1 // -*- C++ -*-
2 //===----------------------------------------------------------------------===//
3 //
4 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
5 // See https://llvm.org/LICENSE.txt for license information.
6 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
7 //
8 //===----------------------------------------------------------------------===//
9
10 #ifndef _LIBCPP___FORMAT_PARSER_STD_FORMAT_SPEC_H
11 #define _LIBCPP___FORMAT_PARSER_STD_FORMAT_SPEC_H
12
13 /// \file Contains the std-format-spec parser.
14 ///
15 /// Most of the code can be reused in the chrono-format-spec.
16 /// This header has some support for the chrono-format-spec since it doesn't
17 /// affect the std-format-spec.
18
19 #include <__algorithm/find_if.h>
20 #include <__algorithm/min.h>
21 #include <__assert>
22 #include <__concepts/arithmetic.h>
23 #include <__concepts/same_as.h>
24 #include <__config>
25 #include <__debug>
26 #include <__format/format_arg.h>
27 #include <__format/format_error.h>
28 #include <__format/format_parse_context.h>
29 #include <__format/format_string.h>
30 #include <__format/unicode.h>
31 #include <__iterator/concepts.h>
32 #include <__iterator/readable_traits.h> // iter_value_t
33 #include <__type_traits/common_type.h>
34 #include <__type_traits/is_trivially_copyable.h>
35 #include <__variant/monostate.h>
36 #include <cstdint>
37 #include <string_view>
38
39 #if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
40 # pragma GCC system_header
41 #endif
42
43 _LIBCPP_PUSH_MACROS
44 #include <__undef_macros>
45
46 _LIBCPP_BEGIN_NAMESPACE_STD
47
48 #if _LIBCPP_STD_VER >= 20
49
50 namespace __format_spec {
51
52 template <contiguous_iterator _Iterator>
53 _LIBCPP_HIDE_FROM_ABI constexpr __format::__parse_number_result<_Iterator>
__parse_arg_id(_Iterator __begin,_Iterator __end,auto & __parse_ctx)54 __parse_arg_id(_Iterator __begin, _Iterator __end, auto& __parse_ctx) {
55 using _CharT = iter_value_t<_Iterator>;
56 // This function is a wrapper to call the real parser. But it does the
57 // validation for the pre-conditions and post-conditions.
58 if (__begin == __end)
59 std::__throw_format_error("End of input while parsing format-spec arg-id");
60
61 __format::__parse_number_result __r = __format::__parse_arg_id(__begin, __end, __parse_ctx);
62
63 if (__r.__last == __end || *__r.__last != _CharT('}'))
64 std::__throw_format_error("Invalid arg-id");
65
66 ++__r.__last;
67 return __r;
68 }
69
70 template <class _Context>
71 _LIBCPP_HIDE_FROM_ABI constexpr uint32_t
__substitute_arg_id(basic_format_arg<_Context> __format_arg)72 __substitute_arg_id(basic_format_arg<_Context> __format_arg) {
73 // [format.string.std]/8
74 // If the corresponding formatting argument is not of integral type...
75 // This wording allows char and bool too. LWG-3720 changes the wording to
76 // If the corresponding formatting argument is not of standard signed or
77 // unsigned integer type,
78 // This means the 128-bit will not be valid anymore.
79 // TODO FMT Verify this resolution is accepted and add a test to verify
80 // 128-bit integrals fail and switch to visit_format_arg.
81 return _VSTD::__visit_format_arg(
82 [](auto __arg) -> uint32_t {
83 using _Type = decltype(__arg);
84 if constexpr (same_as<_Type, monostate>)
85 std::__throw_format_error("Argument index out of bounds");
86
87 // [format.string.std]/8
88 // If { arg-idopt } is used in a width or precision, the value of the
89 // corresponding formatting argument is used in its place. If the
90 // corresponding formatting argument is not of standard signed or unsigned
91 // integer type, or its value is negative for precision or non-positive for
92 // width, an exception of type format_error is thrown.
93 //
94 // When an integral is used in a format function, it is stored as one of
95 // the types checked below. Other integral types are promoted. For example,
96 // a signed char is stored as an int.
97 if constexpr (same_as<_Type, int> || same_as<_Type, unsigned int> || //
98 same_as<_Type, long long> || same_as<_Type, unsigned long long>) {
99 if constexpr (signed_integral<_Type>) {
100 if (__arg < 0)
101 std::__throw_format_error("A format-spec arg-id replacement shouldn't have a negative value");
102 }
103
104 using _CT = common_type_t<_Type, decltype(__format::__number_max)>;
105 if (static_cast<_CT>(__arg) > static_cast<_CT>(__format::__number_max))
106 std::__throw_format_error("A format-spec arg-id replacement exceeds the maximum supported value");
107
108 return __arg;
109 } else
110 std::__throw_format_error("Replacement argument isn't a standard signed or unsigned integer type");
111 },
112 __format_arg);
113 }
114
115 /// These fields are a filter for which elements to parse.
116 ///
117 /// They default to false so when a new field is added it needs to be opted in
118 /// explicitly.
119 // TODO FMT Use an ABI tag for this struct.
