libcpp/include/cpplib.h

   1 /* Definitions for CPP library.
   2    Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
   3    2004, 2005, 2007, 2008, 2009
   4    Free Software Foundation, Inc.
   5    Written by Per Bothner, 1994-95.
   6
   7 This program is free software; you can redistribute it and/or modify it
   8 under the terms of the GNU General Public License as published by the
   9 Free Software Foundation; either version 3, or (at your option) any
  10 later version.
  11
  12 This program is distributed in the hope that it will be useful,
  13 but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15 GNU General Public License for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with this program; see the file COPYING3.  If not see
  19 <http://www.gnu.org/licenses/>.
  20
  21  In other words, you are welcome to use, share and improve this program.
  22  You are forbidden to forbid anyone else to use, share and improve
  23  what you give them.   Help stamp out software-hoarding!  */
  24 #ifndef LIBCPP_CPPLIB_H
  25 #define LIBCPP_CPPLIB_H
  26
  27 #include <sys/types.h>
  28 #include "symtab.h"
  29 #include "line-map.h"
  30
  31 #ifdef __cplusplus
  32 extern "C" {
  33 #endif
  34
  35 typedef struct cpp_reader cpp_reader;
  36 typedef struct cpp_buffer cpp_buffer;
  37 typedef struct cpp_options cpp_options;
  38 typedef struct cpp_token cpp_token;
  39 typedef struct cpp_string cpp_string;
  40 typedef struct cpp_hashnode cpp_hashnode;
  41 typedef struct cpp_macro cpp_macro;
  42 typedef struct cpp_callbacks cpp_callbacks;
  43 typedef struct cpp_dir cpp_dir;
  44
  45 struct answer;
  46 struct _cpp_file;
  47
  48 /* The first three groups, apart from '=', can appear in preprocessor
  49    expressions (+= and -= are used to indicate unary + and - resp.).
  50    This allows a lookup table to be implemented in _cpp_parse_expr.
  51
  52    The first group, to CPP_LAST_EQ, can be immediately followed by an
  53    '='.  The lexer needs operators ending in '=', like ">>=", to be in
  54    the same order as their counterparts without the '=', like ">>".
  55
  56    See the cpp_operator table optab in expr.c if you change the order or
  57    add or remove anything in the first group.  */
  58
  59 #define TTYPE_TABLE                                                     \
  60   OP(EQ,                "=")                                            \
  61   OP(NOT,               "!")                                            \
  62   OP(GREATER,           ">")    /* compare */                           \
  63   OP(LESS,              "<")                                            \
  64   OP(PLUS,              "+")    /* math */                              \
  65   OP(MINUS,             "-")                                            \
  66   OP(MULT,              "*")                                            \
  67   OP(DIV,               "/")                                            \
  68   OP(MOD,               "%")                                            \
  69   OP(AND,               "&")    /* bit ops */                           \
  70   OP(OR,                "|")                                            \
  71   OP(XOR,               "^")                                            \
  72   OP(RSHIFT,            ">>")                                           \
  73   OP(LSHIFT,            "<<")                                           \
  74                                                                         \
  75   OP(COMPL,             "~")                                            \
  76   OP(AND_AND,           "&&")   /* logical */                           \
  77   OP(OR_OR,             "||")                                           \
  78   OP(QUERY,             "?")                                            \
  79   OP(COLON,             ":")                                            \
  80   OP(COMMA,             ",")    /* grouping */                          \
  81   OP(OPEN_PAREN,        "(")                                            \
  82   OP(CLOSE_PAREN,       ")")                                            \
  83   TK(EOF,               NONE)                                           \
  84   OP(EQ_EQ,             "==")   /* compare */                           \
  85   OP(NOT_EQ,            "!=")                                           \
  86   OP(GREATER_EQ,        ">=")                                           \
  87   OP(LESS_EQ,           "<=")                                           \
  88                                                                         \
  89   /* These two are unary + / - in preprocessor expressions.  */         \
  90   OP(PLUS_EQ,           "+=")   /* math */                              \
  91   OP(MINUS_EQ,          "-=")                                           \
  92                                                                         \
  93   OP(MULT_EQ,           "*=")                                           \
  94   OP(DIV_EQ,            "/=")                                           \
  95   OP(MOD_EQ,            "%=")                                           \
  96   OP(AND_EQ,            "&=")   /* bit ops */                           \
  97   OP(OR_EQ,             "|=")                                           \
  98   OP(XOR_EQ,            "^=")                                           \
  99   OP(RSHIFT_EQ,         ">>=")                                          \
 100   OP(LSHIFT_EQ,         "<<=")                                          \
 101   /* Digraphs together, beginning with CPP_FIRST_DIGRAPH.  */           \
 102   OP(HASH,              "#")    /* digraphs */                          \
 103   OP(PASTE,             "##")                                           \
 104   OP(OPEN_SQUARE,       "[")                                            \
 105   OP(CLOSE_SQUARE,      "]")                                            \
 106   OP(OPEN_BRACE,        "{")                                            \
 107   OP(CLOSE_BRACE,       "}")                                            \
 108   /* The remainder of the punctuation.  Order is not significant.  */   \
 109   OP(SEMICOLON,         ";")    /* structure */                         \
 110   OP(ELLIPSIS,          "...")                                          \
 111   OP(PLUS_PLUS,         "++")   /* increment */                         \
 112   OP(MINUS_MINUS,       "--")                                           \
