gcc/cpphash.h

   1 /* Part of CPP library.
   2    Copyright (C) 1997, 1998, 1999, 2000, 2001 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 2, 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; if not, write to the Free Software
  16 Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  17
  18 /* This header defines all the internal data structures and functions
  19    that need to be visible across files.  It's called cpphash.h for
  20    historical reasons.  */
  21
  22 #ifndef __GCC_CPPHASH__
  23 #define __GCC_CPPHASH__
  24
  25 struct directive;               /* These are deliberately incomplete.  */
  26 struct htab;
  27
  28 /* Test if a sign is valid within a preprocessing number.  */
  29 #define VALID_SIGN(c, prevc) \
  30   (((c) == '+' || (c) == '-') && \
  31    ((prevc) == 'e' || (prevc) == 'E' \
  32     || (((prevc) == 'p' || (prevc) == 'P') \
  33         && CPP_OPTION (pfile, extended_numbers))))
  34
  35 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
  36 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
  37 #define CPP_BUF_LINE(BUF) ((BUF)->lineno)
  38 #define CPP_BUF_COLUMN(BUF, CUR) ((CUR) - (BUF)->line_base + (BUF)->col_adjust)
  39 #define CPP_BUF_COL(BUF) CPP_BUF_COLUMN(BUF, (BUF)->cur)
  40
  41 /* Maximum nesting of cpp_buffers.  We use a static limit, partly for
  42    efficiency, and partly to limit runaway recursion.  */
  43 #define CPP_STACK_MAX 200
  44
  45 /* Memory pools.  */
  46 #define POOL_ALIGN(size, align) (((size) + ((align) - 1)) & ~((align) - 1))
  47 #define POOL_FRONT(p) ((p)->cur->front)
  48 #define POOL_LIMIT(p) ((p)->cur->limit)
  49 #define POOL_BASE(p)  ((p)->cur->base)
  50 #define POOL_SIZE(p)  ((p)->cur->limit - (p)->cur->base)
  51 #define POOL_ROOM(p)  ((p)->cur->limit - (p)->cur->front)
  52 #define POOL_USED(p)  ((p)->cur->front - (p)->cur->base)
  53 #define POOL_COMMIT(p, len) do {\
  54   ((p)->cur->front += POOL_ALIGN (len, (p)->align));\
  55   if ((p)->cur->front > (p)->cur->limit) abort ();} while (0)
  56
  57 typedef struct cpp_chunk cpp_chunk;
  58 struct cpp_chunk
  59 {
  60   cpp_chunk *next;
  61   unsigned char *front;
  62   unsigned char *limit;
  63   unsigned char *base;
  64 };
  65
  66 typedef struct cpp_pool cpp_pool;
  67 struct cpp_pool
  68 {
  69   struct cpp_chunk *cur, *locked;
  70   unsigned char *pos;           /* Current position.  */
  71   unsigned int align;
  72   unsigned int locks;
  73 };
  74
  75 /* List of directories to look for include files in.  */
  76 struct search_path
  77 {
  78   struct search_path *next;
  79
  80   /* NOTE: NAME may not be null terminated for the case of the current
  81      file's directory!  */
  82   const char *name;
  83   unsigned int len;
  84   /* We use these to tell if the directory mentioned here is a duplicate
  85      of an earlier directory on the search path.  */
  86   ino_t ino;
  87   dev_t dev;
  88   /* Non-zero if it is a system include directory.  */
  89   int sysp;
  90   /* Mapping of file names for this directory.  Only used on MS-DOS
  91      and related platforms.  */
  92   struct file_name_map *name_map;
  93 };
  94
  95 /* Multiple-include optimisation.  */
  96 enum mi_state {MI_FAILED = 0, MI_OUTSIDE};
  97 enum mi_ind {MI_IND_NONE = 0, MI_IND_NOT};
  98
  99 typedef struct toklist toklist;
 100 struct toklist
 101 {
 102   cpp_token *first;
 103   cpp_token *limit;
 104 };
 105
 106 typedef struct cpp_context cpp_context;
 107 struct cpp_context
 108 {
 109   /* Doubly-linked list.  */
 110   cpp_context *next, *prev;
 111
 112   /* Contexts other than the base context are contiguous tokens.
