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 typedef unsigned char uchar;
  33 #define U (const uchar *)  /* Intended use: U"string" */
  34
  35 #define BITS_PER_CPPCHAR_T (CHAR_BIT * sizeof (cppchar_t))
  36
  37 /* Test if a sign is valid within a preprocessing number.  */
  38 #define VALID_SIGN(c, prevc) \
  39   (((c) == '+' || (c) == '-') && \
  40    ((prevc) == 'e' || (prevc) == 'E' \
  41     || (((prevc) == 'p' || (prevc) == 'P') \
  42         && CPP_OPTION (pfile, extended_numbers))))
  43
  44 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
  45 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
  46 #define CPP_BUF_COLUMN(BUF, CUR) ((CUR) - (BUF)->line_base + (BUF)->col_adjust)
  47 #define CPP_BUF_COL(BUF) CPP_BUF_COLUMN(BUF, (BUF)->cur)
  48
  49 /* Maximum nesting of cpp_buffers.  We use a static limit, partly for
  50    efficiency, and partly to limit runaway recursion.  */
  51 #define CPP_STACK_MAX 200
  52
  53 /* Each macro definition is recorded in a cpp_macro structure.
  54    Variadic macros cannot occur with traditional cpp.  */
  55 struct cpp_macro
  56 {
  57   cpp_hashnode **params;        /* Parameters, if any.  */
  58   union
  59   {
  60     cpp_token *tokens;          /* Tokens of replacement list (ISO).  */
  61     const uchar *text;          /* Expansion text (traditional).  */
  62   } exp;
  63   unsigned int line;            /* Starting line number.  */
  64   unsigned int count;           /* Number of tokens / bytes in expansion.  */
  65   unsigned short paramc;        /* Number of parameters.  */
  66   unsigned int fun_like : 1;    /* If a function-like macro.  */
  67   unsigned int variadic : 1;    /* If a variadic macro.  */
  68   unsigned int syshdr   : 1;    /* If macro defined in system header.  */
  69 };
  70
  71 /* A generic memory buffer, and operations on it.  */
  72 typedef struct _cpp_buff _cpp_buff;
  73 struct _cpp_buff
  74 {
  75   struct _cpp_buff *next;
  76   unsigned char *base, *cur, *limit;
  77 };
  78
  79 extern _cpp_buff *_cpp_get_buff PARAMS ((cpp_reader *, size_t));
  80 extern void _cpp_release_buff PARAMS ((cpp_reader *, _cpp_buff *));
  81 extern void _cpp_extend_buff PARAMS ((cpp_reader *, _cpp_buff **, size_t));
  82 extern _cpp_buff *_cpp_append_extend_buff PARAMS ((cpp_reader *, _cpp_buff *,
  83                                                    size_t));
  84 extern void _cpp_free_buff PARAMS ((_cpp_buff *));
  85 extern unsigned char *_cpp_aligned_alloc PARAMS ((cpp_reader *, size_t));
  86 extern unsigned char *_cpp_unaligned_alloc PARAMS ((cpp_reader *, size_t));
  87
  88 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
  89 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
  90 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
  91
  92 /* List of directories to look for include files in.  */
  93 struct search_path
  94 {
  95   struct search_path *next;
  96
  97   /* NOTE: NAME may not be null terminated for the case of the current
  98      file's directory!  */
  99   const char *name;
 100   unsigned int len;
 101   /* We use these to tell if the directory mentioned here is a duplicate
 102      of an earlier directory on the search path.  */
 103   ino_t ino;
 104   dev_t dev;
 105   /* Non-zero if it is a system include directory.  */
 106   int sysp;
 107   /* Mapping of file names for this directory.  Only used on MS-DOS
 108      and related platforms.  */
 109   struct file_name_map *name_map;
 110 };
 111
 112 /* #include types.  */
 113 enum include_type {IT_INCLUDE, IT_INCLUDE_NEXT, IT_IMPORT, IT_CMDLINE};
 114
 115 union utoken
 116 {
 117   const cpp_token *token;
 118   const cpp_token **ptoken;
 119 };
 120
 121 /* A "run" of tokens; part of a chain of runs.  */
 122 typedef struct tokenrun tokenrun;
 123 struct tokenrun
 124 {
 125   tokenrun *next, *prev;
 126   cpp_token *base, *limit;
 127 };
 128
 129 typedef struct cpp_context cpp_context;
 130 struct cpp_context
 131 {
 132   /* Doubly-linked list.  */
 133   cpp_context *next, *prev;
 134
 135   /* Contexts other than the base context are contiguous tokens.
