Line | Count | Source (jump to first uncovered line) |
1 | | /* |
2 | | * Copyright(c) 2014-2024 Tim Ruehsen |
3 | | * |
4 | | * Permission is hereby granted, free of charge, to any person obtaining a |
5 | | * copy of this software and associated documentation files (the "Software"), |
6 | | * to deal in the Software without restriction, including without limitation |
7 | | * the rights to use, copy, modify, merge, publish, distribute, sublicense, |
8 | | * and/or sell copies of the Software, and to permit persons to whom the |
9 | | * Software is furnished to do so, subject to the following conditions: |
10 | | * |
11 | | * The above copyright notice and this permission notice shall be included in |
12 | | * all copies or substantial portions of the Software. |
13 | | * |
14 | | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
15 | | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
16 | | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE |
17 | | * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER |
18 | | * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING |
19 | | * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER |
20 | | * DEALINGS IN THE SOFTWARE. |
21 | | * |
22 | | * This file is part of libpsl. |
23 | | * |
24 | | * Public Suffix List routines |
25 | | * |
26 | | * Changelog |
27 | | * 19.03.2014 Tim Ruehsen created from libmget/cookie.c |
28 | | * |
29 | | */ |
30 | | |
31 | | #if HAVE_CONFIG_H |
32 | | # include <config.h> |
33 | | #endif |
34 | | |
35 | | #if defined(__GNUC__) && defined(__GNUC_MINOR__) |
36 | | # define GCC_VERSION_AT_LEAST(major, minor) ((__GNUC__ > (major)) || (__GNUC__ == (major) && __GNUC_MINOR__ >= (minor))) |
37 | | #else |
38 | | # define GCC_VERSION_AT_LEAST(major, minor) 0 |
39 | | #endif |
40 | | |
41 | | #include <sys/types.h> |
42 | | #include <sys/stat.h> |
43 | | |
44 | | #if defined(_WIN32) && (defined(WITH_LIBIDN2) || defined(WITH_LIBIDN)) |
45 | | # ifndef WIN32_LEAN_AND_MEAN |
46 | | # define WIN32_LEAN_AND_MEAN |
47 | | # endif |
48 | | # include <windows.h> /* for GetACP() */ |
49 | | #endif |
50 | | |
51 | | #if defined(_MSC_VER) && ! defined(ssize_t) |
52 | | # include <basetsd.h> |
53 | | typedef SSIZE_T ssize_t; |
54 | | #endif |
55 | | |
56 | | #include <stdio.h> |
57 | | #include <stdlib.h> |
58 | | #include <string.h> |
59 | | #include <ctype.h> |
60 | | #include <time.h> |
61 | | #include <errno.h> |
62 | | #include <limits.h> /* for UINT_MAX */ |
63 | | |
64 | | #ifdef HAVE_NL_LANGINFO |
65 | | # include <langinfo.h> |
66 | | #endif |
67 | | |
68 | | #ifdef _WIN32 |
69 | | # include <malloc.h> |
70 | | #endif |
71 | | |
72 | | #ifdef WITH_LIBICU |
73 | | # include <unicode/uversion.h> |
74 | | # include <unicode/ustring.h> |
75 | | # include <unicode/uidna.h> |
76 | | # include <unicode/ucnv.h> |
77 | | #elif defined(WITH_LIBIDN2) |
78 | | # include <iconv.h> |
79 | | # include <idn2.h> |
80 | | # include <unicase.h> |
81 | | # include <unistr.h> |
82 | | #elif defined(WITH_LIBIDN) |
83 | | # include <iconv.h> |
84 | | # include <stringprep.h> |
85 | | # include <idna.h> |
86 | | # include <unicase.h> |
87 | | # include <unistr.h> |
88 | | #endif |
89 | | |
90 | | #ifdef WINICONV_CONST |
91 | | # define ICONV_CONST WINICONV_CONST |
92 | | #endif |
93 | | #ifndef ICONV_CONST |
94 | | # define ICONV_CONST |
95 | | #endif |
96 | | |
97 | | |
98 | | #include <libpsl.h> |
99 | | |
100 | | /** |
101 | | * SECTION:libpsl |
102 | | * @short_description: Public Suffix List library functions |
103 | | * @title: libpsl |
104 | | * @stability: Stable |
105 | | * @include: libpsl.h |
106 | | * |
107 | | * [Public Suffix List](https://publicsuffix.org/) library functions. |
108 | | * |
109 | | */ |
110 | | |
111 | | #define countof(a) (sizeof(a)/sizeof(*(a))) |
112 | | |
113 | 0 | #define PRIV_PSL_FLAG_EXCEPTION (1<<0) |
114 | 0 | #define PRIV_PSL_FLAG_WILDCARD (1<<1) |
115 | 0 | #define PRIV_PSL_FLAG_ICANN (1<<2) /* entry of ICANN section */ |
116 | 0 | #define PRIV_PSL_FLAG_PRIVATE (1<<3) /* entry of PRIVATE section */ |
117 | 0 | #define PRIV_PSL_FLAG_PLAIN (1<<4) /* just used for PSL syntax checking */ |
118 | | |
119 | | typedef struct { |
120 | | char |
121 | | label_buf[128]; |
122 | | const char * |
123 | | label; |
124 | | unsigned short |
125 | | length; |
126 | | unsigned char |
127 | | nlabels, /* number of labels */ |
128 | | flags; |
129 | | } psl_entry_t; |
130 | | |
131 | | /* stripped down version libmget vector routines */ |
132 | | typedef struct { |
133 | | int |
134 | | (*cmp)(const psl_entry_t **, const psl_entry_t **); /* comparison function */ |
135 | | psl_entry_t |
136 | | **entry; /* pointer to array of pointers to elements */ |
137 | | int |
138 | | max, /* allocated elements */ |
139 | | cur; /* number of elements in use */ |
140 | | } psl_vector_t; |
141 | | |
142 | | struct psl_ctx_st { |
143 | | psl_vector_t |
144 | | *suffixes; |
145 | | unsigned char |
146 | | *dafsa; |
147 | | size_t |
148 | | dafsa_size; |
149 | | int |
150 | | nsuffixes, |
151 | | nexceptions, |
152 | | nwildcards; |
153 | | unsigned |
154 | | utf8 : 1; /* 1: data contains UTF-8 + punycode encoded rules */ |
155 | | }; |
156 | | |
157 | | /* include the PSL data generated by psl-make-dafsa */ |
158 | | #ifdef ENABLE_BUILTIN |
159 | | #include "suffixes_dafsa.h" |
160 | | #else |
161 | | static const unsigned char kDafsa[] = ""; |
162 | | static time_t _psl_file_time = 0; |
163 | | static int _psl_nsuffixes = 0; |
164 | | static int _psl_nexceptions = 0; |
165 | | static int _psl_nwildcards = 0; |
166 | | static const char _psl_sha1_checksum[] = ""; |
167 | | static const char _psl_filename[] = ""; |
168 | | #endif |
169 | | |
170 | | /* references to these PSLs will result in lookups to built-in data */ |
171 | | static const psl_ctx_t |
172 | | builtin_psl; |
173 | | |
174 | | #ifdef PSL_DISTFILE |
175 | | static const char _psl_dist_filename[] = PSL_DISTFILE; |
176 | | #else |
177 | | static const char _psl_dist_filename[] = ""; |
178 | | #endif |
179 | | |
180 | | static psl_vector_t *vector_alloc(int max, int (*cmp)(const psl_entry_t **, const psl_entry_t **)) |
181 | 0 | { |
182 | 0 | psl_vector_t *v; |
183 | |
|
184 | 0 | if (!(v = calloc(1, sizeof(psl_vector_t)))) |
185 | 0 | return NULL; |
186 | | |
187 | 0 | if (!(v->entry = malloc(max * sizeof(psl_entry_t *)))) { |
188 | 0 | free(v); |
189 | 0 | return NULL; |
190 | 0 | } |
191 | | |
192 | 0 | v->max = max; |
193 | 0 | v->cmp = cmp; |
194 | 0 | return v; |
195 | 0 | } |
196 | | |
197 | | static void vector_free(psl_vector_t **v) |
198 | 0 | { |
199 | 0 | if (v && *v) { |
200 | 0 | if ((*v)->entry) { |
201 | 0 | int it; |
202 | |
|
203 | 0 | for (it = 0; it < (*v)->cur; it++) |
204 | 0 | free((*v)->entry[it]); |
205 | |
|
206 | 0 | free((*v)->entry); |
207 | 0 | } |
208 | 0 | free(*v); |
209 | 0 | } |
210 | 0 | } |
211 | | |
212 | | static psl_entry_t *vector_get(const psl_vector_t *v, int pos) |
213 | 0 | { |
214 | 0 | if (pos < 0 || !v || pos >= v->cur) return NULL; |
215 | | |
216 | 0 | return v->entry[pos]; |
217 | 0 | } |
218 | | |
219 | | /* the entries must be sorted by */ |
220 | | static int vector_find(const psl_vector_t *v, const psl_entry_t *elem) |
221 | 0 | { |
222 | 0 | if (v) { |
223 | 0 | int l, r, m; |
224 | 0 | int res; |
225 | | |
226 | | /* binary search for element (exact match) */ |
227 | 0 | for (l = 0, r = v->cur - 1; l <= r;) { |
228 | 0 | m = (l + r) / 2; |
229 | 0 | if ((res = v->cmp(&elem, (const psl_entry_t **)&(v->entry[m]))) > 0) l = m + 1; |
230 | 0 | else if (res < 0) r = m - 1; |
231 | 0 | else return m; |
232 | 0 | } |
233 | 0 | } |
234 | | |
235 | 0 | return -1; /* not found */ |
236 | 0 | } |
237 | | |
238 | | static int vector_add(psl_vector_t *v, const psl_entry_t *elem) |
239 | 0 | { |
240 | 0 | if (v) { |
241 | 0 | void *elemp; |
242 | |
|
243 | 0 | if (!(elemp = malloc(sizeof(psl_entry_t)))) |
244 | 0 | return -1; |
245 | | |
246 | 0 | memcpy(elemp, elem, sizeof(psl_entry_t)); |
247 | |
|
248 | 0 | if (v->max == v->cur) { |
249 | 0 | void *m = realloc(v->entry, (v->max *= 2) * sizeof(psl_entry_t *)); |
250 | |
|
251 | 0 | if (m) |
252 | 0 | v->entry = m; |
253 | 0 | else { |
254 | 0 | free(elemp); |
255 | 0 | return -1; |
256 | 0 | } |
257 | 0 | } |
258 | | |
259 | 0 | v->entry[v->cur++] = elemp; |
260 | 0 | return v->cur - 1; |
261 | 0 | } |
262 | | |
263 | 0 | return -1; |
264 | 0 | } |
265 | | |
266 | | static void vector_sort(psl_vector_t *v) |
267 | 0 | { |
268 | 0 | if (v && v->cmp) |
269 | 0 | qsort(v->entry, v->cur, sizeof(psl_vector_t **), (int(*)(const void *, const void *))v->cmp); |
270 | 0 | } |
271 | | |
272 | | /* by this kind of sorting, we can easily see if a domain matches or not */ |
273 | | static int suffix_compare(const psl_entry_t *s1, const psl_entry_t *s2) |
274 | 0 | { |
275 | 0 | int n; |
276 | |
|
277 | 0 | if ((n = s2->nlabels - s1->nlabels)) |
278 | 0 | return n; /* most labels first */ |
279 | | |
280 | 0 | if ((n = s1->length - s2->length)) |
281 | 0 | return n; /* shorter rules first */ |
282 | | |
283 | 0 | return strcmp(s1->label ? s1->label : s1->label_buf, s2->label ? s2->label : s2->label_buf); |
284 | 0 | } |
285 | | |
286 | | /* needed to sort array of pointers, given to qsort() */ |
287 | | static int suffix_compare_array(const psl_entry_t **s1, const psl_entry_t **s2) |
