1 /* Definitions for CPP library.
2 Copyright (C) 1995-2023 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 */ \
132 TK(UNEVAL_STRING, LITERAL) /* unevaluated "string" - C++26 */ \
134 TK(CHAR_USERDEF, LITERAL) /* 'char'_suffix - C++11 */ \
135 TK(WCHAR_USERDEF, LITERAL) /* L'char'_suffix - C++11 */ \
136 TK(CHAR16_USERDEF, LITERAL) /* u'char'_suffix - C++11 */ \
137 TK(CHAR32_USERDEF, LITERAL) /* U'char'_suffix - C++11 */ \
138 TK(UTF8CHAR_USERDEF, LITERAL) /* u8'char'_suffix - C++11 */ \
139 TK(STRING_USERDEF, LITERAL) /* "string"_suffix - C++11 */ \
140 TK(WSTRING_USERDEF, LITERAL) /* L"string"_suffix - C++11 */ \
141 TK(STRING16_USERDEF, LITERAL) /* u"string"_suffix - C++11 */ \
142 TK(STRING32_USERDEF, LITERAL) /* U"string"_suffix - C++11 */ \
143 TK(UTF8STRING_USERDEF,LITERAL) /* u8"string"_suffix - C++11 */ \
145 TK(COMMENT, LITERAL) /* Only if output comments. */ \
146 /* SPELL_LITERAL happens to DTRT. */ \
147 TK(MACRO_ARG, NONE) /* Macro argument. */ \
148 TK(PRAGMA, NONE) /* Only for deferred pragmas. */ \
149 TK(PRAGMA_EOL, NONE) /* End-of-line for deferred pragmas. */ \
150 TK(PADDING, NONE) /* Whitespace for -E. */
152 #define OP(e, s) CPP_ ## e,
153 #define TK(e, s) CPP_ ## e,
159 /* A token type for keywords, as opposed to ordinary identifiers. */
162 /* Positions in the table. */
163 CPP_LAST_EQ
= CPP_LSHIFT
,
164 CPP_FIRST_DIGRAPH
= CPP_HASH
,
165 CPP_LAST_PUNCTUATOR
= CPP_ATSIGN
,
166 CPP_LAST_CPP_OP
= CPP_LESS_EQ
171 /* C language kind, used when calling cpp_create_reader. */
172 enum c_lang
{CLK_GNUC89
= 0, CLK_GNUC99
, CLK_GNUC11
, CLK_GNUC17
, CLK_GNUC2X
,
173 CLK_STDC89
, CLK_STDC94
, CLK_STDC99
, CLK_STDC11
, CLK_STDC17
,
175 CLK_GNUCXX
, CLK_CXX98
, CLK_GNUCXX11
, CLK_CXX11
,
176 CLK_GNUCXX14
, CLK_CXX14
, CLK_GNUCXX17
, CLK_CXX17
,
177 CLK_GNUCXX20
, CLK_CXX20
, CLK_GNUCXX23
, CLK_CXX23
,
178 CLK_GNUCXX26
, CLK_CXX26
, CLK_ASM
};
180 /* Payload of a NUMBER, STRING, CHAR or COMMENT token. */
181 struct GTY(()) cpp_string
{
184 /* TEXT is always null terminated (terminator not included in len); but this
185 GTY markup arranges that PCH streaming works properly even if there is a
186 null byte in the middle of the string. */
187 const unsigned char * GTY((string_length ("1 + %h.len"))) text
;
190 /* Flags for the cpp_token structure. */
191 #define PREV_WHITE (1 << 0) /* If whitespace before this token. */
192 #define DIGRAPH (1 << 1) /* If it was a digraph. */
193 #define STRINGIFY_ARG (1 << 2) /* If macro argument to be stringified. */
194 #define PASTE_LEFT (1 << 3) /* If on LHS of a ## operator. */
195 #define NAMED_OP (1 << 4) /* C++ named operators. */
196 #define PREV_FALLTHROUGH (1 << 5) /* On a token preceeded by FALLTHROUGH
198 #define DECIMAL_INT (1 << 6) /* Decimal integer, set in c-lex.cc. */
199 #define PURE_ZERO (1 << 7) /* Single 0 digit, used by the C++ frontend,
201 #define SP_DIGRAPH (1 << 8) /* # or ## token was a digraph. */
202 #define SP_PREV_WHITE (1 << 9) /* If whitespace before a ##
203 operator, or before this token
204 after a # operator. */
205 #define NO_EXPAND (1 << 10) /* Do not macro-expand this token. */
206 #define PRAGMA_OP (1 << 11) /* _Pragma token. */
207 #define BOL (1 << 12) /* Token at beginning of line. */
209 /* Specify which field, if any, of the cpp_token union is used. */
211 enum cpp_token_fld_kind
{
213 CPP_TOKEN_FLD_SOURCE
,
215 CPP_TOKEN_FLD_ARG_NO
,
216 CPP_TOKEN_FLD_TOKEN_NO
,
217 CPP_TOKEN_FLD_PRAGMA
,
221 /* A macro argument in the cpp_token union. */
222 struct GTY(()) cpp_macro_arg
{
223 /* Argument number. */
225 /* The original spelling of the macro argument token. */
227 GTY ((nested_ptr (union tree_node
,
228 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
229 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
233 /* An identifier in the cpp_token union. */
234 struct GTY(()) cpp_identifier
{
235 /* The canonical (UTF-8) spelling of the identifier. */
237 GTY ((nested_ptr (union tree_node
,
238 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
239 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
241 /* The original spelling of the identifier. */
243 GTY ((nested_ptr (union tree_node
,
244 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
245 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
249 /* A preprocessing token. This has been carefully packed and should
250 occupy 16 bytes on 32-bit hosts and 24 bytes on 64-bit hosts. */
251 struct GTY(()) cpp_token
{
253 /* Location of first char of token, together with range of full token. */
256 ENUM_BITFIELD(cpp_ttype
) type
: CHAR_BIT
; /* token type */
257 unsigned short flags
; /* flags - see above */
262 struct cpp_identifier
GTY ((tag ("CPP_TOKEN_FLD_NODE"))) node
;
264 /* Inherit padding from this token. */
265 cpp_token
* GTY ((tag ("CPP_TOKEN_FLD_SOURCE"))) source
;
267 /* A string, or number. */
268 struct cpp_string
GTY ((tag ("CPP_TOKEN_FLD_STR"))) str
;
270 /* Argument no. (and original spelling) for a CPP_MACRO_ARG. */
271 struct cpp_macro_arg
GTY ((tag ("CPP_TOKEN_FLD_ARG_NO"))) macro_arg
;
273 /* Original token no. for a CPP_PASTE (from a sequence of
274 consecutive paste tokens in a macro expansion). */
275 unsigned int GTY ((tag ("CPP_TOKEN_FLD_TOKEN_NO"))) token_no
;
277 /* Caller-supplied identifier for a CPP_PRAGMA. */
278 unsigned int GTY ((tag ("CPP_TOKEN_FLD_PRAGMA"))) pragma
;
279 } GTY ((desc ("cpp_token_val_index (&%1)"))) val
;
282 /* Say which field is in use. */
283 extern enum cpp_token_fld_kind
cpp_token_val_index (const cpp_token
*tok
);
285 /* A type wide enough to hold any multibyte source character.
286 cpplib's character constant interpreter requires an unsigned type.
