Line data Source code
1 : /* Definitions for CPP library.
2 : Copyright (C) 1995-2026 Free Software Foundation, Inc.
3 : Written by Per Bothner, 1994-95.
4 :
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
8 : later version.
9 :
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.
14 :
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/>.
18 :
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
24 :
25 : #include <sys/types.h>
26 : #include "symtab.h"
27 : #include "line-map.h"
28 :
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;
38 :
39 : struct _cpp_file;
40 :
41 : class rich_location;
42 :
43 : /* The first three groups, apart from '=', can appear in preprocessor
44 : expressions (+= and -= are used to indicate unary + and - resp.).
45 : This allows a lookup table to be implemented in _cpp_parse_expr.
46 :
47 : The first group, to CPP_LAST_EQ, can be immediately followed by an
48 : '='. The lexer needs operators ending in '=', like ">>=", to be in
49 : the same order as their counterparts without the '=', like ">>".
50 :
51 : See the cpp_operator table optab in expr.cc if you change the order or
52 : add or remove anything in the first group. */
53 :
54 : #define TTYPE_TABLE \
55 : OP(EQ, "=") \
56 : OP(NOT, "!") \
57 : OP(GREATER, ">") /* compare */ \
58 : OP(LESS, "<") \
59 : OP(PLUS, "+") /* math */ \
60 : OP(MINUS, "-") \
61 : OP(MULT, "*") \
62 : OP(DIV, "/") \
63 : OP(MOD, "%") \
64 : OP(AND, "&") /* bit ops */ \
65 : OP(OR, "|") \
66 : OP(XOR, "^") \
67 : OP(RSHIFT, ">>") \
68 : OP(LSHIFT, "<<") \
69 : \
70 : OP(COMPL, "~") \
71 : OP(AND_AND, "&&") /* logical */ \
72 : OP(OR_OR, "||") \
73 : OP(QUERY, "?") \
74 : OP(COLON, ":") \
75 : OP(COMMA, ",") /* grouping */ \
76 : OP(OPEN_PAREN, "(") \
77 : OP(CLOSE_PAREN, ")") \
78 : TK(EOF, NONE) \
79 : OP(EQ_EQ, "==") /* compare */ \
80 : OP(NOT_EQ, "!=") \
81 : OP(GREATER_EQ, ">=") \
82 : OP(LESS_EQ, "<=") \
83 : OP(SPACESHIP, "<=>") \
84 : \
85 : /* These two are unary + / - in preprocessor expressions. */ \
86 : OP(PLUS_EQ, "+=") /* math */ \
87 : OP(MINUS_EQ, "-=") \
88 : \
89 : OP(MULT_EQ, "*=") \
90 : OP(DIV_EQ, "/=") \
91 : OP(MOD_EQ, "%=") \
92 : OP(AND_EQ, "&=") /* bit ops */ \
93 : OP(OR_EQ, "|=") \
94 : OP(XOR_EQ, "^=") \
95 : OP(RSHIFT_EQ, ">>=") \
96 : OP(LSHIFT_EQ, "<<=") \
97 : /* Digraphs together, beginning with CPP_FIRST_DIGRAPH. */ \
98 : OP(HASH, "#") /* digraphs */ \
99 : OP(PASTE, "##") \
100 : OP(OPEN_SQUARE, "[") \
101 : OP(CLOSE_SQUARE, "]") \
102 : OP(OPEN_BRACE, "{") \
103 : OP(CLOSE_BRACE, "}") \
104 : OP(OPEN_SPLICE, "[:") \
105 : OP(CLOSE_SPLICE, ":]") \
106 : /* The remainder of the punctuation. Order is not significant. */ \
107 : OP(SEMICOLON, ";") /* structure */ \
108 : OP(ELLIPSIS, "...") \
109 : OP(PLUS_PLUS, "++") /* increment */ \
110 : OP(MINUS_MINUS, "--") \
111 : OP(DEREF, "->") /* accessors */ \
112 : OP(DOT, ".") \
113 : OP(SCOPE, "::") \
114 : OP(DEREF_STAR, "->*") \
115 : OP(DOT_STAR, ".*") \
116 : OP(REFLECT_OP, "^^") \
117 : OP(ATSIGN, "@") /* used in Objective-C */ \
118 : \
119 : TK(NAME, IDENT) /* word */ \
120 : TK(AT_NAME, IDENT) /* @word - Objective-C */ \
121 : TK(NUMBER, LITERAL) /* 34_be+ta */ \
122 : \
123 : TK(CHAR, LITERAL) /* 'char' */ \
124 : TK(WCHAR, LITERAL) /* L'char' */ \
125 : TK(CHAR16, LITERAL) /* u'char' */ \
126 : TK(CHAR32, LITERAL) /* U'char' */ \
127 : TK(UTF8CHAR, LITERAL) /* u8'char' */ \
128 : TK(OTHER, LITERAL) /* stray punctuation */ \
129 : \
130 : TK(STRING, LITERAL) /* "string" */ \
131 : TK(WSTRING, LITERAL) /* L"string" */ \
132 : TK(STRING16, LITERAL) /* u"string" */ \
133 : TK(STRING32, LITERAL) /* U"string" */ \
134 : TK(UTF8STRING, LITERAL) /* u8"string" */ \
135 : TK(OBJC_STRING, LITERAL) /* @"string" - Objective-C */ \
136 : TK(HEADER_NAME, LITERAL) /* <stdio.h> in #include */ \
137 : TK(UNEVAL_STRING, LITERAL) /* unevaluated "string" - C++26 */ \
138 : \
139 : TK(CHAR_USERDEF, LITERAL) /* 'char'_suffix - C++11 */ \
140 : TK(WCHAR_USERDEF, LITERAL) /* L'char'_suffix - C++11 */ \
141 : TK(CHAR16_USERDEF, LITERAL) /* u'char'_suffix - C++11 */ \
142 : TK(CHAR32_USERDEF, LITERAL) /* U'char'_suffix - C++11 */ \
143 : TK(UTF8CHAR_USERDEF, LITERAL) /* u8'char'_suffix - C++11 */ \
144 : TK(STRING_USERDEF, LITERAL) /* "string"_suffix - C++11 */ \
145 : TK(WSTRING_USERDEF, LITERAL) /* L"string"_suffix - C++11 */ \
146 : TK(STRING16_USERDEF, LITERAL) /* u"string"_suffix - C++11 */ \
147 : TK(STRING32_USERDEF, LITERAL) /* U"string"_suffix - C++11 */ \
148 : TK(UTF8STRING_USERDEF,LITERAL) /* u8"string"_suffix - C++11 */ \
149 : \
150 : TK(EMBED, LITERAL) /* #embed - C23 */ \
151 : \
152 : TK(COMMENT, LITERAL) /* Only if output comments. */ \
153 : /* SPELL_LITERAL happens to DTRT. */ \
154 : TK(MACRO_ARG, NONE) /* Macro argument. */ \
155 : TK(PRAGMA, NONE) /* Only for deferred pragmas. */ \
156 : TK(PRAGMA_EOL, NONE) /* End-of-line for deferred pragmas. */ \
157 : TK(PADDING, NONE) /* Whitespace for -E. */
158 :
159 : #define OP(e, s) CPP_ ## e,
160 : #define TK(e, s) CPP_ ## e,
161 : enum cpp_ttype
162 : {
163 : TTYPE_TABLE
164 : N_TTYPES,
165 :
166 : /* A token type for keywords, as opposed to ordinary identifiers. */
167 : CPP_KEYWORD,
168 :
169 : /* Positions in the table. */
170 : CPP_LAST_EQ = CPP_LSHIFT,
171 : CPP_FIRST_DIGRAPH = CPP_HASH,
172 : CPP_LAST_PUNCTUATOR= CPP_ATSIGN,
173 : CPP_LAST_CPP_OP = CPP_LESS_EQ
174 : };
175 : #undef OP
176 : #undef TK
177 :
178 : /* C language kind, used when calling cpp_create_reader. */
179 : enum c_lang {CLK_GNUC89 = 0, CLK_GNUC99, CLK_GNUC11, CLK_GNUC17, CLK_GNUC23,
180 : CLK_GNUC2Y,
181 : CLK_STDC89, CLK_STDC94, CLK_STDC99, CLK_STDC11, CLK_STDC17,
182 : CLK_STDC23, CLK_STDC2Y,
183 : CLK_GNUCXX, CLK_CXX98, CLK_GNUCXX11, CLK_CXX11,
184 : CLK_GNUCXX14, CLK_CXX14, CLK_GNUCXX17, CLK_CXX17,
185 : CLK_GNUCXX20, CLK_CXX20, CLK_GNUCXX23, CLK_CXX23,
186 : CLK_GNUCXX26, CLK_CXX26, CLK_GNUCXX29, CLK_CXX29,
187 : CLK_ASM};
188 :
189 : /* Payload of a NUMBER, STRING, CHAR or COMMENT token. */
190 : struct GTY(()) cpp_string {
191 : unsigned int len;
192 :
193 : /* TEXT is always null terminated (terminator not included in len); but this
194 : GTY markup arranges that PCH streaming works properly even if there is a
195 : null byte in the middle of the string. */
196 : const unsigned char * GTY((string_length ("1 + %h.len"))) text;
197 : };
198 :
199 : /* Flags for the cpp_token structure. */
200 : #define PREV_WHITE (1 << 0) /* If whitespace before this token. */
201 : #define DIGRAPH (1 << 1) /* If it was a digraph. */
202 : #define STRINGIFY_ARG (1 << 2) /* If macro argument to be stringified. */
203 : #define PASTE_LEFT (1 << 3) /* If on LHS of a ## operator. */
204 : #define NAMED_OP (1 << 4) /* C++ named operators. */
205 : #define PREV_FALLTHROUGH (1 << 5) /* On a token preceded by FALLTHROUGH
206 : comment. */
207 : #define DECIMAL_INT (1 << 6) /* Decimal integer, set in c-lex.cc. */
208 : #define PURE_ZERO (1 << 7) /* Single 0 digit, used by the C++ frontend,
209 : set in c-lex.cc. */
210 : #define COLON_SCOPE PURE_ZERO /* Adjacent colons in C < 23. */
211 : #define NO_DOT_COLON PURE_ZERO /* Set on CPP_NAME tokens whose expansion
212 : shouldn't start with CPP_DOT or CPP_COLON
213 : after optional CPP_PADDING. */
214 : #define SP_DIGRAPH (1 << 8) /* # or ## token was a digraph. */
215 : #define SP_PREV_WHITE (1 << 9) /* If whitespace before a ##
216 : operator, or before this token
217 : after a # operator. */
218 : #define NO_EXPAND (1 << 10) /* Do not macro-expand this token. */
219 : #define PRAGMA_OP (1 << 11) /* _Pragma token. */
220 : #define BOL (1 << 12) /* Token at beginning of line. */
221 :
222 : /* Specify which field, if any, of the cpp_token union is used. */
223 :
224 : enum cpp_token_fld_kind {
225 : CPP_TOKEN_FLD_NODE,
226 : CPP_TOKEN_FLD_SOURCE,
227 : CPP_TOKEN_FLD_STR,
228 : CPP_TOKEN_FLD_ARG_NO,
229 : CPP_TOKEN_FLD_TOKEN_NO,
230 : CPP_TOKEN_FLD_PRAGMA,
231 : CPP_TOKEN_FLD_NONE
232 : };
233 :