 113   OP(DEREF,             "->")   /* accessors */                         \
 114   OP(DOT,               ".")                                            \
 115   OP(SCOPE,             "::")                                           \
 116   OP(DEREF_STAR,        "->*")                                          \
 117   OP(DOT_STAR,          ".*")                                           \
 118   OP(ATSIGN,            "@")  /* used in Objective-C */                 \
 119                                                                         \
 120   TK(NAME,              IDENT)   /* word */                             \
 121   TK(AT_NAME,           IDENT)   /* @word - Objective-C */              \
 122   TK(NUMBER,            LITERAL) /* 34_be+ta  */                        \
 123                                                                         \
 124   TK(CHAR,              LITERAL) /* 'char' */                           \
 125   TK(WCHAR,             LITERAL) /* L'char' */                          \
 126   TK(CHAR16,            LITERAL) /* u'char' */                          \
 127   TK(CHAR32,            LITERAL) /* U'char' */                          \
 128   TK(OTHER,             LITERAL) /* stray punctuation */                \
 129                                                                         \
 130   TK(STRING,            LITERAL) /* "string" */                         \
 131   TK(WSTRING,           LITERAL) /* L"string" */                        \
 132   TK(STRING16,          LITERAL) /* u"string" */                        \
 133   TK(STRING32,          LITERAL) /* U"string" */                        \
 134   TK(OBJC_STRING,       LITERAL) /* @"string" - Objective-C */          \
 135   TK(HEADER_NAME,       LITERAL) /* <stdio.h> in #include */            \
 136                                                                         \
 137   TK(COMMENT,           LITERAL) /* Only if output comments.  */        \
 138                                  /* SPELL_LITERAL happens to DTRT.  */  \
 139   TK(MACRO_ARG,         NONE)    /* Macro argument.  */                 \
 140   TK(PRAGMA,            NONE)    /* Only for deferred pragmas.  */      \
 141   TK(PRAGMA_EOL,        NONE)    /* End-of-line for deferred pragmas.  */ \
 142   TK(PADDING,           NONE)    /* Whitespace for -E.  */
 143
 144 #define OP(e, s) CPP_ ## e,
 145 #define TK(e, s) CPP_ ## e,
 146 enum cpp_ttype
 147 {
 148   TTYPE_TABLE
 149   N_TTYPES,
 150
 151   /* Positions in the table.  */
 152   CPP_LAST_EQ        = CPP_LSHIFT,
 153   CPP_FIRST_DIGRAPH  = CPP_HASH,
 154   CPP_LAST_PUNCTUATOR= CPP_ATSIGN,
 155   CPP_LAST_CPP_OP    = CPP_LESS_EQ
 156 };
 157 #undef OP
 158 #undef TK
 159
 160 /* C language kind, used when calling cpp_create_reader.  */
 161 enum c_lang {CLK_GNUC89 = 0, CLK_GNUC99, CLK_STDC89, CLK_STDC94, CLK_STDC99,
 162              CLK_GNUCXX, CLK_CXX98, CLK_GNUCXX0X, CLK_CXX0X, CLK_ASM};
 163
 164 /* Payload of a NUMBER, STRING, CHAR or COMMENT token.  */
 165 struct GTY(()) cpp_string {
 166   unsigned int len;
 167   const unsigned char *text;
 168 };
 169
 170 /* Flags for the cpp_token structure.  */
 171 #define PREV_WHITE      (1 << 0) /* If whitespace before this token.  */
 172 #define DIGRAPH         (1 << 1) /* If it was a digraph.  */
 173 #define STRINGIFY_ARG   (1 << 2) /* If macro argument to be stringified.  */
 174 #define PASTE_LEFT      (1 << 3) /* If on LHS of a ## operator.  */
 175 #define NAMED_OP        (1 << 4) /* C++ named operators.  */
 176 #define NO_EXPAND       (1 << 5) /* Do not macro-expand this token.  */
 177 #define BOL             (1 << 6) /* Token at beginning of line.  */
 178 #define PURE_ZERO       (1 << 7) /* Single 0 digit, used by the C++ frontend,
 179                                     set in c-lex.c.  */
 180 #define SP_DIGRAPH      (1 << 8) /* # or ## token was a digraph.  */
 181 #define SP_PREV_WHITE   (1 << 9) /* If whitespace before a ##
 182                                     operator, or before this token
 183                                     after a # operator.  */
 184
 185 /* Specify which field, if any, of the cpp_token union is used.  */
 186
 187 enum cpp_token_fld_kind {
 188   CPP_TOKEN_FLD_NODE,
 189   CPP_TOKEN_FLD_SOURCE,
 190   CPP_TOKEN_FLD_STR,
 191   CPP_TOKEN_FLD_ARG_NO,
 192   CPP_TOKEN_FLD_TOKEN_NO,
 193   CPP_TOKEN_FLD_PRAGMA,
 194   CPP_TOKEN_FLD_NONE
 195 };
 196
 197 /* A macro argument in the cpp_token union.  */
 198 struct GTY(()) cpp_macro_arg {
 199   /* Argument number.  */
 200   unsigned int arg_no;
 201 };
 202
 203 /* An identifier in the cpp_token union.  */
 204 struct GTY(()) cpp_identifier {
 205   /* The canonical (UTF-8) spelling of the identifier.  */
 206   cpp_hashnode *
 207     GTY ((nested_ptr (union tree_node,
 208                 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
 209                         "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
 210        node;
 211 };
 212
 213 /* A preprocessing token.  This has been carefully packed and should
 214    occupy 16 bytes on 32-bit hosts and 24 bytes on 64-bit hosts.  */
 215 struct GTY(()) cpp_token {
 216   source_location src_loc;      /* Location of first char of token.  */
 217   ENUM_BITFIELD(cpp_ttype) type : CHAR_BIT;  /* token type */
 218   unsigned short flags;         /* flags - see above */
 219
 220   union cpp_token_u
 221   {
 222     /* An identifier.  */
 223     struct cpp_identifier GTY ((tag ("CPP_TOKEN_FLD_NODE"))) node;
 224
 225     /* Inherit padding from this token.  */
 226     cpp_token * GTY ((tag ("CPP_TOKEN_FLD_SOURCE"))) source;
 227
 228     /* A string, or number.  */
 229     struct cpp_string GTY ((tag ("CPP_TOKEN_FLD_STR"))) str;
 230
 231     /* Argument no. for a CPP_MACRO_ARG.  */
 232     struct cpp_macro_arg GTY ((tag ("CPP_TOKEN_FLD_ARG_NO"))) macro_arg;
 233
 234     /* Original token no. for a CPP_PASTE (from a sequence of
 235        consecutive paste tokens in a macro expansion).  */
 236     unsigned int GTY ((tag ("CPP_TOKEN_FLD_TOKEN_NO"))) token_no;
 237
 238     /* Caller-supplied identifier for a CPP_PRAGMA.  */
 239     unsigned int GTY ((tag ("CPP_TOKEN_FLD_PRAGMA"))) pragma;
 240   } GTY ((desc ("cpp_token_val_index (&%1)"))) val;
 241 };
 242
 243 /* Say which field is in use.  */
 244 extern enum cpp_token_fld_kind cpp_token_val_index (cpp_token *tok);
 245
 246 /* A type wide enough to hold any multibyte source character.
