1// -*- C++ -*- 2//===----------------------------------------------------------------------===// 3// 4// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 5// See https://llvm.org/LICENSE.txt for license information. 6// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 7// 8//===----------------------------------------------------------------------===// 9 10#ifndef _LIBCPP_PRINT 11#define _LIBCPP_PRINT 12 13/* 14namespace std { 15 // [print.fun], print functions 16 template<class... Args> 17 void print(format_string<Args...> fmt, Args&&... args); 18 template<class... Args> 19 void print(FILE* stream, format_string<Args...> fmt, Args&&... args); 20 21 template<class... Args> 22 void println(format_string<Args...> fmt, Args&&... args); 23 template<class... Args> 24 void println(FILE* stream, format_string<Args...> fmt, Args&&... args); 25 26 void vprint_unicode(string_view fmt, format_args args); 27 void vprint_unicode(FILE* stream, string_view fmt, format_args args); 28 29 void vprint_nonunicode(string_view fmt, format_args args); 30 void vprint_nonunicode(FILE* stream, string_view fmt, format_args args); 31} 32*/ 33 34#include <__assert> // all public C++ headers provide the assertion handler 35#include <__concepts/same_as.h> 36#include <__config> 37#include <__format/buffer.h> 38#include <__format/format_arg_store.h> 39#include <__format/format_args.h> 40#include <__format/format_context.h> 41#include <__format/format_error.h> 42#include <__format/format_functions.h> 43#include <__format/unicode.h> 44#include <__system_error/system_error.h> 45#include <__utility/forward.h> 46#include <cerrno> 47#include <cstdio> 48#include <string> 49#include <string_view> 50#include <version> 51 52#if __has_include(<unistd.h>) 53# include <unistd.h> 54#endif 55 56#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER) 57# pragma GCC system_header 58#endif 59 60_LIBCPP_BEGIN_NAMESPACE_STD 61 62#ifdef _WIN32 63_LIBCPP_EXPORTED_FROM_ABI bool __is_windows_terminal(FILE* __stream); 64 65# ifndef _LIBCPP_HAS_NO_WIDE_CHARACTERS 66// A wrapper for WriteConsoleW which is used to write to the Windows 67// console. This function is in the dylib to avoid pulling in windows.h 68// in the library headers. The function itself uses some private parts 69// of the dylib too. 70// 71// The function does not depend on the language standard used. Guarding 72// it with C++23 would fail since the dylib is currently built using C++20. 73// 74// Note the function is only implemented on the Windows platform. 75_LIBCPP_EXPORTED_FROM_ABI void __write_to_windows_console(FILE* __stream, wstring_view __view); 76# endif // _LIBCPP_HAS_NO_WIDE_CHARACTERS 77 78#endif // _WIN32 79 80#if _LIBCPP_STD_VER >= 23 81 82# ifndef _LIBCPP_HAS_NO_UNICODE 83// This is the code to transcode UTF-8 to UTF-16. This is used on 84// Windows for the native Unicode API. The code is modeled to make it 85// easier to extend to 86// 87// P2728R0 Unicode in the Library, Part 1: UTF Transcoding 88// 89// This paper is still under heavy development so it makes no sense yet 90// to strictly follow the paper. 91namespace __unicode { 92 93// The names of these concepts are modelled after P2728R0, but the 94// implementation is not. char16_t may contain 32-bits so depending on the 95// number of bits is an issue. 96# ifdef _LIBCPP_SHORT_WCHAR 97template <class _Tp> 98concept __utf16_code_unit = 99 same_as<_Tp, char16_t> 100# ifndef _LIBCPP_HAS_NO_WIDE_CHARACTERS 101 || same_as<_Tp, wchar_t> 102# endif 103 ; 104template <class _Tp> 105concept __utf32_code_unit = same_as<_Tp, char32_t>; 106# else // _LIBCPP_SHORT_WCHAR 107template <class _Tp> 108concept __utf16_code_unit = same_as<_Tp, char16_t>; 109template <class _Tp> 110concept __utf32_code_unit = 111 same_as<_Tp, char32_t> 112# ifndef _LIBCPP_HAS_NO_WIDE_CHARACTERS 113 || same_as<_Tp, wchar_t> 114# endif 115 ; 116# endif // _LIBCPP_SHORT_WCHAR 117 118// Pass by reference since an output_iterator may not be copyable. 