1 /*
2 +----------------------------------------------------------------------+
3 | PHP Version 7 |
4 +----------------------------------------------------------------------+
5 | Copyright (c) 2009 The PHP Group |
6 +----------------------------------------------------------------------+
7 | This source file is subject to version 3.01 of the PHP license, |
8 | that is bundled with this package in the file LICENSE, and is |
9 | available through the world-wide-web at the following url: |
10 | http://www.php.net/license/3_01.txt |
11 | If you did not receive a copy of the PHP license and are unable to |
12 | obtain it through the world-wide-web, please send a note to |
13 | license@php.net so we can mail you a copy immediately. |
14 +----------------------------------------------------------------------+
15 | Author: Pierre A. Joye <pierre@php.net> |
16 | Gustavo Lopes <cataphract@php.net> |
17 +----------------------------------------------------------------------+
18 */
19 /* $Id$ */
20
21 /* {{{ includes */
22 #ifdef HAVE_CONFIG_H
23 #include "config.h"
24 #endif
25
26 #include <php.h>
27
28 #include <unicode/uidna.h>
29 #include <unicode/ustring.h>
30 #include "ext/standard/php_string.h"
31
32 #include "intl_error.h"
33 #include "intl_convert.h"
34 /* }}} */
35
36 #ifdef UIDNA_INFO_INITIALIZER
37 #define HAVE_46_API 1 /* has UTS#46 API (introduced in ICU 4.6) */
38 #endif
39
40 enum {
41 INTL_IDN_VARIANT_2003 = 0,
42 INTL_IDN_VARIANT_UTS46
43 };
44
45 /* {{{ grapheme_register_constants
46 * Register API constants
47 */
idn_register_constants(INIT_FUNC_ARGS)48 void idn_register_constants( INIT_FUNC_ARGS )
49 {
50 /* OPTIONS */
51
52 /* Option to prohibit processing of unassigned codepoints in the input and
53 do not check if the input conforms to STD-3 ASCII rules. */
54 REGISTER_LONG_CONSTANT("IDNA_DEFAULT", UIDNA_DEFAULT, CONST_CS | CONST_PERSISTENT);
55
56 /* Option to allow processing of unassigned codepoints in the input */
57 REGISTER_LONG_CONSTANT("IDNA_ALLOW_UNASSIGNED", UIDNA_ALLOW_UNASSIGNED, CONST_CS | CONST_PERSISTENT);
58
59 /* Option to check if input conforms to STD-3 ASCII rules */
60 REGISTER_LONG_CONSTANT("IDNA_USE_STD3_RULES", UIDNA_USE_STD3_RULES, CONST_CS | CONST_PERSISTENT);
61
62 #ifdef HAVE_46_API
63
64 /* Option to check for whether the input conforms to the BiDi rules.
65 * Ignored by the IDNA2003 implementation. (IDNA2003 always performs a BiDi check.) */
66 REGISTER_LONG_CONSTANT("IDNA_CHECK_BIDI", UIDNA_CHECK_BIDI, CONST_CS | CONST_PERSISTENT);
67
68 /* Option to check for whether the input conforms to the CONTEXTJ rules.
69 * Ignored by the IDNA2003 implementation. (The CONTEXTJ check is new in IDNA2008.) */
70 REGISTER_LONG_CONSTANT("IDNA_CHECK_CONTEXTJ", UIDNA_CHECK_CONTEXTJ, CONST_CS | CONST_PERSISTENT);
71
72 /* Option for nontransitional processing in ToASCII().
73 * By default, ToASCII() uses transitional processing.
74 * Ignored by the IDNA2003 implementation. */
75 REGISTER_LONG_CONSTANT("IDNA_NONTRANSITIONAL_TO_ASCII", UIDNA_NONTRANSITIONAL_TO_ASCII, CONST_CS | CONST_PERSISTENT);
76
77 /* Option for nontransitional processing in ToUnicode().
78 * By default, ToUnicode() uses transitional processing.
