1 /* Definitions for CPP library.
2 Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001
3 Free Software Foundation, Inc.
4 Written by Per Bothner, 1994-95.
6 This program is free software; you can redistribute it and/or modify it
7 under the terms of the GNU General Public License as published by the
8 Free Software Foundation; either version 2, or (at your option) any
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, write to the Free Software
18 Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
20 In other words, you are welcome to use, share and improve this program.
21 You are forbidden to forbid anyone else to use, share and improve
22 what you give them. Help stamp out software-hoarding! */
26 #include <sys/types.h>
27 #include "hashtable.h"
34 /* For complex reasons, cpp_reader is also typedefed in c-pragma.h. */
35 #ifndef GCC_C_PRAGMA_H
36 typedef struct cpp_reader cpp_reader;
38 typedef struct cpp_buffer cpp_buffer;
39 typedef struct cpp_options cpp_options;
40 typedef struct cpp_token cpp_token;
41 typedef struct cpp_string cpp_string;
42 typedef struct cpp_hashnode cpp_hashnode;
43 typedef struct cpp_macro cpp_macro;
44 typedef struct cpp_lexer_pos cpp_lexer_pos;
45 typedef struct cpp_lookahead cpp_lookahead;
46 typedef struct cpp_callbacks cpp_callbacks;
49 struct file_name_map_list;
52 /* The first two groups, apart from '=', can appear in preprocessor
53 expressions. This allows a lookup table to be implemented in
56 The first group, to CPP_LAST_EQ, can be immediately followed by an
57 '='. The lexer needs operators ending in '=', like ">>=", to be in
58 the same order as their counterparts without the '=', like ">>". */
60 /* Positions in the table. */
61 #define CPP_LAST_EQ CPP_MAX
62 #define CPP_FIRST_DIGRAPH CPP_HASH
63 #define CPP_LAST_PUNCTUATOR CPP_DOT_STAR
68 OP(CPP_GREATER, ">") /* compare */ \
70 OP(CPP_PLUS, "+") /* math */ \
75 OP(CPP_AND, "&") /* bit ops */ \
78 OP(CPP_RSHIFT, ">>") \
79 OP(CPP_LSHIFT, "<<") \
80 OP(CPP_MIN, "<?") /* extension */ \
84 OP(CPP_AND_AND, "&&") /* logical */ \
88 OP(CPP_COMMA, ",") /* grouping */ \
89 OP(CPP_OPEN_PAREN, "(") \
90 OP(CPP_CLOSE_PAREN, ")") \
91 OP(CPP_EQ_EQ, "==") /* compare */ \
92 OP(CPP_NOT_EQ, "!=") \
93 OP(CPP_GREATER_EQ, ">=") \
94 OP(CPP_LESS_EQ, "<=") \
96 OP(CPP_PLUS_EQ, "+=") /* math */ \
97 OP(CPP_MINUS_EQ, "-=") \
98 OP(CPP_MULT_EQ, "*=") \
99 OP(CPP_DIV_EQ, "/=") \
100 OP(CPP_MOD_EQ, "%=") \
101 OP(CPP_AND_EQ, "&=") /* bit ops */ \
102 OP(CPP_OR_EQ, "|=") \
103 OP(CPP_XOR_EQ, "^=") \
104 OP(CPP_RSHIFT_EQ, ">>=") \
105 OP(CPP_LSHIFT_EQ, "<<=") \
106 OP(CPP_MIN_EQ, "<?=") /* extension */ \
107 OP(CPP_MAX_EQ, ">?=") \
108 /* Digraphs together, beginning with CPP_FIRST_DIGRAPH. */ \
109 OP(CPP_HASH, "#") /* digraphs */ \
110 OP(CPP_PASTE, "##") \
111 OP(CPP_OPEN_SQUARE, "[") \
112 OP(CPP_CLOSE_SQUARE, "]") \
113 OP(CPP_OPEN_BRACE, "{") \
114 OP(CPP_CLOSE_BRACE, "}") \
115 /* The remainder of the punctuation. Order is not significant. */ \
116 OP(CPP_SEMICOLON, ";") /* structure */ \
