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 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
74 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
75 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
76 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
77 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
78 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
79 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
80 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
81 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
82 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
83 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
84 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
85 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
86 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
87 "verylongverylongverylongverylongverylongverylongverylongverylongverylong"
88 "verylongverylongverylongverylongverylongverylongverylongverylongverylong",
89 };
90 static const gint num_bad_names = G_N_ELEMENTS (bad_names);
91
92 static void
test_to_ascii(void)93 test_to_ascii (void)
94 {
95 gint i;
96 gchar *ascii;
97
98 for (i = 0; i < num_idn_test_domains; i++)
99 {
100 g_assert_true (g_hostname_is_non_ascii (idn_test_domains[i].unicode_name));
101 ascii = g_hostname_to_ascii (idn_test_domains[i].unicode_name);
102 g_assert_cmpstr (idn_test_domains[i].ascii_name, ==, ascii);
103 g_free (ascii);
104
105 ascii = g_hostname_to_ascii (idn_test_domains[i].ascii_name);
106 g_assert_cmpstr (idn_test_domains[i].ascii_name, ==, ascii);
107 g_free (ascii);
108 }
109
110 for (i = 0; i < num_non_round_trip_names; i++)
111 {
112 if (non_round_trip_names[i].orig_is_unicode)
113 g_assert_true (g_hostname_is_non_ascii (non_round_trip_names[i].orig_name));
114 else
115 g_assert_true (!g_hostname_is_non_ascii (non_round_trip_names[i].orig_name));
116
117 if (non_round_trip_names[i].ascii_is_encoded)
118 g_assert_true (g_hostname_is_ascii_encoded (non_round_trip_names[i].ascii_name));
119 else
120 g_assert_true (!g_hostname_is_ascii_encoded (non_round_trip_names[i].ascii_name));
121
122 ascii = g_hostname_to_ascii (non_round_trip_names[i].orig_name);
123 g_assert_cmpstr (non_round_trip_names[i].ascii_name, ==, ascii);
124 g_free (ascii);
125
126 ascii = g_hostname_to_ascii (non_round_trip_names[i].ascii_name);
127 g_assert_cmpstr (non_round_trip_names[i].ascii_name, ==, ascii);
128 g_free (ascii);
129 }
130
131 for (i = 0; i < num_bad_names; i++)
132 {
133 ascii = g_hostname_to_ascii (bad_names[i]);
134 g_assert_cmpstr (ascii, ==, NULL);
135 }
136 }
137
138 static void
test_to_unicode(void)139 test_to_unicode (void)
140 {
141 gint i;
142 gchar *unicode;
143
144 for (i = 0; i < num_idn_test_domains; i++)
145 {
146 g_assert_true (g_hostname_is_ascii_encoded (idn_test_domains[i].ascii_name));
147 unicode = g_hostname_to_unicode (idn_test_domains[i].ascii_name);
148 g_assert_cmpstr (idn_test_domains[i].unicode_name, ==, unicode);
149 g_free (unicode);
150
151 unicode = g_hostname_to_unicode (idn_test_domains[i].unicode_name);
152 g_assert_cmpstr (idn_test_domains[i].unicode_name, ==, unicode);
153 g_free (unicode);
154 }
155
156 for (i = 0; i < num_bad_names; i++)
157 {
158 unicode = g_hostname_to_unicode (bad_names[i]);
159 g_assert_cmpstr (unicode, ==, NULL);
160 }
161 }
162
163 static const struct {
164 const gchar *addr;
165 gboolean is_addr;
166 } ip_addr_tests[] = {
167 /* IPv6 tests */
168
