks.
 538      See identifiers.c */
 539   struct obstack hash_ob;
 540
 541   /* Obstack holding buffer and conditional structures.  This is a
 542      real stack.  See directives.c.  */
 543   struct obstack buffer_ob;
 544
 545   /* Pragma table - dynamic, because a library user can add to the
 546      list of recognized pragmas.  */
 547   struct pragma_entry *pragmas;
 548
 549   /* Call backs to cpplib client.  */
 550   struct cpp_callbacks cb;
 551
 552   /* Identifier hash table.  */
 553   struct ht *hash_table;
 554
 555   /* Expression parser stack.  */
 556   struct op *op_stack, *op_limit;
 557
 558   /* User visible options.  */
 559   struct cpp_options opts;
 560
 561   /* Special nodes - identifiers with predefined significance to the
 562      preprocessor.  */
 563   struct spec_nodes spec_nodes;
 564
 565   /* Whether cpplib owns the hashtable.  */
 566   bool our_hashtable;
 567
 568   /* Traditional preprocessing output buffer (a logical line).  */
 569   struct
 570   {
 571     unsigned char *base;
 572     unsigned char *limit;
 573     unsigned char *cur;
 574     location_t first_line;
 575   } out;
 576
 577   /* Used for buffer overlays by traditional.c.  */
 578   const unsigned char *saved_cur, *saved_rlimit, *saved_line_base;
 579
 580   /* A saved list of the defined macros, for dependency checking
 581      of precompiled headers.  */
 582   struct cpp_savedstate *savedstate;
 583
 584   /* Next value of __COUNTER__ macro. */
 585   unsigned int counter;
 586
 587   /* Table of comments, when state.save_comments is true.  */
 588   cpp_comment_table comments;
 589
 590   /* List of saved macros by push_macro.  */
 591   struct def_pragma_macro *pushed_macros;
 592
 593   /* If non-zero, the lexer will use this location for the next token
 594      instead of getting a location from the linemap.  */
 595   location_t forced_token_location;
 596
 597   /* Location identifying the main source file -- intended to be line
 598      zero of said file.  */
 599   location_t main_loc;
 600
 601   /* Returns true iff we should warn about UTF-8 bidirectional control
 602      characters.  */
 603   bool warn_bidi_p () const
 604   {
 605     return CPP_OPTION (this, cpp_warn_bidirectional) != bidirectional_none;
 606   }
 607 };
 608
 609 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 610    If the definition of `numchar' looks odd to you, please look up the
 611    definition of a pp-number in the C standard [section 6.4.8 of C99].
 612
 613    In the unlikely event that characters other than \r and \n enter
 614    the set is_vspace, the macro handle_newline() in lex.c must be
 615    updated.  */
 616 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 617
 618 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 619 #define is_numchar(x)   ISIDNUM(x)
 620 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 621 #define is_numstart(x)  ISDIGIT(x)
 622 #define is_hspace(x)    ISBLANK(x)
 623 #define is_vspace(x)    IS_VSPACE(x)
 624 #define is_nvspace(x)   IS_NVSPACE(x)
 625 #define is_space(x)     IS_SPACE_OR_NUL(x)
 626
 627 #define SEEN_EOL() (pfile->cur_token[-1].type == CPP_EOF)
 628
 629 /* This table is constant if it can be initialized at compile time,
 630    which is the case if cpp was compiled with GCC >=2.7, or another
 631    compiler that supports C99.  */
 632 #if HAVE_DESIGNATED_INITIALIZERS
 633 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 634 #else
 635 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 636 #endif
 637
 638 #if !defined (HAVE_UCHAR) && !defined (IN_GCC)
 639 typedef unsigned char uchar;
 640 #endif
 641
 642 #define UC (const uchar *)  /* Intended use: UC"string" */
 643
 644 /* Accessors.  */
 645
 646 inline int
 647 _cpp_in_system_header (cpp_reader *pfile)
 648 {
 649   return pfile->buffer ? pfile->buffer->sysp : 0;
 650 }