 247    cpplib's character constant interpreter requires an unsigned type.
 248    Also, a typedef for the signed equivalent.
 249    The width of this type is capped at 32 bits; there do exist targets
 250    where wchar_t is 64 bits, but only in a non-default mode, and there
 251    would be no meaningful interpretation for a wchar_t value greater
 252    than 2^32 anyway -- the widest wide-character encoding around is
 253    ISO 10646, which stops at 2^31.  */
 254 #if CHAR_BIT * SIZEOF_INT >= 32
 255 # define CPPCHAR_SIGNED_T int
 256 #elif CHAR_BIT * SIZEOF_LONG >= 32
 257 # define CPPCHAR_SIGNED_T long
 258 #else
 259 # error "Cannot find a least-32-bit signed integer type"
 260 #endif
 261 typedef unsigned CPPCHAR_SIGNED_T cppchar_t;
 262 typedef CPPCHAR_SIGNED_T cppchar_signed_t;
 263
 264 /* Style of header dependencies to generate.  */
 265 enum cpp_deps_style { DEPS_NONE = 0, DEPS_USER, DEPS_SYSTEM };
 266
 267 /* The possible normalization levels, from most restrictive to least.  */
 268 enum cpp_normalize_level {
 269   /* In NFKC.  */
 270   normalized_KC = 0,
 271   /* In NFC.  */
 272   normalized_C,
 273   /* In NFC, except for subsequences where being in NFC would make
 274      the identifier invalid.  */
 275   normalized_identifier_C,
 276   /* Not normalized at all.  */
 277   normalized_none
 278 };
 279
 280 /* This structure is nested inside struct cpp_reader, and
 281    carries all the options visible to the command line.  */
 282 struct cpp_options
 283 {
 284   /* Characters between tab stops.  */
 285   unsigned int tabstop;
 286
 287   /* The language we're preprocessing.  */
 288   enum c_lang lang;
 289
 290   /* Nonzero means use extra default include directories for C++.  */
 291   unsigned char cplusplus;
 292
 293   /* Nonzero means handle cplusplus style comments.  */
 294   unsigned char cplusplus_comments;
 295
 296   /* Nonzero means define __OBJC__, treat @ as a special token, use
 297      the OBJC[PLUS]_INCLUDE_PATH environment variable, and allow
 298      "#import".  */
 299   unsigned char objc;
 300
 301   /* Nonzero means don't copy comments into the output file.  */
 302   unsigned char discard_comments;
 303
 304   /* Nonzero means don't copy comments into the output file during
 305      macro expansion.  */
 306   unsigned char discard_comments_in_macro_exp;
 307
 308   /* Nonzero means process the ISO trigraph sequences.  */
 309   unsigned char trigraphs;
 310
 311   /* Nonzero means process the ISO digraph sequences.  */
 312   unsigned char digraphs;
 313
 314   /* Nonzero means to allow hexadecimal floats and LL suffixes.  */
 315   unsigned char extended_numbers;
 316
 317   /* Nonzero means process u/U prefix literals (UTF-16/32).  */
 318   unsigned char uliterals;
 319
 320   /* Nonzero means print names of header files (-H).  */
 321   unsigned char print_include_names;
 322
 323   /* Nonzero means complain about deprecated features.  */
 324   unsigned char warn_deprecated;
 325
 326   /* Nonzero means warn if slash-star appears in a comment.  */
 327   unsigned char warn_comments;
 328
 329   /* Nonzero means warn if a user-supplied include directory does not
 330      exist.  */
 331   unsigned char warn_missing_include_dirs;
 332
 333   /* Nonzero means warn if there are any trigraphs.  */
 334   unsigned char warn_trigraphs;
 335
 336   /* Nonzero means warn about multicharacter charconsts.  */
 337   unsigned char warn_multichar;
 338
 339   /* Nonzero means warn about various incompatibilities with
 340      traditional C.  */
 341   unsigned char warn_traditional;
 342
 343   /* Nonzero means warn about long long numeric constants.  */
 344   unsigned char warn_long_long;
 345
 346   /* Nonzero means warn about text after an #endif (or #else).  */
 347   unsigned char warn_endif_labels;
 348
 349   /* Nonzero means warn about implicit sign changes owing to integer
 350      promotions.  */
 351   unsigned char warn_num_sign_change;
 352
 353   /* Zero means don't warn about __VA_ARGS__ usage in c89 pedantic mode.