 136      e.g. macro expansions, expanded argument tokens.  */
 137   union utoken first;
 138   union utoken last;
 139
 140   /* If non-NULL, a buffer used for storage related to this context.
 141      When the context is popped, the buffer is released.  */
 142   _cpp_buff *buff;
 143
 144   /* For a macro context, the macro node, otherwise NULL.  */
 145   cpp_hashnode *macro;
 146
 147   /* True if utoken element is token, else ptoken.  */
 148   bool direct_p;
 149 };
 150
 151 struct lexer_state
 152 {
 153   /* Nonzero if first token on line is CPP_HASH.  */
 154   unsigned char in_directive;
 155
 156   /* True if we are skipping a failed conditional group.  */
 157   unsigned char skipping;
 158
 159   /* Nonzero if in a directive that takes angle-bracketed headers.  */
 160   unsigned char angled_headers;
 161
 162   /* Nonzero to save comments.  Turned off if discard_comments, and in
 163      all directives apart from #define.  */
 164   unsigned char save_comments;
 165
 166   /* Nonzero if we're mid-comment.  */
 167   unsigned char lexing_comment;
 168
 169   /* Nonzero if lexing __VA_ARGS__ is valid.  */
 170   unsigned char va_args_ok;
 171
 172   /* Nonzero if lexing poisoned identifiers is valid.  */
 173   unsigned char poisoned_ok;
 174
 175   /* Nonzero to prevent macro expansion.  */
 176   unsigned char prevent_expansion;
 177
 178   /* Nonzero when parsing arguments to a function-like macro.  */
 179   unsigned char parsing_args;
 180
 181   /* Nonzero to skip evaluating part of an expression.  */
 182   unsigned int skip_eval;
 183 };
 184
 185 /* Special nodes - identifiers with predefined significance.  */
 186 struct spec_nodes
 187 {
 188   cpp_hashnode *n_defined;              /* defined operator */
 189   cpp_hashnode *n_true;                 /* C++ keyword true */
 190   cpp_hashnode *n_false;                /* C++ keyword false */
 191   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 192 };
 193
 194 /* Represents the contents of a file cpplib has read in.  */
 195 struct cpp_buffer
 196 {
 197   const unsigned char *cur;      /* current position */
 198   const unsigned char *backup_to; /* if peeked character is not wanted */
 199   const unsigned char *rlimit; /* end of valid data */
 200   const unsigned char *line_base; /* start of current line */
 201
 202   struct cpp_buffer *prev;
 203
 204   const unsigned char *buf;      /* Entire character buffer.  */
 205
 206   /* Pointer into the include table; non-NULL if this is a file
 207      buffer.  Used for include_next and to record control macros.  */
 208   struct include_file *inc;
 209
 210   /* Value of if_stack at start of this file.
 211      Used to prohibit unmatched #endif (etc) in an include file.  */
 212   struct if_stack *if_stack;
 213
 214   /* Token column position adjustment owing to tabs in whitespace.  */
 215   unsigned int col_adjust;
 216
 217   /* Contains PREV_WHITE and/or AVOID_LPASTE.  */
 218   unsigned char saved_flags;
 219
 220   /* Because of the way the lexer works, -Wtrigraphs can sometimes
 221      warn twice for the same trigraph.  This helps prevent that.  */
 222   const unsigned char *last_Wtrigraphs;
 223
 224   /* True if we have already warned about C++ comments in this file.
