1 //===------------------------- locale.cpp ---------------------------------===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is dual licensed under the MIT and the University of Illinois Open 6 // Source Licenses. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 10 // On Solaris, we need to define something to make the C99 parts of localeconv 11 // visible. 12 #ifdef __sun__ 13 #define _LCONV_C99 14 #endif 15 16 #include "string" 17 #include "locale" 18 #include "codecvt" 19 #include "vector" 20 #include "algorithm" 21 #include "typeinfo" 22 #ifndef _LIBCPP_NO_EXCEPTIONS 23 # include "type_traits" 24 #endif 25 #include "clocale" 26 #include "cstring" 27 #if defined(_LIBCPP_MSVCRT) 28 #define _CTYPE_DISABLE_MACROS 29 #endif 30 #include "cwctype" 31 #include "__sso_allocator" 32 #if defined(_LIBCPP_MSVCRT) || defined(__MINGW32__) 33 #include "support/win32/locale_win32.h" 34 #elif !defined(__BIONIC__) 35 #include <langinfo.h> 36 #endif 37 #include <stdlib.h> 38 #include <stdio.h> 39 #include "include/atomic_support.h" 40 #include "__undef_macros" 41 42 // On Linux, wint_t and wchar_t have different signed-ness, and this causes 43 // lots of noise in the build log, but no bugs that I know of. 44 #if defined(__clang__) 45 #pragma clang diagnostic ignored "-Wsign-conversion" 46 #endif 47 48 _LIBCPP_BEGIN_NAMESPACE_STD 49 50 struct __libcpp_unique_locale { 51 __libcpp_unique_locale(const char* nm) : __loc_(newlocale(LC_ALL_MASK, nm, 0)) {} 52 53 ~__libcpp_unique_locale() { 54 if (__loc_) 55 freelocale(__loc_); 56 } 57 58 explicit operator bool() const { return __loc_; } 59 60 locale_t& get() { return __loc_; } 61 62 locale_t __loc_; 63 private: 64 __libcpp_unique_locale(__libcpp_unique_locale const&); 65 __libcpp_unique_locale& operator=(__libcpp_unique_locale const&); 66 }; 67 68 #ifdef __cloc_defined 69 locale_t __cloc() { 70 // In theory this could create a race condition. In practice 71 // the race condition is non-fatal since it will just create 72 // a little resource leak. Better approach would be appreciated. 73 static locale_t result = newlocale(LC_ALL_MASK, "C", 0); 74 return result; 75 } 76 #endif // __cloc_defined 77 78 namespace { 79 80 struct release 81 { 82 void operator()(locale::facet* p) {p->__release_shared();} 83 }; 84 85 template <class T, class A0> 86 inline 87 T& 88 make(A0 a0) 89 { 90 static typename aligned_storage<sizeof(T)>::type buf; 91 auto *obj = ::new (&buf) T(a0); 92 return *obj; 93 } 94 95 template <class T, class A0, class A1> 96 inline 97 T& 98 make(A0 a0, A1 a1) 99 { 100 static typename aligned_storage<sizeof(T)>::type buf; 101 ::new (&buf) T(a0, a1); 102 return *reinterpret_cast<T*>(&buf); 103 } 104 105 template <class T, class A0, class A1, class A2> 106 inline 107 T& 108 make(A0 a0, A1 a1, A2 a2) 109 { 110 static typename aligned_storage<sizeof(T)>::type buf; 111 auto *obj = ::new (&buf) T(a0, a1, a2); 112 return *obj; 113 } 114 115 template <typename T, size_t N> 116 inline 117 _LIBCPP_CONSTEXPR 118 size_t 119 countof(const T (&)[N]) 120 { 121 return N; 122 } 123 124 template <typename T> 125 inline 126 _LIBCPP_CONSTEXPR 127 size_t 128 countof(const T * const begin, const T * const end) 129 { 130 return static_cast<size_t>(end - begin); 131 } 132 133 _LIBCPP_NORETURN static void __throw_runtime_error(const string &msg) 134 { 135 #ifndef _LIBCPP_NO_EXCEPTIONS 136 throw runtime_error(msg); 137 #else 138 (void)msg; 139 _VSTD::abort(); 140 #endif 141 } 142 143 } 144 145 #if defined(_AIX) 146 // Set priority to INT_MIN + 256 + 150 147 # pragma priority ( -2147483242 ) 148 #endif 149 150 const locale::category locale::none; 151 const locale::category locale::collate; 152 const locale::category locale::ctype; 153 const locale::category locale::monetary; 154 const locale::category locale::numeric; 155 const locale::category locale::time; 156 const locale::category locale::messages; 157 const locale::category locale::all; 158 159 class _LIBCPP_HIDDEN locale::__imp 160 : public facet 161 { 162 enum {N = 28}; 163 #if defined(_LIBCPP_COMPILER_MSVC) 164 // FIXME: MSVC doesn't support aligned parameters by value. 165 // I can't get the __sso_allocator to work here 166 // for MSVC I think for this reason. 167 vector<facet*> facets_; 168 #else 169 vector<facet*, __sso_allocator<facet*, N> > facets_; 170 #endif 171 string name_; 172 public: 173 explicit __imp(size_t refs = 0); 174 explicit __imp(const string& name, size_t refs = 0); 175 __imp(const __imp&); 176 __imp(const __imp&, const string&, locale::category c); 177 __imp(const __imp& other, const __imp& one, locale::category c); 178 __imp(const __imp&, facet* f, long id); 179 ~__imp(); 180 181 const string& name() const {return name_;} 182 bool has_facet(long id) const 183 {return static_cast<size_t>(id) < facets_.size() && facets_[static_cast<size_t>(id)];} 184 const locale::facet* use_facet(long id) const; 185 186 static const locale& make_classic(); 187 static locale& make_global(); 188 private: 189 void install(facet* f, long id); 190 template <class F> void install(F* f) {install(f, f->id.__get());} 191 template <class F> void install_from(const __imp& other); 192 }; 193 194 locale::__imp::__imp(size_t refs) 195 : facet(refs), 196 facets_(N), 197 name_("C") 198 { 199 facets_.clear(); 200 install(&make<_VSTD::collate<char> >(1u)); 201 install(&make<_VSTD::collate<wchar_t> >(1u)); 202 install(&make<_VSTD::ctype<char> >(nullptr, false, 1u)); 203 install(&make<_VSTD::ctype<wchar_t> >(1u)); 204 install(&make<codecvt<char, char, mbstate_t> >(1u)); 205 install(&make<codecvt<wchar_t, char, mbstate_t> >(1u)); 206 install(&make<codecvt<char16_t, char, mbstate_t> >(1u)); 207 install(&make<codecvt<char32_t, char, mbstate_t> >(1u)); 208 install(&make<numpunct<char> >(1u)); 209 install(&make<numpunct<wchar_t> >(1u)); 210 install(&make<num_get<char> >(1u)); 211 install(&make<num_get<wchar_t> >(1u)); 212 install(&make<num_put<char> >(1u)); 213 install(&make<num_put<wchar_t> >(1u)); 214 install(&make<moneypunct<char, false> >(1u)); 215 install(&make<moneypunct<char, true> >(1u)); 216 install(&make<moneypunct<wchar_t, false> >(1u)); 217 install(&make<moneypunct<wchar_t, true> >(1u)); 218 install(&make<money_get<char> >(1u)); 219 install(&make<money_get<wchar_t> >(1u)); 220 install(&make<money_put<char> >(1u)); 221 install(&make<money_put<wchar_t> >(1u)); 222 install(&make<time_get<char> >(1u)); 223 install(&make<time_get<wchar_t> >(1u)); 224 install(&make<time_put<char> >(1u)); 225 install(&make<time_put<wchar_t> >(1u)); 226 install(&make<_VSTD::messages<char> >(1u)); 227 install(&make<_VSTD::messages<wchar_t> >(1u)); 228 } 229 230 locale::__imp::__imp(const string& name, size_t refs) 231 : facet(refs), 232 facets_(N), 233 name_(name) 234 { 235 #ifndef _LIBCPP_NO_EXCEPTIONS 236 try 237 { 238 #endif // _LIBCPP_NO_EXCEPTIONS 239 facets_ = locale::classic().__locale_->facets_; 240 for (unsigned i = 0; i < facets_.size(); ++i) 241 if (facets_[i]) 242 facets_[i]->__add_shared(); 243 install(new collate_byname<char>(name_)); 244 install(new collate_byname<wchar_t>(name_)); 245 install(new ctype_byname<char>(name_)); 246 install(new ctype_byname<wchar_t>(name_)); 247 install(new codecvt_byname<char, char, mbstate_t>(name_)); 248 install(new codecvt_byname<wchar_t, char, mbstate_t>(name_)); 249 install(new codecvt_byname<char16_t, char, mbstate_t>(name_)); 250 install(new codecvt_byname<char32_t, char, mbstate_t>(name_)); 251 install(new numpunct_byname<char>(name_)); 252 install(new numpunct_byname<wchar_t>(name_)); 253 install(new moneypunct_byname<char, false>(name_)); 254 install(new moneypunct_byname<char, true>(name_)); 255 install(new moneypunct_byname<wchar_t, false>(name_)); 256 install(new moneypunct_byname<wchar_t, true>(name_)); 257 install(new time_get_byname<char>(name_)); 258 install(new time_get_byname<wchar_t>(name_)); 259 install(new time_put_byname<char>(name_)); 260 install(new time_put_byname<wchar_t>(name_)); 261 install(new messages_byname<char>(name_)); 262 install(new messages_byname<wchar_t>(name_)); 263 #ifndef _LIBCPP_NO_EXCEPTIONS 264 } 265 catch (...) 266 { 267 for (unsigned i = 0; i < facets_.size(); ++i) 268 if (facets_[i]) 269 facets_[i]->__release_shared(); 270 throw; 271 } 272 #endif // _LIBCPP_NO_EXCEPTIONS 273 } 274 275 // NOTE avoid the `base class should be explicitly initialized in the 276 // copy constructor` warning emitted by GCC 277 #if defined(__clang__) || _GNUC_VER >= 406 278 #pragma GCC diagnostic push 279 #pragma GCC diagnostic ignored "-Wextra" 280 #endif 281 282 locale::__imp::__imp(const __imp& other) 283 : facets_(max<size_t>(N, other.facets_.size())), 284 name_(other.name_) 285 { 286 facets_ = other.facets_; 287 for (unsigned i = 0; i < facets_.size(); ++i) 288 if (facets_[i]) 289 facets_[i]->__add_shared(); 290 } 291 292 #if defined(__clang__) || _GNUC_VER >= 406 293 #pragma GCC diagnostic pop 294 #endif 295 296 locale::__imp::__imp(const __imp& other, const string& name, locale::category c) 297 : facets_(N), 298 name_("*") 299 { 300 facets_ = other.facets_; 301 for (unsigned i = 0; i < facets_.size(); ++i) 302 if (facets_[i]) 303 facets_[i]->__add_shared(); 304 #ifndef _LIBCPP_NO_EXCEPTIONS 305 try 306 { 307 #endif // _LIBCPP_NO_EXCEPTIONS 308 if (c & locale::collate) 309 { 310 install(new collate_byname<char>(name)); 311 install(new collate_byname<wchar_t>(name)); 312 } 313 if (c & locale::ctype) 314 { 315 install(new ctype_byname<char>(name)); 316 install(new ctype_byname<wchar_t>(name)); 317 install(new codecvt_byname<char, char, mbstate_t>(name)); 318 install(new codecvt_byname<wchar_t, char, mbstate_t>(name)); 319 install(new codecvt_byname<char16_t, char, mbstate_t>(name)); 320 install(new codecvt_byname<char32_t, char, mbstate_t>(name)); 321 } 322 if (c & locale::monetary) 323 { 324 install(new moneypunct_byname<char, false>(name)); 325 install(new moneypunct_byname<char, true>(name)); 326 install(new moneypunct_byname<wchar_t, false>(name)); 327 install(new moneypunct_byname<wchar_t, true>(name)); 328 } 329 if (c & locale::numeric) 330 { 331 install(new numpunct_byname<char>(name)); 332 install(new numpunct_byname<wchar_t>(name)); 333 } 334 if (c & locale::time) 335 { 336 install(new time_get_byname<char>(name)); 337 install(new time_get_byname<wchar_t>(name)); 338 install(new time_put_byname<char>(name)); 339 install(new time_put_byname<wchar_t>(name)); 340 } 341 if (c & locale::messages) 342 { 343 install(new messages_byname<char>(name)); 344 install(new messages_byname<wchar_t>(name)); 345 } 346 #ifndef _LIBCPP_NO_EXCEPTIONS 347 } 348 catch (...) 349 { 350 for (unsigned i = 0; i < facets_.size(); ++i) 351 if (facets_[i]) 352 facets_[i]->__release_shared(); 353 throw; 354 } 355 #endif // _LIBCPP_NO_EXCEPTIONS 356 } 357 358 template<class F> 359 inline 360 void 361 locale::__imp::install_from(const locale::__imp& one) 362 { 363 long id = F::id.__get(); 364 install(const_cast<F*>(static_cast<const F*>(one.use_facet(id))), id); 365 } 366 367 locale::__imp::__imp(const __imp& other, const __imp& one, locale::category c) 368 : facets_(N), 369 name_("*") 370 { 371 facets_ = other.facets_; 372 for (unsigned i = 0; i < facets_.size(); ++i) 373 if (facets_[i]) 374 facets_[i]->__add_shared(); 375 #ifndef _LIBCPP_NO_EXCEPTIONS 376 try 377 { 378 #endif // _LIBCPP_NO_EXCEPTIONS 379 if (c & locale::collate) 380 { 381 install_from<_VSTD::collate<char> >(one); 382 install_from<_VSTD::collate<wchar_t> >(one); 383 } 384 if (c & locale::ctype) 385 { 386 install_from<_VSTD::ctype<char> >(one); 387 install_from<_VSTD::ctype<wchar_t> >(one); 388 install_from<_VSTD::codecvt<char, char, mbstate_t> >(one); 389 install_from<_VSTD::codecvt<char16_t, char, mbstate_t> >(one); 390 install_from<_VSTD::codecvt<char32_t, char, mbstate_t> >(one); 391 install_from<_VSTD::codecvt<wchar_t, char, mbstate_t> >(one); 392 } 393 if (c & locale::monetary) 394 { 395 install_from<moneypunct<char, false> >(one); 396 install_from<moneypunct<char, true> >(one); 397 install_from<moneypunct<wchar_t, false> >(one); 398 install_from<moneypunct<wchar_t, true> >(one); 399 install_from<money_get<char> >(one); 400 install_from<money_get<wchar_t> >(one); 401 install_from<money_put<char> >(one); 402 install_from<money_put<wchar_t> >(one); 403 } 404 if (c & locale::numeric) 405 { 406 install_from<numpunct<char> >(one); 407 install_from<numpunct<wchar_t> >(one); 408 install_from<num_get<char> >(one); 409 install_from<num_get<wchar_t> >(one); 410 install_from<num_put<char> >(one); 411 install_from<num_put<wchar_t> >(one); 412 } 413 if (c & locale::time) 414 { 415 install_from<time_get<char> >(one); 416 install_from<time_get<wchar_t> >(one); 417 install_from<time_put<char> >(one); 418 install_from<time_put<wchar_t> >(one); 419 } 420 if (c & locale::messages) 421 { 422 install_from<_VSTD::messages<char> >(one); 423 install_from<_VSTD::messages<wchar_t> >(one); 424 } 425 #ifndef _LIBCPP_NO_EXCEPTIONS 426 } 427 catch (...) 428 { 429 for (unsigned i = 0; i < facets_.size(); ++i) 430 if (facets_[i]) 431 facets_[i]->__release_shared(); 432 throw; 433 } 434 #endif // _LIBCPP_NO_EXCEPTIONS 435 } 436 437 locale::__imp::__imp(const __imp& other, facet* f, long id) 438 : facets_(max<size_t>(N, other.facets_.size()+1)), 439 name_("*") 440 { 441 f->__add_shared(); 442 unique_ptr<facet, release> hold(f); 443 facets_ = other.facets_; 444 for (unsigned i = 0; i < other.facets_.size(); ++i) 445 if (facets_[i]) 446 facets_[i]->__add_shared(); 447 install(hold.get(), id); 448 } 449 450 locale::__imp::~__imp() 451 { 452 for (unsigned i = 0; i < facets_.size(); ++i) 453 if (facets_[i]) 454 facets_[i]->__release_shared(); 455 } 456 457 void 458 locale::__imp::install(facet* f, long id) 459 { 460 f->__add_shared(); 461 unique_ptr<facet, release> hold(f); 462 if (static_cast<size_t>(id) >= facets_.size()) 463 facets_.resize(static_cast<size_t>(id+1)); 464 if (facets_[static_cast<size_t>(id)]) 465 facets_[static_cast<size_t>(id)]->__release_shared(); 466 facets_[static_cast<size_t>(id)] = hold.release(); 467 } 468 469 const locale::facet* 470 locale::__imp::use_facet(long id) const 471 { 472 #ifndef _LIBCPP_NO_EXCEPTIONS 473 if (!has_facet(id)) 474 throw bad_cast(); 475 #endif // _LIBCPP_NO_EXCEPTIONS 476 return facets_[static_cast<size_t>(id)]; 477 } 478 479 // locale 480 481 const locale& 482 locale::__imp::make_classic() 483 { 484 // only one thread can get in here and it only gets in once 485 static aligned_storage<sizeof(locale)>::type buf; 486 locale* c = reinterpret_cast<locale*>(&buf); 487 c->__locale_ = &make<__imp>(1u); 488 return *c; 489 } 490 491 const locale& 492 locale::classic() 493 { 494 static const locale& c = __imp::make_classic(); 495 return c; 496 } 497 498 locale& 499 locale::__imp::make_global() 500 { 501 // only one thread can get in here and it only gets in once 502 static aligned_storage<sizeof(locale)>::type buf; 503 auto *obj = ::new (&buf) locale(locale::classic()); 504 return *obj; 505 } 506 507 locale& 508 locale::__global() 509 { 510 static locale& g = __imp::make_global(); 511 return g; 512 } 513 514 locale::locale() _NOEXCEPT 515 : __locale_(__global().__locale_) 516 { 517 __locale_->__add_shared(); 518 } 519 520 locale::locale(const locale& l) _NOEXCEPT 521 : __locale_(l.__locale_) 522 { 523 __locale_->__add_shared(); 524 } 525 526 locale::~locale() 527 { 528 __locale_->__release_shared(); 529 } 530 531 const locale& 532 locale::operator=(const locale& other) _NOEXCEPT 533 { 534 other.__locale_->__add_shared(); 535 __locale_->__release_shared(); 536 __locale_ = other.__locale_; 537 return *this; 538 } 539 540 locale::locale(const char* name) 541 #ifndef _LIBCPP_NO_EXCEPTIONS 542 : __locale_(name ? new __imp(name) 543 : throw runtime_error("locale constructed with null")) 544 #else // _LIBCPP_NO_EXCEPTIONS 545 : __locale_(new __imp(name)) 546 #endif 547 { 548 __locale_->__add_shared(); 549 } 550 551 locale::locale(const string& name) 552 : __locale_(new __imp(name)) 553 { 554 __locale_->__add_shared(); 555 } 556 557 locale::locale(const locale& other, const char* name, category c) 558 #ifndef _LIBCPP_NO_EXCEPTIONS 559 : __locale_(name ? new __imp(*other.__locale_, name, c) 560 : throw runtime_error("locale constructed with null")) 561 #else // _LIBCPP_NO_EXCEPTIONS 562 : __locale_(new __imp(*other.__locale_, name, c)) 563 #endif 564 { 565 __locale_->__add_shared(); 566 } 567 568 locale::locale(const locale& other, const string& name, category c) 569 : __locale_(new __imp(*other.__locale_, name, c)) 570 { 571 __locale_->__add_shared(); 572 } 573 574 locale::locale(const locale& other, const locale& one, category c) 575 : __locale_(new __imp(*other.__locale_, *one.__locale_, c)) 576 { 577 __locale_->__add_shared(); 578 } 579 580 string 581 locale::name() const 582 { 583 return __locale_->name(); 584 } 585 586 void 587 locale::__install_ctor(const locale& other, facet* f, long id) 588 { 589 if (f) 590 __locale_ = new __imp(*other.__locale_, f, id); 591 else 592 __locale_ = other.__locale_; 593 __locale_->__add_shared(); 594 } 595 596 locale 597 locale::global(const locale& loc) 598 { 599 locale& g = __global(); 600 locale r = g; 601 g = loc; 602 if (g.name() != "*") 603 setlocale(LC_ALL, g.name().c_str()); 604 return r; 605 } 606 607 bool 608 locale::has_facet(id& x) const 609 { 610 return __locale_->has_facet(x.__get()); 611 } 612 613 const locale::facet* 614 locale::use_facet(id& x) const 615 { 616 return __locale_->use_facet(x.__get()); 617 } 618 619 bool 620 locale::operator==(const locale& y) const 621 { 622 return (__locale_ == y.__locale_) 623 || (__locale_->name() != "*" && __locale_->name() == y.__locale_->name()); 624 } 625 626 // locale::facet 627 628 locale::facet::~facet() 629 { 630 } 631 632 void 633 locale::facet::__on_zero_shared() _NOEXCEPT 634 { 635 delete this; 636 } 637 638 // locale::id 639 640 int32_t locale::id::__next_id = 0; 641 642 namespace 643 { 644 645 class __fake_bind 646 { 647 locale::id* id_; 648 void (locale::id::* pmf_)(); 649 public: 650 __fake_bind(void (locale::id::* pmf)(), locale::id* id) 651 : id_(id), pmf_(pmf) {} 652 653 void operator()() const 654 { 655 (id_->*pmf_)(); 656 } 657 }; 658 659 } 660 661 long 662 locale::id::__get() 663 { 664 call_once(__flag_, __fake_bind(&locale::id::__init, this)); 665 return __id_ - 1; 666 } 667 668 void 669 locale::id::__init() 670 { 671 __id_ = __libcpp_atomic_add(&__next_id, 1); 672 } 673 674 // template <> class collate_byname<char> 675 676 collate_byname<char>::collate_byname(const char* n, size_t refs) 677 : collate<char>(refs), 678 __l(newlocale(LC_ALL_MASK, n, 0)) 679 { 680 if (__l == 0) 681 __throw_runtime_error("collate_byname<char>::collate_byname" 682 " failed to construct for " + string(n)); 683 } 684 685 collate_byname<char>::collate_byname(const string& name, size_t refs) 686 : collate<char>(refs), 687 __l(newlocale(LC_ALL_MASK, name.c_str(), 0)) 688 { 689 if (__l == 0) 690 __throw_runtime_error("collate_byname<char>::collate_byname" 691 " failed to construct for " + name); 692 } 693 694 collate_byname<char>::~collate_byname() 695 { 696 freelocale(__l); 697 } 698 699 int 700 collate_byname<char>::do_compare(const char_type* __lo1, const char_type* __hi1, 701 const char_type* __lo2, const char_type* __hi2) const 702 { 703 string_type lhs(__lo1, __hi1); 704 string_type rhs(__lo2, __hi2); 705 int r = strcoll_l(lhs.c_str(), rhs.c_str(), __l); 706 if (r < 0) 707 return -1; 708 if (r > 0) 709 return 1; 710 return r; 711 } 712 713 collate_byname<char>::string_type 714 collate_byname<char>::do_transform(const char_type* lo, const char_type* hi) const 715 { 716 const string_type in(lo, hi); 717 string_type out(strxfrm_l(0, in.c_str(), 0, __l), char()); 718 strxfrm_l(const_cast<char*>(out.c_str()), in.c_str(), out.size()+1, __l); 719 return out; 720 } 721 722 // template <> class collate_byname<wchar_t> 723 724 collate_byname<wchar_t>::collate_byname(const char* n, size_t refs) 725 : collate<wchar_t>(refs), 726 __l(newlocale(LC_ALL_MASK, n, 0)) 727 { 728 if (__l == 0) 729 __throw_runtime_error("collate_byname<wchar_t>::collate_byname(size_t refs)" 730 " failed to construct for " + string(n)); 731 } 732 733 collate_byname<wchar_t>::collate_byname(const string& name, size_t refs) 734 : collate<wchar_t>(refs), 735 __l(newlocale(LC_ALL_MASK, name.c_str(), 0)) 736 { 737 if (__l == 0) 738 __throw_runtime_error("collate_byname<wchar_t>::collate_byname(size_t refs)" 739 " failed to construct for " + name); 740 } 741 742 collate_byname<wchar_t>::~collate_byname() 743 { 744 freelocale(__l); 745 } 746 747 int 748 collate_byname<wchar_t>::do_compare(const char_type* __lo1, const char_type* __hi1, 749 const char_type* __lo2, const char_type* __hi2) const 750 { 751 string_type lhs(__lo1, __hi1); 752 string_type rhs(__lo2, __hi2); 753 int r = wcscoll_l(lhs.c_str(), rhs.c_str(), __l); 754 if (r < 0) 755 return -1; 756 if (r > 0) 757 return 1; 758 return r; 759 } 760 761 collate_byname<wchar_t>::string_type 762 collate_byname<wchar_t>::do_transform(const char_type* lo, const char_type* hi) const 763 { 764 const string_type in(lo, hi); 765 string_type out(wcsxfrm_l(0, in.c_str(), 0, __l), wchar_t()); 766 wcsxfrm_l(const_cast<wchar_t*>(out.c_str()), in.c_str(), out.size()+1, __l); 767 return out; 768 } 769 770 // template <> class ctype<wchar_t>; 771 772 const ctype_base::mask ctype_base::space; 773 const ctype_base::mask ctype_base::print; 774 const ctype_base::mask ctype_base::cntrl; 775 const ctype_base::mask ctype_base::upper; 776 const ctype_base::mask ctype_base::lower; 777 const ctype_base::mask ctype_base::alpha; 778 const ctype_base::mask ctype_base::digit; 779 const ctype_base::mask ctype_base::punct; 780 const ctype_base::mask ctype_base::xdigit; 781 const ctype_base::mask ctype_base::blank; 782 const ctype_base::mask ctype_base::alnum; 783 const ctype_base::mask ctype_base::graph; 784 785 locale::id ctype<wchar_t>::id; 786 787 ctype<wchar_t>::~ctype() 788 { 789 } 790 791 bool 792 ctype<wchar_t>::do_is(mask m, char_type c) const 793 { 794 return isascii(c) ? (ctype<char>::classic_table()[c] & m) != 0 : false; 795 } 796 797 const wchar_t* 798 ctype<wchar_t>::do_is(const char_type* low, const char_type* high, mask* vec) const 799 { 800 for (; low != high; ++low, ++vec) 801 *vec = static_cast<mask>(isascii(*low) ? 802 ctype<char>::classic_table()[*low] : 0); 803 return low; 804 } 805 806 const wchar_t* 807 ctype<wchar_t>::do_scan_is(mask m, const char_type* low, const char_type* high) const 808 { 809 for (; low != high; ++low) 810 if (isascii(*low) && (ctype<char>::classic_table()[*low] & m)) 811 break; 812 return low; 813 } 814 815 const wchar_t* 816 ctype<wchar_t>::do_scan_not(mask m, const char_type* low, const char_type* high) const 817 { 818 for (; low != high; ++low) 819 if (!