1 /*
2 * Copyright 2014 Chris Young <chris@unsatisfactorysoftware.co.uk>
3 *
4 * This file is part of NetSurf, http://www.netsurf-browser.org/
5 *
6 * NetSurf is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published by
8 * the Free Software Foundation; version 2 of the License.
9 *
10 * NetSurf is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License
16 * along with this program. If not, see <http://www.gnu.org/licenses/>.
17 */
18
19 /**
20 * \file
21 * NetSurf international domain name handling implementation.
22 */
23
24 #include <assert.h>
25 #include <stddef.h>
26 #include <stdint.h>
27 #include <stdlib.h>
28 #include <string.h>
29 #include <sys/types.h>
30
31 #include "netsurf/inttypes.h"
32
33 #include "utils/errors.h"
34 #include "utils/idna.h"
35 #include "utils/idna_props.h"
36 #include "utils/log.h"
37 #include "utils/punycode.h"
38 #include "utils/utf8.h"
39 #include "utils/utils.h"
40
41
42 /**
43 * Convert punycode status into nserror.
44 *
45 * \param status The punycode status to convert.
46 * \return The corresponding nserror code for the status.
47 */
punycode_status_to_nserror(enum punycode_status status)48 static nserror punycode_status_to_nserror(enum punycode_status status)
49 {
50 nserror ret = NSERROR_NOMEM;
51
52 switch (status) {
53 case punycode_success:
54 ret = NSERROR_OK;
55 break;
56
57 case punycode_bad_input:
58 NSLOG(netsurf, INFO, "Bad input");
59 ret = NSERROR_BAD_ENCODING;
60 break;
61
62 case punycode_big_output:
63 NSLOG(netsurf, INFO, "Output too big");
64 ret = NSERROR_BAD_SIZE;
65 break;
66
67 case punycode_overflow:
68 NSLOG(netsurf, INFO, "Overflow");
69 ret = NSERROR_NOSPACE;
70 break;
71
72 default:
73 break;
74 }
75 return ret;
76 }
77
78
79 /**
80 * Convert a host label in UCS-4 to an ACE version
81 *
82 * \param ucs4_label UCS-4 NFC string containing host label
83 * \param len Length of host label (in characters/codepoints)
84 * \param ace_label ASCII-compatible encoded version
85 * \param out_len Length of ace_label
86 * \return NSERROR_OK on success, appropriate error otherwise
87 *
88 * If return value != NSERROR_OK, output will be left untouched.
89 */
90 static nserror
idna__ucs4_to_ace(int32_t * ucs4_label,size_t len,char ** ace_label,size_t * out_len)91 idna__ucs4_to_ace(int32_t *ucs4_label,
92 size_t len,
93 char **ace_label,
94 size_t *out_len)
95 {
96 char punycode[65]; /* max length of host label + NULL */
97 size_t output_length = 60; /* punycode length - 4 - 1 */
98 nserror ret;
99
100 punycode[0] = 'x';
101 punycode[1] = 'n';
102 punycode[2] = '-';
103 punycode[3] = '-';
104
105 ret = punycode_status_to_nserror(punycode_encode(len,
106 (const punycode_uint *)ucs4_label, NULL,
107 &output_length, punycode + 4));
108 if (ret != NSERROR_OK) {
109 return ret;
110 }
111
112 output_length += SLEN("xn--");
113 punycode[output_length] = '\0';
114
115 *ace_label = strdup(punycode);
116 *out_len = output_length;
117
118 return NSERROR_OK;
119 }
120
121
122 /**
123 * Convert a host label in ACE format to UCS-4
124 *
125 * \param ace_label ASCII string containing host label
126 * \param ace_len Length of host label
127 * \param ucs4_label Pointer to hold UCS4 decoded version
128 * \param ucs4_len Pointer to hold length of ucs4_label
129 * \return NSERROR_OK on success, appropriate error otherwise
130 *
131 * If return value != NSERROR_OK, output will be left untouched.