120 struct __fields {
121 uint8_t __sign_ : 1 {false};
122 uint8_t __alternate_form_ : 1 {false};
123 uint8_t __zero_padding_ : 1 {false};
124 uint8_t __precision_ : 1 {false};
125 uint8_t __locale_specific_form_ : 1 {false};
126 uint8_t __type_ : 1 {false};
127 // Determines the valid values for fill.
128 //
129 // Originally the fill could be any character except { and }. Range-based
130 // formatters use the colon to mark the beginning of the
131 // underlying-format-spec. To avoid parsing ambiguities these formatter
132 // specializations prohibit the use of the colon as a fill character.
133 uint8_t __use_range_fill_ : 1 {false};
134 };
135
136 // By not placing this constant in the formatter class it's not duplicated for
137 // char and wchar_t.
138 inline constexpr __fields __fields_integral{
139 .__sign_ = true,
140 .__alternate_form_ = true,
141 .__zero_padding_ = true,
142 .__locale_specific_form_ = true,
143 .__type_ = true};
144 inline constexpr __fields __fields_floating_point{
145 .__sign_ = true,
146 .__alternate_form_ = true,
147 .__zero_padding_ = true,
148 .__precision_ = true,
149 .__locale_specific_form_ = true,
150 .__type_ = true};
151 inline constexpr __fields __fields_string{.__precision_ = true, .__type_ = true};
152 inline constexpr __fields __fields_pointer{.__type_ = true};
153
154 # if _LIBCPP_STD_VER >= 23
155 inline constexpr __fields __fields_tuple{.__use_range_fill_ = true};
156 inline constexpr __fields __fields_range{.__use_range_fill_ = true};
157 inline constexpr __fields __fields_fill_align_width{};
158 # endif
159
160 enum class _LIBCPP_ENUM_VIS __alignment : uint8_t {
161 /// No alignment is set in the format string.
162 __default,
163 __left,
164 __center,
165 __right,
166 __zero_padding
167 };
168
169 enum class _LIBCPP_ENUM_VIS __sign : uint8_t {
170 /// No sign is set in the format string.
171 ///
172 /// The sign isn't allowed for certain format-types. By using this value
173 /// it's possible to detect whether or not the user explicitly set the sign
174 /// flag. For formatting purposes it behaves the same as \ref __minus.
175 __default,
176 __minus,
177 __plus,
178 __space
179 };
180
181 enum class _LIBCPP_ENUM_VIS __type : uint8_t {
182 __default,
183 __string,
184 __binary_lower_case,
185 __binary_upper_case,
186 __octal,
187 __decimal,
188 __hexadecimal_lower_case,
189 __hexadecimal_upper_case,
190 __pointer,
191 __char,
192 __hexfloat_lower_case,
193 __hexfloat_upper_case,
194 __scientific_lower_case,
195 __scientific_upper_case,
196 __fixed_lower_case,
197 __fixed_upper_case,
198 __general_lower_case,
199 __general_upper_case,
200 __debug
201 };
202
203 struct __std {
204 __alignment __alignment_ : 3;
205 __sign __sign_ : 2;
206 bool __alternate_form_ : 1;
207 bool __locale_specific_form_ : 1;
208 __type __type_;
209 };
210
211 struct __chrono {
212 __alignment __alignment_ : 3;
213 bool __locale_specific_form_ : 1;
214 bool __hour_ : 1;
215 bool __weekday_name_ : 1;
216 bool __weekday_ : 1;
217 bool __day_of_year_ : 1;
218 bool __week_of_year_ : 1;
219 bool __month_name_ : 1;
220 };
221
222 /// Contains the parsed formatting specifications.
223 ///
224 /// This contains information for both the std-format-spec and the
225 /// chrono-format-spec. This results in some unused members for both
226 /// specifications. However these unused members don't increase the size
227 /// of the structure.
228 ///
229 /// This struct doesn't cross ABI boundaries so its layout doesn't need to be
230 /// kept stable.
231 template <class _CharT>
232 struct __parsed_specifications {
233 union {
234 // The field __alignment_ is the first element in __std_ and __chrono_.
235 // This allows the code to always inspect this value regards which member
236 // of the union is the active member [class.union.general]/2.
237 //
238 // This is needed since the generic output routines handle the alignment of
239 // the output.
240 __alignment __alignment_ : 3;
241 __std __std_;
242 __chrono __chrono_;
243 };
244
245 /// The requested width.
246 ///
247 /// When the format-spec used an arg-id for this field it has already been
248 /// replaced with the value of that arg-id.
249 int32_t __width_;
250
251 /// The requested precision.
252 ///
253 /// When the format-spec used an arg-id for this field it has already been
254 /// replaced with the value of that arg-id.
255 int32_t __precision_;
256
257 _CharT __fill_;
258
__has_width__parsed_specifications259 _LIBCPP_HIDE_FROM_ABI constexpr bool __has_width() const { return __width_ > 0; }
260
__has_precision__parsed_specifications261 _LIBCPP_HIDE_FROM_ABI constexpr bool __has_precision() const { return __precision_ >= 0; }
262 };
263
264 // Validate the struct is small and cheap to copy since the struct is passed by
265 // value in formatting functions.
