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 /* Holds the name of the target (execution) character set. */
530 const char *narrow_charset;
532 /* Holds the name of the target wide character set. */
533 const char *wide_charset;
535 /* Holds the name of the input character set. */
536 const char *input_charset;
538 /* The minimum permitted level of normalization before a warning
539 is generated. See enum cpp_normalize_level. */
542 /* True to warn about precompiled header files we couldn't use. */
543 bool warn_invalid_pch;
545 /* True if dependencies should be restored from a precompiled header. */
546 bool restore_pch_deps;
548 /* True if warn about differences between C90 and C99. */
549 signed char cpp_warn_c90_c99_compat;
551 /* True if warn about differences between C11 and C2X. */
552 signed char cpp_warn_c11_c2x_compat;
554 /* True if warn about differences between C++98 and C++11. */
555 bool cpp_warn_cxx11_compat;
557 /* True if warn about differences between C++17 and C++20. */
558 bool cpp_warn_cxx20_compat;
560 /* Nonzero if bidirectional control characters checking is on. See enum
561 cpp_bidirectional_level. */
562 unsigned char cpp_warn_bidirectional;
564 /* True if libcpp should warn about invalid UTF-8 characters in comments.
565 2 if it should be a pedwarn. */
566 unsigned char cpp_warn_invalid_utf8;
568 /* True if libcpp should warn about invalid forms of delimited or named
570 bool cpp_warn_unicode;
572 /* True if -finput-charset= option has been used explicitly. */
573 bool cpp_input_charset_explicit;
575 /* Dependency generation. */
578 /* Style of header dependencies to generate. */
579 enum cpp_deps_style style;
581 /* Assume missing files are generated files. */
584 /* Generate phony targets for each dependency apart from the first
588 /* Generate dependency info for modules. */
591 /* If true, no dependency is generated on the main file. */
592 bool ignore_main_file;
594 /* If true, intend to use the preprocessor output (e.g., for compilation)
595 in addition to the dependency info. */
596 bool need_preprocessor_output;
599 /* Target-specific features set by the front end or client. */
601 /* Precision for target CPP arithmetic, target characters, target
602 ints and target wide characters, respectively. */
603 size_t precision, char_precision, int_precision, wchar_precision;
605 /* True means chars (wide chars, UTF-8 chars) are unsigned. */
606 bool unsigned_char, unsigned_wchar, unsigned_utf8char;
608 /* True if the most significant byte in a word has the lowest
609 address in memory. */
610 bool bytes_big_endian;
612 /* Nonzero means __STDC__ should have the value 0 in system headers. */
613 unsigned char stdc_0_in_system_headers;
615 /* True disables tokenization outside of preprocessing directives. */
616 bool directives_only;
618 /* True enables canonicalization of system header file paths. */
619 bool canonical_system_headers;
621 /* The maximum depth of the nested #include. */
622 unsigned int max_include_depth;
624 cpp_main_search main_search : 8;
627 /* Diagnostic levels. To get a diagnostic without associating a
628 position in the translation unit with it, use cpp_error_with_line
629 with a line number of zero. */
631 enum cpp_diagnostic_level {
632 /* Warning, an error with -Werror. */
634 /* Same as CPP_DL_WARNING, except it is not suppressed in system headers. */
635 CPP_DL_WARNING_SYSHDR,
636 /* Warning, an error with -pedantic-errors or -Werror. */
640 /* An internal consistency check failed. Prints "internal error: ",
641 otherwise the same as CPP_DL_ERROR. */
643 /* An informative note following a warning. */
649 /* Warning reason codes. Use a reason code of CPP_W_NONE for unclassified
650 warnings and diagnostics that are not warnings. */
652 enum cpp_warning_reason {
656 CPP_W_MISSING_INCLUDE_DIRS,
662 CPP_W_NUM_SIGN_CHANGE,
663 CPP_W_VARIADIC_MACROS,
664 CPP_W_BUILTIN_MACRO_REDEFINED,
668 CPP_W_CXX_OPERATOR_NAMES,
671 CPP_W_WARNING_DIRECTIVE,
672 CPP_W_LITERAL_SUFFIX,
673 CPP_W_SIZE_T_LITERALS,
676 CPP_W_C90_C99_COMPAT,
677 CPP_W_C11_C2X_COMPAT,
680 CPP_W_EXPANSION_TO_DEFINED,
686 /* Callback for header lookup for HEADER, which is the name of a
687 source file. It is used as a method of last resort to find headers
688 that are not otherwise found during the normal include processing.
689 The return value is the malloced name of a header to try and open,
690 if any, or NULL otherwise. This callback is called only if the
691 header is otherwise unfound. */
692 typedef const char *(*missing_header_cb)(cpp_reader *, const char *header, cpp_dir **);
694 /* Call backs to cpplib client. */
697 /* Called when a new line of preprocessed output is started. */
698 void (*line_change) (cpp_reader *, const cpp_token *, int);
700 /* Called when switching to/from a new file.
