1 /* Definitions for CPP library.
2    Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3    2004, 2005
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 2, 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; if not, write to the Free Software
19 Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
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(OTHER,		LITERAL) /* stray punctuation */		\
127 									\
128   TK(STRING,		LITERAL) /* "string" */				\
129   TK(WSTRING,		LITERAL) /* L"string" */			\
130   TK(OBJC_STRING,	LITERAL) /* @"string" - Objective-C */		\
131   TK(HEADER_NAME,	LITERAL) /* <stdio.h> in #include */		\
132 									\
133   TK(COMMENT,		LITERAL) /* Only if output comments.  */	\
134 				 /* SPELL_LITERAL happens to DTRT.  */	\
135   TK(MACRO_ARG,		NONE)	 /* Macro argument.  */			\
136   TK(PRAGMA,		NONE)	 /* Only for deferred pragmas.  */	\
137   TK(PRAGMA_EOL,	NONE)	 /* End-of-line for deferred pragmas.  */ \
138   TK(PADDING,		NONE)	 /* Whitespace for -E.	*/
139 
140 #define OP(e, s) CPP_ ## e,
141 #define TK(e, s) CPP_ ## e,
142 enum cpp_ttype
143 {
144   TTYPE_TABLE
145   N_TTYPES,
146 
147   /* Positions in the table.  */
148   CPP_LAST_EQ        = CPP_LSHIFT,
149   CPP_FIRST_DIGRAPH  = CPP_HASH,
150   CPP_LAST_PUNCTUATOR= CPP_ATSIGN,
151   CPP_LAST_CPP_OP    = CPP_LESS_EQ
152 };
153 #undef OP
154 #undef TK
155 
156 /* C language kind, used when calling cpp_create_reader.  */
157 enum c_lang {CLK_GNUC89 = 0, CLK_GNUC99, CLK_STDC89, CLK_STDC94, CLK_STDC99,
158 	     CLK_GNUCXX, CLK_CXX98, CLK_ASM};
159 
160 /* Payload of a NUMBER, STRING, CHAR or COMMENT token.  */
161 struct cpp_string GTY(())
162 {
163   unsigned int len;
164   const unsigned char *text;
165 };
166 
167 /* Flags for the cpp_token structure.  */
168 #define PREV_WHITE	(1 << 0) /* If whitespace before this token.  */
169 #define DIGRAPH		(1 << 1) /* If it was a digraph.  */
170 #define STRINGIFY_ARG	(1 << 2) /* If macro argument to be stringified.  */
171 #define PASTE_LEFT	(1 << 3) /* If on LHS of a ## operator.  */
172 #define NAMED_OP	(1 << 4) /* C++ named operators.  */
173 #define NO_EXPAND	(1 << 5) /* Do not macro-expand this token.  */
174 #define BOL		(1 << 6) /* Token at beginning of line.  */
175 #define PURE_ZERO	(1 << 7) /* Single 0 digit, used by the C++ frontend,
176 				    set in c-lex.c.  */
177 
178 /* Specify which field, if any, of the cpp_token union is used.  */
179 
180 enum cpp_token_fld_kind {
181   CPP_TOKEN_FLD_NODE,
182   CPP_TOKEN_FLD_SOURCE,
183   CPP_TOKEN_FLD_STR,
184   CPP_TOKEN_FLD_ARG_NO,
185   CPP_TOKEN_FLD_PRAGMA,
186   CPP_TOKEN_FLD_NONE
187 };
188 
189 /* A preprocessing token.  This has been carefully packed and should
190    occupy 16 bytes on 32-bit hosts and 24 bytes on 64-bit hosts.  */
191 struct cpp_token GTY(())
192 {
193   source_location src_loc;	/* Location of first char of token.  */
194   ENUM_BITFIELD(cpp_ttype) type : CHAR_BIT;  /* token type */
195   unsigned char flags;		/* flags - see above */
196 
197   union cpp_token_u
198   {
199     /* An identifier.  */
200     cpp_hashnode *
201       GTY ((nested_ptr (union tree_node,
202 		"%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
203 			"%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL"),
204 	    tag ("CPP_TOKEN_FLD_NODE")))
205 	 node;
206 
207     /* Inherit padding from this token.  */
208     cpp_token * GTY ((tag ("CPP_TOKEN_FLD_SOURCE"))) source;
209 
210     /* A string, or number.  */
211     struct cpp_string GTY ((tag ("CPP_TOKEN_FLD_STR"))) str;
212 
213     /* Argument no. for a CPP_MACRO_ARG.  */
214     unsigned int GTY ((tag ("CPP_TOKEN_FLD_ARG_NO"))) arg_no;
215 
216     /* Caller-supplied identifier for a CPP_PRAGMA.  */
217     unsigned int GTY ((tag ("CPP_TOKEN_FLD_PRAGMA"))) pragma;
218   } GTY ((desc ("cpp_token_val_index (&%1)"))) val;
219 };
220 
221 /* Say which field is in use.  */
222 extern enum cpp_token_fld_kind cpp_token_val_index (cpp_token *tok);
223 
224 /* A type wide enough to hold any multibyte source character.
