tst_idna.c

   1 /* tst_idna.c   Self tests for idna_to_ascii().
   2  * Copyright (C) 2002, 2003  Simon Josefsson
   3  *
   4  * This file is part of GNU Libidn.
   5  *
   6  * GNU Libidn is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * GNU Libidn is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * along with GNU Libidn; if not, write to the Free Software
  18  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  19  *
  20  */
  21
  22 #include "internal.h"
  23
  24 static int debug = 0;
  25 static int error_count = 0;
  26 static int break_on_error = 0;
  27
  28 static void
  29 fail (const char *format, ...)
  30 {
  31   va_list arg_ptr;
  32
  33   va_start (arg_ptr, format);
  34   vfprintf (stderr, format, arg_ptr);
  35   va_end (arg_ptr);
  36   error_count++;
  37   if (break_on_error)
  38     exit (1);
  39 }
  40
  41 static void
  42 escapeprint (char *str, int len)
  43 {
  44   int i;
  45
  46   printf ("(length %d bytes):\n", len);
  47   printf ("\t'");
  48   for (i = 0; i < len; i++)
  49     {
  50       if (((str[i] & 0xFF) >= 'A' && (str[i] & 0xFF) <= 'Z') ||
  51           ((str[i] & 0xFF) >= 'a' && (str[i] & 0xFF) <= 'z') ||
  52           ((str[i] & 0xFF) >= '0' && (str[i] & 0xFF) <= '9')
  53           || (str[i] & 0xFF) == ' ' || (str[i] & 0xFF) == '.')
  54         printf ("%c", (str[i] & 0xFF));
  55       else
  56         printf ("\\x%02X", (str[i] & 0xFF));
  57       if ((i+1)%16 == 0 && (i+1) < len)
  58         printf("'\n\t'");
  59     }
  60 }
  61
  62 static void
  63 ucs4print (unsigned long *str, ssize_t len)
  64 {
  65   int i;
  66
  67   printf ("\t;; ");
  68   for (i = 0; len >= 0 ? i < len : str[i]; i++)
  69     {
  70       printf ("U+%04lu ", str[i] & 0xFFFF);
  71       if ((i + 1) % 4 == 0)
  72         printf (" ");
  73       if ((i + 1) % 8 == 0 && i + 1 < len)
  74         printf ("\n\t;; ");
  75     }
  76   puts ("");
  77 }
  78
  79 struct idna
  80 {
  81   char *name;
  82   size_t inlen;
  83   unsigned long in[100];
  84   char *out;
  85   int allowunassigned;
  86   int usestd3asciirules;
  87   int toasciirc;
  88   int tounicoderc;
  89 } idna[] =
  90 {
  91   {
  92     "Arabic (Egyptian)", 17,
  93     {
  94   0x0644, 0x064A, 0x0647, 0x0645, 0x0627, 0x0628, 0x062A, 0x0643,
  95         0x0644, 0x0645, 0x0648, 0x0634, 0x0639, 0x0631, 0x0628, 0x064A,
  96         0x061F},
  97       IDNA_ACE_PREFIX "egbpdaj6bu4bxfgehfvwxn", 0, 0, IDNA_SUCCESS,
  98       IDNA_SUCCESS},
  99   {
 100     "Chinese (simplified)", 9,
 101     {
 102   0x4ED6, 0x4EEC, 0x4E3A, 0x4EC0, 0x4E48, 0x4E0D, 0x8BF4, 0x4E2D, 0x6587},
 103       IDNA_ACE_PREFIX "ihqwcrb4cv8a8dqg056pqjye", 0, 0, IDNA_SUCCESS,
 104       IDNA_SUCCESS},
 105   {
 106     "Chinese (traditional)", 9,
 107     {
 108   0x4ED6, 0x5011, 0x7232, 0x4EC0, 0x9EBD, 0x4E0D, 0x8AAA, 0x4E2D, 0x6587},
 109       IDNA_ACE_PREFIX "ihqwctvzc91f659drss3x8bo0yb", 0, 0, IDNA_SUCCESS,
 110       IDNA_SUCCESS},
 111   {
 112     "Czech"
 113 #if !defined(DRAFT)
 114     ": Pro[CCARON]prost[ECARON]nemluv[IACUTE][CCARON]esky"
 115 #endif
 116     , 22,
 117     {