 651 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, cpp_pedantic)
 652 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, cpp_warn_traditional)
 653
 654 /* Return true if we're in the main file (unless it's considered to be
 655    an include file in its own right.  */
 656 inline int
 657 _cpp_in_main_source_file (cpp_reader *pfile)
 658 {
 659   return (!CPP_OPTION (pfile, main_search)
 660           && pfile->buffer->file == pfile->main_file);
 661 }
 662
 663 /* True if NODE is a macro for the purposes of ifdef, defined etc.  */
 664 inline bool _cpp_defined_macro_p (cpp_hashnode *node)
 665 {
 666   /* Do not treat conditional macros as being defined.  This is due to
 667      the powerpc port using conditional macros for 'vector', 'bool',
 668      and 'pixel' to act as conditional keywords.  This messes up tests
 669      like #ifndef bool.  */
 670   return cpp_macro_p (node) && !(node->flags & NODE_CONDITIONAL);
 671 }
 672
 673 /* In macro.c */
 674 extern bool _cpp_notify_macro_use (cpp_reader *pfile, cpp_hashnode *node,
 675                                    location_t);
 676 inline bool _cpp_maybe_notify_macro_use (cpp_reader *pfile, cpp_hashnode *node,
 677                                          location_t loc)
 678 {
 679   if (!(node->flags & NODE_USED))
 680     return _cpp_notify_macro_use (pfile, node, loc);
 681   return true;
 682 }
 683 extern cpp_macro *_cpp_new_macro (cpp_reader *, cpp_macro_kind, void *);
 684 extern void _cpp_free_definition (cpp_hashnode *);
 685 extern bool _cpp_create_definition (cpp_reader *, cpp_hashnode *);
 686 extern void _cpp_pop_context (cpp_reader *);
 687 extern void _cpp_push_text_context (cpp_reader *, cpp_hashnode *,
 688                                     const unsigned char *, size_t);
 689 extern bool _cpp_save_parameter (cpp_reader *, unsigned, cpp_hashnode *,
 690                                  cpp_hashnode *);
 691 extern void _cpp_unsave_parameters (cpp_reader *, unsigned);
 692 extern bool _cpp_arguments_ok (cpp_reader *, cpp_macro *, const cpp_hashnode *,
 693                                unsigned int);
 694 extern const unsigned char *_cpp_builtin_macro_text (cpp_reader *,
 695                                                      cpp_hashnode *,
 696                                                      location_t = 0);
 697 extern int _cpp_warn_if_unused_macro (cpp_reader *, cpp_hashnode *, void *);
 698 extern void _cpp_push_token_context (cpp_reader *, cpp_hashnode *,
 699                                      const cpp_token *, unsigned int);
 700 extern void _cpp_backup_tokens_direct (cpp_reader *, unsigned int);
 701
 702 /* In identifiers.c */
 703 extern void _cpp_init_hashtable (cpp_reader *, cpp_hash_table *);
 704 extern void _cpp_destroy_hashtable (cpp_reader *);
 705
 706 /* In files.c */
 707 enum _cpp_find_file_kind
 708   { _cpp_FFK_NORMAL, _cpp_FFK_FAKE, _cpp_FFK_PRE_INCLUDE, _cpp_FFK_HAS_INCLUDE };
 709 extern _cpp_file *_cpp_find_file (cpp_reader *, const char *, cpp_dir *,
 710                                   int angle, _cpp_find_file_kind, location_t);
 711 extern bool _cpp_find_failed (_cpp_file *);
 712 extern void _cpp_mark_file_once_only (cpp_reader *, struct _cpp_file *);
 713 extern const char *_cpp_find_header_unit (cpp_reader *, const char *file,
 714                                           bool angle_p,  location_t);
 715 extern void _cpp_fake_include (cpp_reader *, const char *);
 716 extern bool _cpp_stack_file (cpp_reader *, _cpp_file*, include_type, location_t);
 717 extern bool _cpp_stack_include (cpp_reader *, const char *, int,
 718                                 enum include_type, location_t);
 719 extern int _cpp_compare_file_date (cpp_reader *, const char *, int);
 720 extern void _cpp_report_missing_guards (cpp_reader *);
 721 extern void _cpp_init_files (cpp_reader *);
 722 extern void _cpp_cleanup_files (cpp_reader *);
