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_H
  23 #define GCC_CPPHASH_H
  24
  25 #include "hashtable.h"
  26
  27 struct directive;               /* Deliberately incomplete.  */
  28
  29 /* Test if a sign is valid within a preprocessing number.  */
  30 #define VALID_SIGN(c, prevc) \
  31   (((c) == '+' || (c) == '-') && \
  32    ((prevc) == 'e' || (prevc) == 'E' \
  33     || (((prevc) == 'p' || (prevc) == 'P') \
  34         && CPP_OPTION (pfile, extended_numbers))))
  35
  36 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
  37 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
  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 /* A generic memory buffer, and operations on it.  */
  46
  47 typedef struct _cpp_buff _cpp_buff;
  48 struct _cpp_buff
  49 {
  50   struct _cpp_buff *next;
  51   unsigned char *base, *cur, *limit;
  52 };
  53
  54 extern _cpp_buff *_cpp_get_buff PARAMS ((cpp_reader *, size_t));
  55 extern void _cpp_release_buff PARAMS ((cpp_reader *, _cpp_buff *));
  56 extern void _cpp_extend_buff PARAMS ((cpp_reader *, _cpp_buff **, size_t));
  57 extern _cpp_buff *_cpp_append_extend_buff PARAMS ((cpp_reader *, _cpp_buff *,
  58                                                    size_t));
  59 extern void _cpp_free_buff PARAMS ((_cpp_buff *));
  60 extern unsigned char *_cpp_aligned_alloc PARAMS ((cpp_reader *, size_t));
  61 extern unsigned char *_cpp_unaligned_alloc PARAMS ((cpp_reader *, size_t));
  62
  63 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
  64 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
  65 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
  66
  67 /* List of directories to look for include files in.  */
  68 struct search_path
  69 {
  70   struct search_path *next;
  71
  72   /* NOTE: NAME may not be null terminated for the case of the current
  73      file's directory!  */
  74   const char *name;
  75   unsigned int len;
  76   /* We use these to tell if the directory mentioned here is a duplicate
  77      of an earlier directory on the search path.  */
  78   ino_t ino;
  79   dev_t dev;
  80   /* Non-zero if it is a system include directory.  */
  81   int sysp;
  82   /* Mapping of file names for this directory.  Only used on MS-DOS
  83      and related platforms.  */
  84   struct file_name_map *name_map;
  85 };
  86
  87 /* #include types.  */
  88 enum include_type {IT_INCLUDE, IT_INCLUDE_NEXT, IT_IMPORT, IT_CMDLINE};
  89
  90 union utoken
  91 {
  92   const cpp_token *token;
  93   const cpp_token **ptoken;
  94 };
  95
  96 typedef struct tokenrun tokenrun;
  97 struct tokenrun
  98 {
  99   tokenrun *next, *prev;
 100   cpp_token *base, *limit;
 101 };
 102
 103 typedef struct cpp_context cpp_context;
 104 struct cpp_context
 105 {
 106   /* Doubly-linked list.  */
 107   cpp_context *next, *prev;
 108
 109   /* Contexts other than the base context are contiguous tokens.
 110      e.g. macro expansions, expanded argument tokens.  */
 111   union utoken first;
 112   union utoken last;
 113
 114   /* If non-NULL, a buffer used for storage related to this context.
 115      When the context is popped, the buffer is released.  */
 116   _cpp_buff *buff;
 117
 118   /* For a macro context, the macro node, otherwise NULL.  */
 119   cpp_hashnode *macro;
 120
 121   /* True if utoken element is token, else ptoken.  */
 122   bool direct_p;
 123 };
 124
 125 struct lexer_state
 126 {
 127   /* Nonzero if first token on line is CPP_HASH.  */
 128   unsigned char in_directive;
 129
 130   /* True if we are skipping a failed conditional group.  */
 131   unsigned char skipping;
 132
 133   /* Nonzero if in a directive that takes angle-bracketed headers.  */
 134   unsigned char angled_headers;
 135
 136   /* Nonzero to save comments.  Turned off if discard_comments, and in
 137      all directives apart from #define.  */
 138   unsigned char save_comments;
 139
 140   /* Nonzero if we're mid-comment.  */
 141   unsigned char lexing_comment;
 142
 143   /* Nonzero if lexing __VA_ARGS__ is valid.  */
 144   unsigned char va_args_ok;
 145
 146   /* Nonzero if lexing poisoned identifiers is valid.  */
 147   unsigned char poisoned_ok;
 148
 149   /* Nonzero to prevent macro expansion.  */
 150   unsigned char prevent_expansion;
 151
 152   /* Nonzero when parsing arguments to a function-like macro.  */
 153   unsigned char parsing_args;
 154
 155   /* Nonzero when in a # NUMBER directive.  */
 156   unsigned char line_extension;
 157 };
 158
 159 /* Special nodes - identifiers with predefined significance.  */
 160 struct spec_nodes
 161 {
 162   cpp_hashnode *n_L;                    /* L"str" */
 163   cpp_hashnode *n_defined;              /* defined operator */
 164   cpp_hashnode *n_true;                 /* C++ keyword true */
 165   cpp_hashnode *n_false;                /* C++ keyword false */
