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 ucs4print (unsigned long *str, ssize_t len)
  43 {
  44   int i;
  45
  46   printf ("\t;; ");
  47   for (i = 0; len >= 0 ? i < len : str[i]; i++)
  48     {
  49       printf ("U+%04lux ", str[i]);
  50       if ((i + 1) % 4 == 0)
  51         printf (" ");
  52       if ((i + 1) % 8 == 0 && i + 1 < len)
  53         printf ("\n\t;; ");
  54     }
  55   puts ("");
  56 }
  57
  58 struct idna
  59 {
  60   char *name;
  61   size_t inlen;
  62   unsigned long in[100];
  63   char *out;
  64   int allowunassigned;
  65   int usestd3asciirules;
  66   int toasciirc;
  67   int tounicoderc;
  68 } idna[] =
  69 {
  70   {
  71     "(A) Arabic (Egyptian)", 17,
  72     {
  73   0x0644, 0x064A, 0x0647, 0x0645, 0x0627, 0x0628, 0x062A, 0x0643,
  74         0x0644, 0x0645, 0x0648, 0x0634, 0x0639, 0x0631, 0x0628, 0x064A,
  75         0x061F},
  76       IDNA_ACE_PREFIX "egbpdaj6bu4bxfgehfvwxn", 0, 0, IDNA_SUCCESS,
  77       IDNA_SUCCESS},
  78   {
  79     "(B) Chinese (simplified)", 9,
  80     {
  81   0x4ED6, 0x4EEC, 0x4E3A, 0x4EC0, 0x4E48, 0x4E0D, 0x8BF4, 0x4E2D, 0x6587},
  82       IDNA_ACE_PREFIX "ihqwcrb4cv8a8dqg056pqjye", 0, 0, IDNA_SUCCESS,
  83       IDNA_SUCCESS},
  84   {
  85     "(C) Chinese (traditional)", 9,
  86     {
  87   0x4ED6, 0x5011, 0x7232, 0x4EC0, 0x9EBD, 0x4E0D, 0x8AAA, 0x4E2D, 0x6587},
  88       IDNA_ACE_PREFIX "ihqwctvzc91f659drss3x8bo0yb", 0, 0, IDNA_SUCCESS,
  89       IDNA_SUCCESS},
  90   {
  91     "(D) Czech: Pro<ccaron>prost<ecaron>nemluv<iacute><ccaron>esky", 22,
  92     {
  93   0x0050, 0x0072, 0x006F, 0x010D, 0x0070, 0x0072, 0x006F, 0x0073,
  94         0x0074, 0x011B, 0x006E, 0x0065, 0x006D, 0x006C, 0x0075, 0x0076,
  95         0x00ED, 0x010D, 0x0065, 0x0073, 0x006B, 0x0079},
  96       IDNA_ACE_PREFIX "Proprostnemluvesky-uyb24dma41a", 0, 0, IDNA_SUCCESS,
  97       IDNA_SUCCESS},
  98   {
  99     "(E) Hebrew:", 22,
 100     {
 101   0x05DC, 0x05DE, 0x05D4, 0x05D4, 0x05DD, 0x05E4, 0x05E9, 0x05D5,
 102         0x05D8, 0x05DC, 0x05D0, 0x05DE, 0x05D3, 0x05D1, 0x05E8, 0x05D9,
 103         0x05DD, 0x05E2, 0x05D1, 0x05E8, 0x05D9, 0x05EA},
 104       IDNA_ACE_PREFIX "4dbcagdahymbxekheh6e0a7fei0b", 0, 0, IDNA_SUCCESS,
 105       IDNA_SUCCESS},
 106   {
 107     "(F) Hindi (Devanagari):", 30,
 108     {
 109   0x092F, 0x0939, 0x0932, 0x094B, 0x0917, 0x0939, 0x093F, 0x0928,
 110         0x094D, 0x0926, 0x0940, 0x0915, 0x094D, 0x092F, 0x094B, 0x0902,
 111         0x0928, 0x0939, 0x0940, 0x0902, 0x092C, 0x094B, 0x0932, 0x0938,
 112         0x0915, 0x0924, 0x0947, 0x0939, 0x0948, 0x0902},
 113       IDNA_ACE_PREFIX "i1baa7eci9glrd9b2ae1bj0hfcgg6iyaf8o0a1dig0cd", 0, 0,
 114       IDNA_SUCCESS},
 115   {
 116     "(G) Japanese (kanji and hiragana):", 18,
 117     {
 118   0x306A, 0x305C, 0x307F, 0x3093, 0x306A, 0x65E5, 0x672C, 0x8A9E,
 119         0x3092, 0x8A71, 0x3057, 0x3066, 0x304F, 0x308C, 0x306A, 0x3044,