132 */
133 static nserror
idna__ace_to_ucs4(const char * ace_label,size_t ace_len,int32_t ** ucs4_label,size_t * ucs4_len)134 idna__ace_to_ucs4(const char *ace_label,
135 size_t ace_len,
136 int32_t **ucs4_label,
137 size_t *ucs4_len)
138 {
139 int32_t *ucs4;
140 nserror ret;
141 size_t output_length = ace_len; /* never exceeds input length */
142
143 /* The header should always have been checked before calling */
144 assert((ace_label[0] == 'x') && (ace_label[1] == 'n') &&
145 (ace_label[2] == '-') && (ace_label[3] == '-'));
146
147 ucs4 = malloc(output_length * 4);
148 if (ucs4 == NULL) {
149 return NSERROR_NOMEM;
150 }
151
152 ret = punycode_status_to_nserror(punycode_decode(ace_len - 4,
153 ace_label + 4, &output_length, (punycode_uint *)ucs4, NULL));
154 if (ret != NSERROR_OK) {
155 free(ucs4);
156 return ret;
157 }
158
159 ucs4[output_length] = '\0';
160
161 *ucs4_label = ucs4;
162 *ucs4_len = output_length;
163
164 return NSERROR_OK;
165 }
166
167
168 #ifdef WITH_UTF8PROC
169
170 #include <utf8proc.h>
171
172 int32_t idna_contexto[] = {
173 /* CONTEXTO codepoints which have a rule defined */
174 0x00b7, 0x0375, 0x05f3, 0x05f4, 0x30fb, 0x0660, 0x0661,
175 0x0662, 0x0663, 0x0664, 0x0665, 0x0666, 0x0667, 0x0668,
176 0x0669, 0x06f0, 0x06f1, 0x06f2, 0x06f3, 0x06f4, 0x06f5,
177 0x06f6, 0x06f7, 0x06f8, 0x06f9, 0
178 };
179
180 /**
181 * Find the IDNA property of a UCS-4 codepoint
182 *
183 * \param cp Unicode codepoint
184 * \return IDNA property
185 */
idna__cp_property(int32_t cp)186 static idna_property idna__cp_property(int32_t cp)
187 {
188 const idna_table *t;
189
190 t = idna_derived;
191 while (t->p.property) {
192 if ((cp >= t->start) && (cp <= t->end)) {
193 return t->p.property;
194 }
195 t++;
196 };
197
198 return IDNA_P_DISALLOWED;
199 }
200
201
202 /**
203 * Find the Joining_Type property of a UCS-4 codepoint
204 *
205 * \param cp Unicode codepoint
206 * \return JT property
207 */
idna__jt_property(int32_t cp)208 static idna_unicode_jt idna__jt_property(int32_t cp)
209 {
210 const idna_table *t;
211
212 t = idna_joiningtype;
213 while (t->p.jt) {
214 if ((cp >= t->start) && (cp <= t->end)) {
215 return t->p.jt;
216 }
217 t++;
218 };
219
220 return IDNA_UNICODE_JT_U;
221 }
222
223
224 /**
225 * Check if a CONTEXTO codepoint has a rule defined
226 *
227 * \param cp Unicode codepoint
228 * \return true if a rule is defined
229 */
idna__contexto_rule(int32_t cp)230 static bool idna__contexto_rule(int32_t cp)
231 {
232 int32_t *t;
233 for (t = idna_contexto; *t != 0; t++) {
234 if (*t == cp) {
235 return true;
236 }
237 }
238
239 return false;
240 }
241
242
243 /**
244 * Check if a CONTEXTJ codepoint has a rule defined,
245 * and conforms to that rule.