117 OP(CPP_ELLIPSIS, "...") \
118 OP(CPP_PLUS_PLUS, "++") /* increment */ \
119 OP(CPP_MINUS_MINUS, "--") \
120 OP(CPP_DEREF, "->") /* accessors */ \
122 OP(CPP_SCOPE, "::") \
123 OP(CPP_DEREF_STAR, "->*") \
124 OP(CPP_DOT_STAR, ".*") \
125 OP(CPP_ATSIGN, "@") /* used in Objective C */ \
127 TK(CPP_NAME, SPELL_IDENT) /* word */ \
128 TK(CPP_NUMBER, SPELL_STRING) /* 34_be+ta */ \
130 TK(CPP_CHAR, SPELL_STRING) /* 'char' */ \
131 TK(CPP_WCHAR, SPELL_STRING) /* L'char' */ \
132 TK(CPP_OTHER, SPELL_CHAR) /* stray punctuation */ \
134 TK(CPP_STRING, SPELL_STRING) /* "string" */ \
135 TK(CPP_WSTRING, SPELL_STRING) /* L"string" */ \
136 TK(CPP_HEADER_NAME, SPELL_STRING) /* <stdio.h> in #include */ \
138 TK(CPP_COMMENT, SPELL_STRING) /* Only if output comments. */ \
139 TK(CPP_MACRO_ARG, SPELL_NONE) /* Macro argument. */ \
140 OP(CPP_EOF, "EOL") /* End of line or file. */
152 /* C language kind, used when calling cpp_reader_init. */
153 enum c_lang {CLK_GNUC89 = 0, CLK_GNUC99, CLK_STDC89, CLK_STDC94, CLK_STDC99,
154 CLK_GNUCXX, CLK_CXX98, CLK_OBJC, CLK_OBJCXX, CLK_ASM};
156 /* Payload of a NUMBER, FLOAT, STRING, or COMMENT token. */
160 const unsigned char *text;
163 /* Flags for the cpp_token structure. */
164 #define PREV_WHITE (1 << 0) /* If whitespace before this token. */
165 #define DIGRAPH (1 << 1) /* If it was a digraph. */
166 #define STRINGIFY_ARG (1 << 2) /* If macro argument to be stringified. */
167 #define PASTE_LEFT (1 << 3) /* If on LHS of a ## operator. */
168 #define NAMED_OP (1 << 4) /* C++ named operators. */
169 #define NO_EXPAND (1 << 5) /* Do not macro-expand this token. */
170 #define AVOID_LPASTE (1 << 6) /* Check left for accidental pastes. */
172 /* A preprocessing token. This has been carefully packed and should
173 occupy 12 bytes on 32-bit hosts and 16 bytes on 64-bit hosts. */
176 unsigned int line; /* Logical line of first char of token. */
177 unsigned short col; /* Column of first char of token. */
178 ENUM_BITFIELD(cpp_ttype) type : CHAR_BIT; /* token type */
179 unsigned char flags; /* flags - see above */
183 cpp_hashnode *node; /* An identifier. */
184 struct cpp_string str; /* A string, or number. */
185 unsigned int arg_no; /* Argument no. for a CPP_MACRO_ARG. */
186 unsigned char c; /* Character represented by CPP_OTHER. */
190 /* The position of a token in the current file. */
194 unsigned int output_line;
198 typedef struct cpp_token_with_pos cpp_token_with_pos;
199 struct cpp_token_with_pos
205 /* Token lookahead. */
208 struct cpp_lookahead *next;
209 cpp_token_with_pos *tokens;
211 unsigned int cur, count, cap;
214 /* A standalone character. We may want to make it unsigned for the
215 same reason we use unsigned char - to avoid signedness issues. */
216 typedef int cppchar_t;
218 /* Values for opts.dump_macros.
219 dump_only means inhibit output of the preprocessed text
220 and instead output the definitions of all user-defined
221 macros in a form suitable for use as input to cpp.
222 dump_names means pass #define and the macro name through to output.
