glib/tests/hostutils.c

   1 /*
   2  * Copyright (C) 2008 Red Hat, Inc
   3  *
   4  * This library is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU Lesser General Public
   6  * License as published by the Free Software Foundation; either
   7  * version 2.1 of the License, or (at your option) any later version.
   8  *
   9  * This library 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 GNU
  12  * Lesser General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU Lesser General
  15  * Public License along with this library; if not, see <http://www.gnu.org/licenses/>.
  16  */
  17
  18 #include <glib/glib.h>
  19
  20 #include <stdio.h>
  21 #include <stdlib.h>
  22 #include <string.h>
  23
  24 static const struct {
  25   const gchar *ascii_name, *unicode_name;
  26 } idn_test_domains[] = {
  27   /* "example.test" in various languages */
  28   { "xn--mgbh0fb.xn--kgbechtv", "\xd9\x85\xd8\xab\xd8\xa7\xd9\x84.\xd8\xa5\xd8\xae\xd8\xaa\xd8\xa8\xd8\xa7\xd8\xb1" },
  29   { "xn--fsqu00a.xn--0zwm56d", "\xe4\xbe\x8b\xe5\xad\x90.\xe6\xb5\x8b\xe8\xaf\x95" },
  30   { "xn--fsqu00a.xn--g6w251d", "\xe4\xbe\x8b\xe5\xad\x90.\xe6\xb8\xac\xe8\xa9\xa6" },
  31   { "xn--hxajbheg2az3al.xn--jxalpdlp", "\xcf\x80\xce\xb1\xcf\x81\xce\xac\xce\xb4\xce\xb5\xce\xb9\xce\xb3\xce\xbc\xce\xb1.\xce\xb4\xce\xbf\xce\xba\xce\xb9\xce\xbc\xce\xae" },
  32   { "xn--p1b6ci4b4b3a.xn--11b5bs3a9aj6g", "\xe0\xa4\x89\xe0\xa4\xa6\xe0\xa4\xbe\xe0\xa4\xb9\xe0\xa4\xb0\xe0\xa4\xa3.\xe0\xa4\xaa\xe0\xa4\xb0\xe0\xa5\x80\xe0\xa4\x95\xe0\xa5\x8d\xe0\xa4\xb7\xe0\xa4\xbe" },
  33   { "xn--r8jz45g.xn--zckzah", "\xe4\xbe\x8b\xe3\x81\x88.\xe3\x83\x86\xe3\x82\xb9\xe3\x83\x88" },
  34   { "xn--9n2bp8q.xn--9t4b11yi5a", "\xec\x8b\xa4\xeb\xa1\x80.\xed\x85\x8c\xec\x8a\xa4\xed\x8a\xb8" },
  35   { "xn--mgbh0fb.xn--hgbk6aj7f53bba", "\xd9\x85\xd8\xab\xd8\xa7\xd9\x84.\xd8\xa2\xd8\xb2\xd9\x85\xd8\xa7\xdb\x8c\xd8\xb4\xdb\x8c" },
  36   { "xn--e1afmkfd.xn--80akhbyknj4f", "\xd0\xbf\xd1\x80\xd0\xb8\xd0\xbc\xd0\xb5\xd1\x80.\xd0\xb8\xd1\x81\xd0\xbf\xd1\x8b\xd1\x82\xd0\xb0\xd0\xbd\xd0\xb8\xd0\xb5" },
  37   { "xn--zkc6cc5bi7f6e.xn--hlcj6aya9esc7a", "\xe0\xae\x89\xe0\xae\xa4\xe0\xae\xbe\xe0\xae\xb0\xe0\xae\xa3\xe0\xae\xae\xe0\xaf\x8d.\xe0\xae\xaa\xe0\xae\xb0\xe0\xae\xbf\xe0\xae\x9f\xe0\xaf\x8d\xe0\xae\x9a\xe0\xaf\x88" },
  38   { "xn--fdbk5d8ap9b8a8d.xn--deba0ad", "\xd7\x91\xd7\xb2\xd6\xb7\xd7\xa9\xd7\xa4\xd6\xbc\xd7\x99\xd7\x9c.\xd7\x98\xd7\xa2\xd7\xa1\xd7\x98" },
  39
  40   /* further examples without their own IDN-ized TLD */
  41   { "xn--1xd0bwwra.idn.icann.org", "\xe1\x8a\xa0\xe1\x88\x9b\xe1\x88\xad\xe1\x8a\x9b.idn.icann.org" },
  42   { "xn--54b7fta0cc.idn.icann.org", "\xe0\xa6\xac\xe0\xa6\xbe\xe0\xa6\x82\xe0\xa6\xb2\xe0\xa6\xbe.idn.icann.org" },
  43   { "xn--5dbqzzl.idn.icann.org", "\xd7\xa2\xd7\x91\xd7\xa8\xd7\x99\xd7\xaa.idn.icann.org" },