225    cpplib's character constant interpreter requires an unsigned type.
226    Also, a typedef for the signed equivalent.
227    The width of this type is capped at 32 bits; there do exist targets
228    where wchar_t is 64 bits, but only in a non-default mode, and there
229    would be no meaningful interpretation for a wchar_t value greater
230    than 2^32 anyway -- the widest wide-character encoding around is
231    ISO 10646, which stops at 2^31.  */
232 #if CHAR_BIT * SIZEOF_INT >= 32
233 # define CPPCHAR_SIGNED_T int
234 #elif CHAR_BIT * SIZEOF_LONG >= 32
235 # define CPPCHAR_SIGNED_T long
236 #else
237 # error "Cannot find a least-32-bit signed integer type"
238 #endif
239 typedef unsigned CPPCHAR_SIGNED_T cppchar_t;
240 typedef CPPCHAR_SIGNED_T cppchar_signed_t;
241 
242 /* Style of header dependencies to generate.  */
243 enum cpp_deps_style { DEPS_NONE = 0, DEPS_USER, DEPS_SYSTEM };
244 
245 /* The possible normalization levels, from most restrictive to least.  */
246 enum cpp_normalize_level {
247   /* In NFKC.  */
248   normalized_KC = 0,
249   /* In NFC.  */
250   normalized_C,
251   /* In NFC, except for subsequences where being in NFC would make
252      the identifier invalid.  */
253   normalized_identifier_C,
254   /* Not normalized at all.  */
255   normalized_none
256 };
257 
258 /* This structure is nested inside struct cpp_reader, and
259    carries all the options visible to the command line.  */
260 struct cpp_options
261 {
262   /* Characters between tab stops.  */
263   unsigned int tabstop;
264 
265   /* The language we're preprocessing.  */
266   enum c_lang lang;
267 
268   /* Nonzero means use extra default include directories for C++.  */
269   unsigned char cplusplus;
270 
271   /* Nonzero means handle cplusplus style comments.  */
272   unsigned char cplusplus_comments;
273 
274   /* Nonzero means define __OBJC__, treat @ as a special token, and
275      use the OBJC[PLUS]_INCLUDE_PATH environment variable.  */
276   unsigned char objc;
277 
278   /* Nonzero means don't copy comments into the output file.  */
279   unsigned char discard_comments;
280 
281   /* Nonzero means don't copy comments into the output file during
282      macro expansion.  */
283   unsigned char discard_comments_in_macro_exp;
284 
285   /* Nonzero means process the ISO trigraph sequences.  */
286   unsigned char trigraphs;
287 
288   /* Nonzero means process the ISO digraph sequences.  */
289   unsigned char digraphs;
290 
291   /* Nonzero means to allow hexadecimal floats and LL suffixes.  */
292   unsigned char extended_numbers;
293 
294   /* Nonzero means print names of header files (-H).  */
295   unsigned char print_include_names;
296 
297   /* Nonzero means cpp_pedwarn causes a hard error.  */
298   unsigned char pedantic_errors;
299 
300   /* Nonzero means don't print warning messages.  */
301   unsigned char inhibit_warnings;
302 
303   /* Nonzero means complain about deprecated features.  */
304   unsigned char warn_deprecated;
305 
306   /* Nonzero means don't suppress warnings from system headers.  */
307   unsigned char warn_system_headers;
308 
309   /* Nonzero means don't print error messages.  Has no option to
310      select it, but can be set by a user of cpplib (e.g. fix-header).  */
311   unsigned char inhibit_errors;
312 
313   /* Nonzero means warn if slash-star appears in a comment.  */
314   unsigned char warn_comments;
315 
316   /* Nonzero means warn if a user-supplied include directory does not
317      exist.  */
318   unsigned char warn_missing_include_dirs;
319 
320   /* Nonzero means warn if there are any trigraphs.  */
321   unsigned char warn_trigraphs;
322 
323   /* APPLE LOCAL begin -Wnewline-eof 2001-08-23 --sts */
324   /* Nonzero means warn if no newline at end of file.  */
325   unsigned char warn_newline_at_eof;
326   /* APPLE LOCAL end -Wnewline-eof 2001-08-23 --sts */
327 
328   /* Nonzero means warn about multicharacter charconsts.  */
329   unsigned char warn_multichar;
330 
331   /* Nonzero means warn about various incompatibilities with
332      traditional C.  */
333   unsigned char warn_traditional;
334 
335   /* Nonzero means warn about long long numeric constants.  */
336   unsigned char warn_long_long;
337 
338   /* Nonzero means warn about text after an #endif (or #else).  */