 354      Presumably the usage is protected by the appropriate #ifdef.  */
 355   unsigned char warn_variadic_macros;
 356
 357   /* Nonzero means warn about builtin macros that are redefined or
 358      explicitly undefined.  */
 359   unsigned char warn_builtin_macro_redefined;
 360
 361   /* Nonzero means we should look for header.gcc files that remap file
 362      names.  */
 363   unsigned char remap;
 364
 365   /* Zero means dollar signs are punctuation.  */
 366   unsigned char dollars_in_ident;
 367
 368   /* Nonzero means UCNs are accepted in identifiers.  */
 369   unsigned char extended_identifiers;
 370
 371   /* True if we should warn about dollars in identifiers or numbers
 372      for this translation unit.  */
 373   unsigned char warn_dollars;
 374
 375   /* Nonzero means warn if undefined identifiers are evaluated in an #if.  */
 376   unsigned char warn_undef;
 377
 378   /* Nonzero means warn of unused macros from the main file.  */
 379   unsigned char warn_unused_macros;
 380
 381   /* Nonzero for the 1999 C Standard, including corrigenda and amendments.  */
 382   unsigned char c99;
 383
 384   /* Nonzero if we are conforming to a specific C or C++ standard.  */
 385   unsigned char std;
 386
 387   /* Nonzero means give all the error messages the ANSI standard requires.  */
 388   unsigned char pedantic;
 389
 390   /* Nonzero means we're looking at already preprocessed code, so don't
 391      bother trying to do macro expansion and whatnot.  */
 392   unsigned char preprocessed;
 393
 394   /* Print column number in error messages.  */
 395   unsigned char show_column;
 396
 397   /* Nonzero means handle C++ alternate operator names.  */
 398   unsigned char operator_names;
 399
 400   /* Nonzero means warn about use of C++ alternate operator names.  */
 401   unsigned char warn_cxx_operator_names;
 402
 403   /* True for traditional preprocessing.  */
 404   unsigned char traditional;
 405
 406   /* Holds the name of the target (execution) character set.  */
 407   const char *narrow_charset;
 408
 409   /* Holds the name of the target wide character set.  */
 410   const char *wide_charset;
 411
 412   /* Holds the name of the input character set.  */
 413   const char *input_charset;
 414
 415   /* The minimum permitted level of normalization before a warning
 416      is generated.  */
 417   enum cpp_normalize_level warn_normalize;
 418
 419   /* True to warn about precompiled header files we couldn't use.  */
 420   bool warn_invalid_pch;
 421
 422   /* True if dependencies should be restored from a precompiled header.  */
 423   bool restore_pch_deps;
 424
 425   /* Dependency generation.  */
 426   struct
 427   {
 428     /* Style of header dependencies to generate.  */
 429     enum cpp_deps_style style;
 430
 431     /* Assume missing files are generated files.  */
 432     bool missing_files;
 433
 434     /* Generate phony targets for each dependency apart from the first
 435        one.  */
 436     bool phony_targets;
 437
 438     /* If true, no dependency is generated on the main file.  */
 439     bool ignore_main_file;
 440   } deps;
 441
 442   /* Target-specific features set by the front end or client.  */
 443
 444   /* Precision for target CPP arithmetic, target characters, target
 445      ints and target wide characters, respectively.  */
 446   size_t precision, char_precision, int_precision, wchar_precision;
 447
 448   /* True means chars (wide chars) are unsigned.  */
 449   bool unsigned_char, unsigned_wchar;
 450
 451   /* True if the most significant byte in a word has the lowest
 452      address in memory.  */
 453   bool bytes_big_endian;
 454
 455   /* Nonzero means __STDC__ should have the value 0 in system headers.  */
 456   unsigned char stdc_0_in_system_headers;
 457
 458   /* True disables tokenization outside of preprocessing directives. */
 459   bool directives_only;
 460 };
 461
 462 /* Callback for header lookup for HEADER, which is the name of a
 463    source file.  It is used as a method of last resort to find headers
 464    that are not otherwise found during the normal include processing.
 465    The return value is the malloced name of a header to try and open,
 466    if any, or NULL otherwise.  This callback is called only if the
 467    header is otherwise unfound.  */
 468 typedef const char *(*missing_header_cb)(cpp_reader *, const char *header, cpp_dir **);
 469
 470 /* Call backs to cpplib client.  */
 471 struct cpp_callbacks
 472 {
 473   /* Called when a new line of preprocessed output is started.  */
 474   void (*line_change) (cpp_reader *, const cpp_token *, int);
 475
 476   /* Called when switching to/from a new file.
