]>
Commit | Line | Data |
---|---|---|
14f25df2 | 1 | import base64 |
b87e01c1 | 2 | import collections |
f59f5ef8 | 3 | import contextlib |
54007a45 | 4 | import http.cookiejar |
8817a80d | 5 | import http.cookies |
b87e01c1 | 6 | import io |
982ee69a MB |
7 | import json |
8 | import os | |
9bd13fe5 | 9 | import re |
982ee69a | 10 | import shutil |
982ee69a MB |
11 | import struct |
12 | import subprocess | |
13 | import sys | |
14 | import tempfile | |
2e4585da | 15 | import time |
b87e01c1 | 16 | import urllib.request |
982ee69a | 17 | from datetime import datetime, timedelta, timezone |
f59f5ef8 | 18 | from enum import Enum, auto |
982ee69a MB |
19 | from hashlib import pbkdf2_hmac |
20 | ||
1d3586d0 | 21 | from .aes import ( |
22 | aes_cbc_decrypt_bytes, | |
23 | aes_gcm_decrypt_and_verify_bytes, | |
24 | unpad_pkcs7, | |
25 | ) | |
9b7a48ab | 26 | from .compat import functools |
9b8ee23b | 27 | from .dependencies import ( |
28 | _SECRETSTORAGE_UNAVAILABLE_REASON, | |
29 | secretstorage, | |
30 | sqlite3, | |
31 | ) | |
97ec5bc5 | 32 | from .minicurses import MultilinePrinter, QuietMultilinePrinter |
d2c8aadf | 33 | from .utils import ( |
34 | Popen, | |
d2c8aadf | 35 | error_to_str, |
b87e01c1 | 36 | escape_url, |
d2c8aadf | 37 | expand_path, |
22df97f9 | 38 | is_path_like, |
b87e01c1 | 39 | sanitize_url, |
40 | str_or_none, | |
d2c8aadf | 41 | try_call, |
b87e01c1 | 42 | write_string, |
d2c8aadf | 43 | ) |
982ee69a | 44 | |
982ee69a MB |
45 | CHROMIUM_BASED_BROWSERS = {'brave', 'chrome', 'chromium', 'edge', 'opera', 'vivaldi'} |
46 | SUPPORTED_BROWSERS = CHROMIUM_BASED_BROWSERS | {'firefox', 'safari'} | |
47 | ||
48 | ||
49 | class YDLLogger: | |
50 | def __init__(self, ydl=None): | |
51 | self._ydl = ydl | |
52 | ||
53 | def debug(self, message): | |
54 | if self._ydl: | |
55 | self._ydl.write_debug(message) | |
56 | ||
57 | def info(self, message): | |
58 | if self._ydl: | |
59 | self._ydl.to_screen(f'[Cookies] {message}') | |
60 | ||
61 | def warning(self, message, only_once=False): | |
62 | if self._ydl: | |
63 | self._ydl.report_warning(message, only_once) | |
64 | ||
65 | def error(self, message): | |
66 | if self._ydl: | |
67 | self._ydl.report_error(message) | |
68 | ||
2e4585da | 69 | class ProgressBar(MultilinePrinter): |
70 | _DELAY, _timer = 0.1, 0 | |
71 | ||
72 | def print(self, message): | |
73 | if time.time() - self._timer > self._DELAY: | |
74 | self.print_at_line(f'[Cookies] {message}', 0) | |
75 | self._timer = time.time() | |
76 | ||
97ec5bc5 | 77 | def progress_bar(self): |
78 | """Return a context manager with a print method. (Optional)""" | |
79 | # Do not print to files/pipes, loggers, or when --no-progress is used | |
80 | if not self._ydl or self._ydl.params.get('noprogress') or self._ydl.params.get('logger'): | |
81 | return | |
591bb9d3 | 82 | file = self._ydl._out_files.error |
97ec5bc5 | 83 | try: |
84 | if not file.isatty(): | |
85 | return | |
86 | except BaseException: | |
87 | return | |
2e4585da | 88 | return self.ProgressBar(file, preserve_output=False) |
97ec5bc5 | 89 | |
90 | ||
91 | def _create_progress_bar(logger): | |
92 | if hasattr(logger, 'progress_bar'): | |
93 | printer = logger.progress_bar() | |
94 | if printer: | |
95 | return printer | |
96 | printer = QuietMultilinePrinter() | |
97 | printer.print = lambda _: None | |
98 | return printer | |
99 | ||
982ee69a MB |
100 | |
101 | def load_cookies(cookie_file, browser_specification, ydl): | |
102 | cookie_jars = [] | |
103 | if browser_specification is not None: | |
9bd13fe5 | 104 | browser_name, profile, keyring, container = _parse_browser_specification(*browser_specification) |
