Branch data Line data Source code
1 : : /* Definitions for CPP library.
2 : : Copyright (C) 1995-2025 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 : : /* The remainder of the punctuation. Order is not significant. */ \
105 : : OP(SEMICOLON, ";") /* structure */ \
106 : : OP(ELLIPSIS, "...") \
107 : : OP(PLUS_PLUS, "++") /* increment */ \
108 : : OP(MINUS_MINUS, "--") \
109 : : OP(DEREF, "->") /* accessors */ \
110 : : OP(DOT, ".") \
111 : : OP(SCOPE, "::") \
112 : : OP(DEREF_STAR, "->*") \
113 : : OP(DOT_STAR, ".*") \
114 : : OP(ATSIGN, "@") /* used in Objective-C */ \
115 : : \
116 : : TK(NAME, IDENT) /* word */ \
117 : : TK(AT_NAME, IDENT) /* @word - Objective-C */ \
118 : : TK(NUMBER, LITERAL) /* 34_be+ta */ \
119 : : \
120 : : TK(CHAR, LITERAL) /* 'char' */ \
121 : : TK(WCHAR, LITERAL) /* L'char' */ \
122 : : TK(CHAR16, LITERAL) /* u'char' */ \
123 : : TK(CHAR32, LITERAL) /* U'char' */ \
124 : : TK(UTF8CHAR, LITERAL) /* u8'char' */ \
125 : : TK(OTHER, LITERAL) /* stray punctuation */ \
126 : : \
127 : : TK(STRING, LITERAL) /* "string" */ \
128 : : TK(WSTRING, LITERAL) /* L"string" */ \
129 : : TK(STRING16, LITERAL) /* u"string" */ \
130 : : TK(STRING32, LITERAL) /* U"string" */ \
131 : : TK(UTF8STRING, LITERAL) /* u8"string" */ \
132 : : TK(OBJC_STRING, LITERAL) /* @"string" - Objective-C */ \
133 : : TK(HEADER_NAME, LITERAL) /* <stdio.h> in #include */ \
134 : : TK(UNEVAL_STRING, LITERAL) /* unevaluated "string" - C++26 */ \
135 : : \
136 : : TK(CHAR_USERDEF, LITERAL) /* 'char'_suffix - C++11 */ \
137 : : TK(WCHAR_USERDEF, LITERAL) /* L'char'_suffix - C++11 */ \
138 : : TK(CHAR16_USERDEF, LITERAL) /* u'char'_suffix - C++11 */ \
139 : : TK(CHAR32_USERDEF, LITERAL) /* U'char'_suffix - C++11 */ \
140 : : TK(UTF8CHAR_USERDEF, LITERAL) /* u8'char'_suffix - C++11 */ \
141 : : TK(STRING_USERDEF, LITERAL) /* "string"_suffix - C++11 */ \
142 : : TK(WSTRING_USERDEF, LITERAL) /* L"string"_suffix - C++11 */ \
143 : : TK(STRING16_USERDEF, LITERAL) /* u"string"_suffix - C++11 */ \
144 : : TK(STRING32_USERDEF, LITERAL) /* U"string"_suffix - C++11 */ \
145 : : TK(UTF8STRING_USERDEF,LITERAL) /* u8"string"_suffix - C++11 */ \
146 : : \
147 : : TK(EMBED, LITERAL) /* #embed - C23 */ \
148 : : \
149 : : TK(COMMENT, LITERAL) /* Only if output comments. */ \
150 : : /* SPELL_LITERAL happens to DTRT. */ \
151 : : TK(MACRO_ARG, NONE) /* Macro argument. */ \
152 : : TK(PRAGMA, NONE) /* Only for deferred pragmas. */ \
153 : : TK(PRAGMA_EOL, NONE) /* End-of-line for deferred pragmas. */ \
154 : : TK(PADDING, NONE) /* Whitespace for -E. */
155 : :
156 : : #define OP(e, s) CPP_ ## e,
157 : : #define TK(e, s) CPP_ ## e,
158 : : enum cpp_ttype
159 : : {
160 : : TTYPE_TABLE
161 : : N_TTYPES,
162 : :
163 : : /* A token type for keywords, as opposed to ordinary identifiers. */
164 : : CPP_KEYWORD,
165 : :
166 : : /* Positions in the table. */
167 : : CPP_LAST_EQ = CPP_LSHIFT,
168 : : CPP_FIRST_DIGRAPH = CPP_HASH,
169 : : CPP_LAST_PUNCTUATOR= CPP_ATSIGN,
170 : : CPP_LAST_CPP_OP = CPP_LESS_EQ
171 : : };
172 : : #undef OP
173 : : #undef TK
174 : :
175 : : /* C language kind, used when calling cpp_create_reader. */
176 : : enum c_lang {CLK_GNUC89 = 0, CLK_GNUC99, CLK_GNUC11, CLK_GNUC17, CLK_GNUC23,
177 : : CLK_GNUC2Y,
178 : : CLK_STDC89, CLK_STDC94, CLK_STDC99, CLK_STDC11, CLK_STDC17,
179 : : CLK_STDC23, CLK_STDC2Y,
180 : : CLK_GNUCXX, CLK_CXX98, CLK_GNUCXX11, CLK_CXX11,
181 : : CLK_GNUCXX14, CLK_CXX14, CLK_GNUCXX17, CLK_CXX17,
182 : : CLK_GNUCXX20, CLK_CXX20, CLK_GNUCXX23, CLK_CXX23,
183 : : CLK_GNUCXX26, CLK_CXX26, CLK_ASM};
184 : :
185 : : /* Payload of a NUMBER, STRING, CHAR or COMMENT token. */
186 : : struct GTY(()) cpp_string {
187 : : unsigned int len;
188 : :
189 : : /* TEXT is always null terminated (terminator not included in len); but this
190 : : GTY markup arranges that PCH streaming works properly even if there is a
191 : : null byte in the middle of the string. */
192 : : const unsigned char * GTY((string_length ("1 + %h.len"))) text;
193 : : };
194 : :
195 : : /* Flags for the cpp_token structure. */
196 : : #define PREV_WHITE (1 << 0) /* If whitespace before this token. */
197 : : #define DIGRAPH (1 << 1) /* If it was a digraph. */
198 : : #define STRINGIFY_ARG (1 << 2) /* If macro argument to be stringified. */
199 : : #define PASTE_LEFT (1 << 3) /* If on LHS of a ## operator. */
200 : : #define NAMED_OP (1 << 4) /* C++ named operators. */
201 : : #define PREV_FALLTHROUGH (1 << 5) /* On a token preceeded by FALLTHROUGH
202 : : comment. */
203 : : #define DECIMAL_INT (1 << 6) /* Decimal integer, set in c-lex.cc. */
204 : : #define PURE_ZERO (1 << 7) /* Single 0 digit, used by the C++ frontend,
205 : : set in c-lex.cc. */
206 : : #define COLON_SCOPE PURE_ZERO /* Adjacent colons in C < 23. */
207 : : #define NO_DOT_COLON PURE_ZERO /* Set on CPP_NAME tokens whose expansion
208 : : shouldn't start with CPP_DOT or CPP_COLON
209 : : after optional CPP_PADDING. */
210 : : #define SP_DIGRAPH (1 << 8) /* # or ## token was a digraph. */
211 : : #define SP_PREV_WHITE (1 << 9) /* If whitespace before a ##
212 : : operator, or before this token
213 : : after a # operator. */
214 : : #define NO_EXPAND (1 << 10) /* Do not macro-expand this token. */
215 : : #define PRAGMA_OP (1 << 11) /* _Pragma token. */
216 : : #define BOL (1 << 12) /* Token at beginning of line. */
217 : :
218 : : /* Specify which field, if any, of the cpp_token union is used. */
219 : :
220 : : enum cpp_token_fld_kind {
221 : : CPP_TOKEN_FLD_NODE,
222 : : CPP_TOKEN_FLD_SOURCE,
223 : : CPP_TOKEN_FLD_STR,
224 : : CPP_TOKEN_FLD_ARG_NO,
225 : : CPP_TOKEN_FLD_TOKEN_NO,
226 : : CPP_TOKEN_FLD_PRAGMA,
227 : : CPP_TOKEN_FLD_NONE
228 : : };
229 : :
230 : : /* A macro argument in the cpp_token union. */
231 : : struct GTY(()) cpp_macro_arg {
232 : : /* Argument number. */
233 : : unsigned int arg_no;
234 : : /* The original spelling of the macro argument token. */
235 : : cpp_hashnode *
236 : : GTY ((nested_ptr (union tree_node,
237 : : "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
238 : : "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
239 : : spelling;
240 : : };
241 : :
242 : : /* An identifier in the cpp_token union. */
243 : : struct GTY(()) cpp_identifier {
244 : : /* The canonical (UTF-8) spelling of the identifier. */
245 : : cpp_hashnode *
246 : : GTY ((nested_ptr (union tree_node,
247 : : "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
248 : : "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
249 : : node;
250 : : /* The original spelling of the identifier. */
251 : : cpp_hashnode *
252 : : GTY ((nested_ptr (union tree_node,
253 : : "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
254 : : "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
255 : : spelling;
256 : : };
257 : :
258 : : /* A preprocessing token. This occupies 32 bytes on a 64-bit host. On a
259 : : 32-bit host it occupies 20 or 24 bytes, depending whether a uint64_t
260 : : requires 4- or 8-byte alignment. */
261 : :
262 : : struct GTY(()) cpp_token {
263 : :
264 : : /* Location of first char of token, together with range of full token. */
265 : : location_t src_loc;
266 : :
267 : : ENUM_BITFIELD(cpp_ttype) type : CHAR_BIT; /* token type */
268 : : unsigned short flags; /* flags - see above */
269 : :
270 : : union cpp_token_u
271 : : {
272 : : /* An identifier. */
273 : : struct cpp_identifier GTY ((tag ("CPP_TOKEN_FLD_NODE"))) node;
274 : :
275 : : /* Inherit padding from this token. */
276 : : cpp_token * GTY ((tag ("CPP_TOKEN_FLD_SOURCE"))) source;
277 : :
278 : : /* A string, or number. */
279 : : struct cpp_string GTY ((tag ("CPP_TOKEN_FLD_STR"))) str;
280 : :
281 : : /* Argument no. (and original spelling) for a CPP_MACRO_ARG. */
282 : : struct cpp_macro_arg GTY ((tag ("CPP_TOKEN_FLD_ARG_NO"))) macro_arg;
283 : :
284 : : /* Original token no. for a CPP_PASTE (from a sequence of
285 : : consecutive paste tokens in a macro expansion). */
286 : : unsigned int GTY ((tag ("CPP_TOKEN_FLD_TOKEN_NO"))) token_no;
287 : :
288 : : /* Caller-supplied identifier for a CPP_PRAGMA. */
289 : : unsigned int GTY ((tag ("CPP_TOKEN_FLD_PRAGMA"))) pragma;
290 : : } GTY ((desc ("cpp_token_val_index (&%1)"))) val;
291 : : };
292 : :
293 : : /* Say which field is in use. */
294 : : extern enum cpp_token_fld_kind cpp_token_val_index (const cpp_token *tok);
295 : :
296 : : /* A type wide enough to hold any multibyte source character.
