1 /*************************************************
2 * Exim - an Internet mail transport agent *
3 *************************************************/
5 /* Copyright (c) The Exim Maintainers 2022 */
6 /* Copyright (c) Jeremy Harris 2015 - 2018 */
7 /* See the file NOTICE for conditions of use and distribution. */
14 #ifdef SUPPORT_I18N_2008
21 #include <stringprep.h>
24 string_localpart_alabel_to_utf8_(const uschar * alabel, uschar ** err);
26 /**************************************************/
29 string_is_utf8(const uschar * s)
32 if (s) while ((c = *s++)) if (c & 0x80) return TRUE;
37 string_is_alabel(const uschar * s)
39 return s[0] == 'x' && s[1] == 'n' && s[2] == '-' && s[3] == '-';
42 /**************************************************/
43 /* Domain conversions.
44 The *err string pointer should be null before the call
46 Return NULL for error, with optional errstr pointer filled in
50 string_domain_utf8_to_alabel(const uschar * utf8, uschar ** err)
55 #ifdef SUPPORT_I18N_2008
56 /* Avoid lowercasing plain-ascii domains */
57 if (!string_is_utf8(utf8))
58 return string_copy(utf8);
60 /* Only lowercase is accepted by the library call. A pity since we lose
61 any mixed-case annotation. This does not really matter for a domain. */
64 for (s1 = s = US utf8; (c = *s1); s1++) if (!(c & 0x80) && isupper(c))
66 s = string_copy(utf8);
67 for (s1 = s + (s1 - utf8); (c = *s1); s1++) if (!(c & 0x80) && isupper(c))
72 if ((rc = idn2_lookup_u8((const uint8_t *) s, &s1, IDN2_NFC_INPUT)) != IDN2_OK)
74 if (err) *err = US idn2_strerror(rc);
78 s = US stringprep_utf8_nfkc_normalize(CCS utf8, -1);
79 if ( (rc = idna_to_ascii_8z(CCS s, CSS &s1, IDNA_ALLOW_UNASSIGNED))
83 if (err) *err = US idna_strerror(rc);
96 string_domain_alabel_to_utf8(const uschar * alabel, uschar ** err)
98 #ifdef SUPPORT_I18N_2008
103 while (label = string_nextinlist(&alabel, &sep, NULL, 0))
104 if ( string_is_alabel(label)
105 && !(label = string_localpart_alabel_to_utf8_(label, err))
109 g = string_append_listele(g, '.', label);
110 return string_from_gstring(g);
117 if ( (rc = idna_to_unicode_8z8z(CCS alabel, CSS &s1, IDNA_USE_STD3_ASCII_RULES))
120 if (err) *err = US idna_strerror(rc);
129 /**************************************************/
130 /* localpart conversions */
131 /* the *err string pointer should be null before the call */
135 string_localpart_utf8_to_alabel(const uschar * utf8, uschar ** err)
143 if (!string_is_utf8(utf8)) return string_copy(utf8);
145 p = (punycode_uint *) stringprep_utf8_to_ucs4(CCS utf8, -1, &ucs4_len);
148 if (err) *err = US"l_u2a: bad UTF-8 input";
151 p_len = ucs4_len*4; /* this multiplier is pure guesswork */
152 res = store_get(p_len+5, utf8);
154 res[0] = 'x'; res[1] = 'n'; res[2] = res[3] = '-';
156 if ((rc = punycode_encode(ucs4_len, p, NULL, &p_len, CS res+4)) != PUNYCODE_SUCCESS)
158 DEBUG(D_expand) debug_printf("l_u2a: bad '%s'\n", punycode_strerror(rc));
160 if (err) *err = US punycode_strerror(rc);
171 string_localpart_alabel_to_utf8_(const uschar * alabel, uschar ** err)
178 DEBUG(D_expand) debug_printf("l_a2u: '%s'\n", alabel);
180 p_len = Ustrlen(alabel);
181 p = store_get((p_len+1) * sizeof(*p), alabel);
183 if ((rc = punycode_decode(p_len, CCS alabel, &p_len, p, NULL)) != PUNYCODE_SUCCESS)
185 if (err) *err = US punycode_strerror(rc);
189 s = US stringprep_ucs4_to_utf8(p, p_len, NULL, &p_len);
190 res = string_copyn(s, p_len);
197 string_localpart_alabel_to_utf8(const uschar * alabel, uschar ** err)
199 if (string_is_alabel(alabel))
200 return string_localpart_alabel_to_utf8_(alabel, err);
202 if (err) *err = US"bad alabel prefix";
207 /**************************************************/
208 /* Whole address conversion.
209 The *err string pointer should be null before the call.
211 Return NULL on error, with (optional) errstring pointer filled in
215 string_address_utf8_to_alabel(const uschar * utf8, uschar ** err)
219 if (!*utf8) return string_copy(utf8);
221 DEBUG(D_expand) debug_printf("addr from utf8 <%s>", utf8);
223 for (const uschar * s = utf8; *s; s++)
226 l = string_copyn(utf8, s - utf8);
227 if ( !(l = string_localpart_utf8_to_alabel(l, err))
228 || !(d = string_domain_utf8_to_alabel(++s, err))
231 l = string_sprintf("%s@%s", l, d);
232 DEBUG(D_expand) debug_printf(" -> <%s>\n", l);
236 l = string_localpart_utf8_to_alabel(utf8, err);
237 DEBUG(D_expand) debug_printf(" -> <%s>\n", l);
243 /*************************************************
244 * Report the library versions. *
245 *************************************************/
247 /* See a description in tls-openssl.c for an explanation of why this exists.
249 Arguments: string to append to
254 utf8_version_report(gstring * g)
256 #ifdef SUPPORT_I18N_2008
257 g = string_fmt_append(g, "Library version: IDN2: Compile: %s\n"
260 idn2_check_version(NULL));
261 g = string_fmt_append(g, "Library version: Stringprep: Compile: %s\n"
264 stringprep_check_version(NULL));
266 g = string_fmt_append(g, "Library version: IDN: Compile: %s\n"
269 stringprep_check_version(NULL));
274 #endif /* whole file */