1 /* Definitions for CPP library.
2 Copyright (C) 1995-2023 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 /* The first three groups, apart from '=', can appear in preprocessor
42 expressions (+= and -= are used to indicate unary + and - resp.).
43 This allows a lookup table to be implemented in _cpp_parse_expr.
44
45 The first group, to CPP_LAST_EQ, can be immediately followed by an
46 '='. The lexer needs operators ending in '=', like ">>=", to be in
47 the same order as their counterparts without the '=', like ">>".
48
49 See the cpp_operator table optab in expr.cc if you change the order or
50 add or remove anything in the first group. */
51
52 #define TTYPE_TABLE \
53 OP(EQ, "=") \
54 OP(NOT, "!") \
55 OP(GREATER, ">") /* compare */ \
56 OP(LESS, "<") \
57 OP(PLUS, "+") /* math */ \
58 OP(MINUS, "-") \
59 OP(MULT, "*") \
60 OP(DIV, "/") \
61 OP(MOD, "%") \
62 OP(AND, "&") /* bit ops */ \
63 OP(OR, "|") \
64 OP(XOR, "^") \
65 OP(RSHIFT, ">>") \
66 OP(LSHIFT, "<<") \
67 \
68 OP(COMPL, "~") \
69 OP(AND_AND, "&&") /* logical */ \
70 OP(OR_OR, "||") \
71 OP(QUERY, "?") \
72 OP(COLON, ":") \
73 OP(COMMA, ",") /* grouping */ \
74 OP(OPEN_PAREN, "(") \
75 OP(CLOSE_PAREN, ")") \
76 TK(EOF, NONE) \
77 OP(EQ_EQ, "==") /* compare */ \
78 OP(NOT_EQ, "!=") \
79 OP(GREATER_EQ, ">=") \
80 OP(LESS_EQ, "<=") \
81 OP(SPACESHIP, "<=>") \
82 \
83 /* These two are unary + / - in preprocessor expressions. */ \
84 OP(PLUS_EQ, "+=") /* math */ \
85 OP(MINUS_EQ, "-=") \
86 \
87 OP(MULT_EQ, "*=") \
88 OP(DIV_EQ, "/=") \
89 OP(MOD_EQ, "%=") \
90 OP(AND_EQ, "&=") /* bit ops */ \
91 OP(OR_EQ, "|=") \
92 OP(XOR_EQ, "^=") \
93 OP(RSHIFT_EQ, ">>=") \
94 OP(LSHIFT_EQ, "<<=") \
95 /* Digraphs together, beginning with CPP_FIRST_DIGRAPH. */ \
96 OP(HASH, "#") /* digraphs */ \
97 OP(PASTE, "##") \
98 OP(OPEN_SQUARE, "[") \
99 OP(CLOSE_SQUARE, "]") \
100 OP(OPEN_BRACE, "{") \
101 OP(CLOSE_BRACE, "}") \
102 /* The remainder of the punctuation. Order is not significant. */ \
103 OP(SEMICOLON, ";") /* structure */ \
104 OP(ELLIPSIS, "...") \
105 OP(PLUS_PLUS, "++") /* increment */ \
106 OP(MINUS_MINUS, "--") \
107 OP(DEREF, "->") /* accessors */ \
108 OP(DOT, ".") \
109 OP(SCOPE, "::") \
110 OP(DEREF_STAR, "->*") \
111 OP(DOT_STAR, ".*") \
112 OP(ATSIGN, "@") /* used in Objective-C */ \
113 \
114 TK(NAME, IDENT) /* word */ \
115 TK(AT_NAME, IDENT) /* @word - Objective-C */ \
116 TK(NUMBER, LITERAL) /* 34_be+ta */ \
117 \
118 TK(CHAR, LITERAL) /* 'char' */ \
119 TK(WCHAR, LITERAL) /* L'char' */ \
120 TK(CHAR16, LITERAL) /* u'char' */ \
121 TK(CHAR32, LITERAL) /* U'char' */ \
122 TK(UTF8CHAR, LITERAL) /* u8'char' */ \
123 TK(OTHER, LITERAL) /* stray punctuation */ \
124 \
125 TK(STRING, LITERAL) /* "string" */ \
126 TK(WSTRING, LITERAL) /* L"string" */ \
127 TK(STRING16, LITERAL) /* u"string" */ \
128 TK(STRING32, LITERAL) /* U"string" */ \
129 TK(UTF8STRING, LITERAL) /* u8"string" */ \
130 TK(OBJC_STRING, LITERAL) /* @"string" - Objective-C */ \
131 TK(HEADER_NAME, LITERAL) /* <stdio.h> in #include */ \
132 \
133 TK(CHAR_USERDEF, LITERAL) /* 'char'_suffix - C++-0x */ \
134 TK(WCHAR_USERDEF, LITERAL) /* L'char'_suffix - C++-0x */ \
135 TK(CHAR16_USERDEF, LITERAL) /* u'char'_suffix - C++-0x */ \
136 TK(CHAR32_USERDEF, LITERAL) /* U'char'_suffix - C++-0x */ \
137 TK(UTF8CHAR_USERDEF, LITERAL) /* u8'char'_suffix - C++-0x */ \
138 TK(STRING_USERDEF, LITERAL) /* "string"_suffix - C++-0x */ \
139 TK(WSTRING_USERDEF, LITERAL) /* L"string"_suffix - C++-0x */ \
140 TK(STRING16_USERDEF, LITERAL) /* u"string"_suffix - C++-0x */ \
141 TK(STRING32_USERDEF, LITERAL) /* U"string"_suffix - C++-0x */ \
142 TK(UTF8STRING_USERDEF,LITERAL) /* u8"string"_suffix - C++-0x */ \
143 \
144 TK(COMMENT, LITERAL) /* Only if output comments. */ \
145 /* SPELL_LITERAL happens to DTRT. */ \
146 TK(MACRO_ARG, NONE) /* Macro argument. */ \
147 TK(PRAGMA, NONE) /* Only for deferred pragmas. */ \
148 TK(PRAGMA_EOL, NONE) /* End-of-line for deferred pragmas. */ \
149 TK(PADDING, NONE) /* Whitespace for -E. */
150
151 #define OP(e, s) CPP_ ## e,
152 #define TK(e, s) CPP_ ## e,
153 enum cpp_ttype
154 {
155 TTYPE_TABLE
156 N_TTYPES,
157
158 /* A token type for keywords, as opposed to ordinary identifiers. */
159 CPP_KEYWORD,
160
161 /* Positions in the table. */
162 CPP_LAST_EQ = CPP_LSHIFT,
163 CPP_FIRST_DIGRAPH = CPP_HASH,
164 CPP_LAST_PUNCTUATOR= CPP_ATSIGN,
165 CPP_LAST_CPP_OP = CPP_LESS_EQ
166 };
167 #undef OP
168 #undef TK
169
170 /* C language kind, used when calling cpp_create_reader. */
171 enum c_lang {CLK_GNUC89 = 0, CLK_GNUC99, CLK_GNUC11, CLK_GNUC17, CLK_GNUC2X,
172 CLK_STDC89, CLK_STDC94, CLK_STDC99, CLK_STDC11, CLK_STDC17,
173 CLK_STDC2X,
174 CLK_GNUCXX, CLK_CXX98, CLK_GNUCXX11, CLK_CXX11,
175 CLK_GNUCXX14, CLK_CXX14, CLK_GNUCXX17, CLK_CXX17,
176 CLK_GNUCXX20, CLK_CXX20, CLK_GNUCXX23, CLK_CXX23,
177 CLK_ASM};
178
179 /* Payload of a NUMBER, STRING, CHAR or COMMENT token. */
180 struct GTY(()) cpp_string {
181 unsigned int len;
182
183 /* TEXT is always null terminated (terminator not included in len); but this
184 GTY markup arranges that PCH streaming works properly even if there is a
185 null byte in the middle of the string. */
186 const unsigned char * GTY((string_length ("1 + %h.len"))) text;
187 };
188
189 /* Flags for the cpp_token structure. */
190 #define PREV_WHITE (1 << 0) /* If whitespace before this token. */
191 #define DIGRAPH (1 << 1) /* If it was a digraph. */
192 #define STRINGIFY_ARG (1 << 2) /* If macro argument to be stringified. */
193 #define PASTE_LEFT (1 << 3) /* If on LHS of a ## operator. */
194 #define NAMED_OP (1 << 4) /* C++ named operators. */
195 #define PREV_FALLTHROUGH (1 << 5) /* On a token preceeded by FALLTHROUGH
196 comment. */
197 #define DECIMAL_INT (1 << 6) /* Decimal integer, set in c-lex.cc. */
198 #define PURE_ZERO (1 << 7) /* Single 0 digit, used by the C++ frontend,
199 set in c-lex.cc. */
200 #define SP_DIGRAPH (1 << 8) /* # or ## token was a digraph. */
201 #define SP_PREV_WHITE (1 << 9) /* If whitespace before a ##
202 operator, or before this token
203 after a # operator. */
204 #define NO_EXPAND (1 << 10) /* Do not macro-expand this token. */
205 #define PRAGMA_OP (1 << 11) /* _Pragma token. */
206 #define BOL (1 << 12) /* Token at beginning of line. */
207
208 /* Specify which field, if any, of the cpp_token union is used. */
209
210 enum cpp_token_fld_kind {
211 CPP_TOKEN_FLD_NODE,
212 CPP_TOKEN_FLD_SOURCE,
213 CPP_TOKEN_FLD_STR,
214 CPP_TOKEN_FLD_ARG_NO,
215 CPP_TOKEN_FLD_TOKEN_NO,
216 CPP_TOKEN_FLD_PRAGMA,
217 CPP_TOKEN_FLD_NONE
218 };
219
220 /* A macro argument in the cpp_token union. */
221 struct GTY(()) cpp_macro_arg {
222 /* Argument number. */
223 unsigned int arg_no;
224 /* The original spelling of the macro argument token. */
225 cpp_hashnode *
226 GTY ((nested_ptr (union tree_node,
