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 /* #include types.  */
 100 enum include_type {IT_INCLUDE, IT_INCLUDE_NEXT, IT_IMPORT, IT_CMDLINE};
 101
 102 typedef struct toklist toklist;
 103 struct toklist
 104 {
 105   cpp_token *first;
 106   cpp_token *limit;
 107 };
 108
 109 typedef struct cpp_context cpp_context;
 110 struct cpp_context
 111 {
 112   /* Doubly-linked list.  */
 113   cpp_context *next, *prev;
 114
 115   /* Contexts other than the base context are contiguous tokens.
 116      e.g. macro expansions, expanded argument tokens.  */
 117   struct toklist list;
 118
 119   /* For a macro context, these are the macro and its arguments.  */
 120   cpp_macro *macro;
 121 };
 122
 123 struct lexer_state
 124 {
 125   /* Nonzero if first token on line is CPP_HASH.  */
 126   unsigned char in_directive;
 127
 128   /* Nonzero if in a directive that takes angle-bracketed headers.  */
 129   unsigned char angled_headers;
 130
 131   /* Nonzero to save comments.  Turned off if discard_comments, and in
 132      all directives apart from #define.  */
 133   unsigned char save_comments;
 134
 135   /* If nonzero the next token is at the beginning of the line.  */
 136   unsigned char next_bol;
 137
 138   /* Nonzero if we're mid-comment.  */
 139   unsigned char lexing_comment;
 140
 141   /* Nonzero if lexing __VA_ARGS__ is valid.  */
 142   unsigned char va_args_ok;
 143
 144   /* Nonzero if lexing poisoned identifiers is valid.  */
 145   unsigned char poisoned_ok;
 146
 147   /* Nonzero to prevent macro expansion.  */
 148   unsigned char prevent_expansion;
 149
 150   /* Nonzero when parsing arguments to a function-like macro.  */
 151   unsigned char parsing_args;
 152
 153   /* Nonzero when in a # NUMBER directive.  */
 154   unsigned char line_extension;
 155 };
 156
 157 /* Special nodes - identifiers with predefined significance.  */
 158 struct spec_nodes
 159 {
 160   cpp_hashnode *n_L;                    /* L"str" */
 161   cpp_hashnode *n_defined;              /* defined operator */
 162   cpp_hashnode *n_true;                 /* C++ keyword true */
 163   cpp_hashnode *n_false;                /* C++ keyword false */
 164   cpp_hashnode *n__Pragma;              /* _Pragma operator */
 165   cpp_hashnode *n__STRICT_ANSI__;       /* STDC_0_IN_SYSTEM_HEADERS */
 166   cpp_hashnode *n__CHAR_UNSIGNED__;     /* plain char is unsigned */
 167   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 168 };
 169
 170 struct cpp_buffer
 171 {
 172   const unsigned char *cur;      /* current position */
 173   const unsigned char *rlimit; /* end of valid data */
 174   const unsigned char *line_base; /* start of current line */
 175   cppchar_t read_ahead;         /* read ahead character */
 176   cppchar_t extra_char;         /* extra read-ahead for long tokens.  */
 177
 178   struct cpp_reader *pfile;     /* Owns this buffer.  */
 179   struct cpp_buffer *prev;
 180
 181   const unsigned char *buf;      /* entire buffer */
 182
 183   /* Filename specified with #line command.  */
 184   const char *nominal_fname;
 185
 186   /* Pointer into the include table.  Used for include_next and
 187      to record control macros. */
 188   struct include_file *inc;
 189
 190   /* Value of if_stack at start of this file.
 191      Used to prohibit unmatched #endif (etc) in an include file.  */
 192   struct if_stack *if_stack;
 193
 194   /* Token column position adjustment owing to tabs in whitespace.  */
 195   unsigned int col_adjust;
 196
 197   /* Line number at line_base (above). */
 198   unsigned int lineno;
 199
 200   /* Contains PREV_WHITE and/or AVOID_LPASTE.  */
 201   unsigned char saved_flags;
 202
 203   /* Because of the way the lexer works, -Wtrigraphs can sometimes
 204      warn twice for the same trigraph.  This helps prevent that.  */
 205   const unsigned char *last_Wtrigraphs;
 206
 207   /* True if we have already warned about C++ comments in this file.
