| 1 | /***************************************************************************
|
|---|
| 2 | * _ _ ____ _
|
|---|
| 3 | * Project ___| | | | _ \| |
|
|---|
| 4 | * / __| | | | |_) | |
|
|---|
| 5 | * | (__| |_| | _ <| |___
|
|---|
| 6 | * \___|\___/|_| \_\_____|
|
|---|
| 7 | *
|
|---|
| 8 | * Copyright (C) Daniel Stenberg, <[email protected]>, et al.
|
|---|
| 9 | *
|
|---|
| 10 | * This software is licensed as described in the file COPYING, which
|
|---|
| 11 | * you should have received as part of this distribution. The terms
|
|---|
| 12 | * are also available at https://curl.se/docs/copyright.html.
|
|---|
| 13 | *
|
|---|
| 14 | * You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
|---|
| 15 | * copies of the Software, and permit persons to whom the Software is
|
|---|
| 16 | * furnished to do so, under the terms of the COPYING file.
|
|---|
| 17 | *
|
|---|
| 18 | * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
|---|
| 19 | * KIND, either express or implied.
|
|---|
| 20 | *
|
|---|
| 21 | * SPDX-License-Identifier: curl
|
|---|
| 22 | *
|
|---|
| 23 | ***************************************************************************/
|
|---|
| 24 |
|
|---|
| 25 | /*
|
|---|
| 26 | * IDN conversions
|
|---|
| 27 | */
|
|---|
| 28 |
|
|---|
| 29 | #include "curl_setup.h"
|
|---|
| 30 | #include "urldata.h"
|
|---|
| 31 | #include "idn.h"
|
|---|
| 32 | #include "sendf.h"
|
|---|
| 33 | #include "curl_multibyte.h"
|
|---|
| 34 | #include "warnless.h"
|
|---|
| 35 |
|
|---|
| 36 | #ifdef USE_LIBIDN2
|
|---|
| 37 | #include <idn2.h>
|
|---|
| 38 |
|
|---|
| 39 | #if defined(_WIN32) && defined(UNICODE)
|
|---|
| 40 | #define IDN2_LOOKUP(name, host, flags) \
|
|---|
| 41 | idn2_lookup_u8((const uint8_t *)name, (uint8_t **)host, flags)
|
|---|
| 42 | #else
|
|---|
| 43 | #define IDN2_LOOKUP(name, host, flags) \
|
|---|
| 44 | idn2_lookup_ul((const char *)name, (char **)host, flags)
|
|---|
| 45 | #endif
|
|---|
| 46 | #endif /* USE_LIBIDN2 */
|
|---|
| 47 |
|
|---|
| 48 | /* The last 3 #include files should be in this order */
|
|---|
| 49 | #include "curl_printf.h"
|
|---|
| 50 | #include "curl_memory.h"
|
|---|
| 51 | #include "memdebug.h"
|
|---|
| 52 |
|
|---|
| 53 | /* for macOS and iOS targets */
|
|---|
| 54 | #if defined(USE_APPLE_IDN)
|
|---|
| 55 | #include <unicode/uidna.h>
|
|---|
| 56 | #include <iconv.h>
|
|---|
| 57 | #include <langinfo.h>
|
|---|
| 58 |
|
|---|
| 59 | #define MAX_HOST_LENGTH 512
|
|---|
| 60 |
|
|---|
| 61 | static CURLcode iconv_to_utf8(const char *in, size_t inlen,
|
|---|
| 62 | char **out, size_t *outlen)
|
|---|
| 63 | {
|
|---|
| 64 | iconv_t cd = iconv_open("UTF-8", nl_langinfo(CODESET));
|
|---|
| 65 | if(cd != (iconv_t)-1) {
|
|---|
| 66 | size_t iconv_outlen = *outlen;
|
|---|
| 67 | char *iconv_in = (char *)in;
|
|---|
| 68 | size_t iconv_inlen = inlen;
|
|---|
| 69 | size_t iconv_result = iconv(cd, &iconv_in, &iconv_inlen,
|
|---|
| 70 | out, &iconv_outlen);
|
|---|