227 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
228 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
229 spelling;
230 };
231
232 /* An identifier in the cpp_token union. */
233 struct GTY(()) cpp_identifier {
234 /* The canonical (UTF-8) spelling of the identifier. */
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 node;
240 /* The original spelling of the identifier. */
241 cpp_hashnode *
242 GTY ((nested_ptr (union tree_node,
243 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
244 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL")))
245 spelling;
246 };
247
248 /* A preprocessing token. This has been carefully packed and should
249 occupy 16 bytes on 32-bit hosts and 24 bytes on 64-bit hosts. */
250 struct GTY(()) cpp_token {
251
252 /* Location of first char of token, together with range of full token. */
253 location_t src_loc;
254
255 ENUM_BITFIELD(cpp_ttype) type : CHAR_BIT; /* token type */
256 unsigned short flags; /* flags - see above */
257
258 union cpp_token_u
259 {
260 /* An identifier. */
261 struct cpp_identifier GTY ((tag ("CPP_TOKEN_FLD_NODE"))) node;
262
263 /* Inherit padding from this token. */
264 cpp_token * GTY ((tag ("CPP_TOKEN_FLD_SOURCE"))) source;
265
266 /* A string, or number. */
267 struct cpp_string GTY ((tag ("CPP_TOKEN_FLD_STR"))) str;
268
269 /* Argument no. (and original spelling) for a CPP_MACRO_ARG. */
270 struct cpp_macro_arg GTY ((tag ("CPP_TOKEN_FLD_ARG_NO"))) macro_arg;
271
272 /* Original token no. for a CPP_PASTE (from a sequence of
273 consecutive paste tokens in a macro expansion). */
274 unsigned int GTY ((tag ("CPP_TOKEN_FLD_TOKEN_NO"))) token_no;
275
276 /* Caller-supplied identifier for a CPP_PRAGMA. */
277 unsigned int GTY ((tag ("CPP_TOKEN_FLD_PRAGMA"))) pragma;
278 } GTY ((desc ("cpp_token_val_index (&%1)"))) val;
279 };
280
281 /* Say which field is in use. */
282 extern enum cpp_token_fld_kind cpp_token_val_index (const cpp_token *tok);
283
284 /* A type wide enough to hold any multibyte source character.
285 cpplib's character constant interpreter requires an unsigned type.
286 Also, a typedef for the signed equivalent.
287 The width of this type is capped at 32 bits; there do exist targets
288 where wchar_t is 64 bits, but only in a non-default mode, and there
289 would be no meaningful interpretation for a wchar_t value greater
290 than 2^32 anyway -- the widest wide-character encoding around is
291 ISO 10646, which stops at 2^31. */
292 #if CHAR_BIT * SIZEOF_INT >= 32
293 # define CPPCHAR_SIGNED_T int
294 #elif CHAR_BIT * SIZEOF_LONG >= 32
295 # define CPPCHAR_SIGNED_T long
296 #else
297 # error "Cannot find a least-32-bit signed integer type"
298 #endif
299 typedef unsigned CPPCHAR_SIGNED_T cppchar_t;
300 typedef CPPCHAR_SIGNED_T cppchar_signed_t;
301
302 /* Style of header dependencies to generate. */
303 enum cpp_deps_style { DEPS_NONE = 0, DEPS_USER, DEPS_SYSTEM };
304
305 /* The possible normalization levels, from most restrictive to least. */
306 enum cpp_normalize_level {
307 /* In NFKC. */
308 normalized_KC = 0,
309 /* In NFC. */
310 normalized_C,
311 /* In NFC, except for subsequences where being in NFC would make
312 the identifier invalid. */
313 normalized_identifier_C,
314 /* Not normalized at all. */
315 normalized_none
316 };
317
318 enum cpp_main_search
319 {
320 CMS_none, /* A regular source file. */
321 CMS_header, /* Is a directly-specified header file (eg PCH or
322 header-unit). */
323 CMS_user, /* Search the user INCLUDE path. */
324 CMS_system, /* Search the system INCLUDE path. */
325 };
326
327 /* The possible bidirectional control characters checking levels. */
328 enum cpp_bidirectional_level {
329 /* No checking. */
330 bidirectional_none = 0,
331 /* Only detect unpaired uses of bidirectional control characters. */
332 bidirectional_unpaired = 1,
333 /* Detect any use of bidirectional control characters. */
334 bidirectional_any = 2,
335 /* Also warn about UCNs. */
336 bidirectional_ucn = 4
337 };
338
339 /* This structure is nested inside struct cpp_reader, and
340 carries all the options visible to the command line. */
341 struct cpp_options
342 {
343 /* The language we're preprocessing. */
344 enum c_lang lang;
345
346 /* Nonzero means use extra default include directories for C++. */
347 unsigned char cplusplus;
348
349 /* Nonzero means handle cplusplus style comments. */
350 unsigned char cplusplus_comments;
351
352 /* Nonzero means define __OBJC__, treat @ as a special token, use
353 the OBJC[PLUS]_INCLUDE_PATH environment variable, and allow
354 "#import". */
355 unsigned char objc;
356
357 /* Nonzero means don't copy comments into the output file. */
358 unsigned char discard_comments;
359
360 /* Nonzero means don't copy comments into the output file during
361 macro expansion. */
362 unsigned char discard_comments_in_macro_exp;
363
364 /* Nonzero means process the ISO trigraph sequences. */
365 unsigned char trigraphs;
366
367 /* Nonzero means process the ISO digraph sequences. */
368 unsigned char digraphs;
369
370 /* Nonzero means to allow hexadecimal floats and LL suffixes. */
371 unsigned char extended_numbers;
372
373 /* Nonzero means process u/U prefix literals (UTF-16/32). */
374 unsigned char uliterals;
375
376 /* Nonzero means process u8 prefixed character literals (UTF-8). */
377 unsigned char utf8_char_literals;
378
379 /* Nonzero means process r/R raw strings. If this is set, uliterals
380 must be set as well. */
381 unsigned char rliterals;
382
383 /* Nonzero means print names of header files (-H). */
384 unsigned char print_include_names;
385
386 /* Nonzero means complain about deprecated features. */
387 unsigned char cpp_warn_deprecated;
388
389 /* Nonzero means warn if slash-star appears in a comment. */
390 unsigned char warn_comments;
391
392 /* Nonzero means to warn about __DATA__, __TIME__ and __TIMESTAMP__ usage. */
393 unsigned char warn_date_time;
394
395 /* Nonzero means warn if a user-supplied include directory does not
396 exist. */
397 unsigned char warn_missing_include_dirs;
398
399 /* Nonzero means warn if there are any trigraphs. */
400 unsigned char warn_trigraphs;
401
402 /* Nonzero means warn about multicharacter charconsts. */
403 unsigned char warn_multichar;
404
405 /* Nonzero means warn about various incompatibilities with
406 traditional C. */
407 unsigned char cpp_warn_traditional;
408
409 /* Nonzero means warn about long long numeric constants. */
410 unsigned char cpp_warn_long_long;
411
412 /* Nonzero means warn about text after an #endif (or #else). */
413 unsigned char warn_endif_labels;
414
415 /* Nonzero means warn about implicit sign changes owing to integer
416 promotions. */
417 unsigned char warn_num_sign_change;
418
419 /* Zero means don't warn about __VA_ARGS__ usage in c89 pedantic mode.