701 The line_map is for the new file. It is NULL if there is no new file.
702 (In C this happens when done with <built-in>+<command line> and also
703 when done with a main file.) This can be used for resource cleanup. */
704 void (*file_change) (cpp_reader *, const line_map_ordinary *);
706 void (*dir_change) (cpp_reader *, const char *);
707 void (*include) (cpp_reader *, location_t, const unsigned char *,
708 const char *, int, const cpp_token **);
709 void (*define) (cpp_reader *, location_t, cpp_hashnode *);
710 void (*undef) (cpp_reader *, location_t, cpp_hashnode *);
711 void (*ident) (cpp_reader *, location_t, const cpp_string *);
712 void (*def_pragma) (cpp_reader *, location_t);
713 int (*valid_pch) (cpp_reader *, const char *, int);
714 void (*read_pch) (cpp_reader *, const char *, int, const char *);
715 missing_header_cb missing_header;
717 /* Context-sensitive macro support. Returns macro (if any) that should
719 cpp_hashnode * (*macro_to_expand) (cpp_reader *, const cpp_token *);
721 /* Called to emit a diagnostic. This callback receives the
722 translated message. */
723 bool (*diagnostic) (cpp_reader *,
724 enum cpp_diagnostic_level,
725 enum cpp_warning_reason,
727 const char *, va_list *)
728 ATTRIBUTE_FPTR_PRINTF(5,0);
730 /* Callbacks for when a macro is expanded, or tested (whether
731 defined or not at the time) in #ifdef, #ifndef or "defined". */
732 void (*used_define) (cpp_reader *, location_t, cpp_hashnode *);
733 void (*used_undef) (cpp_reader *, location_t, cpp_hashnode *);
734 /* Called before #define and #undef or other macro definition
735 changes are processed. */
736 void (*before_define) (cpp_reader *);
737 /* Called whenever a macro is expanded or tested.
738 Second argument is the location of the start of the current expansion. */
739 void (*used) (cpp_reader *, location_t, cpp_hashnode *);
741 /* Callback to identify whether an attribute exists. */
742 int (*has_attribute) (cpp_reader *, bool);
744 /* Callback to determine whether a built-in function is recognized. */
745 int (*has_builtin) (cpp_reader *);
747 /* Callback that can change a user lazy into normal macro. */
748 void (*user_lazy_macro) (cpp_reader *, cpp_macro *, unsigned);
750 /* Callback to handle deferred cpp_macros. */
751 cpp_macro *(*user_deferred_macro) (cpp_reader *, location_t, cpp_hashnode *);
753 /* Callback to parse SOURCE_DATE_EPOCH from environment. */
754 time_t (*get_source_date_epoch) (cpp_reader *);
756 /* Callback for providing suggestions for misspelled directives. */
757 const char *(*get_suggestion) (cpp_reader *, const char *, const char *const *);
759 /* Callback for when a comment is encountered, giving the location
760 of the opening slash, a pointer to the content (which is not
761 necessarily 0-terminated), and the length of the content.
762 The content contains the opening slash-star (or slash-slash),
763 and for C-style comments contains the closing star-slash. For
764 C++-style comments it does not include the terminating newline. */
765 void (*comment) (cpp_reader *, location_t, const unsigned char *,
768 /* Callback for filename remapping in __FILE__ and __BASE_FILE__ macro
770 const char *(*remap_filename) (const char*);
772 /* Maybe translate a #include into something else. Return a
773 cpp_buffer containing the translation if translating. */
774 char *(*translate_include) (cpp_reader *, line_maps *, location_t,
779 #define INO_T_CPP ino_t ino[3]
780 #elif defined (_AIX) && SIZEOF_INO_T == 4
781 #define INO_T_CPP ino64_t ino
783 #define INO_T_CPP ino_t ino
786 #if defined (_AIX) && SIZEOF_DEV_T == 4
787 #define DEV_T_CPP dev64_t dev
789 #define DEV_T_CPP dev_t dev
792 /* Chain of directories to look for include files in. */
795 /* NULL-terminated singly-linked list. */
796 struct cpp_dir *next;
798 /* NAME of the directory, NUL-terminated. */
802 /* One if a system header, two if a system header that has extern
803 "C" guards for C++. */
806 /* Is this a user-supplied directory? */
807 bool user_supplied_p;
809 /* The canonicalized NAME as determined by lrealpath. This field
810 is only used by hosts that lack reliable inode numbers. */
811 char *canonical_name;
813 /* Mapping of file names for this directory for MS-DOS and related
814 platforms. A NULL-terminated array of (from, to) pairs. */
815 const char **name_map;
817 /* Routine to construct pathname, given the search path name and the
818 HEADER we are trying to find, return a constructed pathname to
819 try and open. If this is NULL, the constructed pathname is as
820 constructed by append_file_to_dir. */
821 char *(*construct) (const char *header, cpp_dir *dir);
823 /* The C front end uses these to recognize duplicated
824 directories in the search path. */
829 /* The kind of the cpp_macro. */
830 enum cpp_macro_kind {
831 cmk_macro, /* An ISO macro (token expansion). */
832 cmk_assert, /* An assertion. */
833 cmk_traditional /* A traditional macro (text expansion). */
836 /* Each macro definition is recorded in a cpp_macro structure.