297 : : cpplib's character constant interpreter requires an unsigned type.
298 : : Also, a typedef for the signed equivalent.
299 : : The width of this type is capped at 32 bits; there do exist targets
300 : : where wchar_t is 64 bits, but only in a non-default mode, and there
301 : : would be no meaningful interpretation for a wchar_t value greater
302 : : than 2^32 anyway -- the widest wide-character encoding around is
303 : : ISO 10646, which stops at 2^31. */
304 : : #if CHAR_BIT * SIZEOF_INT >= 32
305 : : # define CPPCHAR_SIGNED_T int
306 : : #elif CHAR_BIT * SIZEOF_LONG >= 32
307 : : # define CPPCHAR_SIGNED_T long
308 : : #else
309 : : # error "Cannot find a least-32-bit signed integer type"
310 : : #endif
311 : : typedef unsigned CPPCHAR_SIGNED_T cppchar_t;
312 : : typedef CPPCHAR_SIGNED_T cppchar_signed_t;
313 : :
314 : : /* Style of header dependencies to generate. */
315 : : enum cpp_deps_style { DEPS_NONE = 0, DEPS_USER, DEPS_SYSTEM };
316 : :
317 : : /* Structured format of module dependencies to generate. */
318 : : enum cpp_fdeps_format { FDEPS_FMT_NONE = 0, FDEPS_FMT_P1689R5 };
319 : :
320 : : /* The possible normalization levels, from most restrictive to least. */
321 : : enum cpp_normalize_level {
322 : : /* In NFKC. */
323 : : normalized_KC = 0,
324 : : /* In NFC. */
325 : : normalized_C,
326 : : /* In NFC, except for subsequences where being in NFC would make
327 : : the identifier invalid. */
328 : : normalized_identifier_C,
329 : : /* Not normalized at all. */
330 : : normalized_none
331 : : };
332 : :
333 : : enum cpp_main_search
334 : : {
335 : : CMS_none, /* A regular source file. */
336 : : CMS_header, /* Is a directly-specified header file (eg PCH or
337 : : header-unit). */
338 : : CMS_user, /* Search the user INCLUDE path. */
339 : : CMS_system, /* Search the system INCLUDE path. */
340 : : };
341 : :
342 : : /* The possible bidirectional control characters checking levels. */
343 : : enum cpp_bidirectional_level {
344 : : /* No checking. */
345 : : bidirectional_none = 0,
346 : : /* Only detect unpaired uses of bidirectional control characters. */
347 : : bidirectional_unpaired = 1,
348 : : /* Detect any use of bidirectional control characters. */
349 : : bidirectional_any = 2,
350 : : /* Also warn about UCNs. */
351 : : bidirectional_ucn = 4
352 : : };
353 : :
354 : : /* This structure is nested inside struct cpp_reader, and
355 : : carries all the options visible to the command line. */
356 : : struct cpp_options
357 : : {
358 : : /* The language we're preprocessing. */
359 : : enum c_lang lang;
360 : :
361 : : /* Nonzero means use extra default include directories for C++. */
362 : : unsigned char cplusplus;
363 : :
364 : : /* Nonzero means handle cplusplus style comments. */
365 : : unsigned char cplusplus_comments;
366 : :
367 : : /* Nonzero means define __OBJC__, treat @ as a special token, use
368 : : the OBJC[PLUS]_INCLUDE_PATH environment variable, and allow
369 : : "#import". */
370 : : unsigned char objc;
371 : :
372 : : /* Nonzero means don't copy comments into the output file. */
373 : : unsigned char discard_comments;
374 : :
375 : : /* Nonzero means don't copy comments into the output file during
376 : : macro expansion. */
377 : : unsigned char discard_comments_in_macro_exp;
378 : :
379 : : /* Nonzero means process the ISO trigraph sequences. */
380 : : unsigned char trigraphs;
381 : :
382 : : /* Nonzero means process the ISO digraph sequences. */
383 : : unsigned char digraphs;
384 : :
385 : : /* Nonzero means to allow hexadecimal floats and LL suffixes. */
386 : : unsigned char extended_numbers;
387 : :
388 : : /* Nonzero means process u/U prefix literals (UTF-16/32). */
389 : : unsigned char uliterals;
390 : :
391 : : /* Nonzero means process u8 prefixed character literals (UTF-8). */
392 : : unsigned char utf8_char_literals;
393 : :
394 : : /* Nonzero means process r/R raw strings. If this is set, uliterals
395 : : must be set as well. */
396 : : unsigned char rliterals;
397 : :
398 : : /* Nonzero means print names of header files (-H). */
399 : : unsigned char print_include_names;
400 : :
401 : : /* Nonzero means complain about deprecated features. */
402 : : unsigned char cpp_warn_deprecated;
403 : :
404 : : /* Nonzero means warn if slash-star appears in a comment. */
405 : : unsigned char warn_comments;
406 : :
407 : : /* Nonzero means to warn about __DATA__, __TIME__ and __TIMESTAMP__ usage. */
408 : : unsigned char warn_date_time;
409 : :
410 : : /* Nonzero means warn if a user-supplied include directory does not
411 : : exist. */
412 : : unsigned char warn_missing_include_dirs;
413 : :
414 : : /* Nonzero means warn if there are any trigraphs. */
415 : : unsigned char warn_trigraphs;
416 : :
417 : : /* Nonzero means warn about multicharacter charconsts. */
418 : : unsigned char warn_multichar;
419 : :
420 : : /* Nonzero means warn about various incompatibilities with
421 : : traditional C. */
422 : : unsigned char cpp_warn_traditional;
423 : :
424 : : /* Nonzero means warn about long long numeric constants. */
425 : : unsigned char cpp_warn_long_long;
426 : :
427 : : /* Nonzero means warn about text after an #endif (or #else). */
428 : : unsigned char warn_endif_labels;
429 : :
430 : : /* Nonzero means warn about implicit sign changes owing to integer
431 : : promotions. */
432 : : unsigned char warn_num_sign_change;
433 : :
434 : : /* Zero means don't warn about __VA_ARGS__ usage in c89 pedantic mode.
435 : : Presumably the usage is protected by the appropriate #ifdef. */
436 : : unsigned char warn_variadic_macros;
437 : :
438 : : /* Non-zero means suppress diagnostics for NODE_WARN #define or #undef.