 208      The warning happens only for C89 extended mode with -pedantic on,
 209      or for -Wtraditional, and only once per file (otherwise it would
 210      be far too noisy).  */
 211   unsigned char warned_cplusplus_comments;
 212
 213   /* True if we don't process trigraphs and escaped newlines.  True
 214      for preprocessed input, command line directives, and _Pragma
 215      buffers.  */
 216   unsigned char from_stage3;
 217
 218   /* Temporary storage for pfile->skipping whilst in a directive.  */
 219   unsigned char was_skipping;
 220
 221   /* 1 = system header file, 2 = C system header file used for C++.  */
 222   unsigned char sysp;
 223
 224   /* Nonzero means we have printed (while error reporting) a list of
 225      containing files that matches the current status.  */
 226   unsigned char include_stack_listed;
 227
 228   /* Nonzero means that the directory to start searching for ""
 229      include files has been calculated and stored in "dir" below.  */
 230   unsigned char search_cached;
 231
 232   /* Buffer type.  */
 233   ENUM_BITFIELD (cpp_buffer_type) type : 8;
 234
 235   /* The directory of the this buffer's file.  Its NAME member is not
 236      allocated, so we don't need to worry about freeing it.  */
 237   struct search_path dir;
 238 };
 239
 240 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 241    Applying cpp_get_token repeatedly yields a stream of pre-processor
 242    tokens.  Usually, there is only one cpp_reader object active.  */
 243
 244 struct cpp_reader
 245 {
 246   /* Top of buffer stack.  */
 247   cpp_buffer *buffer;
 248
 249   /* Lexer state.  */
 250   struct lexer_state state;
 251
 252   /* The position of the last lexed token and last lexed directive.  */
 253   cpp_lexer_pos lexer_pos;
 254   cpp_lexer_pos directive_pos;
 255
 256   /* Memory pools.  */
 257   cpp_pool ident_pool;          /* For all identifiers, and permanent
 258                                    numbers and strings.  */
 259   cpp_pool macro_pool;          /* For macro definitions.  Permanent.  */
 260   cpp_pool argument_pool;       /* For macro arguments.  Temporary.   */
 261
 262   /* Context stack.  */
 263   struct cpp_context base_context;
 264   struct cpp_context *context;
 265
 266   /* If in_directive, the directive if known.  */
 267   const struct directive *directive;
 268
 269   /* Multiple inlcude optimisation.  */
 270   enum mi_state mi_state;
 271   enum mi_ind mi_if_not_defined;
 272   unsigned int mi_lexed;
 273   const cpp_hashnode *mi_cmacro;
 274   const cpp_hashnode *mi_ind_cmacro;
 275
 276   /* Token lookahead.  */
 277   struct cpp_lookahead *la_read;        /* Read from this lookahead.  */
 278   struct cpp_lookahead *la_write;       /* Write to this lookahead.  */
 279   struct cpp_lookahead *la_unused;      /* Free store.  */
 280   struct cpp_lookahead *la_saved;       /* Backup when entering directive.  */
 281
 282   /* Error counter for exit code.  */
 283   unsigned int errors;
 284
 285   /* Line and column where a newline was first seen in a string
 286      constant (multi-line strings).  */
 287   cpp_lexer_pos mlstring_pos;
 288
 289   /* Buffer to hold macro definition string.  */
 290   unsigned char *macro_buffer;
 291   unsigned int macro_buffer_len;
 292
 293   /* Current depth in #include directives that use <...>.  */
 294   unsigned int system_include_depth;
 295
 296   /* Current depth of buffer stack.  */
 297   unsigned int buffer_stack_depth;
 298
 299   /* Current depth in #include directives.  */
 300   unsigned int include_depth;
 301
 302   /* Hash table of macros and assertions.  See cpphash.c.  */
 303   struct htab *hashtab;
 304
 305   /* Tree of other included files.  See cppfiles.c.  */
 306   struct splay_tree_s *all_include_files;
 307
 308   /* Current maximum length of directory names in the search path
 309      for include files.  (Altered as we get more of them.)  */
 310   unsigned int max_include_len;
 311
 312   /* Date and time tokens.  Calculated together if either is requested.  */
 313   cpp_token date;
 314   cpp_token time;
 315
 316   /* Opaque handle to the dependencies of mkdeps.c.  Used by -M etc.  */
 317   struct deps *deps;
 318
 319   /* Obstack holding all macro hash nodes.  This never shrinks.