(isascii(*low) && (ctype<char>::classic_table()[*low] & m))) 820 break; 821 return low; 822 } 823 824 wchar_t 825 ctype<wchar_t>::do_toupper(char_type c) const 826 { 827 #ifdef _LIBCPP_HAS_DEFAULTRUNELOCALE 828 return isascii(c) ? _DefaultRuneLocale.__mapupper[c] : c; 829 #elif defined(__GLIBC__) || defined(__EMSCRIPTEN__) || \ 830 defined(__NetBSD__) 831 return isascii(c) ? ctype<char>::__classic_upper_table()[c] : c; 832 #else 833 return (isascii(c) && iswlower_l(c, _LIBCPP_GET_C_LOCALE)) ? c-L'a'+L'A' : c; 834 #endif 835 } 836 837 const wchar_t* 838 ctype<wchar_t>::do_toupper(char_type* low, const char_type* high) const 839 { 840 for (; low != high; ++low) 841 #ifdef _LIBCPP_HAS_DEFAULTRUNELOCALE 842 *low = isascii(*low) ? _DefaultRuneLocale.__mapupper[*low] : *low; 843 #elif defined(__GLIBC__) || defined(__EMSCRIPTEN__) || \ 844 defined(__NetBSD__) 845 *low = isascii(*low) ? ctype<char>::__classic_upper_table()[*low] 846 : *low; 847 #else 848 *low = (isascii(*low) && islower_l(*low, _LIBCPP_GET_C_LOCALE)) ? (*low-L'a'+L'A') : *low; 849 #endif 850 return low; 851 } 852 853 wchar_t 854 ctype<wchar_t>::do_tolower(char_type c) const 855 { 856 #ifdef _LIBCPP_HAS_DEFAULTRUNELOCALE 857 return isascii(c) ? _DefaultRuneLocale.__maplower[c] : c; 858 #elif defined(__GLIBC__) || defined(__EMSCRIPTEN__) || \ 859 defined(__NetBSD__) 860 return isascii(c) ? ctype<char>::__classic_lower_table()[c] : c; 861 #else 862 return (isascii(c) && isupper_l(c, _LIBCPP_GET_C_LOCALE)) ? c-L'A'+'a' : c; 863 #endif 864 } 865 866 const wchar_t* 867 ctype<wchar_t>::do_tolower(char_type* low, const char_type* high) const 868 { 869 for (; low != high; ++low) 870 #ifdef _LIBCPP_HAS_DEFAULTRUNELOCALE 871 *low = isascii(*low) ? _DefaultRuneLocale.__maplower[*low] : *low; 872 #elif defined(__GLIBC__) || defined(__EMSCRIPTEN__) || \ 873 defined(__NetBSD__) 874 *low = isascii(*low) ? ctype<char>::__classic_lower_table()[*low] 875 : *low; 876 #else 877 *low = (isascii(*low) && isupper_l(*low, _LIBCPP_GET_C_LOCALE)) ? *low-L'A'+L'a' : *low; 878 #endif 879 return low; 880 } 881 882 wchar_t 883 ctype<wchar_t>::do_widen(char c) const 884 { 885 return c; 886 } 887 888 const char* 889 ctype<wchar_t>::do_widen(const char* low, const char* high, char_type* dest) const 890 { 891 for (; low != high; ++low, ++dest) 892 *dest = *low; 893 return low; 894 } 895 896 char 897 ctype<wchar_t>::do_narrow(char_type c, char dfault) const 898 { 899 if (isascii(c)) 900 return static_cast<char>(c); 901 return dfault; 902 } 903 904 const wchar_t* 905 ctype<wchar_t>::do_narrow(const char_type* low, const char_type* high, char dfault, char* dest) const 906 { 907 for (; low != high; ++low, ++dest) 908 if (isascii(*low)) 909 *dest = static_cast<char>(*low); 910 else 911 *dest = dfault; 912 return low; 913 } 914 915 // template <> class ctype<char>; 916 917 locale::id ctype<char>::id; 918 919 ctype<char>::ctype(const mask* tab, bool del, size_t refs) 920 : locale::facet(refs), 921 __tab_(tab), 922 __del_(del) 923 { 924 if (__tab_ == 0) 925 __tab_ = classic_table(); 926 } 927 928 ctype<char>::~ctype() 929 { 930 if (__tab_ && __del_) 931 delete [] __tab_; 932 } 933 934 char 935 ctype<char>::do_toupper(char_type c) const 936 { 937 #ifdef _LIBCPP_HAS_DEFAULTRUNELOCALE 938 return isascii(c) ? 939 static_cast<char>(_DefaultRuneLocale.__mapupper[static_cast<ptrdiff_t>(c)]) : c; 940 #elif defined(__NetBSD__) 941 return static_cast<char>(__classic_upper_table()[static_cast<unsigned char>(c)]); 942 #elif defined(__GLIBC__) || defined(__EMSCRIPTEN__) 943 return isascii(c) ? 944 static_cast<char>(__classic_upper_table()[static_cast<unsigned char>(c)]) : c; 945 #else 946 return (isascii(c) && islower_l(c, _LIBCPP_GET_C_LOCALE)) ? c-'a'+'A' : c; 947 #endif 948 } 949 950 const char* 951 ctype<char>::do_toupper(char_type* low, const char_type* high) const 952 { 953 for (; low != high; ++low) 954 #ifdef _LIBCPP_HAS_DEFAULTRUNELOCALE 955 *low = isascii(*low) ? 956 static_cast<char>(_DefaultRuneLocale.__mapupper[static_cast<ptrdiff_t>(*low)]) : *low; 957 #elif defined(__NetBSD__) 958 *low = static_cast<char>(__classic_upper_table()[static_cast<unsigned char>(*low)]); 959 #elif defined(__GLIBC__) || defined(__EMSCRIPTEN__) 960 *low = isascii(*low) ? 961 static_cast<char>(__classic_upper_table()[static_cast<size_t>(*low)]) : *low; 962 #else 963 *low = (isascii(*low) && islower_l(*low, _LIBCPP_GET_C_LOCALE)) ? *low-'a'+'A' : *low; 964 #endif 965 return low; 966 } 967 968 char 969 ctype<char>::do_tolower(char_type c) const 970 { 971 #ifdef _LIBCPP_HAS_DEFAULTRUNELOCALE 972 return isascii(c) ? 973 static_cast<char>(_DefaultRuneLocale.__maplower[static_cast<ptrdiff_t>(c)]) : c; 974 #elif defined(__NetBSD__) 975 return static_cast<char>(__classic_lower_table()[static_cast<unsigned char>(c)]); 976 #elif defined(__GLIBC__) || defined(__EMSCRIPTEN__) 977 return isascii(c) ? 978 static_cast<char>(__classic_lower_table()[static_cast<size_t>(c)]) : c; 979 #else 980 return (isascii(c) && isupper_l(c, _LIBCPP_GET_C_LOCALE)) ? c-'A'+'a' : c; 981 #endif 982 } 983 984 const char* 985 ctype<char>::do_tolower(char_type* low, const char_type* high) const 986 { 987 for (; low != high; ++low) 988 #ifdef _LIBCPP_HAS_DEFAULTRUNELOCALE 989 *low = isascii(*low) ? static_cast<char>(_DefaultRuneLocale.__maplower[static_cast<ptrdiff_t>(*low)]) : *low; 990 #elif defined(__NetBSD__) 991 *low = static_cast<char>(__classic_lower_table()[static_cast<unsigned char>(*low)]); 992 #elif defined(__GLIBC__) || defined(__EMSCRIPTEN__) 993 *low = isascii(*low) ? static_cast<char>(__classic_lower_table()[static_cast<size_t>(*low)]) : *low; 994 #else 995 *low = (isascii(*low) && isupper_l(*low, _LIBCPP_GET_C_LOCALE)) ? *low-'A'+'a' : *low; 996 #endif 997 return low; 998 } 999 1000 char 1001 ctype<char>::do_widen(char c) const 1002 { 1003 return c; 1004 } 1005 1006 const char* 1007 ctype<char>::do_widen(const char* low, const char* high, char_type* dest) const 1008 { 1009 for (; low != high; ++low, ++dest) 1010 *dest = *low; 1011 return low; 1012 } 1013 1014 char 1015 ctype<char>::do_narrow(char_type c, char dfault) const 1016 { 1017 if (isascii(c)) 1018 return static_cast<char>(c); 1019 return dfault; 1020 } 1021 1022 const char* 1023 ctype<char>::do_narrow(const char_type* low, const char_type* high, char dfault, char* dest) const 1024 { 1025 for (; low != high; ++low, ++dest) 1026 if (isascii(*low)) 1027 *dest = *low; 1028 else 1029 *dest = dfault; 1030 return low; 1031 } 1032 1033 #if defined(__EMSCRIPTEN__) 1034 extern "C" const unsigned short ** __ctype_b_loc(); 1035 extern "C" const int ** __ctype_tolower_loc(); 1036 extern "C" const int ** __ctype_toupper_loc(); 1037 #endif 1038 1039 #ifdef _LIBCPP_PROVIDES_DEFAULT_RUNE_TABLE 1040 const ctype<char>::mask* 1041 ctype<char>::classic_table() _NOEXCEPT 1042 { 1043 static _LIBCPP_CONSTEXPR const ctype<char>::mask builtin_table[table_size] = { 1044 cntrl, cntrl, 1045 cntrl, cntrl, 1046 cntrl, cntrl, 1047 cntrl, cntrl, 1048 cntrl, cntrl | space | blank, 1049 cntrl | space, cntrl | space, 1050 cntrl | space, cntrl | space, 1051 cntrl, cntrl, 1052 cntrl, cntrl, 1053 cntrl, cntrl, 1054 cntrl, cntrl, 1055 cntrl, cntrl, 1056 cntrl, cntrl, 1057 cntrl, cntrl, 1058 cntrl, cntrl, 1059 cntrl, cntrl, 1060 space | blank | print, punct | print, 1061 punct | print, punct | print, 1062 punct | print, punct | print, 1063 punct | print, punct | print, 1064 punct | print, punct | print, 1065 punct | print, punct | print, 1066 punct | print, punct | print, 1067 punct | print, punct | print, 1068 digit | print | xdigit, digit | print | xdigit, 1069 digit | print | xdigit, digit | print | xdigit, 1070 digit | print | xdigit, digit | print | xdigit, 1071 digit | print | xdigit, digit | print | xdigit, 1072 digit | print | xdigit, digit | print | xdigit, 1073 punct | print, punct | print, 1074 punct | print, punct | print, 1075 punct | print, punct | print, 1076 punct | print, upper | xdigit | print | alpha, 1077 upper | xdigit | print | alpha, upper | xdigit | print | alpha, 1078 upper | xdigit | print | alpha, upper | xdigit | print | alpha, 1079 upper | xdigit | print | alpha, upper | print | alpha, 1080 upper | print | alpha, upper | print | alpha, 1081 upper | print | alpha, upper | print | alpha, 1082 upper | print | alpha, upper | print | alpha, 1083 upper | print | alpha, upper | print | alpha, 1084 upper | print | alpha, upper | print | alpha, 1085 upper | print | alpha, upper | print | alpha, 1086 upper | print | alpha, upper | print | alpha, 1087 upper | print | alpha, upper | print | alpha, 1088 upper | print | alpha, upper | print | alpha, 1089 upper | print | alpha, punct | print, 1090 punct | print, punct | print, 1091 punct | print, punct | print, 1092 punct | print, lower | xdigit | print | alpha, 1093 lower | xdigit | print | alpha, lower | xdigit | print | alpha, 1094 lower | xdigit | print | alpha, lower | xdigit | print | alpha, 1095 lower | xdigit | print | alpha, lower | print | alpha, 1096 lower | print | alpha, lower | print | alpha, 1097 lower | print | alpha, lower | print | alpha, 1098 lower | print | alpha, lower | print | alpha, 1099 lower | print | alpha, lower | print | alpha, 1100 lower | print | alpha, lower | print | alpha, 1101 lower | print | alpha, lower | print | alpha, 1102 lower | print | alpha, lower | print | alpha, 1103 lower | print | alpha, lower | print | alpha, 1104 lower | print | alpha, lower | print | alpha, 1105 lower | print | alpha, punct | print, 1106 punct | print, punct | print, 1107 punct | print, cntrl, 1108 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1109 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1110 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1111 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1112 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1113 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1114 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1115 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 1116 }; 1117 return builtin_table; 1118 } 1119 #else 1120 const ctype<char>::mask* 1121 ctype<char>::classic_table() _NOEXCEPT 1122 { 1123 #if defined(__APPLE__) || defined(__FreeBSD__) 1124 return _DefaultRuneLocale.__runetype; 1125 #elif defined(__NetBSD__) 1126 return _C_ctype_tab_ + 1; 1127 #elif defined(__GLIBC__) 1128 return _LIBCPP_GET_C_LOCALE->__ctype_b; 1129 #elif __sun__ 1130 return __ctype_mask; 1131 #elif defined(_LIBCPP_MSVCRT) || defined(__MINGW32__) 1132 return __pctype_func(); 1133 #elif defined(__EMSCRIPTEN__) 1134 return *__ctype_b_loc(); 1135 #elif defined(_NEWLIB_VERSION) 1136 // Newlib has a 257-entry table in ctype_.c, where (char)0 starts at [1]. 1137 return _ctype_ + 1; 1138 #elif defined(_AIX) 1139 return (const unsigned int *)__lc_ctype_ptr->obj->mask; 1140 #else 1141 // Platform not supported: abort so the person doing the port knows what to 1142 // fix 1143 # warning ctype<char>::classic_table() is not implemented 1144 printf("ctype<char>::classic_table() is not implemented\n"); 1145 abort(); 1146 return NULL; 1147 #endif 1148 } 1149 #endif 1150 1151 #if defined(__GLIBC__) 1152 const int* 1153 ctype<char>::__classic_lower_table() _NOEXCEPT 1154 { 1155 return _LIBCPP_GET_C_LOCALE->__ctype_tolower; 1156 } 1157 1158 const int* 1159 ctype<char>::__classic_upper_table() _NOEXCEPT 1160 { 1161 return _LIBCPP_GET_C_LOCALE->__ctype_toupper; 1162 } 1163 #elif __NetBSD__ 1164 const short* 1165 ctype<char>::__classic_lower_table() _NOEXCEPT 1166 { 1167 return _C_tolower_tab_ + 1; 1168 } 1169 1170 const short* 1171 ctype<char>::__classic_upper_table() _NOEXCEPT 1172 { 1173 return _C_toupper_tab_ + 1; 1174 } 1175 1176 #elif defined(__EMSCRIPTEN__) 1177 const int* 1178 ctype<char>::__classic_lower_table() _NOEXCEPT 1179 { 1180 return *__ctype_tolower_loc(); 1181 } 1182 1183 const int* 1184 ctype<char>::__classic_upper_table() _NOEXCEPT 1185 { 1186 return *__ctype_toupper_loc(); 1187 } 1188 #endif // __GLIBC__ || __NETBSD__ || __EMSCRIPTEN__ 1189 1190 // template <> class ctype_byname<char> 1191 1192 ctype_byname<char>::ctype_byname(const char* name, size_t refs) 1193 : ctype<char>(0, false, refs), 1194 __l(newlocale(LC_ALL_MASK, name, 0)) 1195 { 1196 if (__l == 0) 1197 __throw_runtime_error("ctype_byname<char>::ctype_byname" 1198 " failed to construct for " + string(name)); 1199 } 1200 1201 ctype_byname<char>::ctype_byname(const string& name, size_t refs) 1202 : ctype<char>(0, false, refs), 1203 __l(newlocale(LC_ALL_MASK, name.c_str(), 0)) 1204 { 1205 if (__l == 0) 1206 __throw_runtime_error("ctype_byname<char>::ctype_byname" 1207 " failed to construct for " + name); 1208 } 1209 1210 ctype_byname<char>::~ctype_byname() 1211 { 1212 freelocale(__l); 1213 } 1214 1215 char 1216 ctype_byname<char>::do_toupper(char_type c) const 1217 { 1218 return static_cast<char>(toupper_l(static_cast<unsigned char>(c), __l)); 1219 } 1220 1221 const char* 1222 ctype_byname<char>::do_toupper(char_type* low, const char_type* high) const 1223 { 1224 for (; low != high; ++low) 1225 *low = static_cast<char>(toupper_l(static_cast<unsigned char>(*low), __l)); 1226 return low; 1227 } 1228 1229 char 1230 ctype_byname<char>::do_tolower(char_type c) const 1231 { 1232 return static_cast<char>(tolower_l(static_cast<unsigned char>(c), __l)); 1233 } 1234 1235 const char* 1236 ctype_byname<char>::do_tolower(char_type* low, const char_type* high) const 1237 { 1238 for (; low != high; ++low) 1239 *low = static_cast<char>(tolower_l(static_cast<unsigned char>(*low), __l)); 1240 return low; 1241 } 1242 1243 // template <> class ctype_byname<wchar_t> 1244 1245 ctype_byname<wchar_t>::ctype_byname(const char* name, size_t refs) 1246 : ctype<wchar_t>(refs), 1247 __l(newlocale(LC_ALL_MASK, name, 0)) 1248 { 1249 if (__l == 0) 1250 __throw_runtime_error("ctype_byname<wchar_t>::ctype_byname" 1251 " failed to construct for " + string(name)); 1252 } 1253 1254 ctype_byname<wchar_t>::ctype_byname(const string& name, size_t refs) 1255 : ctype<wchar_t>(refs), 1256 __l(newlocale(LC_ALL_MASK, name.c_str(), 0)) 1257 { 1258 if (__l == 0) 1259 __throw_runtime_error("ctype_byname<wchar_t>::ctype_byname" 1260 " failed to construct for " + name); 1261 } 1262 1263 ctype_byname<wchar_t>::~ctype_byname() 1264 { 1265 freelocale(__l); 1266 } 1267 1268 bool 1269 ctype_byname<wchar_t>::do_is(mask m, char_type c) const 1270 { 1271 #ifdef _LIBCPP_WCTYPE_IS_MASK 1272 return static_cast<bool>(iswctype_l(c, m, __l)); 1273 #else 1274 bool result = false; 1275 wint_t ch = static_cast<wint_t>(c); 1276 if ((m & space) == space) result |= (iswspace_l(ch, __l) != 0); 1277 if ((m & print) == print) result |= (iswprint_l(ch, __l) != 0); 1278 if ((m & cntrl) == cntrl) result |= (iswcntrl_l(ch, __l) != 0); 1279 if ((m & upper) == upper) result |= (iswupper_l(ch, __l) != 0); 1280 if ((m & lower) == lower) result |= (iswlower_l(ch, __l) != 0); 1281 if ((m & alpha) == alpha) result |= (iswalpha_l(ch, __l) != 0); 1282 if ((m & digit) == digit) result |= (iswdigit_l(ch, __l) != 0); 1283 if ((m & punct) == punct) result |= (iswpunct_l(ch, __l) != 0); 1284 if ((m & xdigit) == xdigit) result |= (iswxdigit_l(ch, __l) != 0); 1285 if ((m & blank) == blank) result |= (iswblank_l(ch, __l) != 0); 1286 return result; 1287 #endif 1288 } 1289 1290 const wchar_t* 1291 ctype_byname<wchar_t>::do_is(const char_type* low, const char_type* high, mask* vec) const 1292 { 1293 for (; low != high; ++low, ++vec) 1294 { 1295 if (isascii(*low)) 1296 *vec = static_cast<mask>(ctype<char>::classic_table()[*low]); 1297 else 1298 { 1299 *vec = 0; 1300 wint_t ch = static_cast<wint_t>(*low); 1301 if (iswspace_l(ch, __l)) 1302 *vec |= space; 1303 #ifndef _LIBCPP_CTYPE_MASK_IS_COMPOSITE_PRINT 1304 if (iswprint_l(ch, __l)) 1305 *vec |= print; 1306 #endif 1307 if (iswcntrl_l(ch, __l)) 1308 *vec |= cntrl; 1309 if (iswupper_l(ch, __l)) 1310 *vec |= upper; 1311 if (iswlower_l(ch, __l)) 1312 *vec |= lower; 1313 #ifndef _LIBCPP_CTYPE_MASK_IS_COMPOSITE_ALPHA 1314 if (iswalpha_l(ch, __l)) 1315 *vec |= alpha; 1316 #endif 1317 if (iswdigit_l(ch, __l)) 1318 *vec |= digit; 1319 if (iswpunct_l(ch, __l)) 1320 *vec |= punct; 1321 #ifndef _LIBCPP_CTYPE_MASK_IS_COMPOSITE_XDIGIT 1322 if (iswxdigit_l(ch, __l)) 1323 *vec |= xdigit; 1324 #endif 1325 #if !defined(__sun__) 1326 if (iswblank_l(ch, __l)) 1327 *vec |= blank; 1328 #endif 1329 } 1330 } 1331 return low; 1332 } 1333 1334 const wchar_t* 1335 ctype_byname<wchar_t>::do_scan_is(mask m, const char_type* low, const char_type* high) const 1336 { 1337 for (; low != high; ++low) 1338 { 1339 #ifdef _LIBCPP_WCTYPE_IS_MASK 1340 if (iswctype_l(*low, m, __l)) 1341 break; 1342 #else 1343 wint_t ch = static_cast<wint_t>(*low); 1344 if ((m & space) == space && iswspace_l(ch, __l)) break; 1345 if ((m & print) == print && iswprint_l(ch, __l)) break; 1346 if ((m & cntrl) == cntrl && iswcntrl_l(ch, __l)) break; 1347 if ((m & upper) == upper && iswupper_l(ch, __l)) break; 1348 if ((m & lower) == lower && iswlower_l(ch, __l)) break; 1349 if ((m & alpha) == alpha && iswalpha_l(ch, __l)) break; 1350 if ((m & digit) == digit && iswdigit_l(ch, __l)) break; 1351 if ((m & punct) == punct && iswpunct_l(ch, __l)) break; 1352 if ((m & xdigit) == xdigit && iswxdigit_l(ch, __l)) break; 1353 if ((m & blank) == blank && iswblank_l(ch, __l)) break; 1354 #endif 1355 } 1356 return low; 1357 } 1358 1359 const wchar_t* 1360 ctype_byname<wchar_t>::do_scan_not(mask m, const char_type* low, const char_type* high) const 1361 { 1362 for (; low != high; ++low) 1363 { 1364 #ifdef _LIBCPP_WCTYPE_IS_MASK 1365 if (!iswctype_l(*low, m, __l)) 1366 break; 1367 #else 1368 wint_t ch = static_cast<wint_t>(*low); 1369 if ((m & space) == space && iswspace_l(ch, __l)) continue; 1370 if ((m & print) == print && iswprint_l(ch, __l)) continue; 1371 if ((m & cntrl) == cntrl && iswcntrl_l(ch, __l)) continue; 1372 if ((m & upper) == upper && iswupper_l(ch, __l)) continue; 1373 if ((m & lower) == lower && iswlower_l(ch, __l)) continue; 1374 if ((m & alpha) == alpha && iswalpha_l(ch, __l)) continue; 1375 if ((m & digit) == digit && iswdigit_l(ch, __l)) continue; 1376 if ((m & punct) == punct && iswpunct_l(ch, __l)) continue; 1377 if ((m & xdigit) == xdigit && iswxdigit_l(ch, __l)) continue; 1378 if ((m & blank) == blank && iswblank_l(ch, __l)) continue; 1379 break; 1380 #endif 1381 } 1382 return low; 1383 } 1384 1385 wchar_t 1386 ctype_byname<wchar_t>::do_toupper(char_type c) const 1387 { 1388 return towupper_l(c, __l); 1389 } 1390 1391 const wchar_t* 1392 ctype_byname<wchar_t>::do_toupper(char_type* low, const char_type* high) const 1393 { 1394 for (; low != high; ++low) 1395 *low = towupper_l(*low, __l); 1396 return low; 1397 } 1398 1399 wchar_t 1400 ctype_byname<wchar_t>::do_tolower(char_type c) const 1401 { 1402 return towlower_l(c, __l); 1403 } 1404 1405 const wchar_t* 1406 ctype_byname<wchar_t>::do_tolower(char_type* low, const char_type* high) const 1407 { 1408 for (; low != high; ++low) 1409 *low = towlower_l(*low, __l); 1410 return low; 1411 } 1412 1413 wchar_t 1414 ctype_byname<wchar_t>::do_widen(char c) const 1415 { 1416 return __libcpp_btowc_l(c, __l); 1417 } 1418 1419 const char* 1420 ctype_byname<wchar_t>::do_widen(const char* low, const char* high, char_type* dest) const 1421 { 1422 for (; low != high; ++low, ++dest) 1423 *dest = __libcpp_btowc_l(*low, __l); 1424 return low; 1425 } 1426 1427 char 1428 ctype_byname<wchar_t>::do_narrow(char_type c, char dfault) const 1429 { 1430 int r = __libcpp_wctob_l(c, __l); 1431 return r != static_cast<int>(WEOF) ? static_cast<char>(r) : dfault; 1432 } 1433 1434 const wchar_t* 1435 ctype_byname<wchar_t>::do_narrow(const char_type* low, const char_type* high, char dfault, char* dest) const 1436 { 1437 for (; low != high; ++low, ++dest) 1438 { 1439 int r = __libcpp_wctob_l(*low, __l); 1440 *dest = r != static_cast<int>(WEOF) ? static_cast<char>(r) : dfault; 1441 } 1442 return low; 1443 } 1444 1445 // template <> class codecvt<char, char, mbstate_t> 1446 1447 locale::id codecvt<char, char, mbstate_t>::id; 1448 1449 codecvt<char, char, mbstate_t>::~codecvt() 1450 { 1451 } 1452 1453 codecvt<char, char, mbstate_t>::result 1454 codecvt<char, char, mbstate_t>::do_out(state_type&, 1455 const intern_type* frm, const intern_type*, const intern_type*& frm_nxt, 1456 extern_type* to, extern_type*, extern_type*& to_nxt) const 1457 { 1458 frm_nxt = frm; 1459 to_nxt = to; 1460 return noconv; 1461 } 1462 1463 codecvt<char, char, mbstate_t>::result 1464 codecvt<char, char, mbstate_t>::do_in(state_type&, 1465 const extern_type* frm, const extern_type*, const extern_type*& frm_nxt, 1466 intern_type* to, intern_type*, intern_type*& to_nxt) const 1467 { 1468 frm_nxt = frm; 1469 to_nxt = to; 1470 return noconv; 1471 } 1472 1473 codecvt<char, char, mbstate_t>::result 1474 codecvt<char, char, mbstate_t>::do_unshift(state_type&, 1475 extern_type* to, extern_type*, extern_type*& to_nxt) const 1476 { 1477 to_nxt = to; 1478 return noconv; 1479 } 1480 1481 int 1482 codecvt<char, char, mbstate_t>::do_encoding() const _NOEXCEPT 1483 { 1484 return 1; 1485 } 1486 1487 bool 1488 codecvt<char, char, mbstate_t>::do_always_noconv() const _NOEXCEPT 1489 { 1490 return true; 1491 } 1492 1493 int 1494 codecvt<char, char, mbstate_t>::do_length(state_type&, 1495 const extern_type* frm, const extern_type* end, size_t mx) const 1496 { 1497 return static_cast<int>(min<size_t>(mx, static_cast<size_t>(end-frm))); 1498 } 1499 1500 int 1501 codecvt<char, char, mbstate_t>::do_max_length() const _NOEXCEPT 1502 { 1503 return 1; 1504 } 1505 1506 // template <> class codecvt<wchar_t, char, mbstate_t> 1507 1508 locale::id codecvt<wchar_t, char, mbstate_t>::id; 1509 1510 codecvt<wchar_t, char, mbstate_t>::codecvt(size_t refs) 1511 : locale::facet(refs), 1512 __l(_LIBCPP_GET_C_LOCALE) 1513 { 1514 } 1515 1516 codecvt<wchar_t, char, mbstate_t>::codecvt(const char* nm, size_t refs) 1517 : locale::facet(refs), 1518 __l(newlocale(LC_ALL_MASK, nm, 0)) 1519 { 1520 if (__l == 0) 1521 __throw_runtime_error("codecvt_byname<wchar_t, char, mbstate_t>::codecvt_byname" 1522 " failed to construct for " + string(nm)); 1523 } 1524 1525 codecvt<wchar_t, char, mbstate_t>::~codecvt() 1526 { 1527 if (__l != _LIBCPP_GET_C_LOCALE) 1528 freelocale(__l); 1529 } 1530 1531 codecvt<wchar_t, char, mbstate_t>::result 1532 codecvt<wchar_t, char, mbstate_t>::do_out(state_type& st, 1533 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 1534 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 1535 { 1536 // look