tst_punycode.c

   1 /* tst_punycode.c       Self tests for punycode.
   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 (uint32_t *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 punycode
  59 {
  60   char *name;
  61   size_t inlen;
  62   uint32_t in[100];
  63   char *out;
  64   int rc;
  65 } punycode[] =
  66 {
  67   {
  68     "(A) Arabic (Egyptian)", 17,
  69     {
  70   0x0644, 0x064A, 0x0647, 0x0645, 0x0627, 0x0628, 0x062A, 0x0643,
  71         0x0644, 0x0645, 0x0648, 0x0634, 0x0639, 0x0631, 0x0628, 0x064A,
  72         0x061F}, "egbpdaj6bu4bxfgehfvwxn", PUNYCODE_SUCCESS},
  73   {
  74     "(B) Chinese (simplified)", 9,
  75     {
  76   0x4ED6, 0x4EEC, 0x4E3A, 0x4EC0, 0x4E48, 0x4E0D, 0x8BF4, 0x4E2D,
  77         0x6587}, "ihqwcrb4cv8a8dqg056pqjye", PUNYCODE_SUCCESS},
  78   {
  79     "(C) Chinese (traditional)", 9,
  80     {
  81   0x4ED6, 0x5011, 0x7232, 0x4EC0, 0x9EBD, 0x4E0D, 0x8AAA, 0x4E2D,
  82         0x6587}, "ihqwctvzc91f659drss3x8bo0yb", PUNYCODE_SUCCESS},
  83   {
  84     "(D) Czech: Pro<ccaron>prost<ecaron>nemluv<iacute><ccaron>esky", 22,
  85     {
  86   0x0050, 0x0072, 0x006F, 0x010D, 0x0070, 0x0072, 0x006F, 0x0073,
  87         0x0074, 0x011B, 0x006E, 0x0065, 0x006D, 0x006C, 0x0075, 0x0076,
  88         0x00ED, 0x010D, 0x0065, 0x0073, 0x006B, 0x0079},
  89       "Proprostnemluvesky-uyb24dma41a", PUNYCODE_SUCCESS},
  90   {
  91     "(E) Hebrew:", 22,
  92     {
  93   0x05DC, 0x05DE, 0x05D4, 0x05D4, 0x05DD, 0x05E4, 0x05E9, 0x05D5,
  94         0x05D8, 0x05DC, 0x05D0, 0x05DE, 0x05D3, 0x05D1, 0x05E8, 0x05D9,
  95         0x05DD, 0x05E2, 0x05D1, 0x05E8, 0x05D9, 0x05EA},
  96       "4dbcagdahymbxekheh6e0a7fei0b", PUNYCODE_SUCCESS},
  97   {
  98     "(F) Hindi (Devanagari):", 30,
  99     {
 100   0x092F, 0x0939, 0x0932, 0x094B, 0x0917, 0x0939, 0x093F, 0x0928,
 101         0x094D, 0x0926, 0x0940, 0x0915, 0x094D, 0x092F, 0x094B, 0x0902,
 102         0x0928, 0x0939, 0x0940, 0x0902, 0x092C, 0x094B, 0x0932, 0x0938,
 103         0x0915, 0x0924, 0x0947, 0x0939, 0x0948, 0x0902},
 104       "i1baa7eci9glrd9b2ae1bj0hfcgg6iyaf8o0a1dig0cd", PUNYCODE_SUCCESS},
 105   {
 106     "(G) Japanese (kanji and hiragana):", 18,
 107     {
 108   0x306A, 0x305C, 0x307F, 0x3093, 0x306A, 0x65E5, 0x672C, 0x8A9E,
 109         0x3092, 0x8A71, 0x3057, 0x3066, 0x304F, 0x308C, 0x306A, 0x3044,
 110         0x306E, 0x304B},
 111       "n8jok5ay5dzabd5bym9f0cm5685rrjetr6pdxa", PUNYCODE_SUCCESS},
 112   {
 113     "(H) Korean (Hangul syllables):", 24,
 114     {
 115   0xC138, 0xACC4, 0xC758, 0xBAA8, 0xB4E0, 0xC0AC, 0xB78C, 0xB4E4,
 116         0xC774, 0xD55C, 0xAD6D, 0xC5B4, 0xB97C, 0xC774, 0xD574, 0xD55C,
 117         0xB2E4, 0xBA74, 0xC5BC, 0xB9C8, 0xB098, 0xC88B, 0xC744, 0xAE4C},
 118       "989aomsvi5e83db1d2a355cv1e0vak1dwrv93d5xbh15a0dt30a5jpsd879ccm6fea98c",
 119       PUNYCODE_SUCCESS},
 120   {
 121     "(I) Russian (Cyrillic):", 28,
 122     {
 123   0x043F, 0x043E, 0x0447, 0x0435, 0x043C, 0x0443, 0x0436, 0x0435,
 124         0x043E, 0x043D, 0x0438, 0x043D, 0x0435, 0x0433, 0x043E, 0x0432,
 125         0x043E, 0x0440, 0x044F, 0x0442, 0x043F, 0x043E, 0x0440, 0x0443,
 126         0x0441, 0x0441, 0x043A, 0x0438},
 127       "b1abfaaepdrnnbgefbadotcwatmq2g4l", PUNYCODE_SUCCESS},
 128   {
 129     "(J) Spanish: Porqu<eacute>nopuedensimplementehablarenEspa<ntilde>ol", 40,
 130     {
 131   0x0050, 0x006F, 0x0072, 0x0071, 0x0075, 0x00E9, 0x006E, 0x006F,