288 | 0 | { |
289 | 0 | return suffix_compare(*s1, *s2); |
290 | 0 | } |
291 | | |
292 | | static int suffix_init(psl_entry_t *suffix, const char *rule, size_t length) |
293 | 0 | { |
294 | 0 | const char *src; |
295 | 0 | char *dst; |
296 | |
|
297 | 0 | suffix->label = suffix->label_buf; |
298 | |
|
299 | 0 | if (length >= sizeof(suffix->label_buf) - 1) { |
300 | 0 | suffix->nlabels = 0; |
301 | | /* fprintf(stderr, "Suffix rule too long (%zd, ignored): %s\n", length, rule); */ |
302 | 0 | return -1; |
303 | 0 | } |
304 | | |
305 | 0 | suffix->length = (unsigned char)length; |
306 | |
|
307 | 0 | suffix->nlabels = 1; |
308 | |
|
309 | 0 | for (dst = suffix->label_buf, src = rule; *src;) { |
310 | 0 | if (*src == '.') |
311 | 0 | suffix->nlabels++; |
312 | 0 | *dst++ = *src++; |
313 | 0 | } |
314 | 0 | *dst = 0; |
315 | |
|
316 | 0 | return 0; |
317 | 0 | } |
318 | | |
319 | | static char *psl_strdup(const char *s) |
320 | 0 | { |
321 | 0 | char *p = malloc(strlen(s) + 1); |
322 | 0 | if (!p) |
323 | 0 | return NULL; |
324 | 0 | return strcpy(p, s); |
325 | 0 | } |
326 | | |
327 | | #if !defined(WITH_LIBIDN) && !defined(WITH_LIBIDN2) && !defined(WITH_LIBICU) |
328 | | /* |
329 | | * When configured without runtime IDNA support (./configure --disable-runtime), we need a pure ASCII |
330 | | * representation of non-ASCII characters in labels as found in UTF-8 domain names. |
331 | | * This is because the current DAFSA format used may only hold character values [21..127]. |
332 | | * |
333 | | Code copied from http://www.nicemice.net/idn/punycode-spec.gz on |
334 | | 2011-01-04 with SHA-1 a966a8017f6be579d74a50a226accc7607c40133 |
335 | | labeled punycode-spec 1.0.3 (2006-Mar-24-Thu). It is modified for |
336 | | libpsl by Tim Rühsen. License on the original code: |
337 | | |
338 | | punycode-spec 1.0.3 (2006-Mar-23-Thu) |
339 | | http://www.nicemice.net/idn/ |
340 | | Adam M. Costello |
341 | | http://www.nicemice.net/amc/ |
342 | | |
343 | | B. Disclaimer and license |
344 | | |
345 | | Regarding this entire document or any portion of it (including |
346 | | the pseudocode and C code), the author makes no guarantees and |
347 | | is not responsible for any damage resulting from its use. The |
348 | | author grants irrevocable permission to anyone to use, modify, |
349 | | and distribute it in any way that does not diminish the rights |
350 | | of anyone else to use, modify, and distribute it, provided that |
351 | | redistributed derivative works do not contain misleading author or |
352 | | version information. Derivative works need not be licensed under |
353 | | similar terms. |
354 | | |
355 | | C. Punycode sample implementation |
356 | | |
357 | | punycode-sample.c 2.0.0 (2004-Mar-21-Sun) |
358 | | http://www.nicemice.net/idn/ |
359 | | Adam M. Costello |
360 | | http://www.nicemice.net/amc/ |
361 | | |
362 | | This is ANSI C code (C89) implementing Punycode 1.0.x. |
363 | | */ |
364 | | enum punycode_status { |
365 | | punycode_success = 0, |
366 | | punycode_bad_input = 1, /* Input is invalid. */ |
367 | | punycode_big_output = 2, /* Output would exceed the space provided. */ |
368 | | punycode_overflow = 3 /* Wider integers needed to process input. */ |
369 | | }; |
370 | | |
371 | | #ifdef PUNYCODE_UINT |
372 | | typedef PUNYCODE_UINT punycode_uint; |
373 | | #elif UINT_MAX >= (1 << 26) - 1 |
374 | | typedef unsigned int punycode_uint; |
375 | | #else |
376 | | typedef unsigned long punycode_uint; |
377 | | #endif |
378 | | |
379 | | /*** Bootstring parameters for Punycode ***/ |
380 | | enum { |
381 | | base = 36, tmin = 1, tmax = 26, skew = 38, damp = 700, |
382 | | initial_bias = 72, initial_n = 0x80, delimiter = 0x2D |
383 | | }; |
384 | | |
385 | | static char encode_digit(punycode_uint d) |
386 | | { |
387 | | return d + 22 + 75 * (d < 26); |
388 | | /* 0..25 map to ASCII a..z or A..Z */ |
389 | | /* 26..35 map to ASCII 0..9 */ |
390 | | } |
391 | | #define flagged(bcp) ((punycode_uint)(bcp) - 65 < 26) |
392 | | static const punycode_uint maxint = -1; |
393 | | |
394 | | static punycode_uint adapt(punycode_uint delta, punycode_uint numpoints, int firsttime) |
395 | | { |
396 | | punycode_uint k; |
397 | | |
398 | | delta = firsttime ? delta / damp : delta >> 1; |
399 | | /* delta >> 1 is a faster way of doing delta / 2 */ |
400 | | delta += delta / numpoints; |
401 | | |
402 | | for (k = 0; delta > ((base - tmin) * tmax) / 2; k += base) { |
403 | | delta /= base - tmin; |
404 | | } |
405 | | |
406 | | return k + (base - tmin + 1) * delta / (delta + skew); |
407 | | } |
408 | | |
409 | | static enum punycode_status punycode_encode( |
410 | | size_t input_length_orig, |
411 | | const punycode_uint input[], |
412 | | size_t *output_length, |
413 | | char output[]) |
414 | | { |
415 | | punycode_uint input_length, n, delta, h, b, bias, j, m, q, k, t; |
416 | | size_t out, max_out; |
417 | | |
418 | | /* The Punycode spec assumes that the input length is the same type */ |
419 | | /* of integer as a code point, so we need to convert the size_t to */ |
420 | | /* a punycode_uint, which could overflow. */ |
421 | | |
422 | | if (input_length_orig > maxint) |
423 | | return punycode_overflow; |
424 | | |
425 | | input_length = (punycode_uint) input_length_orig; |
426 | | |
427 | | /* Initialize the state: */ |
428 | | |
429 | | n = initial_n; |
430 | | delta = 0; |
431 | | out = 0; |
432 | | max_out = *output_length; |
433 | | bias = initial_bias; |
434 | | |
435 | | /* Handle the basic code points: */ |
436 | | for (j = 0; j < input_length; ++j) { |
437 | | if (input[j] < 0x80) { |
438 | | if (max_out - out < 2) |
439 | | return punycode_big_output; |
440 | | output[out++] = (char) input[j]; |
441 | | } |
442 | | /* else if (input[j] < n) return punycode_bad_input; */ |
443 | | /* (not needed for Punycode with unsigned code points) */ |
444 | | } |
445 | | |
446 | | h = b = (punycode_uint) out; |
447 | | /* cannot overflow because out <= input_length <= maxint */ |
448 | | |
449 | | /* h is the number of code points that have been handled, b is the */ |
450 | | /* number of basic code points, and out is the number of ASCII code */ |
451 | | /* points that have been output. */ |
452 | | |
453 | | if (b > 0) |
454 | | output[out++] = delimiter; |
455 | | |
456 | | /* Main encoding loop: */ |
457 | | |
458 | | while (h < input_length) { |
459 | | /* All non-basic code points < n have been */ |
460 | | /* handled already. Find the next larger one: */ |
461 | | |
462 | | for (m = maxint, j = 0; j < input_length; ++j) { |
463 | | /* if (basic(input[j])) continue; */ |
464 | | /* (not needed for Punycode) */ |
465 | | if (input[j] >= n && input[j] < m) |
466 | | m = input[j]; |
467 | | } |
468 | | |
469 | | /* Increase delta enough to advance the decoder's */ |
470 | | /* <n,i> state to <m,0>, but guard against overflow: */ |
471 | | |
472 | | if (m - n > (maxint - delta) / (h + 1)) |
473 | | return punycode_overflow; |
474 | | delta += (m - n) * (h + 1); |
475 | | n = m; |
476 | | |
477 | | for (j = 0; j < input_length; ++j) { |
478 | | /* Punycode does not need to check whether input[j] is basic: */ |
479 | | if (input[j] < n /* || basic(input[j]) */) { |
480 | | if (++delta == 0) |
481 | | return punycode_overflow; |
482 | | } |
483 | | |
484 | | if (input[j] == n) { |
485 | | /* Represent delta as a generalized variable-length integer: */ |
486 | | |
487 | | for (q = delta, k = base;; k += base) { |
488 | | if (out >= max_out) |
489 | | return punycode_big_output; |
490 | | t = k <= bias /* + tmin */ ? tmin : /* +tmin not needed */ |
491 | | k >= bias + tmax ? tmax : k - bias; |
492 | | if (q < t) |
493 | | break; |
494 | | output[out++] = encode_digit(t + (q - t) % (base - t)); |
495 | | q = (q - t) / (base - t); |
496 | | } |
497 | | |
498 | | output[out++] = encode_digit(q); |
499 | | bias = adapt(delta, h + 1, h == b); |
500 | | delta = 0; |
501 | | ++h; |
502 | | } |
503 | | } |
504 | | |
505 | | ++delta, ++n; |
506 | | } |
507 | | |
508 | | *output_length = out; |
509 | | return punycode_success; |
510 | | } |
511 | | |
512 | | static ssize_t utf8_to_utf32(const char *in, size_t inlen, punycode_uint *out, size_t outlen) |
513 | | { |
514 | | size_t n = 0; |
515 | | const unsigned char *s = (void *)in; |
516 | | const unsigned char *e = (void *)(in + inlen); |
517 | | |
518 | | if (!outlen) |
519 | | return -1; |
520 | | |
521 | | outlen--; |
522 | | |
523 | | while (n < outlen) { |
524 | | size_t inleft = e - s; |
525 | | |
526 | | if (inleft >= 1 && (*s & 0x80) == 0) { /* 0xxxxxxx ASCII char */ |
527 | | out[n++] = *s; |
528 | | s++; |
529 | | } else if (inleft >= 2 && (*s & 0xE0) == 0xC0) /* 110xxxxx 10xxxxxx */ { |
530 | | if ((s[1] & 0xC0) != 0x80) |
531 | | return -1; |
532 | | out[n++] = ((*s & 0x1F) << 6) | (s[1] & 0x3F); |
533 | | s += 2; |
534 | | } else if (inleft >= 3 && (*s & 0xF0) == 0xE0) /* 1110xxxx 10xxxxxx 10xxxxxx */ { |
535 | | if ((s[1] & 0xC0) != 0x80 || (s[2] & 0xC0) != 0x80) |
536 | | return -1; |
537 | | out[n++] = ((*s & 0x0F) << 12) | ((s[1] & 0x3F) << 6) | (s[2] & 0x3F); |
538 | | s += 3; |
539 | | } else if (inleft >= 4 && (*s & 0xF8) == 0xF0) /* 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */ { |
540 | | if ((s[1] & 0xC0) != 0x80 || (s[2] & 0xC0) != 0x80 || (s[3] & 0xC0) != 0x80) |
541 | | return -1; |
