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 /* Host alignment handling.  */
  54 struct dummy
  55 {
  56   char c;
  57   union
  58   {
  59     double d;
  60     int *p;
  61   } u;
  62 };
  63
  64 #define DEFAULT_ALIGNMENT offsetof (struct dummy, u)
  65 #define CPP_ALIGN2(size, align) (((size) + ((align) - 1)) & ~((align) - 1))
  66 #define CPP_ALIGN(size) CPP_ALIGN2 (size, DEFAULT_ALIGNMENT)
  67
  68 /* Each macro definition is recorded in a cpp_macro structure.
  69    Variadic macros cannot occur with traditional cpp.  */
  70 struct cpp_macro
  71 {
  72   cpp_hashnode **params;        /* Parameters, if any.  */
  73   union
  74   {
  75     cpp_token *tokens;          /* Tokens of replacement list (ISO).  */
  76     const uchar *text;          /* Expansion text (traditional).  */
  77   } exp;
  78   unsigned int line;            /* Starting line number.  */
  79   unsigned int count;           /* Number of tokens / bytes in expansion.  */
  80   unsigned short paramc;        /* Number of parameters.  */
  81   unsigned int fun_like : 1;    /* If a function-like macro.  */
  82   unsigned int variadic : 1;    /* If a variadic macro.  */
  83   unsigned int syshdr   : 1;    /* If macro defined in system header.  */
  84 };
  85
  86 /* A generic memory buffer, and operations on it.  */
  87 typedef struct _cpp_buff _cpp_buff;
  88 struct _cpp_buff
  89 {
  90   struct _cpp_buff *next;
  91   unsigned char *base, *cur, *limit;
  92 };
  93
  94 extern _cpp_buff *_cpp_get_buff PARAMS ((cpp_reader *, size_t));
  95 extern void _cpp_release_buff PARAMS ((cpp_reader *, _cpp_buff *));
  96 extern void _cpp_extend_buff PARAMS ((cpp_reader *, _cpp_buff **, size_t));
  97 extern _cpp_buff *_cpp_append_extend_buff PARAMS ((cpp_reader *, _cpp_buff *,
  98                                                    size_t));
  99 extern void _cpp_free_buff PARAMS ((_cpp_buff *));
 100 extern unsigned char *_cpp_aligned_alloc PARAMS ((cpp_reader *, size_t));
 101 extern unsigned char *_cpp_unaligned_alloc PARAMS ((cpp_reader *, size_t));
 102
 103 #define BUFF_ROOM(BUFF) (size_t) ((BUFF)->limit - (BUFF)->cur)
 104 #define BUFF_FRONT(BUFF) ((BUFF)->cur)
 105 #define BUFF_LIMIT(BUFF) ((BUFF)->limit)
 106
 107 /* List of directories to look for include files in.  */
 108 struct search_path
 109 {
 110   struct search_path *next;
 111
 112   /* NOTE: NAME may not be null terminated for the case of the current
 113      file's directory!  */
 114   const char *name;
 115   unsigned int len;
 116   /* We use these to tell if the directory mentioned here is a duplicate
 117      of an earlier directory on the search path.  */
 118   ino_t ino;
 119   dev_t dev;
 120   /* Non-zero if it is a system include directory.  */
 121   int sysp;
 122   /* Mapping of file names for this directory.  Only used on MS-DOS
 123      and related platforms.  */
 124   struct file_name_map *name_map;
 125 };
 126
 127 /* #include types.  */
 128 enum include_type {IT_INCLUDE, IT_INCLUDE_NEXT, IT_IMPORT, IT_CMDLINE};
 129
 130 union utoken
 131 {
 132   const cpp_token *token;
 133   const cpp_token **ptoken;
 134 };
 135
 136 /* A "run" of tokens; part of a chain of runs.  */
 137 typedef struct tokenrun tokenrun;
 138 struct tokenrun
 139 {
 140   tokenrun *next, *prev;
 141   cpp_token *base, *limit;
 142 };
 143
 144 /* Accessor macros for struct cpp_context.  */
 145 #define FIRST(c) (c->u.iso.first)
 146 #define LAST(c) (c->u.iso.last)
 147 #define CUR(c) (c->u.trad.cur)
 148 #define RLIMIT(c) (c->u.trad.rlimit)
 149
 150 typedef struct cpp_context cpp_context;
 151 struct cpp_context
 152 {
 153   /* Doubly-linked list.  */
 154   cpp_context *next, *prev;
 155
 156   union
 157   {
 158     /* For ISO macro expansion.  Contexts other than the base context
 159        are contiguous tokens.  e.g. macro expansions, expanded
 160        argument tokens.  */
 161     struct
 162     {
 163       union utoken first;
 164       union utoken last;
 165     } iso;
 166
 167     /* For traditional macro expansion.  */
 168     struct
 169     {
 170       const uchar *cur;
 171       const uchar *rlimit;
 172     } trad;
 173   } u;
 174
 175   /* If non-NULL, a buffer used for storage related to this context.