 113      e.g. macro expansions, expanded argument tokens.  */
 114   struct toklist list;
 115
 116   /* For a macro context, these are the macro and its arguments.  */
 117   cpp_macro *macro;
 118 };
 119
 120 struct lexer_state
 121 {
 122   /* Nonzero if first token on line is CPP_HASH.  */
 123   unsigned char in_directive;
 124
 125   /* Nonzero if in a directive that takes angle-bracketed headers.  */
 126   unsigned char angled_headers;
 127
 128   /* Nonzero to save comments.  Turned off if discard_comments, and in
 129      all directives apart from #define.  */
 130   unsigned char save_comments;
 131
 132   /* If nonzero the next token is at the beginning of the line.  */
 133   unsigned char next_bol;
 134
 135   /* Nonzero if we're mid-comment.  */
 136   unsigned char lexing_comment;
 137
 138   /* Nonzero if lexing __VA_ARGS__ is valid.  */
 139   unsigned char va_args_ok;
 140
 141   /* Nonzero if lexing poisoned identifiers is valid.  */
 142   unsigned char poisoned_ok;
 143
 144   /* Nonzero to prevent macro expansion.  */
 145   unsigned char prevent_expansion;
 146
 147   /* Nonzero when parsing arguments to a function-like macro.  */
 148   unsigned char parsing_args;
 149
 150   /* Nonzero when in a # NUMBER directive.  */
 151   unsigned char line_extension;
 152 };
 153
 154 /* Special nodes - identifiers with predefined significance.  */
 155 struct spec_nodes
 156 {
 157   cpp_hashnode *n_L;                    /* L"str" */
 158   cpp_hashnode *n_defined;              /* defined operator */
 159   cpp_hashnode *n_true;                 /* C++ keyword true */
 160   cpp_hashnode *n_false;                /* C++ keyword false */
 161   cpp_hashnode *n__Pragma;              /* _Pragma operator */
 162   cpp_hashnode *n__STRICT_ANSI__;       /* STDC_0_IN_SYSTEM_HEADERS */
 163   cpp_hashnode *n__CHAR_UNSIGNED__;     /* plain char is unsigned */
 164   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 165 };
 166
 167 struct cpp_buffer
 168 {
 169   const unsigned char *cur;      /* current position */
 170   const unsigned char *rlimit; /* end of valid data */
 171   const unsigned char *line_base; /* start of current line */
 172   cppchar_t read_ahead;         /* read ahead character */
 173   cppchar_t extra_char;         /* extra read-ahead for long tokens.  */
 174
 175   struct cpp_reader *pfile;     /* Owns this buffer.  */
 176   struct cpp_buffer *prev;
 177
 178   const unsigned char *buf;      /* entire buffer */
 179
 180   /* Filename specified with #line command.  */
 181   const char *nominal_fname;
 182
 183   /* Pointer into the include table.  Used for include_next and
 184      to record control macros. */
 185   struct include_file *inc;
 186
 187   /* Value of if_stack at start of this file.
 188      Used to prohibit unmatched #endif (etc) in an include file.  */
 189   struct if_stack *if_stack;
 190
 191   /* Token column position adjustment owing to tabs in whitespace.  */
 192   unsigned int col_adjust;
 193
 194   /* Line number at line_base (above). */
 195   unsigned int lineno;
 196
 197   /* Contains PREV_WHITE and/or AVOID_LPASTE.  */
 198   unsigned char saved_flags;
 199
 200   /* Because of the way the lexer works, -Wtrigraphs can sometimes
 201      warn twice for the same trigraph.  This helps prevent that.  */
 202   const unsigned char *last_Wtrigraphs;
 203
 204   /* True if we have already warned about C++ comments in this file.