420 Presumably the usage is protected by the appropriate #ifdef. */
421 unsigned char warn_variadic_macros;
422
423 /* Nonzero means warn about builtin macros that are redefined or
424 explicitly undefined. */
425 unsigned char warn_builtin_macro_redefined;
426
427 /* Different -Wimplicit-fallthrough= levels. */
428 unsigned char cpp_warn_implicit_fallthrough;
429
430 /* Nonzero means we should look for header.gcc files that remap file
431 names. */
432 unsigned char remap;
433
434 /* Zero means dollar signs are punctuation. */
435 unsigned char dollars_in_ident;
436
437 /* Nonzero means UCNs are accepted in identifiers. */
438 unsigned char extended_identifiers;
439
440 /* True if we should warn about dollars in identifiers or numbers
441 for this translation unit. */
442 unsigned char warn_dollars;
443
444 /* Nonzero means warn if undefined identifiers are evaluated in an #if. */
445 unsigned char warn_undef;
446
447 /* Nonzero means warn if "defined" is encountered in a place other than
448 an #if. */
449 unsigned char warn_expansion_to_defined;
450
451 /* Nonzero means warn of unused macros from the main file. */
452 unsigned char warn_unused_macros;
453
454 /* Nonzero for the 1999 C Standard, including corrigenda and amendments. */
455 unsigned char c99;
456
457 /* Nonzero if we are conforming to a specific C or C++ standard. */
458 unsigned char std;
459
460 /* Nonzero means give all the error messages the ANSI standard requires. */
461 unsigned char cpp_pedantic;
462
463 /* Nonzero means we're looking at already preprocessed code, so don't
464 bother trying to do macro expansion and whatnot. */
465 unsigned char preprocessed;
466
467 /* Nonzero means we are going to emit debugging logs during
468 preprocessing. */
469 unsigned char debug;
470
471 /* Nonzero means we are tracking locations of tokens involved in
472 macro expansion. 1 Means we track the location in degraded mode
473 where we do not track locations of tokens resulting from the
474 expansion of arguments of function-like macro. 2 Means we do
475 track all macro expansions. This last option is the one that
476 consumes the highest amount of memory. */
477 unsigned char track_macro_expansion;
478
479 /* Nonzero means handle C++ alternate operator names. */
480 unsigned char operator_names;
481
482 /* Nonzero means warn about use of C++ alternate operator names. */
483 unsigned char warn_cxx_operator_names;
484
485 /* True for traditional preprocessing. */
486 unsigned char traditional;
487
488 /* Nonzero for C++ 2011 Standard user-defined literals. */
489 unsigned char user_literals;
490
491 /* Nonzero means warn when a string or character literal is followed by a
492 ud-suffix which does not beging with an underscore. */
493 unsigned char warn_literal_suffix;
494
495 /* Nonzero means interpret imaginary, fixed-point, or other gnu extension
496 literal number suffixes as user-defined literal number suffixes. */
497 unsigned char ext_numeric_literals;
498
499 /* Nonzero means extended identifiers allow the characters specified
500 in C11. */
501 unsigned char c11_identifiers;
502
503 /* Nonzero means extended identifiers allow the characters specified
504 by Unicode XID_Start and XID_Continue properties. */
505 unsigned char xid_identifiers;
506
507 /* Nonzero for C++ 2014 Standard binary constants. */
508 unsigned char binary_constants;
509
510 /* Nonzero for C++ 2014 Standard digit separators. */
511 unsigned char digit_separators;
512
513 /* Nonzero for C2X decimal floating-point constants. */
514 unsigned char dfp_constants;
515
516 /* Nonzero for C++20 __VA_OPT__ feature. */
517 unsigned char va_opt;
518
519 /* Nonzero for the '::' token. */
520 unsigned char scope;
521
522 /* Nonzero for the '#elifdef' and '#elifndef' directives. */
523 unsigned char elifdef;
524
525 /* Nonzero for the '#warning' directive. */
526 unsigned char warning_directive;
527
528 /* Nonzero means tokenize C++20 module directives. */
529 unsigned char module_directives;
530
531 /* Nonzero for C++23 size_t literals. */
532 unsigned char size_t_literals;
533
534 /* Nonzero for C++23 delimited escape sequences. */
535 unsigned char delimited_escape_seqs;
536
537 /* Nonzero for 'true' and 'false' in #if expressions. */
538 unsigned char true_false;
539
540 /* Holds the name of the target (execution) character set. */
541 const char *narrow_charset;
542
543 /* Holds the name of the target wide character set. */
544 const char *wide_charset;
545
546 /* Holds the name of the input character set. */
547 const char *input_charset;
548
549 /* The minimum permitted level of normalization before a warning
550 is generated. See enum cpp_normalize_level. */
551 int warn_normalize;
552
553 /* True to warn about precompiled header files we couldn't use. */
554 bool warn_invalid_pch;
555
556 /* True if dependencies should be restored from a precompiled header. */
557 bool restore_pch_deps;
558
559 /* True if warn about differences between C90 and C99. */
560 signed char cpp_warn_c90_c99_compat;
561
562 /* True if warn about differences between C11 and C2X. */
563 signed char cpp_warn_c11_c2x_compat;
564
565 /* True if warn about differences between C++98 and C++11. */
566 bool cpp_warn_cxx11_compat;
567
568 /* True if warn about differences between C++17 and C++20. */
569 bool cpp_warn_cxx20_compat;
570
571 /* Nonzero if bidirectional control characters checking is on. See enum
572 cpp_bidirectional_level. */
573 unsigned char cpp_warn_bidirectional;
574
575 /* True if libcpp should warn about invalid UTF-8 characters in comments.
576 2 if it should be a pedwarn. */
577 unsigned char cpp_warn_invalid_utf8;
578
579 /* True if libcpp should warn about invalid forms of delimited or named
580 escape sequences. */
581 bool cpp_warn_unicode;
582
583 /* True if -finput-charset= option has been used explicitly. */
584 bool cpp_input_charset_explicit;
585
586 /* Dependency generation. */
587 struct
588 {
589 /* Style of header dependencies to generate. */
590 enum cpp_deps_style style;
591
592 /* Assume missing files are generated files. */
593 bool missing_files;
594
595 /* Generate phony targets for each dependency apart from the first
596 one. */
597 bool phony_targets;
598
599 /* Generate dependency info for modules. */
600 bool modules;
601
602 /* If true, no dependency is generated on the main file. */
603 bool ignore_main_file;
604
605 /* If true, intend to use the preprocessor output (e.g., for compilation)
606 in addition to the dependency info. */
607 bool need_preprocessor_output;
608 } deps;
609
610 /* Target-specific features set by the front end or client. */
611
612 /* Precision for target CPP arithmetic, target characters, target
613 ints and target wide characters, respectively. */
614 size_t precision, char_precision, int_precision, wchar_precision;
615
616 /* True means chars (wide chars, UTF-8 chars) are unsigned. */
617 bool unsigned_char, unsigned_wchar, unsigned_utf8char;
618
619 /* True if the most significant byte in a word has the lowest
620 address in memory. */
621 bool bytes_big_endian;
622
623 /* Nonzero means __STDC__ should have the value 0 in system headers. */
624 unsigned char stdc_0_in_system_headers;
625
626 /* True disables tokenization outside of preprocessing directives. */
627 bool directives_only;
628
629 /* True enables canonicalization of system header file paths. */
630 bool canonical_system_headers;
631
632 /* The maximum depth of the nested #include. */
633 unsigned int max_include_depth;
634
635 cpp_main_search main_search : 8;
636 };
637
638 /* Diagnostic levels. To get a diagnostic without associating a
639 position in the translation unit with it, use cpp_error_with_line
640 with a line number of zero. */
641
642 enum cpp_diagnostic_level {
643 /* Warning, an error with -Werror. */
644 CPP_DL_WARNING = 0,
645 /* Same as CPP_DL_WARNING, except it is not suppressed in system headers. */
646 CPP_DL_WARNING_SYSHDR,
647 /* Warning, an error with -pedantic-errors or -Werror. */
648 CPP_DL_PEDWARN,
649 /* An error. */
650 CPP_DL_ERROR,
651 /* An internal consistency check failed. Prints "internal error: ",
652 otherwise the same as CPP_DL_ERROR. */
653 CPP_DL_ICE,
654 /* An informative note following a warning. */
655 CPP_DL_NOTE,
656 /* A fatal error. */
657 CPP_DL_FATAL
658 };
659
660 /* Warning reason codes. Use a reason code of CPP_W_NONE for unclassified
661 warnings and diagnostics that are not warnings. */
662
663 enum cpp_warning_reason {
664 CPP_W_NONE = 0,
665 CPP_W_DEPRECATED,
666 CPP_W_COMMENTS,
667 CPP_W_MISSING_INCLUDE_DIRS,
668 CPP_W_TRIGRAPHS,
669 CPP_W_MULTICHAR,
670 CPP_W_TRADITIONAL,
671 CPP_W_LONG_LONG,
672 CPP_W_ENDIF_LABELS,
673 CPP_W_NUM_SIGN_CHANGE,
674 CPP_W_VARIADIC_MACROS,
675 CPP_W_BUILTIN_MACRO_REDEFINED,
676 CPP_W_DOLLARS,
677 CPP_W_UNDEF,
678 CPP_W_UNUSED_MACROS,
679 CPP_W_CXX_OPERATOR_NAMES,
680 CPP_W_NORMALIZE,
681 CPP_W_INVALID_PCH,
682 CPP_W_WARNING_DIRECTIVE,
683 CPP_W_LITERAL_SUFFIX,
684 CPP_W_SIZE_T_LITERALS,
685 CPP_W_DATE_TIME,
686 CPP_W_PEDANTIC,
687 CPP_W_C90_C99_COMPAT,
688 CPP_W_C11_C2X_COMPAT,
689 CPP_W_CXX11_COMPAT,
690 CPP_W_CXX20_COMPAT,
691 CPP_W_EXPANSION_TO_DEFINED,
692 CPP_W_BIDIRECTIONAL,
693 CPP_W_INVALID_UTF8,
694 CPP_W_UNICODE
695 };
696
697 /* Callback for header lookup for HEADER, which is the name of a
698 source file. It is used as a method of last resort to find headers
699 that are not otherwise found during the normal include processing.
