1 /**
2 * @file
3 * Handling of international domain names
4 *
5 * @authors
6 * Copyright (C) 2003,2005,2008-2009 Thomas Roessler <roessler@does-not-exist.org>
7 *
8 * @copyright
9 * This program is free software: you can redistribute it and/or modify it under
10 * the terms of the GNU General Public License as published by the Free Software
11 * Foundation, either version 2 of the License, or (at your option) any later
12 * version.
13 *
14 * This program is distributed in the hope that it will be useful, but WITHOUT
15 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
16 * FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
17 * details.
18 *
19 * You should have received a copy of the GNU General Public License along with
20 * this program. If not, see <http://www.gnu.org/licenses/>.
21 */
22
23 /**
24 * @page addr_idna International domain names
25 *
26 * Handling of international domain names
27 */
28
29 #include "config.h"
30 #include <stdint.h>
31 #include <stdio.h>
32 #include "mutt/lib.h"
33 #include "config/lib.h"
34 #include "core/lib.h"
35 #include "idna2.h"
36 #ifdef HAVE_LIBIDN
37 #include <stdbool.h>
38 #include <string.h>
39 #endif
40 #ifdef HAVE_STRINGPREP_H
41 #include <stringprep.h>
42 #elif defined(HAVE_IDN_STRINGPREP_H)
43 #include <idn/stringprep.h>
44 #endif
45 #define IDN2_SKIP_LIBIDN_COMPAT
46 #ifdef HAVE_IDN2_H
47 #include <idn2.h>
48 #elif defined(HAVE_IDN_IDN2_H)
49 #include <idn/idn2.h>
50 #elif defined(HAVE_IDNA_H)
51 #include <idna.h>
52 #elif defined(HAVE_IDN_IDNA_H)
53 #include <idn/idna.h>
54 #endif
55
56 #if defined(HAVE_IDN2_H) || defined(HAVE_IDN_IDN2_H)
57 #define IDN_VERSION 2
58 #elif defined(HAVE_IDNA_H) || defined(HAVE_IDN_IDNA_H)
59 #define IDN_VERSION 1
60 #endif
61
62 #ifdef HAVE_LIBIDN
63 /* Work around incompatibilities in the libidn API */
64 #if (!defined(HAVE_IDNA_TO_ASCII_8Z) && defined(HAVE_IDNA_TO_ASCII_FROM_UTF8))
65 #define idna_to_ascii_8z(input, output, flags) \
66 idna_to_ascii_from_utf8(input, output, (flags) &1, ((flags) &2) ? 1 : 0)
67 #endif
68 #if (!defined(HAVE_IDNA_TO_ASCII_LZ) && defined(HAVE_IDNA_TO_ASCII_FROM_LOCALE))
69 #define idna_to_ascii_lz(input, output, flags) \
70 idna_to_ascii_from_locale(input, output, (flags) &1, ((flags) &2) ? 1 : 0)
71 #endif
72 #if (!defined(HAVE_IDNA_TO_UNICODE_8Z8Z) && defined(HAVE_IDNA_TO_UNICODE_UTF8_FROM_UTF8))
73 #define idna_to_unicode_8z8z(input, output, flags) \
74 idna_to_unicode_utf8_from_utf8(input, output, (flags) &1, ((flags) &2) ? 1 : 0)
75 #endif
76 #endif /* HAVE_LIBIDN */
77
78 #ifdef HAVE_LIBIDN
79 /**
80 * check_idn - Is domain in Punycode?
81 * @param domain Domain to test
82 * @retval true At least one part of domain is in Punycode
83 */
check_idn(char * domain)84 static bool check_idn(char *domain)
85 {
86 if (!domain)
87 return false;
88
89 if (mutt_istr_startswith(domain, "xn--"))
90 return true;
91
92 while ((domain = strchr(domain, '.')))
93 {
94 if (mutt_istr_startswith(++domain, "xn--"))
95 return true;
96 }
97
98 return false;
99 }
100
101 /**
102 * mutt_idna_to_ascii_lz - Convert a domain to Punycode
103 * @param[in] input Domain
104 * @param[out] output Result
105 * @param[in] flags Flags, e.g. IDNA_ALLOW_UNASSIGNED
106 * @retval 0 Success
107 * @retval >0 Failure, error code
108 *
109 * Convert a domain from the current locale to Punycode.