439 : : Used for cpp_define/cpp_undef. */
440 : : unsigned char suppress_builtin_macro_warnings;
441 : :
442 : : /* Nonzero means warn about builtin macros that are redefined or
443 : : explicitly undefined. */
444 : : unsigned char warn_builtin_macro_redefined;
445 : :
446 : : /* Different -Wimplicit-fallthrough= levels. */
447 : : unsigned char cpp_warn_implicit_fallthrough;
448 : :
449 : : /* Nonzero means warn about a define of a different macro right after
450 : : #ifndef/#if !defined header guard directive. */
451 : : unsigned char warn_header_guard;
452 : :
453 : : /* Nonzero means we should look for header.gcc files that remap file
454 : : names. */
455 : : unsigned char remap;
456 : :
457 : : /* Zero means dollar signs are punctuation. */
458 : : unsigned char dollars_in_ident;
459 : :
460 : : /* Nonzero means UCNs are accepted in identifiers. */
461 : : unsigned char extended_identifiers;
462 : :
463 : : /* True if we should warn about dollars in identifiers or numbers
464 : : for this translation unit. */
465 : : unsigned char warn_dollars;
466 : :
467 : : /* Nonzero means warn if undefined identifiers are evaluated in an #if. */
468 : : unsigned char warn_undef;
469 : :
470 : : /* Nonzero means warn if "defined" is encountered in a place other than
471 : : an #if. */
472 : : unsigned char warn_expansion_to_defined;
473 : :
474 : : /* Nonzero means warn of unused macros from the main file. */
475 : : unsigned char warn_unused_macros;
476 : :
477 : : /* Nonzero for the 1999 C Standard, including corrigenda and amendments. */
478 : : unsigned char c99;
479 : :
480 : : /* Nonzero if we are conforming to a specific C or C++ standard. */
481 : : unsigned char std;
482 : :
483 : : /* Nonzero means give all the error messages the ANSI standard requires. */
484 : : unsigned char cpp_pedantic;
485 : :
486 : : /* Nonzero means we're looking at already preprocessed code, so don't
487 : : bother trying to do macro expansion and whatnot. */
488 : : unsigned char preprocessed;
489 : :
490 : : /* Nonzero means we are going to emit debugging logs during
491 : : preprocessing. */
492 : : unsigned char debug;
493 : :
494 : : /* Nonzero means we are tracking locations of tokens involved in
495 : : macro expansion. 1 Means we track the location in degraded mode
496 : : where we do not track locations of tokens resulting from the
497 : : expansion of arguments of function-like macro. 2 Means we do
498 : : track all macro expansions. This last option is the one that
499 : : consumes the highest amount of memory. */
500 : : unsigned char track_macro_expansion;
501 : :
502 : : /* Nonzero means handle C++ alternate operator names. */
503 : : unsigned char operator_names;
504 : :
505 : : /* Nonzero means warn about use of C++ alternate operator names. */
506 : : unsigned char warn_cxx_operator_names;
507 : :
508 : : /* True for traditional preprocessing. */
509 : : unsigned char traditional;
510 : :
511 : : /* Nonzero for C++ 2011 Standard user-defined literals. */
512 : : unsigned char user_literals;
513 : :
514 : : /* Nonzero means warn when a string or character literal is followed by a
515 : : ud-suffix which does not beging with an underscore. */
516 : : unsigned char warn_literal_suffix;
517 : :
518 : : /* Nonzero means interpret imaginary, fixed-point, or other gnu extension
519 : : literal number suffixes as user-defined literal number suffixes. */
520 : : unsigned char ext_numeric_literals;
521 : :
522 : : /* Nonzero means extended identifiers allow the characters specified
523 : : in C11. */
524 : : unsigned char c11_identifiers;
525 : :
526 : : /* Nonzero means extended identifiers allow the characters specified
527 : : by Unicode XID_Start and XID_Continue properties. */
528 : : unsigned char xid_identifiers;
529 : :
530 : : /* Nonzero for C++ 2014 Standard binary constants. */
531 : : unsigned char binary_constants;
532 : :
533 : : /* Nonzero for C2Y imaginary (floating) constants. */
534 : : unsigned char imaginary_constants;
535 : :
536 : : /* Nonzero for C++ 2014 Standard digit separators. */
537 : : unsigned char digit_separators;
538 : :
539 : : /* Nonzero for C23 decimal floating-point constants. */
540 : : unsigned char dfp_constants;
541 : :
542 : : /* Nonzero for C++20 __VA_OPT__ feature. */
543 : : unsigned char va_opt;
544 : :
545 : : /* Nonzero for the '::' token. */
546 : : unsigned char scope;
547 : :
548 : : /* Nonzero for the '#elifdef' and '#elifndef' directives. */
549 : : unsigned char elifdef;
550 : :
551 : : /* Nonzero for the '#warning' directive. */
552 : : unsigned char warning_directive;
553 : :
554 : : /* Nonzero means tokenize C++20 module directives. */
555 : : unsigned char module_directives;
556 : :
557 : : /* Nonzero for C++23 size_t literals. */
558 : : unsigned char size_t_literals;
559 : :
560 : : /* Nonzero for C++23 delimited escape sequences. */
561 : : unsigned char delimited_escape_seqs;
562 : :
563 : : /* Nonzero for C++23 named universal character escape sequences. */
564 : : unsigned char named_uc_escape_seqs;
565 : :
566 : : /* Nonzero for C++ and C23 UCNs for characters below 0xa0. */
567 : : unsigned char low_ucns;
568 : :
569 : : /* Nonzero for C2Y 0o prefixed octal integer constants. */
570 : : unsigned char octal_constants;
571 : :
572 : : /* Nonzero for 'true' and 'false' in #if expressions. */
573 : : unsigned char true_false;
574 : :
575 : : /* Nonzero for the '#embed' directive. */
576 : : unsigned char embed;
577 : :
578 : : /* Holds the name of the target (execution) character set. */
579 : : const char *narrow_charset;
580 : :
581 : : /* Holds the name of the target wide character set. */
582 : : const char *wide_charset;
583 : :
584 : : /* Holds the name of the input character set. */
585 : : const char *input_charset;
586 : :
587 : : /* The minimum permitted level of normalization before a warning
588 : : is generated. See enum cpp_normalize_level. */
589 : : int warn_normalize;
590 : :
591 : : /* True to warn about precompiled header files we couldn't use. */
592 : : bool warn_invalid_pch;
593 : :
594 : : /* True if dependencies should be restored from a precompiled header. */
595 : : bool restore_pch_deps;
596 : :
597 : : /* True if warn about differences between C90 and C99. */
598 : : signed char cpp_warn_c90_c99_compat;
599 : :
600 : : /* True if warn about differences between C11 and C23. */
601 : : signed char cpp_warn_c11_c23_compat;
602 : :
603 : : /* True if warn about differences between C23 and C2Y. */
604 : : signed char cpp_warn_c23_c2y_compat;
605 : :
606 : : /* True if warn about differences between C++98 and C++11. */
607 : : bool cpp_warn_cxx11_compat;
608 : :
609 : : /* True if warn about differences between C++17 and C++20. */
610 : : bool cpp_warn_cxx20_compat;
611 : :
612 : : /* Nonzero if bidirectional control characters checking is on. See enum
613 : : cpp_bidirectional_level. */
614 : : unsigned char cpp_warn_bidirectional;
615 : :
616 : : /* True if libcpp should warn about invalid UTF-8 characters in comments.
617 : : 2 if it should be a pedwarn. */
618 : : unsigned char cpp_warn_invalid_utf8;
619 : :
620 : : /* True if libcpp should warn about invalid forms of delimited or named
621 : : escape sequences. */
622 : : bool cpp_warn_unicode;
623 : :
624 : : /* True if -finput-charset= option has been used explicitly. */
625 : : bool cpp_input_charset_explicit;
626 : :
627 : : /* True if -Wkeyword-macro. */
628 : : bool cpp_warn_keyword_macro;
629 : :
630 : : /* -Wleading-whitespace= value. */
631 : : unsigned char cpp_warn_leading_whitespace;
632 : :
633 : : /* -Wtrailing-whitespace= value. */
634 : : unsigned char cpp_warn_trailing_whitespace;
635 : :
636 : : /* -ftabstop= value. */
637 : : unsigned int cpp_tabstop;
638 : :
639 : : /* Dependency generation. */
640 : : struct
641 : : {
642 : : /* Style of header dependencies to generate. */
643 : : enum cpp_deps_style style;
644 : :
645 : : /* Structured format of module dependencies to generate. */
646 : : enum cpp_fdeps_format fdeps_format;
647 : :
648 : : /* Assume missing files are generated files. */
649 : : bool missing_files;
650 : :
651 : : /* Generate phony targets for each dependency apart from the first
652 : : one. */
653 : : bool phony_targets;
654 : :
655 : : /* Generate dependency info for modules. */
656 : : bool modules;
657 : :
658 : : /* If true, no dependency is generated on the main file. */
659 : : bool ignore_main_file;
660 : :
661 : : /* If true, intend to use the preprocessor output (e.g., for compilation)
662 : : in addition to the dependency info. */
663 : : bool need_preprocessor_output;
664 : : } deps;
665 : :
666 : : /* Target-specific features set by the front end or client. */
667 : :
668 : : /* Precision for target CPP arithmetic, target characters, target
669 : : ints and target wide characters, respectively. */
670 : : size_t precision, char_precision, int_precision, wchar_precision;
671 : :
672 : : /* True means chars (wide chars, UTF-8 chars) are unsigned. */
673 : : bool unsigned_char, unsigned_wchar, unsigned_utf8char;
674 : :
675 : : /* True if the most significant byte in a word has the lowest
676 : : address in memory. */
677 : : bool bytes_big_endian;
678 : :
679 : : /* Nonzero means __STDC__ should have the value 0 in system headers. */
680 : : unsigned char stdc_0_in_system_headers;
681 : :
682 : : /* True disables tokenization outside of preprocessing directives. */
683 : : bool directives_only;
684 : :
685 : : /* True enables canonicalization of system header file paths. */
686 : : bool canonical_system_headers;
687 : :
688 : : /* The maximum depth of the nested #include. */
689 : : unsigned int max_include_depth;
690 : :
691 : : cpp_main_search main_search : 8;
692 : : };
693 : :
694 : : #if GCC_VERSION >= 3005
695 : : #define ATTRIBUTE_CPP_PPDIAG(m, n) \
696 : : __attribute__ ((__format__ (__gcc_diag__, m , n))) ATTRIBUTE_NONNULL(m)
697 : : #else
698 : : #define ATTRIBUTE_CPP_PPDIAG(m, n) ATTRIBUTE_NONNULL(m)
699 : : #endif
700 : :
701 : : /* Diagnostic levels. To get a diagnostic without associating a
702 : : position in the translation unit with it, use cpp_error_with_line
703 : : with a line number of zero. */
704 : :
705 : : enum cpp_diagnostic_level {
706 : : /* Warning, an error with -Werror. */
707 : : CPP_DL_WARNING = 0,
708 : : /* Same as CPP_DL_WARNING, except it is not suppressed in system headers. */
709 : : CPP_DL_WARNING_SYSHDR,
710 : : /* Warning, an error with -pedantic-errors or -Werror. */
711 : : CPP_DL_PEDWARN,
712 : : /* An error. */
713 : : CPP_DL_ERROR,
714 : : /* An internal consistency check failed. Prints "internal error: ",
715 : : otherwise the same as CPP_DL_ERROR. */
716 : : CPP_DL_ICE,
717 : : /* An informative note following a warning. */
718 : : CPP_DL_NOTE,
719 : : /* A fatal error. */
720 : : CPP_DL_FATAL
721 : : };
722 : :
723 : : /* Warning reason codes. Use a reason code of CPP_W_NONE for unclassified
724 : : warnings and diagnostics that are not warnings. */
725 : :
726 : : enum cpp_warning_reason {
727 : : CPP_W_NONE = 0,
728 : : CPP_W_DEPRECATED,
729 : : CPP_W_COMMENTS,
730 : : CPP_W_MISSING_INCLUDE_DIRS,
731 : : CPP_W_TRIGRAPHS,
732 : : CPP_W_MULTICHAR,
733 : : CPP_W_TRADITIONAL,
734 : : CPP_W_LONG_LONG,
735 : : CPP_W_ENDIF_LABELS,
736 : : CPP_W_NUM_SIGN_CHANGE,
737 : : CPP_W_VARIADIC_MACROS,
738 : : CPP_W_BUILTIN_MACRO_REDEFINED,
739 : : CPP_W_DOLLARS,
740 : : CPP_W_UNDEF,
741 : : CPP_W_UNUSED_MACROS,
742 : : CPP_W_CXX_OPERATOR_NAMES,
743 : : CPP_W_NORMALIZE,
744 : : CPP_W_INVALID_PCH,
745 : : CPP_W_WARNING_DIRECTIVE,
746 : : CPP_W_LITERAL_SUFFIX,
747 : : CPP_W_SIZE_T_LITERALS,
748 : : CPP_W_DATE_TIME,
749 : : CPP_W_PEDANTIC,
750 : : CPP_W_C90_C99_COMPAT,
751 : : CPP_W_C11_C23_COMPAT,
752 : : CPP_W_C23_C2Y_COMPAT,
753 : : CPP_W_CXX11_COMPAT,
754 : : CPP_W_CXX20_COMPAT,
755 : : CPP_W_CXX14_EXTENSIONS,
756 : : CPP_W_CXX17_EXTENSIONS,
757 : : CPP_W_CXX20_EXTENSIONS,
758 : : CPP_W_CXX23_EXTENSIONS,
759 : : CPP_W_CXX26_EXTENSIONS,
760 : : CPP_W_EXPANSION_TO_DEFINED,
761 : : CPP_W_BIDIRECTIONAL,
762 : : CPP_W_INVALID_UTF8,
763 : : CPP_W_UNICODE,
764 : : CPP_W_HEADER_GUARD,
765 : : CPP_W_PRAGMA_ONCE_OUTSIDE_HEADER,
766 : : CPP_W_LEADING_WHITESPACE,
767 : : CPP_W_TRAILING_WHITESPACE,
768 : : CPP_W_KEYWORD_MACRO
769 : : };
770 : :
771 : : /* Callback for header lookup for HEADER, which is the name of a
772 : : source file. It is used as a method of last resort to find headers
773 : : that are not otherwise found during the normal include processing.