 118   0x0050, 0x0072, 0x006F, 0x010D, 0x0070, 0x0072, 0x006F, 0x0073,
 119         0x0074, 0x011B, 0x006E, 0x0065, 0x006D, 0x006C, 0x0075, 0x0076,
 120         0x00ED, 0x010D, 0x0065, 0x0073, 0x006B, 0x0079},
 121       IDNA_ACE_PREFIX "Proprostnemluvesky-uyb24dma41a", 0, 0, IDNA_SUCCESS,
 122       IDNA_SUCCESS},
 123   {
 124     "Hebrew", 22,
 125     {
 126   0x05DC, 0x05DE, 0x05D4, 0x05D4, 0x05DD, 0x05E4, 0x05E9, 0x05D5,
 127         0x05D8, 0x05DC, 0x05D0, 0x05DE, 0x05D3, 0x05D1, 0x05E8, 0x05D9,
 128         0x05DD, 0x05E2, 0x05D1, 0x05E8, 0x05D9, 0x05EA},
 129       IDNA_ACE_PREFIX "4dbcagdahymbxekheh6e0a7fei0b", 0, 0, IDNA_SUCCESS,
 130       IDNA_SUCCESS},
 131   {
 132     "Hindi (Devanagari)", 30,
 133     {
 134   0x092F, 0x0939, 0x0932, 0x094B, 0x0917, 0x0939, 0x093F, 0x0928,
 135         0x094D, 0x0926, 0x0940, 0x0915, 0x094D, 0x092F, 0x094B, 0x0902,
 136         0x0928, 0x0939, 0x0940, 0x0902, 0x092C, 0x094B, 0x0932, 0x0938,
 137         0x0915, 0x0924, 0x0947, 0x0939, 0x0948, 0x0902},
 138       IDNA_ACE_PREFIX "i1baa7eci9glrd9b2ae1bj0hfcgg6iyaf8o0a1dig0cd", 0, 0,
 139       IDNA_SUCCESS},
 140   {
 141     "Japanese (kanji and hiragana)", 18,
 142     {
 143   0x306A, 0x305C, 0x307F, 0x3093, 0x306A, 0x65E5, 0x672C, 0x8A9E,
 144         0x3092, 0x8A71, 0x3057, 0x3066, 0x304F, 0x308C, 0x306A, 0x3044,
 145         0x306E, 0x304B},
 146       IDNA_ACE_PREFIX "n8jok5ay5dzabd5bym9f0cm5685rrjetr6pdxa", 0, 0,
 147       IDNA_SUCCESS},
 148 #if !defined(DRAFT)
 149   {
 150     "Korean (Hangul syllables)", 24,
 151     {
 152   0xC138, 0xACC4, 0xC758, 0xBAA8, 0xB4E0, 0xC0AC, 0xB78C, 0xB4E4,
 153         0xC774, 0xD55C, 0xAD6D, 0xC5B4, 0xB97C, 0xC774, 0xD574, 0xD55C,
 154         0xB2E4, 0xBA74, 0xC5BC, 0xB9C8, 0xB098, 0xC88B, 0xC744, 0xAE4C},
 155       IDNA_ACE_PREFIX "989aomsvi5e83db1d2a355cv1e0vak1dwrv93d5xbh15a0dt"
 156       "30a5jpsd879ccm6fea98c", 0, 0, IDNA_PUNYCODE_ERROR,
 157       IDNA_PUNYCODE_ERROR},
 158     /* too long output */
 159 #endif
 160   {
 161     "Russian (Cyrillic)", 28,
 162     {
 163   0x043F, 0x043E, 0x0447, 0x0435, 0x043C, 0x0443, 0x0436, 0x0435,
 164         0x043E, 0x043D, 0x0438, 0x043D, 0x0435, 0x0433, 0x043E, 0x0432,
 165         0x043E, 0x0440, 0x044F, 0x0442, 0x043F, 0x043E, 0x0440, 0x0443,
 166         0x0441, 0x0441, 0x043A, 0x0438},
 167       IDNA_ACE_PREFIX "b1abfaaepdrnnbgefbadotcwatmq2g4l", 0, 0,
 168       IDNA_SUCCESS, IDNA_SUCCESS},
 169   {
 170     "Spanish"
 171 #if !defined(DRAFT)
 172     ": Porqu[EACUTE]nopuedensimplementehablarenEspa[NTILDE]ol"
 173 #endif
 174     , 40,
 175     {
 176   0x0050, 0x006F, 0x0072, 0x0071, 0x0075, 0x00E9, 0x006E, 0x006F,
 177         0x0070, 0x0075, 0x0065, 0x0064, 0x0065, 0x006E, 0x0073, 0x0069,
 178         0x006D, 0x0070, 0x006C, 0x0065, 0x006D, 0x0065, 0x006E, 0x0074,
 179         0x0065, 0x0068, 0x0061, 0x0062, 0x006C, 0x0061, 0x0072, 0x0065,
 180         0x006E, 0x0045, 0x0073, 0x0070, 0x0061, 0x00F1, 0x006F, 0x006C},
 181       IDNA_ACE_PREFIX "PorqunopuedensimplementehablarenEspaol-fmd56a", 0, 0,
 182       IDNA_SUCCESS},
 183   {
 184     "Vietnamese", 31,
 185     {
 186   0x0054, 0x1EA1, 0x0069, 0x0073, 0x0061, 0x006F, 0x0068, 0x1ECD,