119template <class _OutIt> 120_LIBCPP_HIDE_FROM_ABI constexpr void __encode(_OutIt&, char32_t) = delete; 121 122template <class _OutIt> 123 requires __utf16_code_unit<iter_value_t<_OutIt>> 124_LIBCPP_HIDE_FROM_ABI constexpr void __encode(_OutIt& __out_it, char32_t __value) { 125 _LIBCPP_ASSERT_UNCATEGORIZED(__is_scalar_value(__value), "an invalid unicode scalar value results in invalid UTF-16"); 126 127 if (__value < 0x10000) { 128 *__out_it++ = __value; 129 return; 130 } 131 132 __value -= 0x10000; 133 *__out_it++ = 0xd800 + (__value >> 10); 134 *__out_it++ = 0xdc00 + (__value & 0x3FF); 135} 136 137template <class _OutIt> 138 requires __utf32_code_unit<iter_value_t<_OutIt>> 139_LIBCPP_HIDE_FROM_ABI constexpr void __encode(_OutIt& __out_it, char32_t __value) { 140 _LIBCPP_ASSERT_UNCATEGORIZED(__is_scalar_value(__value), "an invalid unicode scalar value results in invalid UTF-32"); 141 *__out_it++ = __value; 142} 143 144template <class _OutIt, input_iterator _InIt> 145 requires output_iterator<_OutIt, const iter_value_t<_OutIt>&> && (!same_as<iter_value_t<_OutIt>, iter_value_t<_InIt>>) 146_LIBCPP_HIDE_FROM_ABI constexpr _OutIt __transcode(_InIt __first, _InIt __last, _OutIt __out_it) { 147 // The __code_point_view has a basic_string_view interface. 148 // When transcoding becomes part of the standard we probably want to 149 // look at smarter algorithms. 150 // For example, when processing a code point that is encoded in 151 // 1 to 3 code units in UTF-8, the result will always be encoded 152 // in 1 code unit in UTF-16 (code points that require 4 code 153 // units in UTF-8 will require 2 code units in UTF-16). 154 // 155 // Note if P2728 is accepted types like int may become valid. In that case 156 // the __code_point_view should use a span. Libc++ will remove support for 157 // char_traits<int>. 158 159 // TODO PRINT Validate with clang-tidy 160 // NOLINTNEXTLINE(bugprone-dangling-handle) 161 basic_string_view<iter_value_t<_InIt>> __data{__first, __last}; 162 __code_point_view<iter_value_t<_InIt>> __view{__data.begin(), __data.end()}; 163 while (!__view.__at_end()) 164 __unicode::__encode(__out_it, __view.__consume().__code_point); 165 return __out_it; 166} 167 168} // namespace __unicode 169 170# endif // _LIBCPP_HAS_NO_UNICODE 171 172namespace __print { 173 174// [print.fun]/2 175// Effects: If the ordinary literal encoding ([lex.charset]) is UTF-8, equivalent to: 176// vprint_unicode(stream, fmt.str, make_format_args(args...)); 177// Otherwise, equivalent to: 178// vprint_nonunicode(stream, fmt.str, make_format_args(args...)); 179// 180// Based on the compiler and its compilation flags this value is or is 181// not true. As mentioned in P2093R14 this only affects Windows. The 182// test below could also be done for 183// - GCC using __GNUC_EXECUTION_CHARSET_NAME 184// https://gcc.gnu.org/onlinedocs/cpp/Common-Predefined-Macros.html 185// - Clang using __clang_literal_encoding__ 186// https://clang.llvm.org/docs/LanguageExtensions.html#builtin-macros 187// (note at the time of writing Clang is hard-coded to UTF-8.) 188// 189 190# ifdef _LIBCPP_HAS_NO_UNICODE 191inline constexpr bool __use_unicode = false; 192# elif defined(_MSVC_EXECUTION_CHARACTER_SET) 193// This is the same test MSVC STL uses in their implementation of <print> 194// See: https://learn.microsoft.com/en-us/windows/win32/intl/code-page-identifiers 195inline constexpr bool __use_unicode = _MSVC_EXECUTION_CHARACTER_SET == 65001; 196# else 197inline constexpr bool __use_unicode = true; 198# endif 199 200_LIBCPP_HIDE_FROM_ABI inline bool __is_terminal(FILE* __stream) { 201# ifdef _WIN32 202 return std::__is_windows_terminal(__stream); 203# elif __has_include(<unistd.h>) 204 return isatty(fileno(__stream)); 205# else 206# error "Provide a way to determine whether a FILE* is a terminal" 207# endif 208} 209 210template <class = void> // TODO PRINT template or availability markup fires too eagerly (http://llvm.org/PR61563). 