110 *
111 * @note The caller must free output
112 */
mutt_idna_to_ascii_lz(const char * input,char ** output,uint8_t flags)113 int mutt_idna_to_ascii_lz(const char *input, char **output, uint8_t flags)
114 {
115 if (!input || !output)
116 return 1;
117
118 #if (IDN_VERSION == 2)
119 return idn2_to_ascii_8z(input, output, flags | IDN2_NFC_INPUT | IDN2_NONTRANSITIONAL);
120 #else
121 return idna_to_ascii_lz(input, output, flags);
122 #endif
123 }
124 #endif /* HAVE_LIBIDN */
125
126 /**
127 * mutt_idna_intl_to_local - Convert an email's domain from Punycode
128 * @param user Username
129 * @param domain Domain
130 * @param flags Flags, e.g. #MI_MAY_BE_IRREVERSIBLE
131 * @retval ptr Newly allocated local email address
132 * @retval NULL Error in conversion
133 *
134 * If `$idn_decode` is set, then the domain will be converted from Punycode.
135 * For example, "xn--ls8h.la" becomes the emoji domain: ":poop:.la"
136 * Then the user and domain are changed from 'utf-8' to the encoding in
137 * `$charset`.
138 *
139 * If the flag #MI_MAY_BE_IRREVERSIBLE is NOT given, then the results will be
140 * checked to make sure that the transformation is "undo-able".
141 *
142 * @note The caller must free the returned string.
143 */
mutt_idna_intl_to_local(const char * user,const char * domain,uint8_t flags)144 char *mutt_idna_intl_to_local(const char *user, const char *domain, uint8_t flags)
145 {
146 char *mailbox = NULL;
147 char *reversed_user = NULL, *reversed_domain = NULL;
148 char *tmp = NULL;
149
150 char *local_user = mutt_str_dup(user);
151 char *local_domain = mutt_str_dup(domain);
152
153 #ifdef HAVE_LIBIDN
154 bool is_idn_encoded = check_idn(local_domain);
155 const bool c_idn_decode = cs_subset_bool(NeoMutt->sub, "idn_decode");
156 if (is_idn_encoded && c_idn_decode)
157 {
158 #if (IDN_VERSION == 2)
159 if (idn2_to_unicode_8z8z(local_domain, &tmp, IDN2_ALLOW_UNASSIGNED) != IDN2_OK)
160 #else
161 if (idna_to_unicode_8z8z(local_domain, &tmp, IDNA_ALLOW_UNASSIGNED) != IDNA_SUCCESS)
162 #endif
163 {
164 goto cleanup;
165 }
166 mutt_str_replace(&local_domain, tmp);
167 FREE(&tmp);
168 }
169 #endif /* HAVE_LIBIDN */
170
171 /* we don't want charset-hook effects, so we set flags to 0 */
172 const char *const c_charset = cs_subset_string(NeoMutt->sub, "charset");
173 if (mutt_ch_convert_string(&local_user, "utf-8", c_charset, MUTT_ICONV_NO_FLAGS) != 0)
174 goto cleanup;
175
176 if (mutt_ch_convert_string(&local_domain, "utf-8", c_charset, MUTT_ICONV_NO_FLAGS) != 0)
177 goto cleanup;
178
179 /* make sure that we can convert back and come out with the same
180 * user and domain name. */
181 if ((flags & MI_MAY_BE_IRREVERSIBLE) == 0)
182 {
183 reversed_user = mutt_str_dup(local_user);
184
185 if (mutt_ch_convert_string(&reversed_user, c_charset, "utf-8", MUTT_ICONV_NO_FLAGS) != 0)
186 {
187 mutt_debug(LL_DEBUG1, "Not reversible. Charset conv to utf-8 failed for user = '%s'\n",
188 reversed_user);
189 goto cleanup;
190 }
191
192 if (!mutt_istr_equal(user, reversed_user))
193 {
194 mutt_debug(LL_DEBUG1, "#1 Not reversible. orig = '%s', reversed = '%s'\n",
195 user, reversed_user);
196 goto cleanup;
197 }
198
199 reversed_domain = mutt_str_dup(local_domain);
200
201 if (mutt_ch_convert_string(&reversed_domain, c_charset, "utf-8", MUTT_ICONV_NO_FLAGS) != 0)
202 {
203 mutt_debug(LL_DEBUG1, "Not reversible. Charset conv to utf-8 failed for domain = '%s'\n",
204 reversed_domain);
205 goto cleanup;
206 }
207
208 #ifdef HAVE_LIBIDN
209 /* If the original domain was UTF-8, idna encoding here could
210 * produce a non-matching domain! Thus we only want to do the
211 * idna_to_ascii_8z() if the original domain was IDNA encoded. */
212 if (is_idn_encoded && c_idn_decode)