105 | cookie_jars.append( | |
106 | extract_cookies_from_browser(browser_name, profile, YDLLogger(ydl), keyring=keyring, container=container)) | |
982ee69a MB |
107 | |
108 | if cookie_file is not None: | |
22df97f9 | 109 | is_filename = is_path_like(cookie_file) |
d76fa1f3 | 110 | if is_filename: |
111 | cookie_file = expand_path(cookie_file) | |
112 | ||
982ee69a | 113 | jar = YoutubeDLCookieJar(cookie_file) |
d76fa1f3 | 114 | if not is_filename or os.access(cookie_file, os.R_OK): |
982ee69a MB |
115 | jar.load(ignore_discard=True, ignore_expires=True) |
116 | cookie_jars.append(jar) | |
117 | ||
118 | return _merge_cookie_jars(cookie_jars) | |
119 | ||
120 | ||
9bd13fe5 | 121 | def extract_cookies_from_browser(browser_name, profile=None, logger=YDLLogger(), *, keyring=None, container=None): |
982ee69a | 122 | if browser_name == 'firefox': |
9bd13fe5 | 123 | return _extract_firefox_cookies(profile, container, logger) |
982ee69a MB |
124 | elif browser_name == 'safari': |
125 | return _extract_safari_cookies(profile, logger) | |
126 | elif browser_name in CHROMIUM_BASED_BROWSERS: | |
f59f5ef8 | 127 | return _extract_chrome_cookies(browser_name, profile, keyring, logger) |
982ee69a | 128 | else: |
86e5f3ed | 129 | raise ValueError(f'unknown browser: {browser_name}') |
982ee69a MB |
130 | |
131 | ||
9bd13fe5 | 132 | def _extract_firefox_cookies(profile, container, logger): |
982ee69a | 133 | logger.info('Extracting cookies from firefox') |
9b8ee23b | 134 | if not sqlite3: |
767b02a9 MB |
135 | logger.warning('Cannot extract cookies from firefox without sqlite3 support. ' |
136 | 'Please use a python interpreter compiled with sqlite3 support') | |
137 | return YoutubeDLCookieJar() | |
982ee69a MB |
138 | |
139 | if profile is None: | |
140 | search_root = _firefox_browser_dir() | |
141 | elif _is_path(profile): | |
142 | search_root = profile | |
143 | else: | |
144 | search_root = os.path.join(_firefox_browser_dir(), profile) | |
145 | ||
825d3ce3 | 146 | cookie_database_path = _find_most_recently_used_file(search_root, 'cookies.sqlite', logger) |
147 | if cookie_database_path is None: | |
148 | raise FileNotFoundError(f'could not find firefox cookies database in {search_root}') | |
149 | logger.debug(f'Extracting cookies from: "{cookie_database_path}"') | |
150 | ||
9bd13fe5 | 151 | container_id = None |
825d3ce3 | 152 | if container not in (None, 'none'): |
153 | containers_path = os.path.join(os.path.dirname(cookie_database_path), 'containers.json') | |
9bd13fe5 | 154 | if not os.path.isfile(containers_path) or not os.access(containers_path, os.R_OK): |
155 | raise FileNotFoundError(f'could not read containers.json in {search_root}') | |
d2c8aadf | 156 | with open(containers_path) as containers: |
9bd13fe5 | 157 | identities = json.load(containers).get('identities', []) |
158 | container_id = next((context.get('userContextId') for context in identities if container in ( | |
159 | context.get('name'), | |
160 | try_call(lambda: re.fullmatch(r'userContext([^\.]+)\.label', context['l10nID']).group()) | |
161 | )), None) | |
162 | if not isinstance(container_id, int): | |
163 | raise ValueError(f'could not find firefox container "{container}" in containers.json') | |
164 | ||
0930b11f | 165 | with tempfile.TemporaryDirectory(prefix='yt_dlp') as tmpdir: |
982ee69a MB |
166 | cursor = None |
167 | try: | |
168 | cursor = _open_database_copy(cookie_database_path, tmpdir) | |
9bd13fe5 | 169 | if isinstance(container_id, int): |
9bd13fe5 | 170 | logger.debug( |
171 | f'Only loading cookies from firefox container "{container}", ID {container_id}') | |