 132         0x0070, 0x0075, 0x0065, 0x0064, 0x0065, 0x006E, 0x0073, 0x0069,
 133         0x006D, 0x0070, 0x006C, 0x0065, 0x006D, 0x0065, 0x006E, 0x0074,
 134         0x0065, 0x0068, 0x0061, 0x0062, 0x006C, 0x0061, 0x0072, 0x0065,
 135         0x006E, 0x0045, 0x0073, 0x0070, 0x0061, 0x00F1, 0x006F, 0x006C},
 136       "PorqunopuedensimplementehablarenEspaol-fmd56a", PUNYCODE_SUCCESS},
 137   {
 138     "(K) Vietnamese:", 31,
 139     {
 140   0x0054, 0x1EA1, 0x0069, 0x0073, 0x0061, 0x006F, 0x0068, 0x1ECD,
 141         0x006B, 0x0068, 0x00F4, 0x006E, 0x0067, 0x0074, 0x0068, 0x1EC3,
 142         0x0063, 0x0068, 0x1EC9, 0x006E, 0x00F3, 0x0069, 0x0074, 0x0069,
 143         0x1EBF, 0x006E, 0x0067, 0x0056, 0x0069, 0x1EC7, 0x0074},
 144       "TisaohkhngthchnitingVit-kjcr8268qyxafd2f1b9g", PUNYCODE_SUCCESS},
 145   {
 146     "(L) 3<nen>B<gumi><kinpachi><sensei>", 8,
 147     {
 148   0x0033, 0x5E74, 0x0042, 0x7D44, 0x91D1, 0x516B, 0x5148, 0x751F},
 149       "3B-ww4c5e180e575a65lsy2b", PUNYCODE_SUCCESS},
 150   {
 151     "(M) <amuro><namie>-with-SUPER-MONKEYS", 24,
 152     {
 153   0x5B89, 0x5BA4, 0x5948, 0x7F8E, 0x6075, 0x002D, 0x0077, 0x0069,
 154         0x0074, 0x0068, 0x002D, 0x0053, 0x0055, 0x0050, 0x0045, 0x0052,
 155         0x002D, 0x004D, 0x004F, 0x004E, 0x004B, 0x0045, 0x0059, 0x0053},
 156       "-with-SUPER-MONKEYS-pc58ag80a8qai00g7n9n", PUNYCODE_SUCCESS},
 157   {
 158     "(N) Hello-Another-Way-<sorezore><no><basho>", 25,
 159     {
 160   0x0048, 0x0065, 0x006C, 0x006C, 0x006F, 0x002D, 0x0041, 0x006E,
 161         0x006F, 0x0074, 0x0068, 0x0065, 0x0072, 0x002D, 0x0057, 0x0061,
 162         0x0079, 0x002D, 0x305D, 0x308C, 0x305E, 0x308C, 0x306E, 0x5834,
 163         0x6240},
 164       "Hello-Another-Way--fc4qua05auwb3674vfr0b", PUNYCODE_SUCCESS},
 165   {
 166     "(O) <hitotsu><yane><no><shita>2", 8,
 167     {
 168   0x3072, 0x3068, 0x3064, 0x5C4B, 0x6839, 0x306E, 0x4E0B, 0x0032},
 169       "2-u9tlzr9756bt3uc0v", PUNYCODE_SUCCESS},
 170   {
 171     "(P) Maji<de>Koi<suru>5<byou><mae>", 13,
 172     {
 173   0x004D, 0x0061, 0x006A, 0x0069, 0x3067, 0x004B, 0x006F, 0x0069,
 174         0x3059, 0x308B, 0x0035, 0x79D2, 0x524D},
 175       "MajiKoi5-783gue6qz075azm5e", PUNYCODE_SUCCESS},
 176   {
 177     "(Q) <pafii>de<runba>", 9,
 178     {
 179   0x30D1, 0x30D5, 0x30A3, 0x30FC, 0x0064, 0x0065, 0x30EB, 0x30F3, 0x30D0},
 180       "de-jg4avhby1noc0d", PUNYCODE_SUCCESS},
 181   {
 182     "(R) <sono><supiido><de>", 7,
 183     {
 184   0x305D, 0x306E, 0x30B9, 0x30D4, 0x30FC, 0x30C9, 0x3067},
 185       "d9juau41awczczp", PUNYCODE_SUCCESS},
 186   {
 187     "(S) -> $1.00 <-", 11,
 188     {
 189   0x002D, 0x003E, 0x0020, 0x0024, 0x0031, 0x002E, 0x0030, 0x0030,
 190         0x0020, 0x003C, 0x002D}, "-> $1.00 <--", PUNYCODE_SUCCESS}
 191 };
 192
 193 int
 194 main (int argc, char *argv[])
 195 {
 196   char *p;
 197   uint32_t *q;
 198   int rc;
 199   size_t i, outlen;
 200
 201   do
 202     if (strcmp (argv[argc - 1], "-v") == 0 ||
 203         strcmp (argv[argc - 1], "--verbose") == 0)
 204       debug = 1;
 205     else if (strcmp (argv[argc - 1], "-b") == 0 ||
 206              strcmp (argv[argc - 1], "--break-on-error") == 0)
 207       break_on_error = 1;
 208     else if (strcmp (argv[argc - 1], "-h") == 0 ||
 209              strcmp (argv[argc - 1], "-?") == 0 ||
 210              strcmp (argv[argc - 1], "--help") == 0)
 211       {
 212         printf ("Usage: %s [-vbh?] [--verbose] [--break-on-error] [--help]\n",