700 The return value is the malloced name of a header to try and open,
701 if any, or NULL otherwise. This callback is called only if the
702 header is otherwise unfound. */
703 typedef const char *(*missing_header_cb)(cpp_reader *, const char *header, cpp_dir **);
704
705 /* Call backs to cpplib client. */
706 struct cpp_callbacks
707 {
708 /* Called when a new line of preprocessed output is started. */
709 void (*line_change) (cpp_reader *, const cpp_token *, int);
710
711 /* Called when switching to/from a new file.
712 The line_map is for the new file. It is NULL if there is no new file.
713 (In C this happens when done with <built-in>+<command line> and also
714 when done with a main file.) This can be used for resource cleanup. */
715 void (*file_change) (cpp_reader *, const line_map_ordinary *);
716
717 void (*dir_change) (cpp_reader *, const char *);
718 void (*include) (cpp_reader *, location_t, const unsigned char *,
719 const char *, int, const cpp_token **);
720 void (*define) (cpp_reader *, location_t, cpp_hashnode *);
721 void (*undef) (cpp_reader *, location_t, cpp_hashnode *);
722 void (*ident) (cpp_reader *, location_t, const cpp_string *);
723 void (*def_pragma) (cpp_reader *, location_t);
724 int (*valid_pch) (cpp_reader *, const char *, int);
725 void (*read_pch) (cpp_reader *, const char *, int, const char *);
726 missing_header_cb missing_header;
727
728 /* Context-sensitive macro support. Returns macro (if any) that should
729 be expanded. */
730 cpp_hashnode * (*macro_to_expand) (cpp_reader *, const cpp_token *);
731
732 /* Called to emit a diagnostic. This callback receives the
733 translated message. */
734 bool (*diagnostic) (cpp_reader *,
735 enum cpp_diagnostic_level,
736 enum cpp_warning_reason,
737 rich_location *,
738 const char *, va_list *)
739 ATTRIBUTE_FPTR_PRINTF(5,0);
740
741 /* Callbacks for when a macro is expanded, or tested (whether
742 defined or not at the time) in #ifdef, #ifndef or "defined". */
743 void (*used_define) (cpp_reader *, location_t, cpp_hashnode *);
744 void (*used_undef) (cpp_reader *, location_t, cpp_hashnode *);
745 /* Called before #define and #undef or other macro definition
746 changes are processed. */
747 void (*before_define) (cpp_reader *);
748 /* Called whenever a macro is expanded or tested.
749 Second argument is the location of the start of the current expansion. */
750 void (*used) (cpp_reader *, location_t, cpp_hashnode *);
751
752 /* Callback to identify whether an attribute exists. */
753 int (*has_attribute) (cpp_reader *, bool);
754
755 /* Callback to determine whether a built-in function is recognized. */
756 int (*has_builtin) (cpp_reader *);
757
758 /* Callback that can change a user lazy into normal macro. */
759 void (*user_lazy_macro) (cpp_reader *, cpp_macro *, unsigned);
760
761 /* Callback to handle deferred cpp_macros. */
762 cpp_macro *(*user_deferred_macro) (cpp_reader *, location_t, cpp_hashnode *);
763
764 /* Callback to parse SOURCE_DATE_EPOCH from environment. */
765 time_t (*get_source_date_epoch) (cpp_reader *);
766
767 /* Callback for providing suggestions for misspelled directives. */
768 const char *(*get_suggestion) (cpp_reader *, const char *, const char *const *);
769
770 /* Callback for when a comment is encountered, giving the location
771 of the opening slash, a pointer to the content (which is not
772 necessarily 0-terminated), and the length of the content.
773 The content contains the opening slash-star (or slash-slash),
774 and for C-style comments contains the closing star-slash. For
775 C++-style comments it does not include the terminating newline. */
776 void (*comment) (cpp_reader *, location_t, const unsigned char *,
777 size_t);
778
779 /* Callback for filename remapping in __FILE__ and __BASE_FILE__ macro
780 expansions. */
781 const char *(*remap_filename) (const char*);
782
783 /* Maybe translate a #include into something else. Return a
784 cpp_buffer containing the translation if translating. */
785 char *(*translate_include) (cpp_reader *, line_maps *, location_t,
786 const char *path);
787 };
788
789 #ifdef VMS
790 #define INO_T_CPP ino_t ino[3]
791 #elif defined (_AIX) && SIZEOF_INO_T == 4
792 #define INO_T_CPP ino64_t ino
793 #else
794 #define INO_T_CPP ino_t ino
795 #endif
796
797 #if defined (_AIX) && SIZEOF_DEV_T == 4
798 #define DEV_T_CPP dev64_t dev
799 #else
800 #define DEV_T_CPP dev_t dev
801 #endif
802
803 /* Chain of directories to look for include files in. */
804 struct cpp_dir
805 {
806 /* NULL-terminated singly-linked list. */
807 struct cpp_dir *next;
808
809 /* NAME of the directory, NUL-terminated. */
810 char *name;
811 unsigned int len;
812
813 /* One if a system header, two if a system header that has extern
814 "C" guards for C++. */
815 unsigned char sysp;
816
817 /* Is this a user-supplied directory? */
818 bool user_supplied_p;
819
820 /* The canonicalized NAME as determined by lrealpath. This field
821 is only used by hosts that lack reliable inode numbers. */
822 char *canonical_name;
823
824 /* Mapping of file names for this directory for MS-DOS and related
825 platforms. A NULL-terminated array of (from, to) pairs. */
826 const char **name_map;
827
828 /* Routine to construct pathname, given the search path name and the
829 HEADER we are trying to find, return a constructed pathname to
830 try and open. If this is NULL, the constructed pathname is as
831 constructed by append_file_to_dir. */
832 char *(*construct) (const char *header, cpp_dir *dir);
833
834 /* The C front end uses these to recognize duplicated
835 directories in the search path. */
836 INO_T_CPP;
837 DEV_T_CPP;
838 };
839
840 /* The kind of the cpp_macro. */
841 enum cpp_macro_kind {
842 cmk_macro, /* An ISO macro (token expansion). */
843 cmk_assert, /* An assertion. */
844 cmk_traditional /* A traditional macro (text expansion). */
845 };
846
847 /* Each macro definition is recorded in a cpp_macro structure.