339   unsigned char warn_endif_labels;
340 
341   /* Nonzero means warn about implicit sign changes owing to integer
342      promotions.  */
343   unsigned char warn_num_sign_change;
344 
345   /* Zero means don't warn about __VA_ARGS__ usage in c89 pedantic mode.
346      Presumably the usage is protected by the appropriate #ifdef.  */
347   unsigned char warn_variadic_macros;
348 
349   /* Nonzero means turn warnings into errors.  */
350   unsigned char warnings_are_errors;
351 
352   /* Nonzero means we should look for header.gcc files that remap file
353      names.  */
354   unsigned char remap;
355 
356   /* Zero means dollar signs are punctuation.  */
357   unsigned char dollars_in_ident;
358 
359   /* Nonzero means UCNs are accepted in identifiers.  */
360   unsigned char extended_identifiers;
361 
362   /* True if we should warn about dollars in identifiers or numbers
363      for this translation unit.  */
364   unsigned char warn_dollars;
365 
366   /* Nonzero means warn if undefined identifiers are evaluated in an #if.  */
367   unsigned char warn_undef;
368 
369   /* Nonzero means warn of unused macros from the main file.  */
370   unsigned char warn_unused_macros;
371 
372   /* Nonzero for the 1999 C Standard, including corrigenda and amendments.  */
373   unsigned char c99;
374 
375   /* Nonzero if we are conforming to a specific C or C++ standard.  */
376   unsigned char std;
377 
378   /* Nonzero means give all the error messages the ANSI standard requires.  */
379   unsigned char pedantic;
380 
381   /* Nonzero means we're looking at already preprocessed code, so don't
382      bother trying to do macro expansion and whatnot.  */
383   unsigned char preprocessed;
384 
385   /* Print column number in error messages.  */
386   unsigned char show_column;
387 
388   /* Nonzero means handle C++ alternate operator names.  */
389   unsigned char operator_names;
390 
391   /* True for traditional preprocessing.  */
392   unsigned char traditional;
393 
394   /* Holds the name of the target (execution) character set.  */
395   const char *narrow_charset;
396 
397   /* Holds the name of the target wide character set.  */
398   const char *wide_charset;
399 
400   /* Holds the name of the input character set.  */
401   const char *input_charset;
402 
403   /* The minimum permitted level of normalization before a warning
404      is generated.  */
405   enum cpp_normalize_level warn_normalize;
406 
407   /* True to warn about precompiled header files we couldn't use.  */
408   bool warn_invalid_pch;
409 
410   /* True if dependencies should be restored from a precompiled header.  */
411   bool restore_pch_deps;
412 
413   /* Dependency generation.  */
414   struct
415   {
416     /* Style of header dependencies to generate.  */
417     enum cpp_deps_style style;
418 
419     /* Assume missing files are generated files.  */
420     bool missing_files;
421 
422     /* Generate phony targets for each dependency apart from the first
423        one.  */
424     bool phony_targets;
425 
426     /* If true, no dependency is generated on the main file.  */
427     bool ignore_main_file;
428   } deps;
429 
430   /* Target-specific features set by the front end or client.  */
431 
432   /* Precision for target CPP arithmetic, target characters, target
433      ints and target wide characters, respectively.  */
434   size_t precision, char_precision, int_precision, wchar_precision;
435 
436   /* True means chars (wide chars) are unsigned.  */
437   bool unsigned_char, unsigned_wchar;
438 
439   /* True if the most significant byte in a word has the lowest
440      address in memory.  */
441   bool bytes_big_endian;
442 
443   /* Nonzero means __STDC__ should have the value 0 in system headers.  */
444   unsigned char stdc_0_in_system_headers;
445 
446   /* True means error callback should be used for diagnostics.  */
447   bool client_diagnostic;
448 
449   /* True disables tokenization outside of preprocessing directives. */
450   bool directives_only;
451 };
452 
453 /* Callback for header lookup for HEADER, which is the name of a
454    source file.  It is used as a method of last resort to find headers
455    that are not otherwise found during the normal include processing.