for first internal null in frm 1537 const intern_type* fend = frm; 1538 for (; fend != frm_end; ++fend) 1539 if (*fend == 0) 1540 break; 1541 // loop over all null-terminated sequences in frm 1542 to_nxt = to; 1543 for (frm_nxt = frm; frm != frm_end && to != to_end; frm = frm_nxt, to = to_nxt) 1544 { 1545 // save state in case it is needed to recover to_nxt on error 1546 mbstate_t save_state = st; 1547 size_t n = __libcpp_wcsnrtombs_l(to, &frm_nxt, static_cast<size_t>(fend-frm), 1548 static_cast<size_t>(to_end-to), &st, __l); 1549 if (n == size_t(-1)) 1550 { 1551 // need to recover to_nxt 1552 for (to_nxt = to; frm != frm_nxt; ++frm) 1553 { 1554 n = __libcpp_wcrtomb_l(to_nxt, *frm, &save_state, __l); 1555 if (n == size_t(-1)) 1556 break; 1557 to_nxt += n; 1558 } 1559 frm_nxt = frm; 1560 return error; 1561 } 1562 if (n == 0) 1563 return partial; 1564 to_nxt += n; 1565 if (to_nxt == to_end) 1566 break; 1567 if (fend != frm_end) // set up next null terminated sequence 1568 { 1569 // Try to write the terminating null 1570 extern_type tmp[MB_LEN_MAX]; 1571 n = __libcpp_wcrtomb_l(tmp, intern_type(), &st, __l); 1572 if (n == size_t(-1)) // on error 1573 return error; 1574 if (n > static_cast<size_t>(to_end-to_nxt)) // is there room? 1575 return partial; 1576 for (extern_type* p = tmp; n; --n) // write it 1577 *to_nxt++ = *p++; 1578 ++frm_nxt; 1579 // look for next null in frm 1580 for (fend = frm_nxt; fend != frm_end; ++fend) 1581 if (*fend == 0) 1582 break; 1583 } 1584 } 1585 return frm_nxt == frm_end ? ok : partial; 1586 } 1587 1588 codecvt<wchar_t, char, mbstate_t>::result 1589 codecvt<wchar_t, char, mbstate_t>::do_in(state_type& st, 1590 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 1591 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 1592 { 1593 // look for first internal null in frm 1594 const extern_type* fend = frm; 1595 for (; fend != frm_end; ++fend) 1596 if (*fend == 0) 1597 break; 1598 // loop over all null-terminated sequences in frm 1599 to_nxt = to; 1600 for (frm_nxt = frm; frm != frm_end && to != to_end; frm = frm_nxt, to = to_nxt) 1601 { 1602 // save state in case it is needed to recover to_nxt on error 1603 mbstate_t save_state = st; 1604 size_t n = __libcpp_mbsnrtowcs_l(to, &frm_nxt, static_cast<size_t>(fend-frm), 1605 static_cast<size_t>(to_end-to), &st, __l); 1606 if (n == size_t(-1)) 1607 { 1608 // need to recover to_nxt 1609 for (to_nxt = to; frm != frm_nxt; ++to_nxt) 1610 { 1611 n = __libcpp_mbrtowc_l(to_nxt, frm, static_cast<size_t>(fend-frm), 1612 &save_state, __l); 1613 switch (n) 1614 { 1615 case 0: 1616 ++frm; 1617 break; 1618 case size_t(-1): 1619 frm_nxt = frm; 1620 return error; 1621 case size_t(-2): 1622 frm_nxt = frm; 1623 return partial; 1624 default: 1625 frm += n; 1626 break; 1627 } 1628 } 1629 frm_nxt = frm; 1630 return frm_nxt == frm_end ? ok : partial; 1631 } 1632 if (n == size_t(-1)) 1633 return error; 1634 to_nxt += n; 1635 if (to_nxt == to_end) 1636 break; 1637 if (fend != frm_end) // set up next null terminated sequence 1638 { 1639 // Try to write the terminating null 1640 n = __libcpp_mbrtowc_l(to_nxt, frm_nxt, 1, &st, __l); 1641 if (n != 0) // on error 1642 return error; 1643 ++to_nxt; 1644 ++frm_nxt; 1645 // look for next null in frm 1646 for (fend = frm_nxt; fend != frm_end; ++fend) 1647 if (*fend == 0) 1648 break; 1649 } 1650 } 1651 return frm_nxt == frm_end ? ok : partial; 1652 } 1653 1654 codecvt<wchar_t, char, mbstate_t>::result 1655 codecvt<wchar_t, char, mbstate_t>::do_unshift(state_type& st, 1656 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 1657 { 1658 to_nxt = to; 1659 extern_type tmp[MB_LEN_MAX]; 1660 size_t n = __libcpp_wcrtomb_l(tmp, intern_type(), &st, __l); 1661 if (n == size_t(-1) || n == 0) // on error 1662 return error; 1663 --n; 1664 if (n > static_cast<size_t>(to_end-to_nxt)) // is there room? 1665 return partial; 1666 for (extern_type* p = tmp; n; --n) // write it 1667 *to_nxt++ = *p++; 1668 return ok; 1669 } 1670 1671 int 1672 codecvt<wchar_t, char, mbstate_t>::do_encoding() const _NOEXCEPT 1673 { 1674 if (__libcpp_mbtowc_l(nullptr, nullptr, MB_LEN_MAX, __l) != 0) 1675 return -1; 1676 1677 // stateless encoding 1678 if (__l == 0 || __libcpp_mb_cur_max_l(__l) == 1) // there are no known constant length encodings 1679 return 1; // which take more than 1 char to form a wchar_t 1680 return 0; 1681 } 1682 1683 bool 1684 codecvt<wchar_t, char, mbstate_t>::do_always_noconv() const _NOEXCEPT 1685 { 1686 return false; 1687 } 1688 1689 int 1690 codecvt<wchar_t, char, mbstate_t>::do_length(state_type& st, 1691 const extern_type* frm, const extern_type* frm_end, size_t mx) const 1692 { 1693 int nbytes = 0; 1694 for (size_t nwchar_t = 0; nwchar_t < mx && frm != frm_end; ++nwchar_t) 1695 { 1696 size_t n = __libcpp_mbrlen_l(frm, static_cast<size_t>(frm_end-frm), &st, __l); 1697 switch (n) 1698 { 1699 case 0: 1700 ++nbytes; 1701 ++frm; 1702 break; 1703 case size_t(-1): 1704 case size_t(-2): 1705 return nbytes; 1706 default: 1707 nbytes += n; 1708 frm += n; 1709 break; 1710 } 1711 } 1712 return nbytes; 1713 } 1714 1715 int 1716 codecvt<wchar_t, char, mbstate_t>::do_max_length() const _NOEXCEPT 1717 { 1718 return __l == 0 ? 1 : static_cast<int>(__libcpp_mb_cur_max_l(__l)); 1719 } 1720 1721 // Valid UTF ranges 1722 // UTF-32 UTF-16 UTF-8 # of code points 1723 // first second first second third fourth 1724 // 000000 - 00007F 0000 - 007F 00 - 7F 127 1725 // 000080 - 0007FF 0080 - 07FF C2 - DF, 80 - BF 1920 1726 // 000800 - 000FFF 0800 - 0FFF E0 - E0, A0 - BF, 80 - BF 2048 1727 // 001000 - 00CFFF 1000 - CFFF E1 - EC, 80 - BF, 80 - BF 49152 1728 // 00D000 - 00D7FF D000 - D7FF ED - ED, 80 - 9F, 80 - BF 2048 1729 // 00D800 - 00DFFF invalid 1730 // 00E000 - 00FFFF E000 - FFFF EE - EF, 80 - BF, 80 - BF 8192 1731 // 010000 - 03FFFF D800 - D8BF, DC00 - DFFF F0 - F0, 90 - BF, 80 - BF, 80 - BF 196608 1732 // 040000 - 0FFFFF D8C0 - DBBF, DC00 - DFFF F1 - F3, 80 - BF, 80 - BF, 80 - BF 786432 1733 // 100000 - 10FFFF DBC0 - DBFF, DC00 - DFFF F4 - F4, 80 - 8F, 80 - BF, 80 - BF 65536 1734 1735 static 1736 codecvt_base::result 1737 utf16_to_utf8(const uint16_t* frm, const uint16_t* frm_end, const uint16_t*& frm_nxt, 1738 uint8_t* to, uint8_t* to_end, uint8_t*& to_nxt, 1739 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 1740 { 1741 frm_nxt = frm; 1742 to_nxt = to; 1743 if (mode & generate_header) 1744 { 1745 if (to_end-to_nxt < 3) 1746 return codecvt_base::partial; 1747 *to_nxt++ = static_cast<uint8_t>(0xEF); 1748 *to_nxt++ = static_cast<uint8_t>(0xBB); 1749 *to_nxt++ = static_cast<uint8_t>(0xBF); 1750 } 1751 for (; frm_nxt < frm_end; ++frm_nxt) 1752 { 1753 uint16_t wc1 = *frm_nxt; 1754 if (wc1 > Maxcode) 1755 return codecvt_base::error; 1756 if (wc1 < 0x0080) 1757 { 1758 if (to_end-to_nxt < 1) 1759 return codecvt_base::partial; 1760 *to_nxt++ = static_cast<uint8_t>(wc1); 1761 } 1762 else if (wc1 < 0x0800) 1763 { 1764 if (to_end-to_nxt < 2) 1765 return codecvt_base::partial; 1766 *to_nxt++ = static_cast<uint8_t>(0xC0 | (wc1 >> 6)); 1767 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc1 & 0x03F)); 1768 } 1769 else if (wc1 < 0xD800) 1770 { 1771 if (to_end-to_nxt < 3) 1772 return codecvt_base::partial; 1773 *to_nxt++ = static_cast<uint8_t>(0xE0 | (wc1 >> 12)); 1774 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc1 & 0x0FC0) >> 6)); 1775 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc1 & 0x003F)); 1776 } 1777 else if (wc1 < 0xDC00) 1778 { 1779 if (frm_end-frm_nxt < 2) 1780 return codecvt_base::partial; 1781 uint16_t wc2 = frm_nxt[1]; 1782 if ((wc2 & 0xFC00) != 0xDC00) 1783 return codecvt_base::error; 1784 if (to_end-to_nxt < 4) 1785 return codecvt_base::partial; 1786 if (((((wc1 & 0x03C0UL) >> 6) + 1) << 16) + 1787 ((wc1 & 0x003FUL) << 10) + (wc2 & 0x03FF) > Maxcode) 1788 return codecvt_base::error; 1789 ++frm_nxt; 1790 uint8_t z = ((wc1 & 0x03C0) >> 6) + 1; 1791 *to_nxt++ = static_cast<uint8_t>(0xF0 | (z >> 2)); 1792 *to_nxt++ = static_cast<uint8_t>(0x80 | ((z & 0x03) << 4) | ((wc1 & 0x003C) >> 2)); 1793 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc1 & 0x0003) << 4) | ((wc2 & 0x03C0) >> 6)); 1794 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc2 & 0x003F)); 1795 } 1796 else if (wc1 < 0xE000) 1797 { 1798 return codecvt_base::error; 1799 } 1800 else 1801 { 1802 if (to_end-to_nxt < 3) 1803 return codecvt_base::partial; 1804 *to_nxt++ = static_cast<uint8_t>(0xE0 | (wc1 >> 12)); 1805 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc1 & 0x0FC0) >> 6)); 1806 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc1 & 0x003F)); 1807 } 1808 } 1809 return codecvt_base::ok; 1810 } 1811 1812 static 1813 codecvt_base::result 1814 utf16_to_utf8(const uint32_t* frm, const uint32_t* frm_end, const uint32_t*& frm_nxt, 1815 uint8_t* to, uint8_t* to_end, uint8_t*& to_nxt, 1816 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 1817 { 1818 frm_nxt = frm; 1819 to_nxt = to; 1820 if (mode & generate_header) 1821 { 1822 if (to_end-to_nxt < 3) 1823 return codecvt_base::partial; 1824 *to_nxt++ = static_cast<uint8_t>(0xEF); 1825 *to_nxt++ = static_cast<uint8_t>(0xBB); 1826 *to_nxt++ = static_cast<uint8_t>(0xBF); 1827 } 1828 for (; frm_nxt < frm_end; ++frm_nxt) 1829 { 1830 uint16_t wc1 = static_cast<uint16_t>(*frm_nxt); 1831 if (wc1 > Maxcode) 1832 return codecvt_base::error; 1833 if (wc1 < 0x0080) 1834 { 1835 if (to_end-to_nxt < 1) 1836 return codecvt_base::partial; 1837 *to_nxt++ = static_cast<uint8_t>(wc1); 1838 } 1839 else if (wc1 < 0x0800) 1840 { 1841 if (to_end-to_nxt < 2) 1842 return codecvt_base::partial; 1843 *to_nxt++ = static_cast<uint8_t>(0xC0 | (wc1 >> 6)); 1844 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc1 & 0x03F)); 1845 } 1846 else if (wc1 < 0xD800) 1847 { 1848 if (to_end-to_nxt < 3) 1849 return codecvt_base::partial; 1850 *to_nxt++ = static_cast<uint8_t>(0xE0 | (wc1 >> 12)); 1851 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc1 & 0x0FC0) >> 6)); 1852 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc1 & 0x003F)); 1853 } 1854 else if (wc1 < 0xDC00) 1855 { 1856 if (frm_end-frm_nxt < 2) 1857 return codecvt_base::partial; 1858 uint16_t wc2 = static_cast<uint16_t>(frm_nxt[1]); 1859 if ((wc2 & 0xFC00) != 0xDC00) 1860 return codecvt_base::error; 1861 if (to_end-to_nxt < 4) 1862 return codecvt_base::partial; 1863 if (((((wc1 & 0x03C0UL) >> 6) + 1) << 16) + 1864 ((wc1 & 0x003FUL) << 10) + (wc2 & 0x03FF) > Maxcode) 1865 return codecvt_base::error; 1866 ++frm_nxt; 1867 uint8_t z = ((wc1 & 0x03C0) >> 6) + 1; 1868 *to_nxt++ = static_cast<uint8_t>(0xF0 | (z >> 2)); 1869 *to_nxt++ = static_cast<uint8_t>(0x80 | ((z & 0x03) << 4) | ((wc1 & 0x003C) >> 2)); 1870 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc1 & 0x0003) << 4) | ((wc2 & 0x03C0) >> 6)); 1871 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc2 & 0x003F)); 1872 } 1873 else if (wc1 < 0xE000) 1874 { 1875 return codecvt_base::error; 1876 } 1877 else 1878 { 1879 if (to_end-to_nxt < 3) 1880 return codecvt_base::partial; 1881 *to_nxt++ = static_cast<uint8_t>(0xE0 | (wc1 >> 12)); 1882 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc1 & 0x0FC0) >> 6)); 1883 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc1 & 0x003F)); 1884 } 1885 } 1886 return codecvt_base::ok; 1887 } 1888 1889 static 1890 codecvt_base::result 1891 utf8_to_utf16(const uint8_t* frm, const uint8_t* frm_end, const uint8_t*& frm_nxt, 1892 uint16_t* to, uint16_t* to_end, uint16_t*& to_nxt, 1893 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 1894 { 1895 frm_nxt = frm; 1896 to_nxt = to; 1897 if (mode & consume_header) 1898 { 1899 if (frm_end-frm_nxt >= 3 && frm_nxt[0] == 0xEF && frm_nxt[1] == 0xBB && 1900 frm_nxt[2] == 0xBF) 1901 frm_nxt += 3; 1902 } 1903 for (; frm_nxt < frm_end && to_nxt < to_end; ++to_nxt) 1904 { 1905 uint8_t c1 = *frm_nxt; 1906 if (c1 > Maxcode) 1907 return codecvt_base::error; 1908 if (c1 < 0x80) 1909 { 1910 *to_nxt = static_cast<uint16_t>(c1); 1911 ++frm_nxt; 1912 } 1913 else if (c1 < 0xC2) 1914 { 1915 return codecvt_base::error; 1916 } 1917 else if (c1 < 0xE0) 1918 { 1919 if (frm_end-frm_nxt < 2) 1920 return codecvt_base::partial; 1921 uint8_t c2 = frm_nxt[1]; 1922 if ((c2 & 0xC0) != 0x80) 1923 return codecvt_base::error; 1924 uint16_t t = static_cast<uint16_t>(((c1 & 0x1F) << 6) | (c2 & 0x3F)); 1925 if (t > Maxcode) 1926 return codecvt_base::error; 1927 *to_nxt = t; 1928 frm_nxt += 2; 1929 } 1930 else if (c1 < 0xF0) 1931 { 1932 if (frm_end-frm_nxt < 3) 1933 return codecvt_base::partial; 1934 uint8_t c2 = frm_nxt[1]; 1935 uint8_t c3 = frm_nxt[2]; 1936 switch (c1) 1937 { 1938 case 0xE0: 1939 if ((c2 & 0xE0) != 0xA0) 1940 return codecvt_base::error; 1941 break; 1942 case 0xED: 1943 if ((c2 & 0xE0) != 0x80) 1944 return codecvt_base::error; 1945 break; 1946 default: 1947 if ((c2 & 0xC0) != 0x80) 1948 return codecvt_base::error; 1949 break; 1950 } 1951 if ((c3 & 0xC0) != 0x80) 1952 return codecvt_base::error; 1953 uint16_t t = static_cast<uint16_t>(((c1 & 0x0F) << 12) 1954 | ((c2 & 0x3F) << 6) 1955 | (c3 & 0x3F)); 1956 if (t > Maxcode) 1957 return codecvt_base::error; 1958 *to_nxt = t; 1959 frm_nxt += 3; 1960 } 1961 else if (c1 < 0xF5) 1962 { 1963 if (frm_end-frm_nxt < 4) 1964 return codecvt_base::partial; 1965 uint8_t c2 = frm_nxt[1]; 1966 uint8_t c3 = frm_nxt[2]; 1967 uint8_t c4 = frm_nxt[3]; 1968 switch (c1) 1969 { 1970 case 0xF0: 1971 if (!(0x90 <= c2 && c2 <= 0xBF)) 1972 return codecvt_base::error; 1973 break; 1974 case 0xF4: 1975 if ((c2 & 0xF0) != 0x80) 1976 return codecvt_base::error; 1977 break; 1978 default: 1979 if ((c2 & 0xC0) != 0x80) 1980 return codecvt_base::error; 1981 break; 1982 } 1983 if ((c3 & 0xC0) != 0x80 || (c4 & 0xC0) != 0x80) 1984 return codecvt_base::error; 1985 if (to_end-to_nxt < 2) 1986 return codecvt_base::partial; 1987 if ((((c1 & 7UL) << 18) + 1988 ((c2 & 0x3FUL) << 12) + 1989 ((c3 & 0x3FUL) << 6) + (c4 & 0x3F)) > Maxcode) 1990 return codecvt_base::error; 1991 *to_nxt = static_cast<uint16_t>( 1992 0xD800 1993 | (((((c1 & 0x07) << 2) | ((c2 & 0x30) >> 4)) - 1) << 6) 1994 | ((c2 & 0x0F) << 2) 1995 | ((c3 & 0x30) >> 4)); 1996 *++to_nxt = static_cast<uint16_t>( 1997 0xDC00 1998 | ((c3 & 0x0F) << 6) 1999 | (c4 & 0x3F)); 2000 frm_nxt += 4; 2001 } 2002 else 2003 { 2004 return codecvt_base::error; 2005 } 2006 } 2007 return frm_nxt < frm_end ? codecvt_base::partial : codecvt_base::ok; 2008 } 2009 2010 static 2011 codecvt_base::result 2012 utf8_to_utf16(const uint8_t* frm, const uint8_t* frm_end, const uint8_t*& frm_nxt, 2013 uint32_t* to, uint32_t* to_end, uint32_t*& to_nxt, 2014 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2015 { 2016 frm_nxt = frm; 2017 to_nxt = to; 2018 if (mode & consume_header) 2019 { 2020 if (frm_end-frm_nxt >= 3 && frm_nxt[0] == 0xEF && frm_nxt[1] == 0xBB && 2021 frm_nxt[2] == 0xBF) 2022 frm_nxt += 3; 2023 } 2024 for (; frm_nxt < frm_end && to_nxt < to_end; ++to_nxt) 2025 { 2026 uint8_t c1 = *frm_nxt; 2027 if (c1 > Maxcode) 2028 return codecvt_base::error; 2029 if (c1 < 0x80) 2030 { 2031 *to_nxt = static_cast<uint32_t>(c1); 2032 ++frm_nxt; 2033 } 2034 else if (c1 < 0xC2) 2035 { 2036 return codecvt_base::error; 2037 } 2038 else if (c1 < 0xE0) 2039 { 2040 if (frm_end-frm_nxt < 2) 2041 return codecvt_base::partial; 2042 uint8_t c2 = frm_nxt[1]; 2043 if ((c2 & 0xC0) != 0x80) 2044 return codecvt_base::error; 2045 uint16_t t = static_cast<uint16_t>(((c1 & 0x1F) << 6) | (c2 & 0x3F)); 2046 if (t > Maxcode) 2047 return codecvt_base::error; 2048 *to_nxt = static_cast<uint32_t>(t); 2049 frm_nxt += 2; 2050 } 2051 else if (c1 < 0xF0) 2052 { 2053 if (frm_end-frm_nxt < 3) 2054 return codecvt_base::partial; 2055 uint8_t c2 = frm_nxt[1]; 2056 uint8_t c3 = frm_nxt[2]; 2057 switch (c1) 2058 { 2059 case 0xE0: 2060 if ((c2 & 0xE0) != 0xA0) 2061 return codecvt_base::error; 2062 break; 2063 case 0xED: 2064 if ((c2 & 0xE0) != 0x80) 2065 return codecvt_base::error; 2066 break; 2067 default: 2068 if ((c2 & 0xC0) != 0x80) 2069 return codecvt_base::error; 2070 break; 2071 } 2072 if ((c3 & 0xC0) != 0x80) 2073 return codecvt_base::error; 2074 uint16_t t = static_cast<uint16_t>(((c1 & 0x0F) << 12) 2075 | ((c2 & 0x3F) << 6) 2076 | (c3 & 0x3F)); 2077 if (t > Maxcode) 2078 return codecvt_base::error; 2079 *to_nxt = static_cast<uint32_t>(t); 2080 frm_nxt += 3; 2081 } 2082 else if (c1 < 0xF5) 2083 { 2084 if (frm_end-frm_nxt < 4) 2085 return codecvt_base::partial; 2086 uint8_t c2 = frm_nxt[1]; 2087 uint8_t c3 = frm_nxt[2]; 2088 uint8_t c4 = frm_nxt[3]; 2089 switch (c1) 2090 { 2091 case 0xF0: 2092 if (!(0x90 <= c2 && c2 <= 0xBF)) 2093 return codecvt_base::error; 2094 break; 2095 case 0xF4: 2096 if ((c2 & 0xF0) != 0x80) 2097 return codecvt_base::error; 2098 break; 2099 default: 2100 if ((c2 & 0xC0) != 0x80) 2101 return codecvt_base::error; 2102 break; 2103 } 2104 if ((c3 & 0xC0) != 0x80 || (c4 & 0xC0) != 0x80) 2105 return codecvt_base::error; 2106 if (to_end-to_nxt < 2) 2107 return codecvt_base::partial; 2108 if ((((c1 & 7UL) << 18) + 2109 ((c2 & 0x3FUL) << 12) + 2110 ((c3 & 0x3FUL) << 6) + (c4 & 0x3F)) > Maxcode) 2111 return codecvt_base::error; 2112 *to_nxt = static_cast<uint32_t>( 2113 0xD800 2114 | (((((c1 & 0x07) << 2) | ((c2 & 0x30) >> 4)) - 1) << 6) 2115 | ((c2 & 0x0F) << 2) 2116 | ((c3 & 0x30) >> 4)); 2117 *++to_nxt = static_cast<uint32_t>( 2118 0xDC00 2119 | ((c3 & 0x0F) << 6) 2120 | (c4 & 0x3F)); 2121 frm_nxt += 4; 2122 } 2123 else 2124 { 2125 return codecvt_base::error; 2126 } 2127 } 2128 return frm_nxt < frm_end ? codecvt_base::partial : codecvt_base::ok; 2129 } 2130 2131 static 2132 int 2133 utf8_to_utf16_length(const uint8_t* frm, const uint8_t* frm_end, 2134 size_t mx, unsigned long Maxcode = 0x10FFFF, 2135 codecvt_mode mode = codecvt_mode(0)) 2136 { 2137 const uint8_t* frm_nxt = frm; 2138 if (mode & consume_header) 2139 { 2140 if (frm_end-frm_nxt >= 3 && frm_nxt[0] == 0xEF && frm_nxt[1] == 0xBB && 2141 frm_nxt[2] == 0xBF) 2142 frm_nxt += 3; 2143 } 2144 for (size_t nchar16_t = 0; frm_nxt < frm_end && nchar16_t < mx; ++nchar16_t) 2145 { 2146 uint8_t c1 = *frm_nxt; 2147 if (c1 > Maxcode) 2148 break; 2149 if (c1 < 0x80) 2150 { 2151 ++frm_nxt; 2152 } 2153 else if (c1 < 0xC2) 2154 { 2155 break; 2156 } 2157 else if (c1 < 0xE0) 2158 { 2159 if ((frm_end-frm_nxt < 2) || (frm_nxt[1] & 0xC0) != 0x80) 2160 break; 2161 uint16_t t = static_cast<uint16_t>(((c1 & 0x1F) << 6) | (frm_nxt[1] & 0x3F)); 2162 if (t > Maxcode) 2163 break; 2164 frm_nxt += 2; 2165 } 2166 else if (c1 < 0xF0) 2167 { 2168 if (frm_end-frm_nxt < 3) 2169 break; 2170 uint8_t c2 = frm_nxt[1]; 2171 uint8_t c3 = frm_nxt[2]; 2172 switch (c1) 2173 { 2174 case 0xE0: 2175 if ((c2 & 0xE0) != 0xA0) 2176 return static_cast<int>(frm_nxt - frm); 2177 break; 2178 case 0xED: 2179 if ((c2 & 0xE0) != 0x80) 2180 return static_cast<int>(frm_nxt - frm); 2181 break; 2182 default: 2183 if ((c2 & 0xC0) != 0x80) 2184 return static_cast<int>(frm_nxt - frm); 2185 break; 2186 } 2187 if ((c3 & 0xC0) != 0x80) 2188 break; 2189 if ((((c1 & 0x0Fu) << 12) | ((c2 & 0x3Fu) << 6) | (c3 & 0x3Fu)) > Maxcode) 2190 break; 2191 frm_nxt += 3; 2192 } 2193 else if (c1 < 0xF5) 2194 { 2195 if (frm_end-frm_nxt < 4 || mx-nchar16_t < 2) 2196 break; 2197 uint8_t c2 = frm_nxt[1]; 2198 uint8_t c3 = frm_nxt[2]; 2199 uint8_t c4 = frm_nxt[3]; 2200 switch (c1) 2201 { 2202 case 0xF0: 2203 if (!(0x90 <= c2 && c2 <= 0xBF)) 2204 return static_cast<int>(frm_nxt - frm); 2205 break; 2206 case 0xF4: 2207 if ((c2 & 0xF0) != 0x80) 2208 return static_cast<int>(frm_nxt - frm); 2209 break; 2210 default: 2211 if ((c2 & 0xC0) != 0x80) 2212 return static_cast<int>(frm_nxt - frm); 2213 break; 2214 } 2215 if ((c3 & 0xC0) != 0x80 || (c4 & 0xC0) != 0x80) 2216 break; 2217 if ((((c1 & 7UL) << 18) + 2218 ((c2 & 0x3FUL) << 12) + 2219 ((c3 & 0x3FUL) << 6) + (c4 & 0x3F)) > Maxcode) 2220 break; 2221 ++nchar16_t; 2222 frm_nxt += 4; 2223 } 2224 else 2225 { 2226 break; 2227 } 2228 } 2229 return static_cast<int>(frm_nxt - frm); 2230 } 2231 2232 static 2233 codecvt_base::result 2234 ucs4_to_utf8(const uint32_t* frm, const uint32_t* frm_end, const uint32_t*& frm_nxt, 2235 uint8_t* to, uint8_t* to_end, uint8_t*& to_nxt, 2236 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2237 { 2238 frm_nxt = frm; 2239 to_nxt = to; 2240 if (mode & generate_header) 2241 { 2242 if (to_end-to_nxt < 3) 2243 return codecvt_base::partial; 2244 *to_nxt++ = static_cast<uint8_t>(0xEF); 2245 *to_nxt++ = static_cast<uint8_t>(0xBB); 2246 *to_nxt++ = static_cast<uint8_t>(0xBF); 2247 } 2248 for (; frm_nxt < frm_end; ++frm_nxt) 2249 { 2250 uint32_t wc = *frm_nxt; 2251 if ((wc & 0xFFFFF800) == 0x00D800 || wc > Maxcode) 2252 return codecvt_base::error; 2253 if (wc < 0x000080) 2254 { 2255 if (to_end-to_nxt < 1) 2256 return codecvt_base::partial; 2257 *to_nxt++ = static_cast<uint8_t>(wc); 2258 } 2259 else if (wc < 0x000800) 2260 { 2261 if (to_end-to_nxt < 2) 2262 return codecvt_base::partial; 2263 *to_nxt++ = static_cast<uint8_t>(0xC0 | (wc >> 6)); 2264 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc & 0x03F)); 2265 } 2266 else if (wc < 0x010000) 2267 { 2268 if (to_end-to_nxt < 3) 2269 return codecvt_base::partial; 2270 *to_nxt++ = static_cast<uint8_t>(0xE0 | (wc >> 12)); 2271 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc & 0x0FC0) >> 6)); 2272 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc & 0x003F)); 2273 } 2274 else // if (wc < 0x110000) 2275 { 2276 if (to_end-to_nxt < 4) 2277 return codecvt_base::partial; 2278 *to_nxt++ = static_cast<uint8_t>(0xF0 | (wc >> 18)); 2279 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc & 0x03F000) >> 12)); 2280 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc & 0x000FC0) >> 6)); 2281 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc & 0x00003F)); 2282 } 2283 } 2284 return codecvt_base::ok; 2285 } 2286 2287 static 2288 codecvt_base::result 2289 utf8_to_ucs4(const uint8_t* frm, const uint8_t* frm_end, const uint8_t*& frm_nxt, 2290 uint32_t* to, uint32_t* to_end, uint32_t*& to_nxt, 2291 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2292 { 2293 frm_nxt = frm; 2294 to_nxt = to; 2295 if (mode & consume_header) 2296 { 2297 if (frm_end-frm_nxt >= 3 && frm_nxt[0] == 0xEF && frm_nxt[1] == 0xBB && 2298 frm_nxt[2] == 0xBF) 2299 frm_nxt += 3; 2300 } 2301 for (; frm_nxt < frm_end && to_nxt < to_end; ++to_nxt) 2302 { 2303 uint8_t c1 = static_cast<uint8_t>(*frm_nxt); 2304 if (c1 < 0x80) 2305 { 2306 if (c1 > Maxcode) 2307 return codecvt_base::error; 2308 *to_nxt = static_cast<uint32_t>(c1); 2309 ++frm_nxt; 2310 } 2311 else if (c1 < 0xC2) 2312 { 2313 return codecvt_base::error; 2314 } 2315 else if (c1 < 0xE0) 2316 { 2317 if (frm_end-frm_nxt < 2) 2318 return codecvt_base::partial; 2319 uint8_t c2 = frm_nxt[1]; 2320 if ((c2 & 0xC0) != 0x80) 2321 return codecvt_base::error; 2322 uint32_t t = static_cast<uint32_t>(((c1 & 0x1F) << 6) 2323 | (c2 & 0x3F)); 2324 if (t > Maxcode) 2325 return codecvt_base::error; 2326 *to_nxt = t; 2327 frm_nxt += 2; 2328 } 2329 else if (c1 < 0xF0) 2330 { 2331 if (frm_end-frm_nxt < 3) 2332 return codecvt_base::partial; 2333 uint8_t c2 = frm_nxt[1]; 2334 uint8_t c3 = frm_nxt[2]; 2335 switch (c1) 2336 { 2337 case 0xE0: 2338 if ((c2 & 0xE0) != 0xA0) 2339 return codecvt_base::error; 2340 break; 2341 case 0xED: 2342 if ((c2 & 0xE0) != 0x80) 2343 return codecvt_base::error; 2344 break; 2345 default: 2346 if ((c2 & 0xC0) != 0x80) 2347 return codecvt_base::error; 2348 break; 2349 } 2350 if ((c3 & 0xC0) != 0x80) 2351 return codecvt_base::error; 2352 uint32_t t = static_cast<uint32_t>(((c1 & 0x0F) << 12) 2353 | ((c2 & 0x3F) << 6) 2354 | (c3 & 0x3F)); 2355 if (t > Maxcode) 2356 return codecvt_base::error; 2357 *to_nxt = t; 2358 frm_nxt += 3; 2359 } 2360 else if (c1 < 0xF5) 2361 { 2362 if (frm_end-frm_nxt < 4) 2363 return codecvt_base::partial; 2364 uint8_t c2 = frm_nxt[1]; 2365 uint8_t c3 = frm_nxt[2]; 2366 uint8_t c4 = frm_nxt[3]; 2367 switch (c1) 2368 { 2369 case 0xF0: 2370 if (!