1 /* Definitions for CPP library.
2 Copyright (C) 1995-2022 Free Software Foundation, Inc.
3 Written by Per Bothner, 1994-95.
5 This program is free software; you can redistribute it and/or modify it
6 under the terms of the GNU General Public License as published by the
7 Free Software Foundation; either version 3, or (at your option) any
10 This program is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 GNU General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with this program; see the file COPYING3. If not see
17 <http://www.gnu.org/licenses/>.
19 In other words, you are welcome to use, share and improve this program.
20 You are forbidden to forbid anyone else to use, share and improve
21 what you give them. Help stamp out software-hoarding! */
22 #ifndef LIBCPP_CPPLIB_H
23 #define LIBCPP_CPPLIB_H
25 #include <sys/types.h>
29 typedef struct cpp_reader cpp_reader;
30 typedef struct cpp_buffer cpp_buffer;
31 typedef struct cpp_options cpp_options;
32 typedef struct cpp_token cpp_token;
33 typedef struct cpp_string cpp_string;
34 typedef struct cpp_hashnode cpp_hashnode;
35 typedef struct cpp_macro cpp_macro;
36 typedef struct cpp_callbacks cpp_callbacks;
37 typedef struct cpp_dir cpp_dir;
41 /* The first three groups, apart from '=', can appear in preprocessor
42 expressions (+= and -= are used to indicate unary + and - resp.).
43 This allows a lookup table to be implemented in _cpp_parse_expr.
45 The first group, to CPP_LAST_EQ, can be immediately followed by an
46 '='. The lexer needs operators ending in '=', like ">>=", to be in
47 the same order as their counterparts without the '=', like ">>".
49 See the cpp_operator table optab in expr.cc if you change the order or
50 add or remove anything in the first group. */
55 OP(GREATER, ">") /* compare */ \
57 OP(PLUS, "+") /* math */ \
62 OP(AND, "&") /* bit ops */ \
69 OP(AND_AND, "&&") /* logical */ \
73 OP(COMMA, ",") /* grouping */ \
75 OP(CLOSE_PAREN, ")") \
77 OP(EQ_EQ, "==") /* compare */ \
79 OP(GREATER_EQ, ">=") \
81 OP(SPACESHIP, "<=>") \
83 /* These two are unary + / - in preprocessor expressions. */ \
84 OP(PLUS_EQ, "+=") /* math */ \
90 OP(AND_EQ, "&=") /* bit ops */ \
93 OP(RSHIFT_EQ, ">>=") \
94 OP(LSHIFT_EQ, "<<=") \
95 /* Digraphs together, beginning with CPP_FIRST_DIGRAPH. */ \
96 OP(HASH, "#") /* digraphs */ \
98 OP(OPEN_SQUARE, "[") \
99 OP(CLOSE_SQUARE, "]") \
100 OP(OPEN_BRACE, "{") \
101 OP(CLOSE_BRACE, "}") \
102 /* The remainder of the punctuation. Order is not significant. */ \
103 OP(SEMICOLON, ";") /* structure */ \
104 OP(ELLIPSIS, "...") \
105 OP(PLUS_PLUS, "++") /* increment */ \
106 OP(MINUS_MINUS, "--") \
107 OP(DEREF, "->") /* accessors */ \
110 OP(DEREF_STAR, "->*") \
112 OP(ATSIGN, "@") /* used in Objective-C */ \
114 TK(NAME, IDENT) /* word */ \
115 TK(AT_NAME, IDENT) /* @word - Objective-C */ \
116 TK(NUMBER, LITERAL) /* 34_be+ta */ \
118 TK(CHAR, LITERAL) /* 'char' */ \
119 TK(WCHAR, LITERAL) /* L'char' */ \
120 TK(CHAR16, LITERAL) /* u'char' */ \
121 TK(CHAR32, LITERAL) /* U'char' */ \
122 TK(UTF8CHAR, LITERAL) /* u8'char' */ \
123 TK(OTHER, LITERAL) /* stray punctuation */ \
125 TK(STRING, LITERAL) /* "string" */ \
126 TK(WSTRING, LITERAL) /* L"string" */ \
127 TK(STRING16, LITERAL) /* u"string" */ \
128 TK(STRING32, LITERAL) /* U"string" */ \
129 TK(UTF8STRING, LITERAL) /* u8"string" */ \
130 TK(OBJC_STRING, LITERAL) /* @"string" - Objective-C */ \
131 TK(HEADER_NAME, LITERAL) /* <stdio.h> in #include */ \
133 TK(CHAR_USERDEF, LITERAL) /* 'char'_suffix - C++-0x */ \
134 TK(WCHAR_USERDEF, LITERAL) /* L'char'_suffix - C++-0x */ \
135 TK(CHAR16_USERDEF, LITERAL) /* u'char'_suffix - C++-0x */ \
136 TK(CHAR32_USERDEF, LITERAL) /* U'char'_suffix - C++-0x */ \
137 TK(UTF8CHAR_USERDEF, LITERAL) /* u8'char'_suffix - C++-0x */ \
138 TK(STRING_USERDEF, LITERAL) /* "string"_suffix - C++-0x */ \
139 TK(WSTRING_USERDEF, LITERAL) /* L"string"_suffix - C++-0x */ \
140 TK(STRING16_USERDEF, LITERAL) /* u"string"_suffix - C++-0x */ \
141 TK(STRING32_USERDEF, LITERAL) /* U"string"_suffix - C++-0x */ \
142 TK(UTF8STRING_USERDEF,LITERAL) /* u8"string"_suffix - C++-0x */ \
144 TK(COMMENT, LITERAL) /* Only if output comments. */ \
145 /* SPELL_LITERAL happens to DTRT. */ \
146 TK(MACRO_ARG, NONE) /* Macro argument. */ \
147 TK(PRAGMA, NONE) /* Only for deferred pragmas. */ \
148 TK(PRAGMA_EOL, NONE) /* End-of-line for deferred pragmas. */ \
149 TK(PADDING, NONE) /* Whitespace for -E. */
151 #define OP(e, s) CPP_ ## e,
152 #define TK(e, s) CPP_ ## e,
158 /* A token type for keywords, as opposed to ordinary identifiers. */
161 /* Positions in the table. */
162 CPP_LAST_EQ = CPP_LSHIFT,
163 CPP_FIRST_DIGRAPH = CPP_HASH,
164 CPP_LAST_PUNCTUATOR= CPP_ATSIGN,
165 CPP_LAST_CPP_OP = CPP_LESS_EQ
170 /* C language kind, used when calling cpp_create_reader. */
171 enum c_lang {CLK_GNUC89 = 0, CLK_GNUC99, CLK_GNUC11, CLK_GNUC17, CLK_GNUC2X,
172 CLK_STDC89, CLK_STDC94, CLK_STDC99, CLK_STDC11, CLK_STDC17,
174 CLK_GNUCXX, CLK_CXX98, CLK_GNUCXX11, CLK_CXX11,
175 CLK_GNUCXX14, CLK_CXX14, CLK_GNUCXX17, CLK_CXX17,
176 CLK_GNUCXX20, CLK_CXX20, CLK_GNUCXX23, CLK_CXX23,
179 /* Payload of a NUMBER, STRING, CHAR or COMMENT token. */
180 struct GTY(()) cpp_string {
182 const unsigned char *text;
185 /* Flags for the cpp_token structure. */
186 #define PREV_WHITE (1 << 0) /* If whitespace before this token. */
187 #define DIGRAPH (1 << 1) /* If it was a digraph. */
188 #define STRINGIFY_ARG (1 << 2) /* If macro argument to be stringified. */
189 #define PASTE_LEFT (1 << 3) /* If on LHS of a ## operator. */
190 #define NAMED_OP (1 << 4) /* C++ named operators. */
191 #define PREV_FALLTHROUGH (1 << 5) /* On a token preceeded by FALLTHROUGH
193 #define DECIMAL_INT (1 << 6) /* Decimal integer, set in c-lex.cc. */
194 #define PURE_ZERO (1 << 7) /* Single 0 digit, used by the C++ frontend,
196 #define SP_DIGRAPH (1 << 8) /* # or ## token was a digraph. */
197 #define SP_PREV_WHITE (1 << 9) /* If whitespace before a ##
198 operator, or before this token
199 after a # operator. */
200 #define NO_EXPAND (1 << 10) /* Do not macro-expand this token. */
201 #define PRAGMA_OP (1 << 11) /* _Pragma token. */
202 #define BOL (1 << 12) /* Token at beginning of line. */
204 /* Specify which field, if any, of the cpp_token union is used. */
206 enum cpp_token_fld_kind {
208 CPP_TOKEN_FLD_SOURCE,
210 CPP_TOKEN_FLD_ARG_NO,
211 CPP_TOKEN_FLD_TOKEN_NO,
212 CPP_TOKEN_FLD_PRAGMA,