266 static_assert(sizeof(__parsed_specifications<char>) == 16);
267 static_assert(is_trivially_copyable_v<__parsed_specifications<char>>);
268 # ifndef _LIBCPP_HAS_NO_WIDE_CHARACTERS
269 static_assert(sizeof(__parsed_specifications<wchar_t>) == 16);
270 static_assert(is_trivially_copyable_v<__parsed_specifications<wchar_t>>);
271 # endif
272
273 /// The parser for the std-format-spec.
274 ///
275 /// Note this class is a member of std::formatter specializations. It's
276 /// expected developers will create their own formatter specializations that
277 /// inherit from the std::formatter specializations. This means this class
278 /// must be ABI stable. To aid the stability the unused bits in the class are
279 /// set to zero. That way they can be repurposed if a future revision of the
280 /// Standards adds new fields to std-format-spec.
281 template <class _CharT>
282 class _LIBCPP_TEMPLATE_VIS __parser {
283 public:
284 _LIBCPP_HIDE_FROM_ABI constexpr auto __parse(basic_format_parse_context<_CharT>& __parse_ctx, __fields __fields)
285 -> decltype(__parse_ctx.begin()) {
286
287 auto __begin = __parse_ctx.begin();
288 auto __end = __parse_ctx.end();
289 if (__begin == __end)
290 return __begin;
291
292 if (__parse_fill_align(__begin, __end, __fields.__use_range_fill_) && __begin == __end)
293 return __begin;
294
295 if (__fields.__sign_ && __parse_sign(__begin) && __begin == __end)
296 return __begin;
297
298 if (__fields.__alternate_form_ && __parse_alternate_form(__begin) && __begin == __end)
299 return __begin;
300
301 if (__fields.__zero_padding_ && __parse_zero_padding(__begin) && __begin == __end)
302 return __begin;
303
304 if (__parse_width(__begin, __end, __parse_ctx) && __begin == __end)
305 return __begin;
306
307 if (__fields.__precision_ && __parse_precision(__begin, __end, __parse_ctx) && __begin == __end)
308 return __begin;
309
310 if (__fields.__locale_specific_form_ && __parse_locale_specific_form(__begin) && __begin == __end)
311 return __begin;
312
313 if (__fields.__type_) {
314 __parse_type(__begin);
315
316 // When __type_ is false the calling parser is expected to do additional
317 // parsing. In that case that parser should do the end of format string
318 // validation.
319 if (__begin != __end && *__begin != _CharT('}'))
320 std::__throw_format_error("The format-spec should consume the input or end with a '}'");
321 }
322
323 return __begin;
324 }
325
326 /// \returns the `__parsed_specifications` with the resolved dynamic sizes..
327 _LIBCPP_HIDE_FROM_ABI
__get_parsed_std_specifications(auto & __ctx)328 __parsed_specifications<_CharT> __get_parsed_std_specifications(auto& __ctx) const {
329 return __parsed_specifications<_CharT>{
330 .__std_ = __std{.__alignment_ = __alignment_,
331 .__sign_ = __sign_,
332 .__alternate_form_ = __alternate_form_,
333 .__locale_specific_form_ = __locale_specific_form_,
334 .__type_ = __type_},
335 .__width_{__get_width(__ctx)},
336 .__precision_{__get_precision(__ctx)},
337 .__fill_{__fill_}};
338 }
339
__get_parsed_chrono_specifications(auto & __ctx)340 _LIBCPP_HIDE_FROM_ABI __parsed_specifications<_CharT> __get_parsed_chrono_specifications(auto& __ctx) const {
341 return __parsed_specifications<_CharT>{
342 .__chrono_ =
343 __chrono{.__alignment_ = __alignment_,
344 .__locale_specific_form_ = __locale_specific_form_,
345 .__hour_ = __hour_,
346 .__weekday_name_ = __weekday_name_,
347 .__weekday_ = __weekday_,
348 .__day_of_year_ = __day_of_year_,
349 .__week_of_year_ = __week_of_year_,
350 .__month_name_ = __month_name_},
351 .__width_{__get_width(__ctx)},
352 .__precision_{__get_precision(__ctx)},
353 .__fill_{__fill_}};
354 }
355
356 __alignment __alignment_ : 3 {__alignment::__default};
357 __sign __sign_ : 2 {__sign::__default};
358 bool __alternate_form_ : 1 {false};
359 bool __locale_specific_form_ : 1 {false};
360 bool __reserved_0_ : 1 {false};
361 __type __type_{__type::__default};
362
363 // These flags are only used for formatting chrono. Since the struct has
364 // padding space left it's added to this structure.
365 bool __hour_ : 1 {false};
366
367 bool __weekday_name_ : 1 {false};
368 bool __weekday_ : 1 {false};
369
370 bool __day_of_year_ : 1 {false};
371 bool __week_of_year_ : 1 {false};
372
373 bool __month_name_ : 1 {false};
374
375 uint8_t __reserved_1_ : 2 {0};
376 uint8_t __reserved_2_ : 6 {0};
377 // These two flags are only used internally and not part of the
378 // __parsed_specifications. Therefore put them at the end.