837 Variadic macros cannot occur with traditional cpp. */
838 struct GTY(()) cpp_macro {
841 /* Parameters, if any. If parameter names use extended identifiers,
842 the original spelling of those identifiers, not the canonical
843 UTF-8 spelling, goes here. */
844 cpp_hashnode ** GTY ((tag ("false"),
845 nested_ptr (union tree_node,
846 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
847 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL"),
848 length ("%1.paramc"))) params;
850 /* If this is an assertion, the next one in the chain. */
851 cpp_macro *GTY ((tag ("true"))) next;
852 } GTY ((desc ("%1.kind == cmk_assert"))) parm;
854 /* Definition line number. */
857 /* Number of tokens in body, or bytes for traditional macros. */
858 /* Do we really need 2^32-1 range here? */
861 /* Number of parameters. */
862 unsigned short paramc;
864 /* Non-zero if this is a user-lazy macro, value provided by user. */
867 /* The kind of this macro (ISO, trad or assert) */
870 /* If a function-like macro. */
871 unsigned int fun_like : 1;
873 /* If a variadic macro. */
874 unsigned int variadic : 1;
876 /* If macro defined in system header. */
877 unsigned int syshdr : 1;
879 /* Nonzero if it has been expanded or had its existence tested. */
880 unsigned int used : 1;
882 /* Indicate whether the tokens include extra CPP_PASTE tokens at the
883 end to track invalid redefinitions with consecutive CPP_PASTE
885 unsigned int extra_tokens : 1;
887 /* Imported C++20 macro (from a header unit). */
888 unsigned int imported_p : 1;
890 /* 0 bits spare (32-bit). 32 on 64-bit target. */
894 /* Trailing array of replacement tokens (ISO), or assertion body value. */
895 cpp_token GTY ((tag ("false"), length ("%1.count"))) tokens[1];
897 /* Pointer to replacement text (traditional). See comment at top
898 of cpptrad.c for how traditional function-like macros are
900 const unsigned char *GTY ((tag ("true"))) text;
901 } GTY ((desc ("%1.kind == cmk_traditional"))) exp;
904 /* Poisoned identifiers are flagged NODE_POISONED. NODE_OPERATOR (C++
905 only) indicates an identifier that behaves like an operator such as
906 "xor". NODE_DIAGNOSTIC is for speed in lex_token: it indicates a
907 diagnostic may be required for this node. Currently this only
908 applies to __VA_ARGS__, poisoned identifiers, and -Wc++-compat
909 warnings about NODE_OPERATOR. */
911 /* Hash node flags. */
912 #define NODE_OPERATOR (1 << 0) /* C++ named operator. */
913 #define NODE_POISONED (1 << 1) /* Poisoned identifier. */
914 #define NODE_DIAGNOSTIC (1 << 2) /* Possible diagnostic when lexed. */
915 #define NODE_WARN (1 << 3) /* Warn if redefined or undefined. */
916 #define NODE_DISABLED (1 << 4) /* A disabled macro. */
917 #define NODE_USED (1 << 5) /* Dumped with -dU. */
918 #define NODE_CONDITIONAL (1 << 6) /* Conditional macro */
919 #define NODE_WARN_OPERATOR (1 << 7) /* Warn about C++ named operator. */
920 #define NODE_MODULE (1 << 8) /* C++-20 module-related name. */
922 /* Different flavors of hash node. */
925 NT_VOID = 0, /* Maybe an assert? */
926 NT_MACRO_ARG, /* A macro arg. */
927 NT_USER_MACRO, /* A user macro. */
928 NT_BUILTIN_MACRO, /* A builtin macro. */
929 NT_MACRO_MASK = NT_USER_MACRO /* Mask for either macro kind. */
932 /* Different flavors of builtin macro. _Pragma is an operator, but we
933 handle it with the builtin code for efficiency reasons. */
934 enum cpp_builtin_type
936 BT_SPECLINE = 0, /* `__LINE__' */
937 BT_DATE, /* `__DATE__' */
938 BT_FILE, /* `__FILE__' */
939 BT_FILE_NAME, /* `__FILE_NAME__' */
940 BT_BASE_FILE, /* `__BASE_FILE__' */
941 BT_INCLUDE_LEVEL, /* `__INCLUDE_LEVEL__' */
942 BT_TIME, /* `__TIME__' */
943 BT_STDC, /* `__STDC__' */
944 BT_PRAGMA, /* `_Pragma' operator */
945 BT_TIMESTAMP, /* `__TIMESTAMP__' */
946 BT_COUNTER, /* `__COUNTER__' */
947 BT_HAS_ATTRIBUTE, /* `__has_attribute(x)' */
948 BT_HAS_STD_ATTRIBUTE, /* `__has_c_attribute(x)' */
949 BT_HAS_BUILTIN, /* `__has_builtin(x)' */
950 BT_HAS_INCLUDE, /* `__has_include(x)' */
951 BT_HAS_INCLUDE_NEXT /* `__has_include_next(x)' */
954 #define CPP_HASHNODE(HNODE) ((cpp_hashnode *) (HNODE))
955 #define HT_NODE(NODE) (&(NODE)->ident)
956 #define NODE_LEN(NODE) HT_LEN (HT_NODE (NODE))
957 #define NODE_NAME(NODE) HT_STR (HT_NODE (NODE))
959 /* The common part of an identifier node shared amongst all 3 C front
960 ends. Also used to store CPP identifiers, which are a superset of
961 identifiers in the grammatical sense. */
963 union GTY(()) _cpp_hashnode_value {
964 /* Assert (maybe NULL) */
965 cpp_macro * GTY((tag ("NT_VOID"))) answers;
966 /* Macro (maybe NULL) */
967 cpp_macro * GTY((tag ("NT_USER_MACRO"))) macro;
968 /* Code for a builtin macro. */
969 enum cpp_builtin_type GTY ((tag ("NT_BUILTIN_MACRO"))) builtin;
970 /* Macro argument index. */
971 unsigned short GTY ((tag ("NT_MACRO_ARG"))) arg_index;
974 struct GTY(()) cpp_hashnode {
975 struct ht_identifier ident;
976 unsigned int is_directive : 1;
977 unsigned int directive_index : 7; /* If is_directive,
978 then index into directive table.