 166   cpp_hashnode *n__STRICT_ANSI__;       /* STDC_0_IN_SYSTEM_HEADERS */
 167   cpp_hashnode *n__CHAR_UNSIGNED__;     /* plain char is unsigned */
 168   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 169 };
 170
 171 struct cpp_buffer
 172 {
 173   const unsigned char *cur;      /* current position */
 174   const unsigned char *rlimit; /* end of valid data */
 175   const unsigned char *line_base; /* start of current line */
 176   cppchar_t read_ahead;         /* read ahead character */
 177   cppchar_t extra_char;         /* extra read-ahead for long tokens.  */
 178
 179   struct cpp_buffer *prev;
 180
 181   const unsigned char *buf;      /* Entire character buffer.  */
 182
 183   /* Pointer into the include table; non-NULL if this is a file
 184      buffer.  Used for include_next and 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   /* Contains PREV_WHITE and/or AVOID_LPASTE.  */
 195   unsigned char saved_flags;
 196
 197   /* Because of the way the lexer works, -Wtrigraphs can sometimes
 198      warn twice for the same trigraph.  This helps prevent that.  */
 199   const unsigned char *last_Wtrigraphs;
 200
 201   /* True if we have already warned about C++ comments in this file.
 202      The warning happens only for C89 extended mode with -pedantic on,
 203      or for -Wtraditional, and only once per file (otherwise it would
 204      be far too noisy).  */
 205   unsigned char warned_cplusplus_comments;
 206
 207   /* True if we don't process trigraphs and escaped newlines.  True
 208      for preprocessed input, command line directives, and _Pragma
 209      buffers.  */
 210   unsigned char from_stage3;
 211
 212   /* Nonzero means that the directory to start searching for ""
 213      include files has been calculated and stored in "dir" below.  */
 214   unsigned char search_cached;
 215
 216   /* At EOF, a buffer is automatically popped.  If RETURN_AT_EOF is
 217      true, a CPP_EOF token is then returned.  Otherwise, the next
 218      token from the enclosing buffer is returned.  */
 219   bool return_at_eof;
 220
 221   /* The directory of the this buffer's file.  Its NAME member is not
 222      allocated, so we don't need to worry about freeing it.  */
 223   struct search_path dir;
 224 };
 225
 226 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 227    Applying cpp_get_token repeatedly yields a stream of pre-processor
 228    tokens.  Usually, there is only one cpp_reader object active.  */
 229
 230 struct cpp_reader
 231 {
 232   /* Top of buffer stack.  */
 233   cpp_buffer *buffer;
 234
 235   /* Lexer state.  */
 236   struct lexer_state state;
 237
 238   /* Source line tracking.  */
 239   struct line_maps line_maps;
 240   const struct line_map *map;
 241   unsigned int line;
 242
 243   /* The line of the '#' of the current directive.  */
 244   unsigned int directive_line;
 245
 246   /* Memory buffers.  */
 247   _cpp_buff *a_buff;            /* Aligned permanent storage.  */
 248   _cpp_buff *u_buff;            /* Unaligned permanent storage.  */
 249   _cpp_buff *free_buffs;        /* Free buffer chain.  */
 250
 251   /* Context stack.  */
 252   struct cpp_context base_context;
 253   struct cpp_context *context;
 254
 255   /* If in_directive, the directive if known.  */
 256   const struct directive *directive;
 257
 258   /* Multiple inlcude optimisation.  */
 259   const cpp_hashnode *mi_cmacro;
 260   const cpp_hashnode *mi_ind_cmacro;
 261   bool mi_valid;
 262
 263   /* Lexing.  */
 264   cpp_token *cur_token;
 265   tokenrun base_run, *cur_run;
 266   unsigned int lookaheads;
 267
 268   /* Non-zero prevents the lexer from re-using the token runs.  */
 269   unsigned int keep_tokens;
 270
 271   /* Error counter for exit code.  */
 272   unsigned int errors;
 273
 274   /* Line and column where a newline was first seen in a string
 275      constant (multi-line strings).  */
 276   unsigned int mls_line;
 277   unsigned int mls_col;
 278
 279   /* Buffer to hold macro definition string.  */
 280   unsigned char *macro_buffer;
 281   unsigned int macro_buffer_len;
 282
 283   /* Tree of other included files.  See cppfiles.c.  */
 284   struct splay_tree_s *all_include_files;
 285
 286   /* Current maximum length of directory names in the search path
 287      for include files.  (Altered as we get more of them.)  */
 288   unsigned int max_include_len;
 289
 290   /* Date and time tokens.  Calculated together if either is requested.  */
 291   cpp_token date;
 292   cpp_token time;
 293
 294   /* EOF token, and a token forcing paste avoidance.  */
 295   cpp_token avoid_paste;
 296   cpp_token eof;
 297
 298   /* Opaque handle to the dependencies of mkdeps.c.  Used by -M etc.  */
 299   struct deps *deps;
 300
 301   /* Obstack holding all macro hash nodes.  This never shrinks.