774 : : The return value is the malloced name of a header to try and open,
775 : : if any, or NULL otherwise. This callback is called only if the
776 : : header is otherwise unfound. */
777 : : typedef const char *(*missing_header_cb)(cpp_reader *, const char *header, cpp_dir **);
778 : :
779 : : /* Call backs to cpplib client. */
780 : : struct cpp_callbacks
781 : : {
782 : : /* Called when a new line of preprocessed output is started. */
783 : : void (*line_change) (cpp_reader *, const cpp_token *, int);
784 : :
785 : : /* Called when switching to/from a new file.
786 : : The line_map is for the new file. It is NULL if there is no new file.
787 : : (In C this happens when done with <built-in>+<command line> and also
788 : : when done with a main file.) This can be used for resource cleanup. */
789 : : void (*file_change) (cpp_reader *, const line_map_ordinary *);
790 : :
791 : : void (*dir_change) (cpp_reader *, const char *);
792 : : void (*include) (cpp_reader *, location_t, const unsigned char *,
793 : : const char *, int, const cpp_token **);
794 : : void (*define) (cpp_reader *, location_t, cpp_hashnode *);
795 : : void (*undef) (cpp_reader *, location_t, cpp_hashnode *);
796 : : void (*ident) (cpp_reader *, location_t, const cpp_string *);
797 : : void (*def_pragma) (cpp_reader *, location_t);
798 : : int (*valid_pch) (cpp_reader *, const char *, int);
799 : : void (*read_pch) (cpp_reader *, const char *, int, const char *);
800 : : missing_header_cb missing_header;
801 : :
802 : : /* Context-sensitive macro support. Returns macro (if any) that should
803 : : be expanded. */
804 : : cpp_hashnode * (*macro_to_expand) (cpp_reader *, const cpp_token *);
805 : :
806 : : /* Called to emit a diagnostic. This callback receives the
807 : : translated message. */
808 : : bool (*diagnostic) (cpp_reader *,
809 : : enum cpp_diagnostic_level,
810 : : enum cpp_warning_reason,
811 : : rich_location *,
812 : : const char *, va_list *)
813 : : ATTRIBUTE_CPP_PPDIAG (5,0);
814 : :
815 : : /* Callbacks for when a macro is expanded, or tested (whether
816 : : defined or not at the time) in #ifdef, #ifndef or "defined". */
817 : : void (*used_define) (cpp_reader *, location_t, cpp_hashnode *);
818 : : void (*used_undef) (cpp_reader *, location_t, cpp_hashnode *);
819 : : /* Called before #define and #undef or other macro definition
820 : : changes are processed. */
821 : : void (*before_define) (cpp_reader *);
822 : : /* Called whenever a macro is expanded or tested.
823 : : Second argument is the location of the start of the current expansion. */
824 : : void (*used) (cpp_reader *, location_t, cpp_hashnode *);
825 : :
826 : : /* Callback to identify whether an attribute exists. */
827 : : int (*has_attribute) (cpp_reader *, bool);
828 : :
829 : : /* Callback to determine whether a built-in function is recognized. */
830 : : int (*has_builtin) (cpp_reader *);
831 : :
832 : : /* Callback to determine whether a feature is available. */
833 : : int (*has_feature) (cpp_reader *, bool);
834 : :
835 : : /* Callback that can change a user lazy into normal macro. */
836 : : void (*user_lazy_macro) (cpp_reader *, cpp_macro *, unsigned);
837 : :
838 : : /* Callback to handle deferred cpp_macros. */
839 : : cpp_macro *(*user_deferred_macro) (cpp_reader *, location_t, cpp_hashnode *);
840 : :
841 : : /* Callback to parse SOURCE_DATE_EPOCH from environment. */
842 : : time_t (*get_source_date_epoch) (cpp_reader *);
843 : :
844 : : /* Callback for providing suggestions for misspelled directives. */
845 : : const char *(*get_suggestion) (cpp_reader *, const char *, const char *const *);
846 : :
847 : : /* Callback for when a comment is encountered, giving the location
848 : : of the opening slash, a pointer to the content (which is not
849 : : necessarily 0-terminated), and the length of the content.
850 : : The content contains the opening slash-star (or slash-slash),
851 : : and for C-style comments contains the closing star-slash. For
852 : : C++-style comments it does not include the terminating newline. */
853 : : void (*comment) (cpp_reader *, location_t, const unsigned char *,
854 : : size_t);
855 : :
856 : : /* Callback for filename remapping in __FILE__ and __BASE_FILE__ macro
857 : : expansions. */
858 : : const char *(*remap_filename) (const char*);
859 : :
860 : : /* Maybe translate a #include into something else. Return a
861 : : cpp_buffer containing the translation if translating. */
862 : : char *(*translate_include) (cpp_reader *, line_maps *, location_t,
863 : : const char *path);
864 : : };
865 : :
866 : : #ifdef VMS
867 : : #define INO_T_CPP ino_t ino[3]
868 : : #elif defined (_AIX) && SIZEOF_INO_T == 4
869 : : #define INO_T_CPP ino64_t ino
870 : : #else
871 : : #define INO_T_CPP ino_t ino
872 : : #endif
873 : :
874 : : #if defined (_AIX) && SIZEOF_DEV_T == 4
875 : : #define DEV_T_CPP dev64_t dev
876 : : #else
877 : : #define DEV_T_CPP dev_t dev
878 : : #endif
879 : :
880 : : /* Chain of directories to look for include files in. */
881 : : struct cpp_dir
882 : : {
883 : : /* NULL-terminated singly-linked list. */
884 : : struct cpp_dir *next;
885 : :
886 : : /* NAME of the directory, NUL-terminated. */
887 : : char *name;
888 : : unsigned int len;
889 : :
890 : : /* One if a system header, two if a system header that has extern
891 : : "C" guards for C++. */
892 : : unsigned char sysp;
893 : :
894 : : /* Is this a user-supplied directory? */
895 : : bool user_supplied_p;
896 : :
897 : : /* The canonicalized NAME as determined by lrealpath. This field
898 : : is only used by hosts that lack reliable inode numbers. */
899 : : char *canonical_name;
900 : :
901 : : /* Mapping of file names for this directory for MS-DOS and related
902 : : platforms. A NULL-terminated array of (from, to) pairs. */
903 : : const char **name_map;
904 : :
905 : : /* Routine to construct pathname, given the search path name and the
906 : : HEADER we are trying to find, return a constructed pathname to
907 : : try and open. If this is NULL, the constructed pathname is as
908 : : constructed by append_file_to_dir. */
909 : : char *(*construct) (const char *header, cpp_dir *dir);
910 : :
911 : : /* The C front end uses these to recognize duplicated
912 : : directories in the search path. */
913 : : INO_T_CPP;
914 : : DEV_T_CPP;
915 : : };
916 : :
917 : : /* The kind of the cpp_macro. */
918 : : enum cpp_macro_kind {
919 : : cmk_macro, /* An ISO macro (token expansion). */
920 : : cmk_assert, /* An assertion. */
921 : : cmk_traditional /* A traditional macro (text expansion). */
922 : : };
923 : :
924 : : /* Each macro definition is recorded in a cpp_macro structure.