 205      The warning happens only for C89 extended mode with -pedantic on,
 206      or for -Wtraditional, and only once per file (otherwise it would
 207      be far too noisy).  */
 208   unsigned char warned_cplusplus_comments;
 209
 210   /* True if we don't process trigraphs and escaped newlines.  True
 211      for preprocessed input, command line directives, and _Pragma
 212      buffers.  */
 213   unsigned char from_stage3;
 214
 215   /* Temporary storage for pfile->skipping whilst in a directive.  */
 216   unsigned char was_skipping;
 217
 218   /* 1 = system header file, 2 = C system header file used for C++.  */
 219   unsigned char sysp;
 220
 221   /* Nonzero means we have printed (while error reporting) a list of
 222      containing files that matches the current status.  */
 223   unsigned char include_stack_listed;
 224
 225   /* Buffer type.  */
 226   ENUM_BITFIELD (cpp_buffer_type) type : 8;
 227
 228   /* The directory of the this buffer's file.  Its NAME member is not
 229      allocated, so we don't need to worry about freeing it.  */
 230   struct search_path dir;
 231
 232   /* The directory to start searching for "" include files.  Is either
 233      "dir" above, or options.quote_include, depending upon whether -I-
 234      was on the command line.  */
 235   struct search_path *search_from;
 236 };
 237
 238 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 239    Applying cpp_get_token repeatedly yields a stream of pre-processor
 240    tokens.  Usually, there is only one cpp_reader object active.  */
 241
 242 struct cpp_reader
 243 {
 244   /* Top of buffer stack.  */
 245   cpp_buffer *buffer;
 246
 247   /* Lexer state.  */
 248   struct lexer_state state;
 249
 250   /* The position of the last lexed token and last lexed directive.  */
 251   cpp_lexer_pos lexer_pos;
 252   cpp_lexer_pos directive_pos;
 253
 254   /* Memory pools.  */
 255   cpp_pool ident_pool;          /* For all identifiers, and permanent
 256                                    numbers and strings.  */
 257   cpp_pool macro_pool;          /* For macro definitions.  Permanent.  */
 258   cpp_pool argument_pool;       /* For macro arguments.  Temporary.   */
 259
 260   /* Context stack.  */
 261   struct cpp_context base_context;
 262   struct cpp_context *context;
 263
 264   /* If in_directive, the directive if known.  */
 265   const struct directive *directive;
 266
 267   /* Multiple inlcude optimisation.  */
 268   enum mi_state mi_state;
 269   enum mi_ind mi_if_not_defined;
 270   unsigned int mi_lexed;
 271   const cpp_hashnode *mi_cmacro;
 272   const cpp_hashnode *mi_ind_cmacro;
 273
 274   /* Token lookahead.  */
 275   struct cpp_lookahead *la_read;        /* Read from this lookahead.  */
 276   struct cpp_lookahead *la_write;       /* Write to this lookahead.  */
 277   struct cpp_lookahead *la_unused;      /* Free store.  */
 278   struct cpp_lookahead *la_saved;       /* Backup when entering directive.  */
 279
 280   /* Error counter for exit code.  */
 281   unsigned int errors;
 282
 283   /* Line and column where a newline was first seen in a string
 284      constant (multi-line strings).  */
 285   cpp_lexer_pos mlstring_pos;
 286
 287   /* Buffer to hold macro definition string.  */
 288   unsigned char *macro_buffer;
 289   unsigned int macro_buffer_len;
 290
 291   /* Current depth in #include directives that use <...>.  */
 292   unsigned int system_include_depth;
 293
 294   /* Current depth of buffer stack.  */
 295   unsigned int buffer_stack_depth;
 296
 297   /* Current depth in #include directives.  */
 298   unsigned int include_depth;
 299
 300   /* Hash table of macros and assertions.  See cpphash.c.  */
 301   struct htab *hashtab;
 302
 303   /* Tree of other included files.  See cppfiles.c.  */
 304   struct splay_tree_s *all_include_files;
 305
 306   /* Current maximum length of directory names in the search path
 307      for include files.  (Altered as we get more of them.)  */
 308   unsigned int max_include_len;
 309
 310   /* Date and time tokens.  Calculated together if either is requested.  */
 311   cpp_token date;
 312   cpp_token time;
 313
 314   /* Opaque handle to the dependencies of mkdeps.c.  Used by -M etc.  */
 315   struct deps *deps;
 316
 317   /* Obstack holding all macro hash nodes.  This never shrinks.
 318      See cpphash.c */
 319   struct obstack *hash_ob;
 320
 321   /* Obstack holding buffer and conditional structures.  This is a
 322      real stack.  See cpplib.c */
 323   struct obstack *buffer_ob;
 324
 325   /* Pragma table - dynamic, because a library user can add to the
 326      list of recognized pragmas.  */
 327   struct pragma_entry *pragmas;
 328
 329   /* Call backs.  */
 330   struct cpp_callbacks cb;
 331
 332   /* User visible options.  */
 333   struct cpp_options opts;
 334
 335   /* Special nodes - identifiers with predefined significance to the
 336      preprocessor.  */
 337   struct spec_nodes spec_nodes;
 338
 339   /* We're printed a warning recommending against using #import.  */
 340   unsigned char import_warning;
 341
 342   /* True if we are skipping a failed conditional group.  */
 343   unsigned char skipping;
 344
 345   /* Whether to print our version number.  Done this way so
 346      we don't get it twice for -v -version.  */
 347   unsigned char print_version;
 348 };
 349
 350 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 351    If the definition of `numchar' looks odd to you, please look up the
 352    definition of a pp-number in the C standard [section 6.4.8 of C99].