 176      When the context is popped, the buffer is released.  */
 177   _cpp_buff *buff;
 178
 179   /* For a macro context, the macro node, otherwise NULL.  */
 180   cpp_hashnode *macro;
 181
 182   /* True if utoken element is token, else ptoken.  */
 183   bool direct_p;
 184 };
 185
 186 struct lexer_state
 187 {
 188   /* Nonzero if first token on line is CPP_HASH.  */
 189   unsigned char in_directive;
 190
 191   /* True if we are skipping a failed conditional group.  */
 192   unsigned char skipping;
 193
 194   /* Nonzero if in a directive that takes angle-bracketed headers.  */
 195   unsigned char angled_headers;
 196
 197   /* Nonzero to save comments.  Turned off if discard_comments, and in
 198      all directives apart from #define.  */
 199   unsigned char save_comments;
 200
 201   /* Nonzero if we're mid-comment.  */
 202   unsigned char lexing_comment;
 203
 204   /* Nonzero if lexing __VA_ARGS__ is valid.  */
 205   unsigned char va_args_ok;
 206
 207   /* Nonzero if lexing poisoned identifiers is valid.  */
 208   unsigned char poisoned_ok;
 209
 210   /* Nonzero to prevent macro expansion.  */
 211   unsigned char prevent_expansion;
 212
 213   /* Nonzero when parsing arguments to a function-like macro.  */
 214   unsigned char parsing_args;
 215
 216   /* Nonzero to skip evaluating part of an expression.  */
 217   unsigned int skip_eval;
 218 };
 219
 220 /* Special nodes - identifiers with predefined significance.  */
 221 struct spec_nodes
 222 {
 223   cpp_hashnode *n_defined;              /* defined operator */
 224   cpp_hashnode *n_true;                 /* C++ keyword true */
 225   cpp_hashnode *n_false;                /* C++ keyword false */
 226   cpp_hashnode *n__VA_ARGS__;           /* C99 vararg macros */
 227 };
 228
 229 /* Represents the contents of a file cpplib has read in.  */
 230 struct cpp_buffer
 231 {
 232   const unsigned char *cur;      /* current position */
 233   const unsigned char *backup_to; /* if peeked character is not wanted */
 234   const unsigned char *rlimit; /* end of valid data */
 235   const unsigned char *line_base; /* start of current line */
 236
 237   struct cpp_buffer *prev;
 238
 239   const unsigned char *buf;      /* Entire character buffer.  */
 240
 241   /* Pointer into the include table; non-NULL if this is a file
 242      buffer.  Used for include_next and to record control macros.  */
 243   struct include_file *inc;
 244
 245   /* Value of if_stack at start of this file.
 246      Used to prohibit unmatched #endif (etc) in an include file.  */
 247   struct if_stack *if_stack;
 248
 249   /* Token column position adjustment owing to tabs in whitespace.  */
 250   unsigned int col_adjust;
 251
 252   /* Contains PREV_WHITE and/or AVOID_LPASTE.  */
 253   unsigned char saved_flags;
 254
 255   /* Because of the way the lexer works, -Wtrigraphs can sometimes
 256      warn twice for the same trigraph.  This helps prevent that.  */
 257   const unsigned char *last_Wtrigraphs;
 258
 259   /* True if we have already warned about C++ comments in this file.
 260      The warning happens only for C89 extended mode with -pedantic on,
 261      or for -Wtraditional, and only once per file (otherwise it would
 262      be far too noisy).  */
 263   unsigned char warned_cplusplus_comments;
 264
 265   /* True if we don't process trigraphs and escaped newlines.  True
 266      for preprocessed input, command line directives, and _Pragma
 267      buffers.  */
 268   unsigned char from_stage3;
 269
 270   /* Nonzero means that the directory to start searching for ""
 271      include files has been calculated and stored in "dir" below.  */
 272   unsigned char search_cached;
 273
 274   /* At EOF, a buffer is automatically popped.  If RETURN_AT_EOF is
 275      true, a CPP_EOF token is then returned.  Otherwise, the next
 276      token from the enclosing buffer is returned.  */
 277   bool return_at_eof;
 278
 279   /* The directory of the this buffer's file.  Its NAME member is not
 280      allocated, so we don't need to worry about freeing it.  */
 281   struct search_path dir;
 282
 283   /* Used for buffer overlays by cpptrad.c.  */
 284   const uchar *saved_cur, *saved_rlimit, *saved_line_base;
 285 };
 286
 287 /* A cpp_reader encapsulates the "state" of a pre-processor run.