(0x90 <= c2 && c2 <= 0xBF)) 2371 return codecvt_base::error; 2372 break; 2373 case 0xF4: 2374 if ((c2 & 0xF0) != 0x80) 2375 return codecvt_base::error; 2376 break; 2377 default: 2378 if ((c2 & 0xC0) != 0x80) 2379 return codecvt_base::error; 2380 break; 2381 } 2382 if ((c3 & 0xC0) != 0x80 || (c4 & 0xC0) != 0x80) 2383 return codecvt_base::error; 2384 uint32_t t = static_cast<uint32_t>(((c1 & 0x07) << 18) 2385 | ((c2 & 0x3F) << 12) 2386 | ((c3 & 0x3F) << 6) 2387 | (c4 & 0x3F)); 2388 if (t > Maxcode) 2389 return codecvt_base::error; 2390 *to_nxt = t; 2391 frm_nxt += 4; 2392 } 2393 else 2394 { 2395 return codecvt_base::error; 2396 } 2397 } 2398 return frm_nxt < frm_end ? codecvt_base::partial : codecvt_base::ok; 2399 } 2400 2401 static 2402 int 2403 utf8_to_ucs4_length(const uint8_t* frm, const uint8_t* frm_end, 2404 size_t mx, unsigned long Maxcode = 0x10FFFF, 2405 codecvt_mode mode = codecvt_mode(0)) 2406 { 2407 const uint8_t* frm_nxt = frm; 2408 if (mode & consume_header) 2409 { 2410 if (frm_end-frm_nxt >= 3 && frm_nxt[0] == 0xEF && frm_nxt[1] == 0xBB && 2411 frm_nxt[2] == 0xBF) 2412 frm_nxt += 3; 2413 } 2414 for (size_t nchar32_t = 0; frm_nxt < frm_end && nchar32_t < mx; ++nchar32_t) 2415 { 2416 uint8_t c1 = static_cast<uint8_t>(*frm_nxt); 2417 if (c1 < 0x80) 2418 { 2419 if (c1 > Maxcode) 2420 break; 2421 ++frm_nxt; 2422 } 2423 else if (c1 < 0xC2) 2424 { 2425 break; 2426 } 2427 else if (c1 < 0xE0) 2428 { 2429 if ((frm_end-frm_nxt < 2) || ((frm_nxt[1] & 0xC0) != 0x80)) 2430 break; 2431 if ((((c1 & 0x1Fu) << 6) | (frm_nxt[1] & 0x3Fu)) > Maxcode) 2432 break; 2433 frm_nxt += 2; 2434 } 2435 else if (c1 < 0xF0) 2436 { 2437 if (frm_end-frm_nxt < 3) 2438 break; 2439 uint8_t c2 = frm_nxt[1]; 2440 uint8_t c3 = frm_nxt[2]; 2441 switch (c1) 2442 { 2443 case 0xE0: 2444 if ((c2 & 0xE0) != 0xA0) 2445 return static_cast<int>(frm_nxt - frm); 2446 break; 2447 case 0xED: 2448 if ((c2 & 0xE0) != 0x80) 2449 return static_cast<int>(frm_nxt - frm); 2450 break; 2451 default: 2452 if ((c2 & 0xC0) != 0x80) 2453 return static_cast<int>(frm_nxt - frm); 2454 break; 2455 } 2456 if ((c3 & 0xC0) != 0x80) 2457 break; 2458 if ((((c1 & 0x0Fu) << 12) | ((c2 & 0x3Fu) << 6) | (c3 & 0x3Fu)) > Maxcode) 2459 break; 2460 frm_nxt += 3; 2461 } 2462 else if (c1 < 0xF5) 2463 { 2464 if (frm_end-frm_nxt < 4) 2465 break; 2466 uint8_t c2 = frm_nxt[1]; 2467 uint8_t c3 = frm_nxt[2]; 2468 uint8_t c4 = frm_nxt[3]; 2469 switch (c1) 2470 { 2471 case 0xF0: 2472 if (!(0x90 <= c2 && c2 <= 0xBF)) 2473 return static_cast<int>(frm_nxt - frm); 2474 break; 2475 case 0xF4: 2476 if ((c2 & 0xF0) != 0x80) 2477 return static_cast<int>(frm_nxt - frm); 2478 break; 2479 default: 2480 if ((c2 & 0xC0) != 0x80) 2481 return static_cast<int>(frm_nxt - frm); 2482 break; 2483 } 2484 if ((c3 & 0xC0) != 0x80 || (c4 & 0xC0) != 0x80) 2485 break; 2486 if ((((c1 & 0x07u) << 18) | ((c2 & 0x3Fu) << 12) | 2487 ((c3 & 0x3Fu) << 6) | (c4 & 0x3Fu)) > Maxcode) 2488 break; 2489 frm_nxt += 4; 2490 } 2491 else 2492 { 2493 break; 2494 } 2495 } 2496 return static_cast<int>(frm_nxt - frm); 2497 } 2498 2499 static 2500 codecvt_base::result 2501 ucs2_to_utf8(const uint16_t* frm, const uint16_t* frm_end, const uint16_t*& frm_nxt, 2502 uint8_t* to, uint8_t* to_end, uint8_t*& to_nxt, 2503 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2504 { 2505 frm_nxt = frm; 2506 to_nxt = to; 2507 if (mode & generate_header) 2508 { 2509 if (to_end-to_nxt < 3) 2510 return codecvt_base::partial; 2511 *to_nxt++ = static_cast<uint8_t>(0xEF); 2512 *to_nxt++ = static_cast<uint8_t>(0xBB); 2513 *to_nxt++ = static_cast<uint8_t>(0xBF); 2514 } 2515 for (; frm_nxt < frm_end; ++frm_nxt) 2516 { 2517 uint16_t wc = *frm_nxt; 2518 if ((wc & 0xF800) == 0xD800 || wc > Maxcode) 2519 return codecvt_base::error; 2520 if (wc < 0x0080) 2521 { 2522 if (to_end-to_nxt < 1) 2523 return codecvt_base::partial; 2524 *to_nxt++ = static_cast<uint8_t>(wc); 2525 } 2526 else if (wc < 0x0800) 2527 { 2528 if (to_end-to_nxt < 2) 2529 return codecvt_base::partial; 2530 *to_nxt++ = static_cast<uint8_t>(0xC0 | (wc >> 6)); 2531 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc & 0x03F)); 2532 } 2533 else // if (wc <= 0xFFFF) 2534 { 2535 if (to_end-to_nxt < 3) 2536 return codecvt_base::partial; 2537 *to_nxt++ = static_cast<uint8_t>(0xE0 | (wc >> 12)); 2538 *to_nxt++ = static_cast<uint8_t>(0x80 | ((wc & 0x0FC0) >> 6)); 2539 *to_nxt++ = static_cast<uint8_t>(0x80 | (wc & 0x003F)); 2540 } 2541 } 2542 return codecvt_base::ok; 2543 } 2544 2545 static 2546 codecvt_base::result 2547 utf8_to_ucs2(const uint8_t* frm, const uint8_t* frm_end, const uint8_t*& frm_nxt, 2548 uint16_t* to, uint16_t* to_end, uint16_t*& to_nxt, 2549 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2550 { 2551 frm_nxt = frm; 2552 to_nxt = to; 2553 if (mode & consume_header) 2554 { 2555 if (frm_end-frm_nxt >= 3 && frm_nxt[0] == 0xEF && frm_nxt[1] == 0xBB && 2556 frm_nxt[2] == 0xBF) 2557 frm_nxt += 3; 2558 } 2559 for (; frm_nxt < frm_end && to_nxt < to_end; ++to_nxt) 2560 { 2561 uint8_t c1 = static_cast<uint8_t>(*frm_nxt); 2562 if (c1 < 0x80) 2563 { 2564 if (c1 > Maxcode) 2565 return codecvt_base::error; 2566 *to_nxt = static_cast<uint16_t>(c1); 2567 ++frm_nxt; 2568 } 2569 else if (c1 < 0xC2) 2570 { 2571 return codecvt_base::error; 2572 } 2573 else if (c1 < 0xE0) 2574 { 2575 if (frm_end-frm_nxt < 2) 2576 return codecvt_base::partial; 2577 uint8_t c2 = frm_nxt[1]; 2578 if ((c2 & 0xC0) != 0x80) 2579 return codecvt_base::error; 2580 uint16_t t = static_cast<uint16_t>(((c1 & 0x1F) << 6) 2581 | (c2 & 0x3F)); 2582 if (t > Maxcode) 2583 return codecvt_base::error; 2584 *to_nxt = t; 2585 frm_nxt += 2; 2586 } 2587 else if (c1 < 0xF0) 2588 { 2589 if (frm_end-frm_nxt < 3) 2590 return codecvt_base::partial; 2591 uint8_t c2 = frm_nxt[1]; 2592 uint8_t c3 = frm_nxt[2]; 2593 switch (c1) 2594 { 2595 case 0xE0: 2596 if ((c2 & 0xE0) != 0xA0) 2597 return codecvt_base::error; 2598 break; 2599 case 0xED: 2600 if ((c2 & 0xE0) != 0x80) 2601 return codecvt_base::error; 2602 break; 2603 default: 2604 if ((c2 & 0xC0) != 0x80) 2605 return codecvt_base::error; 2606 break; 2607 } 2608 if ((c3 & 0xC0) != 0x80) 2609 return codecvt_base::error; 2610 uint16_t t = static_cast<uint16_t>(((c1 & 0x0F) << 12) 2611 | ((c2 & 0x3F) << 6) 2612 | (c3 & 0x3F)); 2613 if (t > Maxcode) 2614 return codecvt_base::error; 2615 *to_nxt = t; 2616 frm_nxt += 3; 2617 } 2618 else 2619 { 2620 return codecvt_base::error; 2621 } 2622 } 2623 return frm_nxt < frm_end ? codecvt_base::partial : codecvt_base::ok; 2624 } 2625 2626 static 2627 int 2628 utf8_to_ucs2_length(const uint8_t* frm, const uint8_t* frm_end, 2629 size_t mx, unsigned long Maxcode = 0x10FFFF, 2630 codecvt_mode mode = codecvt_mode(0)) 2631 { 2632 const uint8_t* frm_nxt = frm; 2633 if (mode & consume_header) 2634 { 2635 if (frm_end-frm_nxt >= 3 && frm_nxt[0] == 0xEF && frm_nxt[1] == 0xBB && 2636 frm_nxt[2] == 0xBF) 2637 frm_nxt += 3; 2638 } 2639 for (size_t nchar32_t = 0; frm_nxt < frm_end && nchar32_t < mx; ++nchar32_t) 2640 { 2641 uint8_t c1 = static_cast<uint8_t>(*frm_nxt); 2642 if (c1 < 0x80) 2643 { 2644 if (c1 > Maxcode) 2645 break; 2646 ++frm_nxt; 2647 } 2648 else if (c1 < 0xC2) 2649 { 2650 break; 2651 } 2652 else if (c1 < 0xE0) 2653 { 2654 if ((frm_end-frm_nxt < 2) || ((frm_nxt[1] & 0xC0) != 0x80)) 2655 break; 2656 if ((((c1 & 0x1Fu) << 6) | (frm_nxt[1] & 0x3Fu)) > Maxcode) 2657 break; 2658 frm_nxt += 2; 2659 } 2660 else if (c1 < 0xF0) 2661 { 2662 if (frm_end-frm_nxt < 3) 2663 break; 2664 uint8_t c2 = frm_nxt[1]; 2665 uint8_t c3 = frm_nxt[2]; 2666 switch (c1) 2667 { 2668 case 0xE0: 2669 if ((c2 & 0xE0) != 0xA0) 2670 return static_cast<int>(frm_nxt - frm); 2671 break; 2672 case 0xED: 2673 if ((c2 & 0xE0) != 0x80) 2674 return static_cast<int>(frm_nxt - frm); 2675 break; 2676 default: 2677 if ((c2 & 0xC0) != 0x80) 2678 return static_cast<int>(frm_nxt - frm); 2679 break; 2680 } 2681 if ((c3 & 0xC0) != 0x80) 2682 break; 2683 if ((((c1 & 0x0Fu) << 12) | ((c2 & 0x3Fu) << 6) | (c3 & 0x3Fu)) > Maxcode) 2684 break; 2685 frm_nxt += 3; 2686 } 2687 else 2688 { 2689 break; 2690 } 2691 } 2692 return static_cast<int>(frm_nxt - frm); 2693 } 2694 2695 static 2696 codecvt_base::result 2697 ucs4_to_utf16be(const uint32_t* frm, const uint32_t* frm_end, const uint32_t*& frm_nxt, 2698 uint8_t* to, uint8_t* to_end, uint8_t*& to_nxt, 2699 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2700 { 2701 frm_nxt = frm; 2702 to_nxt = to; 2703 if (mode & generate_header) 2704 { 2705 if (to_end-to_nxt < 2) 2706 return codecvt_base::partial; 2707 *to_nxt++ = static_cast<uint8_t>(0xFE); 2708 *to_nxt++ = static_cast<uint8_t>(0xFF); 2709 } 2710 for (; frm_nxt < frm_end; ++frm_nxt) 2711 { 2712 uint32_t wc = *frm_nxt; 2713 if ((wc & 0xFFFFF800) == 0x00D800 || wc > Maxcode) 2714 return codecvt_base::error; 2715 if (wc < 0x010000) 2716 { 2717 if (to_end-to_nxt < 2) 2718 return codecvt_base::partial; 2719 *to_nxt++ = static_cast<uint8_t>(wc >> 8); 2720 *to_nxt++ = static_cast<uint8_t>(wc); 2721 } 2722 else 2723 { 2724 if (to_end-to_nxt < 4) 2725 return codecvt_base::partial; 2726 uint16_t t = static_cast<uint16_t>( 2727 0xD800 2728 | ((((wc & 0x1F0000) >> 16) - 1) << 6) 2729 | ((wc & 0x00FC00) >> 10)); 2730 *to_nxt++ = static_cast<uint8_t>(t >> 8); 2731 *to_nxt++ = static_cast<uint8_t>(t); 2732 t = static_cast<uint16_t>(0xDC00 | (wc & 0x03FF)); 2733 *to_nxt++ = static_cast<uint8_t>(t >> 8); 2734 *to_nxt++ = static_cast<uint8_t>(t); 2735 } 2736 } 2737 return codecvt_base::ok; 2738 } 2739 2740 static 2741 codecvt_base::result 2742 utf16be_to_ucs4(const uint8_t* frm, const uint8_t* frm_end, const uint8_t*& frm_nxt, 2743 uint32_t* to, uint32_t* to_end, uint32_t*& to_nxt, 2744 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2745 { 2746 frm_nxt = frm; 2747 to_nxt = to; 2748 if (mode & consume_header) 2749 { 2750 if (frm_end-frm_nxt >= 2 && frm_nxt[0] == 0xFE && frm_nxt[1] == 0xFF) 2751 frm_nxt += 2; 2752 } 2753 for (; frm_nxt < frm_end - 1 && to_nxt < to_end; ++to_nxt) 2754 { 2755 uint16_t c1 = static_cast<uint16_t>(frm_nxt[0] << 8 | frm_nxt[1]); 2756 if ((c1 & 0xFC00) == 0xDC00) 2757 return codecvt_base::error; 2758 if ((c1 & 0xFC00) != 0xD800) 2759 { 2760 if (c1 > Maxcode) 2761 return codecvt_base::error; 2762 *to_nxt = static_cast<uint32_t>(c1); 2763 frm_nxt += 2; 2764 } 2765 else 2766 { 2767 if (frm_end-frm_nxt < 4) 2768 return codecvt_base::partial; 2769 uint16_t c2 = static_cast<uint16_t>(frm_nxt[2] << 8 | frm_nxt[3]); 2770 if ((c2 & 0xFC00) != 0xDC00) 2771 return codecvt_base::error; 2772 uint32_t t = static_cast<uint32_t>( 2773 ((((c1 & 0x03C0) >> 6) + 1) << 16) 2774 | ((c1 & 0x003F) << 10) 2775 | (c2 & 0x03FF)); 2776 if (t > Maxcode) 2777 return codecvt_base::error; 2778 *to_nxt = t; 2779 frm_nxt += 4; 2780 } 2781 } 2782 return frm_nxt < frm_end ? codecvt_base::partial : codecvt_base::ok; 2783 } 2784 2785 static 2786 int 2787 utf16be_to_ucs4_length(const uint8_t* frm, const uint8_t* frm_end, 2788 size_t mx, unsigned long Maxcode = 0x10FFFF, 2789 codecvt_mode mode = codecvt_mode(0)) 2790 { 2791 const uint8_t* frm_nxt = frm; 2792 if (mode & consume_header) 2793 { 2794 if (frm_end-frm_nxt >= 2 && frm_nxt[0] == 0xFE && frm_nxt[1] == 0xFF) 2795 frm_nxt += 2; 2796 } 2797 for (size_t nchar32_t = 0; frm_nxt < frm_end - 1 && nchar32_t < mx; ++nchar32_t) 2798 { 2799 uint16_t c1 = static_cast<uint16_t>(frm_nxt[0] << 8 | frm_nxt[1]); 2800 if ((c1 & 0xFC00) == 0xDC00) 2801 break; 2802 if ((c1 & 0xFC00) != 0xD800) 2803 { 2804 if (c1 > Maxcode) 2805 break; 2806 frm_nxt += 2; 2807 } 2808 else 2809 { 2810 if (frm_end-frm_nxt < 4) 2811 break; 2812 uint16_t c2 = static_cast<uint16_t>(frm_nxt[2] << 8 | frm_nxt[3]); 2813 if ((c2 & 0xFC00) != 0xDC00) 2814 break; 2815 uint32_t t = static_cast<uint32_t>( 2816 ((((c1 & 0x03C0) >> 6) + 1) << 16) 2817 | ((c1 & 0x003F) << 10) 2818 | (c2 & 0x03FF)); 2819 if (t > Maxcode) 2820 break; 2821 frm_nxt += 4; 2822 } 2823 } 2824 return static_cast<int>(frm_nxt - frm); 2825 } 2826 2827 static 2828 codecvt_base::result 2829 ucs4_to_utf16le(const uint32_t* frm, const uint32_t* frm_end, const uint32_t*& frm_nxt, 2830 uint8_t* to, uint8_t* to_end, uint8_t*& to_nxt, 2831 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2832 { 2833 frm_nxt = frm; 2834 to_nxt = to; 2835 if (mode & generate_header) 2836 { 2837 if (to_end - to_nxt < 2) 2838 return codecvt_base::partial; 2839 *to_nxt++ = static_cast<uint8_t>(0xFF); 2840 *to_nxt++ = static_cast<uint8_t>(0xFE); 2841 } 2842 for (; frm_nxt < frm_end; ++frm_nxt) 2843 { 2844 uint32_t wc = *frm_nxt; 2845 if ((wc & 0xFFFFF800) == 0x00D800 || wc > Maxcode) 2846 return codecvt_base::error; 2847 if (wc < 0x010000) 2848 { 2849 if (to_end-to_nxt < 2) 2850 return codecvt_base::partial; 2851 *to_nxt++ = static_cast<uint8_t>(wc); 2852 *to_nxt++ = static_cast<uint8_t>(wc >> 8); 2853 } 2854 else 2855 { 2856 if (to_end-to_nxt < 4) 2857 return codecvt_base::partial; 2858 uint16_t t = static_cast<uint16_t>( 2859 0xD800 2860 | ((((wc & 0x1F0000) >> 16) - 1) << 6) 2861 | ((wc & 0x00FC00) >> 10)); 2862 *to_nxt++ = static_cast<uint8_t>(t); 2863 *to_nxt++ = static_cast<uint8_t>(t >> 8); 2864 t = static_cast<uint16_t>(0xDC00 | (wc & 0x03FF)); 2865 *to_nxt++ = static_cast<uint8_t>(t); 2866 *to_nxt++ = static_cast<uint8_t>(t >> 8); 2867 } 2868 } 2869 return codecvt_base::ok; 2870 } 2871 2872 static 2873 codecvt_base::result 2874 utf16le_to_ucs4(const uint8_t* frm, const uint8_t* frm_end, const uint8_t*& frm_nxt, 2875 uint32_t* to, uint32_t* to_end, uint32_t*& to_nxt, 2876 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2877 { 2878 frm_nxt = frm; 2879 to_nxt = to; 2880 if (mode & consume_header) 2881 { 2882 if (frm_end-frm_nxt >= 2 && frm_nxt[0] == 0xFF && frm_nxt[1] == 0xFE) 2883 frm_nxt += 2; 2884 } 2885 for (; frm_nxt < frm_end - 1 && to_nxt < to_end; ++to_nxt) 2886 { 2887 uint16_t c1 = static_cast<uint16_t>(frm_nxt[1] << 8 | frm_nxt[0]); 2888 if ((c1 & 0xFC00) == 0xDC00) 2889 return codecvt_base::error; 2890 if ((c1 & 0xFC00) != 0xD800) 2891 { 2892 if (c1 > Maxcode) 2893 return codecvt_base::error; 2894 *to_nxt = static_cast<uint32_t>(c1); 2895 frm_nxt += 2; 2896 } 2897 else 2898 { 2899 if (frm_end-frm_nxt < 4) 2900 return codecvt_base::partial; 2901 uint16_t c2 = static_cast<uint16_t>(frm_nxt[3] << 8 | frm_nxt[2]); 2902 if ((c2 & 0xFC00) != 0xDC00) 2903 return codecvt_base::error; 2904 uint32_t t = static_cast<uint32_t>( 2905 ((((c1 & 0x03C0) >> 6) + 1) << 16) 2906 | ((c1 & 0x003F) << 10) 2907 | (c2 & 0x03FF)); 2908 if (t > Maxcode) 2909 return codecvt_base::error; 2910 *to_nxt = t; 2911 frm_nxt += 4; 2912 } 2913 } 2914 return frm_nxt < frm_end ? codecvt_base::partial : codecvt_base::ok; 2915 } 2916 2917 static 2918 int 2919 utf16le_to_ucs4_length(const uint8_t* frm, const uint8_t* frm_end, 2920 size_t mx, unsigned long Maxcode = 0x10FFFF, 2921 codecvt_mode mode = codecvt_mode(0)) 2922 { 2923 const uint8_t* frm_nxt = frm; 2924 if (mode & consume_header) 2925 { 2926 if (frm_end-frm_nxt >= 2 && frm_nxt[0] == 0xFF && frm_nxt[1] == 0xFE) 2927 frm_nxt += 2; 2928 } 2929 for (size_t nchar32_t = 0; frm_nxt < frm_end - 1 && nchar32_t < mx; ++nchar32_t) 2930 { 2931 uint16_t c1 = static_cast<uint16_t>(frm_nxt[1] << 8 | frm_nxt[0]); 2932 if ((c1 & 0xFC00) == 0xDC00) 2933 break; 2934 if ((c1 & 0xFC00) != 0xD800) 2935 { 2936 if (c1 > Maxcode) 2937 break; 2938 frm_nxt += 2; 2939 } 2940 else 2941 { 2942 if (frm_end-frm_nxt < 4) 2943 break; 2944 uint16_t c2 = static_cast<uint16_t>(frm_nxt[3] << 8 | frm_nxt[2]); 2945 if ((c2 & 0xFC00) != 0xDC00) 2946 break; 2947 uint32_t t = static_cast<uint32_t>( 2948 ((((c1 & 0x03C0) >> 6) + 1) << 16) 2949 | ((c1 & 0x003F) << 10) 2950 | (c2 & 0x03FF)); 2951 if (t > Maxcode) 2952 break; 2953 frm_nxt += 4; 2954 } 2955 } 2956 return static_cast<int>(frm_nxt - frm); 2957 } 2958 2959 static 2960 codecvt_base::result 2961 ucs2_to_utf16be(const uint16_t* frm, const uint16_t* frm_end, const uint16_t*& frm_nxt, 2962 uint8_t* to, uint8_t* to_end, uint8_t*& to_nxt, 2963 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2964 { 2965 frm_nxt = frm; 2966 to_nxt = to; 2967 if (mode & generate_header) 2968 { 2969 if (to_end-to_nxt < 2) 2970 return codecvt_base::partial; 2971 *to_nxt++ = static_cast<uint8_t>(0xFE); 2972 *to_nxt++ = static_cast<uint8_t>(0xFF); 2973 } 2974 for (; frm_nxt < frm_end; ++frm_nxt) 2975 { 2976 uint16_t wc = *frm_nxt; 2977 if ((wc & 0xF800) == 0xD800 || wc > Maxcode) 2978 return codecvt_base::error; 2979 if (to_end-to_nxt < 2) 2980 return codecvt_base::partial; 2981 *to_nxt++ = static_cast<uint8_t>(wc >> 8); 2982 *to_nxt++ = static_cast<uint8_t>(wc); 2983 } 2984 return codecvt_base::ok; 2985 } 2986 2987 static 2988 codecvt_base::result 2989 utf16be_to_ucs2(const uint8_t* frm, const uint8_t* frm_end, const uint8_t*& frm_nxt, 2990 uint16_t* to, uint16_t* to_end, uint16_t*& to_nxt, 2991 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 2992 { 2993 frm_nxt = frm; 2994 to_nxt = to; 2995 if (mode & consume_header) 2996 { 2997 if (frm_end-frm_nxt >= 2 && frm_nxt[0] == 0xFE && frm_nxt[1] == 0xFF) 2998 frm_nxt += 2; 2999 } 3000 for (; frm_nxt < frm_end - 1 && to_nxt < to_end; ++to_nxt) 3001 { 3002 uint16_t c1 = static_cast<uint16_t>(frm_nxt[0] << 8 | frm_nxt[1]); 3003 if ((c1 & 0xF800) == 0xD800 || c1 > Maxcode) 3004 return codecvt_base::error; 3005 *to_nxt = c1; 3006 frm_nxt += 2; 3007 } 3008 return frm_nxt < frm_end ? codecvt_base::partial : codecvt_base::ok; 3009 } 3010 3011 static 3012 int 3013 utf16be_to_ucs2_length(const uint8_t* frm, const uint8_t* frm_end, 3014 size_t mx, unsigned long Maxcode = 0x10FFFF, 3015 codecvt_mode mode = codecvt_mode(0)) 3016 { 3017 const uint8_t* frm_nxt = frm; 3018 if (mode & consume_header) 3019 { 3020 if (frm_end-frm_nxt >= 2 && frm_nxt[0] == 0xFE && frm_nxt[1] == 0xFF) 3021 frm_nxt += 2; 3022 } 3023 for (size_t nchar16_t = 0; frm_nxt < frm_end - 1 && nchar16_t < mx; ++nchar16_t) 3024 { 3025 uint16_t c1 = static_cast<uint16_t>(frm_nxt[0] << 8 | frm_nxt[1]); 3026 if ((c1 & 0xF800) == 0xD800 || c1 > Maxcode) 3027 break; 3028 frm_nxt += 2; 3029 } 3030 return static_cast<int>(frm_nxt - frm); 3031 } 3032 3033 static 3034 codecvt_base::result 3035 ucs2_to_utf16le(const uint16_t* frm, const uint16_t* frm_end, const uint16_t*& frm_nxt, 3036 uint8_t* to, uint8_t* to_end, uint8_t*& to_nxt, 3037 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 3038 { 3039 frm_nxt = frm; 3040 to_nxt = to; 3041 if (mode & generate_header) 3042 { 3043 if (to_end-to_nxt < 2) 3044 return codecvt_base::partial; 3045 *to_nxt++ = static_cast<uint8_t>(0xFF); 3046 *to_nxt++ = static_cast<uint8_t>(0xFE); 3047 } 3048 for (; frm_nxt < frm_end; ++frm_nxt) 3049 { 3050 uint16_t wc = *frm_nxt; 3051 if ((wc & 0xF800) == 0xD800 || wc > Maxcode) 3052 return codecvt_base::error; 3053 if (to_end-to_nxt < 2) 3054 return codecvt_base::partial; 3055 *to_nxt++ = static_cast<uint8_t>(wc); 3056 *to_nxt++ = static_cast<uint8_t>(wc >> 8); 3057 } 3058 return codecvt_base::ok; 3059 } 3060 3061 static 3062 codecvt_base::result 3063 utf16le_to_ucs2(const uint8_t* frm, const uint8_t* frm_end, const uint8_t*& frm_nxt, 3064 uint16_t* to, uint16_t* to_end, uint16_t*& to_nxt, 3065 unsigned long Maxcode = 0x10FFFF, codecvt_mode mode = codecvt_mode(0)) 3066 { 3067 frm_nxt = frm; 3068 to_nxt = to; 3069 if (mode & consume_header) 3070 { 3071 if (frm_end-frm_nxt >= 2 && frm_nxt[0] == 0xFF && frm_nxt[1] == 0xFE) 3072 frm_nxt += 2; 3073 } 3074 for (; frm_nxt < frm_end - 1 && to_nxt < to_end; ++to_nxt) 3075 { 3076 uint16_t c1 = static_cast<uint16_t>(frm_nxt[1] << 8 | frm_nxt[0]); 3077 if ((c1 & 0xF800) == 0xD800 || c1 > Maxcode) 3078 return codecvt_base::error; 3079 *to_nxt = c1; 3080 frm_nxt += 2; 3081 } 3082 return frm_nxt < frm_end ? codecvt_base::partial : codecvt_base::ok; 3083 } 3084 3085 static 3086 int 3087 utf16le_to_ucs2_length(const uint8_t* frm, const uint8_t* frm_end, 3088 size_t mx, unsigned long Maxcode = 0x10FFFF, 3089 codecvt_mode mode = codecvt_mode(0)) 3090 { 3091 const uint8_t* frm_nxt = frm; 3092 frm_nxt = frm; 3093 if (mode & consume_header) 3094 { 3095 if (frm_end-frm_nxt >= 2 && frm_nxt[0] == 0xFF && frm_nxt[1] == 0xFE) 3096 frm_nxt += 2; 3097 } 3098 for (size_t nchar16_t = 0; frm_nxt < frm_end - 1 && nchar16_t < mx; ++nchar16_t) 3099 { 3100 uint16_t c1 = static_cast<uint16_t>(frm_nxt[1] << 8 | frm_nxt[0]); 3101 if ((c1 & 0xF800) == 0xD800 || c1 > Maxcode) 3102 break; 3103 frm_nxt += 2; 3104 } 3105 return static_cast<int>(frm_nxt - frm); 3106 } 3107 3108 // template <> class codecvt<char16_t, char, mbstate_t> 3109 3110 locale::id codecvt<char16_t, char, mbstate_t>::id; 3111 3112 codecvt<char16_t, char, mbstate_t>::~codecvt() 3113 { 3114 } 3115 3116 codecvt<char16_t, char, mbstate_t>::result 3117 codecvt<char16_t, char, mbstate_t>::do_out(state_type&, 3118 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3119 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3120 { 3121 const uint16_t* _frm = reinterpret_cast<const uint16_t*>(frm); 3122 const uint16_t* _frm_end = reinterpret_cast<const uint16_t*>(frm_end); 3123 const uint16_t* _frm_nxt = _frm; 3124 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3125 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3126 uint8_t* _to_nxt = _to; 3127 result r = utf16_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt); 3128 frm_nxt = frm + (_frm_nxt - _frm); 3129 to_nxt = to + (_to_nxt - _to); 3130 return r; 3131 } 3132 3133 codecvt<char16_t, char, mbstate_t>::result 3134 codecvt<char16_t, char, mbstate_t>::do_in(state_type&, 