848 Variadic macros cannot occur with traditional cpp. */
849 struct GTY(()) cpp_macro {
850 union cpp_parm_u
851 {
852 /* Parameters, if any. If parameter names use extended identifiers,
853 the original spelling of those identifiers, not the canonical
854 UTF-8 spelling, goes here. */
855 cpp_hashnode ** GTY ((tag ("false"),
856 nested_ptr (union tree_node,
857 "%h ? CPP_HASHNODE (GCC_IDENT_TO_HT_IDENT (%h)) : NULL",
858 "%h ? HT_IDENT_TO_GCC_IDENT (HT_NODE (%h)) : NULL"),
859 length ("%1.paramc"))) params;
860
861 /* If this is an assertion, the next one in the chain. */
862 cpp_macro *GTY ((tag ("true"))) next;
863 } GTY ((desc ("%1.kind == cmk_assert"))) parm;
864
865 /* Definition line number. */
866 location_t line;
867
868 /* Number of tokens in body, or bytes for traditional macros. */
869 /* Do we really need 2^32-1 range here? */
870 unsigned int count;
871
872 /* Number of parameters. */
873 unsigned short paramc;
874
875 /* Non-zero if this is a user-lazy macro, value provided by user. */
876 unsigned char lazy;
877
878 /* The kind of this macro (ISO, trad or assert) */
879 unsigned kind : 2;
880
881 /* If a function-like macro. */
882 unsigned int fun_like : 1;
883
884 /* If a variadic macro. */
885 unsigned int variadic : 1;
886
887 /* If macro defined in system header. */
888 unsigned int syshdr : 1;
889
890 /* Nonzero if it has been expanded or had its existence tested. */
891 unsigned int used : 1;
892
893 /* Indicate whether the tokens include extra CPP_PASTE tokens at the
894 end to track invalid redefinitions with consecutive CPP_PASTE
895 tokens. */
896 unsigned int extra_tokens : 1;
897
898 /* Imported C++20 macro (from a header unit). */
899 unsigned int imported_p : 1;
900
901 /* 0 bits spare (32-bit). 32 on 64-bit target. */
902
903 union cpp_exp_u
904 {
905 /* Trailing array of replacement tokens (ISO), or assertion body value. */
906 cpp_token GTY ((tag ("false"), length ("%1.count"))) tokens[1];
907
908 /* Pointer to replacement text (traditional). See comment at top
909 of cpptrad.c for how traditional function-like macros are
910 encoded. */
911 const unsigned char *GTY ((tag ("true"))) text;
912 } GTY ((desc ("%1.kind == cmk_traditional"))) exp;
913 };
914
915 /* Poisoned identifiers are flagged NODE_POISONED. NODE_OPERATOR (C++
916 only) indicates an identifier that behaves like an operator such as
917 "xor". NODE_DIAGNOSTIC is for speed in lex_token: it indicates a
918 diagnostic may be required for this node. Currently this only
919 applies to __VA_ARGS__, poisoned identifiers, and -Wc++-compat
920 warnings about NODE_OPERATOR. */
921
922 /* Hash node flags. */
923 #define NODE_OPERATOR (1 << 0) /* C++ named operator. */
924 #define NODE_POISONED (1 << 1) /* Poisoned identifier. */
925 #define NODE_DIAGNOSTIC (1 << 2) /* Possible diagnostic when lexed. */
926 #define NODE_WARN (1 << 3) /* Warn if redefined or undefined. */
927 #define NODE_DISABLED (1 << 4) /* A disabled macro. */
928 #define NODE_USED (1 << 5) /* Dumped with -dU. */
929 #define NODE_CONDITIONAL (1 << 6) /* Conditional macro */
930 #define NODE_WARN_OPERATOR (1 << 7) /* Warn about C++ named operator. */
931 #define NODE_MODULE (1 << 8) /* C++-20 module-related name. */
932
933 /* Different flavors of hash node. */
934 enum node_type
935 {
936 NT_VOID = 0, /* Maybe an assert? */
937 NT_MACRO_ARG, /* A macro arg. */
938 NT_USER_MACRO, /* A user macro. */
939 NT_BUILTIN_MACRO, /* A builtin macro. */
940 NT_MACRO_MASK = NT_USER_MACRO /* Mask for either macro kind. */
941 };
942
943 /* Different flavors of builtin macro. _Pragma is an operator, but we
944 handle it with the builtin code for efficiency reasons. */
945 enum cpp_builtin_type
946 {
947 BT_SPECLINE = 0, /* `__LINE__' */
948 BT_DATE, /* `__DATE__' */
949 BT_FILE, /* `__FILE__' */
950 BT_FILE_NAME, /* `__FILE_NAME__' */
951 BT_BASE_FILE, /* `__BASE_FILE__' */
952 BT_INCLUDE_LEVEL, /* `__INCLUDE_LEVEL__' */
953 BT_TIME, /* `__TIME__' */
954 BT_STDC, /* `__STDC__' */
955 BT_PRAGMA, /* `_Pragma' operator */
956 BT_TIMESTAMP, /* `__TIMESTAMP__' */
957 BT_COUNTER, /* `__COUNTER__' */
958 BT_HAS_ATTRIBUTE, /* `__has_attribute(x)' */
959 BT_HAS_STD_ATTRIBUTE, /* `__has_c_attribute(x)' */
960 BT_HAS_BUILTIN, /* `__has_builtin(x)' */
961 BT_HAS_INCLUDE, /* `__has_include(x)' */
962 BT_HAS_INCLUDE_NEXT /* `__has_include_next(x)' */
963 };
964
965 #define CPP_HASHNODE(HNODE) ((cpp_hashnode *) (HNODE))
966 #define HT_NODE(NODE) (&(NODE)->ident)
967 #define NODE_LEN(NODE) HT_LEN (HT_NODE (NODE))
968 #define NODE_NAME(NODE) HT_STR (HT_NODE (NODE))
969
970 /* The common part of an identifier node shared amongst all 3 C front
971 ends. Also used to store CPP identifiers, which are a superset of
972 identifiers in the grammatical sense. */
973
974 union GTY(()) _cpp_hashnode_value {
975 /* Assert (maybe NULL) */
976 cpp_macro * GTY((tag ("NT_VOID"))) answers;
977 /* Macro (maybe NULL) */
978 cpp_macro * GTY((tag ("NT_USER_MACRO"))) macro;
979 /* Code for a builtin macro. */
980 enum cpp_builtin_type GTY ((tag ("NT_BUILTIN_MACRO"))) builtin;
981 /* Macro argument index. */
982 unsigned short GTY ((tag ("NT_MACRO_ARG"))) arg_index;
983 };
984
985 struct GTY(()) cpp_hashnode {
986 struct ht_identifier ident;
987 unsigned int is_directive : 1;
988 unsigned int directive_index : 7; /* If is_directive,
989 then index into directive table.
990 Otherwise, a NODE_OPERATOR. */
991 unsigned int rid_code : 8; /* Rid code - for front ends. */
992 unsigned int flags : 9; /* CPP flags. */
993 ENUM_BITFIELD(node_type) type : 2; /* CPP node type. */
994
995 /* 5 bits spare. */
996
997 /* The deferred cookie is applicable to NT_USER_MACRO or NT_VOID.
998 The latter for when a macro had a prevailing undef.
999 On a 64-bit system there would be 32-bits of padding to the value
1000 field. So placing the deferred index here is not costly. */
1001 unsigned deferred; /* Deferred cookie */
1002
1003 union _cpp_hashnode_value GTY ((desc ("%1.type"))) value;
1004 };
1005
1006 /* A class for iterating through the source locations within a
1007 string token (before escapes are interpreted, and before
1008 concatenation). */
1009
1010 class cpp_string_location_reader {
1011 public:
1012 cpp_string_location_reader (location_t src_loc,
1013 line_maps *line_table);
1014
1015 source_range get_next ();
1016
1017 private:
1018 location_t m_loc;
1019 int m_offset_per_column;
1020 };
1021
1022 /* A class for storing the source ranges of all of the characters within
1023 a string literal, after escapes are interpreted, and after
1024 concatenation.
1025
1026 This is not GTY-marked, as instances are intended to be temporary. */
1027
1028 class cpp_substring_ranges
1029 {
1030 public:
1031 cpp_substring_ranges ();
1032 ~cpp_substring_ranges ();
1033
1034 int get_num_ranges () const { return m_num_ranges; }
1035 source_range get_range (int idx) const
1036 {
1037 linemap_assert (idx < m_num_ranges);
1038 return m_ranges[idx];
1039 }
1040
1041 void add_range (source_range range);
1042 void add_n_ranges (int num, cpp_string_location_reader &loc_reader);
1043
1044 private:
1045 source_range *m_ranges;
1046 int m_num_ranges;
1047 int m_alloc_ranges;
1048 };
1049
1050 /* Call this first to get a handle to pass to other functions.