 288    Applying cpp_get_token repeatedly yields a stream of pre-processor
 289    tokens.  Usually, there is only one cpp_reader object active.  */
 290 struct cpp_reader
 291 {
 292   /* Top of buffer stack.  */
 293   cpp_buffer *buffer;
 294
 295   /* Lexer state.  */
 296   struct lexer_state state;
 297
 298   /* Source line tracking.  */
 299   struct line_maps line_maps;
 300   const struct line_map *map;
 301   unsigned int line;
 302
 303   /* The line of the '#' of the current directive.  */
 304   unsigned int directive_line;
 305
 306   /* Memory buffers.  */
 307   _cpp_buff *a_buff;            /* Aligned permanent storage.  */
 308   _cpp_buff *u_buff;            /* Unaligned permanent storage.  */
 309   _cpp_buff *free_buffs;        /* Free buffer chain.  */
 310
 311   /* Context stack.  */
 312   struct cpp_context base_context;
 313   struct cpp_context *context;
 314
 315   /* If in_directive, the directive if known.  */
 316   const struct directive *directive;
 317
 318   /* The next -include-d file; NULL if they all are done.  If it
 319      points to NULL, the last one is in progress, and
 320      _cpp_maybe_push_include_file has yet to restore the line map.  */
 321   struct pending_option **next_include_file;
 322
 323   /* Multiple inlcude optimisation.  */
 324   const cpp_hashnode *mi_cmacro;
 325   const cpp_hashnode *mi_ind_cmacro;
 326   bool mi_valid;
 327
 328   /* Lexing.  */
 329   cpp_token *cur_token;
 330   tokenrun base_run, *cur_run;
 331   unsigned int lookaheads;
 332
 333   /* Non-zero prevents the lexer from re-using the token runs.  */
 334   unsigned int keep_tokens;
 335
 336   /* Error counter for exit code.  */
 337   unsigned int errors;
 338
 339   /* Buffer to hold macro definition string.  */
 340   unsigned char *macro_buffer;
 341   unsigned int macro_buffer_len;
 342
 343   /* Tree of other included files.  See cppfiles.c.  */
 344   struct splay_tree_s *all_include_files;
 345
 346   /* Current maximum length of directory names in the search path
 347      for include files.  (Altered as we get more of them.)  */
 348   unsigned int max_include_len;
 349
 350   /* Date and time tokens.  Calculated together if either is requested.  */
 351   cpp_token date;
 352   cpp_token time;
 353
 354   /* EOF token, and a token forcing paste avoidance.  */
 355   cpp_token avoid_paste;
 356   cpp_token eof;
 357
 358   /* Opaque handle to the dependencies of mkdeps.c.  Used by -M etc.  */
 359   struct deps *deps;
 360
 361   /* Obstack holding all macro hash nodes.  This never shrinks.
 362      See cpphash.c */
 363   struct obstack hash_ob;
 364
 365   /* Obstack holding buffer and conditional structures.  This is a
 366      real stack.  See cpplib.c.  */
 367   struct obstack buffer_ob;
 368
 369   /* Pragma table - dynamic, because a library user can add to the
 370      list of recognized pragmas.  */
 371   struct pragma_entry *pragmas;
 372
 373   /* Call backs.  */
 374   struct cpp_callbacks cb;
 375
 376   /* Identifier hash table.  */
 377   struct ht *hash_table;
 378
 379   /* Expression parser stack.  */
 380   struct op *op_stack, *op_limit;
 381
 382   /* User visible options.  */
 383   struct cpp_options opts;
 384
 385   /* Special nodes - identifiers with predefined significance to the
 386      preprocessor.  */
 387   struct spec_nodes spec_nodes;
 388
 389   /* Whether to print our version number.  Done this way so
 390      we don't get it twice for -v -version.  */
 391   unsigned char print_version;
 392
 393   /* Whether cpplib owns the hashtable.  */
 394   unsigned char our_hashtable;
 395
 396   /* Traditional preprocessing output buffer.  */
 397   uchar *trad_out_base, *trad_out_limit;
 398   uchar *trad_out_cur;
 399   unsigned int trad_line;
 400 };
 401
 402 /* Character classes.  Based on the more primitive macros in safe-ctype.h.