211_LIBCPP_HIDE_FROM_ABI inline void 212__vprint_nonunicode(FILE* __stream, string_view __fmt, format_args __args, bool __write_nl) { 213 _LIBCPP_ASSERT_UNCATEGORIZED(__stream, "__stream is a valid pointer to an output C stream"); 214 string __str = std::vformat(__fmt, __args); 215 if (__write_nl) 216 __str.push_back('\n'); 217 218 size_t __size = fwrite(__str.data(), 1, __str.size(), __stream); 219 if (__size < __str.size()) { 220 if (std::feof(__stream)) 221 std::__throw_system_error(EIO, "EOF while writing the formatted output"); 222 std::__throw_system_error(std::ferror(__stream), "failed to write formatted output"); 223 } 224} 225 226# ifndef _LIBCPP_HAS_NO_UNICODE 227 228// Note these helper functions are mainly used to aid testing. 229// On POSIX systems and Windows the output is no longer considered a 230// terminal when the output is redirected. Typically during testing the 231// output is redirected to be able to capture it. This makes it hard to 232// test this code path. 233template <class = void> // TODO PRINT template or availability markup fires too eagerly (http://llvm.org/PR61563). 234_LIBCPP_HIDE_FROM_ABI inline void 235__vprint_unicode_posix(FILE* __stream, string_view __fmt, format_args __args, bool __write_nl, bool __is_terminal) { 236 // TODO PRINT Should flush errors throw too? 237 if (__is_terminal) 238 std::fflush(__stream); 239 240 __print::__vprint_nonunicode(__stream, __fmt, __args, __write_nl); 241} 242 243# ifndef _LIBCPP_HAS_NO_WIDE_CHARACTERS 244template <class = void> // TODO PRINT template or availability markup fires too eagerly (http://llvm.org/PR61563). 245_LIBCPP_HIDE_FROM_ABI inline void 246__vprint_unicode_windows(FILE* __stream, string_view __fmt, format_args __args, bool __write_nl, bool __is_terminal) { 247 if (!__is_terminal) 248 return __print::__vprint_nonunicode(__stream, __fmt, __args, __write_nl); 249 250 // TODO PRINT Should flush errors throw too? 251 std::fflush(__stream); 252 253 string __str = std::vformat(__fmt, __args); 254 // UTF-16 uses the same number or less code units than UTF-8. 255 // However the size of the code unit is 16 bits instead of 8 bits. 256 // 257 // The buffer uses the worst-case estimate and should never resize. 258 // However when the string is large this could lead to OOM. Using a 259 // smaller size might work, but since the buffer uses a grow factor 260 // the final size might be larger when the estimate is wrong. 261 // 262 // TODO PRINT profile and improve the speed of this code. 263 __format::__retarget_buffer<wchar_t> __buffer{__str.size()}; 264 __unicode::__transcode(__str.begin(), __str.end(), __buffer.__make_output_iterator()); 265 if (__write_nl) 266 __buffer.push_back(L'\n'); 267 268 [[maybe_unused]] wstring_view __view = __buffer.__view(); 269 270 // The macro _LIBCPP_TESTING_PRINT_WRITE_TO_WINDOWS_CONSOLE_FUNCTION is used to change 271 // the behavior in the test. This is not part of the public API. 272# ifdef _LIBCPP_TESTING_PRINT_WRITE_TO_WINDOWS_CONSOLE_FUNCTION 273 _LIBCPP_TESTING_PRINT_WRITE_TO_WINDOWS_CONSOLE_FUNCTION(__stream, __view); 274# elif defined(_WIN32) 275 std::__write_to_windows_console(__stream, __view); 276# else 277 std::__throw_runtime_error("No defintion of _LIBCPP_TESTING_PRINT_WRITE_TO_WINDOWS_CONSOLE_FUNCTION and " 278 "__write_to_windows_console is not available."); 279# endif 280} 281# endif // _LIBCPP_HAS_NO_WIDE_CHARACTERS 282 283template <class = void> // TODO PRINT template or availability markup fires too eagerly (http://llvm.org/PR61563). 284_LIBCPP_HIDE_FROM_ABI inline void 285__vprint_unicode([[maybe_unused]] FILE* __stream, 286 [[maybe_unused]] string_view __fmt, 287 [[maybe_unused]] format_args __args, 288 [[maybe_unused]] bool __write_nl) { 289 _LIBCPP_ASSERT_UNCATEGORIZED(__stream, "__stream is a valid pointer to an output C stream"); 290 291 // [print.fun] 292 // 7 - Effects: If stream refers to a terminal capable of displaying 293 // Unicode, writes out to the terminal using the native Unicode 294 // API; if out contains invalid code units, the behavior is 295 // undefined and implementations are encouraged to diagnose it. 296 // Otherwise writes out to stream unchanged. If the native 297 // Unicode API is used, the function flushes stream before 298 // writing out. 299 // 8 - Throws: Any exception thrown by the call to vformat 300 // ([format.err.report]). system_error if writing to the terminal 301 // or stream fails. May throw bad_alloc. 