tst_punycode.c

   1 /* tst_punycode.c       Self tests for punycode.
   2  * Copyright (C) 2002  Simon Josefsson
   3  *
   4  * This file is part of Libstringprep.
   5  *
   6  * Libstringprep 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  * Libstringprep 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 Libstringprep; 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 punycode
  59 {
  60   char *name;
  61   size_t inlen;
  62   unsigned long in[100];
  63   char *out;
  64   int rc;
  65 }
  66 punycode[] =
  67 {
  68   {
  69     "(A) Arabic (Egyptian)", 17,
  70     {
  71     0x0644, 0x064A, 0x0647, 0x0645, 0x0627, 0x0628, 0x062A, 0x0643,
  72         0x0644, 0x0645, 0x0648, 0x0634, 0x0639, 0x0631, 0x0628, 0x064A,
  73         0x061F}
  74   , "egbpdaj6bu4bxfgehfvwxn", PUNYCODE_SUCCESS}
  75   ,
  76   {
  77     "(B) Chinese (simplified)", 9,
  78     {
  79     0x4ED6, 0x4EEC, 0x4E3A, 0x4EC0, 0x4E48, 0x4E0D, 0x8BF4, 0x4E2D, 0x6587}
  80   , "ihqwcrb4cv8a8dqg056pqjye", PUNYCODE_SUCCESS}
  81   ,
  82   {
  83     "(C) Chinese (traditional)", 9,
  84     {
  85     0x4ED6, 0x5011, 0x7232, 0x4EC0, 0x9EBD, 0x4E0D, 0x8AAA, 0x4E2D, 0x6587}
  86   , "ihqwctvzc91f659drss3x8bo0yb", PUNYCODE_SUCCESS}
  87   ,
  88   {
  89     "(D) Czech: Pro<ccaron>prost<ecaron>nemluv<iacute><ccaron>esky", 22,
  90     {
  91     0x0050, 0x0072, 0x006F, 0x010D, 0x0070, 0x0072, 0x006F, 0x0073,
  92         0x0074, 0x011B, 0x006E, 0x0065, 0x006D, 0x006C, 0x0075, 0x0076,
  93         0x00ED, 0x010D, 0x0065, 0x0073, 0x006B, 0x0079}
  94   , "Proprostnemluvesky-uyb24dma41a", PUNYCODE_SUCCESS}
  95   ,
  96   {
  97     "(E) Hebrew:", 22,
  98     {
  99     0x05DC, 0x05DE, 0x05D4, 0x05D4, 0x05DD, 0x05E4, 0x05E9, 0x05D5,
 100         0x05D8, 0x05DC, 0x05D0, 0x05DE, 0x05D3, 0x05D1, 0x05E8, 0x05D9,
 101         0x05DD, 0x05E2, 0x05D1, 0x05E8, 0x05D9, 0x05EA}
 102   , "4dbcagdahymbxekheh6e0a7fei0b", PUNYCODE_SUCCESS}
 103   ,
 104   {
 105     "(F) Hindi (Devanagari):", 30,
 106     {
 107     0x092F, 0x0939, 0x0932, 0x094B, 0x0917, 0x0939, 0x093F, 0x0928,
 108         0x094D, 0x0926, 0x0940, 0x0915, 0x094D, 0x092F, 0x094B, 0x0902,
 109         0x0928, 0x0939, 0x0940, 0x0902, 0x092C, 0x094B, 0x0932, 0x0938,
 110         0x0915, 0x0924, 0x0947, 0x0939, 0x0948, 0x0902}
 111   , "i1baa7eci9glrd9b2ae1bj0hfcgg6iyaf8o0a1dig0cd", PUNYCODE_SUCCESS}
 112   ,
 113   {
 114     "(G) Japanese (kanji and hiragana):", 18,
 115     {
 116     0x306A, 0x305C, 0x307F, 0x3093, 0x306A, 0x65E5, 0x672C, 0x8A9E,
 117         0x3092, 0x8A71, 0x3057, 0x3066, 0x304F, 0x308C, 0x306A, 0x3044,
 118         0x306E, 0x304B}
 119   , "n8jok5ay5dzabd5bym9f0cm5685rrjetr6pdxa", PUNYCODE_SUCCESS}
 120   ,
 121   {
 122     "(H) Korean (Hangul syllables):", 24,
 123     {
 124     0xC138, 0xACC4, 0xC758, 0xBAA8, 0xB4E0, 0xC0AC, 0xB78C, 0xB4E4,
 125         0xC774, 0xD55C, 0xAD6D, 0xC5B4, 0xB97C, 0xC774, 0xD574, 0xD55C,
 126         0xB2E4, 0xBA74, 0xC5BC, 0xB9C8, 0xB098, 0xC88B, 0xC744, 0xAE4C}
 127   , "989aomsvi5e83db1d2a355cv1e0vak1dwrv93d5xbh15a0dt30a5jpsd879ccm6fea98c",
 128       PUNYCODE_SUCCESS}
 129   ,
 130   {
 131     "(I) Russian (Cyrillic):", 28,
 132     {
 133     0x043F, 0x043E, 0x0447, 0x0435, 0x043C, 0x0443, 0x0436, 0x0435,
 134         0x043E, 0x043D, 0x0438, 0x043D, 0x0435, 0x0433, 0x043E, 0x0432,
 135         0x043E, 0x0440, 0x044F, 0x0442, 0x043F, 0x043E, 0x0440, 0x0443,
 136         0x0441, 0x0441, 0x043A, 0x0438}
 137   , "b1abfaaepdrnnbgefbadotcwatmq2g4l", PUNYCODE_SUCCESS}
 138   ,
 139   {