979 Otherwise, a NODE_OPERATOR. */
980 unsigned int rid_code : 8; /* Rid code - for front ends. */
981 unsigned int flags : 9; /* CPP flags. */
982 ENUM_BITFIELD(node_type) type : 2; /* CPP node type. */
986 /* The deferred cookie is applicable to NT_USER_MACRO or NT_VOID.
987 The latter for when a macro had a prevailing undef.
988 On a 64-bit system there would be 32-bits of padding to the value
989 field. So placing the deferred index here is not costly. */
990 unsigned deferred; /* Deferred cookie */
992 union _cpp_hashnode_value GTY ((desc ("%1.type"))) value;
995 /* A class for iterating through the source locations within a
996 string token (before escapes are interpreted, and before
999 class cpp_string_location_reader {
1001 cpp_string_location_reader (location_t src_loc,
1002 line_maps *line_table);
1004 source_range get_next ();
1008 int m_offset_per_column;
1011 /* A class for storing the source ranges of all of the characters within
1012 a string literal, after escapes are interpreted, and after
1015 This is not GTY-marked, as instances are intended to be temporary. */
1017 class cpp_substring_ranges
1020 cpp_substring_ranges ();
1021 ~cpp_substring_ranges ();
1023 int get_num_ranges () const { return m_num_ranges; }
1024 source_range get_range (int idx) const
1026 linemap_assert (idx < m_num_ranges);
1027 return m_ranges[idx];
1030 void add_range (source_range range);
1031 void add_n_ranges (int num, cpp_string_location_reader &loc_reader);
1034 source_range *m_ranges;
1039 /* Call this first to get a handle to pass to other functions.
1041 If you want cpplib to manage its own hashtable, pass in a NULL
1042 pointer. Otherwise you should pass in an initialized hash table
1043 that cpplib will share; this technique is used by the C front
1045 extern cpp_reader *cpp_create_reader (enum c_lang, struct ht *,
1048 /* Reset the cpp_reader's line_map. This is only used after reading a
1050 extern void cpp_set_line_map (cpp_reader *, class line_maps *);
1052 /* Call this to change the selected language standard (e.g. because of
1053 command line options). */
1054 extern void cpp_set_lang (cpp_reader *, enum c_lang);
1056 /* Set the include paths. */
1057 extern void cpp_set_include_chains (cpp_reader *, cpp_dir *, cpp_dir *, int);
1059 /* Call these to get pointers to the options, callback, and deps
1060 structures for a given reader. These pointers are good until you
1061 call cpp_finish on that reader. You can either edit the callbacks
1062 through the pointer returned from cpp_get_callbacks, or set them
1063 with cpp_set_callbacks. */
1064 extern cpp_options *cpp_get_options (cpp_reader *) ATTRIBUTE_PURE;
1065 extern cpp_callbacks *cpp_get_callbacks (cpp_reader *) ATTRIBUTE_PURE;
1066 extern void cpp_set_callbacks (cpp_reader *, cpp_callbacks *);
1067 extern class mkdeps *cpp_get_deps (cpp_reader *) ATTRIBUTE_PURE;
1069 extern const char *cpp_probe_header_unit (cpp_reader *, const char *file,
1070 bool angle_p, location_t);
1072 /* Call these to get name data about the various compile-time
1074 extern const char *cpp_get_narrow_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1075 extern const char *cpp_get_wide_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1077 /* This function reads the file, but does not start preprocessing. It
1078 returns the name of the original file; this is the same as the
1079 input file, except for preprocessed input. This will generate at
1080 least one file change callback, and possibly a line change callback
1081 too. If there was an error opening the file, it returns NULL. */
1082 extern const char *cpp_read_main_file (cpp_reader *, const char *,
1083 bool injecting = false);
1084 extern location_t cpp_main_loc (const cpp_reader *);
1086 /* Adjust for the main file to be an include. */
1087 extern void cpp_retrofit_as_include (cpp_reader *);
1089 /* Set up built-ins with special behavior. Use cpp_init_builtins()
1090 instead unless your know what you are doing. */
