gcc/cpphash.h

   1 /* Part of CPP library.
   2    Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002
   3    Free Software Foundation, Inc.
   4
   5 This program is free software; you can redistribute it and/or modify it
   6 under the terms of the GNU General Public License as published by the
   7 Free Software Foundation; either version 2, or (at your option) any
   8 later version.
   9
  10 This program is distributed in the hope that it will be useful,
  11 but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13 GNU General Public License for more details.
  14
  15 You should have received a copy of the GNU General Public License
  16 along with this program; if not, write to the Free Software
  17 Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  18
  19 /* This header defines all the internal data structures and functions
  20    that need to be visible across files.  It's called cpphash.h for
  21    historical reasons.  */
  22
  23 #ifndef GCC_CPPHASH_H
  24 #define GCC_CPPHASH_H
  25
  26 #include "hashtable.h"
  27
  28 struct directive;               /* Deliberately incomplete.  */
  29 struct pending_option;
  30 struct op;
  31
  32 /* Test if a sign is valid within a preprocessing number.  */
  33 #define VALID_SIGN(c, prevc) \
  34   (((c) == '+' || (c) == '-') && \
  35    ((prevc) == 'e' || (prevc) == 'E' \
  36     || (((prevc) == 'p' || (prevc) == 'P') \
  37         && CPP_OPTION (pfile, extended_numbers))))
  38
  39 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
  40 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
  41 #define CPP_BUF_COLUMN(BUF, CUR) ((CUR) - (BUF)->line_base + (BUF)->col_adjust)
  42 #define CPP_BUF_COL(BUF) CPP_BUF_COLUMN(BUF, (BUF)->cur)
  43
  44 /* Maximum nesting of cpp_buffers.  We use a static limit, partly for
  45    efficiency, and partly to limit runaway recursion.  */
  46 #define CPP_STACK_MAX 200
  47
  48 /* A generic memory buffer, and operations on it.  */
  49 typedef struct _cpp_buff _cpp_buff;
  50 struct _cpp_buff
  51 {
  52   struct _cpp_buff *next;
  53   unsigned char *base, *cur, *limit;
  54 };
  55
  56 extern _cpp_buff *_cpp_get_buff PARAMS ((cpp_reader *, size_t));
  57 extern void _cpp_release_buff PARAMS ((cpp_reader *, _cpp_buff *));
  58 extern void _cpp_extend_buff PARAMS ((cpp_reader *, _cpp_buff **, size_t));
  59 extern _cpp_buff *_cpp_append_extend_buff PARAMS ((cpp_reader *, _cpp_buff *,
  60                                                    size_t));
  61 extern void _cpp_free_buff PARAMS ((_cpp_buff *));
  62 extern unsigned char *_cpp_aligned_alloc PARAMS ((cpp_reader *, size_t));
  63 extern unsigned char *_cpp_unaligned_alloc PARAMS ((cpp_reader *, size_t));
  64
  65 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
  66 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
  67 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
  68
  69 /* List of directories to look for include files in.  */
  70 struct search_path
  71 {
  72   struct search_path *next;
  73
  74   /* NOTE: NAME may not be null terminated for the case of the current
  75      file's directory!  */
  76   const char *name;
  77   unsigned int len;
  78   /* We use these to tell if the directory mentioned here is a duplicate
  79      of an earlier directory on the search path.  */
  80   ino_t ino;
  81   dev_t dev;
  82   /* Non-zero if it is a system include directory.  */
  83   int sysp;
  84   /* Mapping of file names for this directory.  Only used on MS-DOS
  85      and related platforms.  */
  86   struct file_name_map *name_map;
  87 };
  88
  89 /* #include types.  */
  90 enum include_type {IT_INCLUDE, IT_INCLUDE_NEXT, IT_IMPORT, IT_CMDLINE};
  91
  92 union utoken
  93 {
  94   const cpp_token *token;
  95   const cpp_token **ptoken;
  96 };
  97
  98 /* A "run" of tokens; part of a chain of runs.  */
  99 typedef struct tokenrun tokenrun;
 100 struct tokenrun
 101 {
 102   tokenrun *next, *prev;
 103   cpp_token *base, *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   union utoken first;
 115   union utoken last;
 116
 117   /* If non-NULL, a buffer used for storage related to this context.