216 /* A macro argument in the cpp_token union. */
217 struct GTY(()) cpp_macro_arg {
218 /* Argument number. */
220 /* The original spelling of the macro argument token. */
222 GTY ((nested_ptr (union tree_node,
223 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
224 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
228 /* An identifier in the cpp_token union. */
229 struct GTY(()) cpp_identifier {
230 /* The canonical (UTF-8) spelling of the identifier. */
232 GTY ((nested_ptr (union tree_node,
233 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
234 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
236 /* The original spelling of the identifier. */
238 GTY ((nested_ptr (union tree_node,
239 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
240 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
244 /* A preprocessing token. This has been carefully packed and should
245 occupy 16 bytes on 32-bit hosts and 24 bytes on 64-bit hosts. */
246 struct GTY(()) cpp_token {
248 /* Location of first char of token, together with range of full token. */
251 ENUM_BITFIELD(cpp_ttype) type : CHAR_BIT; /* token type */
252 unsigned short flags; /* flags - see above */
257 struct cpp_identifier GTY ((tag ("CPP_TOKEN_FLD_NODE"))) node;
259 /* Inherit padding from this token. */
260 cpp_token * GTY ((tag ("CPP_TOKEN_FLD_SOURCE"))) source;
262 /* A string, or number. */
263 struct cpp_string GTY ((tag ("CPP_TOKEN_FLD_STR"))) str;
265 /* Argument no. (and original spelling) for a CPP_MACRO_ARG. */
266 struct cpp_macro_arg GTY ((tag ("CPP_TOKEN_FLD_ARG_NO"))) macro_arg;
268 /* Original token no. for a CPP_PASTE (from a sequence of
269 consecutive paste tokens in a macro expansion). */
270 unsigned int GTY ((tag ("CPP_TOKEN_FLD_TOKEN_NO"))) token_no;
272 /* Caller-supplied identifier for a CPP_PRAGMA. */
273 unsigned int GTY ((tag ("CPP_TOKEN_FLD_PRAGMA"))) pragma;
274 } GTY ((desc ("cpp_token_val_index (&%1)"))) val;
277 /* Say which field is in use. */
278 extern enum cpp_token_fld_kind cpp_token_val_index (const cpp_token *tok);
280 /* A type wide enough to hold any multibyte source character.
281 cpplib's character constant interpreter requires an unsigned type.
282 Also, a typedef for the signed equivalent.
283 The width of this type is capped at 32 bits; there do exist targets
284 where wchar_t is 64 bits, but only in a non-default mode, and there
285 would be no meaningful interpretation for a wchar_t value greater
286 than 2^32 anyway -- the widest wide-character encoding around is
287 ISO 10646, which stops at 2^31. */
288 #if CHAR_BIT * SIZEOF_INT >= 32
289 # define CPPCHAR_SIGNED_T int
290 #elif CHAR_BIT * SIZEOF_LONG >= 32
291 # define CPPCHAR_SIGNED_T long
293 # error "Cannot find a least-32-bit signed integer type"
295 typedef unsigned CPPCHAR_SIGNED_T cppchar_t;
296 typedef CPPCHAR_SIGNED_T cppchar_signed_t;
298 /* Style of header dependencies to generate. */
299 enum cpp_deps_style { DEPS_NONE = 0, DEPS_USER, DEPS_SYSTEM };
301 /* The possible normalization levels, from most restrictive to least. */
302 enum cpp_normalize_level {
307 /* In NFC, except for subsequences where being in NFC would make
308 the identifier invalid. */
309 normalized_identifier_C,
310 /* Not normalized at all. */
316 CMS_none, /* A regular source file. */
317 CMS_header, /* Is a directly-specified header file (eg PCH or
319 CMS_user, /* Search the user INCLUDE path. */
320 CMS_system, /* Search the system INCLUDE path. */
323 /* The possible bidirectional control characters checking levels. */
324 enum cpp_bidirectional_level {
326 bidirectional_none = 0,
327 /* Only detect unpaired uses of bidirectional control characters. */
328 bidirectional_unpaired = 1,
329 /* Detect any use of bidirectional control characters. */
330 bidirectional_any = 2,
331 /* Also warn about UCNs. */
332 bidirectional_ucn = 4
335 /* This structure is nested inside struct cpp_reader, and
336 carries all the options visible to the command line. */
339 /* The language we're preprocessing. */
342 /* Nonzero means use extra default include directories for C++. */
343 unsigned char cplusplus;
345 /* Nonzero means handle cplusplus style comments. */
346 unsigned char cplusplus_comments;
348 /* Nonzero means define __OBJC__, treat @ as a special token, use
349 the OBJC[PLUS]_INCLUDE_PATH environment variable, and allow
353 /* Nonzero means don't copy comments into the output file. */
354 unsigned char discard_comments;
356 /* Nonzero means don't copy comments into the output file during
358 unsigned char discard_comments_in_macro_exp;
360 /* Nonzero means process the ISO trigraph sequences. */
361 unsigned char trigraphs;
363 /* Nonzero means process the ISO digraph sequences. */
364 unsigned char digraphs;
366 /* Nonzero means to allow hexadecimal floats and LL suffixes. */
367 unsigned char extended_numbers;
369 /* Nonzero means process u/U prefix literals (UTF-16/32). */
370 unsigned char uliterals;
372 /* Nonzero means process u8 prefixed character literals (UTF-8). */
373 unsigned char utf8_char_literals;
375 /* Nonzero means process r/R raw strings. If this is set, uliterals
376 must be set as well. */
377 unsigned char rliterals;
379 /* Nonzero means print names of header files (-H). */
380 unsigned char print_include_names;
382 /* Nonzero means complain about deprecated features. */
383 unsigned char cpp_warn_deprecated;
385 /* Nonzero means warn if slash-star appears in a comment. */
386 unsigned char warn_comments;
388 /* Nonzero means to warn about __DATA__, __TIME__ and __TIMESTAMP__ usage. */
389 unsigned char warn_date_time;
391 /* Nonzero means warn if a user-supplied include directory does not
393 unsigned char warn_missing_include_dirs;
395 /* Nonzero means warn if there are any trigraphs. */
396 unsigned char warn_trigraphs;
398 /* Nonzero means warn about multicharacter charconsts. */
399 unsigned char warn_multichar;
401 /* Nonzero means warn about various incompatibilities with
403 unsigned char cpp_warn_traditional;
405 /* Nonzero means warn about long long numeric constants. */
406 unsigned char cpp_warn_long_long;
408 /* Nonzero means warn about text after an #endif (or #else). */
409 unsigned char warn_endif_labels;
411 /* Nonzero means warn about implicit sign changes owing to integer
413 unsigned char warn_num_sign_change;
415 /* Zero means don't warn about __VA_ARGS__ usage in c89 pedantic mode.
