tests/test-mbrtoc16.c

   1 /* Test of conversion of multibyte character to 16-bit wide characters.
   2    Copyright (C) 2008-2024 Free Software Foundation, Inc.
   3
   4    This program is free software: you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation, either version 3 of the License, or
   7    (at your option) any later version.
   8
   9    This program 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
  15    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
  16
  17 /* Written by Bruno Haible <bruno@clisp.org>, 2008.  */
  18
  19 #include <config.h>
  20
  21 #include <uchar.h>
  22
  23 #include "signature.h"
  24 SIGNATURE_CHECK (mbrtoc16, size_t,
  25                  (char16_t *, const char *, size_t, mbstate_t *));
  26
  27 #include <locale.h>
  28 #include <stdio.h>
  29 #include <stdlib.h>
  30 #include <string.h>
  31 #include <wchar.h>
  32
  33 #include "macros.h"
  34
  35 int
  36 main (int argc, char *argv[])
  37 {
  38   mbstate_t state;
  39   char16_t wc;
  40   size_t ret;
  41
  42   /* configure should already have checked that the locale is supported.  */
  43   if (setlocale (LC_ALL, "") == NULL)
  44     return 1;
  45
  46   /* Test zero-length input.  */
  47   {
  48     memset (&state, '\0', sizeof (mbstate_t));
  49     wc = (char16_t) 0xBADF;
  50     ret = mbrtoc16 (&wc, "x", 0, &state);
  51     ASSERT (ret == (size_t)(-2));
  52     ASSERT (mbsinit (&state));
  53   }
  54
  55   /* Test NUL byte input.  */
  56   {
  57     memset (&state, '\0', sizeof (mbstate_t));
  58     wc = (char16_t) 0xBADF;
  59     ret = mbrtoc16 (&wc, "", 1, &state);
  60     ASSERT (ret == 0);
  61     ASSERT (wc == 0);
  62     ASSERT (mbsinit (&state));
  63     ret = mbrtoc16 (NULL, "", 1, &state);
  64     ASSERT (ret == 0);
  65     ASSERT (mbsinit (&state));
  66   }
  67
  68   /* Test single-byte input.  */
  69   {
  70     int c;
  71     char buf[1];
  72
  73     memset (&state, '\0', sizeof (mbstate_t));
  74     for (c = 0; c < 0x100; c++)
  75       switch (c)
  76         {
  77         case '\t': case '\v': case '\f':
  78         case ' ': case '!': case '"': case '#': case '%':
  79         case '&': case '\'': case '(': case ')': case '*':
  80         case '+': case ',': case '-': case '.': case '/':
  81         case '0': case '1': case '2': case '3': case '4':
  82         case '5': case '6': case '7': case '8': case '9':
  83         case ':': case ';': case '<': case '=': case '>':
  84         case '?':
  85         case 'A': case 'B': case 'C': case 'D': case 'E':
  86         case 'F': case 'G': case 'H': case 'I': case 'J':
  87         case 'K': case 'L': case 'M': case 'N': case 'O':
  88         case 'P': case 'Q': case 'R': case 'S': case 'T':
  89         case 'U': case 'V': case 'W': case 'X': case 'Y':
  90         case 'Z':
  91         case '[': case '\\': case ']': case '^': case '_':
  92         case 'a': case 'b': case 'c': case 'd': case 'e':
  93         case 'f': case 'g': case 'h': case 'i': case 'j':
  94         case 'k': case 'l': case 'm': case 'n': case 'o':
  95         case 'p': case 'q': case 'r': case 's': case 't':
  96         case 'u': case 'v': case 'w': case 'x': case 'y':
  97         case 'z': case '{': case '|': case '}': case '~':
  98           /* c is in the ISO C "basic character set".  */