1091 extern void cpp_init_special_builtins (cpp_reader *);
1093 /* Set up built-ins like __FILE__. */
1094 extern void cpp_init_builtins (cpp_reader *, int);
1096 /* This is called after options have been parsed, and partially
1098 extern void cpp_post_options (cpp_reader *);
1100 /* Set up translation to the target character set. */
1101 extern void cpp_init_iconv (cpp_reader *);
1103 /* Call this to finish preprocessing. If you requested dependency
1104 generation, pass an open stream to write the information to,
1105 otherwise NULL. It is your responsibility to close the stream. */
1106 extern void cpp_finish (cpp_reader *, FILE *deps_stream);
1108 /* Call this to release the handle at the end of preprocessing. Any
1109 use of the handle after this function returns is invalid. */
1110 extern void cpp_destroy (cpp_reader *);
1112 extern unsigned int cpp_token_len (const cpp_token *);
1113 extern unsigned char *cpp_token_as_text (cpp_reader *, const cpp_token *);
1114 extern unsigned char *cpp_spell_token (cpp_reader *, const cpp_token *,
1115 unsigned char *, bool);
1116 extern void cpp_register_pragma (cpp_reader *, const char *, const char *,
1117 void (*) (cpp_reader *), bool);
1118 extern void cpp_register_deferred_pragma (cpp_reader *, const char *,
1119 const char *, unsigned, bool, bool);
1120 extern int cpp_avoid_paste (cpp_reader *, const cpp_token *,
1122 extern const cpp_token *cpp_get_token (cpp_reader *);
1123 extern const cpp_token *cpp_get_token_with_location (cpp_reader *,
1125 inline bool cpp_user_macro_p (const cpp_hashnode *node)
1127 return node->type == NT_USER_MACRO;
1129 inline bool cpp_builtin_macro_p (const cpp_hashnode *node)
1131 return node->type == NT_BUILTIN_MACRO;
1133 inline bool cpp_macro_p (const cpp_hashnode *node)
1135 return node->type & NT_MACRO_MASK;
1137 inline cpp_macro *cpp_set_deferred_macro (cpp_hashnode *node,
1138 cpp_macro *forced = NULL)
1140 cpp_macro *old = node->value.macro;
1142 node->value.macro = forced;
1143 node->type = NT_USER_MACRO;
1144 node->flags &= ~NODE_USED;
1148 cpp_macro *cpp_get_deferred_macro (cpp_reader *, cpp_hashnode *, location_t);
1150 /* Returns true if NODE is a function-like user macro. */
1151 inline bool cpp_fun_like_macro_p (cpp_hashnode *node)
1153 return cpp_user_macro_p (node) && node->value.macro->fun_like;
1156 extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *);
1157 extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *,
1159 inline location_t cpp_macro_definition_location (cpp_hashnode *node)
1161 const cpp_macro *macro = node->value.macro;
1162 return macro ? macro->line : 0;
1164 /* Return an idempotent time stamp (possibly from SOURCE_DATE_EPOCH). */
1165 enum class CPP_time_kind
1167 FIXED = -1, /* Fixed time via source epoch. */
1168 DYNAMIC = -2, /* Dynamic via time(2). */
1169 UNKNOWN = -3 /* Wibbly wobbly, timey wimey. */
1171 extern CPP_time_kind cpp_get_date (cpp_reader *, time_t *);
1173 extern void _cpp_backup_tokens (cpp_reader *, unsigned int);
1174 extern const cpp_token *cpp_peek_token (cpp_reader *, int);
1176 /* Evaluate a CPP_*CHAR* token. */
1177 extern cppchar_t cpp_interpret_charconst (cpp_reader *, const cpp_token *,
1178 unsigned int *, int *);
1179 /* Evaluate a vector of CPP_*STRING* tokens. */
1180 extern bool cpp_interpret_string (cpp_reader *,
1181 const cpp_string *, size_t,
1182 cpp_string *, enum cpp_ttype);
1183 extern const char *cpp_interpret_string_ranges (cpp_reader *pfile,
1184 const cpp_string *from,
1185 cpp_string_location_reader *,
1187 cpp_substring_ranges *out,
1188 enum cpp_ttype type);
1189 extern bool cpp_interpret_string_notranslate (cpp_reader *,
1190 const cpp_string *, size_t,
1191 cpp_string *, enum cpp_ttype);
1193 /* Convert a host character constant to the execution character set. */
1194 extern cppchar_t cpp_host_to_exec_charset (cpp_reader *, cppchar_t);
1196 /* Used to register macros and assertions, perhaps from the command line.
