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
  30 /* Test if a sign is valid within a preprocessing number.  */
  31 #define VALID_SIGN(c, prevc) \
  32   (((c) == '+' || (c) == '-') && \
  33    ((prevc) == 'e' || (prevc) == 'E' \
  34     || (((prevc) == 'p' || (prevc) == 'P') \
  35         && CPP_OPTION (pfile, extended_numbers))))
  36
  37 #define CPP_OPTION(PFILE, OPTION) ((PFILE)->opts.OPTION)
  38 #define CPP_BUFFER(PFILE) ((PFILE)->buffer)
  39 #define CPP_BUF_COLUMN(BUF, CUR) ((CUR) - (BUF)->line_base + (BUF)->col_adjust)
  40 #define CPP_BUF_COL(BUF) CPP_BUF_COLUMN(BUF, (BUF)->cur)
  41
  42 /* Maximum nesting of cpp_buffers.  We use a static limit, partly for
  43    efficiency, and partly to limit runaway recursion.  */
  44 #define CPP_STACK_MAX 200
  45
  46 /* A generic memory buffer, and operations on it.  */
  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 /* A "run" of tokens; part of a chain of runs.  */
  97 typedef struct tokenrun tokenrun;
  98 struct tokenrun
  99 {
 100   tokenrun *next, *prev;
 101   cpp_token *base, *limit;
 102 };
 103
 104 typedef struct cpp_context cpp_context;
 105 struct cpp_context
 106 {
 107   /* Doubly-linked list.  */
 108   cpp_context *next, *prev;
 109
 110   /* Contexts other than the base context are contiguous tokens.
 111      e.g. macro expansions, expanded argument tokens.  */
 112   union utoken first;
 113   union utoken last;
 114
 115   /* If non-NULL, a buffer used for storage related to this context.
 116      When the context is popped, the buffer is released.  */
 117   _cpp_buff *buff;
 118
 119   /* For a macro context, the macro node, otherwise NULL.  */
 120   cpp_hashnode *macro;
 121
 122   /* True if utoken element is token, else ptoken.  */
 123   bool direct_p;
 124 };
 125
 126 struct lexer_state
 127 {
 128   /* Nonzero if first token on line is CPP_HASH.  */
 129   unsigned char in_directive;
 130
 131   /* True if we are skipping a failed conditional group.  */
 132   unsigned char skipping;
 133
 134   /* Nonzero if in a directive that takes angle-bracketed headers.  */
 135   unsigned char angled_headers;
 136
 137   /* Nonzero to save comments.  Turned off if discard_comments, and in
 138      all directives apart from #define.  */
 139   unsigned char save_comments;
 140
 141   /* Nonzero if we're mid-comment.  */
 142   unsigned char lexing_comment;
 143
 144   /* Nonzero if lexing __VA_ARGS__ is valid.  */
 145   unsigned char va_args_ok;
 146
 147   /* Nonzero if lexing poisoned identifiers is valid.  */
 148   unsigned char poisoned_ok;
 149
 150   /* Nonzero to prevent macro expansion.  */
 151   unsigned char prevent_expansion;
 152
 153   /* Nonzero when parsing arguments to a function-like macro.  */
 154   unsigned char parsing_args;
 155 };
 156
 157 /* Special nodes - identifiers with predefined significance.  */
 158 struct spec_nodes
 159 {
 160   cpp_hashnode *n_defined;              /* defined operator */
 161   cpp_hashnode *n_true;                 /* C++ keyword true */
 162   cpp_hashnode *n_false;                /* C++ keyword false */
 163   cpp_hashnode *n__STRICT_ANSI__;       /* STDC_0_IN_SYSTEM_HEADERS */
 164   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 165 };
 166
 167 /* Represents the contents of a file cpplib has read in.  */
 168 struct cpp_buffer
 169 {
 170   const unsigned char *cur;      /* current position */
 171   const unsigned char *backup_to; /* if peeked character is not wanted */
 172   const unsigned char *rlimit; /* end of valid data */
 173   const unsigned char *line_base; /* start of current line */
 174
 175   struct cpp_buffer *prev;
 176
 177   const unsigned char *buf;      /* Entire character buffer.  */
 178
 179   /* Pointer into the include table; non-NULL if this is a file
 180      buffer.  Used for include_next and to record control macros.  */
 181   struct include_file *inc;
 182
 183   /* Value of if_stack at start of this file.