 723 extern void _cpp_pop_file_buffer (cpp_reader *, struct _cpp_file *,
 724                                   const unsigned char *);
 725 extern bool _cpp_save_file_entries (cpp_reader *pfile, FILE *f);
 726 extern bool _cpp_read_file_entries (cpp_reader *, FILE *);
 727 extern const char *_cpp_get_file_name (_cpp_file *);
 728 extern struct stat *_cpp_get_file_stat (_cpp_file *);
 729 extern bool _cpp_has_header (cpp_reader *, const char *, int,
 730                              enum include_type);
 731
 732 /* In expr.c */
 733 extern bool _cpp_parse_expr (cpp_reader *, bool);
 734 extern struct op *_cpp_expand_op_stack (cpp_reader *);
 735
 736 /* In lex.c */
 737 extern void _cpp_process_line_notes (cpp_reader *, int);
 738 extern void _cpp_clean_line (cpp_reader *);
 739 extern bool _cpp_get_fresh_line (cpp_reader *);
 740 extern bool _cpp_skip_block_comment (cpp_reader *);
 741 extern cpp_token *_cpp_temp_token (cpp_reader *);
 742 extern const cpp_token *_cpp_lex_token (cpp_reader *);
 743 extern cpp_token *_cpp_lex_direct (cpp_reader *);
 744 extern unsigned char *_cpp_spell_ident_ucns (unsigned char *, cpp_hashnode *);
 745 extern int _cpp_equiv_tokens (const cpp_token *, const cpp_token *);
 746 extern void _cpp_init_tokenrun (tokenrun *, unsigned int);
 747 extern cpp_hashnode *_cpp_lex_identifier (cpp_reader *, const char *);
 748 extern int _cpp_remaining_tokens_num_in_context (cpp_context *);
 749 extern void _cpp_init_lexer (void);
 750 static inline void *_cpp_reserve_room (cpp_reader *pfile, size_t have,
 751                                        size_t extra)
 752 {
 753   if (BUFF_ROOM (pfile->a_buff) < (have + extra))
 754     _cpp_extend_buff (pfile, &pfile->a_buff, extra);
 755   return BUFF_FRONT (pfile->a_buff);
 756 }
 757 extern void *_cpp_commit_buff (cpp_reader *pfile, size_t size);
 758
 759 /* In init.c.  */
 760 extern void _cpp_maybe_push_include_file (cpp_reader *);
 761 extern const char *cpp_named_operator2name (enum cpp_ttype type);
 762 extern void _cpp_restore_special_builtin (cpp_reader *pfile,
 763                                           struct def_pragma_macro *);
 764
 765 /* In directives.c */
 766 extern int _cpp_test_assertion (cpp_reader *, unsigned int *);
 767 extern int _cpp_handle_directive (cpp_reader *, bool);
 768 extern void _cpp_define_builtin (cpp_reader *, const char *);
 769 extern char ** _cpp_save_pragma_names (cpp_reader *);
 770 extern void _cpp_restore_pragma_names (cpp_reader *, char **);
 771 extern int _cpp_do__Pragma (cpp_reader *, location_t);
 772 extern void _cpp_init_directives (cpp_reader *);
 773 extern void _cpp_init_internal_pragmas (cpp_reader *);
 774 extern void _cpp_do_file_change (cpp_reader *, enum lc_reason, const char *,
 775                                  linenum_type, unsigned int);
 776 extern void _cpp_pop_buffer (cpp_reader *);
 777 extern char *_cpp_bracket_include (cpp_reader *);
 778
 779 /* In errors.c  */
 780 extern location_t cpp_diagnostic_get_current_location (cpp_reader *);
 781
 782 /* In traditional.c.  */
 783 extern bool _cpp_scan_out_logical_line (cpp_reader *, cpp_macro *, bool);
 784 extern bool _cpp_read_logical_line_trad (cpp_reader *);
 785 extern void _cpp_overlay_buffer (cpp_reader *pfile, const unsigned char *,
 786                                  size_t);
 787 extern void _cpp_remove_overlay (cpp_reader *);
 788 extern cpp_macro *_cpp_create_trad_definition (cpp_reader *);
 789 extern bool _cpp_expansions_different_trad (const cpp_macro *,
 790                                             const cpp_macro *);
 791 extern unsigned char *_cpp_copy_replacement_text (const cpp_macro *,
 792                                                   unsigned char *);
 793 extern size_t _cpp_replacement_text_len (const cpp_macro *);
 794
 795 /* In charset.c.  */
 796
 797 /* The normalization state at this point in the sequence.