3135 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3136 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3137 { 3138 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3139 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3140 const uint8_t* _frm_nxt = _frm; 3141 uint16_t* _to = reinterpret_cast<uint16_t*>(to); 3142 uint16_t* _to_end = reinterpret_cast<uint16_t*>(to_end); 3143 uint16_t* _to_nxt = _to; 3144 result r = utf8_to_utf16(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt); 3145 frm_nxt = frm + (_frm_nxt - _frm); 3146 to_nxt = to + (_to_nxt - _to); 3147 return r; 3148 } 3149 3150 codecvt<char16_t, char, mbstate_t>::result 3151 codecvt<char16_t, char, mbstate_t>::do_unshift(state_type&, 3152 extern_type* to, extern_type*, extern_type*& to_nxt) const 3153 { 3154 to_nxt = to; 3155 return noconv; 3156 } 3157 3158 int 3159 codecvt<char16_t, char, mbstate_t>::do_encoding() const _NOEXCEPT 3160 { 3161 return 0; 3162 } 3163 3164 bool 3165 codecvt<char16_t, char, mbstate_t>::do_always_noconv() const _NOEXCEPT 3166 { 3167 return false; 3168 } 3169 3170 int 3171 codecvt<char16_t, char, mbstate_t>::do_length(state_type&, 3172 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3173 { 3174 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3175 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3176 return utf8_to_utf16_length(_frm, _frm_end, mx); 3177 } 3178 3179 int 3180 codecvt<char16_t, char, mbstate_t>::do_max_length() const _NOEXCEPT 3181 { 3182 return 4; 3183 } 3184 3185 // template <> class codecvt<char32_t, char, mbstate_t> 3186 3187 locale::id codecvt<char32_t, char, mbstate_t>::id; 3188 3189 codecvt<char32_t, char, mbstate_t>::~codecvt() 3190 { 3191 } 3192 3193 codecvt<char32_t, char, mbstate_t>::result 3194 codecvt<char32_t, char, mbstate_t>::do_out(state_type&, 3195 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3196 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3197 { 3198 const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); 3199 const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); 3200 const uint32_t* _frm_nxt = _frm; 3201 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3202 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3203 uint8_t* _to_nxt = _to; 3204 result r = ucs4_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt); 3205 frm_nxt = frm + (_frm_nxt - _frm); 3206 to_nxt = to + (_to_nxt - _to); 3207 return r; 3208 } 3209 3210 codecvt<char32_t, char, mbstate_t>::result 3211 codecvt<char32_t, char, mbstate_t>::do_in(state_type&, 3212 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3213 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3214 { 3215 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3216 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3217 const uint8_t* _frm_nxt = _frm; 3218 uint32_t* _to = reinterpret_cast<uint32_t*>(to); 3219 uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); 3220 uint32_t* _to_nxt = _to; 3221 result r = utf8_to_ucs4(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt); 3222 frm_nxt = frm + (_frm_nxt - _frm); 3223 to_nxt = to + (_to_nxt - _to); 3224 return r; 3225 } 3226 3227 codecvt<char32_t, char, mbstate_t>::result 3228 codecvt<char32_t, char, mbstate_t>::do_unshift(state_type&, 3229 extern_type* to, extern_type*, extern_type*& to_nxt) const 3230 { 3231 to_nxt = to; 3232 return noconv; 3233 } 3234 3235 int 3236 codecvt<char32_t, char, mbstate_t>::do_encoding() const _NOEXCEPT 3237 { 3238 return 0; 3239 } 3240 3241 bool 3242 codecvt<char32_t, char, mbstate_t>::do_always_noconv() const _NOEXCEPT 3243 { 3244 return false; 3245 } 3246 3247 int 3248 codecvt<char32_t, char, mbstate_t>::do_length(state_type&, 3249 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3250 { 3251 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3252 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3253 return utf8_to_ucs4_length(_frm, _frm_end, mx); 3254 } 3255 3256 int 3257 codecvt<char32_t, char, mbstate_t>::do_max_length() const _NOEXCEPT 3258 { 3259 return 4; 3260 } 3261 3262 // __codecvt_utf8<wchar_t> 3263 3264 __codecvt_utf8<wchar_t>::result 3265 __codecvt_utf8<wchar_t>::do_out(state_type&, 3266 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3267 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3268 { 3269 #if defined(_LIBCPP_SHORT_WCHAR) 3270 const uint16_t* _frm = reinterpret_cast<const uint16_t*>(frm); 3271 const uint16_t* _frm_end = reinterpret_cast<const uint16_t*>(frm_end); 3272 const uint16_t* _frm_nxt = _frm; 3273 #else 3274 const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); 3275 const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); 3276 const uint32_t* _frm_nxt = _frm; 3277 #endif 3278 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3279 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3280 uint8_t* _to_nxt = _to; 3281 #if defined(_LIBCPP_SHORT_WCHAR) 3282 result r = ucs2_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3283 _Maxcode_, _Mode_); 3284 #else 3285 result r = ucs4_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3286 _Maxcode_, _Mode_); 3287 #endif 3288 frm_nxt = frm + (_frm_nxt - _frm); 3289 to_nxt = to + (_to_nxt - _to); 3290 return r; 3291 } 3292 3293 __codecvt_utf8<wchar_t>::result 3294 __codecvt_utf8<wchar_t>::do_in(state_type&, 3295 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3296 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3297 { 3298 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3299 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3300 const uint8_t* _frm_nxt = _frm; 3301 #if defined(_LIBCPP_SHORT_WCHAR) 3302 uint16_t* _to = reinterpret_cast<uint16_t*>(to); 3303 uint16_t* _to_end = reinterpret_cast<uint16_t*>(to_end); 3304 uint16_t* _to_nxt = _to; 3305 result r = utf8_to_ucs2(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3306 _Maxcode_, _Mode_); 3307 #else 3308 uint32_t* _to = reinterpret_cast<uint32_t*>(to); 3309 uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); 3310 uint32_t* _to_nxt = _to; 3311 result r = utf8_to_ucs4(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3312 _Maxcode_, _Mode_); 3313 #endif 3314 frm_nxt = frm + (_frm_nxt - _frm); 3315 to_nxt = to + (_to_nxt - _to); 3316 return r; 3317 } 3318 3319 __codecvt_utf8<wchar_t>::result 3320 __codecvt_utf8<wchar_t>::do_unshift(state_type&, 3321 extern_type* to, extern_type*, extern_type*& to_nxt) const 3322 { 3323 to_nxt = to; 3324 return noconv; 3325 } 3326 3327 int 3328 __codecvt_utf8<wchar_t>::do_encoding() const _NOEXCEPT 3329 { 3330 return 0; 3331 } 3332 3333 bool 3334 __codecvt_utf8<wchar_t>::do_always_noconv() const _NOEXCEPT 3335 { 3336 return false; 3337 } 3338 3339 int 3340 __codecvt_utf8<wchar_t>::do_length(state_type&, 3341 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3342 { 3343 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3344 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3345 return utf8_to_ucs4_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 3346 } 3347 3348 int 3349 __codecvt_utf8<wchar_t>::do_max_length() const _NOEXCEPT 3350 { 3351 if (_Mode_ & consume_header) 3352 return 7; 3353 return 4; 3354 } 3355 3356 // __codecvt_utf8<char16_t> 3357 3358 __codecvt_utf8<char16_t>::result 3359 __codecvt_utf8<char16_t>::do_out(state_type&, 3360 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3361 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3362 { 3363 const uint16_t* _frm = reinterpret_cast<const uint16_t*>(frm); 3364 const uint16_t* _frm_end = reinterpret_cast<const uint16_t*>(frm_end); 3365 const uint16_t* _frm_nxt = _frm; 3366 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3367 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3368 uint8_t* _to_nxt = _to; 3369 result r = ucs2_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3370 _Maxcode_, _Mode_); 3371 frm_nxt = frm + (_frm_nxt - _frm); 3372 to_nxt = to + (_to_nxt - _to); 3373 return r; 3374 } 3375 3376 __codecvt_utf8<char16_t>::result 3377 __codecvt_utf8<char16_t>::do_in(state_type&, 3378 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3379 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3380 { 3381 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3382 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3383 const uint8_t* _frm_nxt = _frm; 3384 uint16_t* _to = reinterpret_cast<uint16_t*>(to); 3385 uint16_t* _to_end = reinterpret_cast<uint16_t*>(to_end); 3386 uint16_t* _to_nxt = _to; 3387 result r = utf8_to_ucs2(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3388 _Maxcode_, _Mode_); 3389 frm_nxt = frm + (_frm_nxt - _frm); 3390 to_nxt = to + (_to_nxt - _to); 3391 return r; 3392 } 3393 3394 __codecvt_utf8<char16_t>::result 3395 __codecvt_utf8<char16_t>::do_unshift(state_type&, 3396 extern_type* to, extern_type*, extern_type*& to_nxt) const 3397 { 3398 to_nxt = to; 3399 return noconv; 3400 } 3401 3402 int 3403 __codecvt_utf8<char16_t>::do_encoding() const _NOEXCEPT 3404 { 3405 return 0; 3406 } 3407 3408 bool 3409 __codecvt_utf8<char16_t>::do_always_noconv() const _NOEXCEPT 3410 { 3411 return false; 3412 } 3413 3414 int 3415 __codecvt_utf8<char16_t>::do_length(state_type&, 3416 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3417 { 3418 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3419 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3420 return utf8_to_ucs2_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 3421 } 3422 3423 int 3424 __codecvt_utf8<char16_t>::do_max_length() const _NOEXCEPT 3425 { 3426 if (_Mode_ & consume_header) 3427 return 6; 3428 return 3; 3429 } 3430 3431 // __codecvt_utf8<char32_t> 3432 3433 __codecvt_utf8<char32_t>::result 3434 __codecvt_utf8<char32_t>::do_out(state_type&, 3435 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3436 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3437 { 3438 const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); 3439 const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); 3440 const uint32_t* _frm_nxt = _frm; 3441 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3442 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3443 uint8_t* _to_nxt = _to; 3444 result r = ucs4_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3445 _Maxcode_, _Mode_); 3446 frm_nxt = frm + (_frm_nxt - _frm); 3447 to_nxt = to + (_to_nxt - _to); 3448 return r; 3449 } 3450 3451 __codecvt_utf8<char32_t>::result 3452 __codecvt_utf8<char32_t>::do_in(state_type&, 3453 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3454 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3455 { 3456 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3457 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3458 const uint8_t* _frm_nxt = _frm; 3459 uint32_t* _to = reinterpret_cast<uint32_t*>(to); 3460 uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); 3461 uint32_t* _to_nxt = _to; 3462 result r = utf8_to_ucs4(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3463 _Maxcode_, _Mode_); 3464 frm_nxt = frm + (_frm_nxt - _frm); 3465 to_nxt = to + (_to_nxt - _to); 3466 return r; 3467 } 3468 3469 __codecvt_utf8<char32_t>::result 3470 __codecvt_utf8<char32_t>::do_unshift(state_type&, 3471 extern_type* to, extern_type*, extern_type*& to_nxt) const 3472 { 3473 to_nxt = to; 3474 return noconv; 3475 } 3476 3477 int 3478 __codecvt_utf8<char32_t>::do_encoding() const _NOEXCEPT 3479 { 3480 return 0; 3481 } 3482 3483 bool 3484 __codecvt_utf8<char32_t>::do_always_noconv() const _NOEXCEPT 3485 { 3486 return false; 3487 } 3488 3489 int 3490 __codecvt_utf8<char32_t>::do_length(state_type&, 3491 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3492 { 3493 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3494 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3495 return utf8_to_ucs4_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 3496 } 3497 3498 int 3499 __codecvt_utf8<char32_t>::do_max_length() const _NOEXCEPT 3500 { 3501 if (_Mode_ & consume_header) 3502 return 7; 3503 return 4; 3504 } 3505 3506 // __codecvt_utf16<wchar_t, false> 3507 3508 __codecvt_utf16<wchar_t, false>::result 3509 __codecvt_utf16<wchar_t, false>::do_out(state_type&, 3510 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3511 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3512 { 3513 const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); 3514 const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); 3515 const uint32_t* _frm_nxt = _frm; 3516 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3517 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3518 uint8_t* _to_nxt = _to; 3519 result r = ucs4_to_utf16be(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3520 _Maxcode_, _Mode_); 3521 frm_nxt = frm + (_frm_nxt - _frm); 3522 to_nxt = to + (_to_nxt - _to); 3523 return r; 3524 } 3525 3526 __codecvt_utf16<wchar_t, false>::result 3527 __codecvt_utf16<wchar_t, false>::do_in(state_type&, 3528 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3529 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3530 { 3531 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3532 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3533 const uint8_t* _frm_nxt = _frm; 3534 uint32_t* _to = reinterpret_cast<uint32_t*>(to); 3535 uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); 3536 uint32_t* _to_nxt = _to; 3537 result r = utf16be_to_ucs4(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3538 _Maxcode_, _Mode_); 3539 frm_nxt = frm + (_frm_nxt - _frm); 3540 to_nxt = to + (_to_nxt - _to); 3541 return r; 3542 } 3543 3544 __codecvt_utf16<wchar_t, false>::result 3545 __codecvt_utf16<wchar_t, false>::do_unshift(state_type&, 3546 extern_type* to, extern_type*, extern_type*& to_nxt) const 3547 { 3548 to_nxt = to; 3549 return noconv; 3550 } 3551 3552 int 3553 __codecvt_utf16<wchar_t, false>::do_encoding() const _NOEXCEPT 3554 { 3555 return 0; 3556 } 3557 3558 bool 3559 __codecvt_utf16<wchar_t, false>::do_always_noconv() const _NOEXCEPT 3560 { 3561 return false; 3562 } 3563 3564 int 3565 __codecvt_utf16<wchar_t, false>::do_length(state_type&, 3566 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3567 { 3568 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3569 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3570 return utf16be_to_ucs4_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 3571 } 3572 3573 int 3574 __codecvt_utf16<wchar_t, false>::do_max_length() const _NOEXCEPT 3575 { 3576 if (_Mode_ & consume_header) 3577 return 6; 3578 return 4; 3579 } 3580 3581 // __codecvt_utf16<wchar_t, true> 3582 3583 __codecvt_utf16<wchar_t, true>::result 3584 __codecvt_utf16<wchar_t, true>::do_out(state_type&, 3585 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3586 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3587 { 3588 const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); 3589 const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); 3590 const uint32_t* _frm_nxt = _frm; 3591 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3592 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3593 uint8_t* _to_nxt = _to; 3594 result r = ucs4_to_utf16le(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3595 _Maxcode_, _Mode_); 3596 frm_nxt = frm + (_frm_nxt - _frm); 3597 to_nxt = to + (_to_nxt - _to); 3598 return r; 3599 } 3600 3601 __codecvt_utf16<wchar_t, true>::result 3602 __codecvt_utf16<wchar_t, true>::do_in(state_type&, 3603 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3604 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3605 { 3606 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3607 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3608 const uint8_t* _frm_nxt = _frm; 3609 uint32_t* _to = reinterpret_cast<uint32_t*>(to); 3610 uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); 3611 uint32_t* _to_nxt = _to; 3612 result r = utf16le_to_ucs4(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3613 _Maxcode_, _Mode_); 3614 frm_nxt = frm + (_frm_nxt - _frm); 3615 to_nxt = to + (_to_nxt - _to); 3616 return r; 3617 } 3618 3619 __codecvt_utf16<wchar_t, true>::result 3620 __codecvt_utf16<wchar_t, true>::do_unshift(state_type&, 3621 extern_type* to, extern_type*, extern_type*& to_nxt) const 3622 { 3623 to_nxt = to; 3624 return noconv; 3625 } 3626 3627 int 3628 __codecvt_utf16<wchar_t, true>::do_encoding() const _NOEXCEPT 3629 { 3630 return 0; 3631 } 3632 3633 bool 3634 __codecvt_utf16<wchar_t, true>::do_always_noconv() const _NOEXCEPT 3635 { 3636 return false; 3637 } 3638 3639 int 3640 __codecvt_utf16<wchar_t, true>::do_length(state_type&, 3641 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3642 { 3643 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3644 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3645 return utf16le_to_ucs4_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 3646 } 3647 3648 int 3649 __codecvt_utf16<wchar_t, true>::do_max_length() const _NOEXCEPT 3650 { 3651 if (_Mode_ & consume_header) 3652 return 6; 3653 return 4; 3654 } 3655 3656 // __codecvt_utf16<char16_t, false> 3657 3658 __codecvt_utf16<char16_t, false>::result 3659 __codecvt_utf16<char16_t, false>::do_out(state_type&, 3660 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3661 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3662 { 3663 const uint16_t* _frm = reinterpret_cast<const uint16_t*>(frm); 3664 const uint16_t* _frm_end = reinterpret_cast<const uint16_t*>(frm_end); 3665 const uint16_t* _frm_nxt = _frm; 3666 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3667 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3668 uint8_t* _to_nxt = _to; 3669 result r = ucs2_to_utf16be(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3670 _Maxcode_, _Mode_); 3671 frm_nxt = frm + (_frm_nxt - _frm); 3672 to_nxt = to + (_to_nxt - _to); 3673 return r; 3674 } 3675 3676 __codecvt_utf16<char16_t, false>::result 3677 __codecvt_utf16<char16_t, false>::do_in(state_type&, 3678 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3679 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3680 { 3681 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3682 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3683 const uint8_t* _frm_nxt = _frm; 3684 uint16_t* _to = reinterpret_cast<uint16_t*>(to); 3685 uint16_t* _to_end = reinterpret_cast<uint16_t*>(to_end); 3686 uint16_t* _to_nxt = _to; 3687 result r = utf16be_to_ucs2(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3688 _Maxcode_, _Mode_); 3689 frm_nxt = frm + (_frm_nxt - _frm); 3690 to_nxt = to + (_to_nxt - _to); 3691 return r; 3692 } 3693 3694 __codecvt_utf16<char16_t, false>::result 3695 __codecvt_utf16<char16_t, false>::do_unshift(state_type&, 3696 extern_type* to, extern_type*, extern_type*& to_nxt) const 3697 { 3698 to_nxt = to; 3699 return noconv; 3700 } 3701 3702 int 3703 __codecvt_utf16<char16_t, false>::do_encoding() const _NOEXCEPT 3704 { 3705 return 0; 3706 } 3707 3708 bool 3709 __codecvt_utf16<char16_t, false>::do_always_noconv() const _NOEXCEPT 3710 { 3711 return false; 3712 } 3713 3714 int 3715 __codecvt_utf16<char16_t, false>::do_length(state_type&, 3716 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3717 { 3718 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3719 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3720 return utf16be_to_ucs2_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 3721 } 3722 3723 int 3724 __codecvt_utf16<char16_t, false>::do_max_length() const _NOEXCEPT 3725 { 3726 if (_Mode_ & consume_header) 3727 return 4; 3728 return 2; 3729 } 3730 3731 // __codecvt_utf16<char16_t, true> 3732 3733 __codecvt_utf16<char16_t, true>::result 3734 __codecvt_utf16<char16_t, true>::do_out(state_type&, 3735 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3736 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3737 { 3738 const uint16_t* _frm = reinterpret_cast<const uint16_t*>(frm); 3739 const uint16_t* _frm_end = reinterpret_cast<const uint16_t*>(frm_end); 3740 const uint16_t* _frm_nxt = _frm; 3741 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3742 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3743 uint8_t* _to_nxt = _to; 3744 result r = ucs2_to_utf16le(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3745 _Maxcode_, _Mode_); 3746 frm_nxt = frm + (_frm_nxt - _frm); 3747 to_nxt = to + (_to_nxt - _to); 3748 return r; 3749 } 3750 3751 __codecvt_utf16<char16_t, true>::result 3752 __codecvt_utf16<char16_t, true>::do_in(state_type&, 3753 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3754 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3755 { 3756 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3757 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3758 const uint8_t* _frm_nxt = _frm; 3759 uint16_t* _to = reinterpret_cast<uint16_t*>(to); 3760 uint16_t* _to_end = reinterpret_cast<uint16_t*>(to_end); 3761 uint16_t* _to_nxt = _to; 3762 result r = utf16le_to_ucs2(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3763 _Maxcode_, _Mode_); 3764 frm_nxt = frm + (_frm_nxt - _frm); 3765 to_nxt = to + (_to_nxt - _to); 3766 return r; 3767 } 3768 3769 __codecvt_utf16<char16_t, true>::result 3770 __codecvt_utf16<char16_t, true>::do_unshift(state_type&, 3771 extern_type* to, extern_type*, extern_type*& to_nxt) const 3772 { 3773 to_nxt = to; 3774 return noconv; 3775 } 3776 3777 int 3778 __codecvt_utf16<char16_t, true>::do_encoding() const _NOEXCEPT 3779 { 3780 return 0; 3781 } 3782 3783 bool 3784 __codecvt_utf16<char16_t, true>::do_always_noconv() const _NOEXCEPT 3785 { 3786 return false; 3787 } 3788 3789 int 3790 __codecvt_utf16<char16_t, true>::do_length(state_type&, 3791 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3792 { 3793 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3794 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3795 return utf16le_to_ucs2_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 3796 } 3797 3798 int 3799 __codecvt_utf16<char16_t, true>::do_max_length() const _NOEXCEPT 3800 { 3801 if (_Mode_ & consume_header) 3802 return 4; 3803 return 2; 3804 } 3805 3806 // __codecvt_utf16<char32_t, false> 3807 3808 __codecvt_utf16<char32_t, false>::result 3809 __codecvt_utf16<char32_t, false>::do_out(state_type&, 3810 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3811 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3812 { 3813 const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); 3814 const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); 3815 const uint32_t* _frm_nxt = _frm; 3816 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3817 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3818 