 184      Used to prohibit unmatched #endif (etc) in an include file.  */
 185   struct if_stack *if_stack;
 186
 187   /* Token column position adjustment owing to tabs in whitespace.  */
 188   unsigned int col_adjust;
 189
 190   /* Contains PREV_WHITE and/or AVOID_LPASTE.  */
 191   unsigned char saved_flags;
 192
 193   /* Because of the way the lexer works, -Wtrigraphs can sometimes
 194      warn twice for the same trigraph.  This helps prevent that.  */
 195   const unsigned char *last_Wtrigraphs;
 196
 197   /* True if we have already warned about C++ comments in this file.
 198      The warning happens only for C89 extended mode with -pedantic on,
 199      or for -Wtraditional, and only once per file (otherwise it would
 200      be far too noisy).  */
 201   unsigned char warned_cplusplus_comments;
 202
 203   /* True if we don't process trigraphs and escaped newlines.  True
 204      for preprocessed input, command line directives, and _Pragma
 205      buffers.  */
 206   unsigned char from_stage3;
 207
 208   /* Nonzero means that the directory to start searching for ""
 209      include files has been calculated and stored in "dir" below.  */
 210   unsigned char search_cached;
 211
 212   /* At EOF, a buffer is automatically popped.  If RETURN_AT_EOF is
 213      true, a CPP_EOF token is then returned.  Otherwise, the next
 214      token from the enclosing buffer is returned.  */
 215   bool return_at_eof;
 216
 217   /* The directory of the this buffer's file.  Its NAME member is not
 218      allocated, so we don't need to worry about freeing it.  */
 219   struct search_path dir;
 220 };
 221
 222 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 223    Applying cpp_get_token repeatedly yields a stream of pre-processor
 224    tokens.  Usually, there is only one cpp_reader object active.  */
 225 struct cpp_reader
 226 {
 227   /* Top of buffer stack.  */
 228   cpp_buffer *buffer;
 229
 230   /* Lexer state.  */
 231   struct lexer_state state;
 232
 233   /* Source line tracking.  */
 234   struct line_maps line_maps;
 235   const struct line_map *map;
 236   unsigned int line;
 237
 238   /* The line of the '#' of the current directive.  */
 239   unsigned int directive_line;
 240
 241   /* Memory buffers.  */
 242   _cpp_buff *a_buff;            /* Aligned permanent storage.  */
 243   _cpp_buff *u_buff;            /* Unaligned permanent storage.  */
 244   _cpp_buff *free_buffs;        /* Free buffer chain.  */
 245
 246   /* Context stack.  */
 247   struct cpp_context base_context;
 248   struct cpp_context *context;
 249
 250   /* If in_directive, the directive if known.  */
 251   const struct directive *directive;
 252
 253   /* Multiple inlcude optimisation.  */
 254   const cpp_hashnode *mi_cmacro;
 255   const cpp_hashnode *mi_ind_cmacro;
 256   bool mi_valid;
 257
 258   /* Lexing.  */
 259   cpp_token *cur_token;
 260   tokenrun base_run, *cur_run;
 261   unsigned int lookaheads;
 262
 263   /* Non-zero prevents the lexer from re-using the token runs.  */
 264   unsigned int keep_tokens;
 265
 266   /* Error counter for exit code.  */
 267   unsigned int errors;
 268
 269   /* Line and column where a newline was first seen in a string
 270      constant (multi-line strings).  */
 271   unsigned int mls_line;
 272   unsigned int mls_col;
 273
 274   /* Buffer to hold macro definition string.  */
 275   unsigned char *macro_buffer;
 276   unsigned int macro_buffer_len;
 277
 278   /* Tree of other included files.  See cppfiles.c.  */
 279   struct splay_tree_s *all_include_files;
 280
 281   /* Current maximum length of directory names in the search path
 282      for include files.  (Altered as we get more of them.)  */
 283   unsigned int max_include_len;
 284
 285   /* Date and time tokens.  Calculated together if either is requested.  */
 286   cpp_token date;
 287   cpp_token time;
 288
 289   /* EOF token, and a token forcing paste avoidance.  */
 290   cpp_token avoid_paste;
 291   cpp_token eof;
 292
 293   /* Opaque handle to the dependencies of mkdeps.c.  Used by -M etc.  */
 294   struct deps *deps;
 295
 296   /* Obstack holding all macro hash nodes.  This never shrinks.