 225      The warning happens only for C89 extended mode with -pedantic on,
 226      or for -Wtraditional, and only once per file (otherwise it would
 227      be far too noisy).  */
 228   unsigned char warned_cplusplus_comments;
 229
 230   /* True if we don't process trigraphs and escaped newlines.  True
 231      for preprocessed input, command line directives, and _Pragma
 232      buffers.  */
 233   unsigned char from_stage3;
 234
 235   /* Nonzero means that the directory to start searching for ""
 236      include files has been calculated and stored in "dir" below.  */
 237   unsigned char search_cached;
 238
 239   /* At EOF, a buffer is automatically popped.  If RETURN_AT_EOF is
 240      true, a CPP_EOF token is then returned.  Otherwise, the next
 241      token from the enclosing buffer is returned.  */
 242   bool return_at_eof;
 243
 244   /* The directory of the this buffer's file.  Its NAME member is not
 245      allocated, so we don't need to worry about freeing it.  */
 246   struct search_path dir;
 247
 248   /* Used for buffer overlays by cpptrad.c.  */
 249   const uchar *saved_cur, *saved_rlimit, *saved_line_base;
 250 };
 251
 252 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 253    Applying cpp_get_token repeatedly yields a stream of pre-processor
 254    tokens.  Usually, there is only one cpp_reader object active.  */
 255 struct cpp_reader
 256 {
 257   /* Top of buffer stack.  */
 258   cpp_buffer *buffer;
 259
 260   /* Lexer state.  */
 261   struct lexer_state state;
 262
 263   /* Source line tracking.  */
 264   struct line_maps line_maps;
 265   const struct line_map *map;
 266   unsigned int line;
 267
 268   /* The line of the '#' of the current directive.  */
 269   unsigned int directive_line;
 270
 271   /* Memory buffers.  */
 272   _cpp_buff *a_buff;            /* Aligned permanent storage.  */
 273   _cpp_buff *u_buff;            /* Unaligned permanent storage.  */
 274   _cpp_buff *free_buffs;        /* Free buffer chain.  */
 275
 276   /* Context stack.  */
 277   struct cpp_context base_context;
 278   struct cpp_context *context;
 279
 280   /* If in_directive, the directive if known.  */
 281   const struct directive *directive;
 282
 283   /* The next -include-d file; NULL if they all are done.  If it
 284      points to NULL, the last one is in progress, and
 285      _cpp_maybe_push_include_file has yet to restore the line map.  */
 286   struct pending_option **next_include_file;
 287
 288   /* Multiple inlcude optimisation.  */
 289   const cpp_hashnode *mi_cmacro;
 290   const cpp_hashnode *mi_ind_cmacro;
 291   bool mi_valid;
 292
 293   /* Lexing.  */
 294   cpp_token *cur_token;
 295   tokenrun base_run, *cur_run;
 296   unsigned int lookaheads;
 297
 298   /* Non-zero prevents the lexer from re-using the token runs.  */
 299   unsigned int keep_tokens;
 300
 301   /* Error counter for exit code.  */
 302   unsigned int errors;
 303
 304   /* Buffer to hold macro definition string.  */
 305   unsigned char *macro_buffer;
 306   unsigned int macro_buffer_len;
 307
 308   /* Tree of other included files.  See cppfiles.c.  */
 309   struct splay_tree_s *all_include_files;
 310
 311   /* Current maximum length of directory names in the search path
 312      for include files.  (Altered as we get more of them.)  */
 313   unsigned int max_include_len;
 314
 315   /* Date and time tokens.  Calculated together if either is requested.  */
 316   cpp_token date;
 317   cpp_token time;
 318
 319   /* EOF token, and a token forcing paste avoidance.  */
 320   cpp_token avoid_paste;
 321   cpp_token eof;
 322
 323   /* Opaque handle to the dependencies of mkdeps.c.  Used by -M etc.  */
 324   struct deps *deps;
 325
 326   /* Obstack holding all macro hash nodes.  This never shrinks.