287 Also, a typedef for the signed equivalent.
288 The width of this type is capped at 32 bits; there do exist targets
289 where wchar_t is 64 bits, but only in a non-default mode, and there
290 would be no meaningful interpretation for a wchar_t value greater
291 than 2^32 anyway -- the widest wide-character encoding around is
292 ISO 10646, which stops at 2^31. */
293 #if CHAR_BIT * SIZEOF_INT >= 32
294 # define CPPCHAR_SIGNED_T int
295 #elif CHAR_BIT * SIZEOF_LONG >= 32
296 # define CPPCHAR_SIGNED_T long
298 # error "Cannot find a least-32-bit signed integer type"
300 typedef unsigned CPPCHAR_SIGNED_T cppchar_t
;
301 typedef CPPCHAR_SIGNED_T cppchar_signed_t
;
303 /* Style of header dependencies to generate. */
304 enum cpp_deps_style
{ DEPS_NONE
= 0, DEPS_USER
, DEPS_SYSTEM
};
306 /* Structured format of module dependencies to generate. */
307 enum cpp_fdeps_format
{ FDEPS_FMT_NONE
= 0, FDEPS_FMT_P1689R5
};
309 /* The possible normalization levels, from most restrictive to least. */
310 enum cpp_normalize_level
{
315 /* In NFC, except for subsequences where being in NFC would make
316 the identifier invalid. */
317 normalized_identifier_C
,
318 /* Not normalized at all. */
324 CMS_none
, /* A regular source file. */
325 CMS_header
, /* Is a directly-specified header file (eg PCH or
327 CMS_user
, /* Search the user INCLUDE path. */
328 CMS_system
, /* Search the system INCLUDE path. */
331 /* The possible bidirectional control characters checking levels. */
332 enum cpp_bidirectional_level
{
334 bidirectional_none
= 0,
335 /* Only detect unpaired uses of bidirectional control characters. */
336 bidirectional_unpaired
= 1,
337 /* Detect any use of bidirectional control characters. */
338 bidirectional_any
= 2,
339 /* Also warn about UCNs. */
340 bidirectional_ucn
= 4
343 /* This structure is nested inside struct cpp_reader, and
344 carries all the options visible to the command line. */
347 /* The language we're preprocessing. */
350 /* Nonzero means use extra default include directories for C++. */
351 unsigned char cplusplus
;
353 /* Nonzero means handle cplusplus style comments. */
354 unsigned char cplusplus_comments
;
356 /* Nonzero means define __OBJC__, treat @ as a special token, use
357 the OBJC[PLUS]_INCLUDE_PATH environment variable, and allow
361 /* Nonzero means don't copy comments into the output file. */
362 unsigned char discard_comments
;
364 /* Nonzero means don't copy comments into the output file during
366 unsigned char discard_comments_in_macro_exp
;
368 /* Nonzero means process the ISO trigraph sequences. */
369 unsigned char trigraphs
;
371 /* Nonzero means process the ISO digraph sequences. */
372 unsigned char digraphs
;
374 /* Nonzero means to allow hexadecimal floats and LL suffixes. */
375 unsigned char extended_numbers
;
377 /* Nonzero means process u/U prefix literals (UTF-16/32). */
378 unsigned char uliterals
;
380 /* Nonzero means process u8 prefixed character literals (UTF-8). */
381 unsigned char utf8_char_literals
;
383 /* Nonzero means process r/R raw strings. If this is set, uliterals
384 must be set as well. */
385 unsigned char rliterals
;
387 /* Nonzero means print names of header files (-H). */
388 unsigned char print_include_names
;
390 /* Nonzero means complain about deprecated features. */
391 unsigned char cpp_warn_deprecated
;
393 /* Nonzero means warn if slash-star appears in a comment. */
394 unsigned char warn_comments
;
396 /* Nonzero means to warn about __DATA__, __TIME__ and __TIMESTAMP__ usage. */
397 unsigned char warn_date_time
;
399 /* Nonzero means warn if a user-supplied include directory does not
401 unsigned char warn_missing_include_dirs
;
403 /* Nonzero means warn if there are any trigraphs. */
404 unsigned char warn_trigraphs
;
406 /* Nonzero means warn about multicharacter charconsts. */
407 unsigned char warn_multichar
;
409 /* Nonzero means warn about various incompatibilities with
411 unsigned char cpp_warn_traditional
;
413 /* Nonzero means warn about long long numeric constants. */
414 unsigned char cpp_warn_long_long
;
416 /* Nonzero means warn about text after an #endif (or #else). */
417 unsigned char warn_endif_labels
;
419 /* Nonzero means warn about implicit sign changes owing to integer
421 unsigned char warn_num_sign_change
;
423 /* Zero means don't warn about __VA_ARGS__ usage in c89 pedantic mode.
424 Presumably the usage is protected by the appropriate #ifdef. */
425 unsigned char warn_variadic_macros
;
427 /* Nonzero means warn about builtin macros that are redefined or
428 explicitly undefined. */
429 unsigned char warn_builtin_macro_redefined
;
431 /* Different -Wimplicit-fallthrough= levels. */
432 unsigned char cpp_warn_implicit_fallthrough
;
434 /* Nonzero means we should look for header.gcc files that remap file
438 /* Zero means dollar signs are punctuation. */
439 unsigned char dollars_in_ident
;
441 /* Nonzero means UCNs are accepted in identifiers. */
442 unsigned char extended_identifiers
;
444 /* True if we should warn about dollars in identifiers or numbers
445 for this translation unit. */
446 unsigned char warn_dollars
;
448 /* Nonzero means warn if undefined identifiers are evaluated in an #if. */
449 unsigned char warn_undef
;
451 /* Nonzero means warn if "defined" is encountered in a place other than
453 unsigned char warn_expansion_to_defined
;
455 /* Nonzero means warn of unused macros from the main file. */
456 unsigned char warn_unused_macros
;
458 /* Nonzero for the 1999 C Standard, including corrigenda and amendments. */
461 /* Nonzero if we are conforming to a specific C or C++ standard. */
464 /* Nonzero means give all the error messages the ANSI standard requires. */
465 unsigned char cpp_pedantic
;
467 /* Nonzero means we're looking at already preprocessed code, so don't
468 bother trying to do macro expansion and whatnot. */
469 unsigned char preprocessed
;
471 /* Nonzero means we are going to emit debugging logs during
475 /* Nonzero means we are tracking locations of tokens involved in
476 macro expansion. 1 Means we track the location in degraded mode
477 where we do not track locations of tokens resulting from the
478 expansion of arguments of function-like macro. 2 Means we do
479 track all macro expansions. This last option is the one that
480 consumes the highest amount of memory. */
481 unsigned char track_macro_expansion
;
483 /* Nonzero means handle C++ alternate operator names. */
484 unsigned char operator_names
;
486 /* Nonzero means warn about use of C++ alternate operator names. */
487 unsigned char warn_cxx_operator_names
;
489 /* True for traditional preprocessing. */
490 unsigned char traditional
;
492 /* Nonzero for C++ 2011 Standard user-defined literals. */
493 unsigned char user_literals
;
495 /* Nonzero means warn when a string or character literal is followed by a
496 ud-suffix which does not beging with an underscore. */
497 unsigned char warn_literal_suffix
;
499 /* Nonzero means interpret imaginary, fixed-point, or other gnu extension
500 literal number suffixes as user-defined literal number suffixes. */
501 unsigned char ext_numeric_literals
;
503 /* Nonzero means extended identifiers allow the characters specified
505 unsigned char c11_identifiers
;
507 /* Nonzero means extended identifiers allow the characters specified
508 by Unicode XID_Start and XID_Continue properties. */
509 unsigned char xid_identifiers
;
511 /* Nonzero for C++ 2014 Standard binary constants. */
512 unsigned char binary_constants
;