379 bool __width_as_arg_ : 1 {false};
380 bool __precision_as_arg_ : 1 {false};
381
382 /// The requested width, either the value or the arg-id.
383 int32_t __width_{0};
384
385 /// The requested precision, either the value or the arg-id.
386 int32_t __precision_{-1};
387
388 // LWG 3576 will probably change this to always accept a Unicode code point
389 // To avoid changing the size with that change align the field so when it
390 // becomes 32-bit its alignment will remain the same. That also means the
391 // size will remain the same. (D2572 addresses the solution for LWG 3576.)
392 _CharT __fill_{_CharT(' ')};
393
394 private:
__parse_alignment(_CharT __c)395 _LIBCPP_HIDE_FROM_ABI constexpr bool __parse_alignment(_CharT __c) {
396 switch (__c) {
397 case _CharT('<'):
398 __alignment_ = __alignment::__left;
399 return true;
400
401 case _CharT('^'):
402 __alignment_ = __alignment::__center;
403 return true;
404
405 case _CharT('>'):
406 __alignment_ = __alignment::__right;
407 return true;
408 }
409 return false;
410 }
411
412 // range-fill and tuple-fill are identical
413 template <contiguous_iterator _Iterator>
__parse_fill_align(_Iterator & __begin,_Iterator __end,bool __use_range_fill)414 _LIBCPP_HIDE_FROM_ABI constexpr bool __parse_fill_align(_Iterator& __begin, _Iterator __end, bool __use_range_fill) {
415 _LIBCPP_ASSERT(__begin != __end, "when called with an empty input the function will cause "
416 "undefined behavior by evaluating data not in the input");
417 if (__begin + 1 != __end) {
418 if (__parse_alignment(*(__begin + 1))) {
419 if (__use_range_fill && (*__begin == _CharT('{') || *__begin == _CharT('}') || *__begin == _CharT(':')))
420 std::__throw_format_error("The format-spec range-fill field contains an invalid character");
421 else if (*__begin == _CharT('{') || *__begin == _CharT('}'))
422 std::__throw_format_error("The format-spec fill field contains an invalid character");
423
424 __fill_ = *__begin;
425 __begin += 2;
426 return true;
427 }
428 }
429
430 if (!__parse_alignment(*__begin))
431 return false;
432
433 ++__begin;
434 return true;
435 }
436
437 template <contiguous_iterator _Iterator>
__parse_sign(_Iterator & __begin)438 _LIBCPP_HIDE_FROM_ABI constexpr bool __parse_sign(_Iterator& __begin) {
439 switch (*__begin) {
440 case _CharT('-'):
441 __sign_ = __sign::__minus;
442 break;
443 case _CharT('+'):
444 __sign_ = __sign::__plus;
445 break;
446 case _CharT(' '):
447 __sign_ = __sign::__space;
448 break;
449 default:
450 return false;
451 }
452 ++__begin;
453 return true;
454 }
455
456 template <contiguous_iterator _Iterator>
__parse_alternate_form(_Iterator & __begin)457 _LIBCPP_HIDE_FROM_ABI constexpr bool __parse_alternate_form(_Iterator& __begin) {
458 if (*__begin != _CharT('#'))
459 return false;
460
461 __alternate_form_ = true;
462 ++__begin;
463 return true;
464 }
465
466 template <contiguous_iterator _Iterator>
__parse_zero_padding(_Iterator & __begin)467 _LIBCPP_HIDE_FROM_ABI constexpr bool __parse_zero_padding(_Iterator& __begin) {
468 if (*__begin != _CharT('0'))
469 return false;
470
471 if (__alignment_ == __alignment::__default)
472 __alignment_ = __alignment::__zero_padding;
473 ++__begin;
474 return true;
475 }
476
477 template <contiguous_iterator _Iterator>
__parse_width(_Iterator & __begin,_Iterator __end,auto & __parse_ctx)478 _LIBCPP_HIDE_FROM_ABI constexpr bool __parse_width(_Iterator& __begin, _Iterator __end, auto& __parse_ctx) {
479 if (*__begin == _CharT('0'))
480 std::__throw_format_error("A format-spec width field shouldn't have a leading zero");
481
482 if (*__begin == _CharT('{')) {
483 __format::__parse_number_result __r = __format_spec::__parse_arg_id(++__begin, __end, __parse_ctx);
484 __width_as_arg_ = true;
485 __width_ = __r.__value;
486 __begin = __r.__last;
487 return true;
488 }
489
490 if (*__begin < _CharT('0') || *__begin > _CharT('9'))
491 return false;
492
493 __format::__parse_number_result __r = __format::__parse_number(__begin, __end);
494 __width_ = __r.__value;
495 _LIBCPP_ASSERT(__width_ != 0, "A zero value isn't allowed and should be impossible, "
496 "due to validations in this function");
497 __begin = __r.__last;
498 return true;
499 }
500
501 template <contiguous_iterator _Iterator>
__parse_precision(_Iterator & __begin,_Iterator __end,auto & __parse_ctx)502 _LIBCPP_HIDE_FROM_ABI constexpr bool __parse_precision(_Iterator& __begin, _Iterator __end, auto& __parse_ctx) {
503 if (*__begin != _CharT('.'))