9bd13fe5 | 172 | cursor.execute( |
825d3ce3 | 173 | 'SELECT host, name, value, path, expiry, isSecure FROM moz_cookies WHERE originAttributes LIKE ? OR originAttributes LIKE ?', |
174 | (f'%userContextId={container_id}', f'%userContextId={container_id}&%')) | |
175 | elif container == 'none': | |
176 | logger.debug('Only loading cookies not belonging to any container') | |
177 | cursor.execute( | |
178 | 'SELECT host, name, value, path, expiry, isSecure FROM moz_cookies WHERE NOT INSTR(originAttributes,"userContextId=")') | |
179 | else: | |
9bd13fe5 | 180 | cursor.execute('SELECT host, name, value, path, expiry, isSecure FROM moz_cookies') |
982ee69a | 181 | jar = YoutubeDLCookieJar() |
97ec5bc5 | 182 | with _create_progress_bar(logger) as progress_bar: |
183 | table = cursor.fetchall() | |
184 | total_cookie_count = len(table) | |
185 | for i, (host, name, value, path, expiry, is_secure) in enumerate(table): | |
186 | progress_bar.print(f'Loading cookie {i: 6d}/{total_cookie_count: 6d}') | |
ac668111 | 187 | cookie = http.cookiejar.Cookie( |
97ec5bc5 | 188 | version=0, name=name, value=value, port=None, port_specified=False, |
189 | domain=host, domain_specified=bool(host), domain_initial_dot=host.startswith('.'), | |
190 | path=path, path_specified=bool(path), secure=is_secure, expires=expiry, discard=False, | |
191 | comment=None, comment_url=None, rest={}) | |
192 | jar.set_cookie(cookie) | |
86e5f3ed | 193 | logger.info(f'Extracted {len(jar)} cookies from firefox') |
982ee69a MB |
194 | return jar |
195 | finally: | |
196 | if cursor is not None: | |
197 | cursor.connection.close() | |
198 | ||
199 | ||
200 | def _firefox_browser_dir(): | |
dec30912 | 201 | if sys.platform in ('cygwin', 'win32'): |
19a03940 | 202 | return os.path.expandvars(R'%APPDATA%\Mozilla\Firefox\Profiles') |
982ee69a MB |
203 | elif sys.platform == 'darwin': |
204 | return os.path.expanduser('~/Library/Application Support/Firefox') | |
dec30912 | 205 | return os.path.expanduser('~/.mozilla/firefox') |
982ee69a MB |
206 | |
207 | ||
208 | def _get_chromium_based_browser_settings(browser_name): | |
209 | # https://chromium.googlesource.com/chromium/src/+/HEAD/docs/user_data_dir.md | |
dec30912 | 210 | if sys.platform in ('cygwin', 'win32'): |
982ee69a MB |
211 | appdata_local = os.path.expandvars('%LOCALAPPDATA%') |
212 | appdata_roaming = os.path.expandvars('%APPDATA%') | |
213 | browser_dir = { | |
19a03940 | 214 | 'brave': os.path.join(appdata_local, R'BraveSoftware\Brave-Browser\User Data'), |
215 | 'chrome': os.path.join(appdata_local, R'Google\Chrome\User Data'), | |
216 | 'chromium': os.path.join(appdata_local, R'Chromium\User Data'), | |
217 | 'edge': os.path.join(appdata_local, R'Microsoft\Edge\User Data'), | |
218 | 'opera': os.path.join(appdata_roaming, R'Opera Software\Opera Stable'), | |
219 | 'vivaldi': os.path.join(appdata_local, R'Vivaldi\User Data'), | |
982ee69a MB |
220 | }[browser_name] |
221 | ||
222 | elif sys.platform == 'darwin': | |
223 | appdata = os.path.expanduser('~/Library/Application Support') | |
224 | browser_dir = { | |
225 | 'brave': os.path.join(appdata, 'BraveSoftware/Brave-Browser'), | |
226 | 'chrome': os.path.join(appdata, 'Google/Chrome'), | |
227 | 'chromium': os.path.join(appdata, 'Chromium'), | |
228 | 'edge': os.path.join(appdata, 'Microsoft Edge'), | |
229 | 'opera': os.path.join(appdata, 'com.operasoftware.Opera'), | |
230 | 'vivaldi': os.path.join(appdata, 'Vivaldi'), | |
231 | }[browser_name] | |
232 | ||
233 | else: | |
dec30912 CMT |
234 | config = _config_home() |
235 | browser_dir = { | |
236 | 'brave': os.path.join(config, 'BraveSoftware/Brave-Browser'), | |