246 *
247 * \param label UCS-4 string
248 * \param index character in the string which is CONTEXTJ
249 * \param len The length of the label
250 * \return true if conforming
251 */
idna__contextj_rule(int32_t * label,int index,size_t len)252 static bool idna__contextj_rule(int32_t *label, int index, size_t len)
253 {
254 const utf8proc_property_t *unicode_props;
255 idna_unicode_jt joining_type;
256 int i;
257 bool match;
258
259 /* These CONTEXTJ rules are defined at
260 * http://www.iana.org/assignments/idna-tables-5.2.0/idna-tables-5.2.0.xml
261 */
262
263 if (label[index] == 0x200c) {
264 if (index == 0) {
265 return false; /* No previous character */
266 }
267 unicode_props = utf8proc_get_property(label[index - 1]);
268 if (unicode_props->combining_class == IDNA_UNICODE_CCC_VIRAMA) {
269 return true;
270 }
271
272 match = false;
273 for (i = 0; i < (index - 1); i++) {
274 joining_type = idna__jt_property(label[i]);
275 if (((joining_type == IDNA_UNICODE_JT_L) ||
276 (joining_type == IDNA_UNICODE_JT_D)) &&
277 (idna__jt_property(label[i+1]) == IDNA_UNICODE_JT_T)) {
278 match = true;
279 break;
280 }
281 }
282
283 if (match == false) {
284 return false;
285 }
286
287 if (idna__jt_property(label[index+1]) != IDNA_UNICODE_JT_T) {
288 return false;
289 }
290
291 for (i = (index + 1); i < (int)len; i++) {
292 joining_type = idna__jt_property(label[i]);
293
294 if ((joining_type == IDNA_UNICODE_JT_R) ||
295 (joining_type == IDNA_UNICODE_JT_D)) {
296 return true;
297 }
298 }
299
300 return false;
301
302 } else if (label[index] == 0x200d) {
303 if (index == 0) {
304 return false; /* No previous character */
305 }
306 unicode_props = utf8proc_get_property(label[index - 1]);
307 if (unicode_props->combining_class == IDNA_UNICODE_CCC_VIRAMA) {
308 return true;
309 }
310 return false;
311 }
312
313 /* No rule defined */
314 return false;
315 }
316
317
318 /**
319 * Convert a UTF-8 string to UCS-4
320 *
321 * \param utf8_label UTF-8 string containing host label
322 * \param len Length of host label (in bytes)
323 * \param ucs4_label Pointer to update with the output
324 * \param ucs4_len Pointer to update with the length
325 * \return NSERROR_OK on success, appropriate error otherwise
326 *
327 * If return value != NSERROR_OK, output will be left untouched.
328 */
329 static nserror
idna__utf8_to_ucs4(const char * utf8_label,size_t len,int32_t ** ucs4_label,size_t * ucs4_len)330 idna__utf8_to_ucs4(const char *utf8_label,
331 size_t len,
332 int32_t **ucs4_label,
333 size_t *ucs4_len)
334 {
335 int32_t *nfc_label;
336 ssize_t nfc_size;
337
338 nfc_label = malloc(len * 4);
339 if (nfc_label == NULL) {
340 return NSERROR_NOMEM;
341 }
342
343 nfc_size = utf8proc_decompose((const uint8_t *)utf8_label, len,
344 nfc_label, len * 4, UTF8PROC_STABLE | UTF8PROC_COMPOSE);
345 if (nfc_size < 0) {
346 return NSERROR_NOMEM;
347 }
348
349 nfc_size = utf8proc_normalize_utf32(nfc_label, nfc_size,
350 UTF8PROC_STABLE | UTF8PROC_COMPOSE);
351 if (nfc_size < 0) {
352 return NSERROR_NOMEM;
353 }
354
355 *ucs4_label = nfc_label;
356 *ucs4_len = nfc_size;
357
358 return NSERROR_OK;
359 }
360
361
362 /**
363 * Convert a UCS-4 string to UTF-8
364 *
365 * \param ucs4_label UCS-4 string containing host label
366 * \param ucs4_len Length of host label (in bytes)
367 * \param utf8_label Pointer to update with the output
368 * \param utf8_len Pointer to update with the length
369 * \return NSERROR_OK on success, appropriate error otherwise
370 *
371 * If return value != NSERROR_OK, output will be left untouched.