 302      See cpphash.c */
 303   struct obstack hash_ob;
 304
 305   /* Obstack holding buffer and conditional structures.  This is a
 306      real stack.  See cpplib.c.  */
 307   struct obstack buffer_ob;
 308
 309   /* Pragma table - dynamic, because a library user can add to the
 310      list of recognized pragmas.  */
 311   struct pragma_entry *pragmas;
 312
 313   /* Call backs.  */
 314   struct cpp_callbacks cb;
 315
 316   /* Identifier hash table.  */
 317   struct ht *hash_table;
 318
 319   /* User visible options.  */
 320   struct cpp_options opts;
 321
 322   /* Special nodes - identifiers with predefined significance to the
 323      preprocessor.  */
 324   struct spec_nodes spec_nodes;
 325
 326   /* Whether to print our version number.  Done this way so
 327      we don't get it twice for -v -version.  */
 328   unsigned char print_version;
 329
 330   /* Whether cpplib owns the hashtable.  */
 331   unsigned char our_hashtable;
 332 };
 333
 334 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 335    If the definition of `numchar' looks odd to you, please look up the
 336    definition of a pp-number in the C standard [section 6.4.8 of C99].
 337
 338    In the unlikely event that characters other than \r and \n enter
 339    the set is_vspace, the macro handle_newline() in cpplex.c must be
 340    updated.  */
 341 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 342
 343 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 344 #define is_numchar(x)   ISIDNUM(x)
 345 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 346 #define is_numstart(x)  ISDIGIT(x)
 347 #define is_hspace(x)    ISBLANK(x)
 348 #define is_vspace(x)    IS_VSPACE(x)
 349 #define is_nvspace(x)   IS_NVSPACE(x)
 350 #define is_space(x)     IS_SPACE_OR_NUL(x)
 351
 352 /* This table is constant if it can be initialized at compile time,
 353    which is the case if cpp was compiled with GCC >=2.7, or another
 354    compiler that supports C99.  */
 355 #if HAVE_DESIGNATED_INITIALIZERS
 356 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 357 #else
 358 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 359 #endif
 360
 361 /* Macros.  */
 362
 363 #define CPP_PRINT_DEPS(PFILE) CPP_OPTION (PFILE, print_deps)
 364 #define CPP_IN_SYSTEM_HEADER(PFILE) ((PFILE)->map && (PFILE)->map->sysp)
 365 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
 366 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
 367
 368 /* In cpperror.c  */
 369 enum error_type { WARNING = 0, WARNING_SYSHDR, PEDWARN, ERROR, FATAL, ICE };
 370 extern int _cpp_begin_message PARAMS ((cpp_reader *, enum error_type,
 371                                        unsigned int, unsigned int));
 372
 373 /* In cppmacro.c */
 374 extern void _cpp_free_definition        PARAMS ((cpp_hashnode *));
 375 extern int _cpp_create_definition       PARAMS ((cpp_reader *, cpp_hashnode *));
 376 extern void _cpp_pop_context            PARAMS ((cpp_reader *));
 377
 378 /* In cpphash.c */
 379 extern void _cpp_init_hashtable         PARAMS ((cpp_reader *, hash_table *));
 380 extern void _cpp_destroy_hashtable      PARAMS ((cpp_reader *));
 381
 382 /* In cppfiles.c */
 383 extern void _cpp_fake_include           PARAMS ((cpp_reader *, const char *));
 384 extern void _cpp_never_reread           PARAMS ((struct include_file *));
 385 extern char *_cpp_simplify_pathname     PARAMS ((char *));
 386 extern bool _cpp_read_file              PARAMS ((cpp_reader *, const char *));
 387 extern bool _cpp_execute_include        PARAMS ((cpp_reader *,
 388                                                  const cpp_token *,
 389                                                  enum include_type));
 390 extern int _cpp_compare_file_date       PARAMS ((cpp_reader *,
 391                                                  const cpp_token *));