  99           ASSERT (c < 0x80);
 100           /* c is an ASCII character.  */
 101           buf[0] = c;
 102
 103           wc = (char16_t) 0xBADF;
 104           ret = mbrtoc16 (&wc, buf, 1, &state);
 105           ASSERT (ret == 1);
 106           ASSERT (wc == c);
 107           ASSERT (mbsinit (&state));
 108
 109           ret = mbrtoc16 (NULL, buf, 1, &state);
 110           ASSERT (ret == 1);
 111           ASSERT (mbsinit (&state));
 112
 113           break;
 114         default:
 115           break;
 116         }
 117   }
 118
 119   /* Test special calling convention, passing a NULL pointer.  */
 120   {
 121     memset (&state, '\0', sizeof (mbstate_t));
 122     wc = (char16_t) 0xBADF;
 123     ret = mbrtoc16 (&wc, NULL, 5, &state);
 124     ASSERT (ret == 0);
 125     ASSERT (wc == (char16_t) 0xBADF);
 126     ASSERT (mbsinit (&state));
 127   }
 128
 129 #ifdef __ANDROID__
 130   /* On Android ≥ 5.0, the default locale is the "C.UTF-8" locale, not the
 131      "C" locale.  Furthermore, when you attempt to set the "C" or "POSIX"
 132      locale via setlocale(), what you get is a "C" locale with UTF-8 encoding,
 133      that is, effectively the "C.UTF-8" locale.  */
 134   if (argc > 1 && strcmp (argv[1], "1") == 0 && MB_CUR_MAX > 1)
 135     argv[1] = "3";
 136 #endif
 137
 138   if (argc > 1)
 139     switch (argv[1][0])
 140       {
 141       case '1':
 142         /* C or POSIX locale.  */
 143         {
 144           int c;
 145           char buf[1];
 146
 147           memset (&state, '\0', sizeof (mbstate_t));
 148           for (c = 0; c < 0x100; c++)
 149             if (c != 0)
 150               {
 151                 /* We are testing all nonnull bytes.  */
 152                 buf[0] = c;
 153
 154                 wc = (char16_t) 0xBADF;
 155                 ret = mbrtoc16 (&wc, buf, 1, &state);
 156                 /* POSIX:2018 says regarding mbrtowc: "In the POSIX locale an
 157                    [EILSEQ] error cannot occur since all byte values are valid
 158                    characters."  It is reasonable to expect mbrtoc16 to behave
 159                    in the same way.  */
 160                 ASSERT (ret == 1);
 161                 if (c < 0x80)
 162                   /* c is an ASCII character.  */
 163                   ASSERT (wc == c);
 164                 else
 165                   /* On most platforms, the bytes 0x80..0xFF map to U+0080..U+00FF.
 166                      But on musl libc, the bytes 0x80..0xFF map to U+DF80..U+DFFF.  */
 167                   ASSERT (wc == (btoc32 (c) == 0xDF00 + c ? btoc32 (c) : c));
 168                 ASSERT (mbsinit (&state));
 169
 170                 ret = mbrtoc16 (NULL, buf, 1, &state);
 171                 ASSERT (ret == 1);
 172                 ASSERT (mbsinit (&state));
 173               }
 174         }
 175         return test_exit_status;
 176
 177       case '2':
 178         /* Locale encoding is ISO-8859-1 or ISO-8859-15.  */
 179         {
 180           char input[] = "B\374\337er"; /* "Büßer" */
 181           memset (&state, '\0', sizeof (mbstate_t));
 182
 183           wc = (char16_t) 0xBADF;
 184           ret = mbrtoc16 (&wc, input, 1, &state);
 185           ASSERT (ret == 1);
 186           ASSERT (wc == 'B');
 187           ASSERT (mbsinit (&state));
