libstdc++-v3/testsuite/22_locale/codecvt/codecvt_utf16/misaligned.cc

   1 // Copyright (C) 2017-2025 Free Software Foundation, Inc.
   2 //
   3 // This file is part of the GNU ISO C++ Library.  This library is free
   4 // software; you can redistribute it and/or modify it under the
   5 // terms of the GNU General Public License as published by the
   6 // Free Software Foundation; either version 3, or (at your option)
   7 // any later version.
   8
   9 // This library is distributed in the hope that it will be useful,
  10 // but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12 // GNU General Public License for more details.
  13
  14 // You should have received a copy of the GNU General Public License along
  15 // with this library; see the file COPYING3.  If not see
  16 // <http://www.gnu.org/licenses/>.
  17
  18 // { dg-do run { target c++11 } }
  19
  20 #include <locale>
  21 #include <codecvt>
  22 #include <testsuite_hooks.h>
  23
  24 using std::codecvt_base;
  25 using std::codecvt_mode;
  26 using std::codecvt_utf16;
  27 using std::wstring_convert;
  28 using std::mbstate_t;
  29
  30 constexpr codecvt_mode
  31 operator|(codecvt_mode m1, codecvt_mode m2)
  32 {
  33   using underlying = std::underlying_type<codecvt_mode>::type;
  34   return static_cast<codecvt_mode>(static_cast<underlying>(m1) | m2);
  35 }
  36
  37 // Read/write UTF-16 code units from data not correctly aligned for char16_t
  38
  39 void
  40 test01()
  41 {
  42   mbstate_t st;
  43   constexpr codecvt_mode m = std::consume_header|std::generate_header;
  44   codecvt_utf16<char16_t, 0x10FFFF, m> conv;
  45   const char src[] = "-\xFE\xFF\0\x61\xAB\xCD";
  46   const char* const src_end = src + 7;
  47
  48   int len = conv.length(st, src + 1, src_end, 1);
  49   VERIFY( len == 4 );
  50   len = conv.length(st, src + 1, src_end, 2);
  51   VERIFY( len == 6 );
  52
  53   char16_t dst[2];
  54   char16_t* const dst_end = dst + 2;
  55   char16_t* dst_next;
  56   const char* src_cnext;
  57   auto res = conv.in(st, src + 1, src_end, src_cnext, dst, dst_end, dst_next);
  58   VERIFY( res == codecvt_base::ok );
  59   VERIFY( dst[0] == 0x0061 );
  60   VERIFY( dst[1] == 0xabcd );
  61   VERIFY( src_cnext == src_end );
  62   VERIFY( dst_next == dst_end );
  63
  64   char out[sizeof(src)] = { src[0] };
  65   char* const out_end = out + 7;
  66   char* out_next;
  67   const char16_t* dst_cnext;
  68   res = conv.out(st, dst, dst_end, dst_cnext, out + 1, out_end, out_next);
  69   VERIFY( res == codecvt_base::ok );
  70   VERIFY( out_next == out_end );
  71   VERIFY( dst_cnext == dst_end );
  72   VERIFY( out[1] == src[1] );
  73   VERIFY( out[2] == src[2] );
  74   VERIFY( out[3] == src[3] );
  75   VERIFY( out[4] == src[4] );
  76   VERIFY( out[5] == src[5] );
  77   VERIFY( out[6] == src[6] );
  78
  79   codecvt_utf16<char16_t, 0x10FFFF, m|std::little_endian> conv_le;
  80
  81   len = conv_le.length(st, src + 1, src_end, 1);
  82   VERIFY( len == 4 );
  83   len = conv_le.length(st, src + 1, src_end, 2);
  84   VERIFY( len == 6 );
  85
  86   res = conv_le.in(st, src + 1, src_end, src_cnext, dst, dst_end, dst_next);
  87   VERIFY( res == codecvt_base::ok );
  88   VERIFY( dst[0] == 0x0061 );
  89   VERIFY( dst[1] == 0xabcd );
  90   VERIFY( src_cnext == src_end );
  91   VERIFY( dst_next == dst_end );
  92