 392 extern void _cpp_report_missing_guards  PARAMS ((cpp_reader *));
 393 extern void _cpp_init_includes          PARAMS ((cpp_reader *));
 394 extern void _cpp_cleanup_includes       PARAMS ((cpp_reader *));
 395 extern void _cpp_pop_file_buffer        PARAMS ((cpp_reader *,
 396                                                  struct include_file *));
 397
 398 /* In cppexp.c */
 399 extern int _cpp_parse_expr              PARAMS ((cpp_reader *));
 400
 401 /* In cpplex.c */
 402 extern cpp_token *_cpp_temp_token       PARAMS ((cpp_reader *));
 403 extern const cpp_token *_cpp_lex_token  PARAMS ((cpp_reader *));
 404 extern cpp_token *_cpp_lex_direct       PARAMS ((cpp_reader *));
 405 extern int _cpp_equiv_tokens            PARAMS ((const cpp_token *,
 406                                                  const cpp_token *));
 407 extern void _cpp_init_tokenrun          PARAMS ((tokenrun *, unsigned int));
 408
 409 /* In cppinit.c.  */
 410 extern bool _cpp_push_next_buffer       PARAMS ((cpp_reader *));
 411
 412 /* In cpplib.c */
 413 extern int _cpp_test_assertion PARAMS ((cpp_reader *, int *));
 414 extern int _cpp_handle_directive PARAMS ((cpp_reader *, int));
 415 extern void _cpp_define_builtin PARAMS ((cpp_reader *, const char *));
 416 extern void _cpp_do__Pragma     PARAMS ((cpp_reader *));
 417 extern void _cpp_init_directives PARAMS ((cpp_reader *));
 418 extern void _cpp_init_internal_pragmas PARAMS ((cpp_reader *));
 419 extern void _cpp_do_file_change PARAMS ((cpp_reader *, enum lc_reason,
 420                                          const char *,
 421                                          unsigned int, unsigned int));
 422 extern void _cpp_pop_buffer PARAMS ((cpp_reader *));
 423
 424 /* Utility routines and macros.  */
 425 #define DSC(str) (const U_CHAR *)str, sizeof str - 1
 426 #define xnew(T)         (T *) xmalloc (sizeof(T))
 427 #define xcnew(T)        (T *) xcalloc (1, sizeof(T))
 428 #define xnewvec(T, N)   (T *) xmalloc (sizeof(T) * (N))
 429 #define xcnewvec(T, N)  (T *) xcalloc (N, sizeof(T))
 430 #define xobnew(O, T)    (T *) obstack_alloc (O, sizeof(T))
 431
 432 /* These are inline functions instead of macros so we can get type
 433    checking.  */
 434 typedef unsigned char U_CHAR;
 435 #define U (const U_CHAR *)  /* Intended use: U"string" */
 436
 437 static inline int ustrcmp       PARAMS ((const U_CHAR *, const U_CHAR *));
 438 static inline int ustrncmp      PARAMS ((const U_CHAR *, const U_CHAR *,
 439                                          size_t));
 440 static inline size_t ustrlen    PARAMS ((const U_CHAR *));
 441 static inline U_CHAR *uxstrdup  PARAMS ((const U_CHAR *));
 442 static inline U_CHAR *ustrchr   PARAMS ((const U_CHAR *, int));
 443 static inline int ufputs        PARAMS ((const U_CHAR *, FILE *));
 444
 445 static inline int
 446 ustrcmp (s1, s2)
 447      const U_CHAR *s1, *s2;
 448 {
 449   return strcmp ((const char *)s1, (const char *)s2);
 450 }
 451
 452 static inline int
 453 ustrncmp (s1, s2, n)
 454      const U_CHAR *s1, *s2;
 455      size_t n;
 456 {
 457   return strncmp ((const char *)s1, (const char *)s2, n);
 458 }
 459
 460 static inline size_t
 461 ustrlen (s1)
 462      const U_CHAR *s1;
 463 {
 464   return strlen ((const char *)s1);
 465 }
 466
 467 static inline U_CHAR *
 468 uxstrdup (s1)
 469      const U_CHAR *s1;
 470 {
 471   return (U_CHAR *) xstrdup ((const char *)s1);
 472 }
 473
 474 static inline U_CHAR *
 475 ustrchr (s1, c)
 476      const U_CHAR *s1;
 477      int c;
 478 {
 479   return (U_CHAR *) strchr ((const char *)s1, c);
 480 }
 481
 482 static inline int
 483 ufputs (s, f)
 484      const U_CHAR *s;
 485      FILE *f;
 486 {
 487   return fputs ((const char *)s, f);
 488 }
 489
 490 #endif /* ! GCC_CPPHASH_H */