237 | 'chrome': os.path.join(config, 'google-chrome'), | |
238 | 'chromium': os.path.join(config, 'chromium'), | |
239 | 'edge': os.path.join(config, 'microsoft-edge'), | |
240 | 'opera': os.path.join(config, 'opera'), | |
241 | 'vivaldi': os.path.join(config, 'vivaldi'), | |
242 | }[browser_name] | |
982ee69a MB |
243 | |
244 | # Linux keyring names can be determined by snooping on dbus while opening the browser in KDE: | |
245 | # dbus-monitor "interface='org.kde.KWallet'" "type=method_return" | |
246 | keyring_name = { | |
247 | 'brave': 'Brave', | |
248 | 'chrome': 'Chrome', | |
249 | 'chromium': 'Chromium', | |
29b208f6 | 250 | 'edge': 'Microsoft Edge' if sys.platform == 'darwin' else 'Chromium', |
982ee69a MB |
251 | 'opera': 'Opera' if sys.platform == 'darwin' else 'Chromium', |
252 | 'vivaldi': 'Vivaldi' if sys.platform == 'darwin' else 'Chrome', | |
253 | }[browser_name] | |
254 | ||
255 | browsers_without_profiles = {'opera'} | |
256 | ||
257 | return { | |
258 | 'browser_dir': browser_dir, | |
259 | 'keyring_name': keyring_name, | |
260 | 'supports_profiles': browser_name not in browsers_without_profiles | |
261 | } | |
262 | ||
263 | ||
f59f5ef8 | 264 | def _extract_chrome_cookies(browser_name, profile, keyring, logger): |
86e5f3ed | 265 | logger.info(f'Extracting cookies from {browser_name}') |
767b02a9 | 266 | |
9b8ee23b | 267 | if not sqlite3: |
19a03940 | 268 | logger.warning(f'Cannot extract cookies from {browser_name} without sqlite3 support. ' |
269 | 'Please use a python interpreter compiled with sqlite3 support') | |
767b02a9 MB |
270 | return YoutubeDLCookieJar() |
271 | ||
982ee69a MB |
272 | config = _get_chromium_based_browser_settings(browser_name) |
273 | ||
274 | if profile is None: | |
275 | search_root = config['browser_dir'] | |
276 | elif _is_path(profile): | |
277 | search_root = profile | |
278 | config['browser_dir'] = os.path.dirname(profile) if config['supports_profiles'] else profile | |
279 | else: | |
280 | if config['supports_profiles']: | |
281 | search_root = os.path.join(config['browser_dir'], profile) | |
282 | else: | |
86e5f3ed | 283 | logger.error(f'{browser_name} does not support profiles') |
982ee69a MB |
284 | search_root = config['browser_dir'] |
285 | ||
97ec5bc5 | 286 | cookie_database_path = _find_most_recently_used_file(search_root, 'Cookies', logger) |
982ee69a | 287 | if cookie_database_path is None: |
86e5f3ed | 288 | raise FileNotFoundError(f'could not find {browser_name} cookies database in "{search_root}"') |
289 | logger.debug(f'Extracting cookies from: "{cookie_database_path}"') | |
982ee69a | 290 | |
f59f5ef8 | 291 | decryptor = get_cookie_decryptor(config['browser_dir'], config['keyring_name'], logger, keyring=keyring) |
982ee69a | 292 | |
0930b11f | 293 | with tempfile.TemporaryDirectory(prefix='yt_dlp') as tmpdir: |
982ee69a MB |
294 | cursor = None |
295 | try: | |
296 | cursor = _open_database_copy(cookie_database_path, tmpdir) | |
297 | cursor.connection.text_factory = bytes | |
298 | column_names = _get_column_names(cursor, 'cookies') | |
299 | secure_column = 'is_secure' if 'is_secure' in column_names else 'secure' | |
19a03940 | 300 | cursor.execute(f'SELECT host_key, name, value, encrypted_value, path, expires_utc, {secure_column} FROM cookies') |
982ee69a MB |
301 | jar = YoutubeDLCookieJar() |
302 | failed_cookies = 0 | |
f59f5ef8 | 303 | unencrypted_cookies = 0 |
97ec5bc5 | 304 | with _create_progress_bar(logger) as progress_bar: |
305 | table = cursor.fetchall() | |
306 | total_cookie_count = len(table) | |
307 | for i, line in enumerate(table): | |
308 | progress_bar.print(f'Loading cookie {i: 6d}/{total_cookie_count: 6d}') | |