 353
 354    In the unlikely event that characters other than \r and \n enter
 355    the set is_vspace, the macro handle_newline() in cpplex.c must be
 356    updated.  */
 357 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 358
 359 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 360 #define is_numchar(x)   ISIDNUM(x)
 361 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 362 #define is_numstart(x)  ISDIGIT(x)
 363 #define is_hspace(x)    ISBLANK(x)
 364 #define is_vspace(x)    IS_VSPACE(x)
 365 #define is_nvspace(x)   IS_NVSPACE(x)
 366 #define is_space(x)     IS_SPACE_OR_NUL(x)
 367
 368 /* This table is constant if it can be initialized at compile time,
 369    which is the case if cpp was compiled with GCC >=2.7, or another
 370    compiler that supports C99.  */
 371 #if HAVE_DESIGNATED_INITIALIZERS
 372 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 373 #else
 374 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 375 #endif
 376
 377 /* Macros.  */
 378
 379 #define CPP_PRINT_DEPS(PFILE) CPP_OPTION (PFILE, print_deps)
 380 #define CPP_IN_SYSTEM_HEADER(PFILE) \
 381   (CPP_BUFFER (PFILE) && CPP_BUFFER (PFILE)->sysp)
 382 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
 383 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
 384
 385 /* Hash step.  The hash calculation is duplicated in cpp_lookup and
 386    parse_name.  */
 387 #define HASHSTEP(r, c) ((r) * 67 + (c - 113));
 388
 389 /* In cpperror.c  */
 390 enum error_type { WARNING = 0, PEDWARN, ERROR, FATAL, ICE };
 391 extern int _cpp_begin_message PARAMS ((cpp_reader *, enum error_type,
 392                                        const char *, const cpp_lexer_pos *));
 393
 394 /* In cppmacro.c */
 395 extern void _cpp_free_definition        PARAMS ((cpp_hashnode *));
 396 extern int _cpp_create_definition       PARAMS ((cpp_reader *, cpp_hashnode *));
 397 extern void _cpp_pop_context            PARAMS ((cpp_reader *));
 398 extern void _cpp_free_lookaheads        PARAMS ((cpp_reader *));
 399 extern void _cpp_release_lookahead      PARAMS ((cpp_reader *));
 400 extern void _cpp_push_token             PARAMS ((cpp_reader *, const cpp_token *,
 401                                                  const cpp_lexer_pos *));
 402
 403 /* In cpphash.c */
 404 extern void _cpp_init_hashtable         PARAMS ((cpp_reader *));
 405 extern void _cpp_cleanup_hashtable      PARAMS ((cpp_reader *));
 406 extern cpp_hashnode *_cpp_lookup_with_hash PARAMS ((cpp_reader*, size_t,
 407                                                     unsigned int));
 408
 409 /* In cppfiles.c */
 410 extern void _cpp_fake_include           PARAMS ((cpp_reader *, const char *));
 411 extern void _cpp_never_reread           PARAMS ((struct include_file *));
 412 extern void _cpp_simplify_pathname      PARAMS ((char *));
 413 extern int _cpp_read_file               PARAMS ((cpp_reader *, const char *));
 414 extern void _cpp_execute_include        PARAMS ((cpp_reader *,
 415                                                  const cpp_token *, int, int));
 416 extern int _cpp_compare_file_date       PARAMS ((cpp_reader *,
 417                                                  const cpp_token *));
 418 extern void _cpp_report_missing_guards  PARAMS ((cpp_reader *));
 419 extern void _cpp_init_includes          PARAMS ((cpp_reader *));
 420 extern void _cpp_cleanup_includes       PARAMS ((cpp_reader *));
 421 extern void _cpp_pop_file_buffer        PARAMS ((cpp_reader *, cpp_buffer *));
 422
 423 /* In cppexp.c */
 424 extern int _cpp_parse_expr              PARAMS ((cpp_reader *));
 425