 187         0x006B, 0x0068, 0x00F4, 0x006E, 0x0067, 0x0074, 0x0068, 0x1EC3,
 188         0x0063, 0x0068, 0x1EC9, 0x006E, 0x00F3, 0x0069, 0x0074, 0x0069,
 189         0x1EBF, 0x006E, 0x0067, 0x0056, 0x0069, 0x1EC7, 0x0074},
 190       IDNA_ACE_PREFIX "TisaohkhngthchnitingVit-kjcr8268qyxafd2f1b9g", 0, 0,
 191       IDNA_SUCCESS},
 192   {
 193     "Japanese"
 194 #if !defined(DRAFT)
 195     "3[NEN]B[GUMI][KINPACHI][SENSEI]"
 196 #endif
 197     , 8,
 198     {
 199   0x0033, 0x5E74, 0x0042, 0x7D44, 0x91D1, 0x516B, 0x5148, 0x751F},
 200       IDNA_ACE_PREFIX "3B-ww4c5e180e575a65lsy2b", 0, 0, IDNA_SUCCESS,
 201       IDNA_SUCCESS},
 202   {
 203     "Japanese"
 204 #if !defined(DRAFT)
 205     "[AMURO][NAMIE]-with-SUPER-MONKEYS"
 206 #endif
 207     , 24,
 208     {
 209   0x5B89, 0x5BA4, 0x5948, 0x7F8E, 0x6075, 0x002D, 0x0077, 0x0069,
 210         0x0074, 0x0068, 0x002D, 0x0053, 0x0055, 0x0050, 0x0045, 0x0052,
 211         0x002D, 0x004D, 0x004F, 0x004E, 0x004B, 0x0045, 0x0059, 0x0053},
 212       IDNA_ACE_PREFIX "-with-SUPER-MONKEYS-pc58ag80a8qai00g7n9n", 0, 0,
 213       IDNA_SUCCESS},
 214   {
 215     "Japanese"
 216 #if !defined(DRAFT)
 217     "Hello-Another-Way-[SOREZORE][NO][BASHO]"
 218 #endif
 219     , 25,
 220     {
 221   0x0048, 0x0065, 0x006C, 0x006C, 0x006F, 0x002D, 0x0041, 0x006E,
 222         0x006F, 0x0074, 0x0068, 0x0065, 0x0072, 0x002D, 0x0057, 0x0061,
 223         0x0079, 0x002D, 0x305D, 0x308C, 0x305E, 0x308C, 0x306E, 0x5834,
 224         0x6240},
 225       IDNA_ACE_PREFIX "Hello-Another-Way--fc4qua05auwb3674vfr0b", 0, 0,
 226       IDNA_SUCCESS},
 227   {
 228     "Japanese"
 229 #if !defined(DRAFT)
 230     "[HITOTSU][YANE][NO][SHITA]2"
 231 #endif
 232     , 8,
 233     {
 234   0x3072, 0x3068, 0x3064, 0x5C4B, 0x6839, 0x306E, 0x4E0B, 0x0032},
 235       IDNA_ACE_PREFIX "2-u9tlzr9756bt3uc0v", 0, 0, IDNA_SUCCESS,
 236       IDNA_SUCCESS},
 237   {
 238     "Japanese"
 239 #if !defined(DRAFT)
 240     "Maji[DE]Koi[SURU]5[BYOU][MAE]"
 241 #endif
 242     , 13,
 243     {
 244   0x004D, 0x0061, 0x006A, 0x0069, 0x3067, 0x004B, 0x006F, 0x0069,
 245         0x3059, 0x308B, 0x0035, 0x79D2, 0x524D},
 246       IDNA_ACE_PREFIX "MajiKoi5-783gue6qz075azm5e", 0, 0, IDNA_SUCCESS,
 247       IDNA_SUCCESS},
 248   {
 249     "Japanese"
 250 #if !defined(DRAFT)
 251     "[PAFII]de[RUNBA]"
 252 #endif
 253     , 9,
 254     {
 255   0x30D1, 0x30D5, 0x30A3, 0x30FC, 0x0064, 0x0065, 0x30EB, 0x30F3, 0x30D0},
 256       IDNA_ACE_PREFIX "de-jg4avhby1noc0d", 0, 0, IDNA_SUCCESS, IDNA_SUCCESS},
 257   {
 258     "Japanese"
 259 #if !defined(DRAFT)
 260     "[SONO][SUPIIDO][DE]"
 261 #endif
 262     , 7,
 263     {
 264   0x305D, 0x306E, 0x30B9, 0x30D4, 0x30FC, 0x30C9, 0x3067},
 265       IDNA_ACE_PREFIX "d9juau41awczczp", 0, 0, IDNA_SUCCESS, IDNA_SUCCESS},
 266   {
 267     "Greek", 8,
 268     {0x03b5, 0x03bb, 0x03bb, 0x03b7, 0x03bd, 0x03b9, 0x03ba, 0x03ac},
 269     IDNA_ACE_PREFIX "hxargifdar", 0, 0, IDNA_SUCCESS, IDNA_SUCCESS},
 270   {
 271     "Maltese (Malti)", 10,
 272     {0x0062, 0x006f, 0x006e, 0x0121, 0x0075, 0x0073, 0x0061, 0x0127,
 273      0x0127, 0x0061},
