tests/test-mbrtowc.c

   1 /* Test of conversion of multibyte character to wide character.
   2    Copyright (C) 2008-2025 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 <wchar.h>
  22
  23 #include "signature.h"
  24 SIGNATURE_CHECK (mbrtowc, size_t, (wchar_t *, char const *, size_t,
  25                                    mbstate_t *));
  26
  27 #include <locale.h>
  28 #include <stdio.h>
  29 #include <stdlib.h>
  30 #include <string.h>
  31
  32 #include "macros.h"
  33
  34 int
  35 main (int argc, char *argv[])
  36 {
  37   mbstate_t state;
  38   wchar_t wc;
  39   size_t ret;
  40
  41   /* configure should already have checked that the locale is supported.  */
  42   if (setlocale (LC_ALL, "") == NULL)
  43     return 1;
  44
  45   /* Test zero-length input.  */
  46   {
  47     memset (&state, '\0', sizeof (mbstate_t));
  48     wc = (wchar_t) 0xBADFACE;
  49     ret = mbrtowc (&wc, "x", 0, &state);
  50     ASSERT (ret == (size_t)(-2));
  51     ASSERT (mbsinit (&state));
  52   }
  53
  54   /* Test NUL byte input.  */
  55   {
  56     memset (&state, '\0', sizeof (mbstate_t));
  57     wc = (wchar_t) 0xBADFACE;
  58     ret = mbrtowc (&wc, "", 1, &state);
  59     ASSERT (ret == 0);
  60     ASSERT (wc == 0);
  61     ASSERT (mbsinit (&state));
  62     ret = mbrtowc (NULL, "", 1, &state);
  63     ASSERT (ret == 0);
  64     ASSERT (mbsinit (&state));
  65   }
  66
  67   /* Test single-byte input.  */
  68   {
  69     int c;
  70     char buf[1];
  71
  72     memset (&state, '\0', sizeof (mbstate_t));
  73     for (c = 0; c < 0x100; c++)
  74       switch (c)
  75         {
  76         case '\t': case '\v': case '\f':
  77         case ' ': case '!': case '"': case '#': case '%':
  78         case '&': case '\'': case '(': case ')': case '*':
  79         case '+': case ',': case '-': case '.': case '/':
  80         case '0': case '1': case '2': case '3': case '4':
  81         case '5': case '6': case '7': case '8': case '9':
  82         case ':': case ';': case '<': case '=': case '>':
  83         case '?':
  84         case 'A': case 'B': case 'C': case 'D': case 'E':
  85         case 'F': case 'G': case 'H': case 'I': case 'J':
  86         case 'K': case 'L': case 'M': case 'N': case 'O':
  87         case 'P': case 'Q': case 'R': case 'S': case 'T':
  88         case 'U': case 'V': case 'W': case 'X': case 'Y':
  89         case 'Z':
  90         case '[': case '\\': case ']': case '^': case '_':
  91         case 'a': case 'b': case 'c': case 'd': case 'e':
  92         case 'f': case 'g': case 'h': case 'i': case 'j':
  93         case 'k': case 'l': case 'm': case 'n': case 'o':
  94         case 'p': case 'q': case 'r': case 's': case 't':
  95         case 'u': case 'v': case 'w': case 'x': case 'y':
  96         case 'z': case '{': case '|': case '}': case '~':
  97           /* c is in the ISO C "basic character set".  */
  98           ASSERT (c < 0x80);
  99           /* c is an ASCII character.  */
 100           buf[0] = c;
 101
 102           wc = (wchar_t) 0xBADFACE;
 103           ret = mbrtowc (&wc, buf, 1, &state);
 104           ASSERT (ret == 1);
 105           ASSERT (wc == c);
 106           ASSERT (mbsinit (&state));
 107
 108           ret = mbrtowc (NULL, buf, 1, &state);
 109           ASSERT (ret == 1);
 110           ASSERT (mbsinit (&state));
 111
 112           break;
 113         default:
 114           break;
 115         }
 116   }
 117
 118   /* Test special calling convention, passing a NULL pointer.  */
 119   {
 120     memset (&state, '\0', sizeof (mbstate_t));
 121     wc = (wchar_t) 0xBADFACE;
 122     ret = mbrtowc (&wc, NULL, 5, &state);
 123     ASSERT (ret == 0);
 124     ASSERT (wc == (wchar_t) 0xBADFACE);
 125     ASSERT (mbsinit (&state));
 126   }
 127
 128 #ifdef __ANDROID__
 129   /* On Android ≥ 5.0, the default locale is the "C.UTF-8" locale, not the
 130      "C" locale.  Furthermore, when you attempt to set the "C" or "POSIX"
 131      locale via setlocale(), what you get is a "C" locale with UTF-8 encoding,
 132      that is, effectively the "C.UTF-8" locale.  */
 133   if (argc > 1 && strcmp (argv[1], "1") == 0 && MB_CUR_MAX > 1)
 134     argv[1] = "3";
 135 #endif
 136
 137   if (argc > 1)
 138     switch (argv[1][0])
 139       {
 140       case '1':
 141         /* C or POSIX locale.  */
 142         {
 143           int c;
 144           char buf[1];
 145
 146           memset (&state, '\0', sizeof (mbstate_t));
 147           for (c = 0; c < 0x100; c++)
 148             if (c != 0)
 149               {
 150                 /* We are testing all nonnull bytes.  */
 151                 buf[0] = c;
 152
 153                 wc = (wchar_t) 0xBADFACE;
 154                 ret = mbrtowc (&wc, buf, 1, &state);
 155                 /* POSIX:2018 says: "In the POSIX locale an [EILSEQ] error
 156                    cannot occur since all byte values are valid characters."  */
 157                 ASSERT (ret == 1);
 158                 if (c < 0x80)
 159                   /* c is an ASCII character.  */
 160                   ASSERT (wc == c);
 161                 else
 162                   /* On most platforms, the bytes 0x80..0xFF map to U+0080..U+00FF.