uint8_t* _to_nxt = _to; 3819 result r = ucs4_to_utf16be(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3820 _Maxcode_, _Mode_); 3821 frm_nxt = frm + (_frm_nxt - _frm); 3822 to_nxt = to + (_to_nxt - _to); 3823 return r; 3824 } 3825 3826 __codecvt_utf16<char32_t, false>::result 3827 __codecvt_utf16<char32_t, false>::do_in(state_type&, 3828 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3829 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3830 { 3831 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3832 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3833 const uint8_t* _frm_nxt = _frm; 3834 uint32_t* _to = reinterpret_cast<uint32_t*>(to); 3835 uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); 3836 uint32_t* _to_nxt = _to; 3837 result r = utf16be_to_ucs4(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3838 _Maxcode_, _Mode_); 3839 frm_nxt = frm + (_frm_nxt - _frm); 3840 to_nxt = to + (_to_nxt - _to); 3841 return r; 3842 } 3843 3844 __codecvt_utf16<char32_t, false>::result 3845 __codecvt_utf16<char32_t, false>::do_unshift(state_type&, 3846 extern_type* to, extern_type*, extern_type*& to_nxt) const 3847 { 3848 to_nxt = to; 3849 return noconv; 3850 } 3851 3852 int 3853 __codecvt_utf16<char32_t, false>::do_encoding() const _NOEXCEPT 3854 { 3855 return 0; 3856 } 3857 3858 bool 3859 __codecvt_utf16<char32_t, false>::do_always_noconv() const _NOEXCEPT 3860 { 3861 return false; 3862 } 3863 3864 int 3865 __codecvt_utf16<char32_t, false>::do_length(state_type&, 3866 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3867 { 3868 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3869 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3870 return utf16be_to_ucs4_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 3871 } 3872 3873 int 3874 __codecvt_utf16<char32_t, false>::do_max_length() const _NOEXCEPT 3875 { 3876 if (_Mode_ & consume_header) 3877 return 6; 3878 return 4; 3879 } 3880 3881 // __codecvt_utf16<char32_t, true> 3882 3883 __codecvt_utf16<char32_t, true>::result 3884 __codecvt_utf16<char32_t, true>::do_out(state_type&, 3885 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3886 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3887 { 3888 const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); 3889 const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); 3890 const uint32_t* _frm_nxt = _frm; 3891 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3892 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3893 uint8_t* _to_nxt = _to; 3894 result r = ucs4_to_utf16le(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3895 _Maxcode_, _Mode_); 3896 frm_nxt = frm + (_frm_nxt - _frm); 3897 to_nxt = to + (_to_nxt - _to); 3898 return r; 3899 } 3900 3901 __codecvt_utf16<char32_t, true>::result 3902 __codecvt_utf16<char32_t, true>::do_in(state_type&, 3903 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3904 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3905 { 3906 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3907 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3908 const uint8_t* _frm_nxt = _frm; 3909 uint32_t* _to = reinterpret_cast<uint32_t*>(to); 3910 uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); 3911 uint32_t* _to_nxt = _to; 3912 result r = utf16le_to_ucs4(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3913 _Maxcode_, _Mode_); 3914 frm_nxt = frm + (_frm_nxt - _frm); 3915 to_nxt = to + (_to_nxt - _to); 3916 return r; 3917 } 3918 3919 __codecvt_utf16<char32_t, true>::result 3920 __codecvt_utf16<char32_t, true>::do_unshift(state_type&, 3921 extern_type* to, extern_type*, extern_type*& to_nxt) const 3922 { 3923 to_nxt = to; 3924 return noconv; 3925 } 3926 3927 int 3928 __codecvt_utf16<char32_t, true>::do_encoding() const _NOEXCEPT 3929 { 3930 return 0; 3931 } 3932 3933 bool 3934 __codecvt_utf16<char32_t, true>::do_always_noconv() const _NOEXCEPT 3935 { 3936 return false; 3937 } 3938 3939 int 3940 __codecvt_utf16<char32_t, true>::do_length(state_type&, 3941 const extern_type* frm, const extern_type* frm_end, size_t mx) const 3942 { 3943 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3944 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3945 return utf16le_to_ucs4_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 3946 } 3947 3948 int 3949 __codecvt_utf16<char32_t, true>::do_max_length() const _NOEXCEPT 3950 { 3951 if (_Mode_ & consume_header) 3952 return 6; 3953 return 4; 3954 } 3955 3956 // __codecvt_utf8_utf16<wchar_t> 3957 3958 __codecvt_utf8_utf16<wchar_t>::result 3959 __codecvt_utf8_utf16<wchar_t>::do_out(state_type&, 3960 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 3961 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 3962 { 3963 const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); 3964 const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); 3965 const uint32_t* _frm_nxt = _frm; 3966 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 3967 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 3968 uint8_t* _to_nxt = _to; 3969 result r = utf16_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3970 _Maxcode_, _Mode_); 3971 frm_nxt = frm + (_frm_nxt - _frm); 3972 to_nxt = to + (_to_nxt - _to); 3973 return r; 3974 } 3975 3976 __codecvt_utf8_utf16<wchar_t>::result 3977 __codecvt_utf8_utf16<wchar_t>::do_in(state_type&, 3978 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 3979 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 3980 { 3981 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 3982 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 3983 const uint8_t* _frm_nxt = _frm; 3984 uint32_t* _to = reinterpret_cast<uint32_t*>(to); 3985 uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); 3986 uint32_t* _to_nxt = _to; 3987 result r = utf8_to_utf16(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 3988 _Maxcode_, _Mode_); 3989 frm_nxt = frm + (_frm_nxt - _frm); 3990 to_nxt = to + (_to_nxt - _to); 3991 return r; 3992 } 3993 3994 __codecvt_utf8_utf16<wchar_t>::result 3995 __codecvt_utf8_utf16<wchar_t>::do_unshift(state_type&, 3996 extern_type* to, extern_type*, extern_type*& to_nxt) const 3997 { 3998 to_nxt = to; 3999 return noconv; 4000 } 4001 4002 int 4003 __codecvt_utf8_utf16<wchar_t>::do_encoding() const _NOEXCEPT 4004 { 4005 return 0; 4006 } 4007 4008 bool 4009 __codecvt_utf8_utf16<wchar_t>::do_always_noconv() const _NOEXCEPT 4010 { 4011 return false; 4012 } 4013 4014 int 4015 __codecvt_utf8_utf16<wchar_t>::do_length(state_type&, 4016 const extern_type* frm, const extern_type* frm_end, size_t mx) const 4017 { 4018 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 4019 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 4020 return utf8_to_utf16_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 4021 } 4022 4023 int 4024 __codecvt_utf8_utf16<wchar_t>::do_max_length() const _NOEXCEPT 4025 { 4026 if (_Mode_ & consume_header) 4027 return 7; 4028 return 4; 4029 } 4030 4031 // __codecvt_utf8_utf16<char16_t> 4032 4033 __codecvt_utf8_utf16<char16_t>::result 4034 __codecvt_utf8_utf16<char16_t>::do_out(state_type&, 4035 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 4036 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 4037 { 4038 const uint16_t* _frm = reinterpret_cast<const uint16_t*>(frm); 4039 const uint16_t* _frm_end = reinterpret_cast<const uint16_t*>(frm_end); 4040 const uint16_t* _frm_nxt = _frm; 4041 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 4042 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 4043 uint8_t* _to_nxt = _to; 4044 result r = utf16_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 4045 _Maxcode_, _Mode_); 4046 frm_nxt = frm + (_frm_nxt - _frm); 4047 to_nxt = to + (_to_nxt - _to); 4048 return r; 4049 } 4050 4051 __codecvt_utf8_utf16<char16_t>::result 4052 __codecvt_utf8_utf16<char16_t>::do_in(state_type&, 4053 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 4054 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 4055 { 4056 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 4057 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 4058 const uint8_t* _frm_nxt = _frm; 4059 uint16_t* _to = reinterpret_cast<uint16_t*>(to); 4060 uint16_t* _to_end = reinterpret_cast<uint16_t*>(to_end); 4061 uint16_t* _to_nxt = _to; 4062 result r = utf8_to_utf16(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 4063 _Maxcode_, _Mode_); 4064 frm_nxt = frm + (_frm_nxt - _frm); 4065 to_nxt = to + (_to_nxt - _to); 4066 return r; 4067 } 4068 4069 __codecvt_utf8_utf16<char16_t>::result 4070 __codecvt_utf8_utf16<char16_t>::do_unshift(state_type&, 4071 extern_type* to, extern_type*, extern_type*& to_nxt) const 4072 { 4073 to_nxt = to; 4074 return noconv; 4075 } 4076 4077 int 4078 __codecvt_utf8_utf16<char16_t>::do_encoding() const _NOEXCEPT 4079 { 4080 return 0; 4081 } 4082 4083 bool 4084 __codecvt_utf8_utf16<char16_t>::do_always_noconv() const _NOEXCEPT 4085 { 4086 return false; 4087 } 4088 4089 int 4090 __codecvt_utf8_utf16<char16_t>::do_length(state_type&, 4091 const extern_type* frm, const extern_type* frm_end, size_t mx) const 4092 { 4093 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 4094 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 4095 return utf8_to_utf16_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 4096 } 4097 4098 int 4099 __codecvt_utf8_utf16<char16_t>::do_max_length() const _NOEXCEPT 4100 { 4101 if (_Mode_ & consume_header) 4102 return 7; 4103 return 4; 4104 } 4105 4106 // __codecvt_utf8_utf16<char32_t> 4107 4108 __codecvt_utf8_utf16<char32_t>::result 4109 __codecvt_utf8_utf16<char32_t>::do_out(state_type&, 4110 const intern_type* frm, const intern_type* frm_end, const intern_type*& frm_nxt, 4111 extern_type* to, extern_type* to_end, extern_type*& to_nxt) const 4112 { 4113 const uint32_t* _frm = reinterpret_cast<const uint32_t*>(frm); 4114 const uint32_t* _frm_end = reinterpret_cast<const uint32_t*>(frm_end); 4115 const uint32_t* _frm_nxt = _frm; 4116 uint8_t* _to = reinterpret_cast<uint8_t*>(to); 4117 uint8_t* _to_end = reinterpret_cast<uint8_t*>(to_end); 4118 uint8_t* _to_nxt = _to; 4119 result r = utf16_to_utf8(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 4120 _Maxcode_, _Mode_); 4121 frm_nxt = frm + (_frm_nxt - _frm); 4122 to_nxt = to + (_to_nxt - _to); 4123 return r; 4124 } 4125 4126 __codecvt_utf8_utf16<char32_t>::result 4127 __codecvt_utf8_utf16<char32_t>::do_in(state_type&, 4128 const extern_type* frm, const extern_type* frm_end, const extern_type*& frm_nxt, 4129 intern_type* to, intern_type* to_end, intern_type*& to_nxt) const 4130 { 4131 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 4132 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 4133 const uint8_t* _frm_nxt = _frm; 4134 uint32_t* _to = reinterpret_cast<uint32_t*>(to); 4135 uint32_t* _to_end = reinterpret_cast<uint32_t*>(to_end); 4136 uint32_t* _to_nxt = _to; 4137 result r = utf8_to_utf16(_frm, _frm_end, _frm_nxt, _to, _to_end, _to_nxt, 4138 _Maxcode_, _Mode_); 4139 frm_nxt = frm + (_frm_nxt - _frm); 4140 to_nxt = to + (_to_nxt - _to); 4141 return r; 4142 } 4143 4144 __codecvt_utf8_utf16<char32_t>::result 4145 __codecvt_utf8_utf16<char32_t>::do_unshift(state_type&, 4146 extern_type* to, extern_type*, extern_type*& to_nxt) const 4147 { 4148 to_nxt = to; 4149 return noconv; 4150 } 4151 4152 int 4153 __codecvt_utf8_utf16<char32_t>::do_encoding() const _NOEXCEPT 4154 { 4155 return 0; 4156 } 4157 4158 bool 4159 __codecvt_utf8_utf16<char32_t>::do_always_noconv() const _NOEXCEPT 4160 { 4161 return false; 4162 } 4163 4164 int 4165 __codecvt_utf8_utf16<char32_t>::do_length(state_type&, 4166 const extern_type* frm, const extern_type* frm_end, size_t mx) const 4167 { 4168 const uint8_t* _frm = reinterpret_cast<const uint8_t*>(frm); 4169 const uint8_t* _frm_end = reinterpret_cast<const uint8_t*>(frm_end); 4170 return utf8_to_utf16_length(_frm, _frm_end, mx, _Maxcode_, _Mode_); 4171 } 4172 4173 int 4174 __codecvt_utf8_utf16<char32_t>::do_max_length() const _NOEXCEPT 4175 { 4176 if (_Mode_ & consume_header) 4177 return 7; 4178 return 4; 4179 } 4180 4181 // __narrow_to_utf8<16> 4182 4183 __narrow_to_utf8<16>::~__narrow_to_utf8() 4184 { 4185 } 4186 4187 // __narrow_to_utf8<32> 4188 4189 __narrow_to_utf8<32>::~__narrow_to_utf8() 4190 { 4191 } 4192 4193 // __widen_from_utf8<16> 4194 4195 __widen_from_utf8<16>::~__widen_from_utf8() 4196 { 4197 } 4198 4199 // __widen_from_utf8<32> 4200 4201 __widen_from_utf8<32>::~__widen_from_utf8() 4202 { 4203 } 4204 4205 4206 static bool checked_string_to_wchar_convert(wchar_t& dest, 4207 const char* ptr, 4208 locale_t loc) { 4209 if (*ptr == '\0') 4210 return false; 4211 mbstate_t mb = {}; 4212 wchar_t out; 4213 size_t ret = __libcpp_mbrtowc_l(&out, ptr, strlen(ptr), &mb, loc); 4214 if (ret == static_cast<size_t>(-1) || ret == static_cast<size_t>(-2)) { 4215 return false; 4216 } 4217 dest = out; 4218 return true; 4219 } 4220 4221 static bool checked_string_to_char_convert(char& dest, 4222 const char* ptr, 4223 locale_t __loc) { 4224 if (*ptr == '\0') 4225 return false; 4226 if (!ptr[1]) { 4227 dest = *ptr; 4228 return true; 4229 } 4230 // First convert the MBS into a wide char then attempt to narrow it using 4231 // wctob_l. 4232 wchar_t wout; 4233 if (!checked_string_to_wchar_convert(wout, ptr, __loc)) 4234 return false; 4235 int res; 4236 if ((res = __libcpp_wctob_l(wout, __loc)) != char_traits<char>::eof()) { 4237 dest = res; 4238 return true; 4239 } 4240 // FIXME: Work around specific multibyte sequences that we can reasonable 4241 // translate into a different single byte. 4242 switch (wout) { 4243 case L'\u00A0': // non-breaking space 4244 dest = ' '; 4245 return true; 4246 default: 4247 return false; 4248 } 4249 _LIBCPP_UNREACHABLE(); 4250 } 4251 4252 4253 // numpunct<char> && numpunct<wchar_t> 4254 4255 locale::id numpunct< char >::id; 4256 locale::id numpunct<wchar_t>::id; 4257 4258 numpunct<char>::numpunct(size_t refs) 4259 : locale::facet(refs), 4260 __decimal_point_('.'), 4261 __thousands_sep_(',') 4262 { 4263 } 4264 4265 numpunct<wchar_t>::numpunct(size_t refs) 4266 : locale::facet(refs), 4267 __decimal_point_(L'.'), 4268 __thousands_sep_(L',') 4269 { 4270 } 4271 4272 numpunct<char>::~numpunct() 4273 { 4274 } 4275 4276 numpunct<wchar_t>::~numpunct() 4277 { 4278 } 4279 4280 char numpunct< char >::do_decimal_point() const {return __decimal_point_;} 4281 wchar_t numpunct<wchar_t>::do_decimal_point() const {return __decimal_point_;} 4282 4283 char numpunct< char >::do_thousands_sep() const {return __thousands_sep_;} 4284 wchar_t numpunct<wchar_t>::do_thousands_sep() const {return __thousands_sep_;} 4285 4286 string numpunct< char >::do_grouping() const {return __grouping_;} 4287 string numpunct<wchar_t>::do_grouping() const {return __grouping_;} 4288 4289 string numpunct< char >::do_truename() const {return "true";} 4290 wstring numpunct<wchar_t>::do_truename() const {return L"true";} 4291 4292 string numpunct< char >::do_falsename() const {return "false";} 4293 wstring numpunct<wchar_t>::do_falsename() const {return L"false";} 4294 4295 // numpunct_byname<char> 4296 4297 numpunct_byname<char>::numpunct_byname(const char* nm, size_t refs) 4298 : numpunct<char>(refs) 4299 { 4300 __init(nm); 4301 } 4302 4303 numpunct_byname<char>::numpunct_byname(const string& nm, size_t refs) 4304 : numpunct<char>(refs) 4305 { 4306 __init(nm.c_str()); 4307 } 4308 4309 numpunct_byname<char>::~numpunct_byname() 4310 { 4311 } 4312 4313 void 4314 numpunct_byname<char>::__init(const char* nm) 4315 { 4316 if (strcmp(nm, "C") != 0) 4317 { 4318 __libcpp_unique_locale loc(nm); 4319 if (!loc) 4320 __throw_runtime_error("numpunct_byname<char>::numpunct_byname" 4321 " failed to construct for " + string(nm)); 4322 4323 lconv* lc = __libcpp_localeconv_l(loc.get()); 4324 checked_string_to_char_convert(__decimal_point_, lc->decimal_point, 4325 loc.get()); 4326 checked_string_to_char_convert(__thousands_sep_, lc->thousands_sep, 4327 loc.get()); 4328 __grouping_ = lc->grouping; 4329 // localization for truename and falsename is not available 4330 } 4331 } 4332 4333 // numpunct_byname<wchar_t> 4334 4335 numpunct_byname<wchar_t>::numpunct_byname(const char* nm, size_t refs) 4336 : numpunct<wchar_t>(refs) 4337 { 4338 __init(nm); 4339 } 4340 4341 numpunct_byname<wchar_t>::numpunct_byname(const string& nm, size_t refs) 4342 : numpunct<wchar_t>(refs) 4343 { 4344 __init(nm.c_str()); 4345 } 4346 4347 numpunct_byname<wchar_t>::~numpunct_byname() 4348 { 4349 } 4350 4351 void 4352 numpunct_byname<wchar_t>::__init(const char* nm) 4353 { 4354 if (strcmp(nm, "C") != 0) 4355 { 4356 __libcpp_unique_locale loc(nm); 4357 if (!loc) 4358 __throw_runtime_error("numpunct_byname<wchar_t>::numpunct_byname" 4359 " failed to construct for " + string(nm)); 4360 4361 lconv* lc = __libcpp_localeconv_l(loc.get()); 4362 checked_string_to_wchar_convert(__decimal_point_, lc->decimal_point, 4363 loc.get()); 4364 checked_string_to_wchar_convert(__thousands_sep_, lc->thousands_sep, 4365 loc.get()); 4366 __grouping_ = lc->grouping; 4367 // localization for truename and falsename is not available 4368 } 4369 } 4370 4371 // num_get helpers 4372 4373 int 4374 __num_get_base::__get_base(ios_base& iob) 4375 { 4376 ios_base::fmtflags __basefield = iob.flags() & ios_base::basefield; 4377 if (__basefield == ios_base::oct) 4378 return 8; 4379 else if (__basefield == ios_base::hex) 4380 return 16; 4381 else if (__basefield == 0) 4382 return 0; 4383 return 10; 4384 } 4385 4386 const char __num_get_base::__src[33] = "0123456789abcdefABCDEFxX+-pPiInN"; 4387 4388 void 4389 __check_grouping(const string& __grouping, unsigned* __g, unsigned* __g_end, 4390 ios_base::iostate& __err) 4391 { 4392 if (__grouping.size() != 0) 4393 { 4394 reverse(__g, __g_end); 4395 const char* __ig = __grouping.data(); 4396 const char* __eg = __ig + __grouping.size(); 4397 for (unsigned* __r = __g; __r < __g_end-1; ++__r) 4398 { 4399 if (0 < *__ig && *__ig < numeric_limits<char>::max()) 4400 { 4401 if (static_cast<unsigned>(*__ig) != *__r) 4402 { 4403 __err = ios_base::failbit; 4404 return; 4405 } 4406 } 4407 if (__eg - __ig > 1) 4408 ++__ig; 4409 } 4410 if (0 < *__ig && *__ig < numeric_limits<char>::max()) 4411 { 4412 if (static_cast<unsigned>(*__ig) < __g_end[-1] || __g_end[-1] == 0) 4413 __err = ios_base::failbit; 4414 } 4415 } 4416 } 4417 4418 void 4419 __num_put_base::__format_int(char* __fmtp, const char* __len, bool __signd, 4420 ios_base::fmtflags __flags) 4421 { 4422 if (__flags & ios_base::showpos) 4423 *__fmtp++ = '+'; 4424 if (__flags & ios_base::showbase) 4425 *__fmtp++ = '#'; 4426 while(*__len) 4427 *__fmtp++ = *__len++; 4428 if ((__flags & ios_base::basefield) == ios_base::oct) 4429 *__fmtp = 'o'; 4430 else if ((__flags & ios_base::basefield) == ios_base::hex) 4431 { 4432 if (__flags & ios_base::uppercase) 4433 *__fmtp = 'X'; 4434 else 4435 *__fmtp = 'x'; 4436 } 4437 else if (__signd) 4438 *__fmtp = 'd'; 4439 else 4440 *__fmtp = 'u'; 4441 } 4442 4443 bool 4444 __num_put_base::__format_float(char* __fmtp, const char* __len, 4445 ios_base::fmtflags __flags) 4446 { 4447 bool specify_precision = true; 4448 if (__flags & ios_base::showpos) 4449 *__fmtp++ = '+'; 4450 if (__flags & ios_base::showpoint) 4451 *__fmtp++ = '#'; 4452 ios_base::fmtflags floatfield = __flags & ios_base::floatfield; 4453 bool uppercase = (__flags & ios_base::uppercase) != 0; 4454 if (floatfield == (ios_base::fixed | ios_base::scientific)) 4455 specify_precision = false; 4456 else 4457 { 4458 *__fmtp++ = '.'; 4459 *__fmtp++ = '*'; 4460 } 4461 while(*__len) 4462 *__fmtp++ = *__len++; 4463 if (floatfield == ios_base::fixed) 4464 { 4465 if (uppercase) 4466 *__fmtp = 'F'; 4467 else 4468 *__fmtp = 'f'; 4469 } 4470 else if (floatfield == ios_base::scientific) 4471 { 4472 if (uppercase) 4473 *__fmtp = 'E'; 4474 else 4475 *__fmtp = 'e'; 4476 } 4477 else if (floatfield == (ios_base::fixed | ios_base::scientific)) 4478 { 4479 if (uppercase) 4480 *__fmtp = 'A'; 4481 else 4482 *__fmtp = 'a'; 4483 } 4484 else 4485 { 4486 if (uppercase) 4487 *__fmtp = 'G'; 4488 else 4489 *__fmtp = 'g'; 4490 } 4491 return specify_precision; 4492 } 4493 4494 char* 4495 __num_put_base::__identify_padding(char* __nb, char* __ne, 4496 const ios_base& __iob) 4497 { 4498 switch (__iob.flags() & ios_base::adjustfield) 4499 { 4500 case ios_base::internal: 4501 if (__nb[0] == '-' || __nb[0] == '+') 4502 return __nb+1; 4503 if (__ne - __nb >= 2 && __nb[0] == '0' 4504 && (__nb[1] == 'x' || __nb[1] == 'X')) 4505 return __nb+2; 4506 break; 4507 case ios_base::left: 4508 return __ne; 4509 case ios_base::right: 4510 default: 4511 break; 4512 } 4513 return __nb; 4514 } 4515 4516 // time_get 4517 4518 static 4519 string* 4520 init_weeks() 4521 { 4522 static string weeks[14]; 4523 weeks[0] = "Sunday"; 4524 weeks[1] = "Monday"; 4525 weeks[2] = "Tuesday"; 4526 weeks[3] = "Wednesday"; 4527 weeks[4] = "Thursday"; 4528 weeks[5] = "Friday"; 4529 weeks[6] = "Saturday"; 4530 weeks[7] = "Sun"; 4531 weeks[8] = "Mon"; 4532 weeks[9] = "Tue"; 4533 weeks[10] = "Wed"; 4534 weeks[11] = "Thu"; 4535 weeks[12] = "Fri"; 4536 weeks[13] = "Sat"; 4537 return weeks; 4538 } 4539 4540 static 4541 wstring* 4542 init_wweeks() 4543 { 4544 static wstring weeks[14]; 4545 weeks[0] = L"Sunday"; 4546 weeks[1] = L"Monday"; 4547 weeks[2] = L"Tuesday"; 4548 weeks[3] = L"Wednesday"; 4549 weeks[4] = L"Thursday"; 4550 weeks[5] = L"Friday"; 4551 weeks[6] = L"Saturday"; 4552 weeks[7] = L"Sun"; 4553 weeks[8] = L"Mon"; 4554 weeks[9] = L"Tue"; 4555 weeks[10] = L"Wed"; 4556 weeks[11] = L"Thu"; 4557 weeks[12] = L"Fri"; 4558 weeks[13] = L"Sat"; 4559 return weeks; 4560 } 4561 4562 template <> 4563 const string* 4564 __time_get_c_storage<char>::__weeks() const 4565 { 4566 static const string* weeks = init_weeks(); 4567 return weeks; 4568 } 4569 4570 template <> 4571 const wstring* 4572 __time_get_c_storage<wchar_t>::__weeks() const 4573 { 4574 static const wstring* weeks = init_wweeks(); 4575 return weeks; 4576 } 4577 4578 static 4579 string* 4580 init_months() 4581 { 4582 static string months[24]; 4583 months[0] = "January"; 4584 months[1] = "February"; 4585 months[2] = "March"; 4586 months[3] = "April"; 4587 months[4] = "May"; 4588 months[5] = "June"; 4589 months[6] = "July"; 4590 months[7] = "August"; 4591 months[8] = "September"; 4592 months[9] = "October"; 4593 months[10] = "November"; 4594 months[11] = "December"; 4595 months[12] = "Jan"; 4596 months[13] = "Feb"; 4597 months[14] = "Mar"; 4598 months[15] = "Apr"; 4599 months[16] = "May"; 4600 months[17] = "Jun"; 4601 months[18] = "Jul"; 4602 months[19] = "Aug"; 4603 months[20] = "Sep"; 4604 