 403    If the definition of `numchar' looks odd to you, please look up the
 404    definition of a pp-number in the C standard [section 6.4.8 of C99].
 405
 406    In the unlikely event that characters other than \r and \n enter
 407    the set is_vspace, the macro handle_newline() in cpplex.c must be
 408    updated.  */
 409 #define _dollar_ok(x)   ((x) == '$' && CPP_OPTION (pfile, dollars_in_ident))
 410
 411 #define is_idchar(x)    (ISIDNUM(x) || _dollar_ok(x))
 412 #define is_numchar(x)   ISIDNUM(x)
 413 #define is_idstart(x)   (ISIDST(x) || _dollar_ok(x))
 414 #define is_numstart(x)  ISDIGIT(x)
 415 #define is_hspace(x)    ISBLANK(x)
 416 #define is_vspace(x)    IS_VSPACE(x)
 417 #define is_nvspace(x)   IS_NVSPACE(x)
 418 #define is_space(x)     IS_SPACE_OR_NUL(x)
 419
 420 /* This table is constant if it can be initialized at compile time,
 421    which is the case if cpp was compiled with GCC >=2.7, or another
 422    compiler that supports C99.  */
 423 #if HAVE_DESIGNATED_INITIALIZERS
 424 extern const unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 425 #else
 426 extern unsigned char _cpp_trigraph_map[UCHAR_MAX + 1];
 427 #endif
 428
 429 /* Macros.  */
 430
 431 #define CPP_PRINT_DEPS(PFILE) CPP_OPTION (PFILE, print_deps)
 432 #define CPP_IN_SYSTEM_HEADER(PFILE) ((PFILE)->map && (PFILE)->map->sysp)
 433 #define CPP_PEDANTIC(PF) CPP_OPTION (PF, pedantic)
 434 #define CPP_WTRADITIONAL(PF) CPP_OPTION (PF, warn_traditional)
 435
 436 /* In cpperror.c  */
 437 extern int _cpp_begin_message PARAMS ((cpp_reader *, int,
 438                                        unsigned int, unsigned int));
 439
 440 /* In cppmacro.c */
 441 extern void _cpp_free_definition        PARAMS ((cpp_hashnode *));
 442 extern bool _cpp_create_definition      PARAMS ((cpp_reader *, cpp_hashnode *));
 443 extern void _cpp_pop_context            PARAMS ((cpp_reader *));
 444 extern void _cpp_push_text_context      PARAMS ((cpp_reader *, cpp_hashnode *,
 445                                                  const uchar *, const uchar*));
 446 extern bool _cpp_save_parameter         PARAMS ((cpp_reader *, cpp_macro *,
 447                                                  cpp_hashnode *));
 448
 449 /* In cpphash.c */
 450 extern void _cpp_init_hashtable         PARAMS ((cpp_reader *, hash_table *));
 451 extern void _cpp_destroy_hashtable      PARAMS ((cpp_reader *));
 452
 453 /* In cppfiles.c */
 454 extern void _cpp_fake_include           PARAMS ((cpp_reader *, const char *));
 455 extern void _cpp_never_reread           PARAMS ((struct include_file *));
 456 extern char *_cpp_simplify_pathname     PARAMS ((char *));
 457 extern bool _cpp_read_file              PARAMS ((cpp_reader *, const char *));
 458 extern bool _cpp_execute_include        PARAMS ((cpp_reader *,
 459                                                  const cpp_token *,
 460                                                  enum include_type));
 461 extern int _cpp_compare_file_date       PARAMS ((cpp_reader *,
 462                                                  const cpp_token *));
 463 extern void _cpp_report_missing_guards  PARAMS ((cpp_reader *));
 464 extern void _cpp_init_includes          PARAMS ((cpp_reader *));
 465 extern void _cpp_cleanup_includes       PARAMS ((cpp_reader *));
 466 extern void _cpp_pop_file_buffer        PARAMS ((cpp_reader *,
 467                                                  struct include_file *));
 468
 469 /* In cppexp.c */
 470 extern bool _cpp_parse_expr             PARAMS ((cpp_reader *));
 471 extern struct op *_cpp_expand_op_stack  PARAMS ((cpp_reader *));
 472
 473 /* In cpplex.c */