 163                      But on musl libc, the bytes 0x80..0xFF map to U+DF80..U+DFFF.  */
 164                   ASSERT (wc == (btowc (c) == 0xDF00 + c ? btowc (c) : c));
 165                 ASSERT (mbsinit (&state));
 166
 167                 ret = mbrtowc (NULL, buf, 1, &state);
 168                 ASSERT (ret == 1);
 169                 ASSERT (mbsinit (&state));
 170               }
 171         }
 172         return test_exit_status;
 173
 174       case '2':
 175         /* Locale encoding is ISO-8859-1 or ISO-8859-15.  */
 176         {
 177           char input[] = "B\374\337er"; /* "Büßer" */
 178           memset (&state, '\0', sizeof (mbstate_t));
 179
 180           wc = (wchar_t) 0xBADFACE;
 181           ret = mbrtowc (&wc, input, 1, &state);
 182           ASSERT (ret == 1);
 183           ASSERT (wc == 'B');
 184           ASSERT (mbsinit (&state));
 185           input[0] = '\0';
 186
 187           wc = (wchar_t) 0xBADFACE;
 188           ret = mbrtowc (&wc, input + 1, 1, &state);
 189           ASSERT (ret == 1);
 190           ASSERT (wctob (wc) == (unsigned char) '\374');
 191           ASSERT (mbsinit (&state));
 192           input[1] = '\0';
 193
 194           /* Test support of NULL first argument.  */
 195           ret = mbrtowc (NULL, input + 2, 3, &state);
 196           ASSERT (ret == 1);
 197           ASSERT (mbsinit (&state));
 198
 199           wc = (wchar_t) 0xBADFACE;
 200           ret = mbrtowc (&wc, input + 2, 3, &state);
 201           ASSERT (ret == 1);
 202           ASSERT (wctob (wc) == (unsigned char) '\337');
 203           ASSERT (mbsinit (&state));
 204           input[2] = '\0';
 205
 206           wc = (wchar_t) 0xBADFACE;
 207           ret = mbrtowc (&wc, input + 3, 2, &state);
 208           ASSERT (ret == 1);
 209           ASSERT (wc == 'e');
 210           ASSERT (mbsinit (&state));
 211           input[3] = '\0';
 212
 213           wc = (wchar_t) 0xBADFACE;
 214           ret = mbrtowc (&wc, input + 4, 1, &state);
 215           ASSERT (ret == 1);
 216           ASSERT (wc == 'r');
 217           ASSERT (mbsinit (&state));
 218         }
 219         return test_exit_status;
 220
 221       case '3':
 222         /* Locale encoding is UTF-8.  */
 223         {
 224           char input[] = "B\303\274\303\237er"; /* "Büßer" */
 225           memset (&state, '\0', sizeof (mbstate_t));
 226
 227           wc = (wchar_t) 0xBADFACE;
 228           ret = mbrtowc (&wc, input, 1, &state);
 229           ASSERT (ret == 1);
 230           ASSERT (wc == 'B');
 231           ASSERT (mbsinit (&state));
 232           input[0] = '\0';
 233
 234           wc = (wchar_t) 0xBADFACE;
 235           ret = mbrtowc (&wc, input + 1, 1, &state);
 236           ASSERT (ret == (size_t)(-2));
 237           ASSERT (wc == (wchar_t) 0xBADFACE);
 238           ASSERT (!mbsinit (&state));
 239           input[1] = '\0';
 240
 241           wc = (wchar_t) 0xBADFACE;
 242           ret = mbrtowc (&wc, input + 2, 5, &state);
 243           ASSERT (ret == 1);
 244           ASSERT (wctob (wc) == EOF);
 245           ASSERT (mbsinit (&state));
 246           input[2] = '\0';
 247
 248           /* Test support of NULL first argument.  */
 249           ret = mbrtowc (NULL, input + 3, 4, &state);
 250           ASSERT (ret == 2);
 251           ASSERT (mbsinit (&state));
 252
 253           wc = (wchar_t) 0xBADFACE;
 254           ret = mbrtowc (&wc, input + 3, 4, &state);
 255           ASSERT (ret == 2);
 256           ASSERT (wctob (wc) == EOF);