 188           input[0] = '\0';
 189
 190           wc = (char16_t) 0xBADF;
 191           ret = mbrtoc16 (&wc, input + 1, 1, &state);
 192           ASSERT (ret == 1);
 193           ASSERT (c32tob (wc) == (unsigned char) '\374');
 194           ASSERT (wc == 0x00FC); /* expect UTF-16 encoding */
 195           ASSERT (mbsinit (&state));
 196           input[1] = '\0';
 197
 198           /* Test support of NULL first argument.  */
 199           ret = mbrtoc16 (NULL, input + 2, 3, &state);
 200           ASSERT (ret == 1);
 201           ASSERT (mbsinit (&state));
 202
 203           wc = (char16_t) 0xBADF;
 204           ret = mbrtoc16 (&wc, input + 2, 3, &state);
 205           ASSERT (ret == 1);
 206           ASSERT (c32tob (wc) == (unsigned char) '\337');
 207           ASSERT (wc == 0x00DF); /* expect UTF-16 encoding */
 208           ASSERT (mbsinit (&state));
 209           input[2] = '\0';
 210
 211           wc = (char16_t) 0xBADF;
 212           ret = mbrtoc16 (&wc, input + 3, 2, &state);
 213           ASSERT (ret == 1);
 214           ASSERT (wc == 'e');
 215           ASSERT (mbsinit (&state));
 216           input[3] = '\0';
 217
 218           wc = (char16_t) 0xBADF;
 219           ret = mbrtoc16 (&wc, input + 4, 1, &state);
 220           ASSERT (ret == 1);
 221           ASSERT (wc == 'r');
 222           ASSERT (mbsinit (&state));
 223         }
 224         return test_exit_status;
 225
 226       case '3':
 227         /* Locale encoding is UTF-8.  */
 228         {
 229           char input[] = "s\303\274\303\237\360\237\230\213!"; /* "süß😋!" */
 230           memset (&state, '\0', sizeof (mbstate_t));
 231
 232           wc = (char16_t) 0xBADF;
 233           ret = mbrtoc16 (&wc, input, 1, &state);
 234           ASSERT (ret == 1);
 235           ASSERT (wc == 's');
 236           ASSERT (mbsinit (&state));
 237           input[0] = '\0';
 238
 239           wc = (char16_t) 0xBADF;
 240           ret = mbrtoc16 (&wc, input + 1, 1, &state);
 241           ASSERT (ret == (size_t)(-2));
 242           ASSERT (wc == (char16_t) 0xBADF);
 243           #if 0
 244           ASSERT (!mbsinit (&state));
 245           #endif
 246           input[1] = '\0';
 247
 248           wc = (char16_t) 0xBADF;
 249           ret = mbrtoc16 (&wc, input + 2, 7, &state);
 250           ASSERT (ret == 1);
 251           ASSERT (c32tob (wc) == EOF);
 252           ASSERT (wc == 0x00FC); /* expect UTF-16 encoding */
 253           ASSERT (mbsinit (&state));
 254           input[2] = '\0';
 255
 256           /* Test support of NULL first argument.  */
 257           ret = mbrtoc16 (NULL, input + 3, 6, &state);
 258           ASSERT (ret == 2);
 259           ASSERT (mbsinit (&state));
 260
 261           wc = (char16_t) 0xBADF;
 262           ret = mbrtoc16 (&wc, input + 3, 6, &state);
 263           ASSERT (ret == 2);
 264           ASSERT (c32tob (wc) == EOF);
 265           ASSERT (wc == 0x00DF); /* expect UTF-16 encoding */
 266           ASSERT (mbsinit (&state));
 267           input[3] = '\0';
 268           input[4] = '\0';
 269
 270           /* Test support of NULL first argument.  */
 271           ret = mbrtoc16 (NULL, input + 5, 4, &state);
 272           ASSERT (ret == 4);
 273           ret = mbrtoc16 (NULL, input + 9, 0, &state);
 274           ASSERT (ret == (size_t)(-3));