 798    It starts initialized to all zeros, and at the end
 799    'level' is the normalization level of the sequence.  */
 800
 801 struct normalize_state
 802 {
 803   /* The previous starter character.  */
 804   cppchar_t previous;
 805   /* The combining class of the previous character (whether or not a
 806      starter).  */
 807   unsigned char prev_class;
 808   /* The lowest normalization level so far.  */
 809   enum cpp_normalize_level level;
 810 };
 811 #define INITIAL_NORMALIZE_STATE { 0, 0, normalized_KC }
 812 #define NORMALIZE_STATE_RESULT(st) ((st)->level)
 813
 814 /* We saw a character C that matches ISIDNUM(), update a
 815    normalize_state appropriately.  */
 816 #define NORMALIZE_STATE_UPDATE_IDNUM(st, c)     \
 817   ((st)->previous = (c), (st)->prev_class = 0)
 818
 819 extern bool _cpp_valid_ucn (cpp_reader *, const unsigned char **,
 820                             const unsigned char *, int,
 821                             struct normalize_state *state,
 822                             cppchar_t *,
 823                             source_range *char_range,
 824                             cpp_string_location_reader *loc_reader);
 825
 826 extern bool _cpp_valid_utf8 (cpp_reader *pfile,
 827                              const uchar **pstr,
 828                              const uchar *limit,
 829                              int identifier_pos,
 830                              struct normalize_state *nst,
 831                              cppchar_t *cp);
 832
 833 extern void _cpp_destroy_iconv (cpp_reader *);
 834 extern unsigned char *_cpp_convert_input (cpp_reader *, const char *,
 835                                           unsigned char *, size_t, size_t,
 836                                           const unsigned char **, off_t *);
 837 extern const char *_cpp_default_encoding (void);
 838 extern cpp_hashnode * _cpp_interpret_identifier (cpp_reader *pfile,
 839                                                  const unsigned char *id,
 840                                                  size_t len);
 841
 842 /* Utility routines and macros.  */
 843 #define DSC(str) (const unsigned char *)str, sizeof str - 1
 844
 845 /* These are inline functions instead of macros so we can get type
 846    checking.  */
 847 static inline int ustrcmp (const unsigned char *, const unsigned char *);
 848 static inline int ustrncmp (const unsigned char *, const unsigned char *,
 849                             size_t);
 850 static inline size_t ustrlen (const unsigned char *);
 851 static inline const unsigned char *uxstrdup (const unsigned char *);
 852 static inline const unsigned char *ustrchr (const unsigned char *, int);
 853 static inline int ufputs (const unsigned char *, FILE *);
 854
 855 /* Use a const char for the second parameter since it is usually a literal.  */
 856 static inline int ustrcspn (const unsigned char *, const char *);
 857
 858 static inline int
 859 ustrcmp (const unsigned char *s1, const unsigned char *s2)
 860 {
 861   return strcmp ((const char *)s1, (const char *)s2);
 862 }
 863
 864 static inline int
 865 ustrncmp (const unsigned char *s1, const unsigned char *s2, size_t n)
 866 {
 867   return strncmp ((const char *)s1, (const char *)s2, n);
 868 }
 869
 870 static inline int
 871 ustrcspn (const unsigned char *s1, const char *s2)
 872 {
 873   return strcspn ((const char *)s1, s2);
 874 }
 875
 876 static inline size_t
 877 ustrlen (const unsigned char *s1)
 878 {