223 dump_definitions means pass the whole definition (plus #define) through
225 enum { dump_none = 0, dump_only, dump_names, dump_definitions };
227 /* This structure is nested inside struct cpp_reader, and
228 carries all the options visible to the command line. */
231 /* Name of input and output files. */
232 const char *in_fname;
233 const char *out_fname;
235 /* Characters between tab stops. */
236 unsigned int tabstop;
238 /* Pending options - -D, -U, -A, -I, -ixxx. */
239 struct cpp_pending *pending;
241 /* File name which deps are being written to. This is 0 if deps are
242 being written to stdout. */
243 const char *deps_file;
245 /* Search paths for include files. */
246 struct search_path *quote_include; /* "" */
247 struct search_path *bracket_include; /* <> */
249 /* Map between header names and file names, used only on DOS where
250 file names are limited in length. */
251 struct file_name_map_list *map_list;
253 /* Directory prefix that should replace `/usr/lib/gcc-lib/TARGET/VERSION'
254 in the standard include file directories. */
255 const char *include_prefix;
256 unsigned int include_prefix_len;
258 /* -fleading_underscore sets this to "_". */
259 const char *user_label_prefix;
261 /* The language we're preprocessing. */
264 /* Non-0 means -v, so print the full set of include dirs. */
265 unsigned char verbose;
267 /* Nonzero means use extra default include directories for C++. */
268 unsigned char cplusplus;
270 /* Nonzero means handle cplusplus style comments */
271 unsigned char cplusplus_comments;
273 /* Nonzero means handle #import, for objective C. */
276 /* Nonzero means don't copy comments into the output file. */
277 unsigned char discard_comments;
279 /* Nonzero means process the ISO trigraph sequences. */
280 unsigned char trigraphs;
282 /* Nonzero means process the ISO digraph sequences. */
283 unsigned char digraphs;
285 /* Nonzero means to allow hexadecimal floats and LL suffixes. */
286 unsigned char extended_numbers;
288 /* Nonzero means print the names of included files rather than the
289 preprocessed output. 1 means just the #include "...", 2 means
290 #include <...> as well. */
291 unsigned char print_deps;
293 /* Nonzero if phony targets are created for each header. */
294 unsigned char deps_phony_targets;
296 /* Nonzero if missing .h files in -M output are assumed to be
297 generated files and not errors. */
298 unsigned char print_deps_missing_files;
300 /* If true, fopen (deps_file, "a") else fopen (deps_file, "w"). */
301 unsigned char print_deps_append;
303 /* Nonzero means print names of header files (-H). */
304 unsigned char print_include_names;
306 /* Nonzero means cpp_pedwarn causes a hard error. */
307 unsigned char pedantic_errors;
309 /* Nonzero means don't print warning messages. */
310 unsigned char inhibit_warnings;
312 /* Nonzero means don't suppress warnings from system headers. */
313 unsigned char warn_system_headers;
315 /* Nonzero means don't print error messages. Has no option to
316 select it, but can be set by a user of cpplib (e.g. fix-header). */
317 unsigned char inhibit_errors;
319 /* Nonzero means warn if slash-star appears in a comment. */
320 unsigned char warn_comments;
322 /* Nonzero means warn if there are any trigraphs. */
323 unsigned char warn_trigraphs;
325 /* Nonzero means warn if #import is used. */
326 unsigned char warn_import;
328 /* Nonzero means warn about various incompatibilities with
330 unsigned char warn_traditional;
332 /* Nonzero means turn warnings into errors. */