 477      The line_map is for the new file.  It is NULL if there is no new file.
 478      (In C this happens when done with <built-in>+<command line> and also
 479      when done with a main file.)  This can be used for resource cleanup.  */
 480   void (*file_change) (cpp_reader *, const struct line_map *);
 481
 482   void (*dir_change) (cpp_reader *, const char *);
 483   void (*include) (cpp_reader *, unsigned int, const unsigned char *,
 484                    const char *, int, const cpp_token **);
 485   void (*define) (cpp_reader *, unsigned int, cpp_hashnode *);
 486   void (*undef) (cpp_reader *, unsigned int, cpp_hashnode *);
 487   void (*ident) (cpp_reader *, unsigned int, const cpp_string *);
 488   void (*def_pragma) (cpp_reader *, unsigned int);
 489   int (*valid_pch) (cpp_reader *, const char *, int);
 490   void (*read_pch) (cpp_reader *, const char *, int, const char *);
 491   missing_header_cb missing_header;
 492
 493   /* Context-sensitive macro support.  Returns macro (if any) that should
 494      be expanded.  */
 495   cpp_hashnode * (*macro_to_expand) (cpp_reader *, const cpp_token *);
 496
 497   /* Called to emit a diagnostic.  This callback receives the
 498      translated message.  */
 499   bool (*error) (cpp_reader *, int, source_location, unsigned int,
 500                  const char *, va_list *)
 501        ATTRIBUTE_FPTR_PRINTF(5,0);
 502
 503   /* Callbacks for when a macro is expanded, or tested (whether
 504      defined or not at the time) in #ifdef, #ifndef or "defined".  */
 505   void (*used_define) (cpp_reader *, unsigned int, cpp_hashnode *);
 506   void (*used_undef) (cpp_reader *, unsigned int, cpp_hashnode *);
 507   /* Called before #define and #undef or other macro definition
 508      changes are processed.  */
 509   void (*before_define) (cpp_reader *);
 510 };
 511
 512 /* Chain of directories to look for include files in.  */
 513 struct cpp_dir
 514 {
 515   /* NULL-terminated singly-linked list.  */
 516   struct cpp_dir *next;
 517
 518   /* NAME of the directory, NUL-terminated.  */
 519   char *name;
 520   unsigned int len;
 521
 522   /* One if a system header, two if a system header that has extern
 523      "C" guards for C++.  */
 524   unsigned char sysp;
 525
 526   /* Is this a user-supplied directory? */
 527   bool user_supplied_p;
 528
 529   /* The canonicalized NAME as determined by lrealpath.  This field
 530      is only used by hosts that lack reliable inode numbers.  */
 531   char *canonical_name;
 532
 533   /* Mapping of file names for this directory for MS-DOS and related
 534      platforms.  A NULL-terminated array of (from, to) pairs.  */
 535   const char **name_map;
 536
 537   /* Routine to construct pathname, given the search path name and the
 538      HEADER we are trying to find, return a constructed pathname to
 539      try and open.  If this is NULL, the constructed pathname is as
 540      constructed by append_file_to_dir.  */
 541   char *(*construct) (const char *header, cpp_dir *dir);
 542
 543   /* The C front end uses these to recognize duplicated
 544      directories in the search path.  */
 545   ino_t ino;
 546   dev_t dev;
 547 };
 548
 549 /* The structure of a node in the hash table.  The hash table has
 550    entries for all identifiers: either macros defined by #define
 551    commands (type NT_MACRO), assertions created with #assert
 552    (NT_ASSERTION), or neither of the above (NT_VOID).  Builtin macros
 553    like __LINE__ are flagged NODE_BUILTIN.  Poisoned identifiers are
 554    flagged NODE_POISONED.  NODE_OPERATOR (C++ only) indicates an
 555    identifier that behaves like an operator such as "xor".
 556    NODE_DIAGNOSTIC is for speed in lex_token: it indicates a
 557    diagnostic may be required for this node.  Currently this only
 558    applies to __VA_ARGS__, poisoned identifiers, and -Wc++-compat
 559    warnings about NODE_OPERATOR.  */
 560
 561 /* Hash node flags.  */
 562 #define NODE_OPERATOR   (1 << 0)        /* C++ named operator.  */
 563 #define NODE_POISONED   (1 << 1)        /* Poisoned identifier.  */
 564 #define NODE_BUILTIN    (1 << 2)        /* Builtin macro.  */
 565 #define NODE_DIAGNOSTIC (1 << 3)        /* Possible diagnostic when lexed.  */
 566 #define NODE_WARN       (1 << 4)        /* Warn if redefined or undefined.  */
 567 #define NODE_DISABLED   (1 << 5)        /* A disabled macro.  */
 568 #define NODE_MACRO_ARG  (1 << 6)        /* Used during #define processing.  */
 569 #define NODE_USED       (1 << 7)        /* Dumped with -dU.  */
 570 #define NODE_CONDITIONAL (1 << 8)       /* Conditional macro */
 571 #define NODE_WARN_OPERATOR (1 << 9)     /* Warn about C++ named operator.  */
 572
 573 /* Different flavors of hash node.  */
 574 enum node_type
 575 {
 576   NT_VOID = 0,     /* No definition yet.  */
 577   NT_MACRO,        /* A macro of some form.  */
 578   NT_ASSERTION     /* Predicate for #assert.  */
 579 };
 580
 581 /* Different flavors of builtin macro.  _Pragma is an operator, but we
 582    handle it with the builtin code for efficiency reasons.  */
 583 enum cpp_builtin_type
 584 {
 585   BT_SPECLINE = 0,              /* `__LINE__' */
 586   BT_DATE,                      /* `__DATE__' */
 587   BT_FILE,                      /* `__FILE__' */
 588   BT_BASE_FILE,                 /* `__BASE_FILE__' */
 589   BT_INCLUDE_LEVEL,             /* `__INCLUDE_LEVEL__' */
 590   BT_TIME,                      /* `__TIME__' */
 591   BT_STDC,                      /* `__STDC__' */
 592   BT_PRAGMA,                    /* `_Pragma' operator */
 593   BT_TIMESTAMP,                 /* `__TIMESTAMP__' */
 594   BT_COUNTER                    /* `__COUNTER__' */
 595 };
 596
 597 #define CPP_HASHNODE(HNODE)     ((cpp_hashnode *) (HNODE))
 598 #define HT_NODE(NODE)           ((ht_identifier *) (NODE))
 599 #define NODE_LEN(NODE)          HT_LEN (&(NODE)->ident)
 600 #define NODE_NAME(NODE)         HT_STR (&(NODE)->ident)
 601
 602 /* Specify which field, if any, of the union is used.  */
 603
 604 enum {
 605   NTV_MACRO,
 606   NTV_ANSWER,
 607   NTV_BUILTIN,
 608   NTV_ARGUMENT,
 609   NTV_NONE
 610 };
 611
 612 #define CPP_HASHNODE_VALUE_IDX(HNODE)                           \
 613   ((HNODE.flags & NODE_MACRO_ARG) ? NTV_ARGUMENT                \
 614    : HNODE.type == NT_MACRO ? ((HNODE.flags & NODE_BUILTIN)     \
 615                                ? NTV_BUILTIN : NTV_MACRO)       \
 616    : HNODE.type == NT_ASSERTION ? NTV_ANSWER                    \
 617    : NTV_NONE)
 618
 619 /* The common part of an identifier node shared amongst all 3 C front
 620    ends.  Also used to store CPP identifiers, which are a superset of
 621    identifiers in the grammatical sense.  */
 622
 623 union GTY(()) _cpp_hashnode_value {
 624   /* If a macro.  */
 625   cpp_macro * GTY((tag ("NTV_MACRO"))) macro;
 626   /* Answers to an assertion.  */
 627   struct answer * GTY ((tag ("NTV_ANSWER"))) answers;
 628   /* Code for a builtin macro.  */
 629   enum cpp_builtin_type GTY ((tag ("NTV_BUILTIN"))) builtin;
 630   /* Macro argument index.  */
 631   unsigned short GTY ((tag ("NTV_ARGUMENT"))) arg_index;
 632 };
 633
 634 struct GTY(()) cpp_hashnode {
 635   struct ht_identifier ident;
 636   unsigned int is_directive : 1;
 637   unsigned int directive_index : 7;     /* If is_directive,
 638                                            then index into directive table.