 275           ASSERT (mbsinit (&state));
 276
 277           wc = (char16_t) 0xBADF;
 278           ret = mbrtoc16 (&wc, input + 5, 4, &state);
 279           ASSERT (ret == 4);
 280           ASSERT (c32tob (wc) == EOF);
 281           ASSERT (wc == 0xD83D); /* expect UTF-16 encoding */
 282           input[5] = '\0';
 283           input[6] = '\0';
 284           input[7] = '\0';
 285           input[8] = '\0';
 286           wc = (char16_t) 0xBADF;
 287           ret = mbrtoc16 (&wc, input + 9, 0, &state);
 288           ASSERT (ret == (size_t)(-3));
 289           ASSERT (c32tob (wc) == EOF);
 290           ASSERT (wc == 0xDE0B); /* expect UTF-16 encoding */
 291           ASSERT (mbsinit (&state));
 292
 293           wc = (char16_t) 0xBADF;
 294           ret = mbrtoc16 (&wc, input + 9, 1, &state);
 295           ASSERT (ret == 1);
 296           ASSERT (wc == '!');
 297           ASSERT (mbsinit (&state));
 298         }
 299         return test_exit_status;
 300
 301       case '4':
 302         /* Locale encoding is EUC-JP.  */
 303         {
 304           char input[] = "<\306\374\313\334\270\354>"; /* "<日本語>" */
 305           memset (&state, '\0', sizeof (mbstate_t));
 306
 307           wc = (char16_t) 0xBADF;
 308           ret = mbrtoc16 (&wc, input, 1, &state);
 309           ASSERT (ret == 1);
 310           ASSERT (wc == '<');
 311           ASSERT (mbsinit (&state));
 312           input[0] = '\0';
 313
 314           wc = (char16_t) 0xBADF;
 315           ret = mbrtoc16 (&wc, input + 1, 2, &state);
 316           ASSERT (ret == 2);
 317           ASSERT (c32tob (wc) == EOF);
 318           ASSERT (wc == 0x65E5); /* expect UTF-16 encoding */
 319           ASSERT (mbsinit (&state));
 320           input[1] = '\0';
 321           input[2] = '\0';
 322
 323           wc = (char16_t) 0xBADF;
 324           ret = mbrtoc16 (&wc, input + 3, 1, &state);
 325           ASSERT (ret == (size_t)(-2));
 326           ASSERT (wc == (char16_t) 0xBADF);
 327           #if 0
 328           ASSERT (!mbsinit (&state));
 329           #endif
 330           input[3] = '\0';
 331
 332           wc = (char16_t) 0xBADF;
 333           ret = mbrtoc16 (&wc, input + 4, 4, &state);
 334           ASSERT (ret == 1);
 335           ASSERT (c32tob (wc) == EOF);
 336           ASSERT (wc == 0x672C); /* expect UTF-16 encoding */
 337           ASSERT (mbsinit (&state));
 338           input[4] = '\0';
 339
 340           /* Test support of NULL first argument.  */
 341           ret = mbrtoc16 (NULL, input + 5, 3, &state);
 342           ASSERT (ret == 2);
 343           ASSERT (mbsinit (&state));
 344
 345           wc = (char16_t) 0xBADF;
 346           ret = mbrtoc16 (&wc, input + 5, 3, &state);
 347           ASSERT (ret == 2);
 348           ASSERT (c32tob (wc) == EOF);
 349           ASSERT (wc == 0x8A9E); /* expect UTF-16 encoding */
 350           ASSERT (mbsinit (&state));
 351           input[5] = '\0';
 352           input[6] = '\0';
 353
 354           wc = (char16_t) 0xBADF;
 355           ret = mbrtoc16 (&wc, input + 7, 1, &state);
 356           ASSERT (ret == 1);
 357           ASSERT (wc == '>');
 358           ASSERT (mbsinit (&state));
 359         }
 360         return test_exit_status;
 361
 362       case '5':
 363         /* Locale encoding is GB18030.  */