| 71 | *outlen -= iconv_outlen;
|
|---|
| 72 | iconv_close(cd);
|
|---|
| 73 | if(iconv_result == (size_t)-1) {
|
|---|
| 74 | if(errno == ENOMEM)
|
|---|
| 75 | return CURLE_OUT_OF_MEMORY;
|
|---|
| 76 | else
|
|---|
| 77 | return CURLE_URL_MALFORMAT;
|
|---|
| 78 | }
|
|---|
| 79 |
|
|---|
| 80 | return CURLE_OK;
|
|---|
| 81 | }
|
|---|
| 82 | else {
|
|---|
| 83 | if(errno == ENOMEM)
|
|---|
| 84 | return CURLE_OUT_OF_MEMORY;
|
|---|
| 85 | else
|
|---|
| 86 | return CURLE_FAILED_INIT;
|
|---|
| 87 | }
|
|---|
| 88 | }
|
|---|
| 89 |
|
|---|
| 90 | static CURLcode mac_idn_to_ascii(const char *in, char **out)
|
|---|
| 91 | {
|
|---|
| 92 | size_t inlen = strlen(in);
|
|---|
| 93 | if(inlen < MAX_HOST_LENGTH) {
|
|---|
| 94 | char iconv_buffer[MAX_HOST_LENGTH] = {0};
|
|---|
| 95 | char *iconv_outptr = iconv_buffer;
|
|---|
| 96 | size_t iconv_outlen = sizeof(iconv_buffer);
|
|---|
| 97 | CURLcode iconv_result = iconv_to_utf8(in, inlen,
|
|---|
| 98 | &iconv_outptr, &iconv_outlen);
|
|---|
| 99 | if(!iconv_result) {
|
|---|
| 100 | UErrorCode err = U_ZERO_ERROR;
|
|---|
| 101 | UIDNA* idna = uidna_openUTS46(
|
|---|
| 102 | UIDNA_CHECK_BIDI|UIDNA_NONTRANSITIONAL_TO_ASCII, &err);
|
|---|
| 103 | if(!U_FAILURE(err)) {
|
|---|
| 104 | UIDNAInfo info = UIDNA_INFO_INITIALIZER;
|
|---|
| 105 | char buffer[MAX_HOST_LENGTH] = {0};
|
|---|
| 106 | (void)uidna_nameToASCII_UTF8(idna, iconv_buffer, (int)iconv_outlen,
|
|---|
| 107 | buffer, sizeof(buffer) - 1, &info, &err);
|
|---|
| 108 | uidna_close(idna);
|
|---|
| 109 | if(!U_FAILURE(err) && !info.errors) {
|
|---|
| 110 | *out = strdup(buffer);
|
|---|
| 111 | if(*out)
|
|---|
| 112 | return CURLE_OK;
|
|---|
| 113 | else
|
|---|
| 114 | return CURLE_OUT_OF_MEMORY;
|
|---|
| 115 | }
|
|---|
| 116 | }
|
|---|
| 117 | }
|
|---|
| 118 | else
|
|---|
| 119 | return iconv_result;
|
|---|
| 120 | }
|
|---|
| 121 | return CURLE_URL_MALFORMAT;
|
|---|
| 122 | }
|
|---|
| 123 |
|
|---|
| 124 | static CURLcode mac_ascii_to_idn(const char *in, char **out)
|
|---|
| 125 | {
|
|---|
| 126 | size_t inlen = strlen(in);
|
|---|
| 127 | if(inlen < MAX_HOST_LENGTH) {
|
|---|
| 128 | UErrorCode err = U_ZERO_ERROR;
|
|---|
| 129 | UIDNA* idna = uidna_openUTS46(
|
|---|
| 130 | UIDNA_CHECK_BIDI|UIDNA_NONTRANSITIONAL_TO_UNICODE, &err);
|
|---|
| 131 | if(!U_FAILURE(err)) {
|
|---|
| 132 | UIDNAInfo info = UIDNA_INFO_INITIALIZER;
|
|---|
| 133 | char buffer[MAX_HOST_LENGTH] = {0};
|
|---|
| 134 | (void)uidna_nameToUnicodeUTF8(idna, in, -1, buffer,
|
|---|
| 135 | sizeof(buffer) - 1, &info, &err);
|
|---|
| 136 | uidna_close(idna);
|
|---|
| 137 | if(!U_FAILURE(err)) {
|
|---|
| 138 | *out = strdup(buffer);
|
|---|
| 139 | if(*out)
|
|---|
| 140 | return CURLE_OK;
|
|---|
| 141 | else
|
|---|
| 142 | return CURLE_OUT_OF_MEMORY;
|
|---|
| 143 | }
|
|---|
| 144 | }
|
|---|
| 145 | }
|
|---|
| 146 | return CURLE_URL_MALFORMAT;
|
|---|
| 147 | }
|
|---|
| 148 | #endif
|
|---|
| 149 |
|
|---|