234 : /* A macro argument in the cpp_token union. */
235 : struct GTY(()) cpp_macro_arg {
236 : /* Argument number. */
237 : unsigned int arg_no;
238 : /* The original spelling of the macro argument token. */
239 : cpp_hashnode *
240 : GTY ((nested_ptr (union tree_node,
241 : "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
242 : "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
243 : spelling;
244 : };
245 :
246 : /* An identifier in the cpp_token union. */
247 : struct GTY(()) cpp_identifier {
248 : /* The canonical (UTF-8) spelling of the identifier. */
249 : cpp_hashnode *
250 : GTY ((nested_ptr (union tree_node,
251 : "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
252 : "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
253 : node;
254 : /* The original spelling of the identifier. */
255 : cpp_hashnode *
256 : GTY ((nested_ptr (union tree_node,
257 : "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
258 : "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
259 : spelling;
260 : };
261 :
262 : /* A preprocessing token. This occupies 32 bytes on a 64-bit host. On a
263 : 32-bit host it occupies 20 or 24 bytes, depending whether a uint64_t
264 : requires 4- or 8-byte alignment. */
265 :
266 : struct GTY(()) cpp_token {
267 :
268 : /* Location of first char of token, together with range of full token. */
269 : location_t src_loc;
270 :
271 : ENUM_BITFIELD(cpp_ttype) type : CHAR_BIT; /* token type */
272 : unsigned short flags; /* flags - see above */
273 :
274 : union cpp_token_u
275 : {
276 : /* An identifier. */
277 : struct cpp_identifier GTY ((tag ("CPP_TOKEN_FLD_NODE"))) node;
278 :
279 : /* Inherit padding from this token. */
280 : cpp_token * GTY ((tag ("CPP_TOKEN_FLD_SOURCE"))) source;
281 :
282 : /* A string, or number. */
283 : struct cpp_string GTY ((tag ("CPP_TOKEN_FLD_STR"))) str;
284 :
285 : /* Argument no. (and original spelling) for a CPP_MACRO_ARG. */
286 : struct cpp_macro_arg GTY ((tag ("CPP_TOKEN_FLD_ARG_NO"))) macro_arg;
287 :
288 : /* Original token no. for a CPP_PASTE (from a sequence of
289 : consecutive paste tokens in a macro expansion). */
290 : unsigned int GTY ((tag ("CPP_TOKEN_FLD_TOKEN_NO"))) token_no;
291 :
292 : /* Caller-supplied identifier for a CPP_PRAGMA. */
293 : unsigned int GTY ((tag ("CPP_TOKEN_FLD_PRAGMA"))) pragma;
294 : } GTY ((desc ("cpp_token_val_index (&%1)"))) val;
295 : };
296 :
297 : /* Say which field is in use. */
298 : extern enum cpp_token_fld_kind cpp_token_val_index (const cpp_token *tok);
299 :
300 : /* A type wide enough to hold any multibyte source character.
301 : cpplib's character constant interpreter requires an unsigned type.
302 : Also, a typedef for the signed equivalent.
303 : The width of this type is capped at 32 bits; there do exist targets
304 : where wchar_t is 64 bits, but only in a non-default mode, and there
305 : would be no meaningful interpretation for a wchar_t value greater
306 : than 2^32 anyway -- the widest wide-character encoding around is
307 : ISO 10646, which stops at 2^31. */
308 : #if CHAR_BIT * SIZEOF_INT >= 32
309 : # define CPPCHAR_SIGNED_T int
310 : #elif CHAR_BIT * SIZEOF_LONG >= 32
311 : # define CPPCHAR_SIGNED_T long
312 : #else
313 : # error "Cannot find a least-32-bit signed integer type"
314 : #endif
315 : typedef unsigned CPPCHAR_SIGNED_T cppchar_t;
316 : typedef CPPCHAR_SIGNED_T cppchar_signed_t;
317 :
318 : /* Style of header dependencies to generate. */
319 : enum cpp_deps_style { DEPS_NONE = 0, DEPS_USER, DEPS_SYSTEM };
320 :
321 : /* Structured format of module dependencies to generate. */
322 : enum cpp_fdeps_format { FDEPS_FMT_NONE = 0, FDEPS_FMT_P1689R5 };
323 :
324 : /* The possible normalization levels, from most restrictive to least. */
325 : enum cpp_normalize_level {
326 : /* In NFKC. */
327 : normalized_KC = 0,
328 : /* In NFC. */
329 : normalized_C,
330 : /* In NFC, except for subsequences where being in NFC would make
331 : the identifier invalid. */
332 : normalized_identifier_C,
333 : /* Not normalized at all. */
334 : normalized_none
335 : };
336 :
337 : enum cpp_main_search
338 : {
339 : CMS_none, /* A regular source file. */
340 : CMS_header, /* Is a directly-specified header file (eg PCH or
341 : header-unit). */
342 : CMS_user, /* Search the user INCLUDE path. */
343 : CMS_system, /* Search the system INCLUDE path. */
344 : };
345 :
346 : /* The possible bidirectional control characters checking levels. */
347 : enum cpp_bidirectional_level {
348 : /* No checking. */
349 : bidirectional_none = 0,
350 : /* Only detect unpaired uses of bidirectional control characters. */
351 : bidirectional_unpaired = 1,
352 : /* Detect any use of bidirectional control characters. */
353 : bidirectional_any = 2,
354 : /* Also warn about UCNs. */
355 : bidirectional_ucn = 4
356 : };
357 :
358 : /* This structure is nested inside struct cpp_reader, and
359 : carries all the options visible to the command line. */
360 : struct cpp_options
361 : {
362 : /* The language we're preprocessing. */
363 : enum c_lang lang;
364 :
365 : /* Nonzero means use extra default include directories for C++. */
366 : unsigned char cplusplus;
367 :
368 : /* Nonzero means handle cplusplus style comments. */
369 : unsigned char cplusplus_comments;
370 :
371 : /* Nonzero means define __OBJC__, treat @ as a special token, use
372 : the OBJC[PLUS]_INCLUDE_PATH environment variable, and allow
373 : "#import". */
374 : unsigned char objc;
375 :
376 : /* Nonzero means don't copy comments into the output file. */
377 : unsigned char discard_comments;
378 :
379 : /* Nonzero means don't copy comments into the output file during
380 : macro expansion. */
381 : unsigned char discard_comments_in_macro_exp;
382 :
383 : /* Nonzero means process the ISO trigraph sequences. */
384 : unsigned char trigraphs;
385 :
386 : /* Nonzero means process the ISO digraph sequences. */
387 : unsigned char digraphs;
388 :
389 : /* Nonzero means to allow hexadecimal floats and LL suffixes. */
390 : unsigned char extended_numbers;
391 :
392 : /* Nonzero means process u/U prefix literals (UTF-16/32). */
393 : unsigned char uliterals;
394 :
395 : /* Nonzero means process u8 prefixed character literals (UTF-8). */
396 : unsigned char utf8_char_literals;
397 :
398 : /* Nonzero means process r/R raw strings. If this is set, uliterals
399 : must be set as well. */
400 : unsigned char rliterals;
401 :
402 : /* Nonzero means print names of header files (-H). */
403 : unsigned char print_include_names;
404 :
405 : /* Nonzero means complain about deprecated features. */
406 : unsigned char cpp_warn_deprecated;
407 :
408 : /* Nonzero means warn if slash-star appears in a comment. */
409 : unsigned char warn_comments;
410 :
411 : /* Nonzero means to warn about __DATA__, __TIME__ and __TIMESTAMP__ usage. */
412 : unsigned char warn_date_time;
413 :
414 : /* Nonzero means warn if a user-supplied include directory does not
415 : exist. */
416 : unsigned char warn_missing_include_dirs;
417 :
418 : /* Nonzero means warn if there are any trigraphs. */
419 : unsigned char warn_trigraphs;
420 :
421 : /* Nonzero means warn about multicharacter charconsts. */
422 : unsigned char warn_multichar;
423 :
424 : /* Nonzero means warn about various incompatibilities with
425 : traditional C. */
426 : unsigned char cpp_warn_traditional;
427 :
428 : /* Nonzero means warn about long long numeric constants. */
429 : unsigned char cpp_warn_long_long;
430 :
431 : /* Nonzero means warn about text after an #endif (or #else). */
432 : unsigned char warn_endif_labels;
433 :
434 : /* Nonzero means warn about implicit sign changes owing to integer
435 : promotions. */
436 : unsigned char warn_num_sign_change;
437 :
438 : /* Zero means don't warn about __VA_ARGS__ usage in c89 pedantic mode.
439 : Presumably the usage is protected by the appropriate #ifdef. */
440 : unsigned char warn_variadic_macros;
441 :
442 : /* Non-zero means suppress diagnostics for NODE_WARN #define or #undef.