456    The return value is the malloced name of a header to try and open,
457    if any, or NULL otherwise.  This callback is called only if the
458    header is otherwise unfound.  */
459 typedef const char *(*missing_header_cb)(cpp_reader *, const char *header, cpp_dir **);
460 
461 /* Call backs to cpplib client.  */
462 struct cpp_callbacks
463 {
464   /* Called when a new line of preprocessed output is started.  */
465   void (*line_change) (cpp_reader *, const cpp_token *, int);
466 
467   /* Called when switching to/from a new file.
468      The line_map is for the new file.  It is NULL if there is no new file.
469      (In C this happens when done with <built-in>+<command line> and also
470      when done with a main file.)  This can be used for resource cleanup.  */
471   void (*file_change) (cpp_reader *, const struct line_map *);
472 
473   void (*dir_change) (cpp_reader *, const char *);
474   void (*include) (cpp_reader *, unsigned int, const unsigned char *,
475 		   const char *, int, const cpp_token **);
476   void (*define) (cpp_reader *, unsigned int, cpp_hashnode *);
477   void (*undef) (cpp_reader *, unsigned int, cpp_hashnode *);
478   void (*ident) (cpp_reader *, unsigned int, const cpp_string *);
479   void (*def_pragma) (cpp_reader *, unsigned int);
480   int (*valid_pch) (cpp_reader *, const char *, int);
481   void (*read_pch) (cpp_reader *, const char *, int, const char *);
482   missing_header_cb missing_header;
483 
484   /* Called to emit a diagnostic if client_diagnostic option is true.
485      This callback receives the translated message.  */
486   void (*error) (cpp_reader *, int, const char *, va_list *)
487        ATTRIBUTE_FPTR_PRINTF(3,0);
488 };
489 
490 /* Chain of directories to look for include files in.  */
491 struct cpp_dir
492 {
493   /* NULL-terminated singly-linked list.  */
494   struct cpp_dir *next;
495 
496   /* NAME of the directory, NUL-terminated.  */
497   char *name;
498   unsigned int len;
499 
500   /* One if a system header, two if a system header that has extern
501      "C" guards for C++.  */
502   unsigned char sysp;
503 
504   /* Mapping of file names for this directory for MS-DOS and related
505      platforms.  A NULL-terminated array of (from, to) pairs.  */
506   const char **name_map;
507 
508   /* Routine to construct pathname, given the search path name and the
509      HEADER we are trying to find, return a constructed pathname to
510      try and open.  If this is NULL, the constructed pathname is as
511      constructed by append_file_to_dir.  */
512   char *(*construct) (const char *header, cpp_dir *dir);
513 
514   /* The C front end uses these to recognize duplicated
515      directories in the search path.  */
516   ino_t ino;
517   dev_t dev;
518 
519   /* Is this a user-supplied directory? */
520   bool user_supplied_p;
521 };
522 
523 /* Name under which this program was invoked.  */
524 extern const char *progname;
525 
526 /* The structure of a node in the hash table.  The hash table has
527    entries for all identifiers: either macros defined by #define
528    commands (type NT_MACRO), assertions created with #assert
529    (NT_ASSERTION), or neither of the above (NT_VOID).  Builtin macros
530    like __LINE__ are flagged NODE_BUILTIN.  Poisoned identifiers are
531    flagged NODE_POISONED.  NODE_OPERATOR (C++ only) indicates an
532    identifier that behaves like an operator such as "xor".