542 | | out[n++] = ((*s & 0x07) << 18) | ((s[1] & 0x3F) << 12) | ((s[1] & 0x3F) << 6) | (s[2] & 0x3F); |
543 | | s += 4; |
544 | | } else if (!inleft) { |
545 | | break; |
546 | | } else |
547 | | return -1; |
548 | | } |
549 | | |
550 | | return n; |
551 | | } |
552 | | |
553 | | static int mem_is_ascii(const char *s, size_t n) |
554 | | { |
555 | | for (; n; n--) /* 'while(n--)' generates unsigned integer overflow on n = 0 */ |
556 | | if (*((unsigned char *)s++) >= 128) |
557 | | return 0; |
558 | | |
559 | | return 1; |
560 | | } |
561 | | |
562 | | static int domain_to_punycode(const char *domain, char *out, size_t outsize) |
563 | | { |
564 | | size_t outlen = 0, labellen; |
565 | | punycode_uint input[256]; |
566 | | const char *label, *e; |
567 | | |
568 | | for (e = label = domain; e;) { |
569 | | e = strchr(label, '.'); |
570 | | labellen = e ? (size_t) (e - label) : strlen(label); |
571 | | |
572 | | if (mem_is_ascii(label, labellen)) { |
573 | | if (outlen + labellen + (e != NULL) >= outsize) |
574 | | return 1; |
575 | | |
576 | | memcpy(out + outlen, label, labellen); |
577 | | outlen += labellen; |
578 | | } else { |
579 | | ssize_t inputlen = 0; |
580 | | |
581 | | if (outlen + labellen + (e != NULL) + 4 >= outsize) |
582 | | return 1; |
583 | | |
584 | | if ((inputlen = utf8_to_utf32(label, labellen, input, countof(input))) < 0) |
585 | | return 1; |
586 | | |
587 | | memcpy(out + outlen, "xn--", 4); |
588 | | outlen += 4; |
589 | | |
590 | | labellen = outsize - outlen - (e != NULL) - 1; // -1 to leave space for the trailing \0 |
591 | | if (punycode_encode(inputlen, input, &labellen, out + outlen)) |
592 | | return 1; |
593 | | outlen += labellen; |
594 | | } |
595 | | |
596 | | if (e) { |
597 | | label = e + 1; |
598 | | out[outlen++] = '.'; |
599 | | } |
600 | | out[outlen] = 0; |
601 | | } |
602 | | |
603 | | return 0; |
604 | | } |
605 | | #endif |
606 | | |
607 | | static int isspace_ascii(const char c) |
608 | 0 | { |
609 | 0 | return c == ' ' || c == '\t' || c == '\r' || c == '\n'; |
610 | 0 | } |
611 | | |
612 | | static int str_is_ascii(const char *s) |
613 | 0 | { |
614 | 0 | while (*s && *((unsigned char *)s) < 128) s++; |
615 | |
|
616 | 0 | return !*s; |
617 | 0 | } |
618 | | |
619 | | #if defined(WITH_LIBIDN) |
620 | | /* |
621 | | * Work around a libidn <= 1.30 vulnerability. |
622 | | * |
623 | | * The function checks for a valid UTF-8 character sequence before |
624 | | * passing it to idna_to_ascii_8z(). |
625 | | * |
626 | | * [1] https://lists.gnu.org/archive/html/help-libidn/2015-05/msg00002.html |
627 | | * [2] https://lists.gnu.org/archive/html/bug-wget/2015-06/msg00002.html |
628 | | * [3] https://curl.haxx.se/mail/lib-2015-06/0143.html |
629 | | */ |
630 | | static int utf8_is_valid(const char *utf8) |
631 | | { |
632 | | const unsigned char *s = (const unsigned char *) utf8; |
633 | | |
634 | | while (*s) { |
635 | | if ((*s & 0x80) == 0) /* 0xxxxxxx ASCII char */ |
636 | | s++; |
637 | | else if ((*s & 0xE0) == 0xC0) /* 110xxxxx 10xxxxxx */ { |
638 | | if ((s[1] & 0xC0) != 0x80) |
639 | | return 0; |
640 | | s += 2; |
641 | | } else if ((*s & 0xF0) == 0xE0) /* 1110xxxx 10xxxxxx 10xxxxxx */ { |
642 | | if ((s[1] & 0xC0) != 0x80 || (s[2] & 0xC0) != 0x80) |
643 | | return 0; |
644 | | s += 3; |
645 | | } else if ((*s & 0xF8) == 0xF0) /* 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */ { |
646 | | if ((s[1] & 0xC0) != 0x80 || (s[2] & 0xC0) != 0x80 || (s[3] & 0xC0) != 0x80) |
647 | | return 0; |
648 | | s += 4; |
649 | | } else |
650 | | return 0; |
651 | | } |
652 | | |
653 | | return 1; |
654 | | } |
655 | | #endif |
656 | | |
657 | | typedef void *psl_idna_t; |
658 | | |
659 | | static psl_idna_t *psl_idna_open(void) |
660 | 0 | { |
661 | | #if defined(WITH_LIBICU) |
662 | | UErrorCode status = 0; |
663 | | return (void *)uidna_openUTS46(UIDNA_USE_STD3_RULES | UIDNA_NONTRANSITIONAL_TO_ASCII, &status); |
664 | | #endif |
665 | 0 | return NULL; |
666 | 0 | } |
667 | | |
668 | | static void psl_idna_close(psl_idna_t *idna) |
669 | 0 | { |
670 | 0 | (void) idna; |
671 | |
|
672 | | #if defined(WITH_LIBICU) |
673 | | if (idna) |
674 | | uidna_close((UIDNA *)idna); |
675 | | #endif |
676 | 0 | } |
677 | | |
678 | | static int psl_idna_toASCII(psl_idna_t *idna, const char *utf8, char **ascii) |
679 | 0 | { |
680 | 0 | int ret = -1; |
681 | |
|
682 | | #if defined(WITH_LIBICU) |
683 | | (void) idna; |
684 | | |
685 | | /* IDNA2008 UTS#46 punycode conversion */ |
686 | | if (idna) { |
687 | | char lookupname_buf[128] = "", *lookupname = lookupname_buf; |
688 | | UErrorCode status = 0; |
689 | | UIDNAInfo info = UIDNA_INFO_INITIALIZER; |
690 | | UChar utf16_dst[128], utf16_src_buf[128]; |
691 | | UChar *utf16_src = utf16_src_buf; |
692 | | int32_t utf16_src_length, bytes_written; |
693 | | int32_t utf16_dst_length; |
694 | | |
695 | | u_strFromUTF8(utf16_src, countof(utf16_src_buf), &utf16_src_length, utf8, -1, &status); |
696 | | if (!U_SUCCESS(status)) goto cleanup; /* UTF-8 to UTF-16 conversion failed */ |
697 | | |
698 | | if (utf16_src_length >= (int) countof(utf16_src_buf)) { |
699 | | utf16_src = malloc((utf16_src_length + 1) * sizeof(UChar)); |
700 | | if (!utf16_src) goto cleanup; |
701 | | |
702 | | u_strFromUTF8(utf16_src, utf16_src_length, NULL, utf8, -1, &status); |
703 | | if (!U_SUCCESS(status)) goto cleanup; /* UTF-8 to UTF-16 conversion failed */ |
704 | | |
705 | | utf16_src[utf16_src_length] = 0; /* u_strFromUTF8() doesn't 0-terminate if dest is filled up */ |
706 | | } |
707 | | |
708 | | utf16_dst_length = uidna_nameToASCII((UIDNA *)idna, utf16_src, utf16_src_length, utf16_dst, countof(utf16_dst), &info, &status); |
709 | | if (!U_SUCCESS(status)) goto cleanup; /* to ASCII conversion failed */ |
710 | | |
711 | | u_strToUTF8(lookupname, sizeof(lookupname_buf), &bytes_written, utf16_dst, utf16_dst_length, &status); |
712 | | if (!U_SUCCESS(status)) goto cleanup; /* UTF-16 to UTF-8 conversion failed */ |
713 | | |
714 | | if (bytes_written >= (int) sizeof(lookupname_buf)) { |
715 | | lookupname = malloc(bytes_written + 1); |
716 | | if (!lookupname) goto cleanup; |
717 | | |
718 | | u_strToUTF8(lookupname, bytes_written, NULL, utf16_dst, utf16_dst_length, &status); |
719 | | if (!U_SUCCESS(status)) goto cleanup; /* UTF-16 to UTF-8 conversion failed */ |
720 | | |
721 | | lookupname[bytes_written] = 0; /* u_strToUTF8() doesn't 0-terminate if dest is filled up */ |
722 | | } else { |
723 | | if (!(lookupname = psl_strdup(lookupname))) |
724 | | goto cleanup; |
725 | | } |
726 | | |
727 | | if (ascii) { |
728 | | *ascii = lookupname; |
729 | | lookupname = NULL; |
730 | | } |
731 | | |
732 | | ret = 0; |
733 | | |
734 | | cleanup: |
735 | | if (lookupname != lookupname_buf) |
736 | | free(lookupname); |
737 | | if (utf16_src != utf16_src_buf) |
738 | | free(utf16_src); |
739 | | } |
740 | | #elif defined(WITH_LIBIDN2) |
741 | | #if IDN2_VERSION_NUMBER >= 0x00140000 |
742 | 0 | int rc; |
743 | |
|
744 | 0 | (void) idna; |
745 | | |
746 | | /* IDN2_TRANSITIONAL automatically converts to lowercase |
747 | | * IDN2_NFC_INPUT converts to NFC before toASCII conversion |
748 | | * Since IDN2_TRANSITIONAL implicitly does NFC conversion, we don't need |
749 | | * the additional IDN2_NFC_INPUT. But just for the unlikely case that the linked |
750 | | * library is not matching the headers when building and it doesn't support TR46, |
751 | | * we provide IDN2_NFC_INPUT. */ |
752 | |
|
753 | 0 | if ((rc = idn2_lookup_u8((uint8_t *)utf8, (uint8_t **)ascii, IDN2_NFC_INPUT | IDN2_NONTRANSITIONAL)) == IDN2_OK) |
754 | 0 | ret = 0; |
755 | | /* else |
756 | | fprintf(stderr, "toASCII(%s) failed (%d): %s\n", lower, rc, idn2_strerror(rc)); */ |
757 | | #else |
758 | | int rc; |
759 | | uint8_t *lower; |
760 | | size_t len = u8_strlen((uint8_t *)utf8) + 1; |
761 | | |
762 | | /* we need a conversion to lowercase */ |
763 | | if (!(lower = u8_tolower((uint8_t *)utf8, len, 0, UNINORM_NFKC, NULL, &len))) { |
764 | | /* fprintf(stderr, "u8_tolower(%s) failed (%d)\n", utf8, errno); */ |
765 | | return -1; |
766 | | } |
767 | | |
768 | | if ((rc = idn2_lookup_u8(lower, (uint8_t **)ascii, 0)) == IDN2_OK) { |
769 | | ret = 0; |
770 | | } /* else |
771 | | fprintf(stderr, "toASCII(%s) failed (%d): %s\n", lower, rc, idn2_strerror(rc)); */ |
772 | | |
773 | | free(lower); |
774 | | #endif |
775 | | #elif defined(WITH_LIBIDN) |
776 | | int rc; |
777 | | |
778 | | (void) idna; |
779 | | |
780 | | if (!utf8_is_valid(utf8)) { |
781 | | /* fprintf(stderr, "Invalid UTF-8 sequence not converted: '%s'\n", utf8); */ |
782 | | return -1; |
783 | | } |
784 | | |
785 | | /* idna_to_ascii_8z() automatically converts UTF-8 to lowercase */ |
786 | | |
787 | | if ((rc = idna_to_ascii_8z(utf8, ascii, IDNA_USE_STD3_ASCII_RULES)) == IDNA_SUCCESS) { |
788 | | ret = 0; |
789 | | } /* else |
790 | | fprintf(stderr, "toASCII failed (%d): %s\n", rc, idna_strerror(rc)); */ |
791 | | #else |
792 | | char lookupname[128]; |
793 | | |
794 | | (void) idna; |
795 | | |
796 | | if (domain_to_punycode(utf8, lookupname, sizeof(lookupname)) == 0) { |
797 | | if (ascii) |
798 | | if ((*ascii = psl_strdup(lookupname))) |
799 | | ret = 0; |
800 | | } |
801 | | #endif |
802 | |
|
803 | 0 | return ret; |
804 | 0 | } |
805 | | |