443 : Used for cpp_define/cpp_undef. */
444 : unsigned char suppress_builtin_macro_warnings;
445 :
446 : /* Nonzero means warn about builtin macros that are redefined or
447 : explicitly undefined. */
448 : unsigned char warn_builtin_macro_redefined;
449 :
450 : /* Different -Wimplicit-fallthrough= levels. */
451 : unsigned char cpp_warn_implicit_fallthrough;
452 :
453 : /* Nonzero means warn about a define of a different macro right after
454 : #ifndef/#if !defined header guard directive. */
455 : unsigned char warn_header_guard;
456 :
457 : /* Nonzero means we should look for header.gcc files that remap file
458 : names. */
459 : unsigned char remap;
460 :
461 : /* Zero means dollar signs are punctuation. */
462 : unsigned char dollars_in_ident;
463 :
464 : /* Nonzero means UCNs are accepted in identifiers. */
465 : unsigned char extended_identifiers;
466 :
467 : /* True if we should warn about dollars in identifiers or numbers
468 : for this translation unit. */
469 : unsigned char warn_dollars;
470 :
471 : /* Nonzero means warn if undefined identifiers are evaluated in an #if. */
472 : unsigned char warn_undef;
473 :
474 : /* Nonzero means warn if "defined" is encountered in a place other than
475 : an #if. */
476 : unsigned char warn_expansion_to_defined;
477 :
478 : /* Nonzero means warn of unused macros from the main file. */
479 : unsigned char warn_unused_macros;
480 :
481 : /* Nonzero for the 1999 C Standard, including corrigenda and amendments. */
482 : unsigned char c99;
483 :
484 : /* Nonzero if we are conforming to a specific C or C++ standard. */
485 : unsigned char std;
486 :
487 : /* Nonzero means give all the error messages the ANSI standard requires. */
488 : unsigned char cpp_pedantic;
489 :
490 : /* Nonzero means we're looking at already preprocessed code, so don't
491 : bother trying to do macro expansion and whatnot. */
492 : unsigned char preprocessed;
493 :
494 : /* Nonzero means we are going to emit debugging logs during
495 : preprocessing. */
496 : unsigned char debug;
497 :
498 : /* Nonzero means we are tracking locations of tokens involved in
499 : macro expansion. 1 Means we track the location in degraded mode
500 : where we do not track locations of tokens resulting from the
501 : expansion of arguments of function-like macro. 2 Means we do
502 : track all macro expansions. This last option is the one that
503 : consumes the highest amount of memory. */
504 : unsigned char track_macro_expansion;
505 :
506 : /* Nonzero means handle C++ alternate operator names. */
507 : unsigned char operator_names;
508 :
509 : /* Nonzero means warn about use of C++ alternate operator names. */
510 : unsigned char warn_cxx_operator_names;
511 :
512 : /* True for traditional preprocessing. */
513 : unsigned char traditional;
514 :
515 : /* Nonzero for C++ 2011 Standard user-defined literals. */
516 : unsigned char user_literals;
517 :
518 : /* Nonzero means warn when a string or character literal is followed by a
519 : ud-suffix which does not begin with an underscore. */
520 : unsigned char warn_literal_suffix;
521 :
522 : /* Nonzero means interpret imaginary, fixed-point, or other gnu extension
523 : literal number suffixes as user-defined literal number suffixes. */
524 : unsigned char ext_numeric_literals;
525 :
526 : /* Nonzero means extended identifiers allow the characters specified
527 : in C11. */
528 : unsigned char c11_identifiers;
529 :
530 : /* Nonzero means extended identifiers allow the characters specified
531 : by Unicode XID_Start and XID_Continue properties. */
532 : unsigned char xid_identifiers;
533 :
534 : /* Nonzero for C++ 2014 Standard binary constants. */
535 : unsigned char binary_constants;
536 :
537 : /* Nonzero for C2Y imaginary (floating) constants. */
538 : unsigned char imaginary_constants;
539 :
540 : /* Nonzero for C++ 2014 Standard digit separators. */
541 : unsigned char digit_separators;
542 :
543 : /* Nonzero for C23 decimal floating-point constants. */
544 : unsigned char dfp_constants;
545 :
546 : /* Nonzero for C++20 __VA_OPT__ feature. */
547 : unsigned char va_opt;
548 :
549 : /* Nonzero for the '::' token. */
550 : unsigned char scope;
551 :
552 : /* Nonzero for the '#elifdef' and '#elifndef' directives. */
553 : unsigned char elifdef;
554 :
555 : /* Nonzero for the '#warning' directive. */
556 : unsigned char warning_directive;
557 :
558 : /* Nonzero means tokenize C++20 module directives. */
559 : unsigned char module_directives;
560 :
561 : /* Nonzero for C++23 size_t literals. */
562 : unsigned char size_t_literals;
563 :
564 : /* Nonzero for C++23 delimited escape sequences. */
565 : unsigned char delimited_escape_seqs;
566 :
567 : /* Nonzero for C++23 named universal character escape sequences. */
568 : unsigned char named_uc_escape_seqs;
569 :
570 : /* Nonzero for C++ and C23 UCNs for characters below 0xa0. */
571 : unsigned char low_ucns;
572 :
573 : /* Nonzero for C2Y 0o prefixed octal integer constants. */
574 : unsigned char octal_constants;
575 :
576 : /* Nonzero for 'true' and 'false' in #if expressions. */
577 : unsigned char true_false;
578 :
579 : /* Nonzero for the '#embed' directive. */
580 : unsigned char embed;
581 :
582 : /* Holds the name of the target (execution) character set. */
583 : const char *narrow_charset;
584 :
585 : /* Holds the name of the target wide character set. */
586 : const char *wide_charset;
587 :
588 : /* Holds the name of the input character set. */
589 : const char *input_charset;
590 :
591 : /* The minimum permitted level of normalization before a warning
592 : is generated. See enum cpp_normalize_level. */
593 : int warn_normalize;
594 :
595 : /* True to warn about precompiled header files we couldn't use. */
596 : bool warn_invalid_pch;
597 :
598 : /* True if dependencies should be restored from a precompiled header. */
599 : bool restore_pch_deps;
600 :
601 : /* True if warn about differences between C90 and C99. */
602 : signed char cpp_warn_c90_c99_compat;
603 :
604 : /* True if warn about differences between C11 and C23. */
605 : signed char cpp_warn_c11_c23_compat;
606 :
607 : /* True if warn about differences between C23 and C2Y. */
608 : signed char cpp_warn_c23_c2y_compat;
609 :
610 : /* True if warn about differences between C++98 and C++11. */
611 : bool cpp_warn_cxx11_compat;
612 :
613 : /* True if warn about differences between C++17 and C++20. */
614 : bool cpp_warn_cxx20_compat;
615 :
616 : /* Nonzero if bidirectional control characters checking is on. See enum
617 : cpp_bidirectional_level. */
618 : unsigned char cpp_warn_bidirectional;
619 :
620 : /* True if libcpp should warn about invalid UTF-8 characters in comments.
621 : 2 if it should be a pedwarn. */
622 : unsigned char cpp_warn_invalid_utf8;
623 :
624 : /* True if libcpp should warn about invalid forms of delimited or named
625 : escape sequences. */
626 : bool cpp_warn_unicode;
627 :
628 : /* True if -finput-charset= option has been used explicitly. */
629 : bool cpp_input_charset_explicit;
630 :
631 : /* True if -Wkeyword-macro. */
632 : bool cpp_warn_keyword_macro;
633 :
634 : /* -Wleading-whitespace= value. */
635 : unsigned char cpp_warn_leading_whitespace;
636 :
637 : /* -Wtrailing-whitespace= value. */
638 : unsigned char cpp_warn_trailing_whitespace;
639 :
640 : /* -ftabstop= value. */
641 : unsigned int cpp_tabstop;
642 :
643 : /* Dependency generation. */
644 : struct
645 : {
646 : /* Style of header dependencies to generate. */
647 : enum cpp_deps_style style;
648 :
649 : /* Structured format of module dependencies to generate. */
650 : enum cpp_fdeps_format fdeps_format;
651 :
652 : /* Assume missing files are generated files. */
653 : bool missing_files;
654 :
655 : /* Generate phony targets for each dependency apart from the first
656 : one. */
657 : bool phony_targets;
658 :
659 : /* Generate dependency info for modules. */
660 : bool modules;
661 :
662 : /* If true, no dependency is generated on the main file. */
663 : bool ignore_main_file;
664 :
665 : /* If true, intend to use the preprocessor output (e.g., for compilation)
666 : in addition to the dependency info. */
667 : bool need_preprocessor_output;
668 : } deps;
669 :
670 : /* Target-specific features set by the front end or client. */
671 :
672 : /* Precision for target CPP arithmetic, target characters, target
673 : ints and target wide characters, respectively. */
674 : size_t precision, char_precision, int_precision, wchar_precision;
675 :
676 : /* True means chars (wide chars, UTF-8 chars) are unsigned. */
677 : bool unsigned_char, unsigned_wchar, unsigned_utf8char;
678 :
679 : /* True if the most significant byte in a word has the lowest
680 : address in memory. */
681 : bool bytes_big_endian;
682 :
683 : /* Nonzero means __STDC__ should have the value 0 in system headers. */
684 : unsigned char stdc_0_in_system_headers;
685 :
686 : /* True disables tokenization outside of preprocessing directives. */
687 : bool directives_only;
688 :
689 : /* True enables canonicalization of system header file paths. */
690 : bool canonical_system_headers;
691 :
692 : /* The maximum depth of the nested #include. */