514 /* Nonzero for C++ 2014 Standard digit separators. */
515 unsigned char digit_separators
;
517 /* Nonzero for C2X decimal floating-point constants. */
518 unsigned char dfp_constants
;
520 /* Nonzero for C++20 __VA_OPT__ feature. */
521 unsigned char va_opt
;
523 /* Nonzero for the '::' token. */
526 /* Nonzero for the '#elifdef' and '#elifndef' directives. */
527 unsigned char elifdef
;
529 /* Nonzero for the '#warning' directive. */
530 unsigned char warning_directive
;
532 /* Nonzero means tokenize C++20 module directives. */
533 unsigned char module_directives
;
535 /* Nonzero for C++23 size_t literals. */
536 unsigned char size_t_literals
;
538 /* Nonzero for C++23 delimited escape sequences. */
539 unsigned char delimited_escape_seqs
;
541 /* Nonzero for 'true' and 'false' in #if expressions. */
542 unsigned char true_false
;
544 /* Holds the name of the target (execution) character set. */
545 const char *narrow_charset
;
547 /* Holds the name of the target wide character set. */
548 const char *wide_charset
;
550 /* Holds the name of the input character set. */
551 const char *input_charset
;
553 /* The minimum permitted level of normalization before a warning
554 is generated. See enum cpp_normalize_level. */
557 /* True to warn about precompiled header files we couldn't use. */
558 bool warn_invalid_pch
;
560 /* True if dependencies should be restored from a precompiled header. */
561 bool restore_pch_deps
;
563 /* True if warn about differences between C90 and C99. */
564 signed char cpp_warn_c90_c99_compat
;
566 /* True if warn about differences between C11 and C2X. */
567 signed char cpp_warn_c11_c2x_compat
;
569 /* True if warn about differences between C++98 and C++11. */
570 bool cpp_warn_cxx11_compat
;
572 /* True if warn about differences between C++17 and C++20. */
573 bool cpp_warn_cxx20_compat
;
575 /* Nonzero if bidirectional control characters checking is on. See enum
576 cpp_bidirectional_level. */
577 unsigned char cpp_warn_bidirectional
;
579 /* True if libcpp should warn about invalid UTF-8 characters in comments.
580 2 if it should be a pedwarn. */
581 unsigned char cpp_warn_invalid_utf8
;
583 /* True if libcpp should warn about invalid forms of delimited or named
585 bool cpp_warn_unicode
;
587 /* True if -finput-charset= option has been used explicitly. */
588 bool cpp_input_charset_explicit
;
590 /* Dependency generation. */
593 /* Style of header dependencies to generate. */
594 enum cpp_deps_style style
;
596 /* Structured format of module dependencies to generate. */
597 enum cpp_fdeps_format fdeps_format
;
599 /* Assume missing files are generated files. */
602 /* Generate phony targets for each dependency apart from the first
606 /* Generate dependency info for modules. */
609 /* If true, no dependency is generated on the main file. */
610 bool ignore_main_file
;
612 /* If true, intend to use the preprocessor output (e.g., for compilation)
613 in addition to the dependency info. */
614 bool need_preprocessor_output
;
617 /* Target-specific features set by the front end or client. */
619 /* Precision for target CPP arithmetic, target characters, target
620 ints and target wide characters, respectively. */
621 size_t precision
, char_precision
, int_precision
, wchar_precision
;
623 /* True means chars (wide chars, UTF-8 chars) are unsigned. */
624 bool unsigned_char
, unsigned_wchar
, unsigned_utf8char
;
626 /* True if the most significant byte in a word has the lowest
627 address in memory. */
628 bool bytes_big_endian
;
630 /* Nonzero means __STDC__ should have the value 0 in system headers. */
631 unsigned char stdc_0_in_system_headers
;
633 /* True disables tokenization outside of preprocessing directives. */
634 bool directives_only
;
636 /* True enables canonicalization of system header file paths. */
637 bool canonical_system_headers
;
639 /* The maximum depth of the nested #include. */
640 unsigned int max_include_depth
;
642 cpp_main_search main_search
: 8;
645 /* Diagnostic levels. To get a diagnostic without associating a
646 position in the translation unit with it, use cpp_error_with_line
647 with a line number of zero. */
649 enum cpp_diagnostic_level
{
650 /* Warning, an error with -Werror. */
652 /* Same as CPP_DL_WARNING, except it is not suppressed in system headers. */
653 CPP_DL_WARNING_SYSHDR
,
654 /* Warning, an error with -pedantic-errors or -Werror. */
658 /* An internal consistency check failed. Prints "internal error: ",
659 otherwise the same as CPP_DL_ERROR. */
661 /* An informative note following a warning. */
667 /* Warning reason codes. Use a reason code of CPP_W_NONE for unclassified
668 warnings and diagnostics that are not warnings. */
670 enum cpp_warning_reason
{
674 CPP_W_MISSING_INCLUDE_DIRS
,
680 CPP_W_NUM_SIGN_CHANGE
,
681 CPP_W_VARIADIC_MACROS
,
682 CPP_W_BUILTIN_MACRO_REDEFINED
,
686 CPP_W_CXX_OPERATOR_NAMES
,
689 CPP_W_WARNING_DIRECTIVE
,
690 CPP_W_LITERAL_SUFFIX
,
691 CPP_W_SIZE_T_LITERALS
,
694 CPP_W_C90_C99_COMPAT
,
695 CPP_W_C11_C2X_COMPAT
,
698 CPP_W_EXPANSION_TO_DEFINED
,
704 /* Callback for header lookup for HEADER, which is the name of a
705 source file. It is used as a method of last resort to find headers
706 that are not otherwise found during the normal include processing.
707 The return value is the malloced name of a header to try and open,
708 if any, or NULL otherwise. This callback is called only if the
709 header is otherwise unfound. */
710 typedef const char *(*missing_header_cb
)(cpp_reader
*, const char *header
, cpp_dir
**);
712 /* Call backs to cpplib client. */
715 /* Called when a new line of preprocessed output is started. */
716 void (*line_change
) (cpp_reader
*, const cpp_token
*, int);
718 /* Called when switching to/from a new file.
719 The line_map is for the new file. It is NULL if there is no new file.
720 (In C this happens when done with <built-in>+<command line> and also
721 when done with a main file.) This can be used for resource cleanup. */
722 void (*file_change
) (cpp_reader
*, const line_map_ordinary
*);
724 void (*dir_change
) (cpp_reader
*, const char *);
725 void (*include
) (cpp_reader
*, location_t
, const unsigned char *,
726 const char *, int, const cpp_token
**);
727 void (*define
) (cpp_reader
*, location_t
, cpp_hashnode
*);
728 void (*undef
) (cpp_reader
*, location_t
, cpp_hashnode
*);
729 void (*ident
) (cpp_reader
*, location_t
, const cpp_string
*);
730 void (*def_pragma
) (cpp_reader
*, location_t
);
731 int (*valid_pch
) (cpp_reader
*, const char *, int);
732 void (*read_pch
) (cpp_reader
*, const char *, int, const char *);
733 missing_header_cb missing_header
;
735 /* Context-sensitive macro support. Returns macro (if any) that should
737 cpp_hashnode
* (*macro_to_expand
) (cpp_reader
*, const cpp_token
*);
739 /* Called to emit a diagnostic. This callback receives the
740 translated message. */
741 bool (*diagnostic
) (cpp_reader
*,
742 enum cpp_diagnostic_level
,
743 enum cpp_warning_reason
,
745 const char *, va_list *)
746 ATTRIBUTE_FPTR_PRINTF(5,0);
748 /* Callbacks for when a macro is expanded, or tested (whether
749 defined or not at the time) in #ifdef, #ifndef or "defined". */
750 void (*used_define
) (cpp_reader
*, location_t
, cpp_hashnode
*);
751 void (*used_undef
) (cpp_reader
*, location_t
, cpp_hashnode
*);
752 /* Called before #define and #undef or other macro definition
753 changes are processed. */
754 void (*before_define
) (cpp_reader
*);
755 /* Called whenever a macro is expanded or tested.