372 */
373 static nserror
idna__ucs4_to_utf8(const int32_t * ucs4_label,size_t ucs4_len,char ** utf8_label,size_t * utf8_len)374 idna__ucs4_to_utf8(const int32_t *ucs4_label,
375 size_t ucs4_len,
376 char **utf8_label,
377 size_t *utf8_len)
378 {
379 int32_t *nfc_label;
380 ssize_t nfc_size = ucs4_len;
381
382 nfc_label = malloc(1 + ucs4_len * 4);
383 if (nfc_label == NULL) {
384 return NSERROR_NOMEM;
385 }
386 memcpy(nfc_label, ucs4_label, ucs4_len * 4);
387
388 nfc_size = utf8proc_reencode(nfc_label, ucs4_len,
389 UTF8PROC_STABLE | UTF8PROC_COMPOSE);
390 if (nfc_size < 0) {
391 return NSERROR_NOMEM;
392 }
393
394 *utf8_label = (char *)nfc_label;
395 *utf8_len = nfc_size;
396
397 return NSERROR_OK;
398 }
399
400
401 /**
402 * Check if a host label is valid for IDNA2008
403 *
404 * \param label Host label to check (UCS-4)
405 * \param len Length of host label (in characters/codepoints)
406 * \return true if compliant, false otherwise
407 */
idna__is_valid(int32_t * label,size_t len)408 static bool idna__is_valid(int32_t *label, size_t len)
409 {
410 const utf8proc_property_t *unicode_props;
411 idna_property idna_prop;
412 size_t i = 0;
413
414 /* 1. Check that the string is NFC.
415 * This check is skipped as the conversion to Unicode
416 * does normalisation as part of the conversion.
417 */
418
419 /* 2. Check characters 3 and 4 are not '--'. */
420 if ((len >= 4) && (label[2] == 0x002d) && (label[3] == 0x002d)) {
421 NSLOG(netsurf, INFO,
422 "Check failed: characters 2 and 3 are '--'");
423 return false;
424 }
425
426 /* 3. Check the first character is not a combining mark */
427 unicode_props = utf8proc_get_property(label[0]);
428
429 if ((unicode_props->category == UTF8PROC_CATEGORY_MN) ||
430 (unicode_props->category == UTF8PROC_CATEGORY_MC) ||
431 (unicode_props->category == UTF8PROC_CATEGORY_ME)) {
432 NSLOG(netsurf, INFO,
433 "Check failed: character 0 is a combining mark");
434 return false;
435 }
436
437 for (i = 0; i < len; i++) {
438 idna_prop = idna__cp_property(label[i]);
439
440 /* 4. Check characters not DISALLOWED by RFC5892 */
441 if (idna_prop == IDNA_P_DISALLOWED) {
442 NSLOG(netsurf, INFO,
443 "Check failed: character %"PRIsizet" (%x) is DISALLOWED",
444 i,
445 label[i]);
446 return false;
447 }
448
449 /* 5. Check CONTEXTJ characters conform to defined rules */
450 if (idna_prop == IDNA_P_CONTEXTJ) {
451 if (idna__contextj_rule(label, i, len) == false) {
452 NSLOG(netsurf, INFO,
453 "Check failed: character %"PRIsizet" (%x) does not conform to CONTEXTJ rule",
454 i,
455 label[i]);
456 return false;
457 }
458 }
459
460 /* 6. Check CONTEXTO characters have a rule defined */
461 /** \todo optionally we can check conformance to this rule */
462 if (idna_prop == IDNA_P_CONTEXTO) {
463 if (idna__contexto_rule(label[i]) == false) {
464 NSLOG(netsurf, INFO,