693 : unsigned int max_include_depth;
694 :
695 : cpp_main_search main_search : 8;
696 : };
697 :
698 : #if GCC_VERSION >= 3005
699 : #define ATTRIBUTE_CPP_PPDIAG(m, n) \
700 : __attribute__ ((__format__ (__gcc_diag__, m , n))) ATTRIBUTE_NONNULL(m)
701 : #else
702 : #define ATTRIBUTE_CPP_PPDIAG(m, n) ATTRIBUTE_NONNULL(m)
703 : #endif
704 :
705 : /* Diagnostic levels. To get a diagnostic without associating a
706 : position in the translation unit with it, use cpp_error_with_line
707 : with a line number of zero. */
708 :
709 : enum cpp_diagnostic_level {
710 : /* Warning, an error with -Werror. */
711 : CPP_DL_WARNING = 0,
712 : /* Same as CPP_DL_WARNING, except it is not suppressed in system headers. */
713 : CPP_DL_WARNING_SYSHDR,
714 : /* Warning, an error with -pedantic-errors or -Werror. */
715 : CPP_DL_PEDWARN,
716 : /* An error. */
717 : CPP_DL_ERROR,
718 : /* An internal consistency check failed. Prints "internal error: ",
719 : otherwise the same as CPP_DL_ERROR. */
720 : CPP_DL_ICE,
721 : /* An informative note following a warning. */
722 : CPP_DL_NOTE,
723 : /* A fatal error. */
724 : CPP_DL_FATAL
725 : };
726 :
727 : /* Warning reason codes. Use a reason code of CPP_W_NONE for unclassified
728 : warnings and diagnostics that are not warnings. */
729 :
730 : enum cpp_warning_reason {
731 : CPP_W_NONE = 0,
732 : CPP_W_DEPRECATED,
733 : CPP_W_COMMENTS,
734 : CPP_W_MISSING_INCLUDE_DIRS,
735 : CPP_W_TRIGRAPHS,
736 : CPP_W_MULTICHAR,
737 : CPP_W_TRADITIONAL,
738 : CPP_W_LONG_LONG,
739 : CPP_W_ENDIF_LABELS,
740 : CPP_W_NUM_SIGN_CHANGE,
741 : CPP_W_VARIADIC_MACROS,
742 : CPP_W_BUILTIN_MACRO_REDEFINED,
743 : CPP_W_DOLLARS,
744 : CPP_W_UNDEF,
745 : CPP_W_UNUSED_MACROS,
746 : CPP_W_CXX_OPERATOR_NAMES,
747 : CPP_W_NORMALIZE,
748 : CPP_W_INVALID_PCH,
749 : CPP_W_WARNING_DIRECTIVE,
750 : CPP_W_LITERAL_SUFFIX,
751 : CPP_W_SIZE_T_LITERALS,
752 : CPP_W_DATE_TIME,
753 : CPP_W_PEDANTIC,
754 : CPP_W_C90_C99_COMPAT,
755 : CPP_W_C11_C23_COMPAT,
756 : CPP_W_C23_C2Y_COMPAT,
757 : CPP_W_CXX11_COMPAT,
758 : CPP_W_CXX20_COMPAT,
759 : CPP_W_CXX14_EXTENSIONS,
760 : CPP_W_CXX17_EXTENSIONS,
761 : CPP_W_CXX20_EXTENSIONS,
762 : CPP_W_CXX23_EXTENSIONS,
763 : CPP_W_CXX26_EXTENSIONS,
764 : CPP_W_CXX29_EXTENSIONS,
765 : CPP_W_EXPANSION_TO_DEFINED,
766 : CPP_W_BIDIRECTIONAL,
767 : CPP_W_INVALID_UTF8,
768 : CPP_W_UNICODE,
769 : CPP_W_HEADER_GUARD,
770 : CPP_W_PRAGMA_ONCE_OUTSIDE_HEADER,
771 : CPP_W_LEADING_WHITESPACE,
772 : CPP_W_TRAILING_WHITESPACE,
773 : CPP_W_KEYWORD_MACRO
774 : };
775 :
776 : /* Callback for header lookup for HEADER, which is the name of a
777 : source file. It is used as a method of last resort to find headers
778 : that are not otherwise found during the normal include processing.
779 : The return value is the malloced name of a header to try and open,
780 : if any, or NULL otherwise. This callback is called only if the
781 : header is otherwise unfound. */
782 : typedef const char *(*missing_header_cb)(cpp_reader *, const char *header, cpp_dir **);
783 :
784 : /* Call backs to cpplib client. */
785 : struct cpp_callbacks
786 : {
787 : /* Called when a new line of preprocessed output is started. */
788 : void (*line_change) (cpp_reader *, const cpp_token *, int);
789 :
790 : /* Called when switching to/from a new file.
791 : The line_map is for the new file. It is NULL if there is no new file.
792 : (In C this happens when done with <built-in>+<command line> and also
793 : when done with a main file.) This can be used for resource cleanup. */
794 : void (*file_change) (cpp_reader *, const line_map_ordinary *);
795 :
796 : void (*dir_change) (cpp_reader *, const char *);
797 : void (*include) (cpp_reader *, location_t, const unsigned char *,
798 : const char *, int, const cpp_token **);
799 : void (*define) (cpp_reader *, location_t, cpp_hashnode *);
800 : void (*undef) (cpp_reader *, location_t, cpp_hashnode *);
801 : void (*ident) (cpp_reader *, location_t, const cpp_string *);
802 : void (*def_pragma) (cpp_reader *, location_t);
803 : int (*valid_pch) (cpp_reader *, const char *, int);
804 : void (*read_pch) (cpp_reader *, const char *, int, const char *);
805 : missing_header_cb missing_header;
806 :
807 : /* Context-sensitive macro support. Returns macro (if any) that should
808 : be expanded. */
809 : cpp_hashnode * (*macro_to_expand) (cpp_reader *, const cpp_token *);
810 :
811 : /* Called to emit a diagnostic. This callback receives the
812 : translated message. */
813 : bool (*diagnostic) (cpp_reader *,
814 : enum cpp_diagnostic_level,
815 : enum cpp_warning_reason,
816 : rich_location *,
817 : const char *, va_list *)
818 : ATTRIBUTE_CPP_PPDIAG (5,0);
819 :
820 : /* Callbacks for when a macro is expanded, or tested (whether
821 : defined or not at the time) in #ifdef, #ifndef or "defined". */
822 : void (*used_define) (cpp_reader *, location_t, cpp_hashnode *);
823 : void (*used_undef) (cpp_reader *, location_t, cpp_hashnode *);
824 : /* Called before #define and #undef or other macro definition
825 : changes are processed. */
826 : void (*before_define) (cpp_reader *);
827 : /* Called whenever a macro is expanded or tested.
828 : Second argument is the location of the start of the current expansion. */
829 : void (*used) (cpp_reader *, location_t, cpp_hashnode *);
830 :
831 : /* Callback to identify whether an attribute exists. */
832 : int (*has_attribute) (cpp_reader *, bool);
833 :
834 : /* Callback to determine whether a built-in function is recognized. */
835 : int (*has_builtin) (cpp_reader *);
836 :
837 : /* Callback to determine whether a feature is available. */
838 : int (*has_feature) (cpp_reader *, bool);
839 :
840 : /* Callback that can change a user lazy into normal macro. */
841 : void (*user_lazy_macro) (cpp_reader *, cpp_macro *, unsigned);
842 :
843 : /* Callback to handle deferred cpp_macros. */
844 : cpp_macro *(*user_deferred_macro) (cpp_reader *, location_t, cpp_hashnode *);
845 :
846 : /* Callback to parse SOURCE_DATE_EPOCH from environment. */
847 : time_t (*get_source_date_epoch) (cpp_reader *);
848 :
849 : /* Callback for providing suggestions for misspelled directives. */
850 : const char *(*get_suggestion) (cpp_reader *, const char *, const char *const *);
851 :
852 : /* Callback for when a comment is encountered, giving the location
853 : of the opening slash, a pointer to the content (which is not
854 : necessarily 0-terminated), and the length of the content.
855 : The content contains the opening slash-star (or slash-slash),
856 : and for C-style comments contains the closing star-slash. For
857 : C++-style comments it does not include the terminating newline. */
858 : void (*comment) (cpp_reader *, location_t, const unsigned char *,
859 : size_t);
860 :
861 : /* Callback for filename remapping in __FILE__ and __BASE_FILE__ macro
862 : expansions. */
863 : const char *(*remap_filename) (const char*);
864 :
865 : /* Maybe translate a #include into something else. Return a
866 : cpp_buffer containing the translation if translating. */
867 : char *(*translate_include) (cpp_reader *, line_maps *, location_t,
868 : _cpp_file *file, bool angle_brackets,
869 : const char **alternate);
870 : };
871 :
872 : #ifdef VMS
873 : #define INO_T_CPP ino_t ino[3]
874 : #elif defined (_AIX) && SIZEOF_INO_T == 4
875 : #define INO_T_CPP ino64_t ino
876 : #else
877 : #define INO_T_CPP ino_t ino
878 : #endif
879 :
880 : #if defined (_AIX) && SIZEOF_DEV_T == 4
881 : #define DEV_T_CPP dev64_t dev
882 : #else
883 : #define DEV_T_CPP dev_t dev
884 : #endif
885 :
886 : /* Chain of directories to look for include files in. */
887 : struct cpp_dir
888 : {
889 : /* NULL-terminated singly-linked list. */
890 : struct cpp_dir *next;
891 :
892 : /* NAME of the directory, NUL-terminated. */
893 : char *name;
894 : unsigned int len;
895 :
896 : /* One if a system header, two if a system header that has extern
897 : "C" guards for C++. */
898 : unsigned char sysp;
899 :
900 : /* Is this a user-supplied directory? */
901 : bool user_supplied_p;
902 :
903 : /* The canonicalized NAME as determined by lrealpath. This field
904 : is only used by hosts that lack reliable inode numbers. */
905 : char *canonical_name;
906 :
907 : /* Mapping of file names for this directory for MS-DOS and related
908 : platforms. A NULL-terminated array of (from, to) pairs. */
909 : const char **name_map;
910 :
911 : /* Routine to construct pathname, given the search path name and the
912 : HEADER we are trying to find, return a constructed pathname to
913 : try and open. If this is NULL, the constructed pathname is as
914 : constructed by append_file_to_dir. */
915 : char *(*construct) (const char *header, cpp_dir *dir);
916 :
917 : /* The C front end uses these to recognize duplicated
918 : directories in the search path. */
919 : INO_T_CPP;
920 : DEV_T_CPP;
921 : };
922 :
923 : /* The kind of the cpp_macro. */
924 : enum cpp_macro_kind {
925 : cmk_macro, /* An ISO macro (token expansion). */
926 : cmk_assert, /* An assertion. */
927 : cmk_traditional /* A traditional macro (text expansion). */
928 : };
929 :
930 : /* Each macro definition is recorded in a cpp_macro structure.