79 * Ignored by the IDNA2003 implementation. */
80 REGISTER_LONG_CONSTANT("IDNA_NONTRANSITIONAL_TO_UNICODE", UIDNA_NONTRANSITIONAL_TO_UNICODE, CONST_CS | CONST_PERSISTENT);
81 #endif
82
83 /* VARIANTS */
84 REGISTER_LONG_CONSTANT("INTL_IDNA_VARIANT_2003", INTL_IDN_VARIANT_2003, CONST_CS | CONST_PERSISTENT);
85 #ifdef HAVE_46_API
86 REGISTER_LONG_CONSTANT("INTL_IDNA_VARIANT_UTS46", INTL_IDN_VARIANT_UTS46, CONST_CS | CONST_PERSISTENT);
87 #endif
88
89 #ifdef HAVE_46_API
90 /* PINFO ERROR CODES */
91 REGISTER_LONG_CONSTANT("IDNA_ERROR_EMPTY_LABEL", UIDNA_ERROR_EMPTY_LABEL, CONST_CS | CONST_PERSISTENT);
92 REGISTER_LONG_CONSTANT("IDNA_ERROR_LABEL_TOO_LONG", UIDNA_ERROR_LABEL_TOO_LONG, CONST_CS | CONST_PERSISTENT);
93 REGISTER_LONG_CONSTANT("IDNA_ERROR_DOMAIN_NAME_TOO_LONG", UIDNA_ERROR_DOMAIN_NAME_TOO_LONG, CONST_CS | CONST_PERSISTENT);
94 REGISTER_LONG_CONSTANT("IDNA_ERROR_LEADING_HYPHEN", UIDNA_ERROR_LEADING_HYPHEN, CONST_CS | CONST_PERSISTENT);
95 REGISTER_LONG_CONSTANT("IDNA_ERROR_TRAILING_HYPHEN", UIDNA_ERROR_TRAILING_HYPHEN, CONST_CS | CONST_PERSISTENT);
96 REGISTER_LONG_CONSTANT("IDNA_ERROR_HYPHEN_3_4", UIDNA_ERROR_HYPHEN_3_4, CONST_CS | CONST_PERSISTENT);
97 REGISTER_LONG_CONSTANT("IDNA_ERROR_LEADING_COMBINING_MARK", UIDNA_ERROR_LEADING_COMBINING_MARK, CONST_CS | CONST_PERSISTENT);
98 REGISTER_LONG_CONSTANT("IDNA_ERROR_DISALLOWED", UIDNA_ERROR_DISALLOWED, CONST_CS | CONST_PERSISTENT);
99 REGISTER_LONG_CONSTANT("IDNA_ERROR_PUNYCODE", UIDNA_ERROR_PUNYCODE, CONST_CS | CONST_PERSISTENT);
100 REGISTER_LONG_CONSTANT("IDNA_ERROR_LABEL_HAS_DOT", UIDNA_ERROR_LABEL_HAS_DOT, CONST_CS | CONST_PERSISTENT);
101 REGISTER_LONG_CONSTANT("IDNA_ERROR_INVALID_ACE_LABEL", UIDNA_ERROR_INVALID_ACE_LABEL, CONST_CS | CONST_PERSISTENT);
102 REGISTER_LONG_CONSTANT("IDNA_ERROR_BIDI", UIDNA_ERROR_BIDI, CONST_CS | CONST_PERSISTENT);
103 REGISTER_LONG_CONSTANT("IDNA_ERROR_CONTEXTJ", UIDNA_ERROR_CONTEXTJ, CONST_CS | CONST_PERSISTENT);
104 #endif
105 }
106 /* }}} */
107
108 enum {
109 INTL_IDN_TO_ASCII = 0,
110 INTL_IDN_TO_UTF8
111 };
112
113 /* like INTL_CHECK_STATUS, but as a function and varying the name of the func */
php_intl_idn_check_status(UErrorCode err,const char * msg)114 static int php_intl_idn_check_status(UErrorCode err, const char *msg)
115 {
116 intl_error_set_code(NULL, err);
117 if (U_FAILURE(err)) {
118 char *buff;
119 spprintf(&buff, 0, "%s: %s",
120 get_active_function_name(),
121 msg);
122 intl_error_set_custom_msg(NULL, buff, 1);
123 efree(buff);
124 return FAILURE;
125 }
126
127 return SUCCESS;
128 }
129
php_intl_bad_args(const char * msg)130 static inline void php_intl_bad_args(const char *msg)
131 {
132 php_intl_idn_check_status(U_ILLEGAL_ARGUMENT_ERROR, msg);
133 }
134
135 #ifdef HAVE_46_API
php_intl_idn_to_46(INTERNAL_FUNCTION_PARAMETERS,const zend_string * domain,uint32_t option,int mode,zval * idna_info)136 static void php_intl_idn_to_46(INTERNAL_FUNCTION_PARAMETERS,
137 const zend_string *domain, uint32_t option, int mode, zval *idna_info)
138 {
139 UErrorCode status = U_ZERO_ERROR;
140 UIDNA *uts46;
141 int32_t len;
142 int32_t buffer_capac = 255; /* no domain name may exceed this */
143 zend_string *buffer = zend_string_alloc(buffer_capac, 0);
144 UIDNAInfo info = UIDNA_INFO_INITIALIZER;
145 int buffer_used = 0;
146
147 uts46 = uidna_openUTS46(option, &status);