 213                 argv[0]);
 214         return 1;
 215       }
 216   while (argc-- > 1);
 217
 218   p = malloc (sizeof (*p) * BUFSIZ);
 219   if (p == NULL)
 220     fail ("malloc() returned NULL\n");
 221
 222   q = malloc (sizeof (*q) * BUFSIZ);
 223   if (q == NULL)
 224     fail ("malloc() returned NULL\n");
 225
 226   for (i = 0; i < sizeof (punycode) / sizeof (punycode[0]); i++)
 227     {
 228       if (debug)
 229         printf ("PUNYCODE entry %d: %s\n", i, punycode[i].name);
 230
 231       if (debug)
 232         {
 233           printf ("in:\n");
 234           ucs4print (punycode[i].in, punycode[i].inlen);
 235         }
 236
 237       outlen = BUFSIZ;
 238       rc = punycode_encode (punycode[i].inlen, punycode[i].in,
 239                             NULL, &outlen, p);
 240       if (rc != punycode[i].rc)
 241         {
 242           fail ("punycode_encode() entry %d failed: %d\n", i, rc);
 243           if (debug)
 244             printf ("FATAL\n");
 245           continue;
 246         }
 247
 248       if (rc == PUNYCODE_SUCCESS)
 249         p[outlen] = '\0';
 250
 251       if (debug && rc == PUNYCODE_SUCCESS)
 252         {
 253           printf ("computed out: %s\n", p);
 254           printf ("expected out: %s\n", punycode[i].out);
 255         }
 256       else if (debug)
 257         printf ("returned %d expected %d\n", rc, punycode[i].rc);
 258
 259       if (rc == PUNYCODE_SUCCESS)
 260         {
 261           if (strlen (punycode[i].out) != strlen (p) ||
 262               memcmp (punycode[i].out, p, strlen (p)) != 0)
 263             {
 264               fail ("punycode() entry %d failed\n", i);
 265               if (debug)
 266                 printf ("ERROR\n");
 267             }
 268           else if (debug)
 269             printf ("OK\n\n");
 270         }
 271       else if (debug)
 272         printf ("OK\n\n");
 273
 274       if (debug)
 275         {
 276           printf ("in: %s\n", punycode[i].out);
 277         }
 278
 279       outlen = BUFSIZ;
 280       rc = punycode_decode (strlen (punycode[i].out), punycode[i].out,
 281                             &outlen, q, NULL);
 282       if (rc != punycode[i].rc)
 283         {
 284           fail ("punycode() entry %d failed: %d\n", i, rc);
 285           if (debug)
 286             printf ("FATAL\n");
 287           continue;
 288         }
 289
 290       if (debug && rc == PUNYCODE_SUCCESS)
 291         {
 292           printf ("computed out:\n");
 293           ucs4print (q, outlen);
 294           printf ("expected out:\n");
 295           ucs4print (punycode[i].in, punycode[i].inlen);
 296         }
 297       else if (debug)
 298         printf ("returned %d expected %d\n", rc, punycode[i].rc);
 299
 300       if (rc == PUNYCODE_SUCCESS)
 301         {
 302           if (punycode[i].inlen != outlen ||
 303               memcmp (punycode[i].in, q, outlen) != 0)
 304             {
 305               fail ("punycode_decode() entry %d failed\n", i);
 306               if (debug)
 307                 printf ("ERROR\n");
 308             }
 309           else if (debug)
 310             printf ("OK\n\n");
 311         }
 312       else if (debug)
 313         printf ("OK\n\n");
 314     }
 315
 316   free (q);
 317   free (p);
 318
 319   if (debug)
 320     printf ("Punycode self tests done with %d errors\n", error_count);
 321
 322   return error_count ? 1 : 0;
 323 }