931 : Variadic macros cannot occur with traditional cpp. */
932 : struct GTY(()) cpp_macro {
933 : union cpp_parm_u
934 : {
935 : /* Parameters, if any. If parameter names use extended identifiers,
936 : the original spelling of those identifiers, not the canonical
937 : UTF-8 spelling, goes here. */
938 : cpp_hashnode ** GTY ((tag ("false"),
939 : nested_ptr (union tree_node,
940 : "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
941 : "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL"),
942 : length ("%1.paramc"))) params;
943 :
944 : /* If this is an assertion, the next one in the chain. */
945 : cpp_macro *GTY ((tag ("true"))) next;
946 : } GTY ((desc ("%1.kind == cmk_assert"))) parm;
947 :
948 : /* Definition line number. */
949 : location_t line;
950 :
951 : /* Number of tokens in body, or bytes for traditional macros. */
952 : /* Do we really need 2^32-1 range here? */
953 : unsigned int count;
954 :
955 : /* Number of parameters. */
956 : unsigned short paramc;
957 :
958 : /* Non-zero if this is a user-lazy macro, value provided by user. */
959 : unsigned char lazy;
960 :
961 : /* The kind of this macro (ISO, trad or assert) */
962 : unsigned kind : 2;
963 :
964 : /* If a function-like macro. */
965 : unsigned int fun_like : 1;
966 :
967 : /* If a variadic macro. */
968 : unsigned int variadic : 1;
969 :
970 : /* If macro defined in system header. */
971 : unsigned int syshdr : 1;
972 :
973 : /* Nonzero if it has been expanded or had its existence tested. */
974 : unsigned int used : 1;
975 :
976 : /* Indicate whether the tokens include extra CPP_PASTE tokens at the
977 : end to track invalid redefinitions with consecutive CPP_PASTE
978 : tokens. */
979 : unsigned int extra_tokens : 1;
980 :
981 : /* Imported C++20 macro (from a header unit). */
982 : unsigned int imported_p : 1;
983 :
984 : /* 0 bits spare (32-bit). 32 on 64-bit target. */
985 :
986 : union cpp_exp_u
987 : {
988 : /* Trailing array of replacement tokens (ISO), or assertion body value. */
989 : cpp_token GTY ((tag ("false"), length ("%1.count"))) tokens[1];
990 :
991 : /* Pointer to replacement text (traditional). See comment at top
992 : of cpptrad.c for how traditional function-like macros are
993 : encoded. */
994 : const unsigned char *GTY ((tag ("true"))) text;
995 : } GTY ((desc ("%1.kind == cmk_traditional"))) exp;
996 : };
997 :
998 : /* Poisoned identifiers are flagged NODE_POISONED. NODE_OPERATOR (C++
999 : only) indicates an identifier that behaves like an operator such as
1000 : "xor". NODE_DIAGNOSTIC is for speed in lex_token: it indicates a
1001 : diagnostic may be required for this node. Currently this only
1002 : applies to __VA_ARGS__, poisoned identifiers, and -Wc++-compat
1003 : warnings about NODE_OPERATOR. */
1004 :
1005 : /* Hash node flags. */
1006 : #define NODE_OPERATOR (1 << 0) /* C++ named operator. */
1007 : #define NODE_POISONED (1 << 1) /* Poisoned identifier. */
1008 : #define NODE_DIAGNOSTIC (1 << 2) /* Possible diagnostic when lexed. */
1009 : #define NODE_WARN (1 << 3) /* Warn if redefined or undefined. */
1010 : #define NODE_DISABLED (1 << 4) /* A disabled macro. */
1011 : #define NODE_USED (1 << 5) /* Dumped with -dU. */
1012 : #define NODE_CONDITIONAL (1 << 6) /* Conditional macro */
1013 : #define NODE_WARN_OPERATOR (1 << 7) /* Warn about C++ named operator. */
1014 : #define NODE_MODULE (1 << 8) /* C++-20 module-related name. */
1015 :
1016 : /* Different flavors of hash node. */
1017 : enum node_type
1018 : {
1019 : NT_VOID = 0, /* Maybe an assert? */
1020 : NT_MACRO_ARG, /* A macro arg. */
1021 : NT_USER_MACRO, /* A user macro. */
1022 : NT_BUILTIN_MACRO, /* A builtin macro. */
1023 : NT_MACRO_MASK = NT_USER_MACRO /* Mask for either macro kind. */
1024 : };
1025 :
1026 : /* Different flavors of builtin macro. _Pragma is an operator, but we
1027 : handle it with the builtin code for efficiency reasons. */
1028 : enum cpp_builtin_type
1029 : {
1030 : BT_SPECLINE = 0, /* `__LINE__' */
1031 : BT_DATE, /* `__DATE__' */
1032 : BT_FILE, /* `__FILE__' */
1033 : BT_FILE_NAME, /* `__FILE_NAME__' */
1034 : BT_BASE_FILE, /* `__BASE_FILE__' */
1035 : BT_INCLUDE_LEVEL, /* `__INCLUDE_LEVEL__' */
1036 : BT_TIME, /* `__TIME__' */
1037 : BT_STDC, /* `__STDC__' */
1038 : BT_PRAGMA, /* `_Pragma' operator */
1039 : BT_TIMESTAMP, /* `__TIMESTAMP__' */
1040 : BT_COUNTER, /* `__COUNTER__' */
1041 : BT_HAS_ATTRIBUTE, /* `__has_attribute(x)' */
1042 : BT_HAS_STD_ATTRIBUTE, /* `__has_c_attribute(x)' */
1043 : BT_HAS_BUILTIN, /* `__has_builtin(x)' */
1044 : BT_HAS_INCLUDE, /* `__has_include(x)' */
1045 : BT_HAS_INCLUDE_NEXT, /* `__has_include_next(x)' */
1046 : BT_HAS_EMBED, /* `__has_embed(x)' */
1047 : BT_HAS_FEATURE, /* `__has_feature(x)' */
1048 : BT_HAS_EXTENSION /* `__has_extension(x)' */
1049 : };
1050 :
1051 : #define CPP_HASHNODE(HNODE) ((cpp_hashnode *) (HNODE))
1052 : #define HT_NODE(NODE) (&(NODE)->ident)
1053 : #define NODE_LEN(NODE) HT_LEN (HT_NODE (NODE))
1054 : #define NODE_NAME(NODE) HT_STR (HT_NODE (NODE))
1055 :
1056 : /* The common part of an identifier node shared amongst all 3 C front
1057 : ends. Also used to store CPP identifiers, which are a superset of
1058 : identifiers in the grammatical sense. */
1059 :
1060 : union GTY(()) _cpp_hashnode_value {
1061 : /* Assert (maybe NULL) */
1062 : cpp_macro * GTY((tag ("NT_VOID"))) answers;
1063 : /* Macro (maybe NULL) */
1064 : cpp_macro * GTY((tag ("NT_USER_MACRO"))) macro;
1065 : /* Code for a builtin macro. */
1066 : enum cpp_builtin_type GTY ((tag ("NT_BUILTIN_MACRO"))) builtin;
1067 : /* Macro argument index. */
1068 : unsigned short GTY ((tag ("NT_MACRO_ARG"))) arg_index;
1069 : };
1070 :
1071 : struct GTY(()) cpp_hashnode {
1072 : struct ht_identifier ident;
1073 : unsigned int is_directive : 1;
1074 : unsigned int directive_index : 7; /* If is_directive,
1075 : then index into directive table.
1076 : Otherwise, a NODE_OPERATOR. */
1077 : unsigned int rid_code : 8; /* Rid code - for front ends. */
1078 : unsigned int flags : 9; /* CPP flags. */
1079 : ENUM_BITFIELD(node_type) type : 2; /* CPP node type. */
1080 :
1081 : /* 5 bits spare. */
1082 :
1083 : /* The deferred cookie is applicable to NT_USER_MACRO or NT_VOID.
1084 : The latter for when a macro had a prevailing undef.
1085 : On a 64-bit system there would be 32-bits of padding to the value
1086 : field. So placing the deferred index here is not costly. */
1087 : unsigned deferred; /* Deferred cookie */
1088 :
1089 : union _cpp_hashnode_value GTY ((desc ("%1.type"))) value;
1090 : };
1091 :
1092 : /* Extra information we may need to store per identifier, which is needed rarely
1093 : enough that it's not worth adding directly into the main identifier hash. */
1094 : struct GTY(()) cpp_hashnode_extra
1095 : {
1096 : struct ht_identifier ident;
1097 : location_t poisoned_loc;
1098 : };
1099 :
1100 : /* A class for iterating through the source locations within a
1101 : string token (before escapes are interpreted, and before
1102 : concatenation). */
1103 :
1104 : class cpp_string_location_reader {
1105 : public:
1106 : cpp_string_location_reader (location_t src_loc,
1107 : line_maps *line_table);
1108 :
1109 : source_range get_next ();
1110 :
1111 : private:
1112 : location_t m_loc;
1113 : int m_offset_per_column;
1114 : };
1115 :
1116 : /* A class for storing the source ranges of all of the characters within
1117 : a string literal, after escapes are interpreted, and after
1118 : concatenation.
1119 :
1120 : This is not GTY-marked, as instances are intended to be temporary. */
1121 :
1122 : class cpp_substring_ranges
1123 : {
1124 : public:
1125 : cpp_substring_ranges ();
1126 : ~cpp_substring_ranges ();
1127 :
1128 27951 : int get_num_ranges () const { return m_num_ranges; }
1129 43897 : source_range get_range (int idx) const
1130 : {
1131 43897 : linemap_assert (idx < m_num_ranges);
1132 43897 : return m_ranges[idx];
1133 : }
1134 :
1135 : void add_range (source_range range);
1136 : void add_n_ranges (int num, cpp_string_location_reader &loc_reader);
1137 :
1138 : private:
1139 : source_range *m_ranges;
1140 : int m_num_ranges;
1141 : int m_alloc_ranges;
1142 : };
1143 :
1144 : /* Call this first to get a handle to pass to other functions.