533    NODE_DIAGNOSTIC is for speed in lex_token: it indicates a
534    diagnostic may be required for this node.  Currently this only
535    applies to __VA_ARGS__ and poisoned identifiers.  */
536 
537 /* Hash node flags.  */
538 #define NODE_OPERATOR	(1 << 0)	/* C++ named operator.  */
539 #define NODE_POISONED	(1 << 1)	/* Poisoned identifier.  */
540 #define NODE_BUILTIN	(1 << 2)	/* Builtin macro.  */
541 #define NODE_DIAGNOSTIC (1 << 3)	/* Possible diagnostic when lexed.  */
542 #define NODE_WARN	(1 << 4)	/* Warn if redefined or undefined.  */
543 #define NODE_DISABLED	(1 << 5)	/* A disabled macro.  */
544 #define NODE_MACRO_ARG	(1 << 6)	/* Used during #define processing.  */
545 
546 /* Different flavors of hash node.  */
547 enum node_type
548 {
549   NT_VOID = 0,	   /* No definition yet.  */
550   NT_MACRO,	   /* A macro of some form.  */
551   NT_ASSERTION	   /* Predicate for #assert.  */
552 };
553 
554 /* Different flavors of builtin macro.  _Pragma is an operator, but we
555    handle it with the builtin code for efficiency reasons.  */
556 enum builtin_type
557 {
558   BT_SPECLINE = 0,		/* `__LINE__' */
559   BT_DATE,			/* `__DATE__' */
560   BT_FILE,			/* `__FILE__' */
561   BT_BASE_FILE,			/* `__BASE_FILE__' */
562   BT_INCLUDE_LEVEL,		/* `__INCLUDE_LEVEL__' */
563   BT_TIME,			/* `__TIME__' */
564   BT_STDC,			/* `__STDC__' */
565   BT_PRAGMA,			/* `_Pragma' operator */
566   BT_TIMESTAMP,			/* `__TIMESTAMP__' */
567   BT_COUNTER			/* `__COUNTER__' */
568 };
569 
570 #define CPP_HASHNODE(HNODE)	((cpp_hashnode *) (HNODE))
571 #define HT_NODE(NODE)		((ht_identifier *) (NODE))
572 #define NODE_LEN(NODE)		HT_LEN (&(NODE)->ident)
573 #define NODE_NAME(NODE)		HT_STR (&(NODE)->ident)
574 
575 /* Specify which field, if any, of the union is used.  */
576 
577 enum {
578   NTV_MACRO,
579   NTV_ANSWER,
580   NTV_BUILTIN,
581   NTV_ARGUMENT,
582   NTV_NONE
583 };
584 
585 #define CPP_HASHNODE_VALUE_IDX(HNODE)				\
586   ((HNODE.flags & NODE_MACRO_ARG) ? NTV_ARGUMENT		\
587    : HNODE.type == NT_MACRO ? ((HNODE.flags & NODE_BUILTIN) 	\
588 			       ? NTV_BUILTIN : NTV_MACRO)	\
589    : HNODE.type == NT_ASSERTION ? NTV_ANSWER			\
590    : NTV_NONE)
591 
592 /* The common part of an identifier node shared amongst all 3 C front
593    ends.  Also used to store CPP identifiers, which are a superset of
594    identifiers in the grammatical sense.  */
595 
596 union _cpp_hashnode_value GTY(())
597 {
598   /* If a macro.  */
599   cpp_macro * GTY((tag ("NTV_MACRO"))) macro;
600   /* Answers to an assertion.  */
601   struct answer * GTY ((tag ("NTV_ANSWER"))) answers;
602   /* Code for a builtin macro.  */
603   enum builtin_type GTY ((tag ("NTV_BUILTIN"))) builtin;
604   /* Macro argument index.  */
605   unsigned short GTY ((tag ("NTV_ARGUMENT"))) arg_index;
606 };
607 
608 struct cpp_hashnode GTY(())
609 {
610   struct ht_identifier ident;
611   unsigned int is_directive : 1;
612   unsigned int directive_index : 7;	/* If is_directive,
613 					   then index into directive table.