416 Presumably the usage is protected by the appropriate #ifdef. */
417 unsigned char warn_variadic_macros;
419 /* Nonzero means warn about builtin macros that are redefined or
420 explicitly undefined. */
421 unsigned char warn_builtin_macro_redefined;
423 /* Different -Wimplicit-fallthrough= levels. */
424 unsigned char cpp_warn_implicit_fallthrough;
426 /* Nonzero means we should look for header.gcc files that remap file
430 /* Zero means dollar signs are punctuation. */
431 unsigned char dollars_in_ident;
433 /* Nonzero means UCNs are accepted in identifiers. */
434 unsigned char extended_identifiers;
436 /* True if we should warn about dollars in identifiers or numbers
437 for this translation unit. */
438 unsigned char warn_dollars;
440 /* Nonzero means warn if undefined identifiers are evaluated in an #if. */
441 unsigned char warn_undef;
443 /* Nonzero means warn if "defined" is encountered in a place other than
445 unsigned char warn_expansion_to_defined;
447 /* Nonzero means warn of unused macros from the main file. */
448 unsigned char warn_unused_macros;
450 /* Nonzero for the 1999 C Standard, including corrigenda and amendments. */
453 /* Nonzero if we are conforming to a specific C or C++ standard. */
456 /* Nonzero means give all the error messages the ANSI standard requires. */
457 unsigned char cpp_pedantic;
459 /* Nonzero means we're looking at already preprocessed code, so don't
460 bother trying to do macro expansion and whatnot. */
461 unsigned char preprocessed;
463 /* Nonzero means we are going to emit debugging logs during
467 /* Nonzero means we are tracking locations of tokens involved in
468 macro expansion. 1 Means we track the location in degraded mode
469 where we do not track locations of tokens resulting from the
470 expansion of arguments of function-like macro. 2 Means we do
471 track all macro expansions. This last option is the one that
472 consumes the highest amount of memory. */
473 unsigned char track_macro_expansion;
475 /* Nonzero means handle C++ alternate operator names. */
476 unsigned char operator_names;
478 /* Nonzero means warn about use of C++ alternate operator names. */
479 unsigned char warn_cxx_operator_names;
481 /* True for traditional preprocessing. */
482 unsigned char traditional;
484 /* Nonzero for C++ 2011 Standard user-defined literals. */
485 unsigned char user_literals;
487 /* Nonzero means warn when a string or character literal is followed by a
488 ud-suffix which does not beging with an underscore. */
489 unsigned char warn_literal_suffix;
491 /* Nonzero means interpret imaginary, fixed-point, or other gnu extension
492 literal number suffixes as user-defined literal number suffixes. */
493 unsigned char ext_numeric_literals;
495 /* Nonzero means extended identifiers allow the characters specified
497 unsigned char c11_identifiers;
499 /* Nonzero for C++ 2014 Standard binary constants. */
500 unsigned char binary_constants;
502 /* Nonzero for C++ 2014 Standard digit separators. */
503 unsigned char digit_separators;
505 /* Nonzero for C2X decimal floating-point constants. */
506 unsigned char dfp_constants;
508 /* Nonzero for C++20 __VA_OPT__ feature. */
509 unsigned char va_opt;
511 /* Nonzero for the '::' token. */
514 /* Nonzero for the '#elifdef' and '#elifndef' directives. */
515 unsigned char elifdef;
517 /* Nonzero for the '#warning' directive. */
518 unsigned char warning_directive;
520 /* Nonzero means tokenize C++20 module directives. */
521 unsigned char module_directives;
523 /* Nonzero for C++23 size_t literals. */
524 unsigned char size_t_literals;
526 /* Nonzero for C++23 delimited escape sequences. */
527 unsigned char delimited_escape_seqs;
529 /* Nonzero for 'true' and 'false' in #if expressions. */
530 unsigned char true_false;
532 /* Holds the name of the target (execution) character set. */
533 const char *narrow_charset;
535 /* Holds the name of the target wide character set. */
536 const char *wide_charset;
538 /* Holds the name of the input character set. */
539 const char *input_charset;
541 /* The minimum permitted level of normalization before a warning
542 is generated. See enum cpp_normalize_level. */
545 /* True to warn about precompiled header files we couldn't use. */
546 bool warn_invalid_pch;
548 /* True if dependencies should be restored from a precompiled header. */
549 bool restore_pch_deps;
551 /* True if warn about differences between C90 and C99. */
552 signed char cpp_warn_c90_c99_compat;
554 /* True if warn about differences between C11 and C2X. */
555 signed char cpp_warn_c11_c2x_compat;
557 /* True if warn about differences between C++98 and C++11. */
558 bool cpp_warn_cxx11_compat;
560 /* True if warn about differences between C++17 and C++20. */
561 bool cpp_warn_cxx20_compat;
563 /* Nonzero if bidirectional control characters checking is on. See enum
564 cpp_bidirectional_level. */
565 unsigned char cpp_warn_bidirectional;
567 /* True if libcpp should warn about invalid UTF-8 characters in comments.
568 2 if it should be a pedwarn. */
569 unsigned char cpp_warn_invalid_utf8;
571 /* True if libcpp should warn about invalid forms of delimited or named
573 bool cpp_warn_unicode;
575 /* True if -finput-charset= option has been used explicitly. */
576 bool cpp_input_charset_explicit;
578 /* Dependency generation. */
581 /* Style of header dependencies to generate. */
582 enum cpp_deps_style style;
584 /* Assume missing files are generated files. */
587 /* Generate phony targets for each dependency apart from the first
591 /* Generate dependency info for modules. */
594 /* If true, no dependency is generated on the main file. */
595 bool ignore_main_file;
597 /* If true, intend to use the preprocessor output (e.g., for compilation)
598 in addition to the dependency info. */
599 bool need_preprocessor_output;
602 /* Target-specific features set by the front end or client. */
604 /* Precision for target CPP arithmetic, target characters, target
605 ints and target wide characters, respectively. */
606 size_t precision, char_precision, int_precision, wchar_precision;
608 /* True means chars (wide chars, UTF-8 chars) are unsigned. */
609 bool unsigned_char, unsigned_wchar, unsigned_utf8char;
611 /* True if the most significant byte in a word has the lowest
612 address in memory. */
613 bool bytes_big_endian;
615 /* Nonzero means __STDC__ should have the value 0 in system headers. */
616 unsigned char stdc_0_in_system_headers;
618 /* True disables tokenization outside of preprocessing directives. */
619 bool directives_only;
621 /* True enables canonicalization of system header file paths. */
622 bool canonical_system_headers;
624 /* The maximum depth of the nested #include. */
625 unsigned int max_include_depth;
627 cpp_main_search main_search : 8;
630 /* Diagnostic levels. To get a diagnostic without associating a
631 position in the translation unit with it, use cpp_error_with_line
632 with a line number of zero. */
634 enum cpp_diagnostic_level {
635 /* Warning, an error with -Werror. */
637 /* Same as CPP_DL_WARNING, except it is not suppressed in system headers. */
638 CPP_DL_WARNING_SYSHDR,
639 /* Warning, an error with -pedantic-errors or -Werror. */
643 /* An internal consistency check failed. Prints "internal error: ",
644 otherwise the same as CPP_DL_ERROR. */
646 /* An informative note following a warning. */
652 /* Warning reason codes. Use a reason code of CPP_W_NONE for unclassified
653 warnings and diagnostics that are not warnings. */
655 enum cpp_warning_reason {
659 CPP_W_MISSING_INCLUDE_DIRS,
665 CPP_W_NUM_SIGN_CHANGE,
666 CPP_W_VARIADIC_MACROS,
667 CPP_W_BUILTIN_MACRO_REDEFINED,
671 CPP_W_CXX_OPERATOR_NAMES,
674 CPP_W_WARNING_DIRECTIVE,
675 CPP_W_LITERAL_SUFFIX,
676 CPP_W_SIZE_T_LITERALS,
679 CPP_W_C90_C99_COMPAT,
680 CPP_W_C11_C2X_COMPAT,
683 CPP_W_EXPANSION_TO_DEFINED,
689 /* Callback for header lookup for HEADER, which is the name of a
690 source file. It is used as a method of last resort to find headers
691 that are not otherwise found during the normal include processing.