333 unsigned char warnings_are_errors;
335 /* Nonzero causes output not to be done, but directives such as
336 #define that have side effects are still obeyed. */
337 unsigned char no_output;
339 /* Nonzero means we should look for header.gcc files that remap file
343 /* Nonzero means don't output line number information. */
344 unsigned char no_line_commands;
346 /* Nonzero means -I- has been seen, so don't look for #include "foo"
347 the source-file directory. */
348 unsigned char ignore_srcdir;
350 /* Zero means dollar signs are punctuation. */
351 unsigned char dollars_in_ident;
353 /* Nonzero means warn if undefined identifiers are evaluated in an #if. */
354 unsigned char warn_undef;
356 /* Nonzero for the 1999 C Standard, including corrigenda and amendments. */
359 /* Nonzero means give all the error messages the ANSI standard requires. */
360 unsigned char pedantic;
362 /* Nonzero means we're looking at already preprocessed code, so don't
363 bother trying to do macro expansion and whatnot. */
364 unsigned char preprocessed;
366 /* Nonzero disables all the standard directories for headers. */
367 unsigned char no_standard_includes;
369 /* Nonzero disables the C++-specific standard directories for headers. */
370 unsigned char no_standard_cplusplus_includes;
372 /* Nonzero means dump macros in some fashion - see above. */
373 unsigned char dump_macros;
375 /* Nonzero means pass #include lines through to the output. */
376 unsigned char dump_includes;
378 /* Print column number in error messages. */
379 unsigned char show_column;
381 /* Treat C++ alternate operator names special. */
382 unsigned char operator_names;
384 /* True if --help, --version or --target-help appeared in the
385 options. Stand-alone CPP should then bail out after option
386 parsing; drivers might want to continue printing help. */
387 unsigned char help_only;
390 typedef struct cpp_file_change cpp_file_change;
391 struct cpp_file_change
393 struct line_map *map; /* Line map to use until next callback. */
394 unsigned int line; /* Logical line number of next line. */
395 enum lc_reason reason; /* Reason for change. */
396 unsigned char sysp; /* Nonzero if system header. */
397 unsigned char externc; /* Nonzero if wrapper needed. */
403 void (*file_change) PARAMS ((cpp_reader *, const cpp_file_change *));
404 void (*include) PARAMS ((cpp_reader *, const unsigned char *,
406 void (*define) PARAMS ((cpp_reader *, cpp_hashnode *));
407 void (*undef) PARAMS ((cpp_reader *, cpp_hashnode *));
408 void (*ident) PARAMS ((cpp_reader *, const cpp_string *));
409 void (*def_pragma) PARAMS ((cpp_reader *));
412 #define CPP_FATAL_LIMIT 1000
413 /* True if we have seen a "fatal" error. */
414 #define CPP_FATAL_ERRORS(PFILE) (cpp_errors (PFILE) >= CPP_FATAL_LIMIT)
416 /* Name under which this program was invoked. */
417 extern const char *progname;
419 /* Where does this buffer come from? A faked include, a source file,
420 a builtin macro, a command-line option, or a _Pragma operator. */
421 enum cpp_buffer_type {BUF_FAKE, BUF_FILE, BUF_BUILTIN,
422 BUF_CL_OPTION, BUF_PRAGMA};
424 /* The structure of a node in the hash table. The hash table has
425 entries for all identifiers: either macros defined by #define
426 commands (type NT_MACRO), assertions created with #assert
427 (NT_ASSERTION), or neither of the above (NT_VOID). Builtin macros
428 like __LINE__ are flagged NODE_BUILTIN. Poisioned identifiers are
429 flagged NODE_POISONED. NODE_OPERATOR (C++ only) indicates an
430 identifier that behaves like an operator such as "xor".