  44   { "xn--j2e7beiw1lb2hqg.idn.icann.org", "\xe1\x9e\x97\xe1\x9e\xb6\xe1\x9e\x9f\xe1\x9e\xb6\xe1\x9e\x81\xe1\x9f\x92\xe1\x9e\x98\xe1\x9f\x82\xe1\x9e\x9a.idn.icann.org" },
  45   { "xn--o3cw4h.idn.icann.org", "\xe0\xb9\x84\xe0\xb8\x97\xe0\xb8\xa2.idn.icann.org" },
  46   { "xn--mgbqf7g.idn.icann.org", "\xd8\xa7\xd8\xb1\xd8\xaf\xd9\x88.idn.icann.org" }
  47 };
  48 static const gint num_idn_test_domains = G_N_ELEMENTS (idn_test_domains);
  49
  50 static const struct {
  51   const gchar *orig_name, *ascii_name;
  52   gboolean orig_is_unicode, ascii_is_encoded;
  53 } non_round_trip_names[] = {
  54   /* uppercase characters */
  55   { "EXAMPLE.COM", "example.com", FALSE, FALSE },
  56   { "\xc3\x89XAMPLE.COM", "xn--xample-9ua.com", TRUE, TRUE },
  57
  58   /* unicode that decodes to ascii */
  59   { "\xe2\x93\x94\xe2\x93\xa7\xe2\x93\x90\xe2\x93\x9c\xe2\x93\x9f\xe2\x93\x9b\xe2\x93\x94.com", "example.com", TRUE, FALSE },
  60
  61   /* non-standard dot characters */
  62   { "example\xe3\x80\x82" "com", "example.com", TRUE, FALSE },
  63   { "\xc3\xa9xample\xe3\x80\x82" "com", "xn--xample-9ua.com", TRUE, TRUE },
  64   { "Å.idn.icann.org", "xn--5ca.idn.icann.org", TRUE, TRUE },
  65   { "ℵℶℷ\xcd\x8f.idn.icann.org", "xn--4dbcd.idn.icann.org", TRUE, TRUE }
  66 };
  67 static const gint num_non_round_trip_names = G_N_ELEMENTS (non_round_trip_names);
  68
  69 static const gchar *bad_names[] = {
  70   "disallowed\xef\xbf\xbd" "character",
  71   "non-utf\x88",
  72   "xn--mixed-\xc3\xbcp"
  73 };
  74 static const gint num_bad_names = G_N_ELEMENTS (bad_names);
  75
  76 static void
  77 test_to_ascii (void)
  78 {
  79   gint i;
  80   gchar *ascii;
  81
  82   for (i = 0; i < num_idn_test_domains; i++)
  83     {
  84       g_assert (g_hostname_is_non_ascii (idn_test_domains[i].unicode_name));
  85       ascii = g_hostname_to_ascii (idn_test_domains[i].unicode_name);
  86       g_assert_cmpstr (idn_test_domains[i].ascii_name, ==, ascii);
  87       g_free (ascii);
  88
  89       ascii = g_hostname_to_ascii (idn_test_domains[i].ascii_name);
  90       g_assert_cmpstr (idn_test_domains[i].ascii_name, ==, ascii);
  91       g_free (ascii);
  92     }
  93
  94   for (i = 0; i < num_non_round_trip_names; i++)
  95     {
  96       if (non_round_trip_names[i].orig_is_unicode)
  97         g_assert (g_hostname_is_non_ascii (non_round_trip_names[i].orig_name));
  98       else
  99         g_assert (!g_hostname_is_non_ascii (non_round_trip_names[i].orig_name));
 100
 101       if (non_round_trip_names[i].ascii_is_encoded)
 102         g_assert (g_hostname_is_ascii_encoded (non_round_trip_names[i].ascii_name));
 103       else
 104         g_assert (!g_hostname_is_ascii_encoded (non_round_trip_names[i].ascii_name));
 105
 106       ascii = g_hostname_to_ascii (non_round_trip_names[i].orig_name);
 107       g_assert_cmpstr (non_round_trip_names[i].ascii_name, ==, ascii);
 108       g_free (ascii);
 109
 110       ascii = g_hostname_to_ascii (non_round_trip_names[i].ascii_name);