302 // 9 - Recommended practice: If invoking the native Unicode API 303 // requires transcoding, implementations should substitute 304 // invalid code units with U+FFFD replacement character per the 305 // Unicode Standard, Chapter 3.9 U+FFFD Substitution in 306 // Conversion. 307 308 // On non-Windows platforms the Unicode API is the normal file I/O API 309 // so there the call can be forwarded to the non_unicode API. On 310 // Windows there is a different API. This API requires transcoding. 311 312# ifndef _WIN32 313 __print::__vprint_unicode_posix(__stream, __fmt, __args, __write_nl, __print::__is_terminal(__stream)); 314# elif !defined(_LIBCPP_HAS_NO_WIDE_CHARACTERS) 315 __print::__vprint_unicode_windows(__stream, __fmt, __args, __write_nl, __print::__is_terminal(__stream)); 316# else 317# error "Windows builds with wchar_t disabled are not supported." 318# endif 319} 320 321# endif // _LIBCPP_HAS_NO_UNICODE 322 323} // namespace __print 324 325template <class... _Args> 326_LIBCPP_HIDE_FROM_ABI void print(FILE* __stream, format_string<_Args...> __fmt, _Args&&... __args) { 327# ifndef _LIBCPP_HAS_NO_UNICODE 328 if constexpr (__print::__use_unicode) 329 __print::__vprint_unicode(__stream, __fmt.get(), std::make_format_args(__args...), false); 330 else 331 __print::__vprint_nonunicode(__stream, __fmt.get(), std::make_format_args(__args...), false); 332# else // _LIBCPP_HAS_NO_UNICODE 333 __print::__vprint_nonunicode(__stream, __fmt.get(), std::make_format_args(__args...), false); 334# endif // _LIBCPP_HAS_NO_UNICODE 335} 336 337template <class... _Args> 338_LIBCPP_HIDE_FROM_ABI void print(format_string<_Args...> __fmt, _Args&&... __args) { 339 std::print(stdout, __fmt, std::forward<_Args>(__args)...); 340} 341 342template <class... _Args> 343_LIBCPP_HIDE_FROM_ABI void println(FILE* __stream, format_string<_Args...> __fmt, _Args&&... __args) { 344# ifndef _LIBCPP_HAS_NO_UNICODE 345 // Note the wording in the Standard is inefficient. The output of 346 // std::format is a std::string which is then copied. This solution 347 // just appends a newline at the end of the output. 348 if constexpr (__print::__use_unicode) 349 __print::__vprint_unicode(__stream, __fmt.get(), std::make_format_args(__args...), true); 350 else 351 __print::__vprint_nonunicode(__stream, __fmt.get(), std::make_format_args(__args...), true); 352# else // _LIBCPP_HAS_NO_UNICODE 353 __print::__vprint_nonunicode(__stream, __fmt.get(), std::make_format_args(__args...), true); 354# endif // _LIBCPP_HAS_NO_UNICODE 355} 356 357template <class... _Args> 358_LIBCPP_HIDE_FROM_ABI void println(format_string<_Args...> __fmt, _Args&&... __args) { 359 std::println(stdout, __fmt, std::forward<_Args>(__args)...); 360} 361 362# ifndef _LIBCPP_HAS_NO_UNICODE 363template <class = void> // TODO PRINT template or availability markup fires too eagerly (http://llvm.org/PR61563). 364_LIBCPP_HIDE_FROM_ABI inline void vprint_unicode(FILE* __stream, string_view __fmt, format_args __args) { 365 __print::__vprint_unicode(__stream, __fmt, __args, false); 366} 367 368template <class = void> // TODO PRINT template or availability markup fires too eagerly (http://llvm.org/PR61563). 369_LIBCPP_HIDE_FROM_ABI inline void vprint_unicode(string_view __fmt, format_args __args) { 370 std::vprint_unicode(stdout, __fmt, __args); 371} 372 373# endif // _LIBCPP_HAS_NO_UNICODE 374 375template <class = void> // TODO PRINT template or availability markup fires too eagerly (http://llvm.org/PR61563). 376_LIBCPP_HIDE_FROM_ABI inline void vprint_nonunicode(FILE* __stream, string_view __fmt, format_args __args) { 377 __print::__vprint_nonunicode(__stream, __fmt, __args, false); 378} 379 380template <class = void> // TODO PRINT template or availability markup fires too eagerly (http://llvm.org/PR61563). 381_LIBCPP_HIDE_FROM_ABI inline void vprint_nonunicode(string_view __fmt, format_args __args) { 382 std::vprint_nonunicode(stdout, __fmt, __args); 383} 384 385#endif // _LIBCPP_STD_VER >= 23 386 387_LIBCPP_END_NAMESPACE_STD 388 389#endif // _LIBCPP_PRINT 390