806 | | static void add_punycode_if_needed(psl_idna_t *idna, psl_vector_t *v, psl_entry_t *e) |
807 | 0 | { |
808 | 0 | char *lookupname; |
809 | |
|
810 | 0 | if (str_is_ascii(e->label_buf)) |
811 | 0 | return; |
812 | | |
813 | 0 | if (psl_idna_toASCII(idna, e->label_buf, &lookupname) == 0) { |
814 | 0 | if (strcmp(e->label_buf, lookupname)) { |
815 | 0 | psl_entry_t suffix, *suffixp; |
816 | | |
817 | | /* fprintf(stderr, "toASCII '%s' -> '%s'\n", e->label_buf, lookupname); */ |
818 | 0 | if (suffix_init(&suffix, lookupname, strlen(lookupname)) == 0) { |
819 | 0 | suffix.flags = e->flags; |
820 | 0 | if ((suffixp = vector_get(v, vector_add(v, &suffix)))) |
821 | 0 | suffixp->label = suffixp->label_buf; /* set label to changed address */ |
822 | 0 | } |
823 | 0 | } /* else ignore */ |
824 | |
|
825 | 0 | free(lookupname); |
826 | 0 | } |
827 | 0 | } |
828 | | |
829 | | /* prototypes */ |
830 | | int LookupStringInFixedSet(const unsigned char* graph, size_t length, const char* key, size_t key_length); |
831 | | int GetUtfMode(const unsigned char *graph, size_t length); |
832 | | |
833 | | static int is_public_suffix(const psl_ctx_t *psl, const char *domain, int type) |
834 | 0 | { |
835 | 0 | psl_entry_t suffix; |
836 | 0 | const char *p; |
837 | 0 | char *punycode = NULL; |
838 | 0 | int need_conversion = 0; |
839 | | |
840 | | /* this function should be called without leading dots, just make sure */ |
841 | 0 | if (*domain == '.') |
842 | 0 | domain++; |
843 | |
|
844 | 0 | suffix.nlabels = 1; |
845 | |
|
846 | 0 | for (p = domain; *p; p++) { |
847 | 0 | if (*p == '.') { |
848 | 0 | if (suffix.nlabels == 255) /* weird input, avoid 8bit overflow */ |
849 | 0 | return 0; |
850 | 0 | suffix.nlabels++; |
851 | 0 | } |
852 | 0 | else if (*((unsigned char *)p) >= 128) |
853 | 0 | need_conversion = 1; /* in case domain is non-ascii we need a toASCII conversion */ |
854 | 0 | } |
855 | | |
856 | 0 | if (suffix.nlabels == 1) { |
857 | | /* TLD, this is the prevailing '*' match. If type excludes the '*' rule, continue. |
858 | | */ |
859 | 0 | if (!(type & PSL_TYPE_NO_STAR_RULE)) |
860 | 0 | return 1; |
861 | 0 | } |
862 | | |
863 | 0 | type &= ~PSL_TYPE_NO_STAR_RULE; |
864 | |
|
865 | 0 | if (psl->utf8 || psl == &builtin_psl) |
866 | 0 | need_conversion = 0; |
867 | |
|
868 | 0 | if (need_conversion) { |
869 | 0 | psl_idna_t *idna = psl_idna_open(); |
870 | |
|
871 | 0 | if (psl_idna_toASCII(idna, domain, &punycode) == 0) { |
872 | 0 | suffix.label = punycode; |
873 | 0 | suffix.length = strlen(punycode); |
874 | 0 | } else { |
875 | | /* fallback */ |
876 | |
|
877 | 0 | suffix.label = domain; |
878 | 0 | suffix.length = p - suffix.label; |
879 | 0 | } |
880 | |
|
881 | 0 | psl_idna_close(idna); |
882 | 0 | } else { |
883 | 0 | suffix.label = domain; |
884 | 0 | suffix.length = p - suffix.label; |
885 | 0 | } |
886 | |
|
887 | 0 | if (psl == &builtin_psl || psl->dafsa) { |
888 | 0 | size_t dafsa_size = psl == &builtin_psl ? sizeof(kDafsa) : psl->dafsa_size; |
889 | 0 | const unsigned char *dafsa = psl == &builtin_psl ? kDafsa : psl->dafsa; |
890 | 0 | int rc = LookupStringInFixedSet(dafsa, dafsa_size, suffix.label, suffix.length); |
891 | 0 | if (rc != -1) { |
892 | | /* check for correct rule type */ |
893 | 0 | if (type == PSL_TYPE_ICANN && !(rc & PRIV_PSL_FLAG_ICANN)) |
894 | 0 | goto suffix_no; |
895 | 0 | else if (type == PSL_TYPE_PRIVATE && !(rc & PRIV_PSL_FLAG_PRIVATE)) |
896 | 0 | goto suffix_no; |
897 | | |
898 | 0 | if (rc & PRIV_PSL_FLAG_EXCEPTION) |
899 | 0 | goto suffix_no; |
900 | | |
901 | | /* wildcard *.foo.bar implicitly make foo.bar a public suffix */ |
902 | | /* definitely a match, no matter if the found rule is a wildcard or not */ |
903 | 0 | goto suffix_yes; |
904 | 0 | } |
905 | 0 | if ((suffix.label = strchr(suffix.label, '.'))) { |
906 | 0 | suffix.label++; |
907 | 0 | suffix.length = strlen(suffix.label); |
908 | 0 | suffix.nlabels--; |
909 | |
|
910 | 0 | rc = LookupStringInFixedSet(dafsa, dafsa_size, suffix.label, suffix.length); |
911 | 0 | if (rc != -1) { |
912 | | /* check for correct rule type */ |
913 | 0 | if (type == PSL_TYPE_ICANN && !(rc & PRIV_PSL_FLAG_ICANN)) |
914 | 0 | goto suffix_no; |
915 | 0 | else if (type == PSL_TYPE_PRIVATE && !(rc & PRIV_PSL_FLAG_PRIVATE)) |
916 | 0 | goto suffix_no; |
917 | | |
918 | 0 | if (rc & PRIV_PSL_FLAG_WILDCARD) |
919 | 0 | goto suffix_yes; |
920 | 0 | } |
921 | 0 | } |
922 | 0 | } else { |
923 | 0 | psl_entry_t *rule = vector_get(psl->suffixes, 0); |
924 | |
|
925 | 0 | if (!rule || rule->nlabels < suffix.nlabels - 1) |
926 | 0 | goto suffix_no; |
927 | | |
928 | 0 | rule = vector_get(psl->suffixes, vector_find(psl->suffixes, &suffix)); |
929 | |
|
930 | 0 | if (rule) { |
931 | | /* check for correct rule type */ |
932 | 0 | if (type == PSL_TYPE_ICANN && !(rule->flags & PRIV_PSL_FLAG_ICANN)) |
933 | 0 | goto suffix_no; |
934 | 0 | else if (type == PSL_TYPE_PRIVATE && !(rule->flags & PRIV_PSL_FLAG_PRIVATE)) |
935 | 0 | goto suffix_no; |
936 | | |
937 | 0 | if (rule->flags & PRIV_PSL_FLAG_EXCEPTION) |
938 | 0 | goto suffix_no; |
939 | | |
940 | | /* wildcard *.foo.bar implicitly make foo.bar a public suffix */ |
941 | | /* definitely a match, no matter if the found rule is a wildcard or not */ |
942 | 0 | goto suffix_yes; |
943 | 0 | } |
944 | | |
945 | 0 | if ((suffix.label = strchr(suffix.label, '.'))) { |
946 | 0 | suffix.label++; |
947 | 0 | suffix.length = strlen(suffix.label); |
948 | 0 | suffix.nlabels--; |
949 | |
|
950 | 0 | rule = vector_get(psl->suffixes, vector_find(psl->suffixes, &suffix)); |
951 | |
|
952 | 0 | if (rule) { |
953 | | /* check for correct rule type */ |
954 | 0 | if (type == PSL_TYPE_ICANN && !(rule->flags & PRIV_PSL_FLAG_ICANN)) |
955 | 0 | goto suffix_no; |
956 | 0 | else if (type == PSL_TYPE_PRIVATE && !(rule->flags & PRIV_PSL_FLAG_PRIVATE)) |
957 | 0 | goto suffix_no; |
958 | | |
959 | 0 | if (rule->flags & PRIV_PSL_FLAG_WILDCARD) |
960 | 0 | goto suffix_yes; |
961 | 0 | } |
962 | 0 | } |
963 | 0 | } |
964 | | |
965 | 0 | suffix_no: |
966 | 0 | if (punycode) |
967 | 0 | free(punycode); |
968 | 0 | return 0; |
969 | | |
970 | 0 | suffix_yes: |
971 | 0 | if (punycode) |
972 | 0 | free(punycode); |
973 | 0 | return 1; |
974 | 0 | } |
975 | | |
976 | | /** |
977 | | * psl_is_public_suffix: |
978 | | * @psl: PSL context |
979 | | * @domain: Domain string |
980 | | * |
981 | | * This function checks if @domain is a public suffix by the means of the |
982 | | * [Mozilla Public Suffix List](https://publicsuffix.org). |
983 | | * |
984 | | * For cookie domain checking see psl_is_cookie_domain_acceptable(). |
985 | | * |
986 | | * International @domain names have to be either in UTF-8 (lowercase + NFKC) or in ASCII/ACE format (punycode). |
987 | | * Other encodings likely result in incorrect return values. |
988 | | * Use helper function psl_str_to_utf8lower() for normalization @domain. |
989 | | * |
990 | | * @psl is a context returned by either psl_load_file(), psl_load_fp() or |
991 | | * psl_builtin(). |
992 | | * |
993 | | * Returns: 1 if domain is a public suffix, 0 if not. |
994 | | * |
995 | | * Since: 0.1 |
996 | | */ |
997 | | int psl_is_public_suffix(const psl_ctx_t *psl, const char *domain) |
998 | 0 | { |
999 | 0 | if (!psl || !domain) |
1000 | 0 | return 1; |
1001 | | |
1002 | 0 | return is_public_suffix(psl, domain, PSL_TYPE_ANY); |
1003 | 0 | } |
1004 | | |
1005 | | /** |
1006 | | * psl_is_public_suffix2: |
1007 | | * @psl: PSL context |
1008 | | * @domain: Domain string |
1009 | | * @type: Domain type |
1010 | | * |
1011 | | * This function checks if @domain is a public suffix by the means of the |
1012 | | * [Mozilla Public Suffix List](https://publicsuffix.org). |
1013 | | * |
1014 | | * @type specifies the PSL section where to perform the lookup. Valid values are |
1015 | | * %PSL_TYPE_PRIVATE, %PSL_TYPE_ICANN, %PSL_TYPE_NO_STAR_RULE, and %PSL_TYPE_ANY. |
1016 | | * |
1017 | | * %PSL_TYPE_NO_STAR_RULE switches of the 'prevailing star rule' (see |
1018 | | * [List](https://publicsuffix.org/list) under 'Algorithm' 2.). |
1019 | | * Applying the flag means that TLDs not explicitly listed in the PSL are *not* treated as public suffixes. |
1020 | | * |
1021 | | * International @domain names have to be either in UTF-8 (lowercase + NFKC) or in ASCII/ACE format (punycode). |
1022 | | * Other encodings likely result in incorrect return values. |
1023 | | * Use helper function psl_str_to_utf8lower() for normalization @domain. |
1024 | | * |
1025 | | * @psl is a context returned by either psl_load_file(), psl_load_fp() or |
1026 | | * psl_builtin(). |
1027 | | * |
1028 | | * Returns: 1 if domain is a public suffix, 0 if not. |
1029 | | * |
1030 | | * Since: 0.1 |
1031 | | */ |
1032 | | int psl_is_public_suffix2(const psl_ctx_t *psl, const char *domain, int type) |
1033 | 0 | { |
1034 | 0 | if (!psl || !domain) |
1035 | 0 | return 1; |
1036 | | |
1037 | 0 | return is_public_suffix(psl, domain, type); |
1038 | 0 | } |
1039 | | |
1040 | | /** |
1041 | | * psl_unregistrable_domain: |
1042 | | * @psl: PSL context |
1043 | | * @domain: Domain string |
1044 | | * |
1045 | | * This function finds the longest public suffix part of @domain by the means |
1046 | | * of the [Mozilla Public Suffix List](https://publicsuffix.org). |
1047 | | * |
1048 | | * International @domain names have to be either in UTF-8 (lowercase + NFKC) or in ASCII/ACE format (punycode). |