| 150 | #ifdef USE_WIN32_IDN
|
|---|
| 151 | /* using Windows kernel32 and normaliz libraries. */
|
|---|
| 152 |
|
|---|
| 153 | #if (!defined(_WIN32_WINNT) || _WIN32_WINNT < 0x600) && \
|
|---|
| 154 | (!defined(WINVER) || WINVER < 0x600)
|
|---|
| 155 | WINBASEAPI int WINAPI IdnToAscii(DWORD dwFlags,
|
|---|
| 156 | const WCHAR *lpUnicodeCharStr,
|
|---|
| 157 | int cchUnicodeChar,
|
|---|
| 158 | WCHAR *lpASCIICharStr,
|
|---|
| 159 | int cchASCIIChar);
|
|---|
| 160 | WINBASEAPI int WINAPI IdnToUnicode(DWORD dwFlags,
|
|---|
| 161 | const WCHAR *lpASCIICharStr,
|
|---|
| 162 | int cchASCIIChar,
|
|---|
| 163 | WCHAR *lpUnicodeCharStr,
|
|---|
| 164 | int cchUnicodeChar);
|
|---|
| 165 | #endif
|
|---|
| 166 |
|
|---|
| 167 | #define IDN_MAX_LENGTH 255
|
|---|
| 168 |
|
|---|
| 169 | static CURLcode win32_idn_to_ascii(const char *in, char **out)
|
|---|
| 170 | {
|
|---|
| 171 | wchar_t *in_w = curlx_convert_UTF8_to_wchar(in);
|
|---|
| 172 | *out = NULL;
|
|---|
| 173 | if(in_w) {
|
|---|
| 174 | wchar_t punycode[IDN_MAX_LENGTH];
|
|---|
| 175 | int chars = IdnToAscii(0, in_w, (int)(wcslen(in_w) + 1), punycode,
|
|---|
| 176 | IDN_MAX_LENGTH);
|
|---|
| 177 | curlx_unicodefree(in_w);
|
|---|
| 178 | if(chars) {
|
|---|
| 179 | char *mstr = curlx_convert_wchar_to_UTF8(punycode);
|
|---|
| 180 | if(mstr) {
|
|---|
| 181 | *out = strdup(mstr);
|
|---|
| 182 | curlx_unicodefree(mstr);
|
|---|
| 183 | if(!*out)
|
|---|
| 184 | return CURLE_OUT_OF_MEMORY;
|
|---|
| 185 | }
|
|---|
| 186 | else
|
|---|
| 187 | return CURLE_OUT_OF_MEMORY;
|
|---|
| 188 | }
|
|---|
| 189 | else
|
|---|
| 190 | return CURLE_URL_MALFORMAT;
|
|---|
| 191 | }
|
|---|
| 192 | else
|
|---|
| 193 | return CURLE_URL_MALFORMAT;
|
|---|
| 194 |
|
|---|
| 195 | return CURLE_OK;
|
|---|
| 196 | }
|
|---|
| 197 |
|
|---|
| 198 | static CURLcode win32_ascii_to_idn(const char *in, char **output)
|
|---|
| 199 | {
|
|---|
| 200 | char *out = NULL;
|
|---|
| 201 |
|
|---|
| 202 | wchar_t *in_w = curlx_convert_UTF8_to_wchar(in);
|
|---|
| 203 | if(in_w) {
|
|---|
| 204 | WCHAR idn[IDN_MAX_LENGTH]; /* stores a UTF-16 string */
|
|---|
| 205 | int chars = IdnToUnicode(0, in_w, (int)(wcslen(in_w) + 1), idn,
|
|---|
| 206 | IDN_MAX_LENGTH);
|
|---|
| 207 | if(chars) {
|
|---|
| 208 | /* 'chars' is "the number of characters retrieved" */
|
|---|
| 209 | char *mstr = curlx_convert_wchar_to_UTF8(idn);
|
|---|
| 210 | if(mstr) {
|
|---|
| 211 | out = strdup(mstr);
|
|---|
| 212 | curlx_unicodefree(mstr);
|
|---|
| 213 | if(!out)
|
|---|
| 214 | return CURLE_OUT_OF_MEMORY;
|
|---|
| 215 | }
|
|---|
| 216 | }
|
|---|
| 217 | else
|
|---|
| 218 | return CURLE_URL_MALFORMAT;
|
|---|
| 219 | }
|
|---|
| 220 | else
|
|---|
| 221 | return CURLE_URL_MALFORMAT;
|
|---|
| 222 | *output = out;
|
|---|
| 223 | return CURLE_OK;
|
|---|
| 224 | }
|
|---|
| 225 |
|
|---|
| 226 | #endif /* USE_WIN32_IDN */
|
|---|
| 227 |
|
|---|
| 228 | /*
|
|---|
| 229 | * Helpers for IDNA conversions.