465 "Check failed: character %"PRIsizet" (%x) has no CONTEXTO rule defined",
466 i,
467 label[i]);
468 return false;
469 }
470 }
471
472 /* 7. Check characters are not UNASSIGNED */
473 if (idna_prop == IDNA_P_UNASSIGNED) {
474 NSLOG(netsurf, INFO,
475 "Check failed: character %"PRIsizet" (%x) is UNASSIGNED",
476 i,
477 label[i]);
478 return false;
479 }
480
481 /** \todo 8. (optionally) check Bidi compliance */
482 }
483
484 return true;
485 }
486
487
488 /**
489 * Verify an ACE label is valid
490 *
491 * \param label Host label to check
492 * \param len Length of label
493 * \return true if valid, false otherwise
494 */
idna__verify(const char * label,size_t len)495 static bool idna__verify(const char *label, size_t len)
496 {
497 nserror error;
498 int32_t *ucs4;
499 char *ace;
500 ssize_t ucs4_len;
501 size_t u_ucs4_len, ace_len;
502
503 /* Convert our ACE label back to UCS-4 */
504 error = idna__ace_to_ucs4(label, len, &ucs4, &u_ucs4_len);
505 if (error != NSERROR_OK) {
506 return false;
507 }
508
509 /* Perform NFC normalisation */
510 ucs4_len = utf8proc_normalize_utf32(ucs4, u_ucs4_len,
511 UTF8PROC_STABLE | UTF8PROC_COMPOSE);
512 if (ucs4_len < 0) {
513 free(ucs4);
514 return false;
515 }
516
517 /* Convert the UCS-4 label back to ACE */
518 error = idna__ucs4_to_ace(ucs4, (size_t)ucs4_len,
519 &ace, &ace_len);
520 free(ucs4);
521 if (error != NSERROR_OK) {
522 return false;
523 }
524
525 /* Check if it matches the input */
526 if ((len == ace_len) && (strncmp(label, ace, len) == 0)) {
527 free(ace);
528 return true;
529 }
530
531 NSLOG(netsurf, INFO, "Re-encoded ACE label %s does not match input",
532 ace);
533 free(ace);
534
535 return false;
536 }
537
538
539 #else /* WITH_UTF8PROC */
540
541
542 /**
543 * Convert a UTF-8 string to UCS-4
544 *
545 * \param utf8_label UTF-8 string containing host label
546 * \param len Length of host label (in bytes)
547 * \param ucs4_label Pointer to update with the output
548 * \param ucs4_len Pointer to update with the length
549 * \return NSERROR_OK on success, appropriate error otherwise
550 *
551 * If return value != NSERROR_OK, output will be left untouched.
552 */
553 static nserror
idna__utf8_to_ucs4(const char * utf8_label,size_t len,int32_t ** ucs4_label,size_t * ucs4_len)554 idna__utf8_to_ucs4(const char *utf8_label,
555 size_t len,
556 int32_t **ucs4_label,
557 size_t *ucs4_len)
558 {
559 return NSERROR_NOT_IMPLEMENTED;
560 }
561
562
563 /**
564 * Convert a UCS-4 string to UTF-8
565 *
566 * \param ucs4_label UCS-4 string containing host label
567 * \param ucs4_len Length of host label (in bytes)
568 * \param utf8_label Pointer to update with the output
569 * \param utf8_len Pointer to update with the length
570 * \return NSERROR_OK on success, appropriate error otherwise
571 *
572 * If return value != NSERROR_OK, output will be left untouched.