 327      See cpphash.c */
 328   struct obstack hash_ob;
 329
 330   /* Obstack holding buffer and conditional structures.  This is a
 331      real stack.  See cpplib.c.  */
 332   struct obstack buffer_ob;
 333
 334   /* Pragma table - dynamic, because a library user can add to the
 335      list of recognized pragmas.  */
 336   struct pragma_entry *pragmas;
 337
 338   /* Call backs.  */
 339   struct cpp_callbacks cb;
 340
 341   /* Identifier hash table.  */
 342   struct ht *hash_table;
 343
 344   /* Expression parser stack.  */
 345   struct op *op_stack, *op_limit;
 346
 347   /* User visible options.  */
 348   struct cpp_options opts;
 349
 350   /* Special nodes - identifiers with predefined significance to the
 351      preprocessor.  */
 352   struct spec_nodes spec_nodes;
 353
 354   /* Whether to print our version number.  Done this way so
 355      we don't get it twice for -v -version.  */
 356   unsigned char print_version;
 357
 358   /* Whether cpplib owns the hashtable.  */
 359   unsigned char our_hashtable;
 360
 361   /* Traditional preprocessing output buffer.  */
 362   uchar *trad_out_base, *trad_out_limit;
 363   uchar *trad_out_cur;
 364   unsigned int trad_line;
 365 };
 366
 367 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 368    If the definition of `numchar' looks odd to you, please look up the
 369    definition of a pp-number in the C standard [section 6.4.8 of C99].
 370
 371    In the unlikely event that characters other than \r and \n enter
 372    the set is_vspace, the macro handle_newline() in cpplex.c must be
 373    updated.  */
 374 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 375
 376 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 377 #define is_numchar(x)   ISIDNUM(x)
 378 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 379 #define is_numstart(x)  ISDIGIT(x)
 380 #define is_hspace(x)    ISBLANK(x)
 381 #define is_vspace(x)    IS_VSPACE(x)
 382 #define is_nvspace(x)   IS_NVSPACE(x)
 383 #define is_space(x)     IS_SPACE_OR_NUL(x)
 384
 385 /* This table is constant if it can be initialized at compile time,
 386    which is the case if cpp was compiled with GCC >=2.7, or another
 387    compiler that supports C99.  */
 388 #if HAVE_DESIGNATED_INITIALIZERS
 389 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 390 #else
 391 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 392 #endif
 393
 394 /* Macros.  */
 395
 396 #define CPP_PRINT_DEPS(PFILE) CPP_OPTION (PFILE, print_deps)
 397 #define CPP_IN_SYSTEM_HEADER(PFILE) ((PFILE)->map && (PFILE)->map->sysp)
 398 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
 399 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
 400
 401 /* In cpperror.c  */
 402 extern int _cpp_begin_message PARAMS ((cpp_reader *, int,
 403                                        unsigned int, unsigned int));
 404
 405 /* In cppmacro.c */
 406 extern void _cpp_free_definition        PARAMS ((cpp_hashnode *));
 407 extern int _cpp_create_definition       PARAMS ((cpp_reader *, cpp_hashnode *));
 408 extern void _cpp_pop_context            PARAMS ((cpp_reader *));
 409
 410 /* In cpphash.c */
 411 extern void _cpp_init_hashtable         PARAMS ((cpp_reader *, hash_table *));
 412 extern void _cpp_destroy_hashtable      PARAMS ((cpp_reader *));
 413
 414 /* In cppfiles.c */
 415 extern void _cpp_fake_include           PARAMS ((cpp_reader *, const char *));
 416 extern void _cpp_never_reread           PARAMS ((struct include_file *));
 417 extern char *_cpp_simplify_pathname     PARAMS ((char *));
 418 extern bool _cpp_read_file              PARAMS ((cpp_reader *, const char *));
 419 extern bool _cpp_execute_include        PARAMS ((cpp_reader *,
 420                                                  const cpp_token *,
 421                                                  enum include_type));
 422 extern int _cpp_compare_file_date       PARAMS ((cpp_reader *,
 423                                                  const cpp_token *));