504 return false;
505
506 ++__begin;
507 if (__begin == __end)
508 std::__throw_format_error("End of input while parsing format-spec precision");
509
510 if (*__begin == _CharT('{')) {
511 __format::__parse_number_result __arg_id = __format_spec::__parse_arg_id(++__begin, __end, __parse_ctx);
512 __precision_as_arg_ = true;
513 __precision_ = __arg_id.__value;
514 __begin = __arg_id.__last;
515 return true;
516 }
517
518 if (*__begin < _CharT('0') || *__begin > _CharT('9'))
519 std::__throw_format_error("The format-spec precision field doesn't contain a value or arg-id");
520
521 __format::__parse_number_result __r = __format::__parse_number(__begin, __end);
522 __precision_ = __r.__value;
523 __precision_as_arg_ = false;
524 __begin = __r.__last;
525 return true;
526 }
527
528 template <contiguous_iterator _Iterator>
__parse_locale_specific_form(_Iterator & __begin)529 _LIBCPP_HIDE_FROM_ABI constexpr bool __parse_locale_specific_form(_Iterator& __begin) {
530 if (*__begin != _CharT('L'))
531 return false;
532
533 __locale_specific_form_ = true;
534 ++__begin;
535 return true;
536 }
537
538 template <contiguous_iterator _Iterator>
__parse_type(_Iterator & __begin)539 _LIBCPP_HIDE_FROM_ABI constexpr void __parse_type(_Iterator& __begin) {
540 // Determines the type. It does not validate whether the selected type is
541 // valid. Most formatters have optional fields that are only allowed for
542 // certain types. These parsers need to do validation after the type has
543 // been parsed. So its easier to implement the validation for all types in
544 // the specific parse function.
545 switch (*__begin) {
546 case 'A':
547 __type_ = __type::__hexfloat_upper_case;
548 break;
549 case 'B':
550 __type_ = __type::__binary_upper_case;
551 break;
552 case 'E':
553 __type_ = __type::__scientific_upper_case;
554 break;
555 case 'F':
556 __type_ = __type::__fixed_upper_case;
557 break;
558 case 'G':
559 __type_ = __type::__general_upper_case;
560 break;
561 case 'X':
562 __type_ = __type::__hexadecimal_upper_case;
563 break;
564 case 'a':
565 __type_ = __type::__hexfloat_lower_case;
566 break;
567 case 'b':
568 __type_ = __type::__binary_lower_case;
569 break;
570 case 'c':
571 __type_ = __type::__char;
572 break;
573 case 'd':
574 __type_ = __type::__decimal;
575 break;
576 case 'e':
577 __type_ = __type::__scientific_lower_case;
578 break;
579 case 'f':
580 __type_ = __type::__fixed_lower_case;
581 break;
582 case 'g':
583 __type_ = __type::__general_lower_case;
584 break;
585 case 'o':
586 __type_ = __type::__octal;
587 break;
588 case 'p':
589 __type_ = __type::__pointer;
590 break;
591 case 's':
592 __type_ = __type::__string;
593 break;
594 case 'x':
595 __type_ = __type::__hexadecimal_lower_case;
596 break;
597 # if _LIBCPP_STD_VER >= 23
598 case '?':
599 __type_ = __type::__debug;
600 break;
601 # endif
602 default:
603 return;
604 }
605 ++__begin;
606 }
607
608 _LIBCPP_HIDE_FROM_ABI
__get_width(auto & __ctx)609 int32_t __get_width(auto& __ctx) const {
610 if (!__width_as_arg_)
611 return __width_;
612
613 return __format_spec::__substitute_arg_id(__ctx.arg(__width_));
614 }
615
616 _LIBCPP_HIDE_FROM_ABI
__get_precision(auto & __ctx)617 int32_t __get_precision(auto& __ctx) const {
618 if (!__precision_as_arg_)
619 return __precision_;
620
621 return __format_spec::__substitute_arg_id(__ctx.arg(__precision_));
622 }
623 };
624
625 // Validates whether the reserved bitfields don't change the size.