1145 :
1146 : The first hash table argument is for associating a struct cpp_hashnode
1147 : with each identifier. The second hash table argument is for associating
1148 : a struct cpp_hashnode_extra with each identifier that needs one. For
1149 : either, pass in a NULL pointer if you want cpplib to create and manage
1150 : the hash table itself, or else pass a suitably initialized hash table to
1151 : be managed external to libcpp, as is done by the C-family frontends. */
1152 : extern cpp_reader *cpp_create_reader (enum c_lang, struct ht *,
1153 : class line_maps *,
1154 : struct ht * = nullptr);
1155 :
1156 : /* Reset the cpp_reader's line_map. This is only used after reading a
1157 : PCH file. */
1158 : extern void cpp_set_line_map (cpp_reader *, class line_maps *);
1159 :
1160 : /* Call this to change the selected language standard (e.g. because of
1161 : command line options). */
1162 : extern void cpp_set_lang (cpp_reader *, enum c_lang);
1163 :
1164 : /* Set the include paths. */
1165 : extern void cpp_set_include_chains (cpp_reader *, cpp_dir *, cpp_dir *,
1166 : cpp_dir *, int);
1167 :
1168 : /* Call these to get pointers to the options, callback, and deps
1169 : structures for a given reader. These pointers are good until you
1170 : call cpp_finish on that reader. You can either edit the callbacks
1171 : through the pointer returned from cpp_get_callbacks, or set them
1172 : with cpp_set_callbacks. */
1173 : extern cpp_options *cpp_get_options (cpp_reader *) ATTRIBUTE_PURE;
1174 : extern cpp_callbacks *cpp_get_callbacks (cpp_reader *) ATTRIBUTE_PURE;
1175 : extern void cpp_set_callbacks (cpp_reader *, cpp_callbacks *);
1176 : extern class mkdeps *cpp_get_deps (cpp_reader *) ATTRIBUTE_PURE;
1177 :
1178 : extern const char *cpp_probe_header_unit (cpp_reader *, const char *file,
1179 : bool angle_p, location_t);
1180 :
1181 : /* Call these to get name data about the various compile-time
1182 : charsets. */
1183 : extern const char *cpp_get_narrow_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1184 : extern const char *cpp_get_wide_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1185 :
1186 : extern location_t cpp_get_diagnostic_override_loc (const cpp_reader *);
1187 :
1188 : /* This function reads the file, but does not start preprocessing. It
1189 : returns the name of the original file; this is the same as the
1190 : input file, except for preprocessed input. This will generate at
1191 : least one file change callback, and possibly a line change callback
1192 : too. If there was an error opening the file, it returns NULL. */
1193 : extern const char *cpp_read_main_file (cpp_reader *, const char *,
1194 : bool injecting = false);
1195 : extern location_t cpp_main_loc (const cpp_reader *);
1196 :
1197 : /* Adjust for the main file to be an include. */
1198 : extern void cpp_retrofit_as_include (cpp_reader *);
1199 :
1200 : /* Set up built-ins with special behavior. Use cpp_init_builtins()
1201 : instead unless your know what you are doing. */
1202 : extern void cpp_init_special_builtins (cpp_reader *);
1203 :
1204 : /* Set up built-ins like __FILE__. */
1205 : extern void cpp_init_builtins (cpp_reader *, int);
1206 :
1207 : /* This is called after options have been parsed, and partially
1208 : processed. */
1209 : extern void cpp_post_options (cpp_reader *);
1210 :
1211 : /* Set up translation to the target character set. */
1212 : extern void cpp_init_iconv (cpp_reader *);
1213 :
1214 : /* Call this to finish preprocessing. If you requested dependency
1215 : generation, pass open stream(s) to write the information to,
1216 : otherwise NULL. It is your responsibility to close the stream(s). */
1217 : extern void cpp_finish (cpp_reader *, FILE *deps_stream, FILE *fdeps_stream = NULL);
1218 :
1219 : /* Call this to release the handle at the end of preprocessing. Any
1220 : use of the handle after this function returns is invalid. */
1221 : extern void cpp_destroy (cpp_reader *);
1222 :
1223 : extern unsigned int cpp_token_len (const cpp_token *);
1224 : extern unsigned char *cpp_token_as_text (cpp_reader *, const cpp_token *);
1225 : extern unsigned char *cpp_spell_token (cpp_reader *, const cpp_token *,
1226 : unsigned char *, bool);
1227 : extern void cpp_register_pragma (cpp_reader *, const char *, const char *,
1228 : void (*) (cpp_reader *), bool);
1229 : extern void cpp_register_deferred_pragma (cpp_reader *, const char *,
1230 : const char *, unsigned, bool, bool);
1231 : extern int cpp_avoid_paste (cpp_reader *, const cpp_token *,
1232 : const cpp_token *);
1233 : extern const cpp_token *cpp_get_token (cpp_reader *);
1234 : extern const cpp_token *cpp_get_token_with_location (cpp_reader *,
1235 : location_t *);
1236 7786454 : inline bool cpp_user_macro_p (const cpp_hashnode *node)
1237 : {
1238 7784078 : return node->type == NT_USER_MACRO;
1239 : }
1240 : inline bool cpp_builtin_macro_p (const cpp_hashnode *node)
1241 : {
1242 : return node->type == NT_BUILTIN_MACRO;
1243 : }
1244 17981474 : inline bool cpp_macro_p (const cpp_hashnode *node)
1245 : {
1246 17981474 : return node->type & NT_MACRO_MASK;
1247 : }
1248 87377 : inline cpp_macro *cpp_set_deferred_macro (cpp_hashnode *node,
1249 : cpp_macro *forced = NULL)
1250 : {
1251 87377 : cpp_macro *old = node->value.macro;
1252 :
1253 87377 : node->value.macro = forced;
1254 87377 : node->type = NT_USER_MACRO;
1255 87377 : node->flags &= ~NODE_USED;
1256 :
1257 87377 : return old;
1258 : }
1259 : cpp_macro *cpp_get_deferred_macro (cpp_reader *, cpp_hashnode *, location_t);
1260 :
1261 : /* Returns true if NODE is a function-like user macro. */
1262 4474 : inline bool cpp_fun_like_macro_p (cpp_hashnode *node)
1263 : {
1264 4474 : return cpp_user_macro_p (node) && node->value.macro->fun_like;
1265 : }
1266 :
1267 : /* Return true for nodes marked for -Wkeyword-macro diagnostics. */
1268 : inline bool cpp_keyword_p (cpp_hashnode *node)
1269 : {
1270 : /* As keywords are marked identifiers which don't start with underscore
1271 : or start with underscore followed by capital letter (except for
1272 : _Pragma). */
1273 : return ((node->flags & NODE_WARN)
1274 : && (NODE_NAME (node)[0] != '_'
1275 : || (NODE_NAME (node)[1] != '_' && NODE_NAME (node)[1] != 'P')));
1276 : }
1277 :
1278 : extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *);
1279 : extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *,
1280 : const cpp_macro *);
1281 76 : inline location_t cpp_macro_definition_location (cpp_hashnode *node)
1282 : {
1283 76 : const cpp_macro *macro = node->value.macro;
1284 76 : return macro ? macro->line : 0;
1285 : }
1286 : /* Return an idempotent time stamp (possibly from SOURCE_DATE_EPOCH). */
1287 : enum class CPP_time_kind
1288 : {
1289 : FIXED = -1, /* Fixed time via source epoch. */
1290 : DYNAMIC = -2, /* Dynamic via time(2). */
1291 : UNKNOWN = -3 /* Wibbly wobbly, timey wimey. */
1292 : };
1293 : extern CPP_time_kind cpp_get_date (cpp_reader *, time_t *);
1294 :
1295 : extern void _cpp_backup_tokens (cpp_reader *, unsigned int);
1296 : extern const cpp_token *cpp_peek_token (cpp_reader *, int);
1297 :
1298 : /* Evaluate a CPP_*CHAR* token. */
1299 : extern cppchar_t cpp_interpret_charconst (cpp_reader *, const cpp_token *,
1300 : unsigned int *, int *);
1301 : /* Evaluate a vector of CPP_*STRING* tokens. */
1302 : extern bool cpp_interpret_string (cpp_reader *,
1303 : const cpp_string *, size_t,
1304 : cpp_string *, enum cpp_ttype);
1305 : extern const char *cpp_interpret_string_ranges (cpp_reader *pfile,
1306 : const cpp_string *from,
1307 : cpp_string_location_reader *,
1308 : size_t count,
1309 : cpp_substring_ranges *out,
1310 : enum cpp_ttype type);
1311 : extern bool cpp_interpret_string_notranslate (cpp_reader *,
1312 : const cpp_string *, size_t,
1313 : cpp_string *, enum cpp_ttype);
1314 : extern bool cpp_translate_string (cpp_reader *, const cpp_string *,
1315 : cpp_string *, enum cpp_ttype, bool);
1316 : extern bool cpp_valid_identifier (cpp_reader *, const unsigned char *);
1317 :
1318 : /* Convert a host character constant to the execution character set. */
1319 : extern cppchar_t cpp_host_to_exec_charset (cpp_reader *, cppchar_t);
1320 :
1321 : /* Used to register macros and assertions, perhaps from the command line.
1322 : The text is the same as the command line argument. */
1323 : extern void cpp_define (cpp_reader *, const char *);
1324 : extern void cpp_define_unused (cpp_reader *, const char *);
1325 : extern void cpp_define_formatted (cpp_reader *pfile,
1326 : const char *fmt, ...) ATTRIBUTE_PRINTF_2;
1327 : extern void cpp_define_formatted_unused (cpp_reader *pfile,
1328 : const char *fmt,
1329 : ...) ATTRIBUTE_PRINTF_2;
1330 : extern void cpp_assert (cpp_reader *, const char *);
1331 : extern void cpp_undef (cpp_reader *, const char *);
1332 : extern void cpp_unassert (cpp_reader *, const char *);
1333 :
1334 : /* Mark a node as a lazily defined macro. */
1335 : extern void cpp_define_lazily (cpp_reader *, cpp_hashnode *node, unsigned N);
1336 :
1337 : /* Undefine all macros and assertions. */
1338 : extern void cpp_undef_all (cpp_reader *);
1339 :
1340 : extern cpp_buffer *cpp_push_buffer (cpp_reader *, const unsigned char *,
1341 : size_t, int);
1342 : extern int cpp_defined (cpp_reader *, const unsigned char *, int);
1343 :
1344 : /* A preprocessing number. Code assumes that any unused high bits of
1345 : the double integer are set to zero. */
1346 :
1347 : /* This type has to be equal to unsigned HOST_WIDE_INT, see
1348 : gcc/c-family/c-lex.cc. */
1349 : typedef uint64_t cpp_num_part;
1350 : typedef struct cpp_num cpp_num;
1351 : struct cpp_num
1352 : {
1353 : cpp_num_part high;
1354 : cpp_num_part low;
1355 : bool unsignedp; /* True if value should be treated as unsigned. */
1356 : bool overflow; /* True if the most recent calculation overflowed. */
1357 : };
1358 :
1359 : /* cpplib provides two interfaces for interpretation of preprocessing
1360 : numbers.