 320      See cpphash.c */
 321   struct obstack *hash_ob;
 322
 323   /* Obstack holding buffer and conditional structures.  This is a
 324      real stack.  See cpplib.c */
 325   struct obstack *buffer_ob;
 326
 327   /* Pragma table - dynamic, because a library user can add to the
 328      list of recognized pragmas.  */
 329   struct pragma_entry *pragmas;
 330
 331   /* Call backs.  */
 332   struct cpp_callbacks cb;
 333
 334   /* User visible options.  */
 335   struct cpp_options opts;
 336
 337   /* Special nodes - identifiers with predefined significance to the
 338      preprocessor.  */
 339   struct spec_nodes spec_nodes;
 340
 341   /* We're printed a warning recommending against using #import.  */
 342   unsigned char import_warning;
 343
 344   /* True if we are skipping a failed conditional group.  */
 345   unsigned char skipping;
 346
 347   /* Whether to print our version number.  Done this way so
 348      we don't get it twice for -v -version.  */
 349   unsigned char print_version;
 350 };
 351
 352 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 353    If the definition of `numchar' looks odd to you, please look up the
 354    definition of a pp-number in the C standard [section 6.4.8 of C99].
 355
 356    In the unlikely event that characters other than \r and \n enter
 357    the set is_vspace, the macro handle_newline() in cpplex.c must be
 358    updated.  */
 359 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 360
 361 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 362 #define is_numchar(x)   ISIDNUM(x)
 363 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 364 #define is_numstart(x)  ISDIGIT(x)
 365 #define is_hspace(x)    ISBLANK(x)
 366 #define is_vspace(x)    IS_VSPACE(x)
 367 #define is_nvspace(x)   IS_NVSPACE(x)
 368 #define is_space(x)     IS_SPACE_OR_NUL(x)
 369
 370 /* This table is constant if it can be initialized at compile time,
 371    which is the case if cpp was compiled with GCC >=2.7, or another
 372    compiler that supports C99.  */
 373 #if HAVE_DESIGNATED_INITIALIZERS
 374 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 375 #else
 376 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 377 #endif
 378
 379 /* Macros.  */
 380
 381 #define CPP_PRINT_DEPS(PFILE) CPP_OPTION (PFILE, print_deps)
 382 #define CPP_IN_SYSTEM_HEADER(PFILE) \
 383   (CPP_BUFFER (PFILE) && CPP_BUFFER (PFILE)->sysp)
 384 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
 385 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
 386
 387 /* Hash step.  The hash calculation is duplicated in cpp_lookup and
 388    parse_name.  */
 389 #define HASHSTEP(r, c) ((r) * 67 + (c - 113));
 390
 391 /* In cpperror.c  */
 392 enum error_type { WARNING = 0, PEDWARN, ERROR, FATAL, ICE };
 393 extern int _cpp_begin_message PARAMS ((cpp_reader *, enum error_type,
 394                                        const char *, const cpp_lexer_pos *));
 395
 396 /* In cppmacro.c */
 397 extern void _cpp_free_definition        PARAMS ((cpp_hashnode *));
 398 extern int _cpp_create_definition       PARAMS ((cpp_reader *, cpp_hashnode *));
 399 extern void _cpp_pop_context            PARAMS ((cpp_reader *));
 400 extern void _cpp_free_lookaheads        PARAMS ((cpp_reader *));
 401 extern void _cpp_release_lookahead      PARAMS ((cpp_reader *));
 402 extern void _cpp_push_token             PARAMS ((cpp_reader *, const cpp_token *,
 403                                                  const cpp_lexer_pos *));
 404
 405 /* In cpphash.c */
 406 extern void _cpp_init_hashtable         PARAMS ((cpp_reader *));
 407 extern void _cpp_cleanup_hashtable      PARAMS ((cpp_reader *));
 408 extern cpp_hashnode *_cpp_lookup_with_hash PARAMS ((cpp_reader*, size_t,
 409                                                     unsigned int));
 410
 411 /* In cppfiles.c */
 412 extern void _cpp_fake_include           PARAMS ((cpp_reader *, const char *));
 413 extern void _cpp_never_reread           PARAMS ((struct include_file *));
 414 extern char *_cpp_simplify_pathname     PARAMS ((char *));
 415 extern int _cpp_read_file               PARAMS ((cpp_reader *, const char *));
 416 extern int _cpp_execute_include         PARAMS ((cpp_reader *,
 417                                                  const cpp_token *,