614 					   Otherwise, a NODE_OPERATOR.  */
615   unsigned char rid_code;		/* Rid code - for front ends.  */
616   ENUM_BITFIELD(node_type) type : 8;	/* CPP node type.  */
617   unsigned char flags;			/* CPP flags.  */
618 
619   union _cpp_hashnode_value GTY ((desc ("CPP_HASHNODE_VALUE_IDX (%1)"))) value;
620 };
621 
622 /* Call this first to get a handle to pass to other functions.
623 
624    If you want cpplib to manage its own hashtable, pass in a NULL
625    pointer.  Otherwise you should pass in an initialized hash table
626    that cpplib will share; this technique is used by the C front
627    ends.  */
628 extern cpp_reader *cpp_create_reader (enum c_lang, struct ht *,
629 				      struct line_maps *);
630 
631 /* Call this to change the selected language standard (e.g. because of
632    command line options).  */
633 extern void cpp_set_lang (cpp_reader *, enum c_lang);
634 
635 /* Set the include paths.  */
636 extern void cpp_set_include_chains (cpp_reader *, cpp_dir *, cpp_dir *, int);
637 
638 /* Call these to get pointers to the options, callback, and deps
639    structures for a given reader.  These pointers are good until you
640    call cpp_finish on that reader.  You can either edit the callbacks
641    through the pointer returned from cpp_get_callbacks, or set them
642    with cpp_set_callbacks.  */
643 extern cpp_options *cpp_get_options (cpp_reader *);
644 extern cpp_callbacks *cpp_get_callbacks (cpp_reader *);
645 extern void cpp_set_callbacks (cpp_reader *, cpp_callbacks *);
646 extern struct deps *cpp_get_deps (cpp_reader *);
647 
648 /* This function reads the file, but does not start preprocessing.  It
649    returns the name of the original file; this is the same as the
650    input file, except for preprocessed input.  This will generate at
651    least one file change callback, and possibly a line change callback
652    too.  If there was an error opening the file, it returns NULL.  */
653 extern const char *cpp_read_main_file (cpp_reader *, const char *);
654 
655 /* Set up built-ins with special behavior.  Use cpp_init_builtins()
656    instead unless your know what you are doing.  */
657 extern void cpp_init_special_builtins (cpp_reader *);
658 
659 /* Set up built-ins like __FILE__.  */
660 extern void cpp_init_builtins (cpp_reader *, int);
661 
662 /* This is called after options have been parsed, and partially
663    processed.  */
664 extern void cpp_post_options (cpp_reader *);
665 
666 /* Set up translation to the target character set.  */
667 extern void cpp_init_iconv (cpp_reader *);
668 
669 /* Call this to finish preprocessing.  If you requested dependency
670    generation, pass an open stream to write the information to,
671    otherwise NULL.  It is your responsibility to close the stream.
672 
673    Returns cpp_errors (pfile).  */
674 extern int cpp_finish (cpp_reader *, FILE *deps_stream);
675 
676 /* Call this to release the handle at the end of preprocessing.  Any
677    use of the handle after this function returns is invalid.  Returns
678    cpp_errors (pfile).  */
679 extern void cpp_destroy (cpp_reader *);
680 
681 /* Error count.  */
682 extern unsigned int cpp_errors (cpp_reader *);
683 
684 extern unsigned int cpp_token_len (const cpp_token *);
685 extern unsigned char *cpp_token_as_text (cpp_reader *, const cpp_token *);
686 extern unsigned char *cpp_spell_token (cpp_reader *, const cpp_token *,
687 				       unsigned char *, bool);
688 extern void cpp_register_pragma (cpp_reader *, const char *, const char *,
689 				 void (*) (cpp_reader *), bool);
690 extern void cpp_register_deferred_pragma (cpp_reader *, const char *,
691 					  const char *, unsigned, bool, bool);
692 extern int cpp_avoid_paste (cpp_reader *, const cpp_token *,
693 			    const cpp_token *);
694 extern const cpp_token *cpp_get_token (cpp_reader *);
695 extern const unsigned char *cpp_macro_definition (cpp_reader *,
696 						  const cpp_hashnode *);
697 extern void _cpp_backup_tokens (cpp_reader *, unsigned int);
698 
699 /* Evaluate a CPP_CHAR or CPP_WCHAR token.  */
700 extern cppchar_t cpp_interpret_charconst (cpp_reader *, const cpp_token *,
701 					  unsigned int *, int *);
702 /* Evaluate a vector of CPP_STRING or CPP_WSTRING tokens.  */
703 extern bool cpp_interpret_string (cpp_reader *,
704 				  const cpp_string *, size_t,
705 				  cpp_string *, bool);
706 extern bool cpp_interpret_string_notranslate (cpp_reader *,
707 					      const cpp_string *, size_t,
708 					      cpp_string *, bool);
709 
710 /* Convert a host character constant to the execution character set.  */
711 extern cppchar_t cpp_host_to_exec_charset (cpp_reader *, cppchar_t);
712 
713 /* Used to register macros and assertions, perhaps from the command line.