756 Second argument is the location of the start of the current expansion. */
757 void (*used
) (cpp_reader
*, location_t
, cpp_hashnode
*);
759 /* Callback to identify whether an attribute exists. */
760 int (*has_attribute
) (cpp_reader
*, bool);
762 /* Callback to determine whether a built-in function is recognized. */
763 int (*has_builtin
) (cpp_reader
*);
765 /* Callback that can change a user lazy into normal macro. */
766 void (*user_lazy_macro
) (cpp_reader
*, cpp_macro
*, unsigned);
768 /* Callback to handle deferred cpp_macros. */
769 cpp_macro
*(*user_deferred_macro
) (cpp_reader
*, location_t
, cpp_hashnode
*);
771 /* Callback to parse SOURCE_DATE_EPOCH from environment. */
772 time_t (*get_source_date_epoch
) (cpp_reader
*);
774 /* Callback for providing suggestions for misspelled directives. */
775 const char *(*get_suggestion
) (cpp_reader
*, const char *, const char *const *);
777 /* Callback for when a comment is encountered, giving the location
778 of the opening slash, a pointer to the content (which is not
779 necessarily 0-terminated), and the length of the content.
780 The content contains the opening slash-star (or slash-slash),
781 and for C-style comments contains the closing star-slash. For
782 C++-style comments it does not include the terminating newline. */
783 void (*comment
) (cpp_reader
*, location_t
, const unsigned char *,
786 /* Callback for filename remapping in __FILE__ and __BASE_FILE__ macro
788 const char *(*remap_filename
) (const char*);
790 /* Maybe translate a #include into something else. Return a
791 cpp_buffer containing the translation if translating. */
792 char *(*translate_include
) (cpp_reader
*, line_maps
*, location_t
,
797 #define INO_T_CPP ino_t ino[3]
798 #elif defined (_AIX) && SIZEOF_INO_T == 4
799 #define INO_T_CPP ino64_t ino
801 #define INO_T_CPP ino_t ino
804 #if defined (_AIX) && SIZEOF_DEV_T == 4
805 #define DEV_T_CPP dev64_t dev
807 #define DEV_T_CPP dev_t dev
810 /* Chain of directories to look for include files in. */
813 /* NULL-terminated singly-linked list. */
814 struct cpp_dir
*next
;
816 /* NAME of the directory, NUL-terminated. */
820 /* One if a system header, two if a system header that has extern
821 "C" guards for C++. */
824 /* Is this a user-supplied directory? */
825 bool user_supplied_p
;
827 /* The canonicalized NAME as determined by lrealpath. This field
828 is only used by hosts that lack reliable inode numbers. */
829 char *canonical_name
;
831 /* Mapping of file names for this directory for MS-DOS and related
832 platforms. A NULL-terminated array of (from, to) pairs. */
833 const char **name_map
;
835 /* Routine to construct pathname, given the search path name and the
836 HEADER we are trying to find, return a constructed pathname to
837 try and open. If this is NULL, the constructed pathname is as
838 constructed by append_file_to_dir. */
839 char *(*construct
) (const char *header
, cpp_dir
*dir
);
841 /* The C front end uses these to recognize duplicated
842 directories in the search path. */
847 /* The kind of the cpp_macro. */
848 enum cpp_macro_kind
{
849 cmk_macro
, /* An ISO macro (token expansion). */
850 cmk_assert
, /* An assertion. */
851 cmk_traditional
/* A traditional macro (text expansion). */
854 /* Each macro definition is recorded in a cpp_macro structure.
855 Variadic macros cannot occur with traditional cpp. */
856 struct GTY(()) cpp_macro
{
859 /* Parameters, if any. If parameter names use extended identifiers,
860 the original spelling of those identifiers, not the canonical
861 UTF-8 spelling, goes here. */
862 cpp_hashnode
** GTY ((tag ("false"),
863 nested_ptr (union tree_node
,
864 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
865 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL"),
866 length ("%1.paramc"))) params
;
868 /* If this is an assertion, the next one in the chain. */
869 cpp_macro
*GTY ((tag ("true"))) next
;
870 } GTY ((desc ("%1.kind == cmk_assert"))) parm
;
872 /* Definition line number. */
875 /* Number of tokens in body, or bytes for traditional macros. */
876 /* Do we really need 2^32-1 range here? */
879 /* Number of parameters. */
880 unsigned short paramc
;
882 /* Non-zero if this is a user-lazy macro, value provided by user. */
885 /* The kind of this macro (ISO, trad or assert) */
888 /* If a function-like macro. */
889 unsigned int fun_like
: 1;
891 /* If a variadic macro. */
892 unsigned int variadic
: 1;
894 /* If macro defined in system header. */
895 unsigned int syshdr
: 1;
897 /* Nonzero if it has been expanded or had its existence tested. */
898 unsigned int used
: 1;
900 /* Indicate whether the tokens include extra CPP_PASTE tokens at the
901 end to track invalid redefinitions with consecutive CPP_PASTE
903 unsigned int extra_tokens
: 1;
905 /* Imported C++20 macro (from a header unit). */
906 unsigned int imported_p
: 1;
908 /* 0 bits spare (32-bit). 32 on 64-bit target. */
912 /* Trailing array of replacement tokens (ISO), or assertion body value. */
913 cpp_token
GTY ((tag ("false"), length ("%1.count"))) tokens
[1];
915 /* Pointer to replacement text (traditional). See comment at top
916 of cpptrad.c for how traditional function-like macros are
918 const unsigned char *GTY ((tag ("true"))) text
;
919 } GTY ((desc ("%1.kind == cmk_traditional"))) exp
;
922 /* Poisoned identifiers are flagged NODE_POISONED. NODE_OPERATOR (C++
923 only) indicates an identifier that behaves like an operator such as
924 "xor". NODE_DIAGNOSTIC is for speed in lex_token: it indicates a
925 diagnostic may be required for this node. Currently this only
926 applies to __VA_ARGS__, poisoned identifiers, and -Wc++-compat