1361 :
1362 : cpp_classify_number categorizes numeric constants according to
1363 : their field (integer, floating point, or invalid), radix (decimal,
1364 : octal, hexadecimal), and type suffixes. */
1365 :
1366 : #define CPP_N_CATEGORY 0x000F
1367 : #define CPP_N_INVALID 0x0000
1368 : #define CPP_N_INTEGER 0x0001
1369 : #define CPP_N_FLOATING 0x0002
1370 :
1371 : #define CPP_N_WIDTH 0x00F0
1372 : #define CPP_N_SMALL 0x0010 /* int, float, short _Fract/Accum */
1373 : #define CPP_N_MEDIUM 0x0020 /* long, double, long _Fract/_Accum. */
1374 : #define CPP_N_LARGE 0x0040 /* long long, long double,
1375 : long long _Fract/Accum. */
1376 :
1377 : #define CPP_N_WIDTH_MD 0xF0000 /* machine defined. */
1378 : #define CPP_N_MD_W 0x10000
1379 : #define CPP_N_MD_Q 0x20000
1380 :
1381 : #define CPP_N_RADIX 0x0F00
1382 : #define CPP_N_DECIMAL 0x0100
1383 : #define CPP_N_HEX 0x0200
1384 : #define CPP_N_OCTAL 0x0400
1385 : #define CPP_N_BINARY 0x0800
1386 :
1387 : #define CPP_N_UNSIGNED 0x1000 /* Properties. */
1388 : #define CPP_N_IMAGINARY 0x2000
1389 : #define CPP_N_DFLOAT 0x4000
1390 : #define CPP_N_DEFAULT 0x8000
1391 :
1392 : #define CPP_N_FRACT 0x100000 /* Fract types. */
1393 : #define CPP_N_ACCUM 0x200000 /* Accum types. */
1394 : #define CPP_N_FLOATN 0x400000 /* _FloatN types. */
1395 : #define CPP_N_FLOATNX 0x800000 /* _FloatNx types. */
1396 :
1397 : #define CPP_N_USERDEF 0x1000000 /* C++11 user-defined literal. */
1398 :
1399 : #define CPP_N_SIZE_T 0x2000000 /* C++23 size_t literal. */
1400 : #define CPP_N_BFLOAT16 0x4000000 /* std::bfloat16_t type. */
1401 : #define CPP_N_BITINT 0x8000000 /* C23 _BitInt literal. */
1402 :
1403 : #define CPP_N_WIDTH_FLOATN_NX 0xF0000000 /* _FloatN / _FloatNx value
1404 : of N, divided by 16. */
1405 : #define CPP_FLOATN_SHIFT 24
1406 : #define CPP_FLOATN_MAX 0xF0
1407 :
1408 : /* Classify a CPP_NUMBER token. The return value is a combination of
1409 : the flags from the above sets. */
1410 : extern unsigned cpp_classify_number (cpp_reader *, const cpp_token *,
1411 : const char **, location_t);
1412 :
1413 : /* Return the classification flags for a float suffix. */
1414 : extern unsigned int cpp_interpret_float_suffix (cpp_reader *, const char *,
1415 : size_t);
1416 :
1417 : /* Return the classification flags for an int suffix. */
1418 : extern unsigned int cpp_interpret_int_suffix (cpp_reader *, const char *,
1419 : size_t);
1420 :
1421 : /* Evaluate a token classified as category CPP_N_INTEGER. */
1422 : extern cpp_num cpp_interpret_integer (cpp_reader *, const cpp_token *,
1423 : unsigned int);
1424 :
1425 : /* Sign extend a number, with PRECISION significant bits and all
1426 : others assumed clear, to fill out a cpp_num structure. */
1427 : cpp_num cpp_num_sign_extend (cpp_num, size_t);
1428 :
1429 : /* Output a diagnostic of some kind. */
1430 : extern bool cpp_error (cpp_reader *, enum cpp_diagnostic_level,
1431 : const char *msgid, ...)
1432 : ATTRIBUTE_CPP_PPDIAG (3, 4);
1433 : extern bool cpp_warning (cpp_reader *, enum cpp_warning_reason,
1434 : const char *msgid, ...)
1435 : ATTRIBUTE_CPP_PPDIAG (3, 4);
1436 : extern bool cpp_pedwarning (cpp_reader *, enum cpp_warning_reason,
1437 : const char *msgid, ...)
1438 : ATTRIBUTE_CPP_PPDIAG (3, 4);
1439 : extern bool cpp_warning_syshdr (cpp_reader *, enum cpp_warning_reason reason,
1440 : const char *msgid, ...)
1441 : ATTRIBUTE_CPP_PPDIAG (3, 4);
1442 :
1443 : /* As their counterparts above, but use RICHLOC. */
1444 : extern bool cpp_warning_at (cpp_reader *, enum cpp_warning_reason,
1445 : rich_location *richloc, const char *msgid, ...)
1446 : ATTRIBUTE_CPP_PPDIAG (4, 5);
1447 : extern bool cpp_pedwarning_at (cpp_reader *, enum cpp_warning_reason,
1448 : rich_location *richloc, const char *msgid, ...)
1449 : ATTRIBUTE_CPP_PPDIAG (4, 5);
1450 :
1451 : /* Output a diagnostic with "MSGID: " preceding the
1452 : error string of errno. No location is printed. */
1453 : extern bool cpp_errno (cpp_reader *, enum cpp_diagnostic_level,
1454 : const char *msgid);
1455 : /* Similarly, but with "FILENAME: " instead of "MSGID: ", where
1456 : the filename is not localized. */
1457 : extern bool cpp_errno_filename (cpp_reader *, enum cpp_diagnostic_level,
1458 : const char *filename, location_t loc);
1459 :
1460 : /* Same as cpp_error, except additionally specifies a position as a
1461 : (translation unit) physical line and physical column. If the line is
1462 : zero, then no location is printed. */
1463 : extern bool cpp_error_with_line (cpp_reader *, enum cpp_diagnostic_level,
1464 : location_t, unsigned,
1465 : const char *msgid, ...)
1466 : ATTRIBUTE_CPP_PPDIAG (5, 6);
1467 : extern bool cpp_warning_with_line (cpp_reader *, enum cpp_warning_reason,
1468 : location_t, unsigned,
1469 : const char *msgid, ...)
1470 : ATTRIBUTE_CPP_PPDIAG (5, 6);
1471 : extern bool cpp_pedwarning_with_line (cpp_reader *, enum cpp_warning_reason,
1472 : location_t, unsigned,
1473 : const char *msgid, ...)
1474 : ATTRIBUTE_CPP_PPDIAG (5, 6);
1475 : extern bool cpp_warning_with_line_syshdr (cpp_reader *, enum cpp_warning_reason,
1476 : location_t, unsigned,
1477 : const char *msgid, ...)
1478 : ATTRIBUTE_CPP_PPDIAG (5, 6);
1479 :
1480 : extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1481 : location_t src_loc, const char *msgid, ...)
1482 : ATTRIBUTE_CPP_PPDIAG (4, 5);
1483 :
1484 : extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1485 : rich_location *richloc, const char *msgid, ...)