 257           ASSERT (mbsinit (&state));
 258           input[3] = '\0';
 259           input[4] = '\0';
 260
 261           wc = (wchar_t) 0xBADFACE;
 262           ret = mbrtowc (&wc, input + 5, 2, &state);
 263           ASSERT (ret == 1);
 264           ASSERT (wc == 'e');
 265           ASSERT (mbsinit (&state));
 266           input[5] = '\0';
 267
 268           wc = (wchar_t) 0xBADFACE;
 269           ret = mbrtowc (&wc, input + 6, 1, &state);
 270           ASSERT (ret == 1);
 271           ASSERT (wc == 'r');
 272           ASSERT (mbsinit (&state));
 273         }
 274         if (sizeof (wchar_t) > 2)
 275           { /* \360\237\220\203 = U+0001F403 */
 276             memset (&state, '\0', sizeof (mbstate_t));
 277
 278             wc = (wchar_t) 0xBADFACE;
 279             ret = mbrtowc (&wc, "\360", 1, &state);
 280             ASSERT (ret == (size_t)(-2));
 281             ASSERT (wc == (wchar_t) 0xBADFACE);
 282             ASSERT (!mbsinit (&state));
 283
 284             wc = (wchar_t) 0xBADFACE;
 285             ret = mbrtowc (&wc, "\237", 1, &state);
 286             ASSERT (ret == (size_t)(-2));
 287             ASSERT (wc == (wchar_t) 0xBADFACE);
 288             ASSERT (!mbsinit (&state));
 289
 290             wc = (wchar_t) 0xBADFACE;
 291             ret = mbrtowc (&wc, "\220", 1, &state);
 292             ASSERT (ret == (size_t)(-2));
 293             ASSERT (wc == (wchar_t) 0xBADFACE);
 294             ASSERT (!mbsinit (&state));
 295
 296             wc = (wchar_t) 0xBADFACE;
 297             ret = mbrtowc (&wc, "\203", 1, &state);
 298             ASSERT (ret == 1);
 299             ASSERT (wctob (wc) == EOF);
 300             ASSERT (mbsinit (&state));
 301           }
 302         return test_exit_status;
 303
 304       case '4':
 305         /* Locale encoding is EUC-JP.  */
 306         {
 307           char input[] = "<\306\374\313\334\270\354>"; /* "<日本語>" */
 308           memset (&state, '\0', sizeof (mbstate_t));
 309
 310           wc = (wchar_t) 0xBADFACE;
 311           ret = mbrtowc (&wc, input, 1, &state);
 312           ASSERT (ret == 1);
 313           ASSERT (wc == '<');
 314           ASSERT (mbsinit (&state));
 315           input[0] = '\0';
 316
 317           wc = (wchar_t) 0xBADFACE;
 318           ret = mbrtowc (&wc, input + 1, 2, &state);
 319           ASSERT (ret == 2);
 320           ASSERT (wctob (wc) == EOF);
 321           ASSERT (mbsinit (&state));
 322           input[1] = '\0';
 323           input[2] = '\0';
 324
 325           wc = (wchar_t) 0xBADFACE;
 326           ret = mbrtowc (&wc, input + 3, 1, &state);
 327           ASSERT (ret == (size_t)(-2));
 328           ASSERT (wc == (wchar_t) 0xBADFACE);
 329           ASSERT (!mbsinit (&state));
 330           input[3] = '\0';
 331
 332           wc = (wchar_t) 0xBADFACE;
 333           ret = mbrtowc (&wc, input + 4, 4, &state);
 334           ASSERT (ret == 1);
 335           ASSERT (wctob (wc) == EOF);
 336           ASSERT (mbsinit (&state));
 337           input[4] = '\0';
 338
 339           /* Test support of NULL first argument.  */
 340           ret = mbrtowc (NULL, input + 5, 3, &state);
 341           ASSERT (ret == 2);
 342           ASSERT (mbsinit (&state));
 343
 344           wc = (wchar_t) 0xBADFACE;
 345           ret = mbrtowc (&wc, input + 5, 3, &state);
 346           ASSERT (ret == 2);
 347           ASSERT (wctob (wc) == EOF);
 348           ASSERT (mbsinit (&state));
 349           input[5] = '\0';
 350           input[6] = '\0';