692 The return value is the malloced name of a header to try and open,
693 if any, or NULL otherwise. This callback is called only if the
694 header is otherwise unfound. */
695 typedef const char *(*missing_header_cb)(cpp_reader *, const char *header, cpp_dir **);
697 /* Call backs to cpplib client. */
700 /* Called when a new line of preprocessed output is started. */
701 void (*line_change) (cpp_reader *, const cpp_token *, int);
703 /* Called when switching to/from a new file.
704 The line_map is for the new file. It is NULL if there is no new file.
705 (In C this happens when done with <built-in>+<command line> and also
706 when done with a main file.) This can be used for resource cleanup. */
707 void (*file_change) (cpp_reader *, const line_map_ordinary *);
709 void (*dir_change) (cpp_reader *, const char *);
710 void (*include) (cpp_reader *, location_t, const unsigned char *,
711 const char *, int, const cpp_token **);
712 void (*define) (cpp_reader *, location_t, cpp_hashnode *);
713 void (*undef) (cpp_reader *, location_t, cpp_hashnode *);
714 void (*ident) (cpp_reader *, location_t, const cpp_string *);
715 void (*def_pragma) (cpp_reader *, location_t);
716 int (*valid_pch) (cpp_reader *, const char *, int);
717 void (*read_pch) (cpp_reader *, const char *, int, const char *);
718 missing_header_cb missing_header;
720 /* Context-sensitive macro support. Returns macro (if any) that should
722 cpp_hashnode * (*macro_to_expand) (cpp_reader *, const cpp_token *);
724 /* Called to emit a diagnostic. This callback receives the
725 translated message. */
726 bool (*diagnostic) (cpp_reader *,
727 enum cpp_diagnostic_level,
728 enum cpp_warning_reason,
730 const char *, va_list *)
731 ATTRIBUTE_FPTR_PRINTF(5,0);
733 /* Callbacks for when a macro is expanded, or tested (whether
734 defined or not at the time) in #ifdef, #ifndef or "defined". */
735 void (*used_define) (cpp_reader *, location_t, cpp_hashnode *);
736 void (*used_undef) (cpp_reader *, location_t, cpp_hashnode *);
737 /* Called before #define and #undef or other macro definition
738 changes are processed. */
739 void (*before_define) (cpp_reader *);
740 /* Called whenever a macro is expanded or tested.
741 Second argument is the location of the start of the current expansion. */
742 void (*used) (cpp_reader *, location_t, cpp_hashnode *);
744 /* Callback to identify whether an attribute exists. */
745 int (*has_attribute) (cpp_reader *, bool);
747 /* Callback to determine whether a built-in function is recognized. */
748 int (*has_builtin) (cpp_reader *);
750 /* Callback that can change a user lazy into normal macro. */
751 void (*user_lazy_macro) (cpp_reader *, cpp_macro *, unsigned);
753 /* Callback to handle deferred cpp_macros. */
754 cpp_macro *(*user_deferred_macro) (cpp_reader *, location_t, cpp_hashnode *);
756 /* Callback to parse SOURCE_DATE_EPOCH from environment. */
757 time_t (*get_source_date_epoch) (cpp_reader *);
759 /* Callback for providing suggestions for misspelled directives. */
760 const char *(*get_suggestion) (cpp_reader *, const char *, const char *const *);
762 /* Callback for when a comment is encountered, giving the location
763 of the opening slash, a pointer to the content (which is not
764 necessarily 0-terminated), and the length of the content.
765 The content contains the opening slash-star (or slash-slash),
766 and for C-style comments contains the closing star-slash. For
767 C++-style comments it does not include the terminating newline. */
768 void (*comment) (cpp_reader *, location_t, const unsigned char *,
771 /* Callback for filename remapping in __FILE__ and __BASE_FILE__ macro
773 const char *(*remap_filename) (const char*);
775 /* Maybe translate a #include into something else. Return a
776 cpp_buffer containing the translation if translating. */
777 char *(*translate_include) (cpp_reader *, line_maps *, location_t,
782 #define INO_T_CPP ino_t ino[3]
783 #elif defined (_AIX) && SIZEOF_INO_T == 4
784 #define INO_T_CPP ino64_t ino
786 #define INO_T_CPP ino_t ino
789 #if defined (_AIX) && SIZEOF_DEV_T == 4
790 #define DEV_T_CPP dev64_t dev
792 #define DEV_T_CPP dev_t dev
795 /* Chain of directories to look for include files in. */
798 /* NULL-terminated singly-linked list. */
799 struct cpp_dir *next;
801 /* NAME of the directory, NUL-terminated. */
805 /* One if a system header, two if a system header that has extern
806 "C" guards for C++. */
809 /* Is this a user-supplied directory? */
810 bool user_supplied_p;
812 /* The canonicalized NAME as determined by lrealpath. This field
813 is only used by hosts that lack reliable inode numbers. */
814 char *canonical_name;
816 /* Mapping of file names for this directory for MS-DOS and related
817 platforms. A NULL-terminated array of (from, to) pairs. */
818 const char **name_map;
820 /* Routine to construct pathname, given the search path name and the
821 HEADER we are trying to find, return a constructed pathname to
822 try and open. If this is NULL, the constructed pathname is as
823 constructed by append_file_to_dir. */
824 char *(*construct) (const char *header, cpp_dir *dir);
826 /* The C front end uses these to recognize duplicated
827 directories in the search path. */
832 /* The kind of the cpp_macro. */
833 enum cpp_macro_kind {
834 cmk_macro, /* An ISO macro (token expansion). */
835 cmk_assert, /* An assertion. */
836 cmk_traditional /* A traditional macro (text expansion). */
839 /* Each macro definition is recorded in a cpp_macro structure.