 639                                            Otherwise, a NODE_OPERATOR.  */
 640   unsigned char rid_code;               /* Rid code - for front ends.  */
 641   ENUM_BITFIELD(node_type) type : 6;    /* CPP node type.  */
 642   unsigned int flags : 10;              /* CPP flags.  */
 643
 644   union _cpp_hashnode_value GTY ((desc ("CPP_HASHNODE_VALUE_IDX (%1)"))) value;
 645 };
 646
 647 /* Call this first to get a handle to pass to other functions.
 648
 649    If you want cpplib to manage its own hashtable, pass in a NULL
 650    pointer.  Otherwise you should pass in an initialized hash table
 651    that cpplib will share; this technique is used by the C front
 652    ends.  */
 653 extern cpp_reader *cpp_create_reader (enum c_lang, struct ht *,
 654                                       struct line_maps *);
 655
 656 /* Reset the cpp_reader's line_map.  This is only used after reading a
 657    PCH file.  */
 658 extern void cpp_set_line_map (cpp_reader *, struct line_maps *);
 659
 660 /* Call this to change the selected language standard (e.g. because of
 661    command line options).  */
 662 extern void cpp_set_lang (cpp_reader *, enum c_lang);
 663
 664 /* Set the include paths.  */
 665 extern void cpp_set_include_chains (cpp_reader *, cpp_dir *, cpp_dir *, int);
 666
 667 /* Call these to get pointers to the options, callback, and deps
 668    structures for a given reader.  These pointers are good until you
 669    call cpp_finish on that reader.  You can either edit the callbacks
 670    through the pointer returned from cpp_get_callbacks, or set them
 671    with cpp_set_callbacks.  */
 672 extern cpp_options *cpp_get_options (cpp_reader *);
 673 extern cpp_callbacks *cpp_get_callbacks (cpp_reader *);
 674 extern void cpp_set_callbacks (cpp_reader *, cpp_callbacks *);
 675 extern struct deps *cpp_get_deps (cpp_reader *);
 676
 677 /* This function reads the file, but does not start preprocessing.  It
 678    returns the name of the original file; this is the same as the
 679    input file, except for preprocessed input.  This will generate at
 680    least one file change callback, and possibly a line change callback
 681    too.  If there was an error opening the file, it returns NULL.  */
 682 extern const char *cpp_read_main_file (cpp_reader *, const char *);
 683
 684 /* Set up built-ins with special behavior.  Use cpp_init_builtins()
 685    instead unless your know what you are doing.  */
 686 extern void cpp_init_special_builtins (cpp_reader *);
 687
 688 /* Set up built-ins like __FILE__.  */
 689 extern void cpp_init_builtins (cpp_reader *, int);
 690
 691 /* This is called after options have been parsed, and partially
 692    processed.  */
 693 extern void cpp_post_options (cpp_reader *);
 694
 695 /* Set up translation to the target character set.  */
 696 extern void cpp_init_iconv (cpp_reader *);
 697
 698 /* Call this to finish preprocessing.  If you requested dependency
 699    generation, pass an open stream to write the information to,
 700    otherwise NULL.  It is your responsibility to close the stream.  */
 701 extern void cpp_finish (cpp_reader *, FILE *deps_stream);
 702
 703 /* Call this to release the handle at the end of preprocessing.  Any
 704    use of the handle after this function returns is invalid.  */
 705 extern void cpp_destroy (cpp_reader *);
 706
 707 extern unsigned int cpp_token_len (const cpp_token *);
 708 extern unsigned char *cpp_token_as_text (cpp_reader *, const cpp_token *);
 709 extern unsigned char *cpp_spell_token (cpp_reader *, const cpp_token *,
 710                                        unsigned char *, bool);
 711 extern void cpp_register_pragma (cpp_reader *, const char *, const char *,
 712                                  void (*) (cpp_reader *), bool);
 713 extern void cpp_register_deferred_pragma (cpp_reader *, const char *,
 714                                           const char *, unsigned, bool, bool);
 715 extern int cpp_avoid_paste (cpp_reader *, const cpp_token *,
 716                             const cpp_token *);
 717 extern const cpp_token *cpp_get_token (cpp_reader *);
 718 extern const cpp_token *cpp_get_token_with_location (cpp_reader *,
 719                                                      source_location *);
 720 extern const unsigned char *cpp_macro_definition (cpp_reader *,
 721                                                   const cpp_hashnode *);
 722 extern void _cpp_backup_tokens (cpp_reader *, unsigned int);
 723 extern const cpp_token *cpp_peek_token (cpp_reader *, int);
 724
 725 /* Evaluate a CPP_CHAR or CPP_WCHAR token.  */
 726 extern cppchar_t cpp_interpret_charconst (cpp_reader *, const cpp_token *,
 727                                           unsigned int *, int *);
 728 /* Evaluate a vector of CPP_STRING or CPP_WSTRING tokens.  */
 729 extern bool cpp_interpret_string (cpp_reader *,
 730                                   const cpp_string *, size_t,
 731                                   cpp_string *, enum cpp_ttype);
 732 extern bool cpp_interpret_string_notranslate (cpp_reader *,
 733                                               const cpp_string *, size_t,
 734                                               cpp_string *, enum cpp_ttype);
 735
 736 /* Convert a host character constant to the execution character set.  */
 737 extern cppchar_t cpp_host_to_exec_charset (cpp_reader *, cppchar_t);
 738
 739 /* Used to register macros and assertions, perhaps from the command line.