 424 extern void _cpp_report_missing_guards  PARAMS ((cpp_reader *));
 425 extern void _cpp_init_includes          PARAMS ((cpp_reader *));
 426 extern void _cpp_cleanup_includes       PARAMS ((cpp_reader *));
 427 extern void _cpp_pop_file_buffer        PARAMS ((cpp_reader *,
 428                                                  struct include_file *));
 429
 430 /* In cppexp.c */
 431 extern bool _cpp_parse_expr             PARAMS ((cpp_reader *));
 432 extern struct op *_cpp_expand_op_stack  PARAMS ((cpp_reader *));
 433
 434 /* In cpplex.c */
 435 extern cpp_token *_cpp_temp_token       PARAMS ((cpp_reader *));
 436 extern const cpp_token *_cpp_lex_token  PARAMS ((cpp_reader *));
 437 extern cpp_token *_cpp_lex_direct       PARAMS ((cpp_reader *));
 438 extern int _cpp_equiv_tokens            PARAMS ((const cpp_token *,
 439                                                  const cpp_token *));
 440 extern void _cpp_init_tokenrun          PARAMS ((tokenrun *, unsigned int));
 441
 442 /* In cppinit.c.  */
 443 extern void _cpp_maybe_push_include_file PARAMS ((cpp_reader *));
 444
 445 /* In cpplib.c */
 446 extern int _cpp_test_assertion PARAMS ((cpp_reader *, unsigned int *));
 447 extern int _cpp_handle_directive PARAMS ((cpp_reader *, int));
 448 extern void _cpp_define_builtin PARAMS ((cpp_reader *, const char *));
 449 extern void _cpp_do__Pragma     PARAMS ((cpp_reader *));
 450 extern void _cpp_init_directives PARAMS ((cpp_reader *));
 451 extern void _cpp_init_internal_pragmas PARAMS ((cpp_reader *));
 452 extern void _cpp_do_file_change PARAMS ((cpp_reader *, enum lc_reason,
 453                                          const char *,
 454                                          unsigned int, unsigned int));
 455 extern void _cpp_pop_buffer PARAMS ((cpp_reader *));
 456
 457 /* In cpptrad.c.  */
 458 extern bool _cpp_read_logical_line_trad PARAMS ((cpp_reader *));
 459 extern void _cpp_overlay_buffer PARAMS ((cpp_reader *pfile, const uchar *,
 460                                          size_t));
 461
 462 /* Utility routines and macros.  */
 463 #define DSC(str) (const uchar *)str, sizeof str - 1
 464 #define xnew(T)         (T *) xmalloc (sizeof(T))
 465 #define xcnew(T)        (T *) xcalloc (1, sizeof(T))
 466 #define xnewvec(T, N)   (T *) xmalloc (sizeof(T) * (N))
 467 #define xcnewvec(T, N)  (T *) xcalloc (N, sizeof(T))
 468 #define xobnew(O, T)    (T *) obstack_alloc (O, sizeof(T))
 469
 470 /* These are inline functions instead of macros so we can get type
 471    checking.  */
 472 static inline int ustrcmp       PARAMS ((const uchar *, const uchar *));
 473 static inline int ustrncmp      PARAMS ((const uchar *, const uchar *,
 474                                          size_t));
 475 static inline size_t ustrlen    PARAMS ((const uchar *));
 476 static inline uchar *uxstrdup   PARAMS ((const uchar *));
 477 static inline uchar *ustrchr    PARAMS ((const uchar *, int));
 478 static inline int ufputs        PARAMS ((const uchar *, FILE *));
 479
 480 static inline int
 481 ustrcmp (s1, s2)
 482      const uchar *s1, *s2;
 483 {
 484   return strcmp ((const char *)s1, (const char *)s2);
 485 }
 486
 487 static inline int
 488 ustrncmp (s1, s2, n)
 489      const uchar *s1, *s2;
 490      size_t n;
 491 {
 492   return strncmp ((const char *)s1, (const char *)s2, n);
 493 }
 494
 495 static inline size_t
 496 ustrlen (s1)
 497      const uchar *s1;
 498 {
 499   return strlen ((const char *)s1);
 500 }
 501
 502 static inline uchar *
 503 uxstrdup (s1)
 504      const uchar *s1;
 505 {
 506   return (uchar *) xstrdup ((const char *)s1);
 507 }
 508
 509 static inline uchar *
 510 ustrchr (s1, c)
 511      const uchar *s1;
 512      int c;
 513 {
 514   return (uchar *) strchr ((const char *)s1, c);
 515 }
 516
 517 static inline int
 518 ufputs (s, f)
 519      const uchar *s;
 520      FILE *f;
 521 {
 522   return fputs ((const char *)s, f);
 523 }
 524
 525 #endif /* ! GCC_CPPHASH_H */