840 Variadic macros cannot occur with traditional cpp. */
841 struct GTY(()) cpp_macro {
844 /* Parameters, if any. If parameter names use extended identifiers,
845 the original spelling of those identifiers, not the canonical
846 UTF-8 spelling, goes here. */
847 cpp_hashnode ** GTY ((tag ("false"),
848 nested_ptr (union tree_node,
849 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
850 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL"),
851 length ("%1.paramc"))) params;
853 /* If this is an assertion, the next one in the chain. */
854 cpp_macro *GTY ((tag ("true"))) next;
855 } GTY ((desc ("%1.kind == cmk_assert"))) parm;
857 /* Definition line number. */
860 /* Number of tokens in body, or bytes for traditional macros. */
861 /* Do we really need 2^32-1 range here? */
864 /* Number of parameters. */
865 unsigned short paramc;
867 /* Non-zero if this is a user-lazy macro, value provided by user. */
870 /* The kind of this macro (ISO, trad or assert) */
873 /* If a function-like macro. */
874 unsigned int fun_like : 1;
876 /* If a variadic macro. */
877 unsigned int variadic : 1;
879 /* If macro defined in system header. */
880 unsigned int syshdr : 1;
882 /* Nonzero if it has been expanded or had its existence tested. */
883 unsigned int used : 1;
885 /* Indicate whether the tokens include extra CPP_PASTE tokens at the
886 end to track invalid redefinitions with consecutive CPP_PASTE
888 unsigned int extra_tokens : 1;
890 /* Imported C++20 macro (from a header unit). */
891 unsigned int imported_p : 1;
893 /* 0 bits spare (32-bit). 32 on 64-bit target. */
897 /* Trailing array of replacement tokens (ISO), or assertion body value. */
898 cpp_token GTY ((tag ("false"), length ("%1.count"))) tokens[1];
900 /* Pointer to replacement text (traditional). See comment at top
901 of cpptrad.c for how traditional function-like macros are
903 const unsigned char *GTY ((tag ("true"))) text;
904 } GTY ((desc ("%1.kind == cmk_traditional"))) exp;
907 /* Poisoned identifiers are flagged NODE_POISONED. NODE_OPERATOR (C++
908 only) indicates an identifier that behaves like an operator such as
909 "xor". NODE_DIAGNOSTIC is for speed in lex_token: it indicates a
910 diagnostic may be required for this node. Currently this only
911 applies to __VA_ARGS__, poisoned identifiers, and -Wc++-compat
912 warnings about NODE_OPERATOR. */
914 /* Hash node flags. */
915 #define NODE_OPERATOR (1 << 0) /* C++ named operator. */
916 #define NODE_POISONED (1 << 1) /* Poisoned identifier. */
917 #define NODE_DIAGNOSTIC (1 << 2) /* Possible diagnostic when lexed. */
918 #define NODE_WARN (1 << 3) /* Warn if redefined or undefined. */
919 #define NODE_DISABLED (1 << 4) /* A disabled macro. */
920 #define NODE_USED (1 << 5) /* Dumped with -dU. */
921 #define NODE_CONDITIONAL (1 << 6) /* Conditional macro */
922 #define NODE_WARN_OPERATOR (1 << 7) /* Warn about C++ named operator. */
923 #define NODE_MODULE (1 << 8) /* C++-20 module-related name. */
925 /* Different flavors of hash node. */
928 NT_VOID = 0, /* Maybe an assert? */
929 NT_MACRO_ARG, /* A macro arg. */
930 NT_USER_MACRO, /* A user macro. */
931 NT_BUILTIN_MACRO, /* A builtin macro. */
932 NT_MACRO_MASK = NT_USER_MACRO /* Mask for either macro kind. */
935 /* Different flavors of builtin macro. _Pragma is an operator, but we
936 handle it with the builtin code for efficiency reasons. */
937 enum cpp_builtin_type
939 BT_SPECLINE = 0, /* `__LINE__' */
940 BT_DATE, /* `__DATE__' */
941 BT_FILE, /* `__FILE__' */
942 BT_FILE_NAME, /* `__FILE_NAME__' */
943 BT_BASE_FILE, /* `__BASE_FILE__' */
944 BT_INCLUDE_LEVEL, /* `__INCLUDE_LEVEL__' */
945 BT_TIME, /* `__TIME__' */
946 BT_STDC, /* `__STDC__' */
947 BT_PRAGMA, /* `_Pragma' operator */
948 BT_TIMESTAMP, /* `__TIMESTAMP__' */
949 BT_COUNTER, /* `__COUNTER__' */
950 BT_HAS_ATTRIBUTE, /* `__has_attribute(x)' */
951 BT_HAS_STD_ATTRIBUTE, /* `__has_c_attribute(x)' */
952 BT_HAS_BUILTIN, /* `__has_builtin(x)' */
953 BT_HAS_INCLUDE, /* `__has_include(x)' */
954 BT_HAS_INCLUDE_NEXT /* `__has_include_next(x)' */
957 #define CPP_HASHNODE(HNODE) ((cpp_hashnode *) (HNODE))
958 #define HT_NODE(NODE) (&(NODE)->ident)
959 #define NODE_LEN(NODE) HT_LEN (HT_NODE (NODE))
960 #define NODE_NAME(NODE) HT_STR (HT_NODE (NODE))
962 /* The common part of an identifier node shared amongst all 3 C front
963 ends. Also used to store CPP identifiers, which are a superset of
964 identifiers in the grammatical sense. */
966 union GTY(()) _cpp_hashnode_value {
967 /* Assert (maybe NULL) */
968 cpp_macro * GTY((tag ("NT_VOID"))) answers;
969 /* Macro (maybe NULL) */
970 cpp_macro * GTY((tag ("NT_USER_MACRO"))) macro;
971 /* Code for a builtin macro. */
972 enum cpp_builtin_type GTY ((tag ("NT_BUILTIN_MACRO"))) builtin;
973 /* Macro argument index. */
974 unsigned short GTY ((tag ("NT_MACRO_ARG"))) arg_index;
977 struct GTY(()) cpp_hashnode {
978 struct ht_identifier ident;
979 unsigned int is_directive : 1;
980 unsigned int directive_index : 7; /* If is_directive,
981 then index into directive table.
982 Otherwise, a NODE_OPERATOR. */
983 unsigned int rid_code : 8; /* Rid code - for front ends. */
984 unsigned int flags : 9; /* CPP flags. */
985 ENUM_BITFIELD(node_type) type : 2; /* CPP node type. */
989 /* The deferred cookie is applicable to NT_USER_MACRO or NT_VOID.
990 The latter for when a macro had a prevailing undef.
991 On a 64-bit system there would be 32-bits of padding to the value
992 field. So placing the deferred index here is not costly. */
993 unsigned deferred; /* Deferred cookie */
995 union _cpp_hashnode_value GTY ((desc ("%1.type"))) value;
998 /* A class for iterating through the source locations within a
999 string token (before escapes are interpreted, and before
1002 class cpp_string_location_reader {
1004 cpp_string_location_reader (location_t src_loc,
1005 line_maps *line_table);
1007 source_range get_next ();
1011 int m_offset_per_column;
1014 /* A class for storing the source ranges of all of the characters within
1015 a string literal, after escapes are interpreted, and after
1018 This is not GTY-marked, as instances are intended to be temporary. */
1020 class cpp_substring_ranges
1023 cpp_substring_ranges ();
1024 ~cpp_substring_ranges ();
1026 int get_num_ranges () const { return m_num_ranges; }
1027 source_range get_range (int idx) const
1029 linemap_assert (idx < m_num_ranges);
1030 return m_ranges[idx];
1033 void add_range (source_range range);
1034 void add_n_ranges (int num, cpp_string_location_reader &loc_reader);
1037 source_range *m_ranges;
1042 /* Call this first to get a handle to pass to other functions.