 879   return strlen ((const char *)s1);
 880 }
 881
 882 static inline const unsigned char *
 883 uxstrdup (const unsigned char *s1)
 884 {
 885   return (const unsigned char *) xstrdup ((const char *)s1);
 886 }
 887
 888 static inline const unsigned char *
 889 ustrchr (const unsigned char *s1, int c)
 890 {
 891   return (const unsigned char *) strchr ((const char *)s1, c);
 892 }
 893
 894 static inline int
 895 ufputs (const unsigned char *s, FILE *f)
 896 {
 897   return fputs ((const char *)s, f);
 898 }
 899
 900 /* In line-map.c.  */
 901
 902 /* Create and return a virtual location for a token that is part of a
 903    macro expansion-list at a macro expansion point.  See the comment
 904    inside struct line_map_macro to see what an expansion-list exactly
 905    is.
 906
 907    A call to this function must come after a call to
 908    linemap_enter_macro.
 909
 910    MAP is the map into which the source location is created.  TOKEN_NO
 911    is the index of the token in the macro replacement-list, starting
 912    at number 0.
 913
 914    ORIG_LOC is the location of the token outside of this macro
 915    expansion.  If the token comes originally from the macro
 916    definition, it is the locus in the macro definition; otherwise it
 917    is a location in the context of the caller of this macro expansion
 918    (which is a virtual location or a source location if the caller is
 919    itself a macro expansion or not).
 920
 921    MACRO_DEFINITION_LOC is the location in the macro definition,
 922    either of the token itself or of a macro parameter that it
 923    replaces.  */
 924 location_t linemap_add_macro_token (const line_map_macro *,
 925                                     unsigned int,
 926                                     location_t,
 927                                     location_t);
 928
 929 /* Return the source line number corresponding to source location
 930    LOCATION.  SET is the line map set LOCATION comes from.  If
 931    LOCATION is the location of token that is part of the
 932    expansion-list of a macro expansion return the line number of the
 933    macro expansion point.  */
 934 int linemap_get_expansion_line (class line_maps *,
 935                                 location_t);
 936
 937 /* Return the path of the file corresponding to source code location
 938    LOCATION.
 939
 940    If LOCATION is the location of a token that is part of the
 941    replacement-list of a macro expansion return the file path of the
 942    macro expansion point.
 943
 944    SET is the line map set LOCATION comes from.  */
 945 const char* linemap_get_expansion_filename (class line_maps *,
 946                                             location_t);
 947
 948 /* A subclass of rich_location for emitting a diagnostic
 949    at the current location of the reader, but flagging
 950    it with set_escape_on_output (true).  */
 951 class encoding_rich_location : public rich_location
 952 {
 953  public:
 954   encoding_rich_location (cpp_reader *pfile)
 955   : rich_location (pfile->line_table,
 956                    cpp_diagnostic_get_current_location (pfile))
 957   {
 958     set_escape_on_output (true);
 959   }
 960
 961   encoding_rich_location (cpp_reader *pfile, location_t loc)
 962   : rich_location (pfile->line_table, loc)
 963   {
 964     set_escape_on_output (true);
 965   }
 966 };
 967
 968 #ifdef __cplusplus
 969 }
 970 #endif
 971
 972 #endif /* ! LIBCPP_INTERNAL_H */