 274     IDNA_ACE_PREFIX "bonusaa-5bb1da", 0, 0, IDNA_SUCCESS, IDNA_SUCCESS},
 275   {
 276     "Russian (Cyrillic)", 28,
 277     {0x043f, 0x043e, 0x0447, 0x0435, 0x043c, 0x0443, 0x0436, 0x0435,
 278      0x043e, 0x043d, 0x0438, 0x043d, 0x0435, 0x0433, 0x043e, 0x0432,
 279      0x043e, 0x0440, 0x044f, 0x0442, 0x043f, 0x043e, 0x0440, 0x0443,
 280      0x0441, 0x0441, 0x043a, 0x0438},
 281     IDNA_ACE_PREFIX "b1abfaaepdrnnbgefbadotcwatmq2g4l", 0, 0,
 282     IDNA_SUCCESS, IDNA_SUCCESS},
 283 #if 0
 284   {
 285     "(S) -> $1.00 <-", 11,
 286     {
 287   0x002D, 0x003E, 0x0020, 0x0024, 0x0031, 0x002E, 0x0030, 0x0030,
 288         0x0020, 0x003C, 0x002D},
 289       IDNA_ACE_PREFIX "-> $1.00 <--", 0, 0, IDNA_SUCCESS, IDNA_SUCCESS},
 290 #endif
 291 #if !defined(DRAFT)
 292   {                             /* XXX depends on IDNA_ACE_PREFIX */
 293     "ToASCII() with ACE prefix", 4 + 3,
 294     { 'x', 'n', '-', '-', 'f', 'o', 0x3067},
 295       IDNA_ACE_PREFIX "too long too long too long too long too long too "
 296       "long too long too long too long too long ", 0, 0,
 297       IDNA_CONTAINS_ACE_PREFIX, IDNA_PUNYCODE_ERROR}
 298 #endif
 299 };
 300
 301 int
 302 main (int argc, char *argv[])
 303 {
 304   char label[100];
 305   unsigned long *ucs4label = NULL;
 306   unsigned long tmp[100];
 307   size_t len, len2, i;
 308   int rc;
 309
 310   do
 311     if (strcmp (argv[argc - 1], "-v") == 0 ||
 312         strcmp (argv[argc - 1], "--verbose") == 0)
 313       debug = 1;
 314     else if (strcmp (argv[argc - 1], "-b") == 0 ||
 315              strcmp (argv[argc - 1], "--break-on-error") == 0)
 316       break_on_error = 1;
 317     else if (strcmp (argv[argc - 1], "-h") == 0 ||
 318              strcmp (argv[argc - 1], "-?") == 0 ||
 319              strcmp (argv[argc - 1], "--help") == 0)
 320       {
 321         printf ("Usage: %s [-vbh?] [--verbose] [--break-on-error] [--help]\n",
 322                 argv[0]);
 323         return 1;
 324       }
 325   while (argc-- > 1);
 326
 327   for (i = 0; i < sizeof (idna) / sizeof (idna[0]); i++)
 328     {
 329 #ifdef DRAFT
 330       printf("<section title=\"%s\">\n", idna[i].name);
 331       printf("\n");
 332       printf("<figure>\n");
 333       printf("<artwork>\n");
 334       { char *p;
 335         int j;
 336         size_t t;
 337         p = stringprep_ucs4_to_utf8(idna[i].in, idna[i].inlen, NULL, &t);
 338         printf ("in ");
 339         escapeprint(p, strlen(p));
 340         printf ("\n");
 341         free(p);
 342
 343         printf ("input (length %d):\n\t", t);
 344         for (j = 0; j < idna[i].inlen; j++)
 345           {
 346             printf ("U+%04lx ", idna[i].in[j]);
 347             if ((j+1)%8 == 0)
 348               printf("\n\t");
 349           }
 350         printf ("\n");
 351       }
 352 #endif
 353
 354       if (debug)
 355         printf ("IDNA entry %d: %s\n", i, idna[i].name);
 356
 357       if (debug)
 358         {
 359           printf ("in:\n");
 360           ucs4print (idna[i].in, idna[i].inlen);
 361         }
 362
 363       rc = idna_to_ascii (idna[i].in, idna[i].inlen, label,