626 static_assert(sizeof(__parser<char>) == 16);
627 # ifndef _LIBCPP_HAS_NO_WIDE_CHARACTERS
628 static_assert(sizeof(__parser<wchar_t>) == 16);
629 # endif
630
__process_display_type_string(__format_spec::__type __type)631 _LIBCPP_HIDE_FROM_ABI constexpr void __process_display_type_string(__format_spec::__type __type) {
632 switch (__type) {
633 case __format_spec::__type::__default:
634 case __format_spec::__type::__string:
635 case __format_spec::__type::__debug:
636 break;
637
638 default:
639 std::__throw_format_error("The format-spec type has a type not supported for a string argument");
640 }
641 }
642
643 template <class _CharT>
__process_display_type_bool_string(__parser<_CharT> & __parser)644 _LIBCPP_HIDE_FROM_ABI constexpr void __process_display_type_bool_string(__parser<_CharT>& __parser) {
645 if (__parser.__sign_ != __sign::__default)
646 std::__throw_format_error("A sign field isn't allowed in this format-spec");
647
648 if (__parser.__alternate_form_)
649 std::__throw_format_error("An alternate form field isn't allowed in this format-spec");
650
651 if (__parser.__alignment_ == __alignment::__zero_padding)
652 std::__throw_format_error("A zero-padding field isn't allowed in this format-spec");
653
654 if (__parser.__alignment_ == __alignment::__default)
655 __parser.__alignment_ = __alignment::__left;
656 }
657
658 template <class _CharT>
__process_display_type_char(__parser<_CharT> & __parser)659 _LIBCPP_HIDE_FROM_ABI constexpr void __process_display_type_char(__parser<_CharT>& __parser) {
660 __format_spec::__process_display_type_bool_string(__parser);
661 }
662
663 template <class _CharT>
__process_parsed_bool(__parser<_CharT> & __parser)664 _LIBCPP_HIDE_FROM_ABI constexpr void __process_parsed_bool(__parser<_CharT>& __parser) {
665 switch (__parser.__type_) {
666 case __format_spec::__type::__default:
667 case __format_spec::__type::__string:
668 __format_spec::__process_display_type_bool_string(__parser);
669 break;
670
671 case __format_spec::__type::__binary_lower_case:
672 case __format_spec::__type::__binary_upper_case:
673 case __format_spec::__type::__octal:
674 case __format_spec::__type::__decimal:
675 case __format_spec::__type::__hexadecimal_lower_case:
676 case __format_spec::__type::__hexadecimal_upper_case:
677 break;
678
679 default:
680 std::__throw_format_error("The format-spec type has a type not supported for a bool argument");
681 }
682 }
683
684 template <class _CharT>
__process_parsed_char(__parser<_CharT> & __parser)685 _LIBCPP_HIDE_FROM_ABI constexpr void __process_parsed_char(__parser<_CharT>& __parser) {
686 switch (__parser.__type_) {
687 case __format_spec::__type::__default:
688 case __format_spec::__type::__char:
689 case __format_spec::__type::__debug:
690 __format_spec::__process_display_type_char(__parser);
691 break;
692
693 case __format_spec::__type::__binary_lower_case:
694 case __format_spec::__type::__binary_upper_case:
695 case __format_spec::__type::__octal:
696 case __format_spec::__type::__decimal:
697 case __format_spec::__type::__hexadecimal_lower_case:
698 case __format_spec::__type::__hexadecimal_upper_case:
699 break;
700
701 default:
702 std::__throw_format_error("The format-spec type has a type not supported for a char argument");
703 }
704 }
705
706 template <class _CharT>
__process_parsed_integer(__parser<_CharT> & __parser)707 _LIBCPP_HIDE_FROM_ABI constexpr void __process_parsed_integer(__parser<_CharT>& __parser) {
708 switch (__parser.__type_) {
709 case __format_spec::__type::__default:
710 case __format_spec::__type::__binary_lower_case:
711 case __format_spec::__type::__binary_upper_case:
712 case __format_spec::__type::__octal:
713 case __format_spec::__type::__decimal:
714 case __format_spec::__type::__hexadecimal_lower_case:
715 case __format_spec::__type::__hexadecimal_upper_case:
716 break;
717
718 case __format_spec::__type::__char:
719 __format_spec::__process_display_type_char(__parser);
720 break;
721
722 default:
723 std::__throw_format_error("The format-spec type has a type not supported for an integer argument");
724 }
725 }
726
727 template <class _CharT>
__process_parsed_floating_point(__parser<_CharT> & __parser)728 _LIBCPP_HIDE_FROM_ABI constexpr void __process_parsed_floating_point(__parser<_CharT>& __parser) {
729 switch (__parser.__type_) {
730 case __format_spec::__type::__default:
731 case __format_spec::__type::__hexfloat_lower_case:
732 case __format_spec::__type::__hexfloat_upper_case:
733 // Precision specific behavior will be handled later.
734 break;
735 case __format_spec::__type::__scientific_lower_case:
736 case __format_spec::__type::__scientific_upper_case:
737 case __format_spec::__type::__fixed_lower_case:
738 case __format_spec::__type::__fixed_upper_case:
739 case __format_spec::__type::__general_lower_case:
740 case __format_spec::__type::__general_upper_case:
741 if (!__parser.__precision_as_arg_ && __parser.__precision_ == -1)
742 // Set the default precision for the call to to_chars.
743 __parser.__precision_ = 6;
744 break;
745
746 default:
747 std::__throw_format_error("The format-spec type has a type not supported for a floating-point argument");
748 }
749 }
750
__process_display_type_pointer(__format_spec::__type __type)751 _LIBCPP_HIDE_FROM_ABI constexpr void __process_display_type_pointer(__format_spec::__type __type) {
752 switch (__type) {
753 case __format_spec::__type::__default:
754 case __format_spec::__type::__pointer:
755 break;
756
757 default:
758 std::__throw_format_error("The format-spec type has a type not supported for a pointer argument");
759 }
760 }
761
762 template <contiguous_iterator _Iterator>
763 struct __column_width_result {
764 /// The number of output columns.
765 size_t __width_;
766 /// One beyond the last code unit used in the estimation.
767 ///
768 /// This limits the original output to fit in the wanted number of columns.