573 */
574 static nserror
idna__ucs4_to_utf8(const int32_t * ucs4_label,size_t ucs4_len,char ** utf8_label,size_t * utf8_len)575 idna__ucs4_to_utf8(const int32_t *ucs4_label,
576 size_t ucs4_len,
577 char **utf8_label,
578 size_t *utf8_len)
579 {
580 return NSERROR_NOT_IMPLEMENTED;
581 }
582
583
584 /**
585 * Check if a host label is valid for IDNA2008
586 *
587 * \param label Host label to check (UCS-4)
588 * \param len Length of host label (in characters/codepoints)
589 * \return true if compliant, false otherwise
590 */
idna__is_valid(int32_t * label,size_t len)591 static bool idna__is_valid(int32_t *label, size_t len)
592 {
593 return true;
594 }
595
596
597 /**
598 * Verify an ACE label is valid
599 *
600 * \param label Host label to check
601 * \param len Length of label
602 * \return true if valid, false otherwise
603 */
idna__verify(const char * label,size_t len)604 static bool idna__verify(const char *label, size_t len)
605 {
606 return true;
607 }
608
609
610 #endif /* WITH_UTF8PROC */
611
612
613 /**
614 * Find the length of a host label
615 *
616 * \param host String containing a host or FQDN
617 * \param max_length Length of host string to search (in bytes)
618 * \return Distance to next separator character or end of string
619 */
idna__host_label_length(const char * host,size_t max_length)620 static size_t idna__host_label_length(const char *host, size_t max_length)
621 {
622 const char *p = host;
623 size_t length = 0;
624
625 while (length < max_length) {
626 if ((*p == '.') || (*p == ':') || (*p == '\0')) {
627 break;
628 }
629 length++;
630 p++;
631 }
632
633 return length;
634 }
635
636
637 /**
638 * Check if a host label is LDH
639 *
640 * \param label Host label to check
641 * \param len Length of host label
642 * \return true if LDH compliant, false otherwise
643 */
idna__is_ldh(const char * label,size_t len)644 static bool idna__is_ldh(const char *label, size_t len)
645 {
646 const char *p = label;
647 size_t i = 0;
648
649 /* Check for leading or trailing hyphens */
650 if ((p[0] == '-') || (p[len - 1] == '-'))
651 return false;
652
653 /* Check for non-alphanumeric, non-hyphen characters */
654 for (i = 0; i < len; p++) {
655 i++;
656 if (*p == '-') continue;
657 if ((*p >= '0') && (*p <= '9')) continue;
658 if ((*p >= 'a') && (*p <= 'z')) continue;
659 if ((*p >= 'A') && (*p <= 'Z')) continue;
660
661 return false;
662 }
663
664 return true;
665 }
666
667
668 /**
669 * Check if a host label appears to be ACE
670 *
671 * \param label Host label to check
672 * \param len Length of host label
673 * \return true if ACE compliant, false otherwise
674 */
idna__is_ace(const char * label,size_t len)675 static bool idna__is_ace(const char *label, size_t len)
676 {
677 /* Check it is a valid DNS string */
678 if (idna__is_ldh(label, len) == false) {
679 return false;
680 }
681
682 /* Check the ACE prefix is present */
683 if ((label[0] == 'x') && (label[1] == 'n') &&
684 (label[2] == '-') && (label[3] == '-')) {
685 return true;
686 }
687
688 return false;
689 }
690
691
692 /* exported interface documented in idna.h */
693 nserror
idna_encode(const char * host,size_t len,char ** ace_host,size_t * ace_len)694 idna_encode(const char *host, size_t len, char **ace_host, size_t *ace_len)
695 {
696 nserror error;
697 int32_t *ucs4_host;
698 size_t label_len, output_len, ucs4_len, fqdn_len = 0;