 120         0x306E, 0x304B},
 121       IDNA_ACE_PREFIX "n8jok5ay5dzabd5bym9f0cm5685rrjetr6pdxa", 0, 0,
 122       IDNA_SUCCESS},
 123   {
 124     "(H) Korean (Hangul syllables):", 24,
 125     {
 126   0xC138, 0xACC4, 0xC758, 0xBAA8, 0xB4E0, 0xC0AC, 0xB78C, 0xB4E4,
 127         0xC774, 0xD55C, 0xAD6D, 0xC5B4, 0xB97C, 0xC774, 0xD574, 0xD55C,
 128         0xB2E4, 0xBA74, 0xC5BC, 0xB9C8, 0xB098, 0xC88B, 0xC744, 0xAE4C},
 129       IDNA_ACE_PREFIX "989aomsvi5e83db1d2a355cv1e0vak1dwrv93d5xbh15a0dt"
 130       "30a5jpsd879ccm6fea98c", 0, 0, IDNA_PUNYCODE_ERROR,
 131       IDNA_PUNYCODE_ERROR},
 132     /* too long output */
 133   {
 134     "(I) Russian (Cyrillic):", 28,
 135     {
 136   0x043F, 0x043E, 0x0447, 0x0435, 0x043C, 0x0443, 0x0436, 0x0435,
 137         0x043E, 0x043D, 0x0438, 0x043D, 0x0435, 0x0433, 0x043E, 0x0432,
 138         0x043E, 0x0440, 0x044F, 0x0442, 0x043F, 0x043E, 0x0440, 0x0443,
 139         0x0441, 0x0441, 0x043A, 0x0438},
 140       IDNA_ACE_PREFIX "b1abfaaepdrnnbgefbadotcwatmq2g4l", 0, 0,
 141       IDNA_SUCCESS, IDNA_SUCCESS},
 142   {
 143     "(J) Spanish: Porqu<eacute>nopuedensimplementehablarenEspa<ntilde>ol", 40,
 144     {
 145   0x0050, 0x006F, 0x0072, 0x0071, 0x0075, 0x00E9, 0x006E, 0x006F,
 146         0x0070, 0x0075, 0x0065, 0x0064, 0x0065, 0x006E, 0x0073, 0x0069,
 147         0x006D, 0x0070, 0x006C, 0x0065, 0x006D, 0x0065, 0x006E, 0x0074,
 148         0x0065, 0x0068, 0x0061, 0x0062, 0x006C, 0x0061, 0x0072, 0x0065,
 149         0x006E, 0x0045, 0x0073, 0x0070, 0x0061, 0x00F1, 0x006F, 0x006C},
 150       IDNA_ACE_PREFIX "PorqunopuedensimplementehablarenEspaol-fmd56a", 0, 0,
 151       IDNA_SUCCESS},
 152   {
 153     "(K) Vietnamese:", 31,
 154     {
 155   0x0054, 0x1EA1, 0x0069, 0x0073, 0x0061, 0x006F, 0x0068, 0x1ECD,
 156         0x006B, 0x0068, 0x00F4, 0x006E, 0x0067, 0x0074, 0x0068, 0x1EC3,
 157         0x0063, 0x0068, 0x1EC9, 0x006E, 0x00F3, 0x0069, 0x0074, 0x0069,
 158         0x1EBF, 0x006E, 0x0067, 0x0056, 0x0069, 0x1EC7, 0x0074},
 159       IDNA_ACE_PREFIX "TisaohkhngthchnitingVit-kjcr8268qyxafd2f1b9g", 0, 0,
 160       IDNA_SUCCESS},
 161   {
 162     "(L) 3<nen>B<gumi><kinpachi><sensei>", 8,
 163     {
 164   0x0033, 0x5E74, 0x0042, 0x7D44, 0x91D1, 0x516B, 0x5148, 0x751F},
 165       IDNA_ACE_PREFIX "3B-ww4c5e180e575a65lsy2b", 0, 0, IDNA_SUCCESS,
 166       IDNA_SUCCESS},
 167   {
 168     "(M) <amuro><namie>-with-SUPER-MONKEYS", 24,
 169     {
 170   0x5B89, 0x5BA4, 0x5948, 0x7F8E, 0x6075, 0x002D, 0x0077, 0x0069,
 171         0x0074, 0x0068, 0x002D, 0x0053, 0x0055, 0x0050, 0x0045, 0x0052,
 172         0x002D, 0x004D, 0x004F, 0x004E, 0x004B, 0x0045, 0x0059, 0x0053},
 173       IDNA_ACE_PREFIX "-with-SUPER-MONKEYS-pc58ag80a8qai00g7n9n", 0, 0,
 174       IDNA_SUCCESS},
 175   {
 176     "(N) Hello-Another-Way-<sorezore><no><basho>", 25,
 177     {
 178   0x0048, 0x0065, 0x006C, 0x006C, 0x006F, 0x002D, 0x0041, 0x006E,
 179         0x006F, 0x0074, 0x0068, 0x0065, 0x0072, 0x002D, 0x0057, 0x0061,