169 { "0123:4567:89AB:cdef:3210:7654:ba98:FeDc", TRUE },
170
171 { "0123:4567:89AB:cdef:3210:7654:ba98::", TRUE },
172 { "0123:4567:89AB:cdef:3210:7654::", TRUE },
173 { "0123:4567:89AB:cdef:3210::", TRUE },
174 { "0123:4567:89AB:cdef::", TRUE },
175 { "0123:4567:89AB::", TRUE },
176 { "0123:4567::", TRUE },
177 { "0123::", TRUE },
178
179 { "::4567:89AB:cdef:3210:7654:ba98:FeDc", TRUE },
180 { "::89AB:cdef:3210:7654:ba98:FeDc", TRUE },
181 { "::cdef:3210:7654:ba98:FeDc", TRUE },
182 { "::3210:7654:ba98:FeDc", TRUE },
183 { "::7654:ba98:FeDc", TRUE },
184 { "::ba98:FeDc", TRUE },
185 { "::FeDc", TRUE },
186
187 { "0123::89AB:cdef:3210:7654:ba98:FeDc", TRUE },
188 { "0123:4567::cdef:3210:7654:ba98:FeDc", TRUE },
189 { "0123:4567:89AB::3210:7654:ba98:FeDc", TRUE },
190 { "0123:4567:89AB:cdef::7654:ba98:FeDc", TRUE },
191 { "0123:4567:89AB:cdef:3210::ba98:FeDc", TRUE },
192 { "0123:4567:89AB:cdef:3210:7654::FeDc", TRUE },
193
194 { "0123::cdef:3210:7654:ba98:FeDc", TRUE },
195 { "0123:4567::3210:7654:ba98:FeDc", TRUE },
196 { "0123:4567:89AB::7654:ba98:FeDc", TRUE },
197 { "0123:4567:89AB:cdef::ba98:FeDc", TRUE },
198 { "0123:4567:89AB:cdef:3210::FeDc", TRUE },
199
200 { "0123::3210:7654:ba98:FeDc", TRUE },
201 { "0123:4567::7654:ba98:FeDc", TRUE },
202 { "0123:4567:89AB::ba98:FeDc", TRUE },
203 { "0123:4567:89AB:cdef::FeDc", TRUE },
204
205 { "0123::7654:ba98:FeDc", TRUE },
206 { "0123:4567::ba98:FeDc", TRUE },
207 { "0123:4567:89AB::FeDc", TRUE },
208
209 { "0123::ba98:FeDc", TRUE },
210 { "0123:4567::FeDc", TRUE },
211
212 { "0123::FeDc", TRUE },
213
214 { "::", TRUE },
215
216 { "0:12:345:6789:a:bc:def::", TRUE },
217
218 { "0123:4567:89AB:cdef:3210:7654:123.45.67.89", TRUE },
219 { "0123:4567:89AB:cdef:3210::123.45.67.89", TRUE },
220 { "0123:4567:89AB:cdef::123.45.67.89", TRUE },
221 { "0123:4567:89AB::123.45.67.89", TRUE },
222 { "0123:4567::123.45.67.89", TRUE },
223 { "0123::123.45.67.89", TRUE },
224 { "::123.45.67.89", TRUE },
225
226 /* accept zone-id from rfc6874 */
227 { "0123:4567:89AB:cdef:3210:7654:ba98:FeDc%zoneid0", TRUE },
228 { "fe80::dead:beef%zonÉid0%weird", TRUE },
229 { "fe80::dead:beef%", FALSE },
230
231 /* Contain non-hex chars */
232 { "012x:4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
233 { "0123:45x7:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
234 { "0123:4567:8xAB:cdef:3210:7654:ba98:FeDc", FALSE },
235 { "0123:4567:89AB:xdef:3210:7654:ba98:FeDc", FALSE },
236 { "0123:4567:89AB:cdef:321;:7654:ba98:FeDc", FALSE },
237 { "0123:4567:89AB:cdef:3210:76*4:ba98:FeDc", FALSE },
238 { "0123:4567:89AB:cdef:3210:7654:b-98:FeDc", FALSE },
239 { "0123:4567:89AB:cdef:3210:7654:ba98:+eDc", FALSE },
240 { "0123:4567:89AB:cdef:3210:7654:ba98:FeDc and some trailing junk", FALSE },
241 { " 123:4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
242 { "012 :4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
243 { "0123: 567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
244 { "0123:4567:89AB:cdef:3210:7654:ba98:FeD ", FALSE },
245
246 /* Contains too-long/out-of-range segments */
247 { "00123:4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
248 { "0123:04567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
249 { "0123:4567:189AB:cdef:3210:7654:ba98:FeDc", FALSE },
250
251 /* Too short */
252 { "0123:4567:89AB:cdef:3210:7654:ba98", FALSE },
253 { "0123:4567:89AB:cdef:3210:7654", FALSE },
254 { "0123:4567:89AB:cdef:3210", FALSE },
255 { "0123", FALSE },
256 { "", FALSE },
257
258 /* Too long */