699 char fqdn[256];
700 char *output, *fqdn_p = fqdn;
701
702 label_len = idna__host_label_length(host, len);
703 if (label_len == 0) {
704 return NSERROR_BAD_URL;
705 }
706
707 while (label_len != 0) {
708 if (idna__is_ldh(host, label_len) == false) {
709 /* This string is IDN or invalid */
710
711 /* Convert to Unicode */
712 error = idna__utf8_to_ucs4(host, label_len,
713 &ucs4_host, &ucs4_len);
714 if (error != NSERROR_OK) {
715 return error;
716 }
717
718 /* Check this is valid for conversion */
719 if (idna__is_valid(ucs4_host, ucs4_len) == false) {
720 free(ucs4_host);
721 return NSERROR_BAD_URL;
722 }
723
724 /* Convert to ACE */
725 error = idna__ucs4_to_ace(ucs4_host, ucs4_len,
726 &output, &output_len);
727 free(ucs4_host);
728 if (error != NSERROR_OK) {
729 return error;
730 }
731 strncpy(fqdn_p, output, output_len);
732 free(output);
733 fqdn_p += output_len;
734 fqdn_len += output_len;
735 } else {
736 /* This is already a DNS-valid ASCII string */
737 if ((idna__is_ace(host, label_len) == true) &&
738 (idna__verify(host, label_len) == false)) {
739 NSLOG(netsurf, INFO,
740 "Cannot verify ACE label %s", host);
741 return NSERROR_BAD_URL;
742 }
743 strncpy(fqdn_p, host, label_len);
744 fqdn_p += label_len;
745 fqdn_len += label_len;
746 }
747
748 *fqdn_p = '.';
749 fqdn_p++;
750 fqdn_len++;
751
752 host += label_len;
753 if ((*host == '\0') || (*host == ':')) {
754 break;
755 }
756 host++;
757 len = len - label_len - 1;
758
759 label_len = idna__host_label_length(host, len);
760 }
761
762 fqdn_p--;
763 *fqdn_p = '\0';
764 *ace_host = strdup(fqdn);
765 *ace_len = fqdn_len - 1; /* last character is NULL */
766
767 return NSERROR_OK;
768 }
769
770
771 /* exported interface documented in idna.h */
772 nserror
idna_decode(const char * ace_host,size_t ace_len,char ** host,size_t * host_len)773 idna_decode(const char *ace_host, size_t ace_len, char **host, size_t *host_len)
774 {
775 nserror error;
776 int32_t *ucs4_host;
777 size_t label_len, output_len, ucs4_len, fqdn_len = 0;
778 char fqdn[256];
779 char *output, *fqdn_p = fqdn;
780
781 label_len = idna__host_label_length(ace_host, ace_len);
782 if (label_len == 0) {
783 return NSERROR_BAD_URL;
784 }
785
786 while (label_len != 0) {
787 if (idna__is_ace(ace_host, label_len) == true) {
788 /* This string is DNS-valid and (probably) encoded */
789
790 /* Decode to Unicode */
791 error = idna__ace_to_ucs4(ace_host, label_len,
792 &ucs4_host, &ucs4_len);
793 if (error != NSERROR_OK) {
794 return error;
795 }
796
797 /* Convert to UTF-8 */
798 error = idna__ucs4_to_utf8(ucs4_host, ucs4_len,
799 &output, &output_len);
800 free(ucs4_host);
801 if (error != NSERROR_OK) {
802 return error;
803 }
804
805 memcpy(fqdn_p, output, output_len);
806 free(output);
807 fqdn_p += output_len;
808 fqdn_len += output_len;
809 } else {
810 /* Not ACE */
811 memcpy(fqdn_p, ace_host, label_len);
812 fqdn_p += label_len;
813 fqdn_len += label_len;
814 }
815
816 *fqdn_p = '.';
817 fqdn_p++;
818 fqdn_len++;
819
820 ace_host += label_len;
821 if ((*ace_host == '\0') || (*ace_host == ':')) {
822 break;
823 }
824 ace_host++;
825 ace_len = ace_len - label_len - 1;
826
827 label_len = idna__host_label_length(ace_host, ace_len);
828 }
829
830 fqdn_p--;
831 *fqdn_p = '\0';
832 *host = strdup(fqdn);
833 *host_len = fqdn_len - 1; /* last character is NULL */
834
835 return NSERROR_OK;
836 }
837