 418                                                  enum include_type));
 419 extern int _cpp_compare_file_date       PARAMS ((cpp_reader *,
 420                                                  const cpp_token *));
 421 extern void _cpp_report_missing_guards  PARAMS ((cpp_reader *));
 422 extern void _cpp_init_includes          PARAMS ((cpp_reader *));
 423 extern void _cpp_cleanup_includes       PARAMS ((cpp_reader *));
 424 extern void _cpp_pop_file_buffer        PARAMS ((cpp_reader *, cpp_buffer *));
 425
 426 /* In cppexp.c */
 427 extern int _cpp_parse_expr              PARAMS ((cpp_reader *));
 428
 429 /* In cpplex.c */
 430 extern void _cpp_lex_token              PARAMS ((cpp_reader *, cpp_token *));
 431 extern int _cpp_equiv_tokens            PARAMS ((const cpp_token *,
 432                                                  const cpp_token *));
 433 extern void _cpp_init_pool              PARAMS ((cpp_pool *, unsigned int,
 434                                                   unsigned int, unsigned int));
 435 extern void _cpp_free_pool              PARAMS ((cpp_pool *));
 436 extern unsigned char *_cpp_pool_reserve PARAMS ((cpp_pool *, unsigned int));
 437 extern unsigned char *_cpp_pool_alloc   PARAMS ((cpp_pool *, unsigned int));
 438 extern unsigned char *_cpp_next_chunk   PARAMS ((cpp_pool *, unsigned int,
 439                                                  unsigned char **));
 440 extern void _cpp_lock_pool              PARAMS ((cpp_pool *));
 441 extern void _cpp_unlock_pool            PARAMS ((cpp_pool *));
 442
 443 /* In cpplib.c */
 444 extern int _cpp_test_assertion PARAMS ((cpp_reader *, int *));
 445 extern int _cpp_handle_directive PARAMS ((cpp_reader *, int));
 446 extern void _cpp_define_builtin PARAMS ((cpp_reader *, const char *));
 447 extern void _cpp_do__Pragma     PARAMS ((cpp_reader *));
 448 extern void _cpp_init_stacks    PARAMS ((cpp_reader *));
 449 extern void _cpp_cleanup_stacks PARAMS ((cpp_reader *));
 450 extern void _cpp_init_internal_pragmas PARAMS ((cpp_reader *));
 451 extern void _cpp_do_file_change PARAMS ((cpp_reader *, enum cpp_fc_reason,
 452                                          const char *, unsigned int));
 453
 454 /* Utility routines and macros.  */
 455 #define DSC(str) (const U_CHAR *)str, sizeof str - 1
 456 #define xnew(T)         (T *) xmalloc (sizeof(T))
 457 #define xcnew(T)        (T *) xcalloc (1, sizeof(T))
 458 #define xnewvec(T, N)   (T *) xmalloc (sizeof(T) * (N))
 459 #define xcnewvec(T, N)  (T *) xcalloc (N, sizeof(T))
 460 #define xobnew(O, T)    (T *) obstack_alloc (O, sizeof(T))
 461
 462 /* These are inline functions instead of macros so we can get type
 463    checking.  */
 464 typedef unsigned char U_CHAR;
 465 #define U (const U_CHAR *)  /* Intended use: U"string" */
 466
 467 static inline int ustrcmp       PARAMS ((const U_CHAR *, const U_CHAR *));
 468 static inline int ustrncmp      PARAMS ((const U_CHAR *, const U_CHAR *,
 469                                          size_t));
 470 static inline size_t ustrlen    PARAMS ((const U_CHAR *));
 471 static inline U_CHAR *uxstrdup  PARAMS ((const U_CHAR *));
 472 static inline U_CHAR *ustrchr   PARAMS ((const U_CHAR *, int));
 473 static inline int ufputs        PARAMS ((const U_CHAR *, FILE *));
 474
 475 static inline int
 476 ustrcmp (s1, s2)
 477      const U_CHAR *s1, *s2;
 478 {
 479   return strcmp ((const char *)s1, (const char *)s2);
 480 }
 481
 482 static inline int
 483 ustrncmp (s1, s2, n)
 484      const U_CHAR *s1, *s2;
 485      size_t n;
 486 {
 487   return strncmp ((const char *)s1, (const char *)s2, n);
 488 }
 489
 490 static inline size_t
 491 ustrlen (s1)
 492      const U_CHAR *s1;
 493 {
 494   return strlen ((const char *)s1);
 495 }
 496
 497 static inline U_CHAR *
 498 uxstrdup (s1)
 499      const U_CHAR *s1;
 500 {
 501   return (U_CHAR *) xstrdup ((const char *)s1);
 502 }
 503
 504 static inline U_CHAR *
 505 ustrchr (s1, c)
 506      const U_CHAR *s1;
 507      int c;
 508 {
 509   return (U_CHAR *) strchr ((const char *)s1, c);
 510 }
 511
 512 static inline int
 513 ufputs (s, f)
 514      const U_CHAR *s;
 515      FILE *f;
 516 {
 517   return fputs ((const char *)s, f);
 518 }
 519
 520 #endif