 180         0x0079, 0x002D, 0x305D, 0x308C, 0x305E, 0x308C, 0x306E, 0x5834,
 181         0x6240},
 182       IDNA_ACE_PREFIX "Hello-Another-Way--fc4qua05auwb3674vfr0b", 0, 0,
 183       IDNA_SUCCESS},
 184   {
 185     "(O) <hitotsu><yane><no><shita>2", 8,
 186     {
 187   0x3072, 0x3068, 0x3064, 0x5C4B, 0x6839, 0x306E, 0x4E0B, 0x0032},
 188       IDNA_ACE_PREFIX "2-u9tlzr9756bt3uc0v", 0, 0, IDNA_SUCCESS,
 189       IDNA_SUCCESS},
 190   {
 191     "(P) Maji<de>Koi<suru>5<byou><mae>", 13,
 192     {
 193   0x004D, 0x0061, 0x006A, 0x0069, 0x3067, 0x004B, 0x006F, 0x0069,
 194         0x3059, 0x308B, 0x0035, 0x79D2, 0x524D},
 195       IDNA_ACE_PREFIX "MajiKoi5-783gue6qz075azm5e", 0, 0, IDNA_SUCCESS,
 196       IDNA_SUCCESS},
 197   {
 198     "(Q) <pafii>de<runba>", 9,
 199     {
 200   0x30D1, 0x30D5, 0x30A3, 0x30FC, 0x0064, 0x0065, 0x30EB, 0x30F3, 0x30D0},
 201       IDNA_ACE_PREFIX "de-jg4avhby1noc0d", 0, 0, IDNA_SUCCESS, IDNA_SUCCESS},
 202   {
 203     "(R) <sono><supiido><de>", 7,
 204     {
 205   0x305D, 0x306E, 0x30B9, 0x30D4, 0x30FC, 0x30C9, 0x3067},
 206       IDNA_ACE_PREFIX "d9juau41awczczp", 0, 0, IDNA_SUCCESS, IDNA_SUCCESS},
 207 #if 0
 208   {
 209     "(S) -> $1.00 <-", 11,
 210     {
 211   0x002D, 0x003E, 0x0020, 0x0024, 0x0031, 0x002E, 0x0030, 0x0030,
 212         0x0020, 0x003C, 0x002D},
 213       IDNA_ACE_PREFIX "-> $1.00 <--", 0, 0, IDNA_SUCCESS, IDNA_SUCCESS},
 214 #endif
 215   {                             /* XXX depends on IDNA_ACE_PREFIX */
 216     "ToASCII() with ACE prefix", 4 + 3,
 217     { 'x', 'n', '-', '-', 'f', 'o', 0x3067},
 218       IDNA_ACE_PREFIX "too long too long too long too long too long too "
 219       "long too long too long too long too long ", 0, 0,
 220       IDNA_CONTAINS_ACE_PREFIX, IDNA_PUNYCODE_ERROR}
 221 };
 222
 223 int
 224 main (int argc, char *argv[])
 225 {
 226   char label[100];
 227   unsigned long *ucs4label = NULL;
 228   unsigned long tmp[100];
 229   size_t len, len2, i;
 230   int rc;
 231
 232   do
 233     if (strcmp (argv[argc - 1], "-v") == 0 ||
 234         strcmp (argv[argc - 1], "--verbose") == 0)
 235       debug = 1;
 236     else if (strcmp (argv[argc - 1], "-b") == 0 ||
 237              strcmp (argv[argc - 1], "--break-on-error") == 0)
 238       break_on_error = 1;
 239     else if (strcmp (argv[argc - 1], "-h") == 0 ||
 240              strcmp (argv[argc - 1], "-?") == 0 ||
 241              strcmp (argv[argc - 1], "--help") == 0)
 242       {
 243         printf ("Usage: %s [-vbh?] [--verbose] [--break-on-error] [--help]\n",
 244                 argv[0]);
 245         return 1;
 246       }
 247   while (argc-- > 1);
 248
 249   for (i = 0; i < sizeof (idna) / sizeof (idna[0]); i++)
 250     {
 251       if (debug)
 252         printf ("IDNA entry %d: %s\n", i, idna[i].name);
 253
 254       if (debug)
 255         {
 256           printf ("in:\n");