148 if (php_intl_idn_check_status(status, "failed to open UIDNA instance") == FAILURE) {
149 zend_string_free(buffer);
150 RETURN_FALSE;
151 }
152
153 if (mode == INTL_IDN_TO_ASCII) {
154 len = uidna_nameToASCII_UTF8(uts46, ZSTR_VAL(domain), ZSTR_LEN(domain),
155 ZSTR_VAL(buffer), buffer_capac, &info, &status);
156 } else {
157 len = uidna_nameToUnicodeUTF8(uts46, ZSTR_VAL(domain), ZSTR_LEN(domain),
158 ZSTR_VAL(buffer), buffer_capac, &info, &status);
159 }
160 if (len >= 255 || php_intl_idn_check_status(status, "failed to convert name") == FAILURE) {
161 uidna_close(uts46);
162 zend_string_free(buffer);
163 RETURN_FALSE;
164 }
165
166 ZSTR_VAL(buffer)[len] = '\0';
167 ZSTR_LEN(buffer) = len;
168
169 if (info.errors == 0) {
170 RETVAL_STR(buffer);
171 buffer_used = 1;
172 } else {
173 RETVAL_FALSE;
174 }
175
176 if (idna_info) {
177 if (buffer_used) { /* used in return_value then */
178 zval_addref_p(return_value);
179 add_assoc_zval_ex(idna_info, "result", sizeof("result")-1, return_value);
180 } else {
181 zval zv;
182 ZVAL_NEW_STR(&zv, buffer);
183 buffer_used = 1;
184 add_assoc_zval_ex(idna_info, "result", sizeof("result")-1, &zv);
185 }
186 add_assoc_bool_ex(idna_info, "isTransitionalDifferent",
187 sizeof("isTransitionalDifferent")-1, info.isTransitionalDifferent);
188 add_assoc_long_ex(idna_info, "errors", sizeof("errors")-1, (zend_long)info.errors);
189 }
190
191 if (!buffer_used) {
192 zend_string_free(buffer);
193 }
194
195 uidna_close(uts46);
196 }
197 #endif
198
php_intl_idn_to(INTERNAL_FUNCTION_PARAMETERS,const zend_string * domain,uint32_t option,int mode)199 static void php_intl_idn_to(INTERNAL_FUNCTION_PARAMETERS,
200 const zend_string *domain, uint32_t option, int mode)
201 {
202 UChar* ustring = NULL;
203 int ustring_len = 0;
204 UErrorCode status;
205 zend_string *u8str;
206 UChar converted[MAXPATHLEN];
207 int32_t converted_ret_len;
208
209 /* convert the string to UTF-16. */
210 status = U_ZERO_ERROR;
211 intl_convert_utf8_to_utf16(&ustring, &ustring_len, ZSTR_VAL(domain), ZSTR_LEN(domain), &status);
212
213 if (U_FAILURE(status)) {
214 intl_error_set_code(NULL, status);
215
216 /* Set error messages. */
217 intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 );
218 if (ustring) {
219 efree(ustring);
220 }
221 RETURN_FALSE;
222 } else {
223 UParseError parse_error;
224
225 status = U_ZERO_ERROR;
226 #if defined(__clang__)
227 # pragma clang diagnostic push
228 # pragma clang diagnostic ignored "-Wdeprecated-declarations"
229 #elif ZEND_GCC_VERSION >= 4008
230 # pragma GCC diagnostic push
231 # pragma GCC diagnostic ignored "-Wdeprecated-declarations"
232 #endif
233 if (mode == INTL_IDN_TO_ASCII) {
234 converted_ret_len = uidna_IDNToASCII(ustring, ustring_len, converted, MAXPATHLEN, (int32_t)option, &parse_error, &status);
235 } else {
236 converted_ret_len = uidna_IDNToUnicode(ustring, ustring_len, converted, MAXPATHLEN, (int32_t)option, &parse_error, &status);
237 }
238 #if defined(__clang__)
239 # pragma clang diagnostic pop
240 #elif ZEND_GCC_VERSION >= 4008
241 # pragma GCC diagnostic pop
242 #endif
243 efree(ustring);
244
245 if (U_FAILURE(status)) {
246 intl_error_set( NULL, status, "idn_to_ascii: cannot convert to ASCII", 0 );
247 RETURN_FALSE;
248 }
249
250 status = U_ZERO_ERROR;