 740    The text is the same as the command line argument.  */
 741 extern void cpp_define (cpp_reader *, const char *);
 742 extern void cpp_define_formatted (cpp_reader *pfile,
 743                                   const char *fmt, ...) ATTRIBUTE_PRINTF_2;
 744 extern void cpp_assert (cpp_reader *, const char *);
 745 extern void cpp_undef (cpp_reader *, const char *);
 746 extern void cpp_unassert (cpp_reader *, const char *);
 747
 748 extern cpp_macro *cpp_push_definition (cpp_reader *, const char *);
 749 extern void cpp_pop_definition (cpp_reader *, const char *, cpp_macro *);
 750
 751 /* Undefine all macros and assertions.  */
 752 extern void cpp_undef_all (cpp_reader *);
 753
 754 extern cpp_buffer *cpp_push_buffer (cpp_reader *, const unsigned char *,
 755                                     size_t, int);
 756 extern int cpp_defined (cpp_reader *, const unsigned char *, int);
 757
 758 /* A preprocessing number.  Code assumes that any unused high bits of
 759    the double integer are set to zero.  */
 760 typedef unsigned HOST_WIDE_INT cpp_num_part;
 761 typedef struct cpp_num cpp_num;
 762 struct cpp_num
 763 {
 764   cpp_num_part high;
 765   cpp_num_part low;
 766   bool unsignedp;  /* True if value should be treated as unsigned.  */
 767   bool overflow;   /* True if the most recent calculation overflowed.  */
 768 };
 769
 770 /* cpplib provides two interfaces for interpretation of preprocessing
 771    numbers.
 772
 773    cpp_classify_number categorizes numeric constants according to
 774    their field (integer, floating point, or invalid), radix (decimal,
 775    octal, hexadecimal), and type suffixes.  */
 776
 777 #define CPP_N_CATEGORY  0x000F
 778 #define CPP_N_INVALID   0x0000
 779 #define CPP_N_INTEGER   0x0001
 780 #define CPP_N_FLOATING  0x0002
 781
 782 #define CPP_N_WIDTH     0x00F0
 783 #define CPP_N_SMALL     0x0010  /* int, float, shrot _Fract/Accum  */
 784 #define CPP_N_MEDIUM    0x0020  /* long, double, long _Fract/_Accum.  */
 785 #define CPP_N_LARGE     0x0040  /* long long, long double,
 786                                    long long _Fract/Accum.  */
 787
 788 #define CPP_N_WIDTH_MD  0xF0000 /* machine defined.  */
 789 #define CPP_N_MD_W      0x10000
 790 #define CPP_N_MD_Q      0x20000
 791
 792 #define CPP_N_RADIX     0x0F00
 793 #define CPP_N_DECIMAL   0x0100
 794 #define CPP_N_HEX       0x0200
 795 #define CPP_N_OCTAL     0x0400
 796 #define CPP_N_BINARY    0x0800
 797
 798 #define CPP_N_UNSIGNED  0x1000  /* Properties.  */
 799 #define CPP_N_IMAGINARY 0x2000
 800 #define CPP_N_DFLOAT    0x4000
 801 #define CPP_N_DEFAULT   0x8000
 802
 803 #define CPP_N_FRACT     0x100000 /* Fract types.  */
 804 #define CPP_N_ACCUM     0x200000 /* Accum types.  */
 805
 806 /* Classify a CPP_NUMBER token.  The return value is a combination of
 807    the flags from the above sets.  */
 808 extern unsigned cpp_classify_number (cpp_reader *, const cpp_token *);
 809
 810 /* Evaluate a token classified as category CPP_N_INTEGER.  */
 811 extern cpp_num cpp_interpret_integer (cpp_reader *, const cpp_token *,
 812                                       unsigned int type);
 813
 814 /* Sign extend a number, with PRECISION significant bits and all
 815    others assumed clear, to fill out a cpp_num structure.  */
 816 cpp_num cpp_num_sign_extend (cpp_num, size_t);
 817
 818 /* Diagnostic levels.  To get a diagnostic without associating a
 819    position in the translation unit with it, use cpp_error_with_line
 820    with a line number of zero.  */
 821
 822 /* Warning, an error with -Werror.  */
 823 #define CPP_DL_WARNING          0x00
 824 /* Same as CPP_DL_WARNING, except it is not suppressed in system headers.  */
 825 #define CPP_DL_WARNING_SYSHDR   0x01
 826 /* Warning, an error with -pedantic-errors or -Werror.  */
 827 #define CPP_DL_PEDWARN          0x02
 828 /* An error.  */
 829 #define CPP_DL_ERROR            0x03
 830 /* An internal consistency check failed.  Prints "internal error: ",
 831    otherwise the same as CPP_DL_ERROR.  */
 832 #define CPP_DL_ICE              0x04
 833 /* An informative note following a warning.  */
 834 #define CPP_DL_NOTE             0x05
 835 /* A fatal error.  */
 836 #define CPP_DL_FATAL            0x06
 837
 838 /* Output a diagnostic of some kind.  */
 839 extern bool cpp_error (cpp_reader *, int, const char *msgid, ...)