769 _Iterator __last_;
770 };
771
772 template <contiguous_iterator _Iterator>
773 __column_width_result(size_t, _Iterator) -> __column_width_result<_Iterator>;
774
775 /// Since a column width can be two it's possible that the requested column
776 /// width can't be achieved. Depending on the intended usage the policy can be
777 /// selected.
778 /// - When used as precision the maximum width may not be exceeded and the
779 /// result should be "rounded down" to the previous boundary.
780 /// - When used as a width we're done once the minimum is reached, but
781 /// exceeding is not an issue. Rounding down is an issue since that will
782 /// result in writing fill characters. Therefore the result needs to be
783 /// "rounded up".
784 enum class __column_width_rounding { __down, __up };
785
786 # ifndef _LIBCPP_HAS_NO_UNICODE
787
788 namespace __detail {
789
790 /// Converts a code point to the column width.
791 ///
792 /// The estimations are conforming to [format.string.general]/11
793 ///
794 /// This version expects a value less than 0x1'0000, which is a 3-byte UTF-8
795 /// character.
__column_width_3(uint32_t __c)796 _LIBCPP_HIDE_FROM_ABI constexpr int __column_width_3(uint32_t __c) noexcept {
797 _LIBCPP_ASSERT(__c < 0x10000, "Use __column_width_4 or __column_width for larger values");
798
799 // clang-format off
800 return 1 + (__c >= 0x1100 && (__c <= 0x115f ||
801 (__c >= 0x2329 && (__c <= 0x232a ||
802 (__c >= 0x2e80 && (__c <= 0x303e ||
803 (__c >= 0x3040 && (__c <= 0xa4cf ||
804 (__c >= 0xac00 && (__c <= 0xd7a3 ||
805 (__c >= 0xf900 && (__c <= 0xfaff ||
806 (__c >= 0xfe10 && (__c <= 0xfe19 ||
807 (__c >= 0xfe30 && (__c <= 0xfe6f ||
808 (__c >= 0xff00 && (__c <= 0xff60 ||
809 (__c >= 0xffe0 && (__c <= 0xffe6
810 ))))))))))))))))))));
811 // clang-format on
812 }
813
814 /// @overload
815 ///
816 /// This version expects a value greater than or equal to 0x1'0000, which is a
817 /// 4-byte UTF-8 character.
__column_width_4(uint32_t __c)818 _LIBCPP_HIDE_FROM_ABI constexpr int __column_width_4(uint32_t __c) noexcept {
819 _LIBCPP_ASSERT(__c >= 0x10000, "Use __column_width_3 or __column_width for smaller values");
820
821 // clang-format off
822 return 1 + (__c >= 0x1'f300 && (__c <= 0x1'f64f ||
823 (__c >= 0x1'f900 && (__c <= 0x1'f9ff ||
824 (__c >= 0x2'0000 && (__c <= 0x2'fffd ||
825 (__c >= 0x3'0000 && (__c <= 0x3'fffd
826 ))))))));
827 // clang-format on
828 }
829
830 /// @overload
831 ///
832 /// The general case, accepting all values.
__column_width(uint32_t __c)833 _LIBCPP_HIDE_FROM_ABI constexpr int __column_width(uint32_t __c) noexcept {
834 if (__c < 0x10000)
835 return __detail::__column_width_3(__c);
836
837 return __detail::__column_width_4(__c);
838 }
839
840 template <contiguous_iterator _Iterator>
__estimate_column_width_grapheme_clustering(_Iterator __first,_Iterator __last,size_t __maximum,__column_width_rounding __rounding)841 _LIBCPP_HIDE_FROM_ABI constexpr __column_width_result<_Iterator> __estimate_column_width_grapheme_clustering(
842 _Iterator __first, _Iterator __last, size_t __maximum, __column_width_rounding __rounding) noexcept {
843 using _CharT = iter_value_t<_Iterator>;
844 __unicode::__extended_grapheme_cluster_view<_CharT> __view{__first, __last};
845
846 __column_width_result<_Iterator> __result{0, __first};
847 while (__result.__last_ != __last && __result.__width_ <= __maximum) {
848 typename __unicode::__extended_grapheme_cluster_view<_CharT>::__cluster __cluster = __view.__consume();
849 int __width = __detail::__column_width(__cluster.__code_point_);
850
851 // When the next entry would exceed the maximum width the previous width
852 // might be returned. For example when a width of 100 is requested the
853 // returned width might be 99, since the next code point has an estimated
854 // column width of 2. This depends on the rounding flag.
855 // When the maximum is exceeded the loop will abort the next iteration.
856 if (__rounding == __column_width_rounding::__down && __result.__width_ + __width > __maximum)
857 return __result;
858
859 __result.__width_ += __width;
860 __result.__last_ = __cluster.__last_;
861 }
862
863 return __result;
864 }
865
866 } // namespace __detail
867
868 // Unicode can be stored in several formats: UTF-8, UTF-16, and UTF-32.
869 // Depending on format the relation between the number of code units stored and
870 // the number of output columns differs. The first relation is the number of
871 // code units forming a code point. (The text assumes the code units are
872 // unsigned.)