 140     "(J) Spanish: Porqu<eacute>nopuedensimplementehablarenEspa<ntilde>ol", 40,
 141     {
 142     0x0050, 0x006F, 0x0072, 0x0071, 0x0075, 0x00E9, 0x006E, 0x006F,
 143         0x0070, 0x0075, 0x0065, 0x0064, 0x0065, 0x006E, 0x0073, 0x0069,
 144         0x006D, 0x0070, 0x006C, 0x0065, 0x006D, 0x0065, 0x006E, 0x0074,
 145         0x0065, 0x0068, 0x0061, 0x0062, 0x006C, 0x0061, 0x0072, 0x0065,
 146         0x006E, 0x0045, 0x0073, 0x0070, 0x0061, 0x00F1, 0x006F, 0x006C}
 147   , "PorqunopuedensimplementehablarenEspaol-fmd56a", PUNYCODE_SUCCESS}
 148   ,
 149   {
 150     "(K) Vietnamese:", 31,
 151     {
 152     0x0054, 0x1EA1, 0x0069, 0x0073, 0x0061, 0x006F, 0x0068, 0x1ECD,
 153         0x006B, 0x0068, 0x00F4, 0x006E, 0x0067, 0x0074, 0x0068, 0x1EC3,
 154         0x0063, 0x0068, 0x1EC9, 0x006E, 0x00F3, 0x0069, 0x0074, 0x0069,
 155         0x1EBF, 0x006E, 0x0067, 0x0056, 0x0069, 0x1EC7, 0x0074}
 156   , "TisaohkhngthchnitingVit-kjcr8268qyxafd2f1b9g", PUNYCODE_SUCCESS}
 157   ,
 158   {
 159     "(L) 3<nen>B<gumi><kinpachi><sensei>", 8,
 160     {
 161     0x0033, 0x5E74, 0x0042, 0x7D44, 0x91D1, 0x516B, 0x5148, 0x751F}
 162   , "3B-ww4c5e180e575a65lsy2b", PUNYCODE_SUCCESS}
 163   ,
 164   {
 165     "(M) <amuro><namie>-with-SUPER-MONKEYS", 24,
 166     {
 167     0x5B89, 0x5BA4, 0x5948, 0x7F8E, 0x6075, 0x002D, 0x0077, 0x0069,
 168         0x0074, 0x0068, 0x002D, 0x0053, 0x0055, 0x0050, 0x0045, 0x0052,
 169         0x002D, 0x004D, 0x004F, 0x004E, 0x004B, 0x0045, 0x0059, 0x0053}
 170   , "-with-SUPER-MONKEYS-pc58ag80a8qai00g7n9n", PUNYCODE_SUCCESS}
 171   ,
 172   {
 173     "(N) Hello-Another-Way-<sorezore><no><basho>", 25,
 174     {
 175     0x0048, 0x0065, 0x006C, 0x006C, 0x006F, 0x002D, 0x0041, 0x006E,
 176         0x006F, 0x0074, 0x0068, 0x0065, 0x0072, 0x002D, 0x0057, 0x0061,
 177         0x0079, 0x002D, 0x305D, 0x308C, 0x305E, 0x308C, 0x306E, 0x5834,
 178         0x6240}
 179   , "Hello-Another-Way--fc4qua05auwb3674vfr0b", PUNYCODE_SUCCESS}
 180   ,
 181   {
 182     "(O) <hitotsu><yane><no><shita>2", 8,
 183     {
 184     0x3072, 0x3068, 0x3064, 0x5C4B, 0x6839, 0x306E, 0x4E0B, 0x0032}
 185   , "2-u9tlzr9756bt3uc0v", PUNYCODE_SUCCESS}
 186   ,
 187   {
 188     "(P) Maji<de>Koi<suru>5<byou><mae>", 13,
 189     {
 190     0x004D, 0x0061, 0x006A, 0x0069, 0x3067, 0x004B, 0x006F, 0x0069,
 191         0x3059, 0x308B, 0x0035, 0x79D2, 0x524D}
 192   , "MajiKoi5-783gue6qz075azm5e", PUNYCODE_SUCCESS}
 193   ,
 194   {
 195     "(Q) <pafii>de<runba>", 9,
 196     {
 197     0x30D1, 0x30D5, 0x30A3, 0x30FC, 0x0064, 0x0065, 0x30EB, 0x30F3, 0x30D0}
 198   , "de-jg4avhby1noc0d", PUNYCODE_SUCCESS}
 199   ,
 200   {
 201     "(R) <sono><supiido><de>", 7,
 202     {
 203     0x305D, 0x306E, 0x30B9, 0x30D4, 0x30FC, 0x30C9, 0x3067}
 204   , "d9juau41awczczp", PUNYCODE_SUCCESS}
 205   ,
 206   {
 207     "(S) -> $1.00 <-", 11,
 208     {
 209     0x002D, 0x003E, 0x0020, 0x0024, 0x0031, 0x002E, 0x0030, 0x0030,
 210         0x0020, 0x003C, 0x002D}
 211   , "-> $1.00 <--", PUNYCODE_SUCCESS}
 212 };
 213
 214 int
 215 main (int argc, char *argv[])
 216 {
 217   char *p;
 218   unsigned long *q;
 219   int rc, i;
 220   size_t outlen;
 221
 222   do
 223     if (strcmp (argv[argc - 1], "-v") == 0 ||
 224         strcmp (argv[argc - 1], "--verbose") == 0)
 225       debug = 1;
 226     else if (strcmp (argv[argc - 1], "-b") == 0 ||
 227              strcmp (argv[argc - 1], "--break-on-error") == 0)
 228       break_on_error = 1;
 229     else if (strcmp (argv[argc - 1], "-h") == 0 ||
 230              strcmp (argv[argc - 1], "-?") == 0 ||