925 : : Variadic macros cannot occur with traditional cpp. */
926 : : struct GTY(()) cpp_macro {
927 : : union cpp_parm_u
928 : : {
929 : : /* Parameters, if any. If parameter names use extended identifiers,
930 : : the original spelling of those identifiers, not the canonical
931 : : UTF-8 spelling, goes here. */
932 : : cpp_hashnode ** GTY ((tag ("false"),
933 : : nested_ptr (union tree_node,
934 : : "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
935 : : "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL"),
936 : : length ("%1.paramc"))) params;
937 : :
938 : : /* If this is an assertion, the next one in the chain. */
939 : : cpp_macro *GTY ((tag ("true"))) next;
940 : : } GTY ((desc ("%1.kind == cmk_assert"))) parm;
941 : :
942 : : /* Definition line number. */
943 : : location_t line;
944 : :
945 : : /* Number of tokens in body, or bytes for traditional macros. */
946 : : /* Do we really need 2^32-1 range here? */
947 : : unsigned int count;
948 : :
949 : : /* Number of parameters. */
950 : : unsigned short paramc;
951 : :
952 : : /* Non-zero if this is a user-lazy macro, value provided by user. */
953 : : unsigned char lazy;
954 : :
955 : : /* The kind of this macro (ISO, trad or assert) */
956 : : unsigned kind : 2;
957 : :
958 : : /* If a function-like macro. */
959 : : unsigned int fun_like : 1;
960 : :
961 : : /* If a variadic macro. */
962 : : unsigned int variadic : 1;
963 : :
964 : : /* If macro defined in system header. */
965 : : unsigned int syshdr : 1;
966 : :
967 : : /* Nonzero if it has been expanded or had its existence tested. */
968 : : unsigned int used : 1;
969 : :
970 : : /* Indicate whether the tokens include extra CPP_PASTE tokens at the
971 : : end to track invalid redefinitions with consecutive CPP_PASTE
972 : : tokens. */
973 : : unsigned int extra_tokens : 1;
974 : :
975 : : /* Imported C++20 macro (from a header unit). */
976 : : unsigned int imported_p : 1;
977 : :
978 : : /* 0 bits spare (32-bit). 32 on 64-bit target. */
979 : :
980 : : union cpp_exp_u
981 : : {
982 : : /* Trailing array of replacement tokens (ISO), or assertion body value. */
983 : : cpp_token GTY ((tag ("false"), length ("%1.count"))) tokens[1];
984 : :
985 : : /* Pointer to replacement text (traditional). See comment at top
986 : : of cpptrad.c for how traditional function-like macros are
987 : : encoded. */
988 : : const unsigned char *GTY ((tag ("true"))) text;
989 : : } GTY ((desc ("%1.kind == cmk_traditional"))) exp;
990 : : };
991 : :
992 : : /* Poisoned identifiers are flagged NODE_POISONED. NODE_OPERATOR (C++
993 : : only) indicates an identifier that behaves like an operator such as
994 : : "xor". NODE_DIAGNOSTIC is for speed in lex_token: it indicates a
995 : : diagnostic may be required for this node. Currently this only
996 : : applies to __VA_ARGS__, poisoned identifiers, and -Wc++-compat
997 : : warnings about NODE_OPERATOR. */
998 : :
999 : : /* Hash node flags. */
1000 : : #define NODE_OPERATOR (1 << 0) /* C++ named operator. */
1001 : : #define NODE_POISONED (1 << 1) /* Poisoned identifier. */
1002 : : #define NODE_DIAGNOSTIC (1 << 2) /* Possible diagnostic when lexed. */
1003 : : #define NODE_WARN (1 << 3) /* Warn if redefined or undefined. */
1004 : : #define NODE_DISABLED (1 << 4) /* A disabled macro. */
1005 : : #define NODE_USED (1 << 5) /* Dumped with -dU. */
1006 : : #define NODE_CONDITIONAL (1 << 6) /* Conditional macro */
1007 : : #define NODE_WARN_OPERATOR (1 << 7) /* Warn about C++ named operator. */
1008 : : #define NODE_MODULE (1 << 8) /* C++-20 module-related name. */
1009 : :
1010 : : /* Different flavors of hash node. */
1011 : : enum node_type
1012 : : {
1013 : : NT_VOID = 0, /* Maybe an assert? */
1014 : : NT_MACRO_ARG, /* A macro arg. */
1015 : : NT_USER_MACRO, /* A user macro. */
1016 : : NT_BUILTIN_MACRO, /* A builtin macro. */
1017 : : NT_MACRO_MASK = NT_USER_MACRO /* Mask for either macro kind. */
1018 : : };
1019 : :
1020 : : /* Different flavors of builtin macro. _Pragma is an operator, but we
1021 : : handle it with the builtin code for efficiency reasons. */
1022 : : enum cpp_builtin_type
1023 : : {
1024 : : BT_SPECLINE = 0, /* `__LINE__' */
1025 : : BT_DATE, /* `__DATE__' */
1026 : : BT_FILE, /* `__FILE__' */
1027 : : BT_FILE_NAME, /* `__FILE_NAME__' */
1028 : : BT_BASE_FILE, /* `__BASE_FILE__' */
1029 : : BT_INCLUDE_LEVEL, /* `__INCLUDE_LEVEL__' */
1030 : : BT_TIME, /* `__TIME__' */
1031 : : BT_STDC, /* `__STDC__' */
1032 : : BT_PRAGMA, /* `_Pragma' operator */
1033 : : BT_TIMESTAMP, /* `__TIMESTAMP__' */
1034 : : BT_COUNTER, /* `__COUNTER__' */
1035 : : BT_HAS_ATTRIBUTE, /* `__has_attribute(x)' */
1036 : : BT_HAS_STD_ATTRIBUTE, /* `__has_c_attribute(x)' */
1037 : : BT_HAS_BUILTIN, /* `__has_builtin(x)' */
1038 : : BT_HAS_INCLUDE, /* `__has_include(x)' */
1039 : : BT_HAS_INCLUDE_NEXT, /* `__has_include_next(x)' */
1040 : : BT_HAS_EMBED, /* `__has_embed(x)' */
1041 : : BT_HAS_FEATURE, /* `__has_feature(x)' */
1042 : : BT_HAS_EXTENSION /* `__has_extension(x)' */
1043 : : };
1044 : :
1045 : : #define CPP_HASHNODE(HNODE) ((cpp_hashnode *) (HNODE))
1046 : : #define HT_NODE(NODE) (&(NODE)->ident)
1047 : : #define NODE_LEN(NODE) HT_LEN (HT_NODE (NODE))
1048 : : #define NODE_NAME(NODE) HT_STR (HT_NODE (NODE))
1049 : :
1050 : : /* The common part of an identifier node shared amongst all 3 C front
1051 : : ends. Also used to store CPP identifiers, which are a superset of
1052 : : identifiers in the grammatical sense. */
1053 : :
1054 : : union GTY(()) _cpp_hashnode_value {
1055 : : /* Assert (maybe NULL) */
1056 : : cpp_macro * GTY((tag ("NT_VOID"))) answers;
1057 : : /* Macro (maybe NULL) */
1058 : : cpp_macro * GTY((tag ("NT_USER_MACRO"))) macro;
1059 : : /* Code for a builtin macro. */
1060 : : enum cpp_builtin_type GTY ((tag ("NT_BUILTIN_MACRO"))) builtin;
1061 : : /* Macro argument index. */
1062 : : unsigned short GTY ((tag ("NT_MACRO_ARG"))) arg_index;
1063 : : };
1064 : :
1065 : : struct GTY(()) cpp_hashnode {
1066 : : struct ht_identifier ident;
1067 : : unsigned int is_directive : 1;
1068 : : unsigned int directive_index : 7; /* If is_directive,
1069 : : then index into directive table.
1070 : : Otherwise, a NODE_OPERATOR. */
1071 : : unsigned int rid_code : 8; /* Rid code - for front ends. */
1072 : : unsigned int flags : 9; /* CPP flags. */
1073 : : ENUM_BITFIELD(node_type) type : 2; /* CPP node type. */
1074 : :
1075 : : /* 5 bits spare. */
1076 : :
1077 : : /* The deferred cookie is applicable to NT_USER_MACRO or NT_VOID.
1078 : : The latter for when a macro had a prevailing undef.
1079 : : On a 64-bit system there would be 32-bits of padding to the value
1080 : : field. So placing the deferred index here is not costly. */
1081 : : unsigned deferred; /* Deferred cookie */
1082 : :
1083 : : union _cpp_hashnode_value GTY ((desc ("%1.type"))) value;
1084 : : };
1085 : :
1086 : : /* Extra information we may need to store per identifier, which is needed rarely
1087 : : enough that it's not worth adding directly into the main identifier hash. */
1088 : : struct GTY(()) cpp_hashnode_extra
1089 : : {
1090 : : struct ht_identifier ident;
1091 : : location_t poisoned_loc;
1092 : : };
1093 : :
1094 : : /* A class for iterating through the source locations within a
1095 : : string token (before escapes are interpreted, and before
1096 : : concatenation). */
1097 : :
1098 : : class cpp_string_location_reader {
1099 : : public:
1100 : : cpp_string_location_reader (location_t src_loc,
1101 : : line_maps *line_table);
1102 : :
1103 : : source_range get_next ();
1104 : :
1105 : : private:
1106 : : location_t m_loc;
1107 : : int m_offset_per_column;
1108 : : };
1109 : :
1110 : : /* A class for storing the source ranges of all of the characters within
1111 : : a string literal, after escapes are interpreted, and after
1112 : : concatenation.
1113 : :
1114 : : This is not GTY-marked, as instances are intended to be temporary. */
1115 : :
1116 : : class cpp_substring_ranges
1117 : : {
1118 : : public:
1119 : : cpp_substring_ranges ();
1120 : : ~cpp_substring_ranges ();
1121 : :
1122 : 28130 : int get_num_ranges () const { return m_num_ranges; }
1123 : 44434 : source_range get_range (int idx) const
1124 : : {
1125 : 44434 : linemap_assert (idx < m_num_ranges);
1126 : 44434 : return m_ranges[idx];
1127 : : }
1128 : :
1129 : : void add_range (source_range range);
1130 : : void add_n_ranges (int num, cpp_string_location_reader &loc_reader);
1131 : :
1132 : : private:
1133 : : source_range *m_ranges;
1134 : : int m_num_ranges;
1135 : : int m_alloc_ranges;
1136 : : };
1137 : :
1138 : : /* Call this first to get a handle to pass to other functions.