 351
 352           wc = (wchar_t) 0xBADFACE;
 353           ret = mbrtowc (&wc, input + 7, 1, &state);
 354           ASSERT (ret == 1);
 355           ASSERT (wc == '>');
 356           ASSERT (mbsinit (&state));
 357         }
 358         return test_exit_status;
 359
 360       case '5':
 361         /* Locale encoding is GB18030.  */
 362         {
 363           char input[] = "B\250\271\201\060\211\070er"; /* "Büßer" */
 364           memset (&state, '\0', sizeof (mbstate_t));
 365
 366           wc = (wchar_t) 0xBADFACE;
 367           ret = mbrtowc (&wc, input, 1, &state);
 368           ASSERT (ret == 1);
 369           ASSERT (wc == 'B');
 370           ASSERT (mbsinit (&state));
 371           input[0] = '\0';
 372
 373           wc = (wchar_t) 0xBADFACE;
 374           ret = mbrtowc (&wc, input + 1, 1, &state);
 375           ASSERT (ret == (size_t)(-2));
 376           ASSERT (wc == (wchar_t) 0xBADFACE);
 377           ASSERT (!mbsinit (&state));
 378           input[1] = '\0';
 379
 380           wc = (wchar_t) 0xBADFACE;
 381           ret = mbrtowc (&wc, input + 2, 7, &state);
 382           ASSERT (ret == 1);
 383           ASSERT (wctob (wc) == EOF);
 384           ASSERT (mbsinit (&state));
 385           input[2] = '\0';
 386
 387           /* Test support of NULL first argument.  */
 388           ret = mbrtowc (NULL, input + 3, 6, &state);
 389           ASSERT (ret == 4);
 390           ASSERT (mbsinit (&state));
 391
 392           wc = (wchar_t) 0xBADFACE;
 393           ret = mbrtowc (&wc, input + 3, 6, &state);
 394           ASSERT (ret == 4);
 395           ASSERT (wctob (wc) == EOF);
 396           ASSERT (mbsinit (&state));
 397           input[3] = '\0';
 398           input[4] = '\0';
 399           input[5] = '\0';
 400           input[6] = '\0';
 401
 402           wc = (wchar_t) 0xBADFACE;
 403           ret = mbrtowc (&wc, input + 7, 2, &state);
 404           ASSERT (ret == 1);
 405           ASSERT (wc == 'e');
 406           ASSERT (mbsinit (&state));
 407           input[7] = '\0';
 408
 409           wc = (wchar_t) 0xBADFACE;
 410           ret = mbrtowc (&wc, input + 8, 1, &state);
 411           ASSERT (ret == 1);
 412           ASSERT (wc == 'r');
 413           ASSERT (mbsinit (&state));
 414         }
 415         if (sizeof (wchar_t) > 2)
 416           { /* \224\071\311\067 = U+0001F403 */
 417             memset (&state, '\0', sizeof (mbstate_t));
 418
 419             wc = (wchar_t) 0xBADFACE;
 420             ret = mbrtowc (&wc, "\224", 1, &state);
 421             ASSERT (ret == (size_t)(-2));
 422             ASSERT (wc == (wchar_t) 0xBADFACE);
 423             ASSERT (!mbsinit (&state));
 424
 425             wc = (wchar_t) 0xBADFACE;
 426             ret = mbrtowc (&wc, "\071", 1, &state);
 427             ASSERT (ret == (size_t)(-2));
 428             ASSERT (wc == (wchar_t) 0xBADFACE);
 429             ASSERT (!mbsinit (&state));
 430
 431             wc = (wchar_t) 0xBADFACE;
 432             ret = mbrtowc (&wc, "\311", 1, &state);
 433             ASSERT (ret == (size_t)(-2));
 434             ASSERT (wc == (wchar_t) 0xBADFACE);
 435             ASSERT (!mbsinit (&state));
 436
 437             wc = (wchar_t) 0xBADFACE;
 438             ret = mbrtowc (&wc, "\067", 1, &state);
 439             ASSERT (ret == 1);
 440             ASSERT (wctob (wc) == EOF);
 441             ASSERT (mbsinit (&state));
 442           }
 443         return test_exit_status;
 444       }
 445
 446   return 1;
 447 }