 111       g_assert_cmpstr (non_round_trip_names[i].ascii_name, ==, ascii);
 112       g_free (ascii);
 113     }
 114
 115   for (i = 0; i < num_bad_names; i++)
 116     {
 117       ascii = g_hostname_to_ascii (bad_names[i]);
 118       g_assert_cmpstr (ascii, ==, NULL);
 119     }
 120 }
 121
 122 static void
 123 test_to_unicode (void)
 124 {
 125   gint i;
 126   gchar *unicode;
 127
 128   for (i = 0; i < num_idn_test_domains; i++)
 129     {
 130       g_assert (g_hostname_is_ascii_encoded (idn_test_domains[i].ascii_name));
 131       unicode = g_hostname_to_unicode (idn_test_domains[i].ascii_name);
 132       g_assert_cmpstr (idn_test_domains[i].unicode_name, ==, unicode);
 133       g_free (unicode);
 134
 135       unicode = g_hostname_to_unicode (idn_test_domains[i].unicode_name);
 136       g_assert_cmpstr (idn_test_domains[i].unicode_name, ==, unicode);
 137       g_free (unicode);
 138     }
 139
 140   for (i = 0; i < num_bad_names; i++)
 141     {
 142       unicode = g_hostname_to_unicode (bad_names[i]);
 143       g_assert_cmpstr (unicode, ==, NULL);
 144     }
 145 }
 146
 147 static const struct {
 148   const gchar *addr;
 149   gboolean is_addr;
 150 } ip_addr_tests[] = {
 151   /* IPv6 tests */
 152
 153   { "0123:4567:89AB:cdef:3210:7654:ba98:FeDc", TRUE },
 154
 155   { "0123:4567:89AB:cdef:3210:7654:ba98::", TRUE },
 156   { "0123:4567:89AB:cdef:3210:7654::", TRUE },
 157   { "0123:4567:89AB:cdef:3210::", TRUE },
 158   { "0123:4567:89AB:cdef::", TRUE },
 159   { "0123:4567:89AB::", TRUE },
 160   { "0123:4567::", TRUE },
 161   { "0123::", TRUE },
 162
 163   { "::4567:89AB:cdef:3210:7654:ba98:FeDc", TRUE },
 164   { "::89AB:cdef:3210:7654:ba98:FeDc", TRUE },
 165   { "::cdef:3210:7654:ba98:FeDc", TRUE },
 166   { "::3210:7654:ba98:FeDc", TRUE },
 167   { "::7654:ba98:FeDc", TRUE },
 168   { "::ba98:FeDc", TRUE },
 169   { "::FeDc", TRUE },
 170
 171   { "0123::89AB:cdef:3210:7654:ba98:FeDc", TRUE },
 172   { "0123:4567::cdef:3210:7654:ba98:FeDc", TRUE },
 173   { "0123:4567:89AB::3210:7654:ba98:FeDc", TRUE },
 174   { "0123:4567:89AB:cdef::7654:ba98:FeDc", TRUE },
 175   { "0123:4567:89AB:cdef:3210::ba98:FeDc", TRUE },
 176   { "0123:4567:89AB:cdef:3210:7654::FeDc", TRUE },
 177
 178   { "0123::cdef:3210:7654:ba98:FeDc", TRUE },
 179   { "0123:4567::3210:7654:ba98:FeDc", TRUE },
 180   { "0123:4567:89AB::7654:ba98:FeDc", TRUE },
 181   { "0123:4567:89AB:cdef::ba98:FeDc", TRUE },
 182   { "0123:4567:89AB:cdef:3210::FeDc", TRUE },
 183
 184   { "0123::3210:7654:ba98:FeDc", TRUE },
 185   { "0123:4567::7654:ba98:FeDc", TRUE },
 186   { "0123:4567:89AB::ba98:FeDc", TRUE },
 187   { "0123:4567:89AB:cdef::FeDc", TRUE },
 188
 189   { "0123::7654:ba98:FeDc", TRUE },
 190   { "0123:4567::ba98:FeDc", TRUE },
 191   { "0123:4567:89AB::FeDc", TRUE },
 192
 193   { "0123::ba98:FeDc", TRUE },
 194   { "0123:4567::FeDc", TRUE },
 195