 257           ucs4print (idna[i].in, idna[i].inlen);
 258         }
 259
 260       rc = idna_to_ascii (idna[i].in, idna[i].inlen, label,
 261                           idna[i].allowunassigned, idna[i].usestd3asciirules);
 262       if (rc != idna[i].toasciirc)
 263         {
 264           fail ("IDNA entry %d failed: %d\n", i, rc);
 265           if (debug)
 266             printf ("FATAL\n");
 267           continue;
 268         }
 269
 270       if (debug && rc == IDNA_SUCCESS)
 271         {
 272           printf ("computed out: %s\n", label);
 273           printf ("expected out: %s\n", idna[i].out);
 274         }
 275       else if (debug)
 276         printf ("returned %d expected %d\n", rc, idna[i].toasciirc);
 277
 278       if (rc == IDNA_SUCCESS)
 279         {
 280           if (strlen (idna[i].out) != strlen (label) ||
 281               strcasecmp (idna[i].out, label) != 0)
 282             {
 283               fail ("IDNA entry %d failed\n", i);
 284               if (debug)
 285                 printf ("ERROR\n");
 286             }
 287           else if (debug)
 288             printf ("OK\n");
 289         }
 290       else if (debug)
 291         printf ("OK\n");
 292
 293       if (ucs4label)
 294         free(ucs4label);
 295
 296       ucs4label = stringprep_utf8_to_ucs4 (idna[i].out, -1, &len);
 297
 298       if (debug)
 299         {
 300           printf ("in: %s (%d==%d)\n", idna[i].out, strlen (idna[i].out),
 301                   len);
 302           ucs4print (ucs4label, len);
 303         }
 304
 305       len2 = sizeof (tmp);
 306       rc = idna_to_unicode (ucs4label, len, tmp, &len2,
 307                             idna[i].allowunassigned,
 308                             idna[i].usestd3asciirules);
 309       if (debug)
 310         {
 311           printf ("expected out:\n");
 312           if (rc == IDNA_SUCCESS)
 313             ucs4print (idna[i].in, idna[i].inlen);
 314           else
 315             ucs4print (ucs4label, len);
 316
 317           printf ("computed out:\n");
 318           ucs4print (tmp, len2);
 319         }
 320
 321       if (rc != idna[i].tounicoderc)
 322         {
 323           fail ("IDNA entry %d failed: %d\n", i, rc);
 324           if (debug)
 325             printf ("FATAL\n");
 326           continue;
 327         }
 328
 329       if ((rc == IDNA_SUCCESS && (len2 != idna[i].inlen ||
 330                                   memcmp (idna[i].in, tmp, len2) != 0)) ||
 331           (rc != IDNA_SUCCESS && (len2 != len ||
 332                                   memcmp (ucs4label, tmp, len) != 0)))
 333         {
 334           fail ("IDNA entry %d failed\n", i);
 335           if (debug)
 336             printf ("ERROR\n");
 337         }
 338       else if (debug)
 339         printf ("OK\n\n");
 340
 341     }
 342
 343   if (debug)
 344     printf ("IDNA self tests done with %d errors\n", error_count);
 345
 346   return error_count ? 1 : 0;
 347 }