1197 The text is the same as the command line argument. */
1198 extern void cpp_define (cpp_reader *, const char *);
1199 extern void cpp_define_unused (cpp_reader *, const char *);
1200 extern void cpp_define_formatted (cpp_reader *pfile,
1201 const char *fmt, ...) ATTRIBUTE_PRINTF_2;
1202 extern void cpp_define_formatted_unused (cpp_reader *pfile,
1204 ...) ATTRIBUTE_PRINTF_2;
1205 extern void cpp_assert (cpp_reader *, const char *);
1206 extern void cpp_undef (cpp_reader *, const char *);
1207 extern void cpp_unassert (cpp_reader *, const char *);
1209 /* Mark a node as a lazily defined macro. */
1210 extern void cpp_define_lazily (cpp_reader *, cpp_hashnode *node, unsigned N);
1212 /* Undefine all macros and assertions. */
1213 extern void cpp_undef_all (cpp_reader *);
1215 extern cpp_buffer *cpp_push_buffer (cpp_reader *, const unsigned char *,
1217 extern int cpp_defined (cpp_reader *, const unsigned char *, int);
1219 /* A preprocessing number. Code assumes that any unused high bits of
1220 the double integer are set to zero. */
1222 /* This type has to be equal to unsigned HOST_WIDE_INT, see
1223 gcc/c-family/c-lex.cc. */
1224 typedef uint64_t cpp_num_part;
1225 typedef struct cpp_num cpp_num;
1230 bool unsignedp; /* True if value should be treated as unsigned. */
1231 bool overflow; /* True if the most recent calculation overflowed. */
1234 /* cpplib provides two interfaces for interpretation of preprocessing
1237 cpp_classify_number categorizes numeric constants according to
1238 their field (integer, floating point, or invalid), radix (decimal,
1239 octal, hexadecimal), and type suffixes. */
1241 #define CPP_N_CATEGORY 0x000F
1242 #define CPP_N_INVALID 0x0000
1243 #define CPP_N_INTEGER 0x0001
1244 #define CPP_N_FLOATING 0x0002
1246 #define CPP_N_WIDTH 0x00F0
1247 #define CPP_N_SMALL 0x0010 /* int, float, short _Fract/Accum */
1248 #define CPP_N_MEDIUM 0x0020 /* long, double, long _Fract/_Accum. */
1249 #define CPP_N_LARGE 0x0040 /* long long, long double,
1250 long long _Fract/Accum. */
1252 #define CPP_N_WIDTH_MD 0xF0000 /* machine defined. */
1253 #define CPP_N_MD_W 0x10000
1254 #define CPP_N_MD_Q 0x20000
1256 #define CPP_N_RADIX 0x0F00
1257 #define CPP_N_DECIMAL 0x0100
1258 #define CPP_N_HEX 0x0200
1259 #define CPP_N_OCTAL 0x0400
1260 #define CPP_N_BINARY 0x0800
1262 #define CPP_N_UNSIGNED 0x1000 /* Properties. */
1263 #define CPP_N_IMAGINARY 0x2000
1264 #define CPP_N_DFLOAT 0x4000
1265 #define CPP_N_DEFAULT 0x8000
1267 #define CPP_N_FRACT 0x100000 /* Fract types. */
1268 #define CPP_N_ACCUM 0x200000 /* Accum types. */
1269 #define CPP_N_FLOATN 0x400000 /* _FloatN types. */
1270 #define CPP_N_FLOATNX 0x800000 /* _FloatNx types. */
1272 #define CPP_N_USERDEF 0x1000000 /* C++11 user-defined literal. */
1274 #define CPP_N_SIZE_T 0x2000000 /* C++23 size_t literal. */
1276 #define CPP_N_WIDTH_FLOATN_NX 0xF0000000 /* _FloatN / _FloatNx value
1277 of N, divided by 16. */
1278 #define CPP_FLOATN_SHIFT 24
1279 #define CPP_FLOATN_MAX 0xF0
1281 /* Classify a CPP_NUMBER token. The return value is a combination of
1282 the flags from the above sets. */
1283 extern unsigned cpp_classify_number (cpp_reader *, const cpp_token *,
1284 const char **, location_t);
1286 /* Return the classification flags for a float suffix. */
1287 extern unsigned int cpp_interpret_float_suffix (cpp_reader *, const char *,
1290 /* Return the classification flags for an int suffix. */
1291 extern unsigned int cpp_interpret_int_suffix (cpp_reader *, const char *,
1294 /* Evaluate a token classified as category CPP_N_INTEGER. */
1295 extern cpp_num cpp_interpret_integer (cpp_reader *, const cpp_token *,
1298 /* Sign extend a number, with PRECISION significant bits and all
1299 others assumed clear, to fill out a cpp_num structure. */
1300 cpp_num cpp_num_sign_extend (cpp_num, size_t);
1302 /* Output a diagnostic of some kind. */
1303 extern bool cpp_error (cpp_reader *, enum cpp_diagnostic_level,
1304 const char *msgid, ...)
1306 extern bool cpp_warning (cpp_reader *, enum cpp_warning_reason,
1307 const char *msgid, ...)
1309 extern bool cpp_pedwarning (cpp_reader *, enum cpp_warning_reason,
1310 const char *msgid, ...)
1312 extern bool cpp_warning_syshdr (cpp_reader *, enum cpp_warning_reason reason,
1313 const char *msgid, ...)
1316 /* As their counterparts above, but use RICHLOC. */
1317 extern bool cpp_warning_at (cpp_reader *, enum cpp_warning_reason,
1318 rich_location *richloc, const char *msgid, ...)
1320 extern bool cpp_pedwarning_at (cpp_reader *, enum cpp_warning_reason,
1321 rich_location *richloc, const char *msgid, ...)