 118      When the context is popped, the buffer is released.  */
 119   _cpp_buff *buff;
 120
 121   /* For a macro context, the macro node, otherwise NULL.  */
 122   cpp_hashnode *macro;
 123
 124   /* True if utoken element is token, else ptoken.  */
 125   bool direct_p;
 126 };
 127
 128 struct lexer_state
 129 {
 130   /* Nonzero if first token on line is CPP_HASH.  */
 131   unsigned char in_directive;
 132
 133   /* True if we are skipping a failed conditional group.  */
 134   unsigned char skipping;
 135
 136   /* Nonzero if in a directive that takes angle-bracketed headers.  */
 137   unsigned char angled_headers;
 138
 139   /* Nonzero to save comments.  Turned off if discard_comments, and in
 140      all directives apart from #define.  */
 141   unsigned char save_comments;
 142
 143   /* Nonzero if we're mid-comment.  */
 144   unsigned char lexing_comment;
 145
 146   /* Nonzero if lexing __VA_ARGS__ is valid.  */
 147   unsigned char va_args_ok;
 148
 149   /* Nonzero if lexing poisoned identifiers is valid.  */
 150   unsigned char poisoned_ok;
 151
 152   /* Nonzero to prevent macro expansion.  */
 153   unsigned char prevent_expansion;
 154
 155   /* Nonzero when parsing arguments to a function-like macro.  */
 156   unsigned char parsing_args;
 157
 158   /* Nonzero to skip evaluating part of an expression.  */
 159   unsigned int skip_eval;
 160 };
 161
 162 /* Special nodes - identifiers with predefined significance.  */
 163 struct spec_nodes
 164 {
 165   cpp_hashnode *n_defined;              /* defined operator */
 166   cpp_hashnode *n_true;                 /* C++ keyword true */
 167   cpp_hashnode *n_false;                /* C++ keyword false */
 168   cpp_hashnode *n__STRICT_ANSI__;       /* STDC_0_IN_SYSTEM_HEADERS */
 169   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 170 };
 171
 172 /* Represents the contents of a file cpplib has read in.  */
 173 struct cpp_buffer
 174 {
 175   const unsigned char *cur;      /* current position */
 176   const unsigned char *backup_to; /* if peeked character is not wanted */
 177   const unsigned char *rlimit; /* end of valid data */
 178   const unsigned char *line_base; /* start of current line */
 179
 180   struct cpp_buffer *prev;
 181
 182   const unsigned char *buf;      /* Entire character buffer.  */
 183
 184   /* Pointer into the include table; non-NULL if this is a file
 185      buffer.  Used for include_next and to record control macros.  */
 186   struct include_file *inc;
 187
 188   /* Value of if_stack at start of this file.
 189      Used to prohibit unmatched #endif (etc) in an include file.  */
 190   struct if_stack *if_stack;
 191
 192   /* Token column position adjustment owing to tabs in whitespace.  */
 193   unsigned int col_adjust;
 194
 195   /* Contains PREV_WHITE and/or AVOID_LPASTE.  */
 196   unsigned char saved_flags;
 197
 198   /* Because of the way the lexer works, -Wtrigraphs can sometimes
 199      warn twice for the same trigraph.  This helps prevent that.  */
 200   const unsigned char *last_Wtrigraphs;
 201
 202   /* True if we have already warned about C++ comments in this file.