 196   { "0123::FeDc", TRUE },
 197
 198   { "::", TRUE },
 199
 200   { "0:12:345:6789:a:bc:def::", TRUE },
 201
 202   { "0123:4567:89AB:cdef:3210:7654:123.45.67.89", TRUE },
 203   { "0123:4567:89AB:cdef:3210::123.45.67.89", TRUE },
 204   { "0123:4567:89AB:cdef::123.45.67.89", TRUE },
 205   { "0123:4567:89AB::123.45.67.89", TRUE },
 206   { "0123:4567::123.45.67.89", TRUE },
 207   { "0123::123.45.67.89", TRUE },
 208   { "::123.45.67.89", TRUE },
 209
 210   /* Contain non-hex chars */
 211   { "012x:4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 212   { "0123:45x7:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 213   { "0123:4567:8xAB:cdef:3210:7654:ba98:FeDc", FALSE },
 214   { "0123:4567:89AB:xdef:3210:7654:ba98:FeDc", FALSE },
 215   { "0123:4567:89AB:cdef:321;:7654:ba98:FeDc", FALSE },
 216   { "0123:4567:89AB:cdef:3210:76*4:ba98:FeDc", FALSE },
 217   { "0123:4567:89AB:cdef:3210:7654:b-98:FeDc", FALSE },
 218   { "0123:4567:89AB:cdef:3210:7654:ba98:+eDc", FALSE },
 219   { "0123:4567:89AB:cdef:3210:7654:ba98:FeDc and some trailing junk", FALSE },
 220   { " 123:4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 221   { "012 :4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 222   { "0123: 567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 223   { "0123:4567:89AB:cdef:3210:7654:ba98:FeD ", FALSE },
 224
 225   /* Contains too-long/out-of-range segments */
 226   { "00123:4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 227   { "0123:04567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 228   { "0123:4567:189AB:cdef:3210:7654:ba98:FeDc", FALSE },
 229
 230   /* Too short */
 231   { "0123:4567:89AB:cdef:3210:7654:ba98", FALSE },
 232   { "0123:4567:89AB:cdef:3210:7654", FALSE },
 233   { "0123:4567:89AB:cdef:3210", FALSE },
 234   { "0123", FALSE },
 235   { "", FALSE },
 236
 237   /* Too long */
 238   { "0123:4567:89AB:cdef:3210:7654:ba98:FeDc:9999", FALSE },
 239   { "0123::4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 240   { "0123:4567::89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 241   { "0123:4567:89AB::cdef:3210:7654:ba98:FeDc", FALSE },
 242   { "0123:4567:89AB:cdef::3210:7654:ba98:FeDc", FALSE },
 243   { "0123:4567:89AB:cdef:3210::7654:ba98:FeDc", FALSE },
 244   { "0123:4567:89AB:cdef:3210:7654::ba98:FeDc", FALSE },
 245   { "0123:4567:89AB:cdef:3210:7654:ba98::FeDc", FALSE },
 246
 247   /* Invalid use of ":"s */
 248   { "0123::89AB::3210:7654:ba98:FeDc", FALSE },
 249   { "::4567:89AB:cdef:3210:7654::FeDc", FALSE },
 250   { "0123::89AB:cdef:3210:7654:ba98::", FALSE },
 251   { ":4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 252   { "0123:4567:89AB:cdef:3210:7654:ba98:", FALSE },
 253   { "0123:::cdef:3210:7654:ba98:FeDc", FALSE },
 254   { "0123:4567:89AB:cdef:3210:7654:ba98:FeDc:", FALSE },
 255   { ":0123:4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
 256   { ":::", FALSE },