 840   ATTRIBUTE_PRINTF_3;
 841
 842 /* Output a diagnostic with "MSGID: " preceding the
 843    error string of errno.  No location is printed.  */
 844 extern bool cpp_errno (cpp_reader *, int, const char *msgid);
 845
 846 /* Same as cpp_error, except additionally specifies a position as a
 847    (translation unit) physical line and physical column.  If the line is
 848    zero, then no location is printed.  */
 849 extern bool cpp_error_with_line (cpp_reader *, int, source_location, unsigned,
 850                                  const char *msgid, ...) ATTRIBUTE_PRINTF_5;
 851
 852 /* In lex.c */
 853 extern int cpp_ideq (const cpp_token *, const char *);
 854 extern void cpp_output_line (cpp_reader *, FILE *);
 855 extern unsigned char *cpp_output_line_to_string (cpp_reader *,
 856                                                  const unsigned char *);
 857 extern void cpp_output_token (const cpp_token *, FILE *);
 858 extern const char *cpp_type2name (enum cpp_ttype, unsigned char flags);
 859 /* Returns the value of an escape sequence, truncated to the correct
 860    target precision.  PSTR points to the input pointer, which is just
 861    after the backslash.  LIMIT is how much text we have.  WIDE is true
 862    if the escape sequence is part of a wide character constant or
 863    string literal.  Handles all relevant diagnostics.  */
 864 extern cppchar_t cpp_parse_escape (cpp_reader *, const unsigned char ** pstr,
 865                                    const unsigned char *limit, int wide);
 866
 867 /* Structure used to hold a comment block at a given location in the
 868    source code.  */
 869
 870 typedef struct
 871 {
 872   /* Text of the comment including the terminators.  */
 873   char *comment;
 874
 875   /* source location for the given comment.  */
 876   source_location sloc;
 877 } cpp_comment;
 878
 879 /* Structure holding all comments for a given cpp_reader.  */
 880
 881 typedef struct
 882 {
 883   /* table of comment entries.  */
 884   cpp_comment *entries;
 885
 886   /* number of actual entries entered in the table.  */
 887   int count;
 888
 889   /* number of entries allocated currently.  */
 890   int allocated;
 891 } cpp_comment_table;
 892
 893 /* Returns the table of comments encountered by the preprocessor. This
 894    table is only populated when pfile->state.save_comments is true. */
 895 extern cpp_comment_table *cpp_get_comments (cpp_reader *);
 896
 897 /* In hash.c */
 898
 899 /* Lookup an identifier in the hashtable.  Puts the identifier in the
 900    table if it is not already there.  */
 901 extern cpp_hashnode *cpp_lookup (cpp_reader *, const unsigned char *,
 902                                  unsigned int);
 903
 904 typedef int (*cpp_cb) (cpp_reader *, cpp_hashnode *, void *);
 905 extern void cpp_forall_identifiers (cpp_reader *, cpp_cb, void *);
 906
 907 /* In macro.c */
 908 extern void cpp_scan_nooutput (cpp_reader *);
 909 extern int  cpp_sys_macro_p (cpp_reader *);
 910 extern unsigned char *cpp_quote_string (unsigned char *, const unsigned char *,
 911                                         unsigned int);
 912
 913 /* In files.c */
 914 extern bool cpp_included (cpp_reader *, const char *);
 915 extern bool cpp_included_before (cpp_reader *, const char *, source_location);
 916 extern void cpp_make_system_header (cpp_reader *, int, int);
 917 extern bool cpp_push_include (cpp_reader *, const char *);
 918 extern void cpp_change_file (cpp_reader *, enum lc_reason, const char *);
 919 extern const char *cpp_get_path (struct _cpp_file *);
 920 extern cpp_dir *cpp_get_dir (struct _cpp_file *);
 921 extern cpp_buffer *cpp_get_buffer (cpp_reader *);
 922 extern struct _cpp_file *cpp_get_file (cpp_buffer *);
 923 extern cpp_buffer *cpp_get_prev (cpp_buffer *);
 924 extern void cpp_clear_file_cache (cpp_reader *);
 925
 926 /* In pch.c */
 927 struct save_macro_data;
 928 extern int cpp_save_state (cpp_reader *, FILE *);
 929 extern int cpp_write_pch_deps (cpp_reader *, FILE *);
 930 extern int cpp_write_pch_state (cpp_reader *, FILE *);
 931 extern int cpp_valid_state (cpp_reader *, const char *, int);
 932 extern void cpp_prepare_state (cpp_reader *, struct save_macro_data **);
 933 extern int cpp_read_state (cpp_reader *, const char *, FILE *,
 934                            struct save_macro_data *);
 935
 936 #ifdef __cplusplus
 937 }
 938 #endif
 939
 940 #endif /* ! LIBCPP_CPPLIB_H */