 203      The warning happens only for C89 extended mode with -pedantic on,
 204      or for -Wtraditional, and only once per file (otherwise it would
 205      be far too noisy).  */
 206   unsigned char warned_cplusplus_comments;
 207
 208   /* True if we don't process trigraphs and escaped newlines.  True
 209      for preprocessed input, command line directives, and _Pragma
 210      buffers.  */
 211   unsigned char from_stage3;
 212
 213   /* Nonzero means that the directory to start searching for ""
 214      include files has been calculated and stored in "dir" below.  */
 215   unsigned char search_cached;
 216
 217   /* At EOF, a buffer is automatically popped.  If RETURN_AT_EOF is
 218      true, a CPP_EOF token is then returned.  Otherwise, the next
 219      token from the enclosing buffer is returned.  */
 220   bool return_at_eof;
 221
 222   /* The directory of the this buffer's file.  Its NAME member is not
 223      allocated, so we don't need to worry about freeing it.  */
 224   struct search_path dir;
 225 };
 226
 227 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 228    Applying cpp_get_token repeatedly yields a stream of pre-processor
 229    tokens.  Usually, there is only one cpp_reader object active.  */
 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   /* The next -include-d file; NULL if they all are done.  If it
 259      points to NULL, the last one is in progress, and
 260      _cpp_maybe_push_include_file has yet to restore the line map.  */
 261   struct pending_option **next_include_file;
 262
 263   /* Multiple inlcude optimisation.  */
 264   const cpp_hashnode *mi_cmacro;
 265   const cpp_hashnode *mi_ind_cmacro;
 266   bool mi_valid;
 267
 268   /* Lexing.  */
 269   cpp_token *cur_token;
 270   tokenrun base_run, *cur_run;
 271   unsigned int lookaheads;
 272
 273   /* Non-zero prevents the lexer from re-using the token runs.  */
 274   unsigned int keep_tokens;
 275
 276   /* Error counter for exit code.  */
 277   unsigned int errors;
 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   /* Expression parser stack.  */
 320   struct op *op_stack, *op_limit;
 321
 322   /* User visible options.  */
 323   struct cpp_options opts;
 324
 325   /* Special nodes - identifiers with predefined significance to the
 326      preprocessor.  */
 327   struct spec_nodes spec_nodes;
 328
 329   /* Whether to print our version number.  Done this way so
 330      we don't get it twice for -v -version.  */
 331   unsigned char print_version;
 332
 333   /* Whether cpplib owns the hashtable.  */
 334   unsigned char our_hashtable;
 335 };
 336
 337 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 338    If the definition of `numchar' looks odd to you, please look up the
 339    definition of a pp-number in the C standard [section 6.4.8 of C99].
 340
 341    In the unlikely event that characters other than \r and \n enter
 342    the set is_vspace, the macro handle_newline() in cpplex.c must be
 343    updated.  */
 344 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 345
 346 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 347 #define is_numchar(x)   ISIDNUM(x)
 348 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 349 #define is_numstart(x)  ISDIGIT(x)
 350 #define is_hspace(x)    ISBLANK(x)
 351 #define is_vspace(x)    IS_VSPACE(x)
 352 #define is_nvspace(x)   IS_NVSPACE(x)
 353 #define is_space(x)     IS_SPACE_OR_NUL(x)
 354
 355 /* This table is constant if it can be initialized at compile time,
 356    which is the case if cpp was compiled with GCC >=2.7, or another
 357    compiler that supports C99.  */
 358 #if HAVE_DESIGNATED_INITIALIZERS
 359 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 360 #else
 361 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 362 #endif
 363
 364 /* Macros.  */
 365
 366 #define CPP_PRINT_DEPS(PFILE) CPP_OPTION (PFILE, print_deps)
 367 #define CPP_IN_SYSTEM_HEADER(PFILE) ((PFILE)->map && (PFILE)->map->sysp)
 368 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
 369 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
 370
 371 /* In cpperror.c  */
 372 extern int _cpp_begin_message PARAMS ((cpp_reader *, int,
 373                                        unsigned int, unsigned int));
 374
 375 /* In cppmacro.c */
 376 extern void _cpp_free_definition        PARAMS ((cpp_hashnode *));
 377 extern int _cpp_create_definition       PARAMS ((cpp_reader *, cpp_hashnode *));
 378 extern void _cpp_pop_context            PARAMS ((cpp_reader *));
 379
 380 /* In cpphash.c */
 381 extern void _cpp_init_hashtable         PARAMS ((cpp_reader *, hash_table *));