309 | is_encrypted, cookie = _process_chrome_cookie(decryptor, *line) | |
310 | if not cookie: | |
982ee69a MB |
311 | failed_cookies += 1 |
312 | continue | |
97ec5bc5 | 313 | elif not is_encrypted: |
314 | unencrypted_cookies += 1 | |
315 | jar.set_cookie(cookie) | |
982ee69a | 316 | if failed_cookies > 0: |
86e5f3ed | 317 | failed_message = f' ({failed_cookies} could not be decrypted)' |
982ee69a MB |
318 | else: |
319 | failed_message = '' | |
86e5f3ed | 320 | logger.info(f'Extracted {len(jar)} cookies from {browser_name}{failed_message}') |
24146491 | 321 | counts = decryptor._cookie_counts.copy() |
f59f5ef8 | 322 | counts['unencrypted'] = unencrypted_cookies |
86e5f3ed | 323 | logger.debug(f'cookie version breakdown: {counts}') |
982ee69a MB |
324 | return jar |
325 | finally: | |
326 | if cursor is not None: | |
327 | cursor.connection.close() | |
328 | ||
329 | ||
97ec5bc5 | 330 | def _process_chrome_cookie(decryptor, host_key, name, value, encrypted_value, path, expires_utc, is_secure): |
0f06bcd7 | 331 | host_key = host_key.decode() |
332 | name = name.decode() | |
333 | value = value.decode() | |
334 | path = path.decode() | |
97ec5bc5 | 335 | is_encrypted = not value and encrypted_value |
336 | ||
337 | if is_encrypted: | |
338 | value = decryptor.decrypt(encrypted_value) | |
339 | if value is None: | |
340 | return is_encrypted, None | |
341 | ||
ac668111 | 342 | return is_encrypted, http.cookiejar.Cookie( |
97ec5bc5 | 343 | version=0, name=name, value=value, port=None, port_specified=False, |
344 | domain=host_key, domain_specified=bool(host_key), domain_initial_dot=host_key.startswith('.'), | |
345 | path=path, path_specified=bool(path), secure=is_secure, expires=expires_utc, discard=False, | |
346 | comment=None, comment_url=None, rest={}) | |
347 | ||
348 | ||
982ee69a MB |
349 | class ChromeCookieDecryptor: |
350 | """ | |
351 | Overview: | |
352 | ||
353 | Linux: | |
354 | - cookies are either v10 or v11 | |
355 | - v10: AES-CBC encrypted with a fixed key | |
356 | - v11: AES-CBC encrypted with an OS protected key (keyring) | |
357 | - v11 keys can be stored in various places depending on the activate desktop environment [2] | |
358 | ||
359 | Mac: | |
360 | - cookies are either v10 or not v10 | |
361 | - v10: AES-CBC encrypted with an OS protected key (keyring) and more key derivation iterations than linux | |
362 | - not v10: 'old data' stored as plaintext | |
363 | ||
364 | Windows: | |
365 | - cookies are either v10 or not v10 | |
366 | - v10: AES-GCM encrypted with a key which is encrypted with DPAPI | |
367 | - not v10: encrypted with DPAPI | |
368 | ||
369 | Sources: | |
370 | - [1] https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/ | |
371 | - [2] https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/key_storage_linux.cc | |
372 | - KeyStorageLinux::CreateService | |
373 | """ | |
374 | ||
0fa7d2c8 | 375 | _cookie_counts = {} |
982ee69a | 376 | |
0fa7d2c8 | 377 | def decrypt(self, encrypted_value): |
19a03940 | 378 | raise NotImplementedError('Must be implemented by sub classes') |
f59f5ef8 | 379 | |
982ee69a | 380 | |
f59f5ef8 | 381 | def get_cookie_decryptor(browser_root, browser_keyring_name, logger, *, keyring=None): |
56ba69e4 | 382 | if sys.platform == 'darwin': |
982ee69a | 383 | return MacChromeCookieDecryptor(browser_keyring_name, logger) |
56ba69e4 | 384 | elif sys.platform in ('win32', 'cygwin'): |
982ee69a | 385 | return WindowsChromeCookieDecryptor(browser_root, logger) |
56ba69e4 | 386 | return LinuxChromeCookieDecryptor(browser_keyring_name, logger, keyring=keyring) |
982ee69a MB |