 257
 258   /* IPv4 address at wrong place */
 259   { "0123:4567:89AB:cdef:3210:123.45.67.89", FALSE },
 260   { "0123:4567:89AB:cdef:3210:7654::123.45.67.89", FALSE },
 261   { "0123:4567:89AB:cdef:123.45.67.89", FALSE },
 262   { "0123:4567:89AB:cdef:3210:123.45.67.89:FeDc", FALSE },
 263
 264
 265   /* IPv4 tests */
 266
 267   { "123.45.67.89", TRUE },
 268   { "1.2.3.4", TRUE },
 269   { "1.2.3.0", TRUE },
 270
 271   { "023.045.067.089", FALSE },
 272   { "1234.5.67.89", FALSE },
 273   { "123.45.67.00", FALSE },
 274   { " 1.2.3.4", FALSE },
 275   { "1 .2.3.4", FALSE },
 276   { "1. 2.3.4", FALSE },
 277   { "1.2.3.4 ", FALSE },
 278   { "1.2.3", FALSE },
 279   { "1.2.3.4.5", FALSE },
 280   { "1.b.3.4", FALSE },
 281   { "1.2.3:4", FALSE },
 282   { "1.2.3.4, etc", FALSE },
 283   { "1,2,3,4", FALSE },
 284   { "1.2.3.com", FALSE },
 285   { "1.2.3.4.", FALSE },
 286   { "1.2.3.", FALSE },
 287   { ".1.2.3.4", FALSE },
 288   { ".2.3.4", FALSE },
 289   { "1..2.3.4", FALSE },
 290   { "1..3.4", FALSE }
 291 };
 292 static const gint num_ip_addr_tests = G_N_ELEMENTS (ip_addr_tests);
 293
 294 static void
 295 test_is_ip_addr (void)
 296 {
 297   gint i;
 298
 299   for (i = 0; i < num_ip_addr_tests; i++)
 300     {
 301       if (g_hostname_is_ip_address (ip_addr_tests[i].addr) != ip_addr_tests[i].is_addr)
 302         {
 303           char *msg = g_strdup_printf ("g_hostname_is_ip_address (\"%s\") == %s",
 304                                        ip_addr_tests[i].addr,
 305                                        ip_addr_tests[i].is_addr ? "TRUE" : "FALSE");
 306           g_assertion_message (G_LOG_DOMAIN, __FILE__, __LINE__, G_STRFUNC, msg);
 307         }
 308     }
 309 }
 310
 311 /* FIXME: test names with both unicode and ACE-encoded labels */
 312 /* FIXME: test invalid unicode names */
 313
 314 int
 315 main (int   argc,
 316       char *argv[])
 317 {
 318   g_test_init (&argc, &argv, NULL);
 319
 320   if (argc == 2 && argv[1][0] != '-')
 321     {
 322       const gchar *hostname = argv[1];
 323       gchar *converted;
 324
 325       if (g_hostname_is_non_ascii (hostname))
 326         {
 327           converted = g_hostname_to_ascii (hostname);
 328           printf ("to_ascii: %s\n", converted);
 329           g_free (converted);
 330         }
 331       else if (g_hostname_is_ascii_encoded (hostname))
 332         {
 333           converted = g_hostname_to_unicode (hostname);
 334           printf ("to_unicode: %s\n", converted);
 335           g_free (converted);
 336         }
 337       else
 338         printf ("hostname is neither unicode nor ACE encoded\n");
 339       return 0;
 340     }
 341
 342   g_test_add_func ("/hostutils/to_ascii", test_to_ascii);
 343   g_test_add_func ("/hostutils/to_unicode", test_to_unicode);
 344   g_test_add_func ("/hostutils/is_ip_addr", test_is_ip_addr);
 345
 346   return g_test_run ();
 347 }