431 NODE_DIAGNOSTIC is for speed in lex_token: it indicates a
432 diagnostic may be required for this node. Currently this only
433 applies to __VA_ARGS__ and poisoned identifiers. */
435 /* Hash node flags. */
436 #define NODE_OPERATOR (1 << 0) /* C++ named operator. */
437 #define NODE_POISONED (1 << 1) /* Poisoned identifier. */
438 #define NODE_BUILTIN (1 << 2) /* Builtin macro. */
439 #define NODE_DIAGNOSTIC (1 << 3) /* Possible diagnostic when lexed. */
440 #define NODE_WARN (1 << 4) /* Warn if redefined or undefined. */
442 /* Different flavors of hash node. */
445 NT_VOID = 0, /* No definition yet. */
446 NT_MACRO, /* A macro of some form. */
447 NT_ASSERTION /* Predicate for #assert. */
450 /* Different flavors of builtin macro. */
453 BT_SPECLINE = 0, /* `__LINE__' */
454 BT_DATE, /* `__DATE__' */
455 BT_FILE, /* `__FILE__' */
456 BT_BASE_FILE, /* `__BASE_FILE__' */
457 BT_INCLUDE_LEVEL, /* `__INCLUDE_LEVEL__' */
458 BT_TIME, /* `__TIME__' */
459 BT_STDC /* `__STDC__' */
462 #define CPP_HASHNODE(HNODE) ((cpp_hashnode *) (HNODE))
463 #define HT_NODE(NODE) ((ht_identifier *) (NODE))
464 #define NODE_LEN(NODE) HT_LEN (&(NODE)->ident)
465 #define NODE_NAME(NODE) HT_STR (&(NODE)->ident)
467 /* The common part of an identifier node shared amongst all 3 C front
468 ends. Also used to store CPP identifiers, which are a superset of
469 identifiers in the grammatical sense. */
472 struct ht_identifier ident;
473 unsigned short arg_index; /* Macro argument index. */
474 unsigned char directive_index; /* Index into directive table. */
475 unsigned char rid_code; /* Rid code - for front ends. */
476 ENUM_BITFIELD(node_type) type : 8; /* CPP node type. */
477 unsigned char flags; /* CPP flags. */
481 cpp_macro *macro; /* If a macro. */
482 struct answer *answers; /* Answers to an assertion. */
483 enum cpp_ttype operator; /* Code for a named operator. */
484 enum builtin_type builtin; /* Code for a builtin macro. */
488 /* Call this first to get a handle to pass to other functions. If you
489 want cpplib to manage its own hashtable, pass in a NULL pointer.
490 Or you can pass in an initialised hash table that cpplib will use;
491 this technique is used by the C front ends. */
492 extern cpp_reader *cpp_create_reader PARAMS ((struct ht *,
495 /* Call this to release the handle. Any use of the handle after this
496 function returns is invalid. Returns cpp_errors (pfile). */
497 extern int cpp_destroy PARAMS ((cpp_reader *));
499 /* Call these to get pointers to the options and callback structures
500 for a given reader. These pointers are good until you call
501 cpp_finish on that reader. You can either edit the callbacks
502 through the pointer returned from cpp_get_callbacks, or set them
503 with cpp_set_callbacks. */
504 extern cpp_options *cpp_get_options PARAMS ((cpp_reader *));
505 extern struct line_maps *cpp_get_line_maps PARAMS ((cpp_reader *));
506 extern cpp_callbacks *cpp_get_callbacks PARAMS ((cpp_reader *));
507 extern void cpp_set_callbacks PARAMS ((cpp_reader *, cpp_callbacks *));
509 /* Now call cpp_handle_option[s] to handle 1[or more] switches. The
510 return value is the number of arguments used. If
511 cpp_handle_options returns without using all arguments, it couldn't
512 understand the next switch. When there are no switches left, you
513 must call cpp_post_options before calling cpp_start_read. Only
514 after cpp_post_options are the contents of the cpp_options
515 structure reliable. */
516 extern int cpp_handle_options PARAMS ((cpp_reader *, int, char **));
517 extern int cpp_handle_option PARAMS ((cpp_reader *, int, char **));
518 extern void cpp_post_options PARAMS ((cpp_reader *));
521 extern unsigned int cpp_errors PARAMS ((cpp_reader *));
523 extern unsigned int cpp_token_len PARAMS ((const cpp_token *));
524 extern unsigned char *cpp_token_as_text PARAMS ((cpp_reader *,