  93   res = conv_le.out(st, dst, dst_end, dst_cnext, out + 1, out_end, out_next);
  94   VERIFY( res == codecvt_base::ok );
  95   VERIFY( out_next == out_end );
  96   VERIFY( dst_cnext == dst_end );
  97   VERIFY( out[1] == src[2] );
  98   VERIFY( out[2] == src[1] );
  99   VERIFY( out[3] == src[4] );
 100   VERIFY( out[4] == src[3] );
 101   VERIFY( out[5] == src[6] );
 102   VERIFY( out[6] == src[5] );
 103 }
 104
 105 void
 106 test02()
 107 {
 108   mbstate_t st;
 109   constexpr codecvt_mode m = std::consume_header|std::generate_header;
 110   codecvt_utf16<char32_t, 0x10FFFF, m> conv;
 111   const char src[] = "-\xFE\xFF\0\x61\xAB\xCD\xD8\x08\xDF\x45";
 112   const char* const src_end = src + 11;
 113
 114   int len = conv.length(st, src + 1, src_end, 1);
 115   VERIFY( len == 4 );
 116   len = conv.length(st, src + 1, src_end, 2);
 117   VERIFY( len == 6 );
 118   len = conv.length(st, src + 1, src_end, -1ul);
 119   VERIFY( len == 10 );
 120
 121   char32_t dst[3];
 122   char32_t* const dst_end = dst + 3;
 123   char32_t* dst_next;
 124   const char* src_cnext;
 125   auto res = conv.in(st, src + 1, src_end, src_cnext, dst, dst_end, dst_next);
 126   VERIFY( res == codecvt_base::ok );
 127   VERIFY( dst[0] == 0x0061 );
 128   VERIFY( dst[1] == 0xabcd );
 129   VERIFY( dst[2] == 0x012345 );
 130   VERIFY( src_cnext == src_end );
 131   VERIFY( dst_next == dst_end );
 132
 133   char out[sizeof(src)] = { src[0] };
 134   char* const out_end = out + 11;
 135   char* out_next;
 136   const char32_t* dst_cnext;
 137   res = conv.out(st, dst, dst_end, dst_cnext, out + 1, out_end, out_next);
 138   VERIFY( res == codecvt_base::ok );
 139   VERIFY( out_next == out_end );
 140   VERIFY( dst_cnext == dst_end );
 141   VERIFY( out[1] == src[1] );
 142   VERIFY( out[2] == src[2] );
 143   VERIFY( out[3] == src[3] );
 144   VERIFY( out[4] == src[4] );
 145   VERIFY( out[5] == src[5] );
 146   VERIFY( out[6] == src[6] );
 147   VERIFY( out[7] == src[7] );
 148   VERIFY( out[8] == src[8] );
 149   VERIFY( out[9] == src[9] );
 150   VERIFY( out[10] == src[10] );
 151
 152   codecvt_utf16<char32_t, 0x10FFFF, m|std::little_endian> conv_le;
 153
 154   len = conv_le.length(st, src + 1, src_end, 1);
 155   VERIFY( len == 4 );
 156   len = conv_le.length(st, src + 1, src_end, 2);
 157   VERIFY( len == 6 );
 158   len = conv.length(st, src + 1, src_end, -1ul);
 159   VERIFY( len == 10 );
 160
 161   res = conv_le.in(st, src + 1, src_end, src_cnext, dst, dst_end, dst_next);
 162   VERIFY( res == codecvt_base::ok );
 163   VERIFY( dst[0] == 0x0061 );
 164   VERIFY( dst[1] == 0xabcd );
 165   VERIFY( dst[2] == 0x012345 );
 166   VERIFY( src_cnext == src_end );
 167   VERIFY( dst_next == dst_end );
 168
 169   res = conv_le.out(st, dst, dst_end, dst_cnext, out + 1, out_end, out_next);
 170   VERIFY( res == codecvt_base::ok );
 171   VERIFY( out_next == out_end );
 172   VERIFY( dst_cnext == dst_end );
 173   VERIFY( out[1] == src[2] );
 174   VERIFY( out[2] == src[1] );
 175   VERIFY( out[3] == src[4] );
 176   VERIFY( out[4] == src[3] );
 177   VERIFY( out[5] == src[6] );
 178   VERIFY( out[6] == src[5] );
 179   VERIFY( out[7] == src[8] );
 180   VERIFY( out[8] == src[7] );
 181   VERIFY( out[9] == src[10] );
 182   VERIFY( out[10] == src[9] );
 183 }
 184
 185 void
 186 test03()
 187 {
 188 #ifdef _GLIBCXX_USE_WCHAR_T