927 warnings about NODE_OPERATOR. */
929 /* Hash node flags. */
930 #define NODE_OPERATOR (1 << 0) /* C++ named operator. */
931 #define NODE_POISONED (1 << 1) /* Poisoned identifier. */
932 #define NODE_DIAGNOSTIC (1 << 2) /* Possible diagnostic when lexed. */
933 #define NODE_WARN (1 << 3) /* Warn if redefined or undefined. */
934 #define NODE_DISABLED (1 << 4) /* A disabled macro. */
935 #define NODE_USED (1 << 5) /* Dumped with -dU. */
936 #define NODE_CONDITIONAL (1 << 6) /* Conditional macro */
937 #define NODE_WARN_OPERATOR (1 << 7) /* Warn about C++ named operator. */
938 #define NODE_MODULE (1 << 8) /* C++-20 module-related name. */
940 /* Different flavors of hash node. */
943 NT_VOID
= 0, /* Maybe an assert? */
944 NT_MACRO_ARG
, /* A macro arg. */
945 NT_USER_MACRO
, /* A user macro. */
946 NT_BUILTIN_MACRO
, /* A builtin macro. */
947 NT_MACRO_MASK
= NT_USER_MACRO
/* Mask for either macro kind. */
950 /* Different flavors of builtin macro. _Pragma is an operator, but we
951 handle it with the builtin code for efficiency reasons. */
952 enum cpp_builtin_type
954 BT_SPECLINE
= 0, /* `__LINE__' */
955 BT_DATE
, /* `__DATE__' */
956 BT_FILE
, /* `__FILE__' */
957 BT_FILE_NAME
, /* `__FILE_NAME__' */
958 BT_BASE_FILE
, /* `__BASE_FILE__' */
959 BT_INCLUDE_LEVEL
, /* `__INCLUDE_LEVEL__' */
960 BT_TIME
, /* `__TIME__' */
961 BT_STDC
, /* `__STDC__' */
962 BT_PRAGMA
, /* `_Pragma' operator */
963 BT_TIMESTAMP
, /* `__TIMESTAMP__' */
964 BT_COUNTER
, /* `__COUNTER__' */
965 BT_HAS_ATTRIBUTE
, /* `__has_attribute(x)' */
966 BT_HAS_STD_ATTRIBUTE
, /* `__has_c_attribute(x)' */
967 BT_HAS_BUILTIN
, /* `__has_builtin(x)' */
968 BT_HAS_INCLUDE
, /* `__has_include(x)' */
969 BT_HAS_INCLUDE_NEXT
/* `__has_include_next(x)' */
972 #define CPP_HASHNODE(HNODE) ((cpp_hashnode *) (HNODE))
973 #define HT_NODE(NODE) (&(NODE)->ident)
974 #define NODE_LEN(NODE) HT_LEN (HT_NODE (NODE))
975 #define NODE_NAME(NODE) HT_STR (HT_NODE (NODE))
977 /* The common part of an identifier node shared amongst all 3 C front
978 ends. Also used to store CPP identifiers, which are a superset of
979 identifiers in the grammatical sense. */
981 union GTY(()) _cpp_hashnode_value
{
982 /* Assert (maybe NULL) */
983 cpp_macro
* GTY((tag ("NT_VOID"))) answers
;
984 /* Macro (maybe NULL) */
985 cpp_macro
* GTY((tag ("NT_USER_MACRO"))) macro
;
986 /* Code for a builtin macro. */
987 enum cpp_builtin_type
GTY ((tag ("NT_BUILTIN_MACRO"))) builtin
;
988 /* Macro argument index. */
989 unsigned short GTY ((tag ("NT_MACRO_ARG"))) arg_index
;
992 struct GTY(()) cpp_hashnode
{
993 struct ht_identifier ident
;
994 unsigned int is_directive
: 1;
995 unsigned int directive_index
: 7; /* If is_directive,
996 then index into directive table.
997 Otherwise, a NODE_OPERATOR. */
998 unsigned int rid_code
: 8; /* Rid code - for front ends. */
999 unsigned int flags
: 9; /* CPP flags. */
1000 ENUM_BITFIELD(node_type
) type
: 2; /* CPP node type. */
1004 /* The deferred cookie is applicable to NT_USER_MACRO or NT_VOID.
1005 The latter for when a macro had a prevailing undef.
1006 On a 64-bit system there would be 32-bits of padding to the value
1007 field. So placing the deferred index here is not costly. */
1008 unsigned deferred
; /* Deferred cookie */
1010 union _cpp_hashnode_value
GTY ((desc ("%1.type"))) value
;
1013 /* Extra information we may need to store per identifier, which is needed rarely
1014 enough that it's not worth adding directly into the main identifier hash. */
1015 struct GTY(()) cpp_hashnode_extra
1017 struct ht_identifier ident
;
1018 location_t poisoned_loc
;
1021 /* A class for iterating through the source locations within a
1022 string token (before escapes are interpreted, and before
1025 class cpp_string_location_reader
{
1027 cpp_string_location_reader (location_t src_loc
,
1028 line_maps
*line_table
);
1030 source_range
get_next ();
1034 int m_offset_per_column
;
1037 /* A class for storing the source ranges of all of the characters within
1038 a string literal, after escapes are interpreted, and after
1041 This is not GTY-marked, as instances are intended to be temporary. */
1043 class cpp_substring_ranges
1046 cpp_substring_ranges ();
1047 ~cpp_substring_ranges ();
1049 int get_num_ranges () const { return m_num_ranges
; }
1050 source_range
get_range (int idx
) const
1052 linemap_assert (idx
< m_num_ranges
);
1053 return m_ranges
[idx
];
1056 void add_range (source_range range
);
1057 void add_n_ranges (int num
, cpp_string_location_reader
&loc_reader
);
1060 source_range
*m_ranges
;
1065 /* Call this first to get a handle to pass to other functions.
1067 The first hash table argument is for associating a struct cpp_hashnode
1068 with each identifier. The second hash table argument is for associating
1069 a struct cpp_hashnode_extra with each identifier that needs one. For
1070 either, pass in a NULL pointer if you want cpplib to create and manage
1071 the hash table itself, or else pass a suitably initialized hash table to
1072 be managed external to libcpp, as is done by the C-family frontends. */
1073 extern cpp_reader
*cpp_create_reader (enum c_lang
, struct ht
*,
1075 struct ht
* = nullptr);
1077 /* Reset the cpp_reader's line_map. This is only used after reading a
1079 extern void cpp_set_line_map (cpp_reader
*, class line_maps
*);
1081 /* Call this to change the selected language standard (e.g. because of
1082 command line options). */
1083 extern void cpp_set_lang (cpp_reader
*, enum c_lang
);
1085 /* Set the include paths. */
1086 extern void cpp_set_include_chains (cpp_reader
*, cpp_dir
*, cpp_dir
*, int);
1088 /* Call these to get pointers to the options, callback, and deps