1051
1052 If you want cpplib to manage its own hashtable, pass in a NULL
1053 pointer. Otherwise you should pass in an initialized hash table
1054 that cpplib will share; this technique is used by the C front
1055 ends. */
1056 extern cpp_reader *cpp_create_reader (enum c_lang, struct ht *,
1057 class line_maps *);
1058
1059 /* Reset the cpp_reader's line_map. This is only used after reading a
1060 PCH file. */
1061 extern void cpp_set_line_map (cpp_reader *, class line_maps *);
1062
1063 /* Call this to change the selected language standard (e.g. because of
1064 command line options). */
1065 extern void cpp_set_lang (cpp_reader *, enum c_lang);
1066
1067 /* Set the include paths. */
1068 extern void cpp_set_include_chains (cpp_reader *, cpp_dir *, cpp_dir *, int);
1069
1070 /* Call these to get pointers to the options, callback, and deps
1071 structures for a given reader. These pointers are good until you
1072 call cpp_finish on that reader. You can either edit the callbacks
1073 through the pointer returned from cpp_get_callbacks, or set them
1074 with cpp_set_callbacks. */
1075 extern cpp_options *cpp_get_options (cpp_reader *) ATTRIBUTE_PURE;
1076 extern cpp_callbacks *cpp_get_callbacks (cpp_reader *) ATTRIBUTE_PURE;
1077 extern void cpp_set_callbacks (cpp_reader *, cpp_callbacks *);
1078 extern class mkdeps *cpp_get_deps (cpp_reader *) ATTRIBUTE_PURE;
1079
1080 extern const char *cpp_probe_header_unit (cpp_reader *, const char *file,
1081 bool angle_p, location_t);
1082
1083 /* Call these to get name data about the various compile-time
1084 charsets. */
1085 extern const char *cpp_get_narrow_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1086 extern const char *cpp_get_wide_charset_name (cpp_reader *) ATTRIBUTE_PURE;
1087
1088 /* This function reads the file, but does not start preprocessing. It
1089 returns the name of the original file; this is the same as the
1090 input file, except for preprocessed input. This will generate at
1091 least one file change callback, and possibly a line change callback
1092 too. If there was an error opening the file, it returns NULL. */
1093 extern const char *cpp_read_main_file (cpp_reader *, const char *,
1094 bool injecting = false);
1095 extern location_t cpp_main_loc (const cpp_reader *);
1096
1097 /* Adjust for the main file to be an include. */
1098 extern void cpp_retrofit_as_include (cpp_reader *);
1099
1100 /* Set up built-ins with special behavior. Use cpp_init_builtins()
1101 instead unless your know what you are doing. */
1102 extern void cpp_init_special_builtins (cpp_reader *);
1103
1104 /* Set up built-ins like __FILE__. */
1105 extern void cpp_init_builtins (cpp_reader *, int);
1106
1107 /* This is called after options have been parsed, and partially
1108 processed. */
1109 extern void cpp_post_options (cpp_reader *);
1110
1111 /* Set up translation to the target character set. */
1112 extern void cpp_init_iconv (cpp_reader *);
1113
1114 /* Call this to finish preprocessing. If you requested dependency
1115 generation, pass an open stream to write the information to,
1116 otherwise NULL. It is your responsibility to close the stream. */
1117 extern void cpp_finish (cpp_reader *, FILE *deps_stream);
1118
1119 /* Call this to release the handle at the end of preprocessing. Any
1120 use of the handle after this function returns is invalid. */
1121 extern void cpp_destroy (cpp_reader *);
1122
1123 extern unsigned int cpp_token_len (const cpp_token *);
1124 extern unsigned char *cpp_token_as_text (cpp_reader *, const cpp_token *);
1125 extern unsigned char *cpp_spell_token (cpp_reader *, const cpp_token *,
1126 unsigned char *, bool);
1127 extern void cpp_register_pragma (cpp_reader *, const char *, const char *,
1128 void (*) (cpp_reader *), bool);
1129 extern void cpp_register_deferred_pragma (cpp_reader *, const char *,
1130 const char *, unsigned, bool, bool);
1131 extern int cpp_avoid_paste (cpp_reader *, const cpp_token *,
1132 const cpp_token *);
1133 extern const cpp_token *cpp_get_token (cpp_reader *);
1134 extern const cpp_token *cpp_get_token_with_location (cpp_reader *,
1135 location_t *);
1136 inline bool cpp_user_macro_p (const cpp_hashnode *node)
1137 {
1138 return node->type == NT_USER_MACRO;
1139 }
1140 inline bool cpp_builtin_macro_p (const cpp_hashnode *node)
1141 {
1142 return node->type == NT_BUILTIN_MACRO;
1143 }
1144 inline bool cpp_macro_p (const cpp_hashnode *node)
1145 {
1146 return node->type & NT_MACRO_MASK;
1147 }
1148 inline cpp_macro *cpp_set_deferred_macro (cpp_hashnode *node,
1149 cpp_macro *forced = NULL)
1150 {
1151 cpp_macro *old = node->value.macro;
1152
1153 node->value.macro = forced;
1154 node->type = NT_USER_MACRO;
1155 node->flags &= ~NODE_USED;
1156
1157 return old;
1158 }
1159 cpp_macro *cpp_get_deferred_macro (cpp_reader *, cpp_hashnode *, location_t);
1160
1161 /* Returns true if NODE is a function-like user macro. */
1162 inline bool cpp_fun_like_macro_p (cpp_hashnode *node)
1163 {
1164 return cpp_user_macro_p (node) && node->value.macro->fun_like;
1165 }
1166
1167 extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *);
1168 extern const unsigned char *cpp_macro_definition (cpp_reader *, cpp_hashnode *,
1169 const cpp_macro *);
1170 inline location_t cpp_macro_definition_location (cpp_hashnode *node)
1171 {
1172 const cpp_macro *macro = node->value.macro;
1173 return macro ? macro->line : 0;
1174 }
1175 /* Return an idempotent time stamp (possibly from SOURCE_DATE_EPOCH). */
1176 enum class CPP_time_kind
1177 {
1178 FIXED = -1, /* Fixed time via source epoch. */
1179 DYNAMIC = -2, /* Dynamic via time(2). */
1180 UNKNOWN = -3 /* Wibbly wobbly, timey wimey. */
1181 };
1182 extern CPP_time_kind cpp_get_date (cpp_reader *, time_t *);
1183
1184 extern void _cpp_backup_tokens (cpp_reader *, unsigned int);
1185 extern const cpp_token *cpp_peek_token (cpp_reader *, int);
1186
1187 /* Evaluate a CPP_*CHAR* token. */
1188 extern cppchar_t cpp_interpret_charconst (cpp_reader *, const cpp_token *,
1189 unsigned int *, int *);
1190 /* Evaluate a vector of CPP_*STRING* tokens. */
1191 extern bool cpp_interpret_string (cpp_reader *,
1192 const cpp_string *, size_t,
1193 cpp_string *, enum cpp_ttype);
1194 extern const char *cpp_interpret_string_ranges (cpp_reader *pfile,
1195 const cpp_string *from,
1196 cpp_string_location_reader *,
1197 size_t count,
1198 cpp_substring_ranges *out,
1199 enum cpp_ttype type);
1200 extern bool cpp_interpret_string_notranslate (cpp_reader *,
1201 const cpp_string *, size_t,
1202 cpp_string *, enum cpp_ttype);
1203
1204 /* Convert a host character constant to the execution character set. */
1205 extern cppchar_t cpp_host_to_exec_charset (cpp_reader *, cppchar_t);
1206
1207 /* Used to register macros and assertions, perhaps from the command line.
1208 The text is the same as the command line argument. */
1209 extern void cpp_define (cpp_reader *, const char *);
1210 extern void cpp_define_unused (cpp_reader *, const char *);
1211 extern void cpp_define_formatted (cpp_reader *pfile,
1212 const char *fmt, ...) ATTRIBUTE_PRINTF_2;
1213 extern void cpp_define_formatted_unused (cpp_reader *pfile,
1214 const char *fmt,
1215 ...) ATTRIBUTE_PRINTF_2;
1216 extern void cpp_assert (cpp_reader *, const char *);
1217 extern void cpp_undef (cpp_reader *, const char *);
1218 extern void cpp_unassert (cpp_reader *, const char *);
1219
1220 /* Mark a node as a lazily defined macro. */
1221 extern void cpp_define_lazily (cpp_reader *, cpp_hashnode *node, unsigned N);
1222
1223 /* Undefine all macros and assertions. */
1224 extern void cpp_undef_all (cpp_reader *);
1225
1226 extern cpp_buffer *cpp_push_buffer (cpp_reader *, const unsigned char *,
1227 size_t, int);
1228 extern int cpp_defined (cpp_reader *, const unsigned char *, int);
1229
1230 /* A preprocessing number. Code assumes that any unused high bits of
1231 the double integer are set to zero. */
1232
1233 /* This type has to be equal to unsigned HOST_WIDE_INT, see
1234 gcc/c-family/c-lex.cc. */
1235 typedef uint64_t cpp_num_part;
1236 typedef struct cpp_num cpp_num;
1237 struct cpp_num
1238 {
1239 cpp_num_part high;
1240 cpp_num_part low;
1241 bool unsignedp; /* True if value should be treated as unsigned. */
1242 bool overflow; /* True if the most recent calculation overflowed. */
1243 };
1244
1245 /* cpplib provides two interfaces for interpretation of preprocessing
1246 numbers.