1486 : ATTRIBUTE_CPP_PPDIAG (4, 5);
1487 :
1488 : /* In lex.cc */
1489 : extern int cpp_ideq (const cpp_token *, const char *);
1490 : extern void cpp_output_line (cpp_reader *, FILE *);
1491 : extern unsigned char *cpp_output_line_to_string (cpp_reader *,
1492 : const unsigned char *);
1493 : extern const unsigned char *cpp_alloc_token_string
1494 : (cpp_reader *, const unsigned char *, unsigned);
1495 : extern void cpp_output_token (const cpp_token *, FILE *);
1496 : extern const char *cpp_type2name (enum cpp_ttype, unsigned char flags);
1497 : /* Returns the value of an escape sequence, truncated to the correct
1498 : target precision. PSTR points to the input pointer, which is just
1499 : after the backslash. LIMIT is how much text we have. WIDE is true
1500 : if the escape sequence is part of a wide character constant or
1501 : string literal. Handles all relevant diagnostics. */
1502 : extern cppchar_t cpp_parse_escape (cpp_reader *, const unsigned char ** pstr,
1503 : const unsigned char *limit, int wide);
1504 :
1505 : /* Structure used to hold a comment block at a given location in the
1506 : source code. */
1507 :
1508 : typedef struct
1509 : {
1510 : /* Text of the comment including the terminators. */
1511 : char *comment;
1512 :
1513 : /* source location for the given comment. */
1514 : location_t sloc;
1515 : } cpp_comment;
1516 :
1517 : /* Structure holding all comments for a given cpp_reader. */
1518 :
1519 : typedef struct
1520 : {
1521 : /* table of comment entries. */
1522 : cpp_comment *entries;
1523 :
1524 : /* number of actual entries entered in the table. */
1525 : int count;
1526 :
1527 : /* number of entries allocated currently. */
1528 : int allocated;
1529 : } cpp_comment_table;
1530 :
1531 : /* Returns the table of comments encountered by the preprocessor. This
1532 : table is only populated when pfile->state.save_comments is true. */
1533 : extern cpp_comment_table *cpp_get_comments (cpp_reader *);
1534 :
1535 : /* In hash.c */
1536 :
1537 : /* Lookup an identifier in the hashtable. Puts the identifier in the
1538 : table if it is not already there. */
1539 : extern cpp_hashnode *cpp_lookup (cpp_reader *, const unsigned char *,
1540 : unsigned int);
1541 :
1542 : /* Set NODE_WARN flag for NAME, such that there will be diagnostics
1543 : for #define or #undef of NAME. */
1544 :
1545 : inline void
1546 7155424 : cpp_warn (cpp_reader *pfile, const char *name, unsigned int len)
1547 : {
1548 7155424 : cpp_lookup (pfile, (const unsigned char *) name, len)->flags |= NODE_WARN;
1549 7155424 : }
1550 :
1551 : inline void
1552 401121 : cpp_warn (cpp_reader *pfile, const char *name)
1553 : {
1554 401121 : cpp_warn (pfile, name, strlen (name));
1555 401121 : }
1556 :
1557 : typedef int (*cpp_cb) (cpp_reader *, cpp_hashnode *, void *);
1558 : extern void cpp_forall_identifiers (cpp_reader *, cpp_cb, void *);
1559 :
1560 : /* In macro.cc */
1561 : extern void cpp_scan_nooutput (cpp_reader *);
1562 : extern int cpp_sys_macro_p (cpp_reader *);
1563 : extern unsigned char *cpp_quote_string (unsigned char *, const unsigned char *,
1564 : unsigned int);
1565 : extern bool cpp_compare_macros (const cpp_macro *macro1,
1566 : const cpp_macro *macro2);
1567 :
1568 : /* In files.cc */
1569 : extern bool cpp_included (cpp_reader *, const char *);
1570 : extern bool cpp_included_before (cpp_reader *, const char *, location_t);
1571 : extern void cpp_make_system_header (cpp_reader *, int, int);
1572 : extern bool cpp_push_include (cpp_reader *, const char *);
1573 : extern bool cpp_push_default_include (cpp_reader *, const char *);
1574 : extern void cpp_change_file (cpp_reader *, enum lc_reason, const char *);
1575 : extern const char *_cpp_get_file_path (_cpp_file *);
1576 : extern const char *_cpp_get_file_name (_cpp_file *);
1577 : extern struct stat *_cpp_get_file_stat (_cpp_file *);
1578 : extern struct cpp_dir *_cpp_get_file_dir (_cpp_file *);
1579 : extern cpp_buffer *cpp_get_buffer (cpp_reader *);
1580 : extern struct _cpp_file *cpp_get_file (cpp_buffer *);
1581 : extern cpp_buffer *cpp_get_prev (cpp_buffer *);
1582 : extern void cpp_clear_file_cache (cpp_reader *);
1583 :
1584 : /* cpp_get_converted_source returns the contents of the given file, as it exists
1585 : after cpplib has read it and converted it from the input charset to the
1586 : source charset. Return struct will be zero-filled if the data could not be
1587 : read for any reason. The data starts at the DATA pointer, but the TO_FREE
1588 : pointer is what should be passed to free(), as there may be an offset. */
1589 : struct cpp_converted_source
1590 : {
1591 : char *to_free;
1592 : char *data;
1593 : size_t len;
1594 : };
1595 : cpp_converted_source cpp_get_converted_source (const char *fname,
1596 : const char *input_charset);
1597 :
1598 : /* In pch.cc */
1599 : struct save_macro_data;
1600 : extern int cpp_save_state (cpp_reader *, FILE *);
1601 : extern int cpp_write_pch_deps (cpp_reader *, FILE *);
1602 : extern int cpp_write_pch_state (cpp_reader *, FILE *);
1603 : extern int cpp_valid_state (cpp_reader *, const char *, int);
1604 : extern void cpp_prepare_state (cpp_reader *, struct save_macro_data **);
1605 : extern int cpp_read_state (cpp_reader *, const char *, FILE *,
1606 : struct save_macro_data *);
1607 :
1608 : /* In lex.cc */
1609 : extern void cpp_force_token_locations (cpp_reader *, location_t);
1610 : extern void cpp_stop_forcing_token_locations (cpp_reader *);
1611 : enum CPP_DO_task
1612 : {
1613 : CPP_DO_print,
1614 : CPP_DO_location,
1615 : CPP_DO_token
1616 : };
1617 :
1618 : extern void cpp_directive_only_process (cpp_reader *pfile,
1619 : void *data,
1620 : void (*cb) (cpp_reader *,
1621 : CPP_DO_task,
1622 : void *data, ...));
1623 :
1624 : /* In expr.cc */
1625 : extern enum cpp_ttype cpp_userdef_string_remove_type
1626 : (enum cpp_ttype type);
1627 : extern enum cpp_ttype cpp_userdef_string_add_type
1628 : (enum cpp_ttype type);
1629 : extern enum cpp_ttype cpp_userdef_char_remove_type
1630 : (enum cpp_ttype type);
1631 : extern enum cpp_ttype cpp_userdef_char_add_type
1632 : (enum cpp_ttype type);
1633 : extern bool cpp_userdef_string_p
1634 : (enum cpp_ttype type);
1635 : extern bool cpp_userdef_char_p
1636 : (enum cpp_ttype type);
1637 : extern const char * cpp_get_userdef_suffix
1638 : (const cpp_token *);
1639 :
1640 : /* In charset.cc */
1641 :
1642 : /* The result of attempting to decode a run of UTF-8 bytes. */
1643 :
1644 : struct cpp_decoded_char
1645 : {
1646 : const char *m_start_byte;
1647 : const char *m_next_byte;
1648 :
1649 : bool m_valid_ch;
1650 : cppchar_t m_ch;
1651 : };
1652 :
1653 : /* Information for mapping between code points and display columns.
1654 :
1655 : This is a tabstop value, along with a callback for getting the
1656 : widths of characters. Normally this callback is cpp_wcwidth, but we
1657 : support other schemes for escaping non-ASCII unicode as a series of
1658 : ASCII chars when printing the user's source code in
1659 : gcc/diagnostics/source-printing.cc
1660 :
1661 : For example, consider:
1662 : - the Unicode character U+03C0 "GREEK SMALL LETTER PI" (UTF-8: 0xCF 0x80)
1663 : - the Unicode character U+1F642 "SLIGHTLY SMILING FACE"
1664 : (UTF-8: 0xF0 0x9F 0x99 0x82)
1665 : - the byte 0xBF (a stray trailing byte of a UTF-8 character)
1666 : Normally U+03C0 would occupy one display column, U+1F642
1667 : would occupy two display columns, and the stray byte would be
1668 : printed verbatim as one display column.
1669 :
1670 : However when escaping them as unicode code points as "<U+03C0>"
1671 : and "<U+1F642>" they occupy 8 and 9 display columns respectively,
1672 : and when escaping them as bytes as "<CF><80>" and "<F0><9F><99><82>"
1673 : they occupy 8 and 16 display columns respectively. In both cases
1674 : the stray byte is escaped to <BF> as 4 display columns. */
1675 :
1676 : struct cpp_char_column_policy
1677 : {
1678 355693 : cpp_char_column_policy (int tabstop,
1679 : int (*width_cb) (cppchar_t c))
1680 406903 : : m_tabstop (tabstop),
1681 406903 : m_undecoded_byte_width (1),
1682 406891 : m_width_cb (width_cb)
1683 : {}
1684 :
1685 : int m_tabstop;
1686 : /* Width in display columns of a stray byte that isn't decodable
1687 : as UTF-8. */
1688 : int m_undecoded_byte_width;
1689 : int (*m_width_cb) (cppchar_t c);
1690 : };
1691 :
1692 : /* A class to manage the state while converting a UTF-8 sequence to cppchar_t
1693 : and computing the display width one character at a time. */
1694 : class cpp_display_width_computation {
1695 : public:
1696 : cpp_display_width_computation (const char *data, int data_length,
1697 : const cpp_char_column_policy &policy);
1698 3862351 : const char *next_byte () const { return m_next; }
1699 2937613 : int bytes_processed () const { return m_next - m_begin; }
1700 4096814 : int bytes_left () const { return m_bytes_left; }
1701 4096814 : bool done () const { return !bytes_left (); }
1702 7119947 : int display_cols_processed () const { return m_display_cols; }
1703 :
1704 : int process_next_codepoint (cpp_decoded_char *out);
1705 : int advance_display_cols (int n);
1706 :
1707 : private:
1708 : const char *const m_begin;
1709 : const char *m_next;
1710 : size_t m_bytes_left;
1711 : const cpp_char_column_policy &m_policy;
1712 : int m_display_cols;
1713 : };
1714 :
1715 : /* Convenience functions that are simple use cases for class
1716 : cpp_display_width_computation. Tab characters will be expanded to spaces
1717 : as determined by POLICY.m_tabstop, and non-printable-ASCII characters
1718 : will be escaped as per POLICY. */
1719 :
1720 : int cpp_byte_column_to_display_column (const char *data, int data_length,
1721 : int column,
1722 : const cpp_char_column_policy &policy);
1723 75060 : inline int cpp_display_width (const char *data, int data_length,
1724 : const cpp_char_column_policy &policy)
1725 : {
1726 75060 : return cpp_byte_column_to_display_column (data, data_length, data_length,
1727 : policy);
1728 : }
1729 : int cpp_display_column_to_byte_column (const char *data, int data_length,
1730 : int display_col,
1731 : const cpp_char_column_policy &policy);
1732 : int cpp_wcwidth (cppchar_t c);
1733 :
1734 : bool cpp_input_conversion_is_trivial (const char *input_charset);
1735 : int cpp_check_utf8_bom (const char *data, size_t data_length);
1736 : bool cpp_valid_utf8_p (const char *data, size_t num_bytes);
1737 :
1738 : bool cpp_is_combining_char (cppchar_t c);
1739 : bool cpp_is_printable_char (cppchar_t c);
1740 :
1741 : enum cpp_xid_property {
1742 : CPP_XID_START = 1,
1743 : CPP_XID_CONTINUE = 2
1744 : };
1745 :
1746 : unsigned int cpp_check_xid_property (cppchar_t c);
1747 :
1748 : /* In errors.cc */
1749 :
1750 : /* RAII class to suppress CPP diagnostics in the current scope. */
1751 : class cpp_auto_suppress_diagnostics
1752 : {
1753 : public:
1754 : explicit cpp_auto_suppress_diagnostics (cpp_reader *pfile);
1755 : ~cpp_auto_suppress_diagnostics ();
1756 : private:
1757 : cpp_reader *const m_pfile;
1758 : const decltype (cpp_callbacks::diagnostic) m_cb;
1759 : };
1760 :
1761 : #endif /* ! LIBCPP_CPPLIB_H */
|