1044 If you want cpplib to manage its own hashtable, pass in a NULL
1045 pointer. Otherwise you should pass in an initialized hash table
1046 that cpplib will share; this technique is used by the C front
1048 extern cpp_reader *cpp_create_reader (enum c_lang, struct ht *,
1051 /* Reset the cpp_reader's line_map. This is only used after reading a
1053 extern void cpp_set_line_map (cpp_reader *, class line_maps *);
1055 /* Call this to change the selected language standard (e.g. because of
1056 command line options). */
1057 extern void cpp_set_lang (cpp_reader *, enum c_lang);
1059 /* Set the include paths. */
1060 extern void cpp_set_include_chains (cpp_reader *, cpp_dir *, cpp_dir *, int);
1062 /* Call these to get pointers to the options, callback, and deps
1063 structures for a given reader. These pointers are good until you
1064 call cpp_finish on that reader. You can either edit the callbacks
1065 through the pointer returned from cpp_get_callbacks, or set them
1066 with cpp_set_callbacks. */
1067 extern cpp_options *cpp_get_options (cpp_reader *) ATTRIBUTE_PURE;
1068 extern cpp_callbacks *cpp_get_callbacks (cpp_reader *) ATTRIBUTE_PURE;
1069 extern void cpp_set_callbacks (cpp_reader *, cpp_callbacks *);
1070 extern class mkdeps *cpp_get_deps (cpp_reader *) ATTRIBUTE_PURE;
1072 extern const char *cpp_probe_header_unit (cpp_reader *, const char *file,
1073 bool angle_p, location_t);
1075 /* Call these to get name data about the various compile-time
1077 extern const char *cpp_get_narrow_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1078 extern const char *cpp_get_wide_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1080 /* This function reads the file, but does not start preprocessing. It
1081 returns the name of the original file; this is the same as the
1082 input file, except for preprocessed input. This will generate at
1083 least one file change callback, and possibly a line change callback
1084 too. If there was an error opening the file, it returns NULL. */
1085 extern const char *cpp_read_main_file (cpp_reader *, const char *,
1086 bool injecting = false);
1087 extern location_t cpp_main_loc (const cpp_reader *);
1089 /* Adjust for the main file to be an include. */
1090 extern void cpp_retrofit_as_include (cpp_reader *);
1092 /* Set up built-ins with special behavior. Use cpp_init_builtins()
1093 instead unless your know what you are doing. */
1094 extern void cpp_init_special_builtins (cpp_reader *);
1096 /* Set up built-ins like __FILE__. */
1097 extern void cpp_init_builtins (cpp_reader *, int);
1099 /* This is called after options have been parsed, and partially
1101 extern void cpp_post_options (cpp_reader *);
1103 /* Set up translation to the target character set. */
1104 extern void cpp_init_iconv (cpp_reader *);
1106 /* Call this to finish preprocessing. If you requested dependency
1107 generation, pass an open stream to write the information to,
1108 otherwise NULL. It is your responsibility to close the stream. */
1109 extern void cpp_finish (cpp_reader *, FILE *deps_stream);
1111 /* Call this to release the handle at the end of preprocessing. Any
1112 use of the handle after this function returns is invalid. */
1113 extern void cpp_destroy (cpp_reader *);
1115 extern unsigned int cpp_token_len (const cpp_token *);
1116 extern unsigned char *cpp_token_as_text (cpp_reader *, const cpp_token *);
1117 extern unsigned char *cpp_spell_token (cpp_reader *, const cpp_token *,
1118 unsigned char *, bool);
1119 extern void cpp_register_pragma (cpp_reader *, const char *, const char *,
1120 void (*) (cpp_reader *), bool);
1121 extern void cpp_register_deferred_pragma (cpp_reader *, const char *,
1122 const char *, unsigned, bool, bool);
1123 extern int cpp_avoid_paste (cpp_reader *, const cpp_token *,
1125 extern const cpp_token *cpp_get_token (cpp_reader *);
1126 extern const cpp_token *cpp_get_token_with_location (cpp_reader *,
1128 inline bool cpp_user_macro_p (const cpp_hashnode *node)
1130 return node->type == NT_USER_MACRO;
1132 inline bool cpp_builtin_macro_p (const cpp_hashnode *node)
1134 return node->type == NT_BUILTIN_MACRO;
1136 inline bool cpp_macro_p (const cpp_hashnode *node)
1138 return node->type & NT_MACRO_MASK;
1140 inline cpp_macro *cpp_set_deferred_macro (cpp_hashnode *node,
1141 cpp_macro *forced = NULL)
1143 cpp_macro *old = node->value.macro;
1145 node->value.macro = forced;
1146 node->type = NT_USER_MACRO;
1147 node->flags &= ~NODE_USED;
1151 cpp_macro *cpp_get_deferred_macro (cpp_reader *, cpp_hashnode *, location_t);
1153 /* Returns true if NODE is a function-like user macro. */
1154 inline bool cpp_fun_like_macro_p (cpp_hashnode *node)
1156 return cpp_user_macro_p (node) && node->value.macro->fun_like;
1159 extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *);
1160 extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *,
1162 inline location_t cpp_macro_definition_location (cpp_hashnode *node)
1164 const cpp_macro *macro = node->value.macro;
1165 return macro ? macro->line : 0;
1167 /* Return an idempotent time stamp (possibly from SOURCE_DATE_EPOCH). */
1168 enum class CPP_time_kind
1170 FIXED = -1, /* Fixed time via source epoch. */
1171 DYNAMIC = -2, /* Dynamic via time(2). */
1172 UNKNOWN = -3 /* Wibbly wobbly, timey wimey. */
1174 extern CPP_time_kind cpp_get_date (cpp_reader *, time_t *);
1176 extern void _cpp_backup_tokens (cpp_reader *, unsigned int);
1177 extern const cpp_token *cpp_peek_token (cpp_reader *, int);
1179 /* Evaluate a CPP_*CHAR* token. */
1180 extern cppchar_t cpp_interpret_charconst (cpp_reader *, const cpp_token *,
1181 unsigned int *, int *);
1182 /* Evaluate a vector of CPP_*STRING* tokens. */
1183 extern bool cpp_interpret_string (cpp_reader *,
1184 const cpp_string *, size_t,
1185 cpp_string *, enum cpp_ttype);
1186 extern const char *cpp_interpret_string_ranges (cpp_reader *pfile,
1187 const cpp_string *from,
1188 cpp_string_location_reader *,
1190 cpp_substring_ranges *out,
1191 enum cpp_ttype type);
1192 extern bool cpp_interpret_string_notranslate (cpp_reader *,
1193 const cpp_string *, size_t,
1194 cpp_string *, enum cpp_ttype);
1196 /* Convert a host character constant to the execution character set. */
1197 extern cppchar_t cpp_host_to_exec_charset (cpp_reader *, cppchar_t);
1199 /* Used to register macros and assertions, perhaps from the command line.