1089 structures for a given reader. These pointers are good until you
1090 call cpp_finish on that reader. You can either edit the callbacks
1091 through the pointer returned from cpp_get_callbacks, or set them
1092 with cpp_set_callbacks. */
1093 extern cpp_options
*cpp_get_options (cpp_reader
*) ATTRIBUTE_PURE
;
1094 extern cpp_callbacks
*cpp_get_callbacks (cpp_reader
*) ATTRIBUTE_PURE
;
1095 extern void cpp_set_callbacks (cpp_reader
*, cpp_callbacks
*);
1096 extern class mkdeps
*cpp_get_deps (cpp_reader
*) ATTRIBUTE_PURE
;
1098 extern const char *cpp_probe_header_unit (cpp_reader
*, const char *file
,
1099 bool angle_p
, location_t
);
1101 /* Call these to get name data about the various compile-time
1103 extern const char *cpp_get_narrow_charset_name (cpp_reader
*) ATTRIBUTE_PURE
;
1104 extern const char *cpp_get_wide_charset_name (cpp_reader
*) ATTRIBUTE_PURE
;
1106 /* This function reads the file, but does not start preprocessing. It
1107 returns the name of the original file; this is the same as the
1108 input file, except for preprocessed input. This will generate at
1109 least one file change callback, and possibly a line change callback
1110 too. If there was an error opening the file, it returns NULL. */
1111 extern const char *cpp_read_main_file (cpp_reader
*, const char *,
1112 bool injecting
= false);
1113 extern location_t
cpp_main_loc (const cpp_reader
*);
1115 /* Adjust for the main file to be an include. */
1116 extern void cpp_retrofit_as_include (cpp_reader
*);
1118 /* Set up built-ins with special behavior. Use cpp_init_builtins()
1119 instead unless your know what you are doing. */
1120 extern void cpp_init_special_builtins (cpp_reader
*);
1122 /* Set up built-ins like __FILE__. */
1123 extern void cpp_init_builtins (cpp_reader
*, int);
1125 /* This is called after options have been parsed, and partially
1127 extern void cpp_post_options (cpp_reader
*);
1129 /* Set up translation to the target character set. */
1130 extern void cpp_init_iconv (cpp_reader
*);
1132 /* Call this to finish preprocessing. If you requested dependency
1133 generation, pass open stream(s) to write the information to,
1134 otherwise NULL. It is your responsibility to close the stream(s). */
1135 extern void cpp_finish (cpp_reader
*, FILE *deps_stream
, FILE *fdeps_stream
= NULL
);
1137 /* Call this to release the handle at the end of preprocessing. Any
1138 use of the handle after this function returns is invalid. */
1139 extern void cpp_destroy (cpp_reader
*);
1141 extern unsigned int cpp_token_len (const cpp_token
*);
1142 extern unsigned char *cpp_token_as_text (cpp_reader
*, const cpp_token
*);
1143 extern unsigned char *cpp_spell_token (cpp_reader
*, const cpp_token
*,
1144 unsigned char *, bool);
1145 extern void cpp_register_pragma (cpp_reader
*, const char *, const char *,
1146 void (*) (cpp_reader
*), bool);
1147 extern void cpp_register_deferred_pragma (cpp_reader
*, const char *,
1148 const char *, unsigned, bool, bool);
1149 extern int cpp_avoid_paste (cpp_reader
*, const cpp_token
*,
1151 extern const cpp_token
*cpp_get_token (cpp_reader
*);
1152 extern const cpp_token
*cpp_get_token_with_location (cpp_reader
*,
1154 inline bool cpp_user_macro_p (const cpp_hashnode
*node
)
1156 return node
->type
== NT_USER_MACRO
;
1158 inline bool cpp_builtin_macro_p (const cpp_hashnode
*node
)
1160 return node
->type
== NT_BUILTIN_MACRO
;
1162 inline bool cpp_macro_p (const cpp_hashnode
*node
)
1164 return node
->type
& NT_MACRO_MASK
;
1166 inline cpp_macro
*cpp_set_deferred_macro (cpp_hashnode
*node
,
1167 cpp_macro
*forced
= NULL
)
1169 cpp_macro
*old
= node
->value
.macro
;
1171 node
->value
.macro
= forced
;
1172 node
->type
= NT_USER_MACRO
;
1173 node
->flags
&= ~NODE_USED
;
1177 cpp_macro
*cpp_get_deferred_macro (cpp_reader
*, cpp_hashnode
*, location_t
);
1179 /* Returns true if NODE is a function-like user macro. */
1180 inline bool cpp_fun_like_macro_p (cpp_hashnode
*node
)
1182 return cpp_user_macro_p (node
) && node
->value
.macro
->fun_like
;
1185 extern const unsigned char *cpp_macro_definition (cpp_reader
*, cpp_hashnode
*);
1186 extern const unsigned char *cpp_macro_definition (cpp_reader
*, cpp_hashnode
*,
1188 inline location_t
cpp_macro_definition_location (cpp_hashnode
*node
)
1190 const cpp_macro
*macro
= node
->value
.macro
;
1191 return macro
? macro
->line
: 0;
1193 /* Return an idempotent time stamp (possibly from SOURCE_DATE_EPOCH). */
1194 enum class CPP_time_kind
1196 FIXED
= -1, /* Fixed time via source epoch. */
1197 DYNAMIC
= -2, /* Dynamic via time(2). */
1198 UNKNOWN
= -3 /* Wibbly wobbly, timey wimey. */
1200 extern CPP_time_kind
cpp_get_date (cpp_reader
*, time_t *);
1202 extern void _cpp_backup_tokens (cpp_reader
*, unsigned int);
1203 extern const cpp_token
*cpp_peek_token (cpp_reader
*, int);
1205 /* Evaluate a CPP_*CHAR* token. */
1206 extern cppchar_t
cpp_interpret_charconst (cpp_reader
*, const cpp_token
*,
1207 unsigned int *, int *);
1208 /* Evaluate a vector of CPP_*STRING* tokens. */
1209 extern bool cpp_interpret_string (cpp_reader
*,
1210 const cpp_string
*, size_t,
1211 cpp_string
*, enum cpp_ttype
);
1212 extern const char *cpp_interpret_string_ranges (cpp_reader
*pfile
,
1213 const cpp_string
*from
,
1214 cpp_string_location_reader
*,
1216 cpp_substring_ranges
*out
,
1217 enum cpp_ttype type
);
1218 extern bool cpp_interpret_string_notranslate (cpp_reader
*,
1219 const cpp_string
*, size_t,
1220 cpp_string
*, enum cpp_ttype
);
1222 /* Convert a host character constant to the execution character set. */
1223 extern cppchar_t
cpp_host_to_exec_charset (cpp_reader
*, cppchar_t
);
1225 /* Used to register macros and assertions, perhaps from the command line.