 231              strcmp (argv[argc - 1], "--help") == 0)
 232       {
 233         printf ("Usage: %s [-vbh?] [--verbose] [--break-on-error] [--help]\n",
 234                 argv[0]);
 235         return 1;
 236       }
 237   while (argc-- > 1);
 238
 239   p = malloc (sizeof(*p) * BUFSIZ);
 240   if (p == NULL)
 241     fail ("malloc() returned NULL\n");
 242
 243   q = malloc (sizeof(*q) * BUFSIZ);
 244   if (q == NULL)
 245     fail ("malloc() returned NULL\n");
 246
 247   for (i = 0; i < sizeof (punycode) / sizeof (punycode[0]); i++)
 248     {
 249       if (debug)
 250         printf ("PUNYCODE entry %d: %s\n", i, punycode[i].name);
 251
 252       if (debug)
 253         {
 254           printf ("in:\n");
 255           ucs4print (punycode[i].in, punycode[i].inlen);
 256         }
 257
 258       outlen = BUFSIZ;
 259       rc = punycode_encode (punycode[i].inlen, punycode[i].in,
 260                             NULL, &outlen, p);
 261       if (rc != punycode[i].rc)
 262         {
 263           fail ("punycode_encode() entry %d failed: %d\n", i, rc);
 264           if (debug)
 265             printf ("FATAL\n");
 266           continue;
 267         }
 268
 269       if (rc == PUNYCODE_SUCCESS)
 270         p[outlen] = '\0';
 271
 272       if (debug && rc == PUNYCODE_SUCCESS)
 273         {
 274           printf ("computed out: %s\n", p);
 275           printf ("expected out: %s\n", punycode[i].out);
 276         }
 277       else if (debug)
 278         printf ("returned %d expected %d\n", rc, punycode[i].rc);
 279
 280       if (rc == PUNYCODE_SUCCESS)
 281         {
 282           if (strlen (punycode[i].out) != strlen (p) ||
 283               memcmp (punycode[i].out, p, strlen (p)) != 0)
 284             {
 285               fail ("punycode() entry %d failed\n", i);
 286               if (debug)
 287                 printf ("ERROR\n");
 288             }
 289           else if (debug)
 290             printf ("OK\n\n");
 291         }
 292       else if (debug)
 293         printf ("OK\n\n");
 294
 295       if (debug)
 296         {
 297           printf ("in: %s\n", punycode[i].out);
 298         }
 299
 300       outlen = BUFSIZ;
 301       rc = punycode_decode (strlen(punycode[i].out), punycode[i].out,
 302                             &outlen, q, NULL);
 303       if (rc != punycode[i].rc)
 304         {
 305           fail ("punycode() entry %d failed: %d\n", i, rc);
 306           if (debug)
 307             printf ("FATAL\n");
 308           continue;
 309         }
 310
 311       if (debug && rc == PUNYCODE_SUCCESS)
 312         {
 313           printf ("computed out:\n");
 314           ucs4print (q, outlen);
 315           printf ("expected out:\n");
 316           ucs4print (punycode[i].in, punycode[i].inlen);
 317         }
 318       else if (debug)
 319         printf ("returned %d expected %d\n", rc, punycode[i].rc);
 320
 321       if (rc == PUNYCODE_SUCCESS)
 322         {
 323           if (punycode[i].inlen != outlen ||
 324               memcmp (punycode[i].in, q, outlen) != 0)
 325             {
 326               fail ("punycode_decode() entry %d failed\n", i);
 327               if (debug)
 328                 printf ("ERROR\n");
 329             }
 330           else if (debug)
 331             printf ("OK\n\n");
 332         }
 333       else if (debug)
 334         printf ("OK\n\n");
 335     }
 336
 337   free (q);
 338   free (p);
 339
 340   if (debug)
 341     printf ("Punycode self tests done with %d errors\n", error_count);
 342
 343   return error_count ? 1 : 0;
 344 }