|
|---|
| 230 | */
|
|---|
| 231 | bool Curl_is_ASCII_name(const char *hostname)
|
|---|
| 232 | {
|
|---|
| 233 | /* get an UNSIGNED local version of the pointer */
|
|---|
| 234 | const unsigned char *ch = (const unsigned char *)hostname;
|
|---|
| 235 |
|
|---|
| 236 | if(!hostname) /* bad input, consider it ASCII! */
|
|---|
| 237 | return TRUE;
|
|---|
| 238 |
|
|---|
| 239 | while(*ch) {
|
|---|
| 240 | if(*ch++ & 0x80)
|
|---|
| 241 | return FALSE;
|
|---|
| 242 | }
|
|---|
| 243 | return TRUE;
|
|---|
| 244 | }
|
|---|
| 245 |
|
|---|
| 246 | #ifdef USE_IDN
|
|---|
| 247 | /*
|
|---|
| 248 | * Curl_idn_decode() returns an allocated IDN decoded string if it was
|
|---|
| 249 | * possible. NULL on error.
|
|---|
| 250 | *
|
|---|
| 251 | * CURLE_URL_MALFORMAT - the hostname could not be converted
|
|---|
| 252 | * CURLE_OUT_OF_MEMORY - memory problem
|
|---|
| 253 | *
|
|---|
| 254 | */
|
|---|
| 255 | static CURLcode idn_decode(const char *input, char **output)
|
|---|
| 256 | {
|
|---|
| 257 | char *decoded = NULL;
|
|---|
| 258 | CURLcode result = CURLE_OK;
|
|---|
| 259 | #ifdef USE_LIBIDN2
|
|---|
| 260 | if(idn2_check_version(IDN2_VERSION)) {
|
|---|
| 261 | int flags = IDN2_NFC_INPUT
|
|---|
| 262 | #if IDN2_VERSION_NUMBER >= 0x00140000
|
|---|
| 263 | /* IDN2_NFC_INPUT: Normalize input string using normalization form C.
|
|---|
| 264 | IDN2_NONTRANSITIONAL: Perform Unicode TR46 non-transitional
|
|---|
| 265 | processing. */
|
|---|
| 266 | | IDN2_NONTRANSITIONAL
|
|---|
| 267 | #endif
|
|---|
| 268 | ;
|
|---|
| 269 | int rc = IDN2_LOOKUP(input, &decoded, flags);
|
|---|
| 270 | if(rc != IDN2_OK)
|
|---|
| 271 | /* fallback to TR46 Transitional mode for better IDNA2003
|
|---|
| 272 | compatibility */
|
|---|
| 273 | rc = IDN2_LOOKUP(input, &decoded, IDN2_TRANSITIONAL);
|
|---|
| 274 | if(rc != IDN2_OK)
|
|---|
| 275 | result = CURLE_URL_MALFORMAT;
|
|---|
| 276 | }
|
|---|
| 277 | else
|
|---|
| 278 | /* a too old libidn2 version */
|
|---|
| 279 | result = CURLE_NOT_BUILT_IN;
|
|---|
| 280 | #elif defined(USE_WIN32_IDN)
|
|---|
| 281 | result = win32_idn_to_ascii(input, &decoded);
|
|---|
| 282 | #elif defined(USE_APPLE_IDN)
|
|---|
| 283 | result = mac_idn_to_ascii(input, &decoded);
|
|---|
| 284 | #endif
|
|---|
| 285 | if(!result)
|
|---|
| 286 | *output = decoded;
|
|---|
| 287 | return result;
|
|---|
| 288 | }
|
|---|
| 289 |
|
|---|
| 290 | static CURLcode idn_encode(const char *puny, char **output)
|
|---|
| 291 | {
|
|---|
| 292 | char *enc = NULL;
|
|---|
| 293 | #ifdef USE_LIBIDN2
|
|---|
| 294 | int rc = idn2_to_unicode_8z8z(puny, &enc, 0);
|
|---|
| 295 | if(rc != IDNA_SUCCESS)
|
|---|
| 296 | return rc == IDNA_MALLOC_ERROR ? CURLE_OUT_OF_MEMORY : CURLE_URL_MALFORMAT;
|
|---|