1226 The text is the same as the command line argument. */
1227 extern void cpp_define (cpp_reader
*, const char *);
1228 extern void cpp_define_unused (cpp_reader
*, const char *);
1229 extern void cpp_define_formatted (cpp_reader
*pfile
,
1230 const char *fmt
, ...) ATTRIBUTE_PRINTF_2
;
1231 extern void cpp_define_formatted_unused (cpp_reader
*pfile
,
1233 ...) ATTRIBUTE_PRINTF_2
;
1234 extern void cpp_assert (cpp_reader
*, const char *);
1235 extern void cpp_undef (cpp_reader
*, const char *);
1236 extern void cpp_unassert (cpp_reader
*, const char *);
1238 /* Mark a node as a lazily defined macro. */
1239 extern void cpp_define_lazily (cpp_reader
*, cpp_hashnode
*node
, unsigned N
);
1241 /* Undefine all macros and assertions. */
1242 extern void cpp_undef_all (cpp_reader
*);
1244 extern cpp_buffer
*cpp_push_buffer (cpp_reader
*, const unsigned char *,
1246 extern int cpp_defined (cpp_reader
*, const unsigned char *, int);
1248 /* A preprocessing number. Code assumes that any unused high bits of
1249 the double integer are set to zero. */
1251 /* This type has to be equal to unsigned HOST_WIDE_INT, see
1252 gcc/c-family/c-lex.cc. */
1253 typedef uint64_t cpp_num_part
;
1254 typedef struct cpp_num cpp_num
;
1259 bool unsignedp
; /* True if value should be treated as unsigned. */
1260 bool overflow
; /* True if the most recent calculation overflowed. */
1263 /* cpplib provides two interfaces for interpretation of preprocessing
1266 cpp_classify_number categorizes numeric constants according to
1267 their field (integer, floating point, or invalid), radix (decimal,
1268 octal, hexadecimal), and type suffixes. */
1270 #define CPP_N_CATEGORY 0x000F
1271 #define CPP_N_INVALID 0x0000
1272 #define CPP_N_INTEGER 0x0001
1273 #define CPP_N_FLOATING 0x0002
1275 #define CPP_N_WIDTH 0x00F0
1276 #define CPP_N_SMALL 0x0010 /* int, float, short _Fract/Accum */
1277 #define CPP_N_MEDIUM 0x0020 /* long, double, long _Fract/_Accum. */
1278 #define CPP_N_LARGE 0x0040 /* long long, long double,
1279 long long _Fract/Accum. */
1281 #define CPP_N_WIDTH_MD 0xF0000 /* machine defined. */
1282 #define CPP_N_MD_W 0x10000
1283 #define CPP_N_MD_Q 0x20000
1285 #define CPP_N_RADIX 0x0F00
1286 #define CPP_N_DECIMAL 0x0100
1287 #define CPP_N_HEX 0x0200
1288 #define CPP_N_OCTAL 0x0400
1289 #define CPP_N_BINARY 0x0800
1291 #define CPP_N_UNSIGNED 0x1000 /* Properties. */
1292 #define CPP_N_IMAGINARY 0x2000
1293 #define CPP_N_DFLOAT 0x4000
1294 #define CPP_N_DEFAULT 0x8000
1296 #define CPP_N_FRACT 0x100000 /* Fract types. */
1297 #define CPP_N_ACCUM 0x200000 /* Accum types. */
1298 #define CPP_N_FLOATN 0x400000 /* _FloatN types. */
1299 #define CPP_N_FLOATNX 0x800000 /* _FloatNx types. */
1301 #define CPP_N_USERDEF 0x1000000 /* C++11 user-defined literal. */
1303 #define CPP_N_SIZE_T 0x2000000 /* C++23 size_t literal. */
1304 #define CPP_N_BFLOAT16 0x4000000 /* std::bfloat16_t type. */
1305 #define CPP_N_BITINT 0x8000000 /* C2X _BitInt literal. */
1307 #define CPP_N_WIDTH_FLOATN_NX 0xF0000000 /* _FloatN / _FloatNx value
1308 of N, divided by 16. */
1309 #define CPP_FLOATN_SHIFT 24
1310 #define CPP_FLOATN_MAX 0xF0
1312 /* Classify a CPP_NUMBER token. The return value is a combination of
1313 the flags from the above sets. */
1314 extern unsigned cpp_classify_number (cpp_reader
*, const cpp_token
*,
1315 const char **, location_t
);
1317 /* Return the classification flags for a float suffix. */
1318 extern unsigned int cpp_interpret_float_suffix (cpp_reader
*, const char *,
1321 /* Return the classification flags for an int suffix. */
1322 extern unsigned int cpp_interpret_int_suffix (cpp_reader
*, const char *,
1325 /* Evaluate a token classified as category CPP_N_INTEGER. */
1326 extern cpp_num
cpp_interpret_integer (cpp_reader
*, const cpp_token
*,
1329 /* Sign extend a number, with PRECISION significant bits and all
1330 others assumed clear, to fill out a cpp_num structure. */
1331 cpp_num
cpp_num_sign_extend (cpp_num
, size_t);
1333 /* Output a diagnostic of some kind. */
1334 extern bool cpp_error (cpp_reader
*, enum cpp_diagnostic_level
,
1335 const char *msgid
, ...)
1337 extern bool cpp_warning (cpp_reader
*, enum cpp_warning_reason
,
1338 const char *msgid
, ...)
1340 extern bool cpp_pedwarning (cpp_reader
*, enum cpp_warning_reason
,
1341 const char *msgid
, ...)
1343 extern bool cpp_warning_syshdr (cpp_reader
*, enum cpp_warning_reason reason
,
1344 const char *msgid
, ...)
1347 /* As their counterparts above, but use RICHLOC. */
1348 extern bool cpp_warning_at (cpp_reader
*, enum cpp_warning_reason
,
1349 rich_location
*richloc
, const char *msgid
, ...)
1351 extern bool cpp_pedwarning_at (cpp_reader
*, enum cpp_warning_reason
,
1352 rich_location
*richloc
, const char *msgid
, ...)
1355 /* Output a diagnostic with "MSGID: " preceding the
1356 error string of errno. No location is printed. */
1357 extern bool cpp_errno (cpp_reader
*, enum cpp_diagnostic_level
,
1359 /* Similarly, but with "FILENAME: " instead of "MSGID: ", where
1360 the filename is not localized. */
1361 extern bool cpp_errno_filename (cpp_reader
*, enum cpp_diagnostic_level
,
1362 const char *filename
, location_t loc
);
1364 /* Same as cpp_error, except additionally specifies a position as a
1365 (translation unit) physical line and physical column. If the line is
1366 zero, then no location is printed. */
1367 extern bool cpp_error_with_line (cpp_reader
*, enum cpp_diagnostic_level
,
1368 location_t
, unsigned,
1369 const char *msgid
, ...)
1371 extern bool cpp_warning_with_line (cpp_reader
*, enum cpp_warning_reason
,
1372 location_t
, unsigned,
1373 const char *msgid
, ...)
1375 extern bool cpp_pedwarning_with_line (cpp_reader
*, enum cpp_warning_reason
,
1376 location_t
, unsigned,
1377 const char *msgid
, ...)
1379 extern bool cpp_warning_with_line_syshdr (cpp_reader
*, enum cpp_warning_reason
,
1380 location_t
, unsigned,
1381 const char *msgid
, ...)
1384 extern bool cpp_error_at (cpp_reader
* pfile
, enum cpp_diagnostic_level
,
1385 location_t src_loc
, const char *msgid
, ...)
1388 extern bool cpp_error_at (cpp_reader
* pfile
, enum cpp_diagnostic_level
,
1389 rich_location
*richloc
, const char *msgid
, ...)