 382 extern void _cpp_destroy_hashtable      PARAMS ((cpp_reader *));
 383
 384 /* In cppfiles.c */
 385 extern void _cpp_fake_include           PARAMS ((cpp_reader *, const char *));
 386 extern void _cpp_never_reread           PARAMS ((struct include_file *));
 387 extern char *_cpp_simplify_pathname     PARAMS ((char *));
 388 extern bool _cpp_read_file              PARAMS ((cpp_reader *, const char *));
 389 extern bool _cpp_execute_include        PARAMS ((cpp_reader *,
 390                                                  const cpp_token *,
 391                                                  enum include_type));
 392 extern int _cpp_compare_file_date       PARAMS ((cpp_reader *,
 393                                                  const cpp_token *));
 394 extern void _cpp_report_missing_guards  PARAMS ((cpp_reader *));
 395 extern void _cpp_init_includes          PARAMS ((cpp_reader *));
 396 extern void _cpp_cleanup_includes       PARAMS ((cpp_reader *));
 397 extern void _cpp_pop_file_buffer        PARAMS ((cpp_reader *,
 398                                                  struct include_file *));
 399
 400 /* In cppexp.c */
 401 extern bool _cpp_parse_expr             PARAMS ((cpp_reader *));
 402 extern struct op *_cpp_expand_op_stack  PARAMS ((cpp_reader *));
 403
 404 /* In cpplex.c */
 405 extern cpp_token *_cpp_temp_token       PARAMS ((cpp_reader *));
 406 extern const cpp_token *_cpp_lex_token  PARAMS ((cpp_reader *));
 407 extern cpp_token *_cpp_lex_direct       PARAMS ((cpp_reader *));
 408 extern int _cpp_equiv_tokens            PARAMS ((const cpp_token *,
 409                                                  const cpp_token *));
 410 extern void _cpp_init_tokenrun          PARAMS ((tokenrun *, unsigned int));
 411
 412 /* In cppinit.c.  */
 413 extern void _cpp_maybe_push_include_file PARAMS ((cpp_reader *));
 414
 415 /* In cpplib.c */
 416 extern int _cpp_test_assertion PARAMS ((cpp_reader *, unsigned int *));
 417 extern int _cpp_handle_directive PARAMS ((cpp_reader *, int));
 418 extern void _cpp_define_builtin PARAMS ((cpp_reader *, const char *));
 419 extern void _cpp_do__Pragma     PARAMS ((cpp_reader *));
 420 extern void _cpp_init_directives PARAMS ((cpp_reader *));
 421 extern void _cpp_init_internal_pragmas PARAMS ((cpp_reader *));
 422 extern void _cpp_do_file_change PARAMS ((cpp_reader *, enum lc_reason,
 423                                          const char *,
 424                                          unsigned int, unsigned int));
 425 extern void _cpp_pop_buffer PARAMS ((cpp_reader *));
 426
 427 /* Utility routines and macros.  */
 428 #define DSC(str) (const uchar *)str, sizeof str - 1
 429 #define xnew(T)         (T *) xmalloc (sizeof(T))
 430 #define xcnew(T)        (T *) xcalloc (1, sizeof(T))
 431 #define xnewvec(T, N)   (T *) xmalloc (sizeof(T) * (N))
 432 #define xcnewvec(T, N)  (T *) xcalloc (N, sizeof(T))
 433 #define xobnew(O, T)    (T *) obstack_alloc (O, sizeof(T))
 434
 435 /* These are inline functions instead of macros so we can get type
 436    checking.  */
 437 typedef unsigned char uchar;
 438 #define U (const uchar *)  /* Intended use: U"string" */
 439
 440 static inline int ustrcmp       PARAMS ((const uchar *, const uchar *));
 441 static inline int ustrncmp      PARAMS ((const uchar *, const uchar *,
 442                                          size_t));
 443 static inline size_t ustrlen    PARAMS ((const uchar *));
 444 static inline uchar *uxstrdup   PARAMS ((const uchar *));
 445 static inline uchar *ustrchr    PARAMS ((const uchar *, int));
 446 static inline int ufputs        PARAMS ((const uchar *, FILE *));
 447
 448 static inline int
 449 ustrcmp (s1, s2)
 450      const uchar *s1, *s2;
 451 {
 452   return strcmp ((const char *)s1, (const char *)s2);
 453 }
 454
 455 static inline int
 456 ustrncmp (s1, s2, n)
 457      const uchar *s1, *s2;
 458      size_t n;
 459 {
 460   return strncmp ((const char *)s1, (const char *)s2, n);
 461 }
 462
 463 static inline size_t
 464 ustrlen (s1)
 465      const uchar *s1;
 466 {
 467   return strlen ((const char *)s1);
 468 }
 469
 470 static inline uchar *
 471 uxstrdup (s1)
 472      const uchar *s1;
 473 {
 474   return (uchar *) xstrdup ((const char *)s1);
 475 }
 476
 477 static inline uchar *
 478 ustrchr (s1, c)
 479      const uchar *s1;
 480      int c;
 481 {
 482   return (uchar *) strchr ((const char *)s1, c);
 483 }
 484
 485 static inline int
 486 ufputs (s, f)
 487      const uchar *s;
 488      FILE *f;
 489 {
 490   return fputs ((const char *)s, f);
 491 }
 492
 493 #endif /* ! GCC_CPPHASH_H */