714    The text is the same as the command line argument.  */
715 extern void cpp_define (cpp_reader *, const char *);
716 extern void cpp_assert (cpp_reader *, const char *);
717 extern void cpp_undef (cpp_reader *, const char *);
718 extern void cpp_unassert (cpp_reader *, const char *);
719 
720 /* Undefine all macros and assertions.  */
721 extern void cpp_undef_all (cpp_reader *);
722 
723 extern cpp_buffer *cpp_push_buffer (cpp_reader *, const unsigned char *,
724 				    size_t, int);
725 extern int cpp_defined (cpp_reader *, const unsigned char *, int);
726 
727 /* A preprocessing number.  Code assumes that any unused high bits of
728    the double integer are set to zero.  */
729 typedef unsigned HOST_WIDE_INT cpp_num_part;
730 typedef struct cpp_num cpp_num;
731 struct cpp_num
732 {
733   cpp_num_part high;
734   cpp_num_part low;
735   bool unsignedp;  /* True if value should be treated as unsigned.  */
736   bool overflow;   /* True if the most recent calculation overflowed.  */
737 };
738 
739 /* cpplib provides two interfaces for interpretation of preprocessing
740    numbers.
741 
742    cpp_classify_number categorizes numeric constants according to
743    their field (integer, floating point, or invalid), radix (decimal,
744    octal, hexadecimal), and type suffixes.  */
745 
746 #define CPP_N_CATEGORY  0x000F
747 #define CPP_N_INVALID	0x0000
748 #define CPP_N_INTEGER	0x0001
749 #define CPP_N_FLOATING	0x0002
750 
751 #define CPP_N_WIDTH	0x00F0
752 #define CPP_N_SMALL	0x0010	/* int, float.  */
753 #define CPP_N_MEDIUM	0x0020	/* long, double.  */
754 #define CPP_N_LARGE	0x0040	/* long long, long double.  */
755 
756 #define CPP_N_RADIX	0x0F00
757 #define CPP_N_DECIMAL	0x0100
758 #define CPP_N_HEX	0x0200
759 #define CPP_N_OCTAL	0x0400
760 #define CPP_N_BINARY	0x0800
761 
762 #define CPP_N_UNSIGNED	0x1000	/* Properties.  */
763 #define CPP_N_IMAGINARY	0x2000
764 #define CPP_N_DFLOAT	0x4000
765 #define CPP_N_DEFAULT	0x8000
766 
767 /* Classify a CPP_NUMBER token.  The return value is a combination of
768    the flags from the above sets.  */
769 extern unsigned cpp_classify_number (cpp_reader *, const cpp_token *);
770 
771 /* Evaluate a token classified as category CPP_N_INTEGER.  */
772 extern cpp_num cpp_interpret_integer (cpp_reader *, const cpp_token *,
773 				      unsigned int type);
774 
775 /* Sign extend a number, with PRECISION significant bits and all
776    others assumed clear, to fill out a cpp_num structure.  */
777 cpp_num cpp_num_sign_extend (cpp_num, size_t);
778 
779 /* Diagnostic levels.  To get a diagnostic without associating a
780    position in the translation unit with it, use cpp_error_with_line
781    with a line number of zero.  */
782 
783 /* Warning, an error with -Werror.  */
784 #define CPP_DL_WARNING		0x00
785 /* Same as CPP_DL_WARNING, except it is not suppressed in system headers.  */
786 #define CPP_DL_WARNING_SYSHDR	0x01
787 /* Warning, an error with -pedantic-errors or -Werror.  */
788 #define CPP_DL_PEDWARN		0x02
789 /* An error.  */
790 #define CPP_DL_ERROR		0x03
791 /* An internal consistency check failed.  Prints "internal error: ",
792    otherwise the same as CPP_DL_ERROR.  */
793 #define CPP_DL_ICE		0x04
794 /* Extracts a diagnostic level from an int.  */
795 #define CPP_DL_EXTRACT(l)	(l & 0xf)
796 /* Nonzero if a diagnostic level is one of the warnings.  */
797 #define CPP_DL_WARNING_P(l)	(CPP_DL_EXTRACT (l) >= CPP_DL_WARNING \
798 				 && CPP_DL_EXTRACT (l) <= CPP_DL_PEDWARN)
799 
800 /* Output a diagnostic of some kind.  */
801 extern void cpp_error (cpp_reader *, int, const char *msgid, ...)