| 297 | #elif defined(USE_WIN32_IDN)
|
|---|
| 298 | CURLcode result = win32_ascii_to_idn(puny, &enc);
|
|---|
| 299 | if(result)
|
|---|
| 300 | return result;
|
|---|
| 301 | #elif defined(USE_APPLE_IDN)
|
|---|
| 302 | CURLcode result = mac_ascii_to_idn(puny, &enc);
|
|---|
| 303 | if(result)
|
|---|
| 304 | return result;
|
|---|
| 305 | #endif
|
|---|
| 306 | *output = enc;
|
|---|
| 307 | return CURLE_OK;
|
|---|
| 308 | }
|
|---|
| 309 |
|
|---|
| 310 | CURLcode Curl_idn_decode(const char *input, char **output)
|
|---|
| 311 | {
|
|---|
| 312 | char *d = NULL;
|
|---|
| 313 | CURLcode result = idn_decode(input, &d);
|
|---|
| 314 | #ifdef USE_LIBIDN2
|
|---|
| 315 | if(!result) {
|
|---|
| 316 | char *c = strdup(d);
|
|---|
| 317 | idn2_free(d);
|
|---|
| 318 | if(c)
|
|---|
| 319 | d = c;
|
|---|
| 320 | else
|
|---|
| 321 | result = CURLE_OUT_OF_MEMORY;
|
|---|
| 322 | }
|
|---|
| 323 | #endif
|
|---|
| 324 | if(!result)
|
|---|
| 325 | *output = d;
|
|---|
| 326 | return result;
|
|---|
| 327 | }
|
|---|
| 328 |
|
|---|
| 329 | CURLcode Curl_idn_encode(const char *puny, char **output)
|
|---|
| 330 | {
|
|---|
| 331 | char *d = NULL;
|
|---|
| 332 | CURLcode result = idn_encode(puny, &d);
|
|---|
| 333 | #ifdef USE_LIBIDN2
|
|---|
| 334 | if(!result) {
|
|---|
| 335 | char *c = strdup(d);
|
|---|
| 336 | idn2_free(d);
|
|---|
| 337 | if(c)
|
|---|
| 338 | d = c;
|
|---|
| 339 | else
|
|---|
| 340 | result = CURLE_OUT_OF_MEMORY;
|
|---|
| 341 | }
|
|---|
| 342 | #endif
|
|---|
| 343 | if(!result)
|
|---|
| 344 | *output = d;
|
|---|
| 345 | return result;
|
|---|
| 346 | }
|
|---|
| 347 |
|
|---|
| 348 | /*
|
|---|
| 349 | * Frees data allocated by idnconvert_hostname()
|
|---|
| 350 | */
|
|---|
| 351 | void Curl_free_idnconverted_hostname(struct hostname *host)
|
|---|
| 352 | {
|
|---|
| 353 | Curl_safefree(host->encalloc);
|
|---|
| 354 | }
|
|---|
| 355 |
|
|---|
| 356 | #endif /* USE_IDN */
|
|---|
| 357 |
|
|---|
| 358 | /*
|
|---|
| 359 | * Perform any necessary IDN conversion of hostname
|
|---|
| 360 | */
|
|---|
| 361 | CURLcode Curl_idnconvert_hostname(struct hostname *host)
|
|---|
| 362 | {
|
|---|
| 363 | /* set the name we use to display the hostname */
|
|---|
| 364 | host->dispname = host->name;
|
|---|
| 365 |
|
|---|
| 366 | #ifdef USE_IDN
|
|---|
| 367 | /* Check name for non-ASCII and convert hostname if we can */
|
|---|
| 368 | if(!Curl_is_ASCII_name(host->name)) {
|
|---|
| 369 | char *decoded;
|
|---|
| 370 | CURLcode result = Curl_idn_decode(host->name, &decoded);
|
|---|
| 371 | if(result)
|
|---|
| 372 | return result;
|
|---|
| 373 | /* successful */
|
|---|
| 374 | host->name = host->encalloc = decoded;
|
|---|
| 375 | }
|
|---|
| 376 | #endif
|
|---|
| 377 | return CURLE_OK;
|
|---|
| 378 | }
|
|---|