 426 /* In cpplex.c */
 427 extern void _cpp_lex_token              PARAMS ((cpp_reader *, cpp_token *));
 428 extern int _cpp_equiv_tokens            PARAMS ((const cpp_token *,
 429                                                  const cpp_token *));
 430 extern void _cpp_init_pool              PARAMS ((cpp_pool *, unsigned int,
 431                                                   unsigned int, unsigned int));
 432 extern void _cpp_free_pool              PARAMS ((cpp_pool *));
 433 extern unsigned char *_cpp_pool_reserve PARAMS ((cpp_pool *, unsigned int));
 434 extern unsigned char *_cpp_pool_alloc   PARAMS ((cpp_pool *, unsigned int));
 435 extern unsigned char *_cpp_next_chunk   PARAMS ((cpp_pool *, unsigned int,
 436                                                  unsigned char **));
 437 extern void _cpp_lock_pool              PARAMS ((cpp_pool *));
 438 extern void _cpp_unlock_pool            PARAMS ((cpp_pool *));
 439
 440 /* In cpplib.c */
 441 extern int _cpp_test_assertion PARAMS ((cpp_reader *, int *));
 442 extern int _cpp_handle_directive PARAMS ((cpp_reader *, int));
 443 extern void _cpp_define_builtin PARAMS ((cpp_reader *, const char *));
 444 extern void _cpp_do__Pragma     PARAMS ((cpp_reader *));
 445 extern void _cpp_init_stacks    PARAMS ((cpp_reader *));
 446 extern void _cpp_cleanup_stacks PARAMS ((cpp_reader *));
 447 extern void _cpp_init_internal_pragmas PARAMS ((cpp_reader *));
 448 extern void _cpp_do_file_change PARAMS ((cpp_reader *, enum cpp_fc_reason,
 449                                          const char *, unsigned int));
 450
 451 /* Utility routines and macros.  */
 452 #define DSC(str) (const U_CHAR *)str, sizeof str - 1
 453 #define xnew(T)         (T *) xmalloc (sizeof(T))
 454 #define xcnew(T)        (T *) xcalloc (1, sizeof(T))
 455 #define xnewvec(T, N)   (T *) xmalloc (sizeof(T) * (N))
 456 #define xcnewvec(T, N)  (T *) xcalloc (N, sizeof(T))
 457 #define xobnew(O, T)    (T *) obstack_alloc (O, sizeof(T))
 458
 459 /* These are inline functions instead of macros so we can get type
 460    checking.  */
 461 typedef unsigned char U_CHAR;
 462 #define U (const U_CHAR *)  /* Intended use: U"string" */
 463
 464 static inline int ustrcmp       PARAMS ((const U_CHAR *, const U_CHAR *));
 465 static inline int ustrncmp      PARAMS ((const U_CHAR *, const U_CHAR *,
 466                                          size_t));
 467 static inline size_t ustrlen    PARAMS ((const U_CHAR *));
 468 static inline U_CHAR *uxstrdup  PARAMS ((const U_CHAR *));
 469 static inline U_CHAR *ustrchr   PARAMS ((const U_CHAR *, int));
 470 static inline int ufputs        PARAMS ((const U_CHAR *, FILE *));
 471
 472 static inline int
 473 ustrcmp (s1, s2)
 474      const U_CHAR *s1, *s2;
 475 {
 476   return strcmp ((const char *)s1, (const char *)s2);
 477 }
 478
 479 static inline int
 480 ustrncmp (s1, s2, n)
 481      const U_CHAR *s1, *s2;
 482      size_t n;
 483 {
 484   return strncmp ((const char *)s1, (const char *)s2, n);
 485 }
 486
 487 static inline size_t
 488 ustrlen (s1)
 489      const U_CHAR *s1;
 490 {
 491   return strlen ((const char *)s1);
 492 }
 493
 494 static inline U_CHAR *
 495 uxstrdup (s1)
 496      const U_CHAR *s1;
 497 {
 498   return (U_CHAR *) xstrdup ((const char *)s1);
 499 }
 500
 501 static inline U_CHAR *
 502 ustrchr (s1, c)
 503      const U_CHAR *s1;
 504      int c;
 505 {
 506   return (U_CHAR *) strchr ((const char *)s1, c);
 507 }
 508
 509 static inline int
 510 ufputs (s, f)
 511      const U_CHAR *s;
 512      FILE *f;
 513 {
 514   return fputs ((const char *)s, f);
 515 }
 516
 517 #endif