 297      See cpphash.c */
 298   struct obstack hash_ob;
 299
 300   /* Obstack holding buffer and conditional structures.  This is a
 301      real stack.  See cpplib.c.  */
 302   struct obstack buffer_ob;
 303
 304   /* Pragma table - dynamic, because a library user can add to the
 305      list of recognized pragmas.  */
 306   struct pragma_entry *pragmas;
 307
 308   /* Call backs.  */
 309   struct cpp_callbacks cb;
 310
 311   /* Identifier hash table.  */
 312   struct ht *hash_table;
 313
 314   /* User visible options.  */
 315   struct cpp_options opts;
 316
 317   /* Special nodes - identifiers with predefined significance to the
 318      preprocessor.  */
 319   struct spec_nodes spec_nodes;
 320
 321   /* Whether to print our version number.  Done this way so
 322      we don't get it twice for -v -version.  */
 323   unsigned char print_version;
 324
 325   /* Whether cpplib owns the hashtable.  */
 326   unsigned char our_hashtable;
 327 };
 328
 329 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 330    If the definition of `numchar' looks odd to you, please look up the
 331    definition of a pp-number in the C standard [section 6.4.8 of C99].
 332
 333    In the unlikely event that characters other than \r and \n enter
 334    the set is_vspace, the macro handle_newline() in cpplex.c must be
 335    updated.  */
 336 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 337
 338 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 339 #define is_numchar(x)   ISIDNUM(x)
 340 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 341 #define is_numstart(x)  ISDIGIT(x)
 342 #define is_hspace(x)    ISBLANK(x)
 343 #define is_vspace(x)    IS_VSPACE(x)
 344 #define is_nvspace(x)   IS_NVSPACE(x)
 345 #define is_space(x)     IS_SPACE_OR_NUL(x)
 346
 347 /* This table is constant if it can be initialized at compile time,
 348    which is the case if cpp was compiled with GCC >=2.7, or another
 349    compiler that supports C99.  */
 350 #if HAVE_DESIGNATED_INITIALIZERS
 351 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 352 #else
 353 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 354 #endif
 355
 356 /* Macros.  */
 357
 358 #define CPP_PRINT_DEPS(PFILE) CPP_OPTION (PFILE, print_deps)
 359 #define CPP_IN_SYSTEM_HEADER(PFILE) ((PFILE)->map && (PFILE)->map->sysp)
 360 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
 361 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
 362
 363 /* In cpperror.c  */
 364 enum error_type { WARNING = 0, WARNING_SYSHDR, PEDWARN, ERROR, FATAL, ICE };
 365 extern int _cpp_begin_message PARAMS ((cpp_reader *, enum error_type,
 366                                        unsigned int, unsigned int));
 367
 368 /* In cppmacro.c */
 369 extern void _cpp_free_definition        PARAMS ((cpp_hashnode *));
 370 extern int _cpp_create_definition       PARAMS ((cpp_reader *, cpp_hashnode *));
 371 extern void _cpp_pop_context            PARAMS ((cpp_reader *));
 372
 373 /* In cpphash.c */
 374 extern void _cpp_init_hashtable         PARAMS ((cpp_reader *, hash_table *));
 375 extern void _cpp_destroy_hashtable      PARAMS ((cpp_reader *));
 376
 377 /* In cppfiles.c */
 378 extern void _cpp_fake_include           PARAMS ((cpp_reader *, const char *));
 379 extern void _cpp_never_reread           PARAMS ((struct include_file *));
 380 extern char *_cpp_simplify_pathname     PARAMS ((char *));
 381 extern bool _cpp_read_file              PARAMS ((cpp_reader *, const char *));
 382 extern bool _cpp_execute_include        PARAMS ((cpp_reader *,
 383                                                  const cpp_token *,
 384                                                  enum include_type));
 385 extern int _cpp_compare_file_date       PARAMS ((cpp_reader *,
 386                                                  const cpp_token *));