 474 extern cpp_token *_cpp_temp_token       PARAMS ((cpp_reader *));
 475 extern const cpp_token *_cpp_lex_token  PARAMS ((cpp_reader *));
 476 extern cpp_token *_cpp_lex_direct       PARAMS ((cpp_reader *));
 477 extern int _cpp_equiv_tokens            PARAMS ((const cpp_token *,
 478                                                  const cpp_token *));
 479 extern void _cpp_init_tokenrun          PARAMS ((tokenrun *, unsigned int));
 480
 481 /* In cppinit.c.  */
 482 extern void _cpp_maybe_push_include_file PARAMS ((cpp_reader *));
 483
 484 /* In cpplib.c */
 485 extern int _cpp_test_assertion PARAMS ((cpp_reader *, unsigned int *));
 486 extern int _cpp_handle_directive PARAMS ((cpp_reader *, int));
 487 extern void _cpp_define_builtin PARAMS ((cpp_reader *, const char *));
 488 extern void _cpp_do__Pragma     PARAMS ((cpp_reader *));
 489 extern void _cpp_init_directives PARAMS ((cpp_reader *));
 490 extern void _cpp_init_internal_pragmas PARAMS ((cpp_reader *));
 491 extern void _cpp_do_file_change PARAMS ((cpp_reader *, enum lc_reason,
 492                                          const char *,
 493                                          unsigned int, unsigned int));
 494 extern void _cpp_pop_buffer PARAMS ((cpp_reader *));
 495
 496 /* In cpptrad.c.  */
 497 extern bool _cpp_read_logical_line_trad PARAMS ((cpp_reader *));
 498 extern void _cpp_overlay_buffer PARAMS ((cpp_reader *pfile, const uchar *,
 499                                          size_t));
 500 extern cpp_hashnode *_cpp_lex_identifier_trad PARAMS ((cpp_reader *));
 501 extern void _cpp_set_trad_context PARAMS ((cpp_reader *));
 502 extern bool _cpp_create_trad_definition PARAMS ((cpp_reader *, cpp_macro *));
 503
 504 /* Utility routines and macros.  */
 505 #define DSC(str) (const uchar *)str, sizeof str - 1
 506 #define xnew(T)         (T *) xmalloc (sizeof(T))
 507 #define xcnew(T)        (T *) xcalloc (1, sizeof(T))
 508 #define xnewvec(T, N)   (T *) xmalloc (sizeof(T) * (N))
 509 #define xcnewvec(T, N)  (T *) xcalloc (N, sizeof(T))
 510 #define xobnew(O, T)    (T *) obstack_alloc (O, sizeof(T))
 511
 512 /* These are inline functions instead of macros so we can get type
 513    checking.  */
 514 static inline int ustrcmp       PARAMS ((const uchar *, const uchar *));
 515 static inline int ustrncmp      PARAMS ((const uchar *, const uchar *,
 516                                          size_t));
 517 static inline size_t ustrlen    PARAMS ((const uchar *));
 518 static inline uchar *uxstrdup   PARAMS ((const uchar *));
 519 static inline uchar *ustrchr    PARAMS ((const uchar *, int));
 520 static inline int ufputs        PARAMS ((const uchar *, FILE *));
 521
 522 static inline int
 523 ustrcmp (s1, s2)
 524      const uchar *s1, *s2;
 525 {
 526   return strcmp ((const char *)s1, (const char *)s2);
 527 }
 528
 529 static inline int
 530 ustrncmp (s1, s2, n)
 531      const uchar *s1, *s2;
 532      size_t n;
 533 {
 534   return strncmp ((const char *)s1, (const char *)s2, n);
 535 }
 536
 537 static inline size_t
 538 ustrlen (s1)
 539      const uchar *s1;
 540 {
 541   return strlen ((const char *)s1);
 542 }
 543
 544 static inline uchar *
 545 uxstrdup (s1)
 546      const uchar *s1;
 547 {
 548   return (uchar *) xstrdup ((const char *)s1);
 549 }
 550
 551 static inline uchar *
 552 ustrchr (s1, c)
 553      const uchar *s1;
 554      int c;
 555 {
 556   return (uchar *) strchr ((const char *)s1, c);
 557 }
 558
 559 static inline int
 560 ufputs (s, f)
 561      const uchar *s;
 562      FILE *f;
 563 {
 564   return fputs ((const char *)s, f);
 565 }
 566
 567 #endif /* ! GCC_CPPHASH_H */