1049 | | * Other encodings likely result in incorrect return values. |
1050 | | * Use helper function psl_str_to_utf8lower() for normalization @domain. |
1051 | | * |
1052 | | * @psl is a context returned by either psl_load_file(), psl_load_fp() or |
1053 | | * psl_builtin(). |
1054 | | * |
1055 | | * Returns: Pointer to longest public suffix part of @domain or %NULL if @domain |
1056 | | * does not contain a public suffix (or if @psl is %NULL). |
1057 | | * |
1058 | | * Since: 0.1 |
1059 | | */ |
1060 | | const char *psl_unregistrable_domain(const psl_ctx_t *psl, const char *domain) |
1061 | 0 | { |
1062 | 0 | int nlabels = 0; |
1063 | 0 | const char *p; |
1064 | |
|
1065 | 0 | if (!psl || !domain) |
1066 | 0 | return NULL; |
1067 | | |
1068 | | /* |
1069 | | * In the main loop we introduce a O(N^2) behavior to avoid code duplication. |
1070 | | * To avoid nasty CPU hogging, we limit the lookup to max. 8 domain labels to the right. |
1071 | | */ |
1072 | 0 | for (p = domain + strlen(domain) - 1; p >= domain; p--) { |
1073 | 0 | if (*p == '.' && ++nlabels > 8) { |
1074 | 0 | domain = p + 1; |
1075 | 0 | break; |
1076 | 0 | } |
1077 | 0 | } |
1078 | | |
1079 | | /* |
1080 | | * We check from left to right to catch special PSL entries like 'forgot.his.name': |
1081 | | * 'forgot.his.name' and 'name' are in the PSL while 'his.name' is not. |
1082 | | */ |
1083 | |
|
1084 | 0 | while (!is_public_suffix(psl, domain, 0)) { |
1085 | 0 | if ((domain = strchr(domain, '.'))) |
1086 | 0 | domain++; |
1087 | 0 | else |
1088 | 0 | break; /* prevent endless loop if is_public_suffix() is broken. */ |
1089 | 0 | } |
1090 | |
|
1091 | 0 | return domain; |
1092 | 0 | } |
1093 | | |
1094 | | /** |
1095 | | * psl_registrable_domain: |
1096 | | * @psl: PSL context |
1097 | | * @domain: Domain string |
1098 | | * |
1099 | | * This function finds the shortest private suffix part of @domain by the means |
1100 | | * of the [Mozilla Public Suffix List](https://publicsuffix.org). |
1101 | | * |
1102 | | * International @domain names have to be either in UTF-8 (lowercase + NFKC) or in ASCII/ACE format (punycode). |
1103 | | * Other encodings likely result in incorrect return values. |
1104 | | * Use helper function psl_str_to_utf8lower() for normalization @domain. |
1105 | | * |
1106 | | * @psl is a context returned by either psl_load_file(), psl_load_fp() or |
1107 | | * psl_builtin(). |
1108 | | * |
1109 | | * Returns: Pointer to shortest private suffix part of @domain or %NULL if @domain |
1110 | | * does not contain a private suffix (or if @psl is %NULL). |
1111 | | * |
1112 | | * Since: 0.1 |
1113 | | */ |
1114 | | const char *psl_registrable_domain(const psl_ctx_t *psl, const char *domain) |
1115 | 0 | { |
1116 | 0 | const char *p, *regdom = NULL; |
1117 | 0 | int nlabels = 0; |
1118 | |
|
1119 | 0 | if (!psl || !domain || *domain == '.') |
1120 | 0 | return NULL; |
1121 | | |
1122 | | /* |
1123 | | * In the main loop we introduce a O(N^2) behavior to avoid code duplication. |
1124 | | * To avoid nasty CPU hogging, we limit the lookup to max. 8 domain labels to the right. |
1125 | | */ |
1126 | 0 | for (p = domain + strlen(domain) - 1; p >= domain; p--) { |
1127 | 0 | if (*p == '.' && ++nlabels > 8) { |
1128 | 0 | domain = p + 1; |
1129 | 0 | break; |
1130 | 0 | } |
1131 | 0 | } |
1132 | | |
1133 | | /* |
1134 | | * We check from left to right to catch special PSL entries like 'forgot.his.name': |
1135 | | * 'forgot.his.name' and 'name' are in the PSL while 'his.name' is not. |
1136 | | */ |
1137 | |
|
1138 | 0 | while (!is_public_suffix(psl, domain, 0)) { |
1139 | 0 | if ((p = strchr(domain, '.'))) { |
1140 | 0 | regdom = domain; |
1141 | 0 | domain = p + 1; |
1142 | 0 | } else |
1143 | 0 | break; /* prevent endless loop if is_public_suffix() is broken. */ |
1144 | 0 | } |
1145 | |
|
1146 | 0 | return regdom; |
1147 | 0 | } |
1148 | | |
1149 | | /** |
1150 | | * psl_load_file: |
1151 | | * @fname: Name of PSL file |
1152 | | * |
1153 | | * This function loads the public suffixes file named @fname. |
1154 | | * To free the allocated resources, call psl_free(). |
1155 | | * |
1156 | | * The suffixes are expected to be UTF-8 encoded (lowercase + NFKC) if they are international. |
1157 | | * |
1158 | | * Returns: Pointer to a PSL context or %NULL on failure. |
1159 | | * |
1160 | | * Since: 0.1 |
1161 | | */ |
1162 | | psl_ctx_t *psl_load_file(const char *fname) |
1163 | 0 | { |
1164 | 0 | FILE *fp; |
1165 | 0 | psl_ctx_t *psl = NULL; |
1166 | |
|
1167 | 0 | if (!fname) |
1168 | 0 | return NULL; |
1169 | | |
1170 | 0 | if ((fp = fopen(fname, "rb"))) { |
1171 | 0 | psl = psl_load_fp(fp); |
1172 | 0 | fclose(fp); |
1173 | 0 | } |
1174 | |
|
1175 | 0 | return psl; |
1176 | 0 | } |
1177 | | |
1178 | | /** |
1179 | | * psl_load_fp: |
1180 | | * @fp: %FILE pointer |
1181 | | * |
1182 | | * This function loads the public suffixes from a %FILE pointer. |
1183 | | * To free the allocated resources, call psl_free(). |
1184 | | * |
1185 | | * The suffixes are expected to be UTF-8 encoded (lowercase + NFKC) if they are international. |
1186 | | * |
1187 | | * Returns: Pointer to a PSL context or %NULL on failure. |
1188 | | * |
1189 | | * Since: 0.1 |
1190 | | */ |
1191 | | psl_ctx_t *psl_load_fp(FILE *fp) |
1192 | 0 | { |
1193 | 0 | psl_ctx_t *psl; |
1194 | 0 | psl_entry_t suffix, *suffixp; |
1195 | 0 | char buf[256], *linep, *p; |
1196 | 0 | int type = 0, is_dafsa; |
1197 | 0 | psl_idna_t *idna; |
1198 | |
|
1199 | 0 | if (!fp) |
1200 | 0 | return NULL; |
1201 | | |
1202 | 0 | if (!(psl = calloc(1, sizeof(psl_ctx_t)))) |
1203 | 0 | return NULL; |
1204 | | |
1205 | | /* read first line to allow ASCII / DAFSA detection */ |
1206 | 0 | if (!(linep = fgets(buf, sizeof(buf) - 1, fp))) |
1207 | 0 | goto fail; |
1208 | | |
1209 | 0 | is_dafsa = strlen(buf) == 16 && !strncmp(buf, ".DAFSA@PSL_", 11); |
1210 | |
|
1211 | 0 | if (is_dafsa) { |
1212 | 0 | void *m; |
1213 | 0 | size_t size = 65536, n, len = 0; |
1214 | 0 | int version = atoi(buf + 11); |
1215 | |
|
1216 | 0 | if (version != 0) |
1217 | 0 | goto fail; |
1218 | | |
1219 | 0 | if (!(psl->dafsa = malloc(size))) |
1220 | 0 | goto fail; |
1221 | | |
1222 | 0 | memcpy(psl->dafsa, buf, len); |
1223 | |
|
1224 | 0 | while ((n = fread(psl->dafsa + len, 1, size - len, fp)) > 0) { |
1225 | 0 | len += n; |
1226 | 0 | if (len >= size) { |
1227 | 0 | if (!(m = realloc(psl->dafsa, size *= 2))) |
1228 | 0 | goto fail; |
1229 | 0 | psl->dafsa = m; |
1230 | 0 | } |
1231 | 0 | } |
1232 | | |
1233 | | /* release unused memory */ |
1234 | 0 | if ((m = realloc(psl->dafsa, len))) |
1235 | 0 | psl->dafsa = m; |
1236 | 0 | else if (!len) |
1237 | 0 | psl->dafsa = NULL; /* realloc() just free'd psl->dafsa */ |
1238 | |
|
1239 | 0 | psl->dafsa_size = len; |
1240 | 0 | psl->utf8 = !!GetUtfMode(psl->dafsa, len); |
1241 | |
|
1242 | 0 | return psl; |
1243 | 0 | } |
1244 | | |
1245 | 0 | idna = psl_idna_open(); |
1246 | | |
1247 | | /* |
1248 | | * as of 02.11.2012, the list at https://publicsuffix.org/list/ contains ~6000 rules and 40 exceptions. |
1249 | | * as of 19.02.2014, the list at https://publicsuffix.org/list/ contains ~6500 rules and 19 exceptions. |
1250 | | * as of 07.10.2018, the list at https://publicsuffix.org/list/ contains ~8600 rules and 8 exceptions. |
1251 | | */ |
1252 | 0 | psl->suffixes = vector_alloc(8*1024, suffix_compare_array); |
1253 | 0 | psl->utf8 = 1; /* we put UTF-8 and punycode rules in the lookup vector */ |
1254 | |
|
1255 | 0 | do { |
1256 | 0 | while (isspace_ascii(*linep)) linep++; /* ignore leading whitespace */ |
1257 | 0 | if (!*linep) continue; /* skip empty lines */ |
1258 | | |
1259 | 0 | if (*linep == '/' && linep[1] == '/') { |
1260 | 0 | if (!type) { |
1261 | 0 | if (strstr(linep + 2, "===BEGIN ICANN DOMAINS===")) |
1262 | 0 | type = PRIV_PSL_FLAG_ICANN; |
1263 | 0 | else if (!type && strstr(linep + 2, "===BEGIN PRIVATE DOMAINS===")) |
1264 | 0 | type = PRIV_PSL_FLAG_PRIVATE; |
1265 | 0 | } |
1266 | 0 | else if (type == PRIV_PSL_FLAG_ICANN && strstr(linep + 2, "===END ICANN DOMAINS===")) |
1267 | 0 | type = 0; |
1268 | 0 | else if (type == PRIV_PSL_FLAG_PRIVATE && strstr(linep + 2, "===END PRIVATE DOMAINS===")) |
1269 | 0 | type = 0; |
1270 | |
|
1271 | 0 | continue; /* skip comments */ |
1272 | 0 | } |
1273 | | |
1274 | | /* parse suffix rule */ |
1275 | 0 | for (p = linep; *linep && !isspace_ascii(*linep);) linep++; |
1276 | 0 | *linep = 0; |
1277 | |
|
1278 | 0 | if (*p == '!') { |
1279 | 0 | p++; |
1280 | 0 | suffix.flags = PRIV_PSL_FLAG_EXCEPTION | type; |
1281 | 0 | psl->nexceptions++; |
1282 | 0 | } else if (*p == '*') { |
1283 | 0 | if (*++p != '.') { |
1284 | | /* fprintf(stderr, "Unsupported kind of rule (ignored): %s\n", p - 1); */ |
1285 | 0 | continue; |
1286 | 0 | } |
1287 | 0 | p++; |
1288 | | /* wildcard *.foo.bar implicitly make foo.bar a public suffix */ |
1289 | 0 | suffix.flags = PRIV_PSL_FLAG_WILDCARD | PRIV_PSL_FLAG_PLAIN | type; |
1290 | 0 | psl->nwildcards++; |
1291 | 0 | psl->nsuffixes++; |
1292 | 0 | } else { |
1293 | 0 | suffix.flags = PRIV_PSL_FLAG_PLAIN | type; |
1294 | 0 | psl->nsuffixes++; |
1295 | 0 | } |
1296 | | |
1297 | 0 | if (suffix_init(&suffix, p, linep - p) == 0) { |
1298 | 0 | int index; |
1299 | |
|
1300 | 0 | if ((index = vector_find(psl->suffixes, &suffix)) >= 0) { |