1200 The text is the same as the command line argument. */
1201 extern void cpp_define (cpp_reader *, const char *);
1202 extern void cpp_define_unused (cpp_reader *, const char *);
1203 extern void cpp_define_formatted (cpp_reader *pfile,
1204 const char *fmt, ...) ATTRIBUTE_PRINTF_2;
1205 extern void cpp_define_formatted_unused (cpp_reader *pfile,
1207 ...) ATTRIBUTE_PRINTF_2;
1208 extern void cpp_assert (cpp_reader *, const char *);
1209 extern void cpp_undef (cpp_reader *, const char *);
1210 extern void cpp_unassert (cpp_reader *, const char *);
1212 /* Mark a node as a lazily defined macro. */
1213 extern void cpp_define_lazily (cpp_reader *, cpp_hashnode *node, unsigned N);
1215 /* Undefine all macros and assertions. */
1216 extern void cpp_undef_all (cpp_reader *);
1218 extern cpp_buffer *cpp_push_buffer (cpp_reader *, const unsigned char *,
1220 extern int cpp_defined (cpp_reader *, const unsigned char *, int);
1222 /* A preprocessing number. Code assumes that any unused high bits of
1223 the double integer are set to zero. */
1225 /* This type has to be equal to unsigned HOST_WIDE_INT, see
1226 gcc/c-family/c-lex.cc. */
1227 typedef uint64_t cpp_num_part;
1228 typedef struct cpp_num cpp_num;
1233 bool unsignedp; /* True if value should be treated as unsigned. */
1234 bool overflow; /* True if the most recent calculation overflowed. */
1237 /* cpplib provides two interfaces for interpretation of preprocessing
1240 cpp_classify_number categorizes numeric constants according to
1241 their field (integer, floating point, or invalid), radix (decimal,
1242 octal, hexadecimal), and type suffixes. */
1244 #define CPP_N_CATEGORY 0x000F
1245 #define CPP_N_INVALID 0x0000
1246 #define CPP_N_INTEGER 0x0001
1247 #define CPP_N_FLOATING 0x0002
1249 #define CPP_N_WIDTH 0x00F0
1250 #define CPP_N_SMALL 0x0010 /* int, float, short _Fract/Accum */
1251 #define CPP_N_MEDIUM 0x0020 /* long, double, long _Fract/_Accum. */
1252 #define CPP_N_LARGE 0x0040 /* long long, long double,
1253 long long _Fract/Accum. */
1255 #define CPP_N_WIDTH_MD 0xF0000 /* machine defined. */
1256 #define CPP_N_MD_W 0x10000
1257 #define CPP_N_MD_Q 0x20000
1259 #define CPP_N_RADIX 0x0F00
1260 #define CPP_N_DECIMAL 0x0100
1261 #define CPP_N_HEX 0x0200
1262 #define CPP_N_OCTAL 0x0400
1263 #define CPP_N_BINARY 0x0800
1265 #define CPP_N_UNSIGNED 0x1000 /* Properties. */
1266 #define CPP_N_IMAGINARY 0x2000
1267 #define CPP_N_DFLOAT 0x4000
1268 #define CPP_N_DEFAULT 0x8000
1270 #define CPP_N_FRACT 0x100000 /* Fract types. */
1271 #define CPP_N_ACCUM 0x200000 /* Accum types. */
1272 #define CPP_N_FLOATN 0x400000 /* _FloatN types. */
1273 #define CPP_N_FLOATNX 0x800000 /* _FloatNx types. */
1275 #define CPP_N_USERDEF 0x1000000 /* C++11 user-defined literal. */
1277 #define CPP_N_SIZE_T 0x2000000 /* C++23 size_t literal. */
1279 #define CPP_N_WIDTH_FLOATN_NX 0xF0000000 /* _FloatN / _FloatNx value
1280 of N, divided by 16. */
1281 #define CPP_FLOATN_SHIFT 24
1282 #define CPP_FLOATN_MAX 0xF0
1284 /* Classify a CPP_NUMBER token. The return value is a combination of
1285 the flags from the above sets. */
1286 extern unsigned cpp_classify_number (cpp_reader *, const cpp_token *,
1287 const char **, location_t);
1289 /* Return the classification flags for a float suffix. */
1290 extern unsigned int cpp_interpret_float_suffix (cpp_reader *, const char *,
1293 /* Return the classification flags for an int suffix. */
1294 extern unsigned int cpp_interpret_int_suffix (cpp_reader *, const char *,
1297 /* Evaluate a token classified as category CPP_N_INTEGER. */
1298 extern cpp_num cpp_interpret_integer (cpp_reader *, const cpp_token *,
1301 /* Sign extend a number, with PRECISION significant bits and all
1302 others assumed clear, to fill out a cpp_num structure. */
1303 cpp_num cpp_num_sign_extend (cpp_num, size_t);
1305 /* Output a diagnostic of some kind. */
1306 extern bool cpp_error (cpp_reader *, enum cpp_diagnostic_level,
1307 const char *msgid, ...)
1309 extern bool cpp_warning (cpp_reader *, enum cpp_warning_reason,
1310 const char *msgid, ...)
1312 extern bool cpp_pedwarning (cpp_reader *, enum cpp_warning_reason,
1313 const char *msgid, ...)
1315 extern bool cpp_warning_syshdr (cpp_reader *, enum cpp_warning_reason reason,
1316 const char *msgid, ...)
1319 /* As their counterparts above, but use RICHLOC. */
1320 extern bool cpp_warning_at (cpp_reader *, enum cpp_warning_reason,
1321 rich_location *richloc, const char *msgid, ...)
1323 extern bool cpp_pedwarning_at (cpp_reader *, enum cpp_warning_reason,
1324 rich_location *richloc, const char *msgid, ...)
1327 /* Output a diagnostic with "MSGID: " preceding the
1328 error string of errno. No location is printed. */
1329 extern bool cpp_errno (cpp_reader *, enum cpp_diagnostic_level,
1331 /* Similarly, but with "FILENAME: " instead of "MSGID: ", where
1332 the filename is not localized. */
1333 extern bool cpp_errno_filename (cpp_reader *, enum cpp_diagnostic_level,
1334 const char *filename, location_t loc);
1336 /* Same as cpp_error, except additionally specifies a position as a
1337 (translation unit) physical line and physical column. If the line is
1338 zero, then no location is printed. */
1339 extern bool cpp_error_with_line (cpp_reader *, enum cpp_diagnostic_level,
1340 location_t, unsigned,
1341 const char *msgid, ...)
1343 extern bool cpp_warning_with_line (cpp_reader *, enum cpp_warning_reason,
1344 location_t, unsigned,
1345 const char *msgid, ...)
1347 extern bool cpp_pedwarning_with_line (cpp_reader *, enum cpp_warning_reason,
1348 location_t, unsigned,
1349 const char *msgid, ...)
1351 extern bool cpp_warning_with_line_syshdr (cpp_reader *, enum cpp_warning_reason,
1352 location_t, unsigned,
1353 const char *msgid, ...)
1356 extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1357 location_t src_loc, const char *msgid, ...)
1360 extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1361 rich_location *richloc, const char *msgid, ...)