873 // - UTF-8 The number of code units is between one and four. The first 127
874 // Unicode code points match the ASCII character set. When the highest bit is
875 // set it means the code point has more than one code unit.
876 // - UTF-16: The number of code units is between 1 and 2. When the first
877 // code unit is in the range [0xd800,0xdfff) it means the code point uses two
878 // code units.
879 // - UTF-32: The number of code units is always one.
880 //
881 // The code point to the number of columns is specified in
882 // [format.string.std]/11. This list might change in the future.
883 //
884 // Another thing to be taken into account is Grapheme clustering. This means
885 // that in some cases multiple code points are combined one element in the
886 // output. For example:
887 // - an ASCII character with a combined diacritical mark
888 // - an emoji with a skin tone modifier
889 // - a group of combined people emoji to create a family
890 // - a combination of flag emoji
891 //
892 // See also:
893 // - [format.string.general]/11
894 // - https://en.wikipedia.org/wiki/UTF-8#Encoding
895 // - https://en.wikipedia.org/wiki/UTF-16#U+D800_to_U+DFFF
896
__is_ascii(char32_t __c)897 _LIBCPP_HIDE_FROM_ABI constexpr bool __is_ascii(char32_t __c) { return __c < 0x80; }
898
899 /// Determines the number of output columns needed to render the input.
900 ///
901 /// \note When the scanner encounters malformed Unicode it acts as-if every
902 /// code unit is a one column code point. Typically a terminal uses the same
903 /// strategy and replaces every malformed code unit with a one column
904 /// replacement character.
905 ///
906 /// \param __first Points to the first element of the input range.
907 /// \param __last Points beyond the last element of the input range.
908 /// \param __maximum The maximum number of output columns. The returned number
909 /// of estimated output columns will not exceed this value.
910 /// \param __rounding Selects the rounding method.
911 /// \c __down result.__width_ <= __maximum
912 /// \c __up result.__width_ <= __maximum + 1
913 template <class _CharT, class _Iterator = typename basic_string_view<_CharT>::const_iterator>
__estimate_column_width(basic_string_view<_CharT> __str,size_t __maximum,__column_width_rounding __rounding)914 _LIBCPP_HIDE_FROM_ABI constexpr __column_width_result<_Iterator> __estimate_column_width(
915 basic_string_view<_CharT> __str, size_t __maximum, __column_width_rounding __rounding) noexcept {
916 // The width estimation is done in two steps:
917 // - Quickly process for the ASCII part. ASCII has the following properties
918 // - One code unit is one code point
919 // - Every code point has an estimated width of one
920 // - When needed it will a Unicode Grapheme clustering algorithm to find
921 // the proper place for truncation.
922
923 if (__str.empty() || __maximum == 0)
924 return {0, __str.begin()};
925
926 // ASCII has one caveat; when an ASCII character is followed by a non-ASCII
927 // character they might be part of an extended grapheme cluster. For example:
928 // an ASCII letter and a COMBINING ACUTE ACCENT
929 // The truncate should happen after the COMBINING ACUTE ACCENT. Therefore we
930 // need to scan one code unit beyond the requested precision. When this code
931 // unit is non-ASCII we omit the current code unit and let the Grapheme
932 // clustering algorithm do its work.
933 auto __it = __str.begin();
934 if (__format_spec::__is_ascii(*__it)) {
935 do {
936 --__maximum;
937 ++__it;
938 if (__it == __str.end())
939 return {__str.size(), __str.end()};
940
941 if (__maximum == 0) {
942 if (__format_spec::__is_ascii(*__it))
943 return {static_cast<size_t>(__it - __str.begin()), __it};
944
945 break;
946 }
947 } while (__format_spec::__is_ascii(*__it));
948 --__it;
949 ++__maximum;
950 }
951
952 ptrdiff_t __ascii_size = __it - __str.begin();
953 __column_width_result __result =
954 __detail::__estimate_column_width_grapheme_clustering(__it, __str.end(), __maximum, __rounding);
955
956 __result.__width_ += __ascii_size;
957 return __result;
958 }
959 # else // !defined(_LIBCPP_HAS_NO_UNICODE)
960 template <class _CharT>
961 _LIBCPP_HIDE_FROM_ABI constexpr __column_width_result<typename basic_string_view<_CharT>::const_iterator>
__estimate_column_width(basic_string_view<_CharT> __str,size_t __maximum,__column_width_rounding)962 __estimate_column_width(basic_string_view<_CharT> __str, size_t __maximum, __column_width_rounding) noexcept {
963 // When Unicode isn't supported assume ASCII and every code unit is one code
964 // point. In ASCII the estimated column width is always one. Thus there's no
965 // need for rounding.
966 size_t __width_ = _VSTD::min(__str.size(), __maximum);
967 return {__width_, __str.begin() + __width_};
968 }
969
970 # endif // !defined(_LIBCPP_HAS_NO_UNICODE)
971
972 } // namespace __format_spec
973
974 #endif //_LIBCPP_STD_VER >= 20
975
976 _LIBCPP_END_NAMESPACE_STD
977
978 _LIBCPP_POP_MACROS
979
980 #endif // _LIBCPP___FORMAT_PARSER_STD_FORMAT_SPEC_H
981