1301 | | /* Found existing entry: |
1302 | | * Combination of exception and plain rule is ambiguous |
1303 | | * !foo.bar |
1304 | | * foo.bar |
1305 | | * |
1306 | | * Allowed: |
1307 | | * !foo.bar + *.foo.bar |
1308 | | * foo.bar + *.foo.bar |
1309 | | * |
1310 | | * We do not check here, let's do it later. |
1311 | | */ |
1312 | |
|
1313 | 0 | suffixp = vector_get(psl->suffixes, index); |
1314 | 0 | suffixp->flags |= suffix.flags; |
1315 | 0 | } else { |
1316 | | /* New entry */ |
1317 | 0 | suffixp = vector_get(psl->suffixes, vector_add(psl->suffixes, &suffix)); |
1318 | 0 | } |
1319 | |
|
1320 | 0 | if (suffixp) { |
1321 | 0 | suffixp->label = suffixp->label_buf; /* set label to changed address */ |
1322 | 0 | add_punycode_if_needed(idna, psl->suffixes, suffixp); |
1323 | 0 | } |
1324 | 0 | } |
1325 | 0 | } while ((linep = fgets(buf, sizeof(buf), fp))); |
1326 | | |
1327 | 0 | vector_sort(psl->suffixes); |
1328 | |
|
1329 | 0 | psl_idna_close(idna); |
1330 | |
|
1331 | 0 | return psl; |
1332 | | |
1333 | 0 | fail: |
1334 | 0 | psl_free(psl); |
1335 | 0 | return NULL; |
1336 | 0 | } |
1337 | | |
1338 | | /** |
1339 | | * psl_free: |
1340 | | * @psl: PSL context pointer |
1341 | | * |
1342 | | * This function frees the the PSL context that has been retrieved via |
1343 | | * psl_load_fp() or psl_load_file(). |
1344 | | * |
1345 | | * Since: 0.1 |
1346 | | */ |
1347 | | void psl_free(psl_ctx_t *psl) |
1348 | 0 | { |
1349 | 0 | if (psl && psl != &builtin_psl) { |
1350 | 0 | vector_free(&psl->suffixes); |
1351 | 0 | free(psl->dafsa); |
1352 | 0 | free(psl); |
1353 | 0 | } |
1354 | 0 | } |
1355 | | |
1356 | | /** |
1357 | | * psl_builtin: |
1358 | | * |
1359 | | * This function returns the PSL context that has been generated and built in at compile-time. |
1360 | | * You don't have to free the returned context explicitly. |
1361 | | * |
1362 | | * The builtin data also contains punycode entries, one for each international domain name. |
1363 | | * |
1364 | | * If the generation of built-in data has been disabled during compilation, %NULL will be returned. |
1365 | | * When using the builtin psl context, you can provide UTF-8 (lowercase + NFKC) or ASCII/ACE (punycode) |
1366 | | * representations of domains to functions like psl_is_public_suffix(). |
1367 | | * |
1368 | | * Returns: Pointer to the built in PSL data or %NULL if this data is not available. |
1369 | | * |
1370 | | * Since: 0.1 |
1371 | | */ |
1372 | | const psl_ctx_t *psl_builtin(void) |
1373 | 0 | { |
1374 | 0 | #ifdef ENABLE_BUILTIN |
1375 | 0 | return &builtin_psl; |
1376 | | #else |
1377 | | return NULL; |
1378 | | #endif |
1379 | 0 | } |
1380 | | |
1381 | | /** |
1382 | | * psl_suffix_count: |
1383 | | * @psl: PSL context pointer |
1384 | | * |
1385 | | * This function returns number of public suffixes maintained by @psl. |
1386 | | * The number of exceptions within the Public Suffix List are not included. |
1387 | | * |
1388 | | * If the information is not available, the return value is -1 (since 0.19). |
1389 | | * This is the case with DAFSA blobs or if @psl is %NULL. |
1390 | | * |
1391 | | * Returns: Number of public suffixes entries in PSL context or -1 if this information is not available. |
1392 | | * |
1393 | | * Since: 0.1 |
1394 | | */ |
1395 | | int psl_suffix_count(const psl_ctx_t *psl) |
1396 | 0 | { |
1397 | 0 | if (psl == &builtin_psl) |
1398 | 0 | return _psl_nsuffixes; |
1399 | 0 | else if (psl) |
1400 | 0 | return psl->dafsa ? -1 : psl->nsuffixes; |
1401 | 0 | else |
1402 | 0 | return -1; |
1403 | 0 | } |
1404 | | |
1405 | | /** |
1406 | | * psl_suffix_exception_count: |
1407 | | * @psl: PSL context pointer |
1408 | | * |
1409 | | * This function returns number of public suffix exceptions maintained by @psl. |
1410 | | * |
1411 | | * If the information is not available, the return value is -1 (since 0.19). |
1412 | | * This is the case with DAFSA blobs or if @psl is %NULL. |
1413 | | * |
1414 | | * Returns: Number of public suffix exceptions in PSL context or -1 if this information is not available. |
1415 | | * |
1416 | | * Since: 0.1 |
1417 | | */ |
1418 | | int psl_suffix_exception_count(const psl_ctx_t *psl) |
1419 | 0 | { |
1420 | 0 | if (psl == &builtin_psl) |
1421 | 0 | return _psl_nexceptions; |
1422 | 0 | else if (psl) |
1423 | 0 | return psl->dafsa ? -1 : psl->nexceptions; |
1424 | 0 | else |
1425 | 0 | return -1; |
1426 | 0 | } |
1427 | | |
1428 | | /** |
1429 | | * psl_suffix_wildcard_count: |
1430 | | * @psl: PSL context pointer |
1431 | | * |
1432 | | * This function returns number of public suffix wildcards maintained by @psl. |
1433 | | * |
1434 | | * If the information is not available, the return value is -1 (since 0.19). |
1435 | | * This is the case with DAFSA blobs or if @psl is %NULL. |
1436 | | * |
1437 | | * Returns: Number of public suffix wildcards in PSL context or -1 if this information is not available. |
1438 | | * |
1439 | | * Since: 0.10.0 |
1440 | | */ |
1441 | | int psl_suffix_wildcard_count(const psl_ctx_t *psl) |
1442 | 0 | { |
1443 | 0 | if (psl == &builtin_psl) |
1444 | 0 | return _psl_nwildcards; |
1445 | 0 | else if (psl) |
1446 | 0 | return psl->dafsa ? -1 : psl->nwildcards; |
1447 | 0 | else |
1448 | 0 | return -1; |
1449 | 0 | } |
1450 | | |
1451 | | /** |
1452 | | * psl_builtin_file_time: |
1453 | | * |
1454 | | * This function returns the mtime of the Public Suffix List file that has been built in. |
1455 | | * |
1456 | | * If the generation of built-in data has been disabled during compilation, 0 will be returned. |
1457 | | * |
1458 | | * Returns: time_t value or 0. |
1459 | | * |
1460 | | * Since: 0.1 |
1461 | | */ |
1462 | | time_t psl_builtin_file_time(void) |
1463 | 0 | { |
1464 | 0 | return _psl_file_time; |
1465 | 0 | } |
1466 | | |
1467 | | /** |
1468 | | * psl_builtin_sha1sum: |
1469 | | * |
1470 | | * This function returns the SHA1 checksum of the Public Suffix List file that has been built in. |
1471 | | * The returned string is in lowercase hex encoding, e.g. "2af1e9e3044eda0678bb05949d7cca2f769901d8". |
1472 | | * |
1473 | | * If the generation of built-in data has been disabled during compilation, an empty string will be returned. |
1474 | | * |
1475 | | * Returns: String containing SHA1 checksum or an empty string. |
1476 | | * |
1477 | | * Since: 0.1 |
1478 | | */ |
1479 | | const char *psl_builtin_sha1sum(void) |
1480 | 0 | { |
1481 | 0 | return _psl_sha1_checksum; |
1482 | 0 | } |
1483 | | |
1484 | | /** |
1485 | | * psl_builtin_filename: |
1486 | | * |
1487 | | * This function returns the file name of the Public Suffix List file that has been built in. |
1488 | | * |
1489 | | * If the generation of built-in data has been disabled during compilation, an empty string will be returned. |
1490 | | * |
1491 | | * Returns: String containing the PSL file name or an empty string. |
1492 | | * |
1493 | | * Since: 0.1 |
1494 | | */ |
1495 | | const char *psl_builtin_filename(void) |
1496 | 0 | { |
1497 | 0 | return _psl_filename; |
1498 | 0 | } |
1499 | | |
1500 | | /** |
1501 | | * psl_builtin_outdated: |
1502 | | * |
1503 | | * This function checks if the built-in data is older than the file it has been created from. |
1504 | | * If it is, it might be a good idea for the application to reload the PSL. |
1505 | | * The mtime is taken as reference. |
1506 | | * |
1507 | | * If the PSL file does not exist, it is assumed that the built-in data is not outdated. |
1508 | | * |
1509 | | * Returns: 1 if the built-in is outdated, 0 otherwise. |
1510 | | * |
1511 | | * Since: 0.10.0 |
1512 | | */ |
1513 | | int psl_builtin_outdated(void) |
1514 | 0 | { |
1515 | 0 | struct stat st; |
1516 | |
|
1517 | 0 | if (stat(_psl_filename, &st) == 0 && st.st_mtime > _psl_file_time) |
1518 | 0 | return 1; |
1519 | | |
1520 | 0 | return 0; |
1521 | 0 | } |
1522 | | |
1523 | | /** |
1524 | | * psl_dist_filename: |
1525 | | * |
1526 | | * This function returns the file name of the distribution/system PSL data file. |
1527 | | * This file will be considered by psl_latest(). |
1528 | | * |
1529 | | * Return the filename that is set by ./configure --with-psl-distfile, or an empty string. |
1530 | | * |
1531 | | * Returns: String containing a PSL file name or an empty string. |
1532 | | * |
1533 | | * Since: 0.16 |
1534 | | */ |
1535 | | const char *psl_dist_filename(void) |
1536 | 0 | { |
1537 | 0 | return _psl_dist_filename; |
1538 | 0 | } |
1539 | | |
1540 | | /** |
1541 | | * psl_get_version: |
1542 | | * |
1543 | | * Get libpsl version. |
1544 | | * |
1545 | | * Returns: String containing version of libpsl. |
1546 | | * |
1547 | | * Since: 0.2.5 |
1548 | | **/ |
1549 | | const char *psl_get_version(void) |
1550 | 0 | { |
1551 | | #ifdef WITH_LIBICU |
1552 | | return PACKAGE_VERSION " (+libicu/" U_ICU_VERSION ")"; |
1553 | | #elif defined(WITH_LIBIDN2) |
1554 | 0 | return PACKAGE_VERSION " (+libidn2/" IDN2_VERSION ")"; |
1555 | | #elif defined(WITH_LIBIDN) |
1556 | | return PACKAGE_VERSION " (+libidn/" STRINGPREP_VERSION ")"; |
1557 | | #else |
1558 | | return PACKAGE_VERSION " (no IDNA support)"; |
1559 | | #endif |
1560 | 0 | } |
1561 | | |
1562 | | /** |
1563 | | * psl_check_version_number: |
1564 | | * @version: Version number (hex) to check against. |
1565 | | * |
1566 | | * Check the given version number is at minimum the current library version number. |