1139 : :
1140 : : The first hash table argument is for associating a struct cpp_hashnode
1141 : : with each identifier. The second hash table argument is for associating
1142 : : a struct cpp_hashnode_extra with each identifier that needs one. For
1143 : : either, pass in a NULL pointer if you want cpplib to create and manage
1144 : : the hash table itself, or else pass a suitably initialized hash table to
1145 : : be managed external to libcpp, as is done by the C-family frontends. */
1146 : : extern cpp_reader *cpp_create_reader (enum c_lang, struct ht *,
1147 : : class line_maps *,
1148 : : struct ht * = nullptr);
1149 : :
1150 : : /* Reset the cpp_reader's line_map. This is only used after reading a
1151 : : PCH file. */
1152 : : extern void cpp_set_line_map (cpp_reader *, class line_maps *);
1153 : :
1154 : : /* Call this to change the selected language standard (e.g. because of
1155 : : command line options). */
1156 : : extern void cpp_set_lang (cpp_reader *, enum c_lang);
1157 : :
1158 : : /* Set the include paths. */
1159 : : extern void cpp_set_include_chains (cpp_reader *, cpp_dir *, cpp_dir *,
1160 : : cpp_dir *, int);
1161 : :
1162 : : /* Call these to get pointers to the options, callback, and deps
1163 : : structures for a given reader. These pointers are good until you
1164 : : call cpp_finish on that reader. You can either edit the callbacks
1165 : : through the pointer returned from cpp_get_callbacks, or set them
1166 : : with cpp_set_callbacks. */
1167 : : extern cpp_options *cpp_get_options (cpp_reader *) ATTRIBUTE_PURE;
1168 : : extern cpp_callbacks *cpp_get_callbacks (cpp_reader *) ATTRIBUTE_PURE;
1169 : : extern void cpp_set_callbacks (cpp_reader *, cpp_callbacks *);
1170 : : extern class mkdeps *cpp_get_deps (cpp_reader *) ATTRIBUTE_PURE;
1171 : :
1172 : : extern const char *cpp_probe_header_unit (cpp_reader *, const char *file,
1173 : : bool angle_p, location_t);
1174 : :
1175 : : /* Call these to get name data about the various compile-time
1176 : : charsets. */
1177 : : extern const char *cpp_get_narrow_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1178 : : extern const char *cpp_get_wide_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1179 : :
1180 : : extern location_t cpp_get_diagnostic_override_loc (const cpp_reader *);
1181 : :
1182 : : /* This function reads the file, but does not start preprocessing. It
1183 : : returns the name of the original file; this is the same as the
1184 : : input file, except for preprocessed input. This will generate at
1185 : : least one file change callback, and possibly a line change callback
1186 : : too. If there was an error opening the file, it returns NULL. */
1187 : : extern const char *cpp_read_main_file (cpp_reader *, const char *,
1188 : : bool injecting = false);
1189 : : extern location_t cpp_main_loc (const cpp_reader *);
1190 : :
1191 : : /* Adjust for the main file to be an include. */
1192 : : extern void cpp_retrofit_as_include (cpp_reader *);
1193 : :
1194 : : /* Set up built-ins with special behavior. Use cpp_init_builtins()
1195 : : instead unless your know what you are doing. */
1196 : : extern void cpp_init_special_builtins (cpp_reader *);
1197 : :
1198 : : /* Set up built-ins like __FILE__. */
1199 : : extern void cpp_init_builtins (cpp_reader *, int);
1200 : :
1201 : : /* This is called after options have been parsed, and partially
1202 : : processed. */
1203 : : extern void cpp_post_options (cpp_reader *);
1204 : :
1205 : : /* Set up translation to the target character set. */
1206 : : extern void cpp_init_iconv (cpp_reader *);
1207 : :
1208 : : /* Call this to finish preprocessing. If you requested dependency
1209 : : generation, pass open stream(s) to write the information to,
1210 : : otherwise NULL. It is your responsibility to close the stream(s). */
1211 : : extern void cpp_finish (cpp_reader *, FILE *deps_stream, FILE *fdeps_stream = NULL);
1212 : :
1213 : : /* Call this to release the handle at the end of preprocessing. Any
1214 : : use of the handle after this function returns is invalid. */
1215 : : extern void cpp_destroy (cpp_reader *);
1216 : :
1217 : : extern unsigned int cpp_token_len (const cpp_token *);
1218 : : extern unsigned char *cpp_token_as_text (cpp_reader *, const cpp_token *);
1219 : : extern unsigned char *cpp_spell_token (cpp_reader *, const cpp_token *,
1220 : : unsigned char *, bool);
1221 : : extern void cpp_register_pragma (cpp_reader *, const char *, const char *,
1222 : : void (*) (cpp_reader *), bool);
1223 : : extern void cpp_register_deferred_pragma (cpp_reader *, const char *,
1224 : : const char *, unsigned, bool, bool);
1225 : : extern int cpp_avoid_paste (cpp_reader *, const cpp_token *,
1226 : : const cpp_token *);
1227 : : extern const cpp_token *cpp_get_token (cpp_reader *);
1228 : : extern const cpp_token *cpp_get_token_with_location (cpp_reader *,
1229 : : location_t *);
1230 : 7557302 : inline bool cpp_user_macro_p (const cpp_hashnode *node)
1231 : : {
1232 : 7557302 : return node->type == NT_USER_MACRO;
1233 : : }
1234 : : inline bool cpp_builtin_macro_p (const cpp_hashnode *node)
1235 : : {
1236 : : return node->type == NT_BUILTIN_MACRO;
1237 : : }
1238 : 18253432 : inline bool cpp_macro_p (const cpp_hashnode *node)
1239 : : {
1240 : 18253432 : return node->type & NT_MACRO_MASK;
1241 : : }
1242 : 72038 : inline cpp_macro *cpp_set_deferred_macro (cpp_hashnode *node,
1243 : : cpp_macro *forced = NULL)
1244 : : {
1245 : 72038 : cpp_macro *old = node->value.macro;
1246 : :
1247 : 72038 : node->value.macro = forced;
1248 : 72038 : node->type = NT_USER_MACRO;
1249 : 72038 : node->flags &= ~NODE_USED;
1250 : :
1251 : 72038 : return old;
1252 : : }
1253 : : cpp_macro *cpp_get_deferred_macro (cpp_reader *, cpp_hashnode *, location_t);
1254 : :
1255 : : /* Returns true if NODE is a function-like user macro. */
1256 : 37 : inline bool cpp_fun_like_macro_p (cpp_hashnode *node)
1257 : : {
1258 : 37 : return cpp_user_macro_p (node) && node->value.macro->fun_like;
1259 : : }
1260 : :
1261 : : /* Return true for nodes marked for -Wkeyword-macro diagnostics. */
1262 : : inline bool cpp_keyword_p (cpp_hashnode *node)
1263 : : {
1264 : : /* As keywords are marked identifiers which don't start with underscore
1265 : : or start with underscore followed by capital letter (except for
1266 : : _Pragma). */
1267 : : return ((node->flags & NODE_WARN)
1268 : : && (NODE_NAME (node)[0] != '_'
1269 : : || (NODE_NAME (node)[1] != '_' && NODE_NAME (node)[1] != 'P')));
1270 : : }
1271 : :
1272 : : extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *);
1273 : : extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *,
1274 : : const cpp_macro *);
1275 : 97 : inline location_t cpp_macro_definition_location (cpp_hashnode *node)
1276 : : {
1277 : 97 : const cpp_macro *macro = node->value.macro;
1278 : 97 : return macro ? macro->line : 0;
1279 : : }
1280 : : /* Return an idempotent time stamp (possibly from SOURCE_DATE_EPOCH). */
1281 : : enum class CPP_time_kind
1282 : : {
1283 : : FIXED = -1, /* Fixed time via source epoch. */
1284 : : DYNAMIC = -2, /* Dynamic via time(2). */
1285 : : UNKNOWN = -3 /* Wibbly wobbly, timey wimey. */
1286 : : };
1287 : : extern CPP_time_kind cpp_get_date (cpp_reader *, time_t *);
1288 : :
1289 : : extern void _cpp_backup_tokens (cpp_reader *, unsigned int);
1290 : : extern const cpp_token *cpp_peek_token (cpp_reader *, int);
1291 : :
1292 : : /* Evaluate a CPP_*CHAR* token. */
1293 : : extern cppchar_t cpp_interpret_charconst (cpp_reader *, const cpp_token *,
1294 : : unsigned int *, int *);
1295 : : /* Evaluate a vector of CPP_*STRING* tokens. */
1296 : : extern bool cpp_interpret_string (cpp_reader *,
1297 : : const cpp_string *, size_t,
1298 : : cpp_string *, enum cpp_ttype);
1299 : : extern const char *cpp_interpret_string_ranges (cpp_reader *pfile,
1300 : : const cpp_string *from,
1301 : : cpp_string_location_reader *,
1302 : : size_t count,
1303 : : cpp_substring_ranges *out,
1304 : : enum cpp_ttype type);
1305 : : extern bool cpp_interpret_string_notranslate (cpp_reader *,
1306 : : const cpp_string *, size_t,
1307 : : cpp_string *, enum cpp_ttype);
1308 : :
1309 : : /* Convert a host character constant to the execution character set. */
1310 : : extern cppchar_t cpp_host_to_exec_charset (cpp_reader *, cppchar_t);
1311 : :
1312 : : /* Used to register macros and assertions, perhaps from the command line.
1313 : : The text is the same as the command line argument. */
1314 : : extern void cpp_define (cpp_reader *, const char *);
1315 : : extern void cpp_define_unused (cpp_reader *, const char *);
1316 : : extern void cpp_define_formatted (cpp_reader *pfile,
1317 : : const char *fmt, ...) ATTRIBUTE_PRINTF_2;
1318 : : extern void cpp_define_formatted_unused (cpp_reader *pfile,
1319 : : const char *fmt,
1320 : : ...) ATTRIBUTE_PRINTF_2;
1321 : : extern void cpp_assert (cpp_reader *, const char *);
1322 : : extern void cpp_undef (cpp_reader *, const char *);
1323 : : extern void cpp_unassert (cpp_reader *, const char *);
1324 : :
1325 : : /* Mark a node as a lazily defined macro. */
1326 : : extern void cpp_define_lazily (cpp_reader *, cpp_hashnode *node, unsigned N);
1327 : :
1328 : : /* Undefine all macros and assertions. */
1329 : : extern void cpp_undef_all (cpp_reader *);
1330 : :
1331 : : extern cpp_buffer *cpp_push_buffer (cpp_reader *, const unsigned char *,
1332 : : size_t, int);
1333 : : extern int cpp_defined (cpp_reader *, const unsigned char *, int);
1334 : :
1335 : : /* A preprocessing number. Code assumes that any unused high bits of
1336 : : the double integer are set to zero. */
1337 : :
1338 : : /* This type has to be equal to unsigned HOST_WIDE_INT, see
1339 : : gcc/c-family/c-lex.cc. */
1340 : : typedef uint64_t cpp_num_part;
1341 : : typedef struct cpp_num cpp_num;
1342 : : struct cpp_num
1343 : : {
1344 : : cpp_num_part high;
1345 : : cpp_num_part low;
1346 : : bool unsignedp; /* True if value should be treated as unsigned. */
1347 : : bool overflow; /* True if the most recent calculation overflowed. */
1348 : : };
1349 : :
1350 : : /* cpplib provides two interfaces for interpretation of preprocessing
1351 : : numbers.