 189   mbstate_t st;
 190   constexpr codecvt_mode m = std::consume_header|std::generate_header;
 191   codecvt_utf16<wchar_t, 0x10FFFF, m> conv;
 192   const char src[] = "-\xFE\xFF\0\x61\xAB\xCD\xD8\x08\xDF\x45";
 193   const size_t in_len = sizeof(wchar_t) == 4 ? 11 : 7;
 194   const size_t out_len = sizeof(wchar_t) == 4 ? 3 : 2;
 195   const char* const src_end = src + in_len;
 196
 197   int len = conv.length(st, src + 1, src_end, 1);
 198   VERIFY( len == 4 );
 199   len = conv.length(st, src + 1, src_end, 2);
 200   VERIFY( len == 6 );
 201   if (sizeof(wchar_t) == 4)
 202   {
 203     len = conv.length(st, src + 1, src_end, -1ul);
 204     VERIFY( len == 10 );
 205   }
 206
 207   wchar_t dst[out_len];
 208   wchar_t* const dst_end = dst + out_len;
 209   wchar_t* dst_next;
 210   const char* src_cnext;
 211   auto res = conv.in(st, src + 1, src_end, src_cnext, dst, dst_end, dst_next);
 212   VERIFY( res == codecvt_base::ok );
 213   VERIFY( dst[0] == 0x0061 );
 214   VERIFY( dst[1] == 0xabcd );
 215   if (sizeof(wchar_t) == 4)
 216     VERIFY( dst[2] == 0x012345 );
 217   VERIFY( src_cnext == src_end );
 218   VERIFY( dst_next == dst_end );
 219
 220   char out[sizeof(src)] = { src[0] };
 221   char* const out_end = out + in_len;
 222   char* out_next;
 223   const wchar_t* dst_cnext;
 224   res = conv.out(st, dst, dst_end, dst_cnext, out + 1, out_end, out_next);
 225   VERIFY( res == codecvt_base::ok );
 226   VERIFY( out_next == out_end );
 227   VERIFY( dst_cnext == dst_end );
 228   VERIFY( out[1] == src[1] );
 229   VERIFY( out[2] == src[2] );
 230   VERIFY( out[3] == src[3] );
 231   VERIFY( out[4] == src[4] );
 232   VERIFY( out[5] == src[5] );
 233   VERIFY( out[6] == src[6] );
 234   if (sizeof(wchar_t) == 4)
 235   {
 236     VERIFY( out[7] == src[7] );
 237     VERIFY( out[8] == src[8] );
 238     VERIFY( out[9] == src[9] );
 239     VERIFY( out[10] == src[10] );
 240   }
 241
 242   codecvt_utf16<wchar_t, 0x10FFFF, m|std::little_endian> conv_le;
 243
 244   len = conv_le.length(st, src + 1, src_end, 1);
 245   VERIFY( len == 4 );
 246   len = conv_le.length(st, src + 1, src_end, 2);
 247   VERIFY( len == 6 );
 248   if (sizeof(wchar_t) == 4)
 249   {
 250     len = conv.length(st, src + 1, src_end, -1ul);
 251     VERIFY( len == 10 );
 252   }
 253
 254   res = conv_le.in(st, src + 1, src_end, src_cnext, dst, dst_end, dst_next);
 255   VERIFY( res == codecvt_base::ok );
 256   VERIFY( dst[0] == 0x0061 );
 257   VERIFY( dst[1] == 0xabcd );
 258   if (sizeof(wchar_t) == 4)
 259     VERIFY( dst[2] == 0x012345 );
 260   VERIFY( src_cnext == src_end );
 261   VERIFY( dst_next == dst_end );
 262
 263   res = conv_le.out(st, dst, dst_end, dst_cnext, out + 1, out_end, out_next);
 264   VERIFY( res == codecvt_base::ok );
 265   VERIFY( out_next == out_end );
 266   VERIFY( dst_cnext == dst_end );
 267   VERIFY( out[1] == src[2] );
 268   VERIFY( out[2] == src[1] );
 269   VERIFY( out[3] == src[4] );
 270   VERIFY( out[4] == src[3] );
 271   VERIFY( out[5] == src[6] );
 272   VERIFY( out[6] == src[5] );
 273   if (sizeof(wchar_t) == 4)
 274   {
 275     VERIFY( out[7] == src[8] );
 276     VERIFY( out[8] == src[7] );
 277     VERIFY( out[9] == src[10] );
 278     VERIFY( out[10] == src[9] );
 279   }
 280 #endif
 281 }
 282
 283 int
 284 main()
 285 {
 286   test01();
 287   test02();
 288   test03();
 289 }