1567 | | * The version number must be a hexadecimal number like 0x000a01 (V0.10.1). |
1568 | | * |
1569 | | * Returns: Returns the library version number if the given version number is at least |
1570 | | * the version of the library, else return 0; If the argument is 0, the function returns |
1571 | | * the library version number without performing a check. |
1572 | | * |
1573 | | * Since: 0.11.0 |
1574 | | **/ |
1575 | | int psl_check_version_number(int version) |
1576 | 0 | { |
1577 | 0 | if (version) { |
1578 | 0 | int major = version >> 16; |
1579 | 0 | int minor = (version >> 8) & 0xFF; |
1580 | 0 | int patch = version & 0xFF; |
1581 | |
|
1582 | 0 | if (major < PSL_VERSION_MAJOR |
1583 | 0 | || (major == PSL_VERSION_MAJOR && minor < PSL_VERSION_MINOR) |
1584 | 0 | || (major == PSL_VERSION_MAJOR && minor == PSL_VERSION_MINOR && patch < PSL_VERSION_PATCH)) |
1585 | 0 | { |
1586 | 0 | return 0; |
1587 | 0 | } |
1588 | 0 | } |
1589 | | |
1590 | 0 | return PSL_VERSION_NUMBER; |
1591 | 0 | } |
1592 | | /* |
1593 | | * Return true if 'src' is a valid dotted quad, else false. |
1594 | | * Assume that characters '0'..'9' have consecutive byte values. |
1595 | | * credit: |
1596 | | * inspired by Paul Vixie |
1597 | | */ |
1598 | | static int is_ip4(const char *s) |
1599 | 0 | { |
1600 | 0 | int i, n; |
1601 | 0 | unsigned char c; |
1602 | |
|
1603 | 0 | for (i = 0; i < 4; i++) { |
1604 | 0 | if (!(c = *s++) || c < '0' || c > '9') |
1605 | 0 | return 0; |
1606 | | |
1607 | 0 | n = c - '0'; |
1608 | 0 | if ((c = *s++) && c >= '0' && c <= '9') { |
1609 | 0 | n = n * 10 + c - '0'; |
1610 | 0 | if ((c = *s++) && c >= '0' && c <= '9') { |
1611 | 0 | n = n * 10 + c - '0'; |
1612 | 0 | if ((c = *s++) && c >= '0' && c <= '9') { |
1613 | 0 | n = n * 10 + c - '0'; |
1614 | 0 | c = *s++; |
1615 | 0 | } |
1616 | 0 | } |
1617 | 0 | } |
1618 | |
|
1619 | 0 | if (n > 255) |
1620 | 0 | return 0; |
1621 | | |
1622 | 0 | if (i < 3 && c != '.') |
1623 | 0 | return 0; |
1624 | 0 | } |
1625 | | |
1626 | 0 | return !c; |
1627 | 0 | } |
1628 | | |
1629 | | static int hexval(unsigned c) |
1630 | 0 | { |
1631 | 0 | if (c - '0' < 10) return c - '0'; |
1632 | 0 | c |= 32; |
1633 | 0 | if (c - 'a' < 6) return c - 'a' + 10; |
1634 | 0 | return -1; |
1635 | 0 | } |
1636 | | |
1637 | | /* |
1638 | | * Original code taken from musl inet_pton(), |
1639 | | * which has a standard MIT license (https://git.musl-libc.org/cgit/musl/tree/COPYRIGHT). |
1640 | | * Amended and simplified to out needs. |
1641 | | */ |
1642 | | static int is_ip6(const char *s) |
1643 | 0 | { |
1644 | 0 | int i, j, n, d, brk = -1, need_v4 = 0; |
1645 | |
|
1646 | 0 | if (*s == ':' && *++s != ':') return 0; |
1647 | | |
1648 | 0 | for (i = 0; ; i++) { |
1649 | 0 | if (s[0] == ':' && brk < 0) { |
1650 | 0 | brk = i; |
1651 | 0 | if (!*++s) break; |
1652 | 0 | continue; |
1653 | 0 | } |
1654 | 0 | for (n = j = 0; j < 4 && (d = hexval(s[j])) >= 0; j++) |
1655 | 0 | n = n * 16 + d; |
1656 | 0 | if (j == 0) return 0; |
1657 | 0 | if (!s[j] && (brk >= 0 || i == 7)) break; |
1658 | 0 | if (i == 7) return 0; |
1659 | 0 | if (s[j] != ':') { |
1660 | 0 | if (s[j] != '.' || (i < 6 && brk < 0)) return 0; |
1661 | 0 | need_v4 = 1; |
1662 | 0 | i++; |
1663 | 0 | break; |
1664 | 0 | } |
1665 | 0 | s += j + 1; |
1666 | 0 | } |
1667 | | |
1668 | 0 | if (need_v4 && !is_ip4(s)) return 0; |
1669 | 0 | return 1; |
1670 | 0 | } |
1671 | | |
1672 | | /* return whether hostname is an IP address or not */ |
1673 | | static int isip(const char *hostname) |
1674 | 0 | { |
1675 | 0 | return is_ip4(hostname) || is_ip6(hostname); |
1676 | 0 | } |
1677 | | |
1678 | | /** |
1679 | | * psl_is_cookie_domain_acceptable: |
1680 | | * @psl: PSL context pointer |
1681 | | * @hostname: The request hostname. |
1682 | | * @cookie_domain: The domain value from a cookie |
1683 | | * |
1684 | | * This helper function checks whether @cookie_domain is an acceptable cookie domain value for the request |
1685 | | * @hostname. |
1686 | | * |
1687 | | * For international domain names both, @hostname and @cookie_domain, have to be either in UTF-8 (lowercase + NFKC) |
1688 | | * or in ASCII/ACE (punycode) format. Other encodings or mixing UTF-8 and punycode likely result in incorrect return values. |
1689 | | * |
1690 | | * Use helper function psl_str_to_utf8lower() for normalization of @hostname and @cookie_domain. |
1691 | | * |
1692 | | * Hint for Windows users: |
1693 | | * Please make sure the calling application has called WSAStartup() before calling psl_is_cookie_domain_acceptable(). |
1694 | | * |
1695 | | * Examples: |
1696 | | * 1. Cookie domain 'example.com' would be acceptable for hostname 'www.example.com', |
1697 | | * but '.com' or 'com' would NOT be acceptable since 'com' is a public suffix. |
1698 | | * |
1699 | | * 2. Cookie domain 'his.name' would be acceptable for hostname 'remember.his.name', |
1700 | | * but NOT for 'forgot.his.name' since 'forgot.his.name' is a public suffix. |
1701 | | * |
1702 | | * Returns: 1 if acceptable, 0 if not acceptable. |
1703 | | * |
1704 | | * Since: 0.1 |
1705 | | */ |
1706 | | int psl_is_cookie_domain_acceptable(const psl_ctx_t *psl, const char *hostname, const char *cookie_domain) |
1707 | 0 | { |
1708 | 0 | const char *p; |
1709 | 0 | size_t hostname_length, cookie_domain_length; |
1710 | |
|
1711 | 0 | if (!psl || !hostname || !cookie_domain) |
1712 | 0 | return 0; |
1713 | | |
1714 | 0 | while (*cookie_domain == '.') |
1715 | 0 | cookie_domain++; |
1716 | |
|
1717 | 0 | if (!strcmp(hostname, cookie_domain)) |
1718 | 0 | return 1; /* an exact match is acceptable (and pretty common) */ |
1719 | | |
1720 | 0 | if (isip(hostname)) |
1721 | 0 | return 0; /* Hostname is an IP address and these must match fully (RFC 6265, 5.1.3) */ |
1722 | | |
1723 | 0 | cookie_domain_length = strlen(cookie_domain); |
1724 | 0 | hostname_length = strlen(hostname); |
1725 | |
|
1726 | 0 | if (cookie_domain_length >= hostname_length) |
1727 | 0 | return 0; /* cookie_domain is too long */ |
1728 | | |
1729 | 0 | p = hostname + hostname_length - cookie_domain_length; |
1730 | 0 | if (!strcmp(p, cookie_domain) && p[-1] == '.') { |
1731 | | /* OK, cookie_domain matches, but it must be longer than the longest public suffix in 'hostname' */ |
1732 | |
|
1733 | 0 | if (!(p = psl_unregistrable_domain(psl, hostname))) |
1734 | 0 | return 1; |
1735 | | |
1736 | 0 | if (cookie_domain_length > strlen(p)) |
1737 | 0 | return 1; |
1738 | 0 | } |
1739 | | |
1740 | 0 | return 0; |
1741 | 0 | } |
1742 | | |
1743 | | /** |
1744 | | * psl_free_string: |
1745 | | * @str: pointer to lowercase string returned by psl_str_to_utf8lower() |
1746 | | * |
1747 | | * This function free()'s the memory allocated by psl_str_to_utf8lower() when |
1748 | | * returning a lowercase string |
1749 | | * |
1750 | | * Since: 0.19 |
1751 | | */ |
1752 | | void psl_free_string(char *str) |
1753 | 0 | { |
1754 | 0 | if (str) |
1755 | 0 | free(str); |
1756 | 0 | } |
1757 | | |
1758 | | #if defined(WITH_LIBIDN2) || defined(WITH_LIBIDN) |
1759 | | /* Avoid using strcasecmp() or _stricmp() */ |
1760 | 0 | static int isUTF8(const char *s) { |
1761 | 0 | return (s[0] == 'u' || s[0] == 'U') |
1762 | 0 | && (s[1] == 't' || s[1] == 'T') |
1763 | 0 | && (s[2] == 'f' || s[2] == 'F') |
1764 | 0 | && s[3] == '-' && s[4] == 0; |
1765 | 0 | } |
1766 | | #endif |
1767 | | |
1768 | | /** |
1769 | | * psl_str_to_utf8lower: |
1770 | | * @str: string to convert |
1771 | | * @encoding: charset encoding of @str, e.g. 'iso-8859-1' or %NULL |
1772 | | * @locale: locale of @str for to lowercase conversion, e.g. 'de' or %NULL |
1773 | | * @lower: return value containing the converted string |
1774 | | * |
1775 | | * This helper function converts a string to UTF-8 lowercase + NFKC representation. |
1776 | | * Lowercase + NFKC UTF-8 is needed as input to the domain checking functions. |
1777 | | * |
1778 | | * @lower stays unchanged on error. |
1779 | | * |
1780 | | * When returning PSL_SUCCESS, the return value 'lower' must be freed after usage. |
1781 | | * |
1782 | | * Returns: psl_error_t value. |
1783 | | * PSL_SUCCESS: Success |
1784 | | * PSL_ERR_INVALID_ARG: @str is a %NULL value. |
1785 | | * PSL_ERR_CONVERTER: Failed to open the unicode converter with name @encoding |
1786 | | * PSL_ERR_TO_UTF16: Failed to convert @str to unicode |
1787 | | * PSL_ERR_TO_LOWER: Failed to convert unicode to lowercase |
1788 | | * PSL_ERR_TO_UTF8: Failed to convert unicode to UTF-8 |
1789 | | * PSL_ERR_NO_MEM: Failed to allocate memory |
1790 | | * |
1791 | | * Since: 0.4 |
1792 | | */ |
1793 | | psl_error_t psl_str_to_utf8lower(const char *str, const char *encoding, const char *locale, char **lower) |
1794 | 0 | { |
1795 | 0 | int ret = PSL_ERR_INVALID_ARG; |
1796 | |
|
1797 | 0 | (void) encoding; |
1798 | 0 | (void) locale; |
1799 | |
|
1800 | 0 | if (!str) |
1801 | 0 | return PSL_ERR_INVALID_ARG; |
1802 | | |
1803 | | /* shortcut to avoid costly conversion */ |
1804 | 0 | if (str_is_ascii(str)) { |
1805 | 0 | if (lower) { |
1806 | 0 | char *p, *tmp; |
1807 | |
|
1808 | 0 | if (!(tmp = psl_strdup(str))) |
1809 | 0 | return PSL_ERR_NO_MEM; |
1810 | | |
1811 | 0 | *lower = tmp; |
1812 | | |
1813 | | /* convert ASCII string to lowercase */ |
1814 | 0 | for (p = *lower; *p; p++) |