1352 : :
1353 : : cpp_classify_number categorizes numeric constants according to
1354 : : their field (integer, floating point, or invalid), radix (decimal,
1355 : : octal, hexadecimal), and type suffixes. */
1356 : :
1357 : : #define CPP_N_CATEGORY 0x000F
1358 : : #define CPP_N_INVALID 0x0000
1359 : : #define CPP_N_INTEGER 0x0001
1360 : : #define CPP_N_FLOATING 0x0002
1361 : :
1362 : : #define CPP_N_WIDTH 0x00F0
1363 : : #define CPP_N_SMALL 0x0010 /* int, float, short _Fract/Accum */
1364 : : #define CPP_N_MEDIUM 0x0020 /* long, double, long _Fract/_Accum. */
1365 : : #define CPP_N_LARGE 0x0040 /* long long, long double,
1366 : : long long _Fract/Accum. */
1367 : :
1368 : : #define CPP_N_WIDTH_MD 0xF0000 /* machine defined. */
1369 : : #define CPP_N_MD_W 0x10000
1370 : : #define CPP_N_MD_Q 0x20000
1371 : :
1372 : : #define CPP_N_RADIX 0x0F00
1373 : : #define CPP_N_DECIMAL 0x0100
1374 : : #define CPP_N_HEX 0x0200
1375 : : #define CPP_N_OCTAL 0x0400
1376 : : #define CPP_N_BINARY 0x0800
1377 : :
1378 : : #define CPP_N_UNSIGNED 0x1000 /* Properties. */
1379 : : #define CPP_N_IMAGINARY 0x2000
1380 : : #define CPP_N_DFLOAT 0x4000
1381 : : #define CPP_N_DEFAULT 0x8000
1382 : :
1383 : : #define CPP_N_FRACT 0x100000 /* Fract types. */
1384 : : #define CPP_N_ACCUM 0x200000 /* Accum types. */
1385 : : #define CPP_N_FLOATN 0x400000 /* _FloatN types. */
1386 : : #define CPP_N_FLOATNX 0x800000 /* _FloatNx types. */
1387 : :
1388 : : #define CPP_N_USERDEF 0x1000000 /* C++11 user-defined literal. */
1389 : :
1390 : : #define CPP_N_SIZE_T 0x2000000 /* C++23 size_t literal. */
1391 : : #define CPP_N_BFLOAT16 0x4000000 /* std::bfloat16_t type. */
1392 : : #define CPP_N_BITINT 0x8000000 /* C23 _BitInt literal. */
1393 : :
1394 : : #define CPP_N_WIDTH_FLOATN_NX 0xF0000000 /* _FloatN / _FloatNx value
1395 : : of N, divided by 16. */
1396 : : #define CPP_FLOATN_SHIFT 24
1397 : : #define CPP_FLOATN_MAX 0xF0
1398 : :
1399 : : /* Classify a CPP_NUMBER token. The return value is a combination of
1400 : : the flags from the above sets. */
1401 : : extern unsigned cpp_classify_number (cpp_reader *, const cpp_token *,
1402 : : const char **, location_t);
1403 : :
1404 : : /* Return the classification flags for a float suffix. */
1405 : : extern unsigned int cpp_interpret_float_suffix (cpp_reader *, const char *,
1406 : : size_t);
1407 : :
1408 : : /* Return the classification flags for an int suffix. */
1409 : : extern unsigned int cpp_interpret_int_suffix (cpp_reader *, const char *,
1410 : : size_t);
1411 : :
1412 : : /* Evaluate a token classified as category CPP_N_INTEGER. */
1413 : : extern cpp_num cpp_interpret_integer (cpp_reader *, const cpp_token *,
1414 : : unsigned int);
1415 : :
1416 : : /* Sign extend a number, with PRECISION significant bits and all
1417 : : others assumed clear, to fill out a cpp_num structure. */
1418 : : cpp_num cpp_num_sign_extend (cpp_num, size_t);
1419 : :
1420 : : /* Output a diagnostic of some kind. */
1421 : : extern bool cpp_error (cpp_reader *, enum cpp_diagnostic_level,
1422 : : const char *msgid, ...)
1423 : : ATTRIBUTE_CPP_PPDIAG (3, 4);
1424 : : extern bool cpp_warning (cpp_reader *, enum cpp_warning_reason,
1425 : : const char *msgid, ...)
1426 : : ATTRIBUTE_CPP_PPDIAG (3, 4);
1427 : : extern bool cpp_pedwarning (cpp_reader *, enum cpp_warning_reason,
1428 : : const char *msgid, ...)
1429 : : ATTRIBUTE_CPP_PPDIAG (3, 4);
1430 : : extern bool cpp_warning_syshdr (cpp_reader *, enum cpp_warning_reason reason,
1431 : : const char *msgid, ...)
1432 : : ATTRIBUTE_CPP_PPDIAG (3, 4);
1433 : :
1434 : : /* As their counterparts above, but use RICHLOC. */
1435 : : extern bool cpp_warning_at (cpp_reader *, enum cpp_warning_reason,
1436 : : rich_location *richloc, const char *msgid, ...)
1437 : : ATTRIBUTE_CPP_PPDIAG (4, 5);
1438 : : extern bool cpp_pedwarning_at (cpp_reader *, enum cpp_warning_reason,
1439 : : rich_location *richloc, const char *msgid, ...)
1440 : : ATTRIBUTE_CPP_PPDIAG (4, 5);
1441 : :
1442 : : /* Output a diagnostic with "MSGID: " preceding the
1443 : : error string of errno. No location is printed. */
1444 : : extern bool cpp_errno (cpp_reader *, enum cpp_diagnostic_level,
1445 : : const char *msgid);
1446 : : /* Similarly, but with "FILENAME: " instead of "MSGID: ", where
1447 : : the filename is not localized. */
1448 : : extern bool cpp_errno_filename (cpp_reader *, enum cpp_diagnostic_level,
1449 : : const char *filename, location_t loc);
1450 : :
1451 : : /* Same as cpp_error, except additionally specifies a position as a
1452 : : (translation unit) physical line and physical column. If the line is
1453 : : zero, then no location is printed. */
1454 : : extern bool cpp_error_with_line (cpp_reader *, enum cpp_diagnostic_level,
1455 : : location_t, unsigned,
1456 : : const char *msgid, ...)
1457 : : ATTRIBUTE_CPP_PPDIAG (5, 6);
1458 : : extern bool cpp_warning_with_line (cpp_reader *, enum cpp_warning_reason,
1459 : : location_t, unsigned,
1460 : : const char *msgid, ...)
1461 : : ATTRIBUTE_CPP_PPDIAG (5, 6);
1462 : : extern bool cpp_pedwarning_with_line (cpp_reader *, enum cpp_warning_reason,
1463 : : location_t, unsigned,
1464 : : const char *msgid, ...)
1465 : : ATTRIBUTE_CPP_PPDIAG (5, 6);
1466 : : extern bool cpp_warning_with_line_syshdr (cpp_reader *, enum cpp_warning_reason,
1467 : : location_t, unsigned,
1468 : : const char *msgid, ...)
1469 : : ATTRIBUTE_CPP_PPDIAG (5, 6);
1470 : :
1471 : : extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1472 : : location_t src_loc, const char *msgid, ...)
1473 : : ATTRIBUTE_CPP_PPDIAG (4, 5);
1474 : :
1475 : : extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1476 : : rich_location *richloc, const char *msgid, ...)
1477 : : ATTRIBUTE_CPP_PPDIAG (4, 5);
1478 : :
1479 : : /* In lex.cc */
1480 : : extern int cpp_ideq (const cpp_token *, const char *);
1481 : : extern void cpp_output_line (cpp_reader *, FILE *);
1482 : : extern unsigned char *cpp_output_line_to_string (cpp_reader *,
1483 : : const unsigned char *);
1484 : : extern const unsigned char *cpp_alloc_token_string
1485 : : (cpp_reader *, const unsigned char *, unsigned);
1486 : : extern void cpp_output_token (const cpp_token *, FILE *);
1487 : : extern const char *cpp_type2name (enum cpp_ttype, unsigned char flags);
1488 : : /* Returns the value of an escape sequence, truncated to the correct
1489 : : target precision. PSTR points to the input pointer, which is just
1490 : : after the backslash. LIMIT is how much text we have. WIDE is true
1491 : : if the escape sequence is part of a wide character constant or
1492 : : string literal. Handles all relevant diagnostics. */
1493 : : extern cppchar_t cpp_parse_escape (cpp_reader *, const unsigned char ** pstr,
1494 : : const unsigned char *limit, int wide);
1495 : :
1496 : : /* Structure used to hold a comment block at a given location in the
1497 : : source code. */
1498 : :
1499 : : typedef struct
1500 : : {
1501 : : /* Text of the comment including the terminators. */
1502 : : char *comment;
1503 : :
1504 : : /* source location for the given comment. */
1505 : : location_t sloc;
1506 : : } cpp_comment;
1507 : :
1508 : : /* Structure holding all comments for a given cpp_reader. */
1509 : :
1510 : : typedef struct
1511 : : {
1512 : : /* table of comment entries. */
1513 : : cpp_comment *entries;
1514 : :
1515 : : /* number of actual entries entered in the table. */
1516 : : int count;
1517 : :
1518 : : /* number of entries allocated currently. */
1519 : : int allocated;
1520 : : } cpp_comment_table;
1521 : :
1522 : : /* Returns the table of comments encountered by the preprocessor. This
1523 : : table is only populated when pfile->state.save_comments is true. */
1524 : : extern cpp_comment_table *cpp_get_comments (cpp_reader *);
1525 : :
1526 : : /* In hash.c */
1527 : :
1528 : : /* Lookup an identifier in the hashtable. Puts the identifier in the
1529 : : table if it is not already there. */
1530 : : extern cpp_hashnode *cpp_lookup (cpp_reader *, const unsigned char *,
1531 : : unsigned int);
1532 : :
1533 : : /* Set NODE_WARN flag for NAME, such that there will be diagnostics
1534 : : for #define or #undef of NAME. */
1535 : :
1536 : : inline void
1537 : 3989398 : cpp_warn (cpp_reader *pfile, const char *name, unsigned int len)
1538 : : {
1539 : 3989398 : cpp_lookup (pfile, (const unsigned char *) name, len)->flags |= NODE_WARN;
1540 : 3989398 : }
1541 : :
1542 : : inline void
1543 : 310463 : cpp_warn (cpp_reader *pfile, const char *name)
1544 : : {
1545 : 310463 : cpp_warn (pfile, name, strlen (name));
1546 : 310463 : }
1547 : :
1548 : : typedef int (*cpp_cb) (cpp_reader *, cpp_hashnode *, void *);