1247
1248 cpp_classify_number categorizes numeric constants according to
1249 their field (integer, floating point, or invalid), radix (decimal,
1250 octal, hexadecimal), and type suffixes. */
1251
1252 #define CPP_N_CATEGORY 0x000F
1253 #define CPP_N_INVALID 0x0000
1254 #define CPP_N_INTEGER 0x0001
1255 #define CPP_N_FLOATING 0x0002
1256
1257 #define CPP_N_WIDTH 0x00F0
1258 #define CPP_N_SMALL 0x0010 /* int, float, short _Fract/Accum */
1259 #define CPP_N_MEDIUM 0x0020 /* long, double, long _Fract/_Accum. */
1260 #define CPP_N_LARGE 0x0040 /* long long, long double,
1261 long long _Fract/Accum. */
1262
1263 #define CPP_N_WIDTH_MD 0xF0000 /* machine defined. */
1264 #define CPP_N_MD_W 0x10000
1265 #define CPP_N_MD_Q 0x20000
1266
1267 #define CPP_N_RADIX 0x0F00
1268 #define CPP_N_DECIMAL 0x0100
1269 #define CPP_N_HEX 0x0200
1270 #define CPP_N_OCTAL 0x0400
1271 #define CPP_N_BINARY 0x0800
1272
1273 #define CPP_N_UNSIGNED 0x1000 /* Properties. */
1274 #define CPP_N_IMAGINARY 0x2000
1275 #define CPP_N_DFLOAT 0x4000
1276 #define CPP_N_DEFAULT 0x8000
1277
1278 #define CPP_N_FRACT 0x100000 /* Fract types. */
1279 #define CPP_N_ACCUM 0x200000 /* Accum types. */
1280 #define CPP_N_FLOATN 0x400000 /* _FloatN types. */
1281 #define CPP_N_FLOATNX 0x800000 /* _FloatNx types. */
1282
1283 #define CPP_N_USERDEF 0x1000000 /* C++11 user-defined literal. */
1284
1285 #define CPP_N_SIZE_T 0x2000000 /* C++23 size_t literal. */
1286 #define CPP_N_BFLOAT16 0x4000000 /* std::bfloat16_t type. */
1287
1288 #define CPP_N_WIDTH_FLOATN_NX 0xF0000000 /* _FloatN / _FloatNx value
1289 of N, divided by 16. */
1290 #define CPP_FLOATN_SHIFT 24
1291 #define CPP_FLOATN_MAX 0xF0
1292
1293 /* Classify a CPP_NUMBER token. The return value is a combination of
1294 the flags from the above sets. */
1295 extern unsigned cpp_classify_number (cpp_reader *, const cpp_token *,
1296 const char **, location_t);
1297
1298 /* Return the classification flags for a float suffix. */
1299 extern unsigned int cpp_interpret_float_suffix (cpp_reader *, const char *,
1300 size_t);
1301
1302 /* Return the classification flags for an int suffix. */
1303 extern unsigned int cpp_interpret_int_suffix (cpp_reader *, const char *,
1304 size_t);
1305
1306 /* Evaluate a token classified as category CPP_N_INTEGER. */
1307 extern cpp_num cpp_interpret_integer (cpp_reader *, const cpp_token *,
1308 unsigned int);
1309
1310 /* Sign extend a number, with PRECISION significant bits and all
1311 others assumed clear, to fill out a cpp_num structure. */
1312 cpp_num cpp_num_sign_extend (cpp_num, size_t);
1313
1314 /* Output a diagnostic of some kind. */
1315 extern bool cpp_error (cpp_reader *, enum cpp_diagnostic_level,
1316 const char *msgid, ...)
1317 ATTRIBUTE_PRINTF_3;
1318 extern bool cpp_warning (cpp_reader *, enum cpp_warning_reason,
1319 const char *msgid, ...)
1320 ATTRIBUTE_PRINTF_3;
1321 extern bool cpp_pedwarning (cpp_reader *, enum cpp_warning_reason,
1322 const char *msgid, ...)
1323 ATTRIBUTE_PRINTF_3;
1324 extern bool cpp_warning_syshdr (cpp_reader *, enum cpp_warning_reason reason,
1325 const char *msgid, ...)
1326 ATTRIBUTE_PRINTF_3;
1327
1328 /* As their counterparts above, but use RICHLOC. */
1329 extern bool cpp_warning_at (cpp_reader *, enum cpp_warning_reason,
1330 rich_location *richloc, const char *msgid, ...)
1331 ATTRIBUTE_PRINTF_4;
1332 extern bool cpp_pedwarning_at (cpp_reader *, enum cpp_warning_reason,
1333 rich_location *richloc, const char *msgid, ...)
1334 ATTRIBUTE_PRINTF_4;
1335
1336 /* Output a diagnostic with "MSGID: " preceding the
1337 error string of errno. No location is printed. */
1338 extern bool cpp_errno (cpp_reader *, enum cpp_diagnostic_level,
1339 const char *msgid);
1340 /* Similarly, but with "FILENAME: " instead of "MSGID: ", where
1341 the filename is not localized. */
1342 extern bool cpp_errno_filename (cpp_reader *, enum cpp_diagnostic_level,
1343 const char *filename, location_t loc);
1344
1345 /* Same as cpp_error, except additionally specifies a position as a
1346 (translation unit) physical line and physical column. If the line is
1347 zero, then no location is printed. */
1348 extern bool cpp_error_with_line (cpp_reader *, enum cpp_diagnostic_level,
1349 location_t, unsigned,
1350 const char *msgid, ...)
1351 ATTRIBUTE_PRINTF_5;
1352 extern bool cpp_warning_with_line (cpp_reader *, enum cpp_warning_reason,
1353 location_t, unsigned,
1354 const char *msgid, ...)
1355 ATTRIBUTE_PRINTF_5;
1356 extern bool cpp_pedwarning_with_line (cpp_reader *, enum cpp_warning_reason,
1357 location_t, unsigned,
1358 const char *msgid, ...)
1359 ATTRIBUTE_PRINTF_5;
1360 extern bool cpp_warning_with_line_syshdr (cpp_reader *, enum cpp_warning_reason,
1361 location_t, unsigned,
1362 const char *msgid, ...)
1363 ATTRIBUTE_PRINTF_5;
1364
1365 extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1366 location_t src_loc, const char *msgid, ...)
1367 ATTRIBUTE_PRINTF_4;
1368
1369 extern bool cpp_error_at (cpp_reader * pfile, enum cpp_diagnostic_level,
1370 rich_location *richloc, const char *msgid, ...)