 387 extern void _cpp_report_missing_guards  PARAMS ((cpp_reader *));
 388 extern void _cpp_init_includes          PARAMS ((cpp_reader *));
 389 extern void _cpp_cleanup_includes       PARAMS ((cpp_reader *));
 390 extern bool _cpp_pop_file_buffer        PARAMS ((cpp_reader *,
 391                                                  struct include_file *));
 392
 393 /* In cppexp.c */
 394 extern int _cpp_parse_expr              PARAMS ((cpp_reader *));
 395
 396 /* In cpplex.c */
 397 extern cpp_token *_cpp_temp_token       PARAMS ((cpp_reader *));
 398 extern const cpp_token *_cpp_lex_token  PARAMS ((cpp_reader *));
 399 extern cpp_token *_cpp_lex_direct       PARAMS ((cpp_reader *));
 400 extern int _cpp_equiv_tokens            PARAMS ((const cpp_token *,
 401                                                  const cpp_token *));
 402 extern void _cpp_init_tokenrun          PARAMS ((tokenrun *, unsigned int));
 403
 404 /* In cppinit.c.  */
 405 extern bool _cpp_push_next_buffer       PARAMS ((cpp_reader *));
 406
 407 /* In cpplib.c */
 408 extern int _cpp_test_assertion PARAMS ((cpp_reader *, int *));
 409 extern int _cpp_handle_directive PARAMS ((cpp_reader *, int));
 410 extern void _cpp_define_builtin PARAMS ((cpp_reader *, const char *));
 411 extern void _cpp_do__Pragma     PARAMS ((cpp_reader *));
 412 extern void _cpp_init_directives PARAMS ((cpp_reader *));
 413 extern void _cpp_init_internal_pragmas PARAMS ((cpp_reader *));
 414 extern void _cpp_do_file_change PARAMS ((cpp_reader *, enum lc_reason,
 415                                          const char *,
 416                                          unsigned int, unsigned int));
 417 extern void _cpp_pop_buffer PARAMS ((cpp_reader *));
 418
 419 /* Utility routines and macros.  */
 420 #define DSC(str) (const U_CHAR *)str, sizeof str - 1
 421 #define xnew(T)         (T *) xmalloc (sizeof(T))
 422 #define xcnew(T)        (T *) xcalloc (1, sizeof(T))
 423 #define xnewvec(T, N)   (T *) xmalloc (sizeof(T) * (N))
 424 #define xcnewvec(T, N)  (T *) xcalloc (N, sizeof(T))
 425 #define xobnew(O, T)    (T *) obstack_alloc (O, sizeof(T))
 426
 427 /* These are inline functions instead of macros so we can get type
 428    checking.  */
 429 typedef unsigned char U_CHAR;
 430 #define U (const U_CHAR *)  /* Intended use: U"string" */
 431
 432 static inline int ustrcmp       PARAMS ((const U_CHAR *, const U_CHAR *));
 433 static inline int ustrncmp      PARAMS ((const U_CHAR *, const U_CHAR *,
 434                                          size_t));
 435 static inline size_t ustrlen    PARAMS ((const U_CHAR *));
 436 static inline U_CHAR *uxstrdup  PARAMS ((const U_CHAR *));
 437 static inline U_CHAR *ustrchr   PARAMS ((const U_CHAR *, int));
 438 static inline int ufputs        PARAMS ((const U_CHAR *, FILE *));
 439
 440 static inline int
 441 ustrcmp (s1, s2)
 442      const U_CHAR *s1, *s2;
 443 {
 444   return strcmp ((const char *)s1, (const char *)s2);
 445 }
 446
 447 static inline int
 448 ustrncmp (s1, s2, n)
 449      const U_CHAR *s1, *s2;
 450      size_t n;
 451 {
 452   return strncmp ((const char *)s1, (const char *)s2, n);
 453 }
 454
 455 static inline size_t
 456 ustrlen (s1)
 457      const U_CHAR *s1;
 458 {
 459   return strlen ((const char *)s1);
 460 }
 461
 462 static inline U_CHAR *
 463 uxstrdup (s1)
 464      const U_CHAR *s1;
 465 {
 466   return (U_CHAR *) xstrdup ((const char *)s1);
 467 }
 468
 469 static inline U_CHAR *
 470 ustrchr (s1, c)
 471      const U_CHAR *s1;
 472      int c;
 473 {
 474   return (U_CHAR *) strchr ((const char *)s1, c);
 475 }
 476
 477 static inline int
 478 ufputs (s, f)
 479      const U_CHAR *s;
 480      FILE *f;
 481 {
 482   return fputs ((const char *)s, f);
 483 }
 484
 485 #endif /* ! GCC_CPPHASH_H */