 364                           idna[i].allowunassigned, idna[i].usestd3asciirules);
 365       if (rc != idna[i].toasciirc)
 366         {
 367           fail ("IDNA entry %d failed: %d\n", i, rc);
 368           if (debug)
 369             printf ("FATAL\n");
 370           continue;
 371         }
 372
 373 #ifdef DRAFT
 374       if (rc == IDNA_SUCCESS)
 375         printf ("\nout: %s\n", label);
 376 #endif
 377
 378       if (debug && rc == IDNA_SUCCESS)
 379         {
 380           printf ("computed out: %s\n", label);
 381           printf ("expected out: %s\n", idna[i].out);
 382         }
 383       else if (debug)
 384         printf ("returned %d expected %d\n", rc, idna[i].toasciirc);
 385
 386       if (rc == IDNA_SUCCESS)
 387         {
 388           if (strlen (idna[i].out) != strlen (label) ||
 389               strcasecmp (idna[i].out, label) != 0)
 390             {
 391               fail ("IDNA entry %d failed\n", i);
 392               if (debug)
 393                 printf ("ERROR\n");
 394             }
 395           else if (debug)
 396             printf ("OK\n");
 397         }
 398       else if (debug)
 399         printf ("OK\n");
 400
 401       if (ucs4label)
 402         free(ucs4label);
 403
 404       ucs4label = stringprep_utf8_to_ucs4 (idna[i].out, -1, &len);
 405
 406       if (debug)
 407         {
 408           printf ("in: %s (%d==%d)\n", idna[i].out, strlen (idna[i].out),
 409                   len);
 410           ucs4print (ucs4label, len);
 411         }
 412
 413       len2 = sizeof (tmp);
 414       rc = idna_to_unicode (ucs4label, len, tmp, &len2,
 415                             idna[i].allowunassigned,
 416                             idna[i].usestd3asciirules);
 417       if (debug)
 418         {
 419           printf ("expected out:\n");
 420           if (rc == IDNA_SUCCESS)
 421             ucs4print (idna[i].in, idna[i].inlen);
 422           else
 423             ucs4print (ucs4label, len);
 424
 425           printf ("computed out:\n");
 426           ucs4print (tmp, len2);
 427         }
 428
 429       if (rc != idna[i].tounicoderc)
 430         {
 431           fail ("IDNA entry %d failed: %d\n", i, rc);
 432           if (debug)
 433             printf ("FATAL\n");
 434           continue;
 435         }
 436
 437       if ((rc == IDNA_SUCCESS && (len2 != idna[i].inlen ||
 438                                   memcmp (idna[i].in, tmp, len2) != 0)) ||
 439           (rc != IDNA_SUCCESS && (len2 != len ||
 440                                   memcmp (ucs4label, tmp, len) != 0)))
 441         {
 442           fail ("IDNA entry %d failed\n", i);
 443           if (debug)
 444             printf ("ERROR\n");
 445         }
 446       else if (debug)
 447         printf ("OK\n\n");
 448
 449 #ifdef DRAFT
 450       printf("</artwork>\n");
 451       printf("</figure>\n");
 452       printf("\n");
 453       printf("</section>\n");
 454 #endif
 455     }
 456
 457   if (ucs4label)
 458     free(ucs4label);
 459
 460   if (debug)
 461     printf ("IDNA self tests done with %d errors\n", error_count);
 462
 463   return error_count ? 1 : 0;
 464 }