1371 ATTRIBUTE_PRINTF_4;
1372
1373 /* In lex.cc */
1374 extern int cpp_ideq (const cpp_token *, const char *);
1375 extern void cpp_output_line (cpp_reader *, FILE *);
1376 extern unsigned char *cpp_output_line_to_string (cpp_reader *,
1377 const unsigned char *);
1378 extern const unsigned char *cpp_alloc_token_string
1379 (cpp_reader *, const unsigned char *, unsigned);
1380 extern void cpp_output_token (const cpp_token *, FILE *);
1381 extern const char *cpp_type2name (enum cpp_ttype, unsigned char flags);
1382 /* Returns the value of an escape sequence, truncated to the correct
1383 target precision. PSTR points to the input pointer, which is just
1384 after the backslash. LIMIT is how much text we have. WIDE is true
1385 if the escape sequence is part of a wide character constant or
1386 string literal. Handles all relevant diagnostics. */
1387 extern cppchar_t cpp_parse_escape (cpp_reader *, const unsigned char ** pstr,
1388 const unsigned char *limit, int wide);
1389
1390 /* Structure used to hold a comment block at a given location in the
1391 source code. */
1392
1393 typedef struct
1394 {
1395 /* Text of the comment including the terminators. */
1396 char *comment;
1397
1398 /* source location for the given comment. */
1399 location_t sloc;
1400 } cpp_comment;
1401
1402 /* Structure holding all comments for a given cpp_reader. */
1403
1404 typedef struct
1405 {
1406 /* table of comment entries. */
1407 cpp_comment *entries;
1408
1409 /* number of actual entries entered in the table. */
1410 int count;
1411
1412 /* number of entries allocated currently. */
1413 int allocated;
1414 } cpp_comment_table;
1415
1416 /* Returns the table of comments encountered by the preprocessor. This
1417 table is only populated when pfile->state.save_comments is true. */
1418 extern cpp_comment_table *cpp_get_comments (cpp_reader *);
1419
1420 /* In hash.c */
1421
1422 /* Lookup an identifier in the hashtable. Puts the identifier in the
1423 table if it is not already there. */
1424 extern cpp_hashnode *cpp_lookup (cpp_reader *, const unsigned char *,
1425 unsigned int);
1426
1427 typedef int (*cpp_cb) (cpp_reader *, cpp_hashnode *, void *);
1428 extern void cpp_forall_identifiers (cpp_reader *, cpp_cb, void *);
1429
1430 /* In macro.cc */
1431 extern void cpp_scan_nooutput (cpp_reader *);
1432 extern int cpp_sys_macro_p (cpp_reader *);
1433 extern unsigned char *cpp_quote_string (unsigned char *, const unsigned char *,
1434 unsigned int);
1435 extern bool cpp_compare_macros (const cpp_macro *macro1,
1436 const cpp_macro *macro2);
1437
1438 /* In files.cc */
1439 extern bool cpp_included (cpp_reader *, const char *);
1440 extern bool cpp_included_before (cpp_reader *, const char *, location_t);
1441 extern void cpp_make_system_header (cpp_reader *, int, int);
1442 extern bool cpp_push_include (cpp_reader *, const char *);
1443 extern bool cpp_push_default_include (cpp_reader *, const char *);
1444 extern void cpp_change_file (cpp_reader *, enum lc_reason, const char *);
1445 extern const char *cpp_get_path (struct _cpp_file *);
1446 extern cpp_dir *cpp_get_dir (struct _cpp_file *);
1447 extern cpp_buffer *cpp_get_buffer (cpp_reader *);
1448 extern struct _cpp_file *cpp_get_file (cpp_buffer *);
1449 extern cpp_buffer *cpp_get_prev (cpp_buffer *);
1450 extern void cpp_clear_file_cache (cpp_reader *);
1451
1452 /* cpp_get_converted_source returns the contents of the given file, as it exists
1453 after cpplib has read it and converted it from the input charset to the
1454 source charset. Return struct will be zero-filled if the data could not be
1455 read for any reason. The data starts at the DATA pointer, but the TO_FREE
1456 pointer is what should be passed to free(), as there may be an offset. */
1457 struct cpp_converted_source
1458 {
1459 char *to_free;
1460 char *data;
1461 size_t len;
1462 };
1463 cpp_converted_source cpp_get_converted_source (const char *fname,
1464 const char *input_charset);
1465
1466 /* In pch.cc */
1467 struct save_macro_data;
1468 extern int cpp_save_state (cpp_reader *, FILE *);
1469 extern int cpp_write_pch_deps (cpp_reader *, FILE *);
1470 extern int cpp_write_pch_state (cpp_reader *, FILE *);
1471 extern int cpp_valid_state (cpp_reader *, const char *, int);
1472 extern void cpp_prepare_state (cpp_reader *, struct save_macro_data **);
1473 extern int cpp_read_state (cpp_reader *, const char *, FILE *,
1474 struct save_macro_data *);
1475
1476 /* In lex.cc */
1477 extern void cpp_force_token_locations (cpp_reader *, location_t);
1478 extern void cpp_stop_forcing_token_locations (cpp_reader *);
1479 enum CPP_DO_task
1480 {
1481 CPP_DO_print,
1482 CPP_DO_location,
1483 CPP_DO_token
1484 };
1485
1486 extern void cpp_directive_only_process (cpp_reader *pfile,
1487 void *data,
1488 void (*cb) (cpp_reader *,
1489 CPP_DO_task,
1490 void *data, ...));
1491
1492 /* In expr.cc */
1493 extern enum cpp_ttype cpp_userdef_string_remove_type
1494 (enum cpp_ttype type);
1495 extern enum cpp_ttype cpp_userdef_string_add_type
1496 (enum cpp_ttype type);
1497 extern enum cpp_ttype cpp_userdef_char_remove_type
1498 (enum cpp_ttype type);
1499 extern enum cpp_ttype cpp_userdef_char_add_type
1500 (enum cpp_ttype type);
1501 extern bool cpp_userdef_string_p
1502 (enum cpp_ttype type);
1503 extern bool cpp_userdef_char_p
1504 (enum cpp_ttype type);
1505 extern const char * cpp_get_userdef_suffix
1506 (const cpp_token *);
1507
1508 /* In charset.cc */
1509
1510 /* The result of attempting to decode a run of UTF-8 bytes. */
1511
1512 struct cpp_decoded_char
1513 {
1514 const char *m_start_byte;
1515 const char *m_next_byte;
1516
1517 bool m_valid_ch;
1518 cppchar_t m_ch;
1519 };
1520
1521 /* Information for mapping between code points and display columns.
1522
1523 This is a tabstop value, along with a callback for getting the
1524 widths of characters. Normally this callback is cpp_wcwidth, but we
1525 support other schemes for escaping non-ASCII unicode as a series of
1526 ASCII chars when printing the user's source code in diagnostic-show-locus.cc
1527
1528 For example, consider:
1529 - the Unicode character U+03C0 "GREEK SMALL LETTER PI" (UTF-8: 0xCF 0x80)
1530 - the Unicode character U+1F642 "SLIGHTLY SMILING FACE"
1531 (UTF-8: 0xF0 0x9F 0x99 0x82)
1532 - the byte 0xBF (a stray trailing byte of a UTF-8 character)
1533 Normally U+03C0 would occupy one display column, U+1F642
1534 would occupy two display columns, and the stray byte would be
1535 printed verbatim as one display column.
1536
1537 However when escaping them as unicode code points as "<U+03C0>"
1538 and "<U+1F642>" they occupy 8 and 9 display columns respectively,
1539 and when escaping them as bytes as "<CF><80>" and "<F0><9F><99><82>"
1540 they occupy 8 and 16 display columns respectively. In both cases
1541 the stray byte is escaped to <BF> as 4 display columns. */
1542
1543 struct cpp_char_column_policy
1544 {
1545 cpp_char_column_policy (int tabstop,
1546 int (*width_cb) (cppchar_t c))
1547 : m_tabstop (tabstop),
1548 m_undecoded_byte_width (1),
1549 m_width_cb (width_cb)
1550 {}
1551
1552 int m_tabstop;
1553 /* Width in display columns of a stray byte that isn't decodable
1554 as UTF-8. */
1555 int m_undecoded_byte_width;
1556 int (*m_width_cb) (cppchar_t c);
1557 };
1558
1559 /* A class to manage the state while converting a UTF-8 sequence to cppchar_t
1560 and computing the display width one character at a time. */
1561 class cpp_display_width_computation {
1562 public:
1563 cpp_display_width_computation (const char *data, int data_length,
1564 const cpp_char_column_policy &policy);
1565 const char *next_byte () const { return m_next; }
1566 int bytes_processed () const { return m_next - m_begin; }
1567 int bytes_left () const { return m_bytes_left; }
1568 bool done () const { return !bytes_left (); }
1569 int display_cols_processed () const { return m_display_cols; }
1570
1571 int process_next_codepoint (cpp_decoded_char *out);
1572 int advance_display_cols (int n);
1573
1574 private:
1575 const char *const m_begin;
1576 const char *m_next;
1577 size_t m_bytes_left;
1578 const cpp_char_column_policy &m_policy;
1579 int m_display_cols;
1580 };
1581
1582 /* Convenience functions that are simple use cases for class
1583 cpp_display_width_computation. Tab characters will be expanded to spaces
1584 as determined by POLICY.m_tabstop, and non-printable-ASCII characters
1585 will be escaped as per POLICY. */
1586
1587 int cpp_byte_column_to_display_column (const char *data, int data_length,
1588 int column,
1589 const cpp_char_column_policy &policy);
1590 inline int cpp_display_width (const char *data, int data_length,
1591 const cpp_char_column_policy &policy)
1592 {
1593 return cpp_byte_column_to_display_column (data, data_length, data_length,
1594 policy);
1595 }
1596 int cpp_display_column_to_byte_column (const char *data, int data_length,
1597 int display_col,
1598 const cpp_char_column_policy &policy);
1599 int cpp_wcwidth (cppchar_t c);
1600
1601 bool cpp_input_conversion_is_trivial (const char *input_charset);
1602 int cpp_check_utf8_bom (const char *data, size_t data_length);
1603 bool cpp_valid_utf8_p (const char *data, size_t num_bytes);
1604
1605 #endif /* ! LIBCPP_CPPLIB_H */