1365 extern int cpp_ideq (const cpp_token *, const char *);
1366 extern void cpp_output_line (cpp_reader *, FILE *);
1367 extern unsigned char *cpp_output_line_to_string (cpp_reader *,
1368 const unsigned char *);
1369 extern const unsigned char *cpp_alloc_token_string
1370 (cpp_reader *, const unsigned char *, unsigned);
1371 extern void cpp_output_token (const cpp_token *, FILE *);
1372 extern const char *cpp_type2name (enum cpp_ttype, unsigned char flags);
1373 /* Returns the value of an escape sequence, truncated to the correct
1374 target precision. PSTR points to the input pointer, which is just
1375 after the backslash. LIMIT is how much text we have. WIDE is true
1376 if the escape sequence is part of a wide character constant or
1377 string literal. Handles all relevant diagnostics. */
1378 extern cppchar_t cpp_parse_escape (cpp_reader *, const unsigned char ** pstr,
1379 const unsigned char *limit, int wide);
1381 /* Structure used to hold a comment block at a given location in the
1386 /* Text of the comment including the terminators. */
1389 /* source location for the given comment. */
1393 /* Structure holding all comments for a given cpp_reader. */
1397 /* table of comment entries. */
1398 cpp_comment *entries;
1400 /* number of actual entries entered in the table. */
1403 /* number of entries allocated currently. */
1405 } cpp_comment_table;
1407 /* Returns the table of comments encountered by the preprocessor. This
1408 table is only populated when pfile->state.save_comments is true. */
1409 extern cpp_comment_table *cpp_get_comments (cpp_reader *);
1413 /* Lookup an identifier in the hashtable. Puts the identifier in the
1414 table if it is not already there. */
1415 extern cpp_hashnode *cpp_lookup (cpp_reader *, const unsigned char *,
1418 typedef int (*cpp_cb) (cpp_reader *, cpp_hashnode *, void *);
1419 extern void cpp_forall_identifiers (cpp_reader *, cpp_cb, void *);
1422 extern void cpp_scan_nooutput (cpp_reader *);
1423 extern int cpp_sys_macro_p (cpp_reader *);
1424 extern unsigned char *cpp_quote_string (unsigned char *, const unsigned char *,
1426 extern bool cpp_compare_macros (const cpp_macro *macro1,
1427 const cpp_macro *macro2);
1430 extern bool cpp_included (cpp_reader *, const char *);
1431 extern bool cpp_included_before (cpp_reader *, const char *, location_t);
1432 extern void cpp_make_system_header (cpp_reader *, int, int);
1433 extern bool cpp_push_include (cpp_reader *, const char *);
1434 extern bool cpp_push_default_include (cpp_reader *, const char *);
1435 extern void cpp_change_file (cpp_reader *, enum lc_reason, const char *);
1436 extern const char *cpp_get_path (struct _cpp_file *);
1437 extern cpp_dir *cpp_get_dir (struct _cpp_file *);
1438 extern cpp_buffer *cpp_get_buffer (cpp_reader *);
1439 extern struct _cpp_file *cpp_get_file (cpp_buffer *);
1440 extern cpp_buffer *cpp_get_prev (cpp_buffer *);
1441 extern void cpp_clear_file_cache (cpp_reader *);
1443 /* cpp_get_converted_source returns the contents of the given file, as it exists
1444 after cpplib has read it and converted it from the input charset to the
1445 source charset. Return struct will be zero-filled if the data could not be
1446 read for any reason. The data starts at the DATA pointer, but the TO_FREE
1447 pointer is what should be passed to free(), as there may be an offset. */
1448 struct cpp_converted_source
1454 cpp_converted_source cpp_get_converted_source (const char *fname,
1455 const char *input_charset);
1458 struct save_macro_data;
1459 extern int cpp_save_state (cpp_reader *, FILE *);
1460 extern int cpp_write_pch_deps (cpp_reader *, FILE *);
1461 extern int cpp_write_pch_state (cpp_reader *, FILE *);
1462 extern int cpp_valid_state (cpp_reader *, const char *, int);
1463 extern void cpp_prepare_state (cpp_reader *, struct save_macro_data **);
1464 extern int cpp_read_state (cpp_reader *, const char *, FILE *,
1465 struct save_macro_data *);
1468 extern void cpp_force_token_locations (cpp_reader *, location_t);
1469 extern void cpp_stop_forcing_token_locations (cpp_reader *);
1477 extern void cpp_directive_only_process (cpp_reader *pfile,
1479 void (*cb) (cpp_reader *,
1484 extern enum cpp_ttype cpp_userdef_string_remove_type
1485 (enum cpp_ttype type);
1486 extern enum cpp_ttype cpp_userdef_string_add_type
1487 (enum cpp_ttype type);
1488 extern enum cpp_ttype cpp_userdef_char_remove_type
1489 (enum cpp_ttype type);
1490 extern enum cpp_ttype cpp_userdef_char_add_type
1491 (enum cpp_ttype type);
1492 extern bool cpp_userdef_string_p
1493 (enum cpp_ttype type);
1494 extern bool cpp_userdef_char_p
1495 (enum cpp_ttype type);
1496 extern const char * cpp_get_userdef_suffix
1497 (const cpp_token *);
1501 /* The result of attempting to decode a run of UTF-8 bytes. */
1503 struct cpp_decoded_char
1505 const char *m_start_byte;
1506 const char *m_next_byte;
1512 /* Information for mapping between code points and display columns.
1514 This is a tabstop value, along with a callback for getting the
1515 widths of characters. Normally this callback is cpp_wcwidth, but we
1516 support other schemes for escaping non-ASCII unicode as a series of
1517 ASCII chars when printing the user's source code in diagnostic-show-locus.cc
1519 For example, consider:
1520 - the Unicode character U+03C0 "GREEK SMALL LETTER PI" (UTF-8: 0xCF 0x80)
1521 - the Unicode character U+1F642 "SLIGHTLY SMILING FACE"
1522 (UTF-8: 0xF0 0x9F 0x99 0x82)
1523 - the byte 0xBF (a stray trailing byte of a UTF-8 character)
1524 Normally U+03C0 would occupy one display column, U+1F642
1525 would occupy two display columns, and the stray byte would be
1526 printed verbatim as one display column.
1528 However when escaping them as unicode code points as "<U+03C0>"
1529 and "<U+1F642>" they occupy 8 and 9 display columns respectively,
1530 and when escaping them as bytes as "<CF><80>" and "<F0><9F><99><82>"
1531 they occupy 8 and 16 display columns respectively. In both cases
1532 the stray byte is escaped to <BF> as 4 display columns. */
1534 struct cpp_char_column_policy
1536 cpp_char_column_policy (int tabstop,
1537 int (*width_cb) (cppchar_t c))
1538 : m_tabstop (tabstop),
1539 m_undecoded_byte_width (1),
1540 m_width_cb (width_cb)
1544 /* Width in display columns of a stray byte that isn't decodable
1546 int m_undecoded_byte_width;
1547 int (*m_width_cb) (cppchar_t c);
1550 /* A class to manage the state while converting a UTF-8 sequence to cppchar_t
1551 and computing the display width one character at a time. */
1552 class cpp_display_width_computation {
1554 cpp_display_width_computation (const char *data, int data_length,
1555 const cpp_char_column_policy &policy);
1556 const char *next_byte () const { return m_next; }
1557 int bytes_processed () const { return m_next - m_begin; }
1558 int bytes_left () const { return m_bytes_left; }
1559 bool done () const { return !bytes_left (); }
1560 int display_cols_processed () const { return m_display_cols; }
1562 int process_next_codepoint (cpp_decoded_char *out);
1563 int advance_display_cols (int n);
1566 const char *const m_begin;
1568 size_t m_bytes_left;
1569 const cpp_char_column_policy &m_policy;
1573 /* Convenience functions that are simple use cases for class
1574 cpp_display_width_computation. Tab characters will be expanded to spaces
1575 as determined by POLICY.m_tabstop, and non-printable-ASCII characters
1576 will be escaped as per POLICY. */
1578 int cpp_byte_column_to_display_column (const char *data, int data_length,
1580 const cpp_char_column_policy &policy);
1581 inline int cpp_display_width (const char *data, int data_length,
1582 const cpp_char_column_policy &policy)
1584 return cpp_byte_column_to_display_column (data, data_length, data_length,
1587 int cpp_display_column_to_byte_column (const char *data, int data_length,
1589 const cpp_char_column_policy &policy);
1590 int cpp_wcwidth (cppchar_t c);
1592 bool cpp_input_conversion_is_trivial (const char *input_charset);
1593 int cpp_check_utf8_bom (const char *data, size_t data_length);
1595 #endif /* ! LIBCPP_CPPLIB_H */
projects / platform / upstream / gcc.git / blob