259 { "0123:4567:89AB:cdef:3210:7654:ba98:FeDc:9999", FALSE },
260 { "0123::4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
261 { "0123:4567::89AB:cdef:3210:7654:ba98:FeDc", FALSE },
262 { "0123:4567:89AB::cdef:3210:7654:ba98:FeDc", FALSE },
263 { "0123:4567:89AB:cdef::3210:7654:ba98:FeDc", FALSE },
264 { "0123:4567:89AB:cdef:3210::7654:ba98:FeDc", FALSE },
265 { "0123:4567:89AB:cdef:3210:7654::ba98:FeDc", FALSE },
266 { "0123:4567:89AB:cdef:3210:7654:ba98::FeDc", FALSE },
267
268 /* Invalid use of ":"s */
269 { "0123::89AB::3210:7654:ba98:FeDc", FALSE },
270 { "::4567:89AB:cdef:3210:7654::FeDc", FALSE },
271 { "0123::89AB:cdef:3210:7654:ba98::", FALSE },
272 { ":4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
273 { "0123:4567:89AB:cdef:3210:7654:ba98:", FALSE },
274 { "0123:::cdef:3210:7654:ba98:FeDc", FALSE },
275 { "0123:4567:89AB:cdef:3210:7654:ba98:FeDc:", FALSE },
276 { ":0123:4567:89AB:cdef:3210:7654:ba98:FeDc", FALSE },
277 { ":::", FALSE },
278
279 /* IPv4 address at wrong place */
280 { "0123:4567:89AB:cdef:3210:123.45.67.89", FALSE },
281 { "0123:4567:89AB:cdef:3210:7654::123.45.67.89", FALSE },
282 { "0123:4567:89AB:cdef:123.45.67.89", FALSE },
283 { "0123:4567:89AB:cdef:3210:123.45.67.89:FeDc", FALSE },
284
285 /* IPv4 tests */
286
287 { "123.45.67.89", TRUE },
288 { "1.2.3.4", TRUE },
289 { "1.2.3.0", TRUE },
290
291 { "023.045.067.089", FALSE },
292 { "1234.5.67.89", FALSE },
293 { "123.45.67.00", FALSE },
294 { " 1.2.3.4", FALSE },
295 { "1 .2.3.4", FALSE },
296 { "1. 2.3.4", FALSE },
297 { "1.2.3.4 ", FALSE },
298 { "1.2.3", FALSE },
299 { "1.2.3.4.5", FALSE },
300 { "1.b.3.4", FALSE },
301 { "1.2.3:4", FALSE },
302 { "1.2.3.4, etc", FALSE },
303 { "1,2,3,4", FALSE },
304 { "1.2.3.com", FALSE },
305 { "1.2.3.4.", FALSE },
306 { "1.2.3.", FALSE },
307 { ".1.2.3.4", FALSE },
308 { ".2.3.4", FALSE },
309 { "1..2.3.4", FALSE },
310 { "1..3.4", FALSE }
311 };
312 static const gint num_ip_addr_tests = G_N_ELEMENTS (ip_addr_tests);
313
314 static void
test_is_ip_addr(void)315 test_is_ip_addr (void)
316 {
317 gint i;
318
319 for (i = 0; i < num_ip_addr_tests; i++)
320 {
321 if (g_hostname_is_ip_address (ip_addr_tests[i].addr) != ip_addr_tests[i].is_addr)
322 {
323 char *msg = g_strdup_printf ("g_hostname_is_ip_address (\"%s\") == %s",
324 ip_addr_tests[i].addr,
325 ip_addr_tests[i].is_addr ? "TRUE" : "FALSE");
326 g_assertion_message (G_LOG_DOMAIN, __FILE__, __LINE__, G_STRFUNC, msg);
327 }
328 }
329 }
330
331 /* FIXME: test names with both unicode and ACE-encoded labels */
332 /* FIXME: test invalid unicode names */
333
334 int
main(int argc,char * argv[])335 main (int argc,
336 char *argv[])
337 {
338 g_test_init (&argc, &argv, NULL);
339
340 if (argc == 2 && argv[1][0] != '-')
341 {
342 const gchar *hostname = argv[1];
343 gchar *converted;
344
345 if (g_hostname_is_non_ascii (hostname))
346 {
347 converted = g_hostname_to_ascii (hostname);
348 printf ("to_ascii: %s\n", converted);
349 g_free (converted);
350 }
351 else if (g_hostname_is_ascii_encoded (hostname))
352 {
353 converted = g_hostname_to_unicode (hostname);
354 printf ("to_unicode: %s\n", converted);
355 g_free (converted);
356 }
357 else
358 printf ("hostname is neither unicode nor ACE encoded\n");
359 return 0;
360 }
361
362 g_test_add_func ("/hostutils/to_ascii", test_to_ascii);
363 g_test_add_func ("/hostutils/to_unicode", test_to_unicode);
364 g_test_add_func ("/hostutils/is_ip_addr", test_is_ip_addr);
365
366 return g_test_run ();
367 }
368