1549 : : extern void cpp_forall_identifiers (cpp_reader *, cpp_cb, void *);
1550 : :
1551 : : /* In macro.cc */
1552 : : extern void cpp_scan_nooutput (cpp_reader *);
1553 : : extern int cpp_sys_macro_p (cpp_reader *);
1554 : : extern unsigned char *cpp_quote_string (unsigned char *, const unsigned char *,
1555 : : unsigned int);
1556 : : extern bool cpp_compare_macros (const cpp_macro *macro1,
1557 : : const cpp_macro *macro2);
1558 : :
1559 : : /* In files.cc */
1560 : : extern bool cpp_included (cpp_reader *, const char *);
1561 : : extern bool cpp_included_before (cpp_reader *, const char *, location_t);
1562 : : extern void cpp_make_system_header (cpp_reader *, int, int);
1563 : : extern bool cpp_push_include (cpp_reader *, const char *);
1564 : : extern bool cpp_push_default_include (cpp_reader *, const char *);
1565 : : extern void cpp_change_file (cpp_reader *, enum lc_reason, const char *);
1566 : : extern const char *cpp_get_path (struct _cpp_file *);
1567 : : extern cpp_dir *cpp_get_dir (struct _cpp_file *);
1568 : : extern cpp_buffer *cpp_get_buffer (cpp_reader *);
1569 : : extern struct _cpp_file *cpp_get_file (cpp_buffer *);
1570 : : extern cpp_buffer *cpp_get_prev (cpp_buffer *);
1571 : : extern void cpp_clear_file_cache (cpp_reader *);
1572 : :
1573 : : /* cpp_get_converted_source returns the contents of the given file, as it exists
1574 : : after cpplib has read it and converted it from the input charset to the
1575 : : source charset. Return struct will be zero-filled if the data could not be
1576 : : read for any reason. The data starts at the DATA pointer, but the TO_FREE
1577 : : pointer is what should be passed to free(), as there may be an offset. */
1578 : : struct cpp_converted_source
1579 : : {
1580 : : char *to_free;
1581 : : char *data;
1582 : : size_t len;
1583 : : };
1584 : : cpp_converted_source cpp_get_converted_source (const char *fname,
1585 : : const char *input_charset);
1586 : :
1587 : : /* In pch.cc */
1588 : : struct save_macro_data;
1589 : : extern int cpp_save_state (cpp_reader *, FILE *);
1590 : : extern int cpp_write_pch_deps (cpp_reader *, FILE *);
1591 : : extern int cpp_write_pch_state (cpp_reader *, FILE *);
1592 : : extern int cpp_valid_state (cpp_reader *, const char *, int);
1593 : : extern void cpp_prepare_state (cpp_reader *, struct save_macro_data **);
1594 : : extern int cpp_read_state (cpp_reader *, const char *, FILE *,
1595 : : struct save_macro_data *);
1596 : :
1597 : : /* In lex.cc */
1598 : : extern void cpp_force_token_locations (cpp_reader *, location_t);
1599 : : extern void cpp_stop_forcing_token_locations (cpp_reader *);
1600 : : enum CPP_DO_task
1601 : : {
1602 : : CPP_DO_print,
1603 : : CPP_DO_location,
1604 : : CPP_DO_token
1605 : : };
1606 : :
1607 : : extern void cpp_directive_only_process (cpp_reader *pfile,
1608 : : void *data,
1609 : : void (*cb) (cpp_reader *,
1610 : : CPP_DO_task,
1611 : : void *data, ...));
1612 : :
1613 : : /* In expr.cc */
1614 : : extern enum cpp_ttype cpp_userdef_string_remove_type
1615 : : (enum cpp_ttype type);
1616 : : extern enum cpp_ttype cpp_userdef_string_add_type
1617 : : (enum cpp_ttype type);
1618 : : extern enum cpp_ttype cpp_userdef_char_remove_type
1619 : : (enum cpp_ttype type);
1620 : : extern enum cpp_ttype cpp_userdef_char_add_type
1621 : : (enum cpp_ttype type);
1622 : : extern bool cpp_userdef_string_p
1623 : : (enum cpp_ttype type);
1624 : : extern bool cpp_userdef_char_p
1625 : : (enum cpp_ttype type);
1626 : : extern const char * cpp_get_userdef_suffix
1627 : : (const cpp_token *);
1628 : :
1629 : : /* In charset.cc */
1630 : :
1631 : : /* The result of attempting to decode a run of UTF-8 bytes. */
1632 : :
1633 : : struct cpp_decoded_char
1634 : : {
1635 : : const char *m_start_byte;
1636 : : const char *m_next_byte;
1637 : :
1638 : : bool m_valid_ch;
1639 : : cppchar_t m_ch;
1640 : : };
1641 : :
1642 : : /* Information for mapping between code points and display columns.
1643 : :
1644 : : This is a tabstop value, along with a callback for getting the
1645 : : widths of characters. Normally this callback is cpp_wcwidth, but we
1646 : : support other schemes for escaping non-ASCII unicode as a series of
1647 : : ASCII chars when printing the user's source code in
1648 : : gcc/diagnostics/source-printing.cc
1649 : :
1650 : : For example, consider:
1651 : : - the Unicode character U+03C0 "GREEK SMALL LETTER PI" (UTF-8: 0xCF 0x80)
1652 : : - the Unicode character U+1F642 "SLIGHTLY SMILING FACE"
1653 : : (UTF-8: 0xF0 0x9F 0x99 0x82)
1654 : : - the byte 0xBF (a stray trailing byte of a UTF-8 character)
1655 : : Normally U+03C0 would occupy one display column, U+1F642
1656 : : would occupy two display columns, and the stray byte would be
1657 : : printed verbatim as one display column.
1658 : :
1659 : : However when escaping them as unicode code points as "<U+03C0>"
1660 : : and "<U+1F642>" they occupy 8 and 9 display columns respectively,
1661 : : and when escaping them as bytes as "<CF><80>" and "<F0><9F><99><82>"
1662 : : they occupy 8 and 16 display columns respectively. In both cases
1663 : : the stray byte is escaped to <BF> as 4 display columns. */
1664 : :
1665 : : struct cpp_char_column_policy
1666 : : {
1667 : 324773 : cpp_char_column_policy (int tabstop,
1668 : : int (*width_cb) (cppchar_t c))
1669 : 374580 : : m_tabstop (tabstop),
1670 : 374580 : m_undecoded_byte_width (1),
1671 : 374568 : m_width_cb (width_cb)
1672 : : {}
1673 : :
1674 : : int m_tabstop;
1675 : : /* Width in display columns of a stray byte that isn't decodable
1676 : : as UTF-8. */
1677 : : int m_undecoded_byte_width;
1678 : : int (*m_width_cb) (cppchar_t c);
1679 : : };
1680 : :
1681 : : /* A class to manage the state while converting a UTF-8 sequence to cppchar_t
1682 : : and computing the display width one character at a time. */
1683 : : class cpp_display_width_computation {
1684 : : public:
1685 : : cpp_display_width_computation (const char *data, int data_length,
1686 : : const cpp_char_column_policy &policy);
1687 : 3812598 : const char *next_byte () const { return m_next; }
1688 : 2938312 : int bytes_processed () const { return m_next - m_begin; }
1689 : 4049508 : int bytes_left () const { return m_bytes_left; }
1690 : 4049508 : bool done () const { return !bytes_left (); }
1691 : 7029431 : int display_cols_processed () const { return m_display_cols; }
1692 : :
1693 : : int process_next_codepoint (cpp_decoded_char *out);
1694 : : int advance_display_cols (int n);
1695 : :
1696 : : private:
1697 : : const char *const m_begin;
1698 : : const char *m_next;
1699 : : size_t m_bytes_left;
1700 : : const cpp_char_column_policy &m_policy;
1701 : : int m_display_cols;
1702 : : };
1703 : :
1704 : : /* Convenience functions that are simple use cases for class
1705 : : cpp_display_width_computation. Tab characters will be expanded to spaces
1706 : : as determined by POLICY.m_tabstop, and non-printable-ASCII characters
1707 : : will be escaped as per POLICY. */
1708 : :
1709 : : int cpp_byte_column_to_display_column (const char *data, int data_length,
1710 : : int column,
1711 : : const cpp_char_column_policy &policy);
1712 : 73595 : inline int cpp_display_width (const char *data, int data_length,
1713 : : const cpp_char_column_policy &policy)
1714 : : {
1715 : 73595 : return cpp_byte_column_to_display_column (data, data_length, data_length,
1716 : : policy);
1717 : : }
1718 : : int cpp_display_column_to_byte_column (const char *data, int data_length,
1719 : : int display_col,
1720 : : const cpp_char_column_policy &policy);
1721 : : int cpp_wcwidth (cppchar_t c);
1722 : :
1723 : : bool cpp_input_conversion_is_trivial (const char *input_charset);
1724 : : int cpp_check_utf8_bom (const char *data, size_t data_length);
1725 : : bool cpp_valid_utf8_p (const char *data, size_t num_bytes);
1726 : :
1727 : : bool cpp_is_combining_char (cppchar_t c);
1728 : : bool cpp_is_printable_char (cppchar_t c);
1729 : :
1730 : : enum cpp_xid_property {
1731 : : CPP_XID_START = 1,
1732 : : CPP_XID_CONTINUE = 2
1733 : : };
1734 : :
1735 : : unsigned int cpp_check_xid_property (cppchar_t c);
1736 : :
1737 : : /* In errors.cc */
1738 : :
1739 : : /* RAII class to suppress CPP diagnostics in the current scope. */
1740 : : class cpp_auto_suppress_diagnostics
1741 : : {
1742 : : public:
1743 : : explicit cpp_auto_suppress_diagnostics (cpp_reader *pfile);
1744 : : ~cpp_auto_suppress_diagnostics ();
1745 : : private:
1746 : : cpp_reader *const m_pfile;
1747 : : const decltype (cpp_callbacks::diagnostic) m_cb;
1748 : : };
1749 : :
1750 : : #endif /* ! LIBCPP_CPPLIB_H */
|