1324 /* Output a diagnostic with "MSGID: " preceding the
1325 error string of errno. No location is printed. */
1326 extern bool cpp_errno (cpp_reader *, enum cpp_diagnostic_level,
1328 /* Similarly, but with "FILENAME: " instead of "MSGID: ", where
1329 the filename is not localized. */
1330 extern bool cpp_errno_filename (cpp_reader *, enum cpp_diagnostic_level,
1331 const char *filename, location_t loc);
1333 /* Same as cpp_error, except additionally specifies a position as a
1334 (translation unit) physical line and physical column. If the line is
1335 zero, then no location is printed. */
1336 extern bool cpp_error_with_line (cpp_reader *, enum cpp_diagnostic_level,
1337 location_t, unsigned,
1338 const char *msgid, ...)
1340 extern bool cpp_warning_with_line (cpp_reader *, enum cpp_warning_reason,
1341 location_t, unsigned,
1342 const char *msgid, ...)
1344 extern bool cpp_pedwarning_with_line (cpp_reader *, enum cpp_warning_reason,
1345 location_t, unsigned,
1346 const char *msgid, ...)
1348 extern bool cpp_warning_with_line_syshdr (cpp_reader *, enum cpp_warning_reason,
1349 location_t, unsigned,
1350 const char *msgid, ...)
1353 extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1354 location_t src_loc, const char *msgid, ...)
1357 extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1358 rich_location *richloc, const char *msgid, ...)
1362 extern int cpp_ideq (const cpp_token *, const char *);
1363 extern void cpp_output_line (cpp_reader *, FILE *);
1364 extern unsigned char *cpp_output_line_to_string (cpp_reader *,
1365 const unsigned char *);
1366 extern const unsigned char *cpp_alloc_token_string
1367 (cpp_reader *, const unsigned char *, unsigned);
1368 extern void cpp_output_token (const cpp_token *, FILE *);
1369 extern const char *cpp_type2name (enum cpp_ttype, unsigned char flags);
1370 /* Returns the value of an escape sequence, truncated to the correct
1371 target precision. PSTR points to the input pointer, which is just
1372 after the backslash. LIMIT is how much text we have. WIDE is true
1373 if the escape sequence is part of a wide character constant or
1374 string literal. Handles all relevant diagnostics. */
1375 extern cppchar_t cpp_parse_escape (cpp_reader *, const unsigned char ** pstr,
1376 const unsigned char *limit, int wide);
1378 /* Structure used to hold a comment block at a given location in the
1383 /* Text of the comment including the terminators. */
1386 /* source location for the given comment. */
1390 /* Structure holding all comments for a given cpp_reader. */
1394 /* table of comment entries. */
1395 cpp_comment *entries;
1397 /* number of actual entries entered in the table. */
1400 /* number of entries allocated currently. */
1402 } cpp_comment_table;
1404 /* Returns the table of comments encountered by the preprocessor. This
1405 table is only populated when pfile->state.save_comments is true. */
1406 extern cpp_comment_table *cpp_get_comments (cpp_reader *);
1410 /* Lookup an identifier in the hashtable. Puts the identifier in the
1411 table if it is not already there. */
1412 extern cpp_hashnode *cpp_lookup (cpp_reader *, const unsigned char *,
1415 typedef int (*cpp_cb) (cpp_reader *, cpp_hashnode *, void *);
1416 extern void cpp_forall_identifiers (cpp_reader *, cpp_cb, void *);
1419 extern void cpp_scan_nooutput (cpp_reader *);
1420 extern int cpp_sys_macro_p (cpp_reader *);
1421 extern unsigned char *cpp_quote_string (unsigned char *, const unsigned char *,
1423 extern bool cpp_compare_macros (const cpp_macro *macro1,
1424 const cpp_macro *macro2);
1427 extern bool cpp_included (cpp_reader *, const char *);
1428 extern bool cpp_included_before (cpp_reader *, const char *, location_t);
1429 extern void cpp_make_system_header (cpp_reader *, int, int);
1430 extern bool cpp_push_include (cpp_reader *, const char *);
1431 extern bool cpp_push_default_include (cpp_reader *, const char *);
1432 extern void cpp_change_file (cpp_reader *, enum lc_reason, const char *);
1433 extern const char *cpp_get_path (struct _cpp_file *);
1434 extern cpp_dir *cpp_get_dir (struct _cpp_file *);
1435 extern cpp_buffer *cpp_get_buffer (cpp_reader *);
1436 extern struct _cpp_file *cpp_get_file (cpp_buffer *);
1437 extern cpp_buffer *cpp_get_prev (cpp_buffer *);
1438 extern void cpp_clear_file_cache (cpp_reader *);
1440 /* cpp_get_converted_source returns the contents of the given file, as it exists
1441 after cpplib has read it and converted it from the input charset to the
1442 source charset. Return struct will be zero-filled if the data could not be
1443 read for any reason. The data starts at the DATA pointer, but the TO_FREE
1444 pointer is what should be passed to free(), as there may be an offset. */
1445 struct cpp_converted_source