months[21] = "Oct"; 4605 months[22] = "Nov"; 4606 months[23] = "Dec"; 4607 return months; 4608 } 4609 4610 static 4611 wstring* 4612 init_wmonths() 4613 { 4614 static wstring months[24]; 4615 months[0] = L"January"; 4616 months[1] = L"February"; 4617 months[2] = L"March"; 4618 months[3] = L"April"; 4619 months[4] = L"May"; 4620 months[5] = L"June"; 4621 months[6] = L"July"; 4622 months[7] = L"August"; 4623 months[8] = L"September"; 4624 months[9] = L"October"; 4625 months[10] = L"November"; 4626 months[11] = L"December"; 4627 months[12] = L"Jan"; 4628 months[13] = L"Feb"; 4629 months[14] = L"Mar"; 4630 months[15] = L"Apr"; 4631 months[16] = L"May"; 4632 months[17] = L"Jun"; 4633 months[18] = L"Jul"; 4634 months[19] = L"Aug"; 4635 months[20] = L"Sep"; 4636 months[21] = L"Oct"; 4637 months[22] = L"Nov"; 4638 months[23] = L"Dec"; 4639 return months; 4640 } 4641 4642 template <> 4643 const string* 4644 __time_get_c_storage<char>::__months() const 4645 { 4646 static const string* months = init_months(); 4647 return months; 4648 } 4649 4650 template <> 4651 const wstring* 4652 __time_get_c_storage<wchar_t>::__months() const 4653 { 4654 static const wstring* months = init_wmonths(); 4655 return months; 4656 } 4657 4658 static 4659 string* 4660 init_am_pm() 4661 { 4662 static string am_pm[2]; 4663 am_pm[0] = "AM"; 4664 am_pm[1] = "PM"; 4665 return am_pm; 4666 } 4667 4668 static 4669 wstring* 4670 init_wam_pm() 4671 { 4672 static wstring am_pm[2]; 4673 am_pm[0] = L"AM"; 4674 am_pm[1] = L"PM"; 4675 return am_pm; 4676 } 4677 4678 template <> 4679 const string* 4680 __time_get_c_storage<char>::__am_pm() const 4681 { 4682 static const string* am_pm = init_am_pm(); 4683 return am_pm; 4684 } 4685 4686 template <> 4687 const wstring* 4688 __time_get_c_storage<wchar_t>::__am_pm() const 4689 { 4690 static const wstring* am_pm = init_wam_pm(); 4691 return am_pm; 4692 } 4693 4694 template <> 4695 const string& 4696 __time_get_c_storage<char>::__x() const 4697 { 4698 static string s("%m/%d/%y"); 4699 return s; 4700 } 4701 4702 template <> 4703 const wstring& 4704 __time_get_c_storage<wchar_t>::__x() const 4705 { 4706 static wstring s(L"%m/%d/%y"); 4707 return s; 4708 } 4709 4710 template <> 4711 const string& 4712 __time_get_c_storage<char>::__X() const 4713 { 4714 static string s("%H:%M:%S"); 4715 return s; 4716 } 4717 4718 template <> 4719 const wstring& 4720 __time_get_c_storage<wchar_t>::__X() const 4721 { 4722 static wstring s(L"%H:%M:%S"); 4723 return s; 4724 } 4725 4726 template <> 4727 const string& 4728 __time_get_c_storage<char>::__c() const 4729 { 4730 static string s("%a %b %d %H:%M:%S %Y"); 4731 return s; 4732 } 4733 4734 template <> 4735 const wstring& 4736 __time_get_c_storage<wchar_t>::__c() const 4737 { 4738 static wstring s(L"%a %b %d %H:%M:%S %Y"); 4739 return s; 4740 } 4741 4742 template <> 4743 const string& 4744 __time_get_c_storage<char>::__r() const 4745 { 4746 static string s("%I:%M:%S %p"); 4747 return s; 4748 } 4749 4750 template <> 4751 const wstring& 4752 __time_get_c_storage<wchar_t>::__r() const 4753 { 4754 static wstring s(L"%I:%M:%S %p"); 4755 return s; 4756 } 4757 4758 // time_get_byname 4759 4760 __time_get::__time_get(const char* nm) 4761 : __loc_(newlocale(LC_ALL_MASK, nm, 0)) 4762 { 4763 if (__loc_ == 0) 4764 __throw_runtime_error("time_get_byname" 4765 " failed to construct for " + string(nm)); 4766 } 4767 4768 __time_get::__time_get(const string& nm) 4769 : __loc_(newlocale(LC_ALL_MASK, nm.c_str(), 0)) 4770 { 4771 if (__loc_ == 0) 4772 __throw_runtime_error("time_get_byname" 4773 " failed to construct for " + nm); 4774 } 4775 4776 __time_get::~__time_get() 4777 { 4778 freelocale(__loc_); 4779 } 4780 #if defined(__clang__) 4781 #pragma clang diagnostic ignored "-Wmissing-field-initializers" 4782 #endif 4783 #if defined(__GNUG__) 4784 #pragma GCC diagnostic ignored "-Wmissing-field-initializers" 4785 #endif 4786 4787 template <> 4788 string 4789 __time_get_storage<char>::__analyze(char fmt, const ctype<char>& ct) 4790 { 4791 tm t = {0}; 4792 t.tm_sec = 59; 4793 t.tm_min = 55; 4794 t.tm_hour = 23; 4795 t.tm_mday = 31; 4796 t.tm_mon = 11; 4797 t.tm_year = 161; 4798 t.tm_wday = 6; 4799 t.tm_yday = 364; 4800 t.tm_isdst = -1; 4801 char buf[100]; 4802 char f[3] = {0}; 4803 f[0] = '%'; 4804 f[1] = fmt; 4805 size_t n = strftime_l(buf, countof(buf), f, &t, __loc_); 4806 char* bb = buf; 4807 char* be = buf + n; 4808 string result; 4809 while (bb != be) 4810 { 4811 if (ct.is(ctype_base::space, *bb)) 4812 { 4813 result.push_back(' '); 4814 for (++bb; bb != be && ct.is(ctype_base::space, *bb); ++bb) 4815 ; 4816 continue; 4817 } 4818 char* w = bb; 4819 ios_base::iostate err = ios_base::goodbit; 4820 ptrdiff_t i = __scan_keyword(w, be, this->__weeks_, this->__weeks_+14, 4821 ct, err, false) 4822 - this->__weeks_; 4823 if (i < 14) 4824 { 4825 result.push_back('%'); 4826 if (i < 7) 4827 result.push_back('A'); 4828 else 4829 result.push_back('a'); 4830 bb = w; 4831 continue; 4832 } 4833 w = bb; 4834 i = __scan_keyword(w, be, this->__months_, this->__months_+24, 4835 ct, err, false) 4836 - this->__months_; 4837 if (i < 24) 4838 { 4839 result.push_back('%'); 4840 if (i < 12) 4841 result.push_back('B'); 4842 else 4843 result.push_back('b'); 4844 if (fmt == 'x' && ct.is(ctype_base::digit, this->__months_[i][0])) 4845 result.back() = 'm'; 4846 bb = w; 4847 continue; 4848 } 4849 if (this->__am_pm_[0].size() + this->__am_pm_[1].size() > 0) 4850 { 4851 w = bb; 4852 i = __scan_keyword(w, be, this->__am_pm_, this->__am_pm_+2, 4853 ct, err, false) - this->__am_pm_; 4854 if (i < 2) 4855 { 4856 result.push_back('%'); 4857 result.push_back('p'); 4858 bb = w; 4859 continue; 4860 } 4861 } 4862 w = bb; 4863 if (ct.is(ctype_base::digit, *bb)) 4864 { 4865 switch(__get_up_to_n_digits(bb, be, err, ct, 4)) 4866 { 4867 case 6: 4868 result.push_back('%'); 4869 result.push_back('w'); 4870 break; 4871 case 7: 4872 result.push_back('%'); 4873 result.push_back('u'); 4874 break; 4875 case 11: 4876 result.push_back('%'); 4877 result.push_back('I'); 4878 break; 4879 case 12: 4880 result.push_back('%'); 4881 result.push_back('m'); 4882 break; 4883 case 23: 4884 result.push_back('%'); 4885 result.push_back('H'); 4886 break; 4887 case 31: 4888 result.push_back('%'); 4889 result.push_back('d'); 4890 break; 4891 case 55: 4892 result.push_back('%'); 4893 result.push_back('M'); 4894 break; 4895 case 59: 4896 result.push_back('%'); 4897 result.push_back('S'); 4898 break; 4899 case 61: 4900 result.push_back('%'); 4901 result.push_back('y'); 4902 break; 4903 case 364: 4904 result.push_back('%'); 4905 result.push_back('j'); 4906 break; 4907 case 2061: 4908 result.push_back('%'); 4909 result.push_back('Y'); 4910 break; 4911 default: 4912 for (; w != bb; ++w) 4913 result.push_back(*w); 4914 break; 4915 } 4916 continue; 4917 } 4918 if (*bb == '%') 4919 { 4920 result.push_back('%'); 4921 result.push_back('%'); 4922 ++bb; 4923 continue; 4924 } 4925 result.push_back(*bb); 4926 ++bb; 4927 } 4928 return result; 4929 } 4930 4931 #if defined(__clang__) 4932 #pragma clang diagnostic ignored "-Wmissing-braces" 4933 #endif 4934 4935 template <> 4936 wstring 4937 __time_get_storage<wchar_t>::__analyze(char fmt, const ctype<wchar_t>& ct) 4938 { 4939 tm t = {0}; 4940 t.tm_sec = 59; 4941 t.tm_min = 55; 4942 t.tm_hour = 23; 4943 t.tm_mday = 31; 4944 t.tm_mon = 11; 4945 t.tm_year = 161; 4946 t.tm_wday = 6; 4947 t.tm_yday = 364; 4948 t.tm_isdst = -1; 4949 char buf[100]; 4950 char f[3] = {0}; 4951 f[0] = '%'; 4952 f[1] = fmt; 4953 strftime_l(buf, countof(buf), f, &t, __loc_); 4954 wchar_t wbuf[100]; 4955 wchar_t* wbb = wbuf; 4956 mbstate_t mb = {0}; 4957 const char* bb = buf; 4958 size_t j = __libcpp_mbsrtowcs_l( wbb, &bb, countof(wbuf), &mb, __loc_); 4959 if (j == size_t(-1)) 4960 __throw_runtime_error("locale not supported"); 4961 wchar_t* wbe = wbb + j; 4962 wstring result; 4963 while (wbb != wbe) 4964 { 4965 if (ct.is(ctype_base::space, *wbb)) 4966 { 4967 result.push_back(L' '); 4968 for (++wbb; wbb != wbe && ct.is(ctype_base::space, *wbb); ++wbb) 4969 ; 4970 continue; 4971 } 4972 wchar_t* w = wbb; 4973 ios_base::iostate err = ios_base::goodbit; 4974 ptrdiff_t i = __scan_keyword(w, wbe, this->__weeks_, this->__weeks_+14, 4975 ct, err, false) 4976 - this->__weeks_; 4977 if (i < 14) 4978 { 4979 result.push_back(L'%'); 4980 if (i < 7) 4981 result.push_back(L'A'); 4982 else 4983 result.push_back(L'a'); 4984 wbb = w; 4985 continue; 4986 } 4987 w = wbb; 4988 i = __scan_keyword(w, wbe, this->__months_, this->__months_+24, 4989 ct, err, false) 4990 - this->__months_; 4991 if (i < 24) 4992 { 4993 result.push_back(L'%'); 4994 if (i < 12) 4995 result.push_back(L'B'); 4996 else 4997 result.push_back(L'b'); 4998 if (fmt == 'x' && ct.is(ctype_base::digit, this->__months_[i][0])) 4999 result.back() = L'm'; 5000 wbb = w; 5001 continue; 5002 } 5003 if (this->__am_pm_[0].size() + this->__am_pm_[1].size() > 0) 5004 { 5005 w = wbb; 5006 i = __scan_keyword(w, wbe, this->__am_pm_, this->__am_pm_+2, 5007 ct, err, false) - this->__am_pm_; 5008 if (i < 2) 5009 { 5010 result.push_back(L'%'); 5011 result.push_back(L'p'); 5012 wbb = w; 5013 continue; 5014 } 5015 } 5016 w = wbb; 5017 if (ct.is(ctype_base::digit, *wbb)) 5018 { 5019 switch(__get_up_to_n_digits(wbb, wbe, err, ct, 4)) 5020 { 5021 case 6: 5022 result.push_back(L'%'); 5023 result.push_back(L'w'); 5024 break; 5025 case 7: 5026 result.push_back(L'%'); 5027 result.push_back(L'u'); 5028 break; 5029 case 11: 5030 result.push_back(L'%'); 5031 result.push_back(L'I'); 5032 break; 5033 case 12: 5034 result.push_back(L'%'); 5035 result.push_back(L'm'); 5036 break; 5037 case 23: 5038 result.push_back(L'%'); 5039 result.push_back(L'H'); 5040 break; 5041 case 31: 5042 result.push_back(L'%'); 5043 result.push_back(L'd'); 5044 break; 5045 case 55: 5046 result.push_back(L'%'); 5047 result.push_back(L'M'); 5048 break; 5049 case 59: 5050 result.push_back(L'%'); 5051 result.push_back(L'S'); 5052 break; 5053 case 61: 5054 result.push_back(L'%'); 5055 result.push_back(L'y'); 5056 break; 5057 case 364: 5058 result.push_back(L'%'); 5059 result.push_back(L'j'); 5060 break; 5061 case 2061: 5062 result.push_back(L'%'); 5063 result.push_back(L'Y'); 5064 break; 5065 default: 5066 for (; w != wbb; ++w) 5067 result.push_back(*w); 5068 break; 5069 } 5070 continue; 5071 } 5072 if (ct.narrow(*wbb, 0) == '%') 5073 { 5074 result.push_back(L'%'); 5075 result.push_back(L'%'); 5076 ++wbb; 5077 continue; 5078 } 5079 result.push_back(*wbb); 5080 ++wbb; 5081 } 5082 return result; 5083 } 5084 5085 template <> 5086 void 5087 __time_get_storage<char>::init(const ctype<char>& ct) 5088 { 5089 tm t = {0}; 5090 char buf[100]; 5091 // __weeks_ 5092 for (int i = 0; i < 7; ++i) 5093 { 5094 t.tm_wday = i; 5095 strftime_l(buf, countof(buf), "%A", &t, __loc_); 5096 __weeks_[i] = buf; 5097 strftime_l(buf, countof(buf), "%a", &t, __loc_); 5098 __weeks_[i+7] = buf; 5099 } 5100 // __months_ 5101 for (int i = 0; i < 12; ++i) 5102 { 5103 t.tm_mon = i; 5104 strftime_l(buf, countof(buf), "%B", &t, __loc_); 5105 __months_[i] = buf; 5106 strftime_l(buf, countof(buf), "%b", &t, __loc_); 5107 __months_[i+12] = buf; 5108 } 5109 // __am_pm_ 5110 t.tm_hour = 1; 5111 strftime_l(buf, countof(buf), "%p", &t, __loc_); 5112 __am_pm_[0] = buf; 5113 t.tm_hour = 13; 5114 strftime_l(buf, countof(buf), "%p", &t, __loc_); 5115 __am_pm_[1] = buf; 5116 __c_ = __analyze('c', ct); 5117 __r_ = __analyze('r', ct); 5118 __x_ = __analyze('x', ct); 5119 __X_ = __analyze('X', ct); 5120 } 5121 5122 template <> 5123 void 5124 __time_get_storage<wchar_t>::init(const ctype<wchar_t>& ct) 5125 { 5126 tm t = {0}; 5127 char buf[100]; 5128 wchar_t wbuf[100]; 5129 wchar_t* wbe; 5130 mbstate_t mb = {0}; 5131 // __weeks_ 5132 for (int i = 0; i < 7; ++i) 5133 { 5134 t.tm_wday = i; 5135 strftime_l(buf, countof(buf), "%A", &t, __loc_); 5136 mb = mbstate_t(); 5137 const char* bb = buf; 5138 size_t j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); 5139 if (j == size_t(-1)) 5140 __throw_runtime_error("locale not supported"); 5141 wbe = wbuf + j; 5142 __weeks_[i].assign(wbuf, wbe); 5143 strftime_l(buf, countof(buf), "%a", &t, __loc_); 5144 mb = mbstate_t(); 5145 bb = buf; 5146 j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); 5147 if (j == size_t(-1)) 5148 __throw_runtime_error("locale not supported"); 5149 wbe = wbuf + j; 5150 __weeks_[i+7].assign(wbuf, wbe); 5151 } 5152 // __months_ 5153 for (int i = 0; i < 12; ++i) 5154 { 5155 t.tm_mon = i; 5156 strftime_l(buf, countof(buf), "%B", &t, __loc_); 5157 mb = mbstate_t(); 5158 const char* bb = buf; 5159 size_t j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); 5160 if (j == size_t(-1)) 5161 __throw_runtime_error("locale not supported"); 5162 wbe = wbuf + j; 5163 __months_[i].assign(wbuf, wbe); 5164 strftime_l(buf, countof(buf), "%b", &t, __loc_); 5165 mb = mbstate_t(); 5166 bb = buf; 5167 j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); 5168 if (j == size_t(-1)) 5169 __throw_runtime_error("locale not supported"); 5170 wbe = wbuf + j; 5171 __months_[i+12].assign(wbuf, wbe); 5172 } 5173 // __am_pm_ 5174 t.tm_hour = 1; 5175 strftime_l(buf, countof(buf), "%p", &t, __loc_); 5176 mb = mbstate_t(); 5177 const char* bb = buf; 5178 size_t j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); 5179 if (j == size_t(-1)) 5180 __throw_runtime_error("locale not supported"); 5181 wbe = wbuf + j; 5182 __am_pm_[0].assign(wbuf, wbe); 5183 t.tm_hour = 13; 5184 strftime_l(buf, countof(buf), "%p", &t, __loc_); 5185 mb = mbstate_t(); 5186 bb = buf; 5187 j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, __loc_); 5188 if (j == size_t(-1)) 5189 __throw_runtime_error("locale not supported"); 5190 wbe = wbuf + j; 5191 __am_pm_[1].assign(wbuf, wbe); 5192 __c_ = __analyze('c', ct); 5193 __r_ = __analyze('r', ct); 5194 __x_ = __analyze('x', ct); 5195 __X_ = __analyze('X', ct); 5196 } 5197 5198 template <class CharT> 5199 struct _LIBCPP_HIDDEN __time_get_temp 5200 : public ctype_byname<CharT> 5201 { 5202 explicit __time_get_temp(const char* nm) 5203 : ctype_byname<CharT>(nm, 1) {} 5204 explicit __time_get_temp(const string& nm) 5205 : ctype_byname<CharT>(nm, 1) {} 5206 }; 5207 5208 template <> 5209 __time_get_storage<char>::__time_get_storage(const char* __nm) 5210 : __time_get(__nm) 5211 { 5212 const __time_get_temp<char> ct(__nm); 5213 init(ct); 5214 } 5215 5216 template <> 5217 __time_get_storage<char>::__time_get_storage(const string& __nm) 5218 : __time_get(__nm) 5219 { 5220 const __time_get_temp<char> ct(__nm); 5221 init(ct); 5222 } 5223 5224 template <> 5225 __time_get_storage<wchar_t>::__time_get_storage(const char* __nm) 5226 : __time_get(__nm) 5227 { 5228 const __time_get_temp<wchar_t> ct(__nm); 5229 init(ct); 5230 } 5231 5232 template <> 5233 __time_get_storage<wchar_t>::__time_get_storage(const string& __nm) 5234 : __time_get(__nm) 5235 { 5236 const __time_get_temp<wchar_t> ct(__nm); 5237 init(ct); 5238 } 5239 5240 template <> 5241 time_base::dateorder 5242 __time_get_storage<char>::__do_date_order() const 5243 { 5244 unsigned i; 5245 for (i = 0; i < __x_.size(); ++i) 5246 if (__x_[i] == '%') 5247 break; 5248 ++i; 5249 switch (__x_[i]) 5250 { 5251 case 'y': 5252 case 'Y': 5253 for (++i; i < __x_.size(); ++i) 5254 if (__x_[i] == '%') 5255 break; 5256 if (i == __x_.size()) 5257 break; 5258 ++i; 5259 switch (__x_[i]) 5260 { 5261 case 'm': 5262 for (++i; i < __x_.size(); ++i) 5263 if (__x_[i] == '%') 5264 break; 5265 if (i == __x_.size()) 5266 break; 5267 ++i; 5268 if (__x_[i] == 'd') 5269 return time_base::ymd; 5270 break; 5271 case 'd': 5272 for (++i; i < __x_.size(); ++i) 5273 if (__x_[i] == '%') 5274 break; 5275 if (i == __x_.size()) 5276 break; 5277 ++i; 5278 if (__x_[i] == 'm') 5279 return time_base::ydm; 5280 break; 5281 } 5282 break; 5283 case 'm': 5284 for (++i; i < __x_.size(); ++i) 5285 if (__x_[i] == '%') 5286 break; 5287 if (i == __x_.size()) 5288 break; 5289 ++i; 5290 if (__x_[i] == 'd') 5291 { 5292 for (++i; i < __x_.size(); ++i) 5293 if (__x_[i] == '%') 5294 break; 5295 if (i == __x_.size()) 5296 break; 5297 ++i; 5298 if (__x_[i] == 'y' || __x_[i] == 'Y') 5299 return time_base::mdy; 5300 break; 5301 } 5302 break; 5303 case 'd': 5304 for (++i; i < __x_.size(); ++i) 5305 if (__x_[i] == '%') 5306 break; 5307 if (i == __x_.size()) 5308 break; 5309 ++i; 5310 if (__x_[i] == 'm') 5311 { 5312 for (++i; i < __x_.size(); ++i) 5313 if (__x_[i] == '%') 5314 break; 5315 if (i == __x_.size()) 5316 break; 5317 ++i; 5318 if (__x_[i] == 'y' || __x_[i] == 'Y') 5319 return time_base::dmy; 5320 break; 5321 } 5322 break; 5323 } 5324 return time_base::no_order; 5325 } 5326 5327 template <> 5328 time_base::dateorder 5329 __time_get_storage<wchar_t>::__do_date_order() const 5330 { 5331 unsigned i; 5332 for (i = 0; i < __x_.size(); ++i) 5333 if (__x_[i] == L'%') 5334 break; 5335 ++i; 5336 switch (__x_[i]) 5337 { 5338 case L'y': 5339 case L'Y': 5340 for (++i; i < __x_.size(); ++i) 5341 if (__x_[i] == L'%') 5342 break; 5343 if (i == __x_.size()) 5344 break; 5345 ++i; 5346 switch (__x_[i]) 5347 { 5348 case L'm': 5349 for (++i; i < __x_.size(); ++i) 5350 if (__x_[i] == L'%') 5351 break; 5352 if (i == __x_.size()) 5353 break; 5354 ++i; 5355 if (__x_[i] == L'd') 5356 return time_base::ymd; 5357 break; 5358 case L'd': 5359 for (++i; i < __x_.size(); ++i) 5360 if (__x_[i] == L'%') 5361 break; 5362 if (i == __x_.size()) 5363 break; 5364 ++i; 5365 if (__x_[i] == L'm') 5366 return time_base::ydm; 5367 break; 5368 } 5369 break; 5370 case L'm': 5371 for (++i; i < __x_.size(); ++i) 5372 if (__x_[i] == L'%') 5373 break; 5374 if (i == __x_.size()) 5375 break; 5376 ++i; 5377 if (__x_[i] == L'd') 5378 { 5379 for (++i; i < __x_.size(); ++i) 5380 if (__x_[i] == L'%') 5381 break; 5382 if (i == __x_.size()) 5383 break; 5384 ++i; 5385 if (__x_[i] == L'y' || __x_[i] == L'Y') 5386 return time_base::mdy; 5387 break; 5388 } 5389 break; 5390 case L'd': 5391 for (++i; i < __x_.size(); ++i) 5392 if (__x_[i] == L'%') 5393 break; 5394 if (i == __x_.size()) 5395 break; 5396 ++i; 5397 if (__x_[i] == L'm') 5398 { 5399 for (++i; i < __x_.size(); ++i) 5400 if (__x_[i] == L'%') 5401 break; 5402 if (i == __x_.size()) 5403 break; 5404 ++i; 5405 if (__x_[i] == L'y' || __x_[i] == L'Y') 5406 return time_base::dmy; 5407 break; 5408 } 5409 break; 5410 } 5411 return time_base::no_order; 5412 } 5413 5414 // time_put 5415 5416 __time_put::__time_put(const char* nm) 5417 : __loc_(newlocale(LC_ALL_MASK, nm, 0)) 5418 { 5419 if (__loc_ == 0) 5420 __throw_runtime_error("time_put_byname" 5421 " failed to construct for " + string(nm)); 5422 } 5423 5424 __time_put::__time_put(const string& nm) 5425 : __loc_(newlocale(LC_ALL_MASK, nm.c_str(), 0)) 5426 { 5427 if (__loc_ == 0) 5428 __throw_runtime_error("time_put_byname" 5429 " failed to construct for " + nm); 5430 } 5431 5432 __time_put::~__time_put() 5433 { 5434 if (__loc_ != _LIBCPP_GET_C_LOCALE) 5435 freelocale(__loc_); 5436 } 5437 5438 void 5439 __time_put::__do_put(char* __nb, char*& __ne, const tm* __tm, 5440 char __fmt, char __mod) const 5441 { 5442 char fmt[] = {'%', __fmt, __mod, 0}; 5443 if (__mod != 0) 5444 swap(fmt[1], fmt[2]); 5445 size_t n = strftime_l(__nb, countof(__nb, __ne), fmt, __tm, __loc_); 5446 __ne = __nb + n; 5447 } 5448 5449 void 5450 __time_put::__do_put(wchar_t* __wb, wchar_t*& __we, const tm* __tm, 5451 char __fmt, char __mod) const 5452 { 5453 char __nar[100]; 5454 char* __ne = __nar + 100; 5455 __do_put(__nar, __ne, __tm, __fmt, __mod); 5456 mbstate_t mb = {0}; 5457 const char* __nb = __nar; 5458 size_t j = __libcpp_mbsrtowcs_l(__wb, &__nb, countof(__wb, __we), &mb, __loc_); 5459 if (j == size_t(-1)) 5460 __throw_runtime_error("locale not supported"); 5461 __we = __wb + j; 5462 } 5463 5464 // moneypunct_byname 5465 5466 template <class charT> 5467 static 5468 void 5469 __init_pat(money_base::pattern& pat, basic_string<charT>& __curr_symbol_, 5470 bool intl, char cs_precedes, char sep_by_space, char sign_posn, 5471 charT space_char) 5472 { 5473 const char sign = static_cast<char>(money_base::sign); 5474 const char space = static_cast<char>(money_base::space); 5475 const char none = static_cast<char>(money_base::none); 5476 const char symbol = static_cast<char>(money_base::symbol); 5477 const char value = static_cast<char>(money_base::value); 5478 const bool symbol_contains_sep = intl && __curr_symbol_.size() == 4; 5479 5480 // Comments on case branches reflect 'C11 7.11.2.1 The localeconv 5481 // function'. "Space between sign and symbol or value" means that 5482 // if the sign is adjacent to the symbol, there's a space between 5483 // them, and otherwise there's a space between the sign and value. 5484 // 5485 // C11's localeconv specifies that the fourth character of an 5486 // international curr_symbol is used to separate the sign and 5487 // value when sep_by_space says to do so. C++ can't represent 5488 // that, so we just use a space. When sep_by_space says to 5489 // separate the symbol and value-or-sign with a space, we rearrange the 5490 // curr_symbol to put its spacing character on the correct side of 5491 // the symbol. 5492 // 5493 // We also need to avoid adding an extra space between the sign 5494 // and value when the currency symbol is suppressed (by not 5495 // setting showbase). We match glibc's strfmon by interpreting 5496 // sep_by_space==1 as "omit the space when the currency symbol is 5497 // absent". 5498 // 5499 // Users who want to get this right should use ICU instead. 5500 5501 switch (cs_precedes) 5502 { 5503 case 0: // value before curr_symbol 5504 if (symbol_contains_sep) { 5505 // Move the separator to before the symbol, to place it 5506 // between the value and symbol. 5507 rotate(__curr_symbol_.begin(), __curr_symbol_.begin() + 3, 5508 __curr_symbol_.end()); 5509 } 5510 switch (sign_posn) 5511 { 5512 case 0: // Parentheses surround the quantity and currency symbol. 5513 pat.field[0] = sign; 5514 pat.field[1] = value; 5515 pat.field[2] = none; // Any space appears in the symbol. 5516 pat.field[3] = symbol; 5517 switch (sep_by_space) 5518 { 5519 case 0: // No space separates the currency symbol and value. 5520 // This case may have changed between C99 and C11; 5521 // assume the currency symbol matches the intention. 5522 case 2: // Space between sign and currency or value. 5523 // The "sign" is two parentheses, so no space here either. 5524 return; 5525 case 1: // Space between currency-and-sign or currency and value. 5526 if (!symbol_contains_sep) { 5527 // We insert the space into the symbol instead of 5528 // setting pat.field[2]=space so that when 5529 // showbase is not set, the space goes away too. 5530 __curr_symbol_.insert(0, 1, space_char); 5531 } 5532 return; 5533 default: 5534 break; 5535 } 5536 break; 5537 case 1: // The sign string precedes the quantity and currency symbol. 