213 {
214 #if (IDN_VERSION == 2)
215 if (idn2_to_ascii_8z(reversed_domain, &tmp,
216 IDN2_ALLOW_UNASSIGNED | IDN2_NFC_INPUT | IDN2_NONTRANSITIONAL) != IDN2_OK)
217 #else
218 if (idna_to_ascii_8z(reversed_domain, &tmp, IDNA_ALLOW_UNASSIGNED) != IDNA_SUCCESS)
219 #endif
220 {
221 mutt_debug(LL_DEBUG1, "Not reversible. idna_to_ascii_8z failed for domain = '%s'\n",
222 reversed_domain);
223 goto cleanup;
224 }
225 mutt_str_replace(&reversed_domain, tmp);
226 }
227 #endif /* HAVE_LIBIDN */
228
229 if (!mutt_istr_equal(domain, reversed_domain))
230 {
231 mutt_debug(LL_DEBUG1, "#2 Not reversible. orig = '%s', reversed = '%s'\n",
232 domain, reversed_domain);
233 goto cleanup;
234 }
235 }
236
237 mailbox = mutt_mem_malloc(mutt_str_len(local_user) + mutt_str_len(local_domain) + 2);
238 sprintf(mailbox, "%s@%s", NONULL(local_user), NONULL(local_domain));
239
240 cleanup:
241 FREE(&local_user);
242 FREE(&local_domain);
243 FREE(&tmp);
244 FREE(&reversed_domain);
245 FREE(&reversed_user);
246
247 return mailbox;
248 }
249
250 /**
251 * mutt_idna_local_to_intl - Convert an email's domain to Punycode
252 * @param user Username
253 * @param domain Domain
254 * @retval ptr Newly allocated Punycode email address
255 * @retval NULL Error in conversion
256 *
257 * The user and domain are assumed to be encoded according to `$charset`.
258 * They are converted to 'utf-8'. If `$idn_encode` is set, then the domain
259 * will be converted to Punycode. For example, the emoji domain:
260 * ":poop:.la" becomes "xn--ls8h.la"
261 *
262 * @note The caller must free the returned string.
263 */
mutt_idna_local_to_intl(const char * user,const char * domain)264 char *mutt_idna_local_to_intl(const char *user, const char *domain)
265 {
266 char *mailbox = NULL;
267 char *tmp = NULL;
268
269 char *intl_user = mutt_str_dup(user);
270 char *intl_domain = mutt_str_dup(domain);
271
272 /* we don't want charset-hook effects, so we set flags to 0 */
273 const char *const c_charset = cs_subset_string(NeoMutt->sub, "charset");
274 if (mutt_ch_convert_string(&intl_user, c_charset, "utf-8", MUTT_ICONV_NO_FLAGS) != 0)
275 goto cleanup;
276
277 if (mutt_ch_convert_string(&intl_domain, c_charset, "utf-8", MUTT_ICONV_NO_FLAGS) != 0)
278 goto cleanup;
279
280 #ifdef HAVE_LIBIDN
281 const bool c_idn_encode = cs_subset_bool(NeoMutt->sub, "idn_encode");
282 if (c_idn_encode)
283 {
284 #if (IDN_VERSION == 2)
285 if (idn2_to_ascii_8z(intl_domain, &tmp,
286 IDN2_ALLOW_UNASSIGNED | IDN2_NFC_INPUT | IDN2_NONTRANSITIONAL) != IDN2_OK)
287 #else
288 if (idna_to_ascii_8z(intl_domain, &tmp, IDNA_ALLOW_UNASSIGNED) != IDNA_SUCCESS)
289 #endif
290 {
291 goto cleanup;
292 }
293 mutt_str_replace(&intl_domain, tmp);
294 }
295 #endif /* HAVE_LIBIDN */
296
297 mailbox = mutt_mem_malloc(mutt_str_len(intl_user) + mutt_str_len(intl_domain) + 2);
298 sprintf(mailbox, "%s@%s", NONULL(intl_user), NONULL(intl_domain));
299
300 cleanup:
301 FREE(&intl_user);
302 FREE(&intl_domain);
303 FREE(&tmp);
304
305 return mailbox;
306 }
307
308 /**
309 * mutt_idna_print_version - Create an IDN version string
310 * @retval ptr Version string
311 *
312 * @note This is a static string and must not be freed.
313 */
mutt_idna_print_version(void)314 const char *mutt_idna_print_version(void)
315 {
316 static char vstring[256];
317
318 #ifdef HAVE_LIBIDN
319 #if (IDN_VERSION == 2)
320 snprintf(vstring, sizeof(vstring), "libidn2: %s (compiled with %s)",
321 idn2_check_version(NULL), IDN2_VERSION);
322 #elif (IDN_VERSION == 1)
323 snprintf(vstring, sizeof(vstring), "libidn: %s (compiled with %s)",
324 stringprep_check_version(NULL), STRINGPREP_VERSION);
325 #endif
326 #endif /* HAVE_LIBIDN */
327
328 return vstring;
329 }
330