526 extern unsigned char *cpp_spell_token PARAMS ((cpp_reader *, const cpp_token *,
528 extern void cpp_register_pragma PARAMS ((cpp_reader *,
529 const char *, const char *,
530 void (*) PARAMS ((cpp_reader *))));
531 extern void cpp_register_pragma_space PARAMS ((cpp_reader *, const char *));
533 extern int cpp_start_read PARAMS ((cpp_reader *, const char *));
534 extern void cpp_finish PARAMS ((cpp_reader *));
535 extern int cpp_avoid_paste PARAMS ((cpp_reader *, const cpp_token *,
537 extern enum cpp_ttype cpp_can_paste PARAMS ((cpp_reader *, const cpp_token *,
538 const cpp_token *, int *));
539 extern void cpp_get_token PARAMS ((cpp_reader *, cpp_token *));
540 extern const cpp_lexer_pos *cpp_get_line PARAMS ((cpp_reader *));
541 extern const unsigned char *cpp_macro_definition PARAMS ((cpp_reader *,
542 const cpp_hashnode *));
544 /* Evaluate a CPP_CHAR or CPP_WCHAR token. */
546 cpp_interpret_charconst PARAMS ((cpp_reader *, const cpp_token *,
547 int, int, unsigned int *));
549 extern void cpp_define PARAMS ((cpp_reader *, const char *));
550 extern void cpp_assert PARAMS ((cpp_reader *, const char *));
551 extern void cpp_undef PARAMS ((cpp_reader *, const char *));
552 extern void cpp_unassert PARAMS ((cpp_reader *, const char *));
554 extern cpp_buffer *cpp_push_buffer PARAMS ((cpp_reader *,
555 const unsigned char *, size_t,
556 enum cpp_buffer_type,
558 extern int cpp_defined PARAMS ((cpp_reader *, const unsigned char *, int));
560 /* N.B. The error-message-printer prototypes have not been nicely
561 formatted because exgettext needs to see 'msgid' on the same line
562 as the name of the function in order to work properly. Only the
563 string argument gets a name in an effort to keep the lines from
564 getting ridiculously oversized. */
566 extern void cpp_ice PARAMS ((cpp_reader *, const char *msgid, ...))
568 extern void cpp_fatal PARAMS ((cpp_reader *, const char *msgid, ...))
570 extern void cpp_error PARAMS ((cpp_reader *, const char *msgid, ...))
572 extern void cpp_warning PARAMS ((cpp_reader *, const char *msgid, ...))
574 extern void cpp_pedwarn PARAMS ((cpp_reader *, const char *msgid, ...))
576 extern void cpp_notice PARAMS ((cpp_reader *, const char *msgid, ...))
578 extern void cpp_error_with_line PARAMS ((cpp_reader *, int, int, const char *msgid, ...))
580 extern void cpp_warning_with_line PARAMS ((cpp_reader *, int, int, const char *msgid, ...))
582 extern void cpp_pedwarn_with_line PARAMS ((cpp_reader *, int, int, const char *msgid, ...))
584 extern void cpp_pedwarn_with_file_and_line PARAMS ((cpp_reader *, const char *, int, int, const char *msgid, ...))
586 extern void cpp_error_from_errno PARAMS ((cpp_reader *, const char *));
587 extern void cpp_notice_from_errno PARAMS ((cpp_reader *, const char *));
590 extern int cpp_ideq PARAMS ((const cpp_token *,
592 extern void cpp_output_line PARAMS ((cpp_reader *, FILE *));
593 extern void cpp_output_token PARAMS ((const cpp_token *, FILE *));
594 extern const char *cpp_type2name PARAMS ((enum cpp_ttype));
595 extern unsigned int cpp_parse_escape PARAMS ((cpp_reader *,
596 const unsigned char **,
597 const unsigned char *,
598 unsigned HOST_WIDE_INT, int));
602 /* Lookup an identifier in the hashtable. Puts the identifier in the
603 table if it is not already there. */
604 extern cpp_hashnode *cpp_lookup PARAMS ((cpp_reader *,
605 const unsigned char *,
608 typedef int (*cpp_cb) PARAMS ((cpp_reader *, cpp_hashnode *, void *));
609 extern void cpp_forall_identifiers PARAMS ((cpp_reader *,
613 extern void cpp_scan_nooutput PARAMS ((cpp_reader *));
614 extern void cpp_start_lookahead PARAMS ((cpp_reader *));
615 extern void cpp_stop_lookahead PARAMS ((cpp_reader *, int));
616 extern int cpp_sys_macro_p PARAMS ((cpp_reader *));
619 extern int cpp_included PARAMS ((cpp_reader *, const char *));
620 extern void cpp_make_system_header PARAMS ((cpp_reader *, int, int));
626 #endif /* ! GCC_CPPLIB_H */