5538 pat.field[0] = sign; 5539 pat.field[3] = symbol; 5540 switch (sep_by_space) 5541 { 5542 case 0: // No space separates the currency symbol and value. 5543 pat.field[1] = value; 5544 pat.field[2] = none; 5545 return; 5546 case 1: // Space between currency-and-sign or currency and value. 5547 pat.field[1] = value; 5548 pat.field[2] = none; 5549 if (!symbol_contains_sep) { 5550 // We insert the space into the symbol instead of 5551 // setting pat.field[2]=space so that when 5552 // showbase is not set, the space goes away too. 5553 __curr_symbol_.insert(0, 1, space_char); 5554 } 5555 return; 5556 case 2: // Space between sign and currency or value. 5557 pat.field[1] = space; 5558 pat.field[2] = value; 5559 if (symbol_contains_sep) { 5560 // Remove the separator from the symbol, since it 5561 // has already appeared after the sign. 5562 __curr_symbol_.erase(__curr_symbol_.begin()); 5563 } 5564 return; 5565 default: 5566 break; 5567 } 5568 break; 5569 case 2: // The sign string succeeds the quantity and currency symbol. 5570 pat.field[0] = value; 5571 pat.field[3] = sign; 5572 switch (sep_by_space) 5573 { 5574 case 0: // No space separates the currency symbol and value. 5575 pat.field[1] = none; 5576 pat.field[2] = symbol; 5577 return; 5578 case 1: // Space between currency-and-sign or currency and value. 5579 if (!symbol_contains_sep) { 5580 // We insert the space into the symbol instead of 5581 // setting pat.field[1]=space so that when 5582 // showbase is not set, the space goes away too. 5583 __curr_symbol_.insert(0, 1, space_char); 5584 } 5585 pat.field[1] = none; 5586 pat.field[2] = symbol; 5587 return; 5588 case 2: // Space between sign and currency or value. 5589 pat.field[1] = symbol; 5590 pat.field[2] = space; 5591 if (symbol_contains_sep) { 5592 // Remove the separator from the symbol, since it 5593 // should not be removed if showbase is absent. 5594 __curr_symbol_.erase(__curr_symbol_.begin()); 5595 } 5596 return; 5597 default: 5598 break; 5599 } 5600 break; 5601 case 3: // The sign string immediately precedes the currency symbol. 5602 pat.field[0] = value; 5603 pat.field[3] = symbol; 5604 switch (sep_by_space) 5605 { 5606 case 0: // No space separates the currency symbol and value. 5607 pat.field[1] = none; 5608 pat.field[2] = sign; 5609 return; 5610 case 1: // Space between currency-and-sign or currency and value. 5611 pat.field[1] = space; 5612 pat.field[2] = sign; 5613 if (symbol_contains_sep) { 5614 // Remove the separator from the symbol, since it 5615 // has already appeared before the sign. 5616 __curr_symbol_.erase(__curr_symbol_.begin()); 5617 } 5618 return; 5619 case 2: // Space between sign and currency or value. 5620 pat.field[1] = sign; 5621 pat.field[2] = none; 5622 if (!symbol_contains_sep) { 5623 // We insert the space into the symbol instead of 5624 // setting pat.field[2]=space so that when 5625 // showbase is not set, the space goes away too. 5626 __curr_symbol_.insert(0, 1, space_char); 5627 } 5628 return; 5629 default: 5630 break; 5631 } 5632 break; 5633 case 4: // The sign string immediately succeeds the currency symbol. 5634 pat.field[0] = value; 5635 pat.field[3] = sign; 5636 switch (sep_by_space) 5637 { 5638 case 0: // No space separates the currency symbol and value. 5639 pat.field[1] = none; 5640 pat.field[2] = symbol; 5641 return; 5642 case 1: // Space between currency-and-sign or currency and value. 5643 pat.field[1] = none; 5644 pat.field[2] = symbol; 5645 if (!symbol_contains_sep) { 5646 // We insert the space into the symbol instead of 5647 // setting pat.field[1]=space so that when 5648 // showbase is not set, the space goes away too. 5649 __curr_symbol_.insert(0, 1, space_char); 5650 } 5651 return; 5652 case 2: // Space between sign and currency or value. 5653 pat.field[1] = symbol; 5654 pat.field[2] = space; 5655 if (symbol_contains_sep) { 5656 // Remove the separator from the symbol, since it 5657 // should not disappear when showbase is absent. 5658 __curr_symbol_.erase(__curr_symbol_.begin()); 5659 } 5660 return; 5661 default: 5662 break; 5663 } 5664 break; 5665 default: 5666 break; 5667 } 5668 break; 5669 case 1: // curr_symbol before value 5670 switch (sign_posn) 5671 { 5672 case 0: // Parentheses surround the quantity and currency symbol. 5673 pat.field[0] = sign; 5674 pat.field[1] = symbol; 5675 pat.field[2] = none; // Any space appears in the symbol. 5676 pat.field[3] = value; 5677 switch (sep_by_space) 5678 { 5679 case 0: // No space separates the currency symbol and value. 5680 // This case may have changed between C99 and C11; 5681 // assume the currency symbol matches the intention. 5682 case 2: // Space between sign and currency or value. 5683 // The "sign" is two parentheses, so no space here either. 5684 return; 5685 case 1: // Space between currency-and-sign or currency and value. 5686 if (!symbol_contains_sep) { 5687 // We insert the space into the symbol instead of 5688 // setting pat.field[2]=space so that when 5689 // showbase is not set, the space goes away too. 5690 __curr_symbol_.insert(0, 1, space_char); 5691 } 5692 return; 5693 default: 5694 break; 5695 } 5696 break; 5697 case 1: // The sign string precedes the quantity and currency symbol. 5698 pat.field[0] = sign; 5699 pat.field[3] = value; 5700 switch (sep_by_space) 5701 { 5702 case 0: // No space separates the currency symbol and value. 5703 pat.field[1] = symbol; 5704 pat.field[2] = none; 5705 return; 5706 case 1: // Space between currency-and-sign or currency and value. 5707 pat.field[1] = symbol; 5708 pat.field[2] = none; 5709 if (!symbol_contains_sep) { 5710 // We insert the space into the symbol instead of 5711 // setting pat.field[2]=space so that when 5712 // showbase is not set, the space goes away too. 5713 __curr_symbol_.push_back(space_char); 5714 } 5715 return; 5716 case 2: // Space between sign and currency or value. 5717 pat.field[1] = space; 5718 pat.field[2] = symbol; 5719 if (symbol_contains_sep) { 5720 // Remove the separator from the symbol, since it 5721 // has already appeared after the sign. 5722 __curr_symbol_.pop_back(); 5723 } 5724 return; 5725 default: 5726 break; 5727 } 5728 break; 5729 case 2: // The sign string succeeds the quantity and currency symbol. 5730 pat.field[0] = symbol; 5731 pat.field[3] = sign; 5732 switch (sep_by_space) 5733 { 5734 case 0: // No space separates the currency symbol and value. 5735 pat.field[1] = none; 5736 pat.field[2] = value; 5737 return; 5738 case 1: // Space between currency-and-sign or currency and value. 5739 pat.field[1] = none; 5740 pat.field[2] = value; 5741 if (!symbol_contains_sep) { 5742 // We insert the space into the symbol instead of 5743 // setting pat.field[1]=space so that when 5744 // showbase is not set, the space goes away too. 5745 __curr_symbol_.push_back(space_char); 5746 } 5747 return; 5748 case 2: // Space between sign and currency or value. 5749 pat.field[1] = value; 5750 pat.field[2] = space; 5751 if (symbol_contains_sep) { 5752 // Remove the separator from the symbol, since it 5753 // will appear before the sign. 5754 __curr_symbol_.pop_back(); 5755 } 5756 return; 5757 default: 5758 break; 5759 } 5760 break; 5761 case 3: // The sign string immediately precedes the currency symbol. 5762 pat.field[0] = sign; 5763 pat.field[3] = value; 5764 switch (sep_by_space) 5765 { 5766 case 0: // No space separates the currency symbol and value. 5767 pat.field[1] = symbol; 5768 pat.field[2] = none; 5769 return; 5770 case 1: // Space between currency-and-sign or currency and value. 5771 pat.field[1] = symbol; 5772 pat.field[2] = none; 5773 if (!symbol_contains_sep) { 5774 // We insert the space into the symbol instead of 5775 // setting pat.field[2]=space so that when 5776 // showbase is not set, the space goes away too. 5777 __curr_symbol_.push_back(space_char); 5778 } 5779 return; 5780 case 2: // Space between sign and currency or value. 5781 pat.field[1] = space; 5782 pat.field[2] = symbol; 5783 if (symbol_contains_sep) { 5784 // Remove the separator from the symbol, since it 5785 // has already appeared after the sign. 5786 __curr_symbol_.pop_back(); 5787 } 5788 return; 5789 default: 5790 break; 5791 } 5792 break; 5793 case 4: // The sign string immediately succeeds the currency symbol. 5794 pat.field[0] = symbol; 5795 pat.field[3] = value; 5796 switch (sep_by_space) 5797 { 5798 case 0: // No space separates the currency symbol and value. 5799 pat.field[1] = sign; 5800 pat.field[2] = none; 5801 return; 5802 case 1: // Space between currency-and-sign or currency and value. 5803 pat.field[1] = sign; 5804 pat.field[2] = space; 5805 if (symbol_contains_sep) { 5806 // Remove the separator from the symbol, since it 5807 // should not disappear when showbase is absent. 5808 __curr_symbol_.pop_back(); 5809 } 5810 return; 5811 case 2: // Space between sign and currency or value. 5812 pat.field[1] = none; 5813 pat.field[2] = sign; 5814 if (!symbol_contains_sep) { 5815 // We insert the space into the symbol instead of 5816 // setting pat.field[1]=space so that when 5817 // showbase is not set, the space goes away too. 5818 __curr_symbol_.push_back(space_char); 5819 } 5820 return; 5821 default: 5822 break; 5823 } 5824 break; 5825 default: 5826 break; 5827 } 5828 break; 5829 default: 5830 break; 5831 } 5832 pat.field[0] = symbol; 5833 pat.field[1] = sign; 5834 pat.field[2] = none; 5835 pat.field[3] = value; 5836 } 5837 5838 template<> 5839 void 5840 moneypunct_byname<char, false>::init(const char* nm) 5841 { 5842 typedef moneypunct<char, false> base; 5843 __libcpp_unique_locale loc(nm); 5844 if (!loc) 5845 __throw_runtime_error("moneypunct_byname" 5846 " failed to construct for " + string(nm)); 5847 5848 lconv* lc = __libcpp_localeconv_l(loc.get()); 5849 if (!checked_string_to_char_convert(__decimal_point_, 5850 lc->mon_decimal_point, 5851 loc.get())) 5852 __decimal_point_ = base::do_decimal_point(); 5853 if (!checked_string_to_char_convert(__thousands_sep_, 5854 lc->mon_thousands_sep, 5855 loc.get())) 5856 __thousands_sep_ = base::do_thousands_sep(); 5857 5858 __grouping_ = lc->mon_grouping; 5859 __curr_symbol_ = lc->currency_symbol; 5860 if (lc->frac_digits != CHAR_MAX) 5861 __frac_digits_ = lc->frac_digits; 5862 else 5863 __frac_digits_ = base::do_frac_digits(); 5864 if (lc->p_sign_posn == 0) 5865 __positive_sign_ = "()"; 5866 else 5867 __positive_sign_ = lc->positive_sign; 5868 if (lc->n_sign_posn == 0) 5869 __negative_sign_ = "()"; 5870 else 5871 __negative_sign_ = lc->negative_sign; 5872 // Assume the positive and negative formats will want spaces in 5873 // the same places in curr_symbol since there's no way to 5874 // represent anything else. 5875 string_type __dummy_curr_symbol = __curr_symbol_; 5876 __init_pat(__pos_format_, __dummy_curr_symbol, false, 5877 lc->p_cs_precedes, lc->p_sep_by_space, lc->p_sign_posn, ' '); 5878 __init_pat(__neg_format_, __curr_symbol_, false, 5879 lc->n_cs_precedes, lc->n_sep_by_space, lc->n_sign_posn, ' '); 5880 } 5881 5882 template<> 5883 void 5884 moneypunct_byname<char, true>::init(const char* nm) 5885 { 5886 typedef moneypunct<char, true> base; 5887 __libcpp_unique_locale loc(nm); 5888 if (!loc) 5889 __throw_runtime_error("moneypunct_byname" 5890 " failed to construct for " + string(nm)); 5891 5892 lconv* lc = __libcpp_localeconv_l(loc.get()); 5893 if (!checked_string_to_char_convert(__decimal_point_, 5894 lc->mon_decimal_point, 5895 loc.get())) 5896 __decimal_point_ = base::do_decimal_point(); 5897 if (!checked_string_to_char_convert(__thousands_sep_, 5898 lc->mon_thousands_sep, 5899 loc.get())) 5900 __thousands_sep_ = base::do_thousands_sep(); 5901 __grouping_ = lc->mon_grouping; 5902 __curr_symbol_ = lc->int_curr_symbol; 5903 if (lc->int_frac_digits != CHAR_MAX) 5904 __frac_digits_ = lc->int_frac_digits; 5905 else 5906 __frac_digits_ = base::do_frac_digits(); 5907 #if defined(_LIBCPP_MSVCRT) || defined(__MINGW32__) 5908 if (lc->p_sign_posn == 0) 5909 #else // _LIBCPP_MSVCRT 5910 if (lc->int_p_sign_posn == 0) 5911 #endif // !_LIBCPP_MSVCRT 5912 __positive_sign_ = "()"; 5913 else 5914 __positive_sign_ = lc->positive_sign; 5915 #if defined(_LIBCPP_MSVCRT) || defined(__MINGW32__) 5916 if(lc->n_sign_posn == 0) 5917 #else // _LIBCPP_MSVCRT 5918 if (lc->int_n_sign_posn == 0) 5919 #endif // !_LIBCPP_MSVCRT 5920 __negative_sign_ = "()"; 5921 else 5922 __negative_sign_ = lc->negative_sign; 5923 // Assume the positive and negative formats will want spaces in 5924 // the same places in curr_symbol since there's no way to 5925 // represent anything else. 5926 string_type __dummy_curr_symbol = __curr_symbol_; 5927 #if defined(_LIBCPP_MSVCRT) || defined(__MINGW32__) 5928 __init_pat(__pos_format_, __dummy_curr_symbol, true, 5929 lc->p_cs_precedes, lc->p_sep_by_space, lc->p_sign_posn, ' '); 5930 __init_pat(__neg_format_, __curr_symbol_, true, 5931 lc->n_cs_precedes, lc->n_sep_by_space, lc->n_sign_posn, ' '); 5932 #else // _LIBCPP_MSVCRT 5933 __init_pat(__pos_format_, __dummy_curr_symbol, true, 5934 lc->int_p_cs_precedes, lc->int_p_sep_by_space, 5935 lc->int_p_sign_posn, ' '); 5936 __init_pat(__neg_format_, __curr_symbol_, true, 5937 lc->int_n_cs_precedes, lc->int_n_sep_by_space, 5938 lc->int_n_sign_posn, ' '); 5939 #endif // !_LIBCPP_MSVCRT 5940 } 5941 5942 template<> 5943 void 5944 moneypunct_byname<wchar_t, false>::init(const char* nm) 5945 { 5946 typedef moneypunct<wchar_t, false> base; 5947 __libcpp_unique_locale loc(nm); 5948 if (!loc) 5949 __throw_runtime_error("moneypunct_byname" 5950 " failed to construct for " + string(nm)); 5951 lconv* lc = __libcpp_localeconv_l(loc.get()); 5952 if (!checked_string_to_wchar_convert(__decimal_point_, 5953 lc->mon_decimal_point, 5954 loc.get())) 5955 __decimal_point_ = base::do_decimal_point(); 5956 if (!checked_string_to_wchar_convert(__thousands_sep_, 5957 lc->mon_thousands_sep, 5958 loc.get())) 5959 __thousands_sep_ = base::do_thousands_sep(); 5960 __grouping_ = lc->mon_grouping; 5961 wchar_t wbuf[100]; 5962 mbstate_t mb = {0}; 5963 const char* bb = lc->currency_symbol; 5964 size_t j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, loc.get()); 5965 if (j == size_t(-1)) 5966 __throw_runtime_error("locale not supported"); 5967 wchar_t* wbe = wbuf + j; 5968 __curr_symbol_.assign(wbuf, wbe); 5969 if (lc->frac_digits != CHAR_MAX) 5970 __frac_digits_ = lc->frac_digits; 5971 else 5972 __frac_digits_ = base::do_frac_digits(); 5973 if (lc->p_sign_posn == 0) 5974 __positive_sign_ = L"()"; 5975 else 5976 { 5977 mb = mbstate_t(); 5978 bb = lc->positive_sign; 5979 j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, loc.get()); 5980 if (j == size_t(-1)) 5981 __throw_runtime_error("locale not supported"); 5982 wbe = wbuf + j; 5983 __positive_sign_.assign(wbuf, wbe); 5984 } 5985 if (lc->n_sign_posn == 0) 5986 __negative_sign_ = L"()"; 5987 else 5988 { 5989 mb = mbstate_t(); 5990 bb = lc->negative_sign; 5991 j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, loc.get()); 5992 if (j == size_t(-1)) 5993 __throw_runtime_error("locale not supported"); 5994 wbe = wbuf + j; 5995 __negative_sign_.assign(wbuf, wbe); 5996 } 5997 // Assume the positive and negative formats will want spaces in 5998 // the same places in curr_symbol since there's no way to 5999 // represent anything else. 6000 string_type __dummy_curr_symbol = __curr_symbol_; 6001 __init_pat(__pos_format_, __dummy_curr_symbol, false, 6002 lc->p_cs_precedes, lc->p_sep_by_space, lc->p_sign_posn, L' '); 6003 __init_pat(__neg_format_, __curr_symbol_, false, 6004 lc->n_cs_precedes, lc->n_sep_by_space, lc->n_sign_posn, L' '); 6005 } 6006 6007 template<> 6008 void 6009 moneypunct_byname<wchar_t, true>::init(const char* nm) 6010 { 6011 typedef moneypunct<wchar_t, true> base; 6012 __libcpp_unique_locale loc(nm); 6013 if (!loc) 6014 __throw_runtime_error("moneypunct_byname" 6015 " failed to construct for " + string(nm)); 6016 6017 lconv* lc = __libcpp_localeconv_l(loc.get()); 6018 if (!checked_string_to_wchar_convert(__decimal_point_, 6019 lc->mon_decimal_point, 6020 loc.get())) 6021 __decimal_point_ = base::do_decimal_point(); 6022 if (!checked_string_to_wchar_convert(__thousands_sep_, 6023 lc->mon_thousands_sep, 6024 loc.get())) 6025 __thousands_sep_ = base::do_thousands_sep(); 6026 __grouping_ = lc->mon_grouping; 6027 wchar_t wbuf[100]; 6028 mbstate_t mb = {0}; 6029 const char* bb = lc->int_curr_symbol; 6030 size_t j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, loc.get()); 6031 if (j == size_t(-1)) 6032 __throw_runtime_error("locale not supported"); 6033 wchar_t* wbe = wbuf + j; 6034 __curr_symbol_.assign(wbuf, wbe); 6035 if (lc->int_frac_digits != CHAR_MAX) 6036 __frac_digits_ = lc->int_frac_digits; 6037 else 6038 __frac_digits_ = base::do_frac_digits(); 6039 #if defined(_LIBCPP_MSVCRT) || defined(__MINGW32__) 6040 if (lc->p_sign_posn == 0) 6041 #else // _LIBCPP_MSVCRT 6042 if (lc->int_p_sign_posn == 0) 6043 #endif // !_LIBCPP_MSVCRT 6044 __positive_sign_ = L"()"; 6045 else 6046 { 6047 mb = mbstate_t(); 6048 bb = lc->positive_sign; 6049 j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, loc.get()); 6050 if (j == size_t(-1)) 6051 __throw_runtime_error("locale not supported"); 6052 wbe = wbuf + j; 6053 __positive_sign_.assign(wbuf, wbe); 6054 } 6055 #if defined(_LIBCPP_MSVCRT) || defined(__MINGW32__) 6056 if (lc->n_sign_posn == 0) 6057 #else // _LIBCPP_MSVCRT 6058 if (lc->int_n_sign_posn == 0) 6059 #endif // !_LIBCPP_MSVCRT 6060 __negative_sign_ = L"()"; 6061 else 6062 { 6063 mb = mbstate_t(); 6064 bb = lc->negative_sign; 6065 j = __libcpp_mbsrtowcs_l(wbuf, &bb, countof(wbuf), &mb, loc.get()); 6066 if (j == size_t(-1)) 6067 __throw_runtime_error("locale not supported"); 6068 wbe = wbuf + j; 6069 __negative_sign_.assign(wbuf, wbe); 6070 } 6071 // Assume the positive and negative formats will want spaces in 6072 // the same places in curr_symbol since there's no way to 6073 // represent anything else. 6074 string_type __dummy_curr_symbol = __curr_symbol_; 6075 #if defined(_LIBCPP_MSVCRT) || defined(__MINGW32__) 6076 __init_pat(__pos_format_, __dummy_curr_symbol, true, 6077 lc->p_cs_precedes, lc->p_sep_by_space, lc->p_sign_posn, L' '); 6078 __init_pat(__neg_format_, __curr_symbol_, true, 6079 lc->n_cs_precedes, lc->n_sep_by_space, lc->n_sign_posn, L' '); 6080 #else // _LIBCPP_MSVCRT 6081 __init_pat(__pos_format_, __dummy_curr_symbol, true, 6082 lc->int_p_cs_precedes, lc->int_p_sep_by_space, 6083 lc->int_p_sign_posn, L' '); 6084 __init_pat(__neg_format_, __curr_symbol_, true, 6085 lc->int_n_cs_precedes, lc->int_n_sep_by_space, 6086 lc->int_n_sign_posn, L' '); 6087 #endif // !_LIBCPP_MSVCRT 6088 } 6089 6090 void __do_nothing(void*) {} 6091 6092 void __throw_runtime_error(const char* msg) 6093 { 6094 #ifndef _LIBCPP_NO_EXCEPTIONS 6095 throw runtime_error(msg); 6096 #else 6097 (void)msg; 6098 _VSTD::abort(); 6099 #endif 6100 } 6101 6102 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS collate<char>; 6103 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS collate<wchar_t>; 6104 6105 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS num_get<char>; 6106 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS num_get<wchar_t>; 6107 6108 template struct _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS __num_get<char>; 6109 template struct _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS __num_get<wchar_t>; 6110 6111 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS num_put<char>; 6112 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS num_put<wchar_t>; 6113 6114 template struct _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS __num_put<char>; 6115 template struct _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS __num_put<wchar_t>; 6116 6117 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS time_get<char>; 6118 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS time_get<wchar_t>; 6119 6120 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS time_get_byname<char>; 6121 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS time_get_byname<wchar_t>; 6122 6123 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS time_put<char>; 6124 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS time_put<wchar_t>; 6125 6126 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS time_put_byname<char>; 6127 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS time_put_byname<wchar_t>; 6128 6129 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS moneypunct<char, false>; 6130 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS moneypunct<char, true>; 6131 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS moneypunct<wchar_t, false>; 6132 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS moneypunct<wchar_t, true>; 6133 6134 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS moneypunct_byname<char, false>; 6135 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS moneypunct_byname<char, true>; 6136 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS moneypunct_byname<wchar_t, false>; 6137 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS moneypunct_byname<wchar_t, true>; 6138 6139 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS money_get<char>; 6140 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS money_get<wchar_t>; 6141 6142 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS __money_get<char>; 6143 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS __money_get<wchar_t>; 6144 6145 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS money_put<char>; 6146 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS money_put<wchar_t>; 6147 6148 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS __money_put<char>; 6149 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS __money_put<wchar_t>; 6150 6151 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS messages<char>; 6152 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS messages<wchar_t>; 6153 6154 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS messages_byname<char>; 6155 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS messages_byname<wchar_t>; 6156 6157 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char, char, mbstate_t>; 6158 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<wchar_t, char, mbstate_t>; 6159 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char16_t, char, mbstate_t>; 6160 template class _LIBCPP_CLASS_TEMPLATE_INSTANTIATION_VIS codecvt_byname<char32_t, char, mbstate_t>; 6161 6162 _LIBCPP_END_NAMESPACE_STD 6163