251 u8str = intl_convert_utf16_to_utf8(converted, converted_ret_len, &status);
252
253 if (!u8str) {
254 /* Set global error code. */
255 intl_error_set_code(NULL, status);
256
257 /* Set error messages. */
258 intl_error_set_custom_msg( NULL, "Error converting output string to UTF-8", 0 );
259 RETURN_FALSE;
260 }
261 }
262
263 /* return the allocated string, not a duplicate */
264 RETVAL_NEW_STR(u8str);
265 }
266
php_intl_idn_handoff(INTERNAL_FUNCTION_PARAMETERS,int mode)267 static void php_intl_idn_handoff(INTERNAL_FUNCTION_PARAMETERS, int mode)
268 {
269 zend_string *domain;
270 zend_long option = 0,
271 variant = INTL_IDN_VARIANT_2003;
272 zval *idna_info = NULL;
273
274 intl_error_reset(NULL);
275
276 if (zend_parse_parameters(ZEND_NUM_ARGS(), "S|llz/",
277 &domain, &option, &variant, &idna_info) == FAILURE) {
278 php_intl_bad_args("bad arguments");
279 RETURN_NULL(); /* don't set FALSE because that's not the way it was before... */
280 }
281
282 #ifdef HAVE_46_API
283 if (variant != INTL_IDN_VARIANT_2003 && variant != INTL_IDN_VARIANT_UTS46) {
284 php_intl_bad_args("invalid variant, must be one of {"
285 "INTL_IDNA_VARIANT_2003, INTL_IDNA_VARIANT_UTS46}");
286 RETURN_FALSE;
287 }
288 #else
289 if (variant != INTL_IDN_VARIANT_2003) {
290 php_intl_bad_args("invalid variant, PHP was compiled against "
291 "an old version of ICU and only supports INTL_IDN_VARIANT_2003");
292 RETURN_FALSE;
293 }
294 #endif
295
296 if (ZSTR_LEN(domain) < 1) {
297 php_intl_bad_args("empty domain name");
298 RETURN_FALSE;
299 }
300 if (ZSTR_LEN(domain) > INT32_MAX - 1) {
301 php_intl_bad_args("domain name too large");
302 RETURN_FALSE;
303 }
304 /* don't check options; it wasn't checked before */
305
306 if (variant == INTL_IDN_VARIANT_2003) {
307 php_error_docref(NULL, E_DEPRECATED, "INTL_IDNA_VARIANT_2003 is deprecated");
308 }
309
310 if (idna_info != NULL) {
311 if (variant == INTL_IDN_VARIANT_2003) {
312 php_error_docref0(NULL, E_NOTICE,
313 "4 arguments were provided, but INTL_IDNA_VARIANT_2003 only "
314 "takes 3 - extra argument ignored");
315 } else {
316 zval_dtor(idna_info);
317 array_init(idna_info);
318 }
319 }
320
321 if (variant == INTL_IDN_VARIANT_2003) {
322 php_intl_idn_to(INTERNAL_FUNCTION_PARAM_PASSTHRU, domain, (uint32_t)option, mode);
323 }
324 #ifdef HAVE_46_API
325 else {
326 php_intl_idn_to_46(INTERNAL_FUNCTION_PARAM_PASSTHRU, domain, (uint32_t)option, mode, idna_info);
327 }
328 #endif
329 }
330
331 /* {{{ proto string idn_to_ascii(string domain[, int options[, int variant[, array &idna_info]]])
332 Converts an Unicode domain to ASCII representation, as defined in the IDNA RFC */
PHP_FUNCTION(idn_to_ascii)333 PHP_FUNCTION(idn_to_ascii)
334 {
335 php_intl_idn_handoff(INTERNAL_FUNCTION_PARAM_PASSTHRU, INTL_IDN_TO_ASCII);
336 }
337 /* }}} */
338
339
340 /* {{{ proto string idn_to_utf8(string domain[, int options[, int variant[, array &idna_info]]])
341 Converts an ASCII representation of the domain to Unicode (UTF-8), as defined in the IDNA RFC */
PHP_FUNCTION(idn_to_utf8)342 PHP_FUNCTION(idn_to_utf8)
343 {
344 php_intl_idn_handoff(INTERNAL_FUNCTION_PARAM_PASSTHRU, INTL_IDN_TO_UTF8);
345 }
346 /* }}} */
347
348
349 /*
350 * Local variables:
351 * tab-width: 4
352 * c-basic-offset: 4
353 * End:
354 * vim600: fdm=marker
355 * vim: noet sw=4 ts=4
356 */
357