 364         #if (defined __GLIBC__ && __GLIBC__ == 2 && __GLIBC_MINOR__ >= 13 && __GLIBC_MINOR__ <= 15) || (GL_CHAR32_T_IS_UNICODE && (defined __FreeBSD__ || defined __NetBSD__ || defined __sun))
 365         if (test_exit_status != EXIT_SUCCESS)
 366           return test_exit_status;
 367         fputs ("Skipping test: The GB18030 converter in this system's iconv is broken.\n", stderr);
 368         return 77;
 369         #endif
 370         {
 371           char input[] = "s\250\271\201\060\211\070\224\071\375\067!"; /* "süß😋!" */
 372           memset (&state, '\0', sizeof (mbstate_t));
 373
 374           wc = (char16_t) 0xBADF;
 375           ret = mbrtoc16 (&wc, input, 1, &state);
 376           ASSERT (ret == 1);
 377           ASSERT (wc == 's');
 378           ASSERT (mbsinit (&state));
 379           input[0] = '\0';
 380
 381           wc = (char16_t) 0xBADF;
 382           ret = mbrtoc16 (&wc, input + 1, 1, &state);
 383           ASSERT (ret == (size_t)(-2));
 384           ASSERT (wc == (char16_t) 0xBADF);
 385           #if 0
 386           ASSERT (!mbsinit (&state));
 387           #endif
 388           input[1] = '\0';
 389
 390           wc = (char16_t) 0xBADF;
 391           ret = mbrtoc16 (&wc, input + 2, 9, &state);
 392           ASSERT (ret == 1);
 393           ASSERT (c32tob (wc) == EOF);
 394           ASSERT (wc == 0x00FC); /* expect UTF-16 encoding */
 395           ASSERT (mbsinit (&state));
 396           input[2] = '\0';
 397
 398           /* Test support of NULL first argument.  */
 399           ret = mbrtoc16 (NULL, input + 3, 8, &state);
 400           ASSERT (ret == 4);
 401           ASSERT (mbsinit (&state));
 402
 403           wc = (char16_t) 0xBADF;
 404           ret = mbrtoc16 (&wc, input + 3, 8, &state);
 405           ASSERT (ret == 4);
 406           ASSERT (c32tob (wc) == EOF);
 407           ASSERT (wc == 0x00DF); /* expect UTF-16 encoding */
 408           ASSERT (mbsinit (&state));
 409           input[3] = '\0';
 410           input[4] = '\0';
 411           input[5] = '\0';
 412           input[6] = '\0';
 413
 414           /* Test support of NULL first argument.  */
 415           ret = mbrtoc16 (NULL, input + 7, 4, &state);
 416           ASSERT (ret == 4);
 417           ret = mbrtoc16 (NULL, input + 11, 0, &state);
 418           ASSERT (ret == (size_t)(-3));
 419           ASSERT (mbsinit (&state));
 420
 421           wc = (char16_t) 0xBADF;
 422           ret = mbrtoc16 (&wc, input + 7, 4, &state);
 423           ASSERT (ret == 4);
 424           ASSERT (c32tob (wc) == EOF);
 425           ASSERT (wc == 0xD83D); /* expect UTF-16 encoding */
 426           input[7] = '\0';
 427           input[8] = '\0';
 428           input[9] = '\0';
 429           input[10] = '\0';
 430           wc = (char16_t) 0xBADF;
 431           ret = mbrtoc16 (&wc, input + 11, 0, &state);
 432           ASSERT (ret == (size_t)(-3));
 433           ASSERT (c32tob (wc) == EOF);
 434           ASSERT (wc == 0xDE0B); /* expect UTF-16 encoding */
 435           ASSERT (mbsinit (&state));
 436
 437           wc = (char16_t) 0xBADF;
 438           ret = mbrtoc16 (&wc, input + 11, 1, &state);
 439           ASSERT (ret == 1);
 440           ASSERT (wc == '!');
 441           ASSERT (mbsinit (&state));
 442         }
 443         return test_exit_status;
 444       }
 445
 446   return 1;
 447 }