1451 cpp_converted_source cpp_get_converted_source (const char *fname,
1452 const char *input_charset);
1455 struct save_macro_data;
1456 extern int cpp_save_state (cpp_reader *, FILE *);
1457 extern int cpp_write_pch_deps (cpp_reader *, FILE *);
1458 extern int cpp_write_pch_state (cpp_reader *, FILE *);
1459 extern int cpp_valid_state (cpp_reader *, const char *, int);
1460 extern void cpp_prepare_state (cpp_reader *, struct save_macro_data **);
1461 extern int cpp_read_state (cpp_reader *, const char *, FILE *,
1462 struct save_macro_data *);
1465 extern void cpp_force_token_locations (cpp_reader *, location_t);
1466 extern void cpp_stop_forcing_token_locations (cpp_reader *);
1474 extern void cpp_directive_only_process (cpp_reader *pfile,
1476 void (*cb) (cpp_reader *,
1481 extern enum cpp_ttype cpp_userdef_string_remove_type
1482 (enum cpp_ttype type);
1483 extern enum cpp_ttype cpp_userdef_string_add_type
1484 (enum cpp_ttype type);
1485 extern enum cpp_ttype cpp_userdef_char_remove_type
1486 (enum cpp_ttype type);
1487 extern enum cpp_ttype cpp_userdef_char_add_type
1488 (enum cpp_ttype type);
1489 extern bool cpp_userdef_string_p
1490 (enum cpp_ttype type);
1491 extern bool cpp_userdef_char_p
1492 (enum cpp_ttype type);
1493 extern const char * cpp_get_userdef_suffix
1494 (const cpp_token *);
1498 /* The result of attempting to decode a run of UTF-8 bytes. */
1500 struct cpp_decoded_char
1502 const char *m_start_byte;
1503 const char *m_next_byte;
1509 /* Information for mapping between code points and display columns.
1511 This is a tabstop value, along with a callback for getting the
1512 widths of characters. Normally this callback is cpp_wcwidth, but we
1513 support other schemes for escaping non-ASCII unicode as a series of
1514 ASCII chars when printing the user's source code in diagnostic-show-locus.cc
1516 For example, consider:
1517 - the Unicode character U+03C0 "GREEK SMALL LETTER PI" (UTF-8: 0xCF 0x80)
1518 - the Unicode character U+1F642 "SLIGHTLY SMILING FACE"
1519 (UTF-8: 0xF0 0x9F 0x99 0x82)
1520 - the byte 0xBF (a stray trailing byte of a UTF-8 character)
1521 Normally U+03C0 would occupy one display column, U+1F642
1522 would occupy two display columns, and the stray byte would be
1523 printed verbatim as one display column.
1525 However when escaping them as unicode code points as "<U+03C0>"
1526 and "<U+1F642>" they occupy 8 and 9 display columns respectively,
1527 and when escaping them as bytes as "<CF><80>" and "<F0><9F><99><82>"
1528 they occupy 8 and 16 display columns respectively. In both cases
1529 the stray byte is escaped to <BF> as 4 display columns. */
1531 struct cpp_char_column_policy
1533 cpp_char_column_policy (int tabstop,
1534 int (*width_cb) (cppchar_t c))
1535 : m_tabstop (tabstop),
1536 m_undecoded_byte_width (1),
1537 m_width_cb (width_cb)
1541 /* Width in display columns of a stray byte that isn't decodable
1543 int m_undecoded_byte_width;
1544 int (*m_width_cb) (cppchar_t c);
1547 /* A class to manage the state while converting a UTF-8 sequence to cppchar_t
1548 and computing the display width one character at a time. */
1549 class cpp_display_width_computation {
1551 cpp_display_width_computation (const char *data, int data_length,
1552 const cpp_char_column_policy &policy);
1553 const char *next_byte () const { return m_next; }
1554 int bytes_processed () const { return m_next - m_begin; }
1555 int bytes_left () const { return m_bytes_left; }
1556 bool done () const { return !bytes_left (); }
1557 int display_cols_processed () const { return m_display_cols; }
1559 int process_next_codepoint (cpp_decoded_char *out);
1560 int advance_display_cols (int n);
1563 const char *const m_begin;
1565 size_t m_bytes_left;
1566 const cpp_char_column_policy &m_policy;
1570 /* Convenience functions that are simple use cases for class
1571 cpp_display_width_computation. Tab characters will be expanded to spaces
1572 as determined by POLICY.m_tabstop, and non-printable-ASCII characters
1573 will be escaped as per POLICY. */
1575 int cpp_byte_column_to_display_column (const char *data, int data_length,
1577 const cpp_char_column_policy &policy);
1578 inline int cpp_display_width (const char *data, int data_length,
1579 const cpp_char_column_policy &policy)
1581 return cpp_byte_column_to_display_column (data, data_length, data_length,
1584 int cpp_display_column_to_byte_column (const char *data, int data_length,
1586 const cpp_char_column_policy &policy);
1587 int cpp_wcwidth (cppchar_t c);
1589 bool cpp_input_conversion_is_trivial (const char *input_charset);
1590 int cpp_check_utf8_bom (const char *data, size_t data_length);
1592 #endif /* ! LIBCPP_CPPLIB_H */
projects / platform / upstream / gcc.git / blob