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