1815 | 0 | if (isupper(*p)) |
1816 | 0 | *p = tolower(*p); |
1817 | 0 | } |
1818 | 0 | return PSL_SUCCESS; |
1819 | 0 | } |
1820 | | |
1821 | | #ifdef WITH_LIBICU |
1822 | | #define STACK_STRLENGTH 256 |
1823 | | do { |
1824 | | UErrorCode status = 0; |
1825 | | UChar *utf16_dst, *utf16_lower; |
1826 | | char *utf8_lower; |
1827 | | int32_t utf16_dst_length, utf16_dst_size, utf16_lower_size, utf8_lower_size; |
1828 | | UConverter *uconv; |
1829 | | UChar utf16_dst_buf[STACK_STRLENGTH * 2 + 1]; |
1830 | | UChar utf16_lower_buf[STACK_STRLENGTH * 2 + 1]; |
1831 | | char utf8_lower_buf[STACK_STRLENGTH * 6 + 1]; |
1832 | | size_t str_length = strlen(str); |
1833 | | |
1834 | | if (str_length <= STACK_STRLENGTH) { |
1835 | | utf16_dst_size = countof(utf16_dst_buf); |
1836 | | utf16_lower_size = countof(utf16_lower_buf); |
1837 | | utf8_lower_size = countof(utf8_lower_buf); |
1838 | | utf16_dst = utf16_dst_buf; |
1839 | | utf16_lower = utf16_lower_buf; |
1840 | | utf8_lower = utf8_lower_buf; |
1841 | | } else { |
1842 | | utf16_dst_size = utf16_lower_size = str_length * 2 + 1; |
1843 | | utf8_lower_size = str_length * 6 + 1; |
1844 | | utf16_dst = malloc(sizeof(UChar) * utf16_dst_size); |
1845 | | utf16_lower = malloc(sizeof(UChar) * utf16_lower_size); |
1846 | | utf8_lower = malloc(sizeof(char) * utf8_lower_size); |
1847 | | |
1848 | | if (!utf16_dst || !utf16_lower || !utf8_lower) { |
1849 | | ret = PSL_ERR_NO_MEM; |
1850 | | goto out; |
1851 | | } |
1852 | | } |
1853 | | |
1854 | | uconv = ucnv_open(encoding, &status); |
1855 | | if (U_SUCCESS(status)) { |
1856 | | utf16_dst_length = ucnv_toUChars(uconv, utf16_dst, utf16_dst_size, str, str_length, &status); |
1857 | | ucnv_close(uconv); |
1858 | | |
1859 | | if (U_SUCCESS(status)) { |
1860 | | int32_t utf16_lower_length = u_strToLower(utf16_lower, utf16_lower_size, utf16_dst, utf16_dst_length, locale, &status); |
1861 | | if (U_SUCCESS(status)) { |
1862 | | u_strToUTF8(utf8_lower, utf8_lower_size, NULL, utf16_lower, utf16_lower_length, &status); |
1863 | | if (U_SUCCESS(status)) { |
1864 | | ret = PSL_SUCCESS; |
1865 | | if (lower) { |
1866 | | char *tmp = psl_strdup(utf8_lower); |
1867 | | |
1868 | | if (tmp) |
1869 | | *lower = tmp; |
1870 | | else |
1871 | | ret = PSL_ERR_NO_MEM; |
1872 | | } |
1873 | | } else { |
1874 | | ret = PSL_ERR_TO_UTF8; |
1875 | | /* fprintf(stderr, "Failed to convert UTF-16 to UTF-8 (status %d)\n", status); */ |
1876 | | } |
1877 | | } else { |
1878 | | ret = PSL_ERR_TO_LOWER; |
1879 | | /* fprintf(stderr, "Failed to convert UTF-16 to lowercase (status %d)\n", status); */ |
1880 | | } |
1881 | | } else { |
1882 | | ret = PSL_ERR_TO_UTF16; |
1883 | | /* fprintf(stderr, "Failed to convert string to UTF-16 (status %d)\n", status); */ |
1884 | | } |
1885 | | } else { |
1886 | | ret = PSL_ERR_CONVERTER; |
1887 | | /* fprintf(stderr, "Failed to open converter for '%s' (status %d)\n", encoding, status); */ |
1888 | | } |
1889 | | out: |
1890 | | if (utf16_dst != utf16_dst_buf) |
1891 | | free(utf16_dst); |
1892 | | if (utf16_lower != utf16_lower_buf) |
1893 | | free(utf16_lower); |
1894 | | if (utf8_lower != utf8_lower_buf) |
1895 | | free(utf8_lower); |
1896 | | |
1897 | | } while (0); |
1898 | | #elif defined(WITH_LIBIDN2) || defined(WITH_LIBIDN) |
1899 | 0 | do { |
1900 | | /* find out local charset encoding */ |
1901 | 0 | if (!encoding) { |
1902 | 0 | #ifdef HAVE_NL_LANGINFO |
1903 | 0 | encoding = nl_langinfo(CODESET); |
1904 | | #elif defined _WIN32 |
1905 | | static char buf[16]; |
1906 | | snprintf(buf, sizeof(buf), "CP%u", GetACP()); |
1907 | | encoding = buf; |
1908 | | #endif |
1909 | 0 | if (!encoding || !*encoding) |
1910 | 0 | encoding = "ASCII"; |
1911 | 0 | } |
1912 | | |
1913 | | /* convert to UTF-8 */ |
1914 | 0 | if (!isUTF8(encoding)) { |
1915 | 0 | iconv_t cd = iconv_open("utf-8", encoding); |
1916 | |
|
1917 | 0 | if (cd != (iconv_t)-1) { |
1918 | 0 | char *tmp = (char *)str; /* iconv won't change where str points to, but changes tmp itself */ |
1919 | 0 | size_t tmp_len = strlen(str) + 1; |
1920 | 0 | size_t dst_len = tmp_len * 6, dst_len_tmp = dst_len; |
1921 | 0 | char *dst = malloc(dst_len + 1), *dst_tmp = dst; |
1922 | |
|
1923 | 0 | if (!dst) { |
1924 | 0 | ret = PSL_ERR_NO_MEM; |
1925 | 0 | } |
1926 | 0 | else if (iconv(cd, (ICONV_CONST char **)&tmp, &tmp_len, &dst_tmp, &dst_len_tmp) != (size_t)-1 |
1927 | 0 | && iconv(cd, NULL, NULL, &dst_tmp, &dst_len_tmp) != (size_t)-1) |
1928 | 0 | { |
1929 | | /* start size for u8_tolower internal memory allocation. |
1930 | | * u8_tolower() does not terminate the result string. we have 0 byte included in above tmp_len |
1931 | | * and thus in len. */ |
1932 | 0 | size_t len = dst_len - dst_len_tmp; |
1933 | |
|
1934 | 0 | if ((tmp = (char *)u8_tolower((uint8_t *)dst, len, 0, UNINORM_NFKC, NULL, &len))) { |
1935 | 0 | ret = PSL_SUCCESS; |
1936 | 0 | if (lower) { |
1937 | 0 | *lower = tmp; |
1938 | 0 | tmp = NULL; |
1939 | 0 | } else |
1940 | 0 | free(tmp); |
1941 | 0 | } else { |
1942 | 0 | ret = PSL_ERR_TO_LOWER; |
1943 | | /* fprintf(stderr, "Failed to convert UTF-8 to lowercase (errno %d)\n", errno); */ |
1944 | 0 | } |
1945 | 0 | } else { |
1946 | 0 | ret = PSL_ERR_TO_UTF8; |
1947 | | /* fprintf(stderr, "Failed to convert '%s' string into '%s' (%d)\n", src_encoding, dst_encoding, errno); */ |
1948 | 0 | } |
1949 | |
|
1950 | 0 | free(dst); |
1951 | 0 | iconv_close(cd); |
1952 | 0 | } else { |
1953 | 0 | ret = PSL_ERR_TO_UTF8; |
1954 | | /* fprintf(stderr, "Failed to prepare encoding '%s' into '%s' (%d)\n", src_encoding, dst_encoding, errno); */ |
1955 | 0 | } |
1956 | 0 | } else { |
1957 | | /* we need a conversion to lowercase */ |
1958 | 0 | uint8_t *tmp; |
1959 | | |
1960 | | /* start size for u8_tolower internal memory allocation. |
1961 | | * u8_tolower() does not terminate the result string, so include terminating 0 byte in len. */ |
1962 | 0 | size_t len = u8_strlen((uint8_t *)str) + 1; |
1963 | |
|
1964 | 0 | if ((tmp = u8_tolower((uint8_t *)str, len, 0, UNINORM_NFKC, NULL, &len))) { |
1965 | 0 | ret = PSL_SUCCESS; |
1966 | 0 | if (lower) { |
1967 | 0 | *lower = (char*)tmp; |
1968 | 0 | tmp = NULL; |
1969 | 0 | } else |
1970 | 0 | free(tmp); |
1971 | 0 | } else { |
1972 | 0 | ret = PSL_ERR_TO_LOWER; |
1973 | | /* fprintf(stderr, "Failed to convert UTF-8 to lowercase (errno %d)\n", errno); */ |
1974 | 0 | } |
1975 | 0 | } |
1976 | |
|
1977 | 0 | } while (0); |
1978 | 0 | #endif |
1979 | |
|
1980 | 0 | return ret; |
1981 | 0 | } |
1982 | | |
1983 | | /* if file is newer than the builtin data, insert it reverse sorted by mtime */ |
1984 | | static int insert_file(const char *fname, const char **psl_fname, time_t *psl_mtime, int n) |
1985 | 0 | { |
1986 | 0 | struct stat st; |
1987 | 0 | int it; |
1988 | |
|
1989 | 0 | if (fname && *fname && stat(fname, &st) == 0 && st.st_mtime > _psl_file_time) { |
1990 | | /* add file name and mtime to end of array */ |
1991 | 0 | psl_fname[n] = fname; |
1992 | 0 | psl_mtime[n++] = st.st_mtime; |
1993 | | |
1994 | | /* move the new entry to it's correct position */ |
1995 | 0 | for (it = n - 2; it >= 0 && st.st_mtime > psl_mtime[it]; it--) { |
1996 | 0 | psl_fname[it + 1] = psl_fname[it]; |
1997 | 0 | psl_mtime[it + 1] = psl_mtime[it]; |
1998 | 0 | psl_fname[it] = fname; |
1999 | 0 | psl_mtime[it] = st.st_mtime; |
2000 | 0 | } |
2001 | 0 | } |
2002 | |
|
2003 | 0 | return n; |
2004 | 0 | } |
2005 | | |
2006 | | /** |
2007 | | * psl_latest: |
2008 | | * @fname: Name of PSL file or %NULL |
2009 | | * |
2010 | | * This function loads the the latest available PSL data from either |
2011 | | * - @fname (application specific filename, may be %NULL) |
2012 | | * - location specified during built-time (filename from ./configure --with-psl-distfile) |
2013 | | * - built-in PSL data (generated from ./configure --with-psl-file) |
2014 | | * - location of built-in data (filename from ./configure --with-psl-file) |
2015 | | * |
2016 | | * If none of the above is available, the function returns %NULL. |
2017 | | * |
2018 | | * To free the allocated resources, call psl_free(). |
2019 | | * |
2020 | | * Returns: Pointer to a PSL context or %NULL on failure. |
2021 | | * |
2022 | | * Since: 0.16 |
2023 | | */ |
2024 | | psl_ctx_t *psl_latest(const char *fname) |
2025 | 0 | { |
2026 | 0 | psl_ctx_t *psl; |
2027 | 0 | const char *psl_fname[3]; |
2028 | 0 | time_t psl_mtime[3]; |
2029 | 0 | int it, ntimes; |
2030 | |
|
2031 | 0 | psl_fname[0] = NULL; /* silence gcc 6.2 false warning */ |
2032 | | |
2033 | | /* create array of PSL files reverse sorted by mtime (latest first) */ |
2034 | 0 | ntimes = insert_file(fname, psl_fname, psl_mtime, 0); |
2035 | 0 | ntimes = insert_file(_psl_dist_filename, psl_fname, psl_mtime, ntimes); |
2036 | 0 | ntimes = insert_file(_psl_filename, psl_fname, psl_mtime, ntimes); |
2037 | | |
2038 | | /* load PSL data from the latest file, falling back to the second recent, ... */ |
2039 | 0 | for (psl = NULL, it = 0; it < ntimes; it++) { |
2040 | 0 | if (psl_mtime[it] > _psl_file_time) |
2041 | 0 | if ((psl = psl_load_file(psl_fname[it]))) |
2042 | 0 | break; |
2043 | 0 | } |
2044 | | |
2045 | | /* if file loading failed or there is no file newer than the builtin data, |
2046 | | * then return the builtin data. */ |
2047 | 0 | return psl ? psl : (psl_ctx_t *) psl_builtin(); |
2048 | 0 | } |