1393 extern int cpp_ideq (const cpp_token
*, const char *);
1394 extern void cpp_output_line (cpp_reader
*, FILE *);
1395 extern unsigned char *cpp_output_line_to_string (cpp_reader
*,
1396 const unsigned char *);
1397 extern const unsigned char *cpp_alloc_token_string
1398 (cpp_reader
*, const unsigned char *, unsigned);
1399 extern void cpp_output_token (const cpp_token
*, FILE *);
1400 extern const char *cpp_type2name (enum cpp_ttype
, unsigned char flags
);
1401 /* Returns the value of an escape sequence, truncated to the correct
1402 target precision. PSTR points to the input pointer, which is just
1403 after the backslash. LIMIT is how much text we have. WIDE is true
1404 if the escape sequence is part of a wide character constant or
1405 string literal. Handles all relevant diagnostics. */
1406 extern cppchar_t
cpp_parse_escape (cpp_reader
*, const unsigned char ** pstr
,
1407 const unsigned char *limit
, int wide
);
1409 /* Structure used to hold a comment block at a given location in the
1414 /* Text of the comment including the terminators. */
1417 /* source location for the given comment. */
1421 /* Structure holding all comments for a given cpp_reader. */
1425 /* table of comment entries. */
1426 cpp_comment
*entries
;
1428 /* number of actual entries entered in the table. */
1431 /* number of entries allocated currently. */
1433 } cpp_comment_table
;
1435 /* Returns the table of comments encountered by the preprocessor. This
1436 table is only populated when pfile->state.save_comments is true. */
1437 extern cpp_comment_table
*cpp_get_comments (cpp_reader
*);
1441 /* Lookup an identifier in the hashtable. Puts the identifier in the
1442 table if it is not already there. */
1443 extern cpp_hashnode
*cpp_lookup (cpp_reader
*, const unsigned char *,
1446 typedef int (*cpp_cb
) (cpp_reader
*, cpp_hashnode
*, void *);
1447 extern void cpp_forall_identifiers (cpp_reader
*, cpp_cb
, void *);
1450 extern void cpp_scan_nooutput (cpp_reader
*);
1451 extern int cpp_sys_macro_p (cpp_reader
*);
1452 extern unsigned char *cpp_quote_string (unsigned char *, const unsigned char *,
1454 extern bool cpp_compare_macros (const cpp_macro
*macro1
,
1455 const cpp_macro
*macro2
);
1458 extern bool cpp_included (cpp_reader
*, const char *);
1459 extern bool cpp_included_before (cpp_reader
*, const char *, location_t
);
1460 extern void cpp_make_system_header (cpp_reader
*, int, int);
1461 extern bool cpp_push_include (cpp_reader
*, const char *);
1462 extern bool cpp_push_default_include (cpp_reader
*, const char *);
1463 extern void cpp_change_file (cpp_reader
*, enum lc_reason
, const char *);
1464 extern const char *cpp_get_path (struct _cpp_file
*);
1465 extern cpp_dir
*cpp_get_dir (struct _cpp_file
*);
1466 extern cpp_buffer
*cpp_get_buffer (cpp_reader
*);
1467 extern struct _cpp_file
*cpp_get_file (cpp_buffer
*);
1468 extern cpp_buffer
*cpp_get_prev (cpp_buffer
*);
1469 extern void cpp_clear_file_cache (cpp_reader
*);
1471 /* cpp_get_converted_source returns the contents of the given file, as it exists
1472 after cpplib has read it and converted it from the input charset to the
1473 source charset. Return struct will be zero-filled if the data could not be
1474 read for any reason. The data starts at the DATA pointer, but the TO_FREE
1475 pointer is what should be passed to free(), as there may be an offset. */
1476 struct cpp_converted_source
1482 cpp_converted_source
cpp_get_converted_source (const char *fname
,
1483 const char *input_charset
);
1486 struct save_macro_data
;
1487 extern int cpp_save_state (cpp_reader
*, FILE *);
1488 extern int cpp_write_pch_deps (cpp_reader
*, FILE *);
1489 extern int cpp_write_pch_state (cpp_reader
*, FILE *);
1490 extern int cpp_valid_state (cpp_reader
*, const char *, int);
1491 extern void cpp_prepare_state (cpp_reader
*, struct save_macro_data
**);
1492 extern int cpp_read_state (cpp_reader
*, const char *, FILE *,
1493 struct save_macro_data
*);
1496 extern void cpp_force_token_locations (cpp_reader
*, location_t
);
1497 extern void cpp_stop_forcing_token_locations (cpp_reader
*);
1505 extern void cpp_directive_only_process (cpp_reader
*pfile
,
1507 void (*cb
) (cpp_reader
*,
1512 extern enum cpp_ttype cpp_userdef_string_remove_type
1513 (enum cpp_ttype type
);
1514 extern enum cpp_ttype cpp_userdef_string_add_type
1515 (enum cpp_ttype type
);
1516 extern enum cpp_ttype cpp_userdef_char_remove_type
1517 (enum cpp_ttype type
);
1518 extern enum cpp_ttype cpp_userdef_char_add_type
1519 (enum cpp_ttype type
);
1520 extern bool cpp_userdef_string_p
1521 (enum cpp_ttype type
);
1522 extern bool cpp_userdef_char_p
1523 (enum cpp_ttype type
);
1524 extern const char * cpp_get_userdef_suffix
1525 (const cpp_token
*);
1529 /* The result of attempting to decode a run of UTF-8 bytes. */
1531 struct cpp_decoded_char
1533 const char *m_start_byte
;
1534 const char *m_next_byte
;
1540 /* Information for mapping between code points and display columns.
1542 This is a tabstop value, along with a callback for getting the
1543 widths of characters. Normally this callback is cpp_wcwidth, but we
1544 support other schemes for escaping non-ASCII unicode as a series of
1545 ASCII chars when printing the user's source code in diagnostic-show-locus.cc
1547 For example, consider:
1548 - the Unicode character U+03C0 "GREEK SMALL LETTER PI" (UTF-8: 0xCF 0x80)
1549 - the Unicode character U+1F642 "SLIGHTLY SMILING FACE"
1550 (UTF-8: 0xF0 0x9F 0x99 0x82)
1551 - the byte 0xBF (a stray trailing byte of a UTF-8 character)
1552 Normally U+03C0 would occupy one display column, U+1F642
1553 would occupy two display columns, and the stray byte would be
1554 printed verbatim as one display column.
1556 However when escaping them as unicode code points as "<U+03C0>"
1557 and "<U+1F642>" they occupy 8 and 9 display columns respectively,
1558 and when escaping them as bytes as "<CF><80>" and "<F0><9F><99><82>"
1559 they occupy 8 and 16 display columns respectively. In both cases
1560 the stray byte is escaped to <BF> as 4 display columns. */
1562 struct cpp_char_column_policy
1564 cpp_char_column_policy (int tabstop
,
1565 int (*width_cb
) (cppchar_t c
))
1566 : m_tabstop (tabstop
),
1567 m_undecoded_byte_width (1),
1568 m_width_cb (width_cb
)
1572 /* Width in display columns of a stray byte that isn't decodable
1574 int m_undecoded_byte_width
;
1575 int (*m_width_cb
) (cppchar_t c
);
1578 /* A class to manage the state while converting a UTF-8 sequence to cppchar_t
1579 and computing the display width one character at a time. */
1580 class cpp_display_width_computation
{
1582 cpp_display_width_computation (const char *data
, int data_length
,
1583 const cpp_char_column_policy
&policy
);
1584 const char *next_byte () const { return m_next
; }
1585 int bytes_processed () const { return m_next
- m_begin
; }
1586 int bytes_left () const { return m_bytes_left
; }
1587 bool done () const { return !bytes_left (); }
1588 int display_cols_processed () const { return m_display_cols
; }
1590 int process_next_codepoint (cpp_decoded_char
*out
);
1591 int advance_display_cols (int n
);
1594 const char *const m_begin
;
1596 size_t m_bytes_left
;
1597 const cpp_char_column_policy
&m_policy
;
1601 /* Convenience functions that are simple use cases for class
1602 cpp_display_width_computation. Tab characters will be expanded to spaces
1603 as determined by POLICY.m_tabstop, and non-printable-ASCII characters
1604 will be escaped as per POLICY. */
1606 int cpp_byte_column_to_display_column (const char *data
, int data_length
,
1608 const cpp_char_column_policy
&policy
);
1609 inline int cpp_display_width (const char *data
, int data_length
,
1610 const cpp_char_column_policy
&policy
)
1612 return cpp_byte_column_to_display_column (data
, data_length
, data_length
,
1615 int cpp_display_column_to_byte_column (const char *data
, int data_length
,
1617 const cpp_char_column_policy
&policy
);
1618 int cpp_wcwidth (cppchar_t c
);
1620 bool cpp_input_conversion_is_trivial (const char *input_charset
);
1621 int cpp_check_utf8_bom (const char *data
, size_t data_length
);
1622 bool cpp_valid_utf8_p (const char *data
, size_t num_bytes
);
1624 bool cpp_is_combining_char (cppchar_t c
);
1625 bool cpp_is_printable_char (cppchar_t c
);
1627 #endif /* ! LIBCPP_CPPLIB_H */