802   ATTRIBUTE_PRINTF_3;
803 
804 /* Output a diagnostic with "MSGID: " preceding the
805    error string of errno.  No location is printed.  */
806 extern void cpp_errno (cpp_reader *, int, const char *msgid);
807 
808 /* Same as cpp_error, except additionally specifies a position as a
809    (translation unit) physical line and physical column.  If the line is
810    zero, then no location is printed.  */
811 extern void cpp_error_with_line (cpp_reader *, int, source_location, unsigned,
812 				 const char *msgid, ...) ATTRIBUTE_PRINTF_5;
813 
814 /* In cpplex.c */
815 extern int cpp_ideq (const cpp_token *, const char *);
816 extern void cpp_output_line (cpp_reader *, FILE *);
817 extern void cpp_output_token (const cpp_token *, FILE *);
818 extern const char *cpp_type2name (enum cpp_ttype);
819 /* Returns the value of an escape sequence, truncated to the correct
820    target precision.  PSTR points to the input pointer, which is just
821    after the backslash.  LIMIT is how much text we have.  WIDE is true
822    if the escape sequence is part of a wide character constant or
823    string literal.  Handles all relevant diagnostics.  */
824 extern cppchar_t cpp_parse_escape (cpp_reader *, const unsigned char ** pstr,
825 				   const unsigned char *limit, int wide);
826 
827 /* In cpphash.c */
828 
829 /* Lookup an identifier in the hashtable.  Puts the identifier in the
830    table if it is not already there.  */
831 extern cpp_hashnode *cpp_lookup (cpp_reader *, const unsigned char *,
832 				 unsigned int);
833 
834 typedef int (*cpp_cb) (cpp_reader *, cpp_hashnode *, void *);
835 extern void cpp_forall_identifiers (cpp_reader *, cpp_cb, void *);
836 
837 /* In cppmacro.c */
838 extern void cpp_scan_nooutput (cpp_reader *);
839 extern int  cpp_sys_macro_p (cpp_reader *);
840 extern unsigned char *cpp_quote_string (unsigned char *, const unsigned char *,
841 					unsigned int);
842 
843 /* In cppfiles.c */
844 extern bool cpp_included (cpp_reader *, const char *);
845 extern void cpp_make_system_header (cpp_reader *, int, int);
846 extern bool cpp_push_include (cpp_reader *, const char *);
847 extern void cpp_change_file (cpp_reader *, enum lc_reason, const char *);
848 extern const char *cpp_get_path (struct _cpp_file *);
849 extern cpp_dir *cpp_get_dir (struct _cpp_file *);
850 extern cpp_buffer *cpp_get_buffer (cpp_reader *);
851 extern struct _cpp_file *cpp_get_file (cpp_buffer *);
852 extern cpp_buffer *cpp_get_prev (cpp_buffer *);
853 
854 /* In cpppch.c */
855 struct save_macro_data;
856 extern int cpp_save_state (cpp_reader *, FILE *);
857 extern int cpp_write_pch_deps (cpp_reader *, FILE *);
858 extern int cpp_write_pch_state (cpp_reader *, FILE *);
859 extern int cpp_valid_state (cpp_reader *, const char *, int);
860 extern void cpp_prepare_state (cpp_reader *, struct save_macro_data **);
861 extern int cpp_read_state (cpp_reader *, const char *, FILE *,
862 			   struct save_macro_data *);
863 
864 #ifdef __cplusplus
865 }
866 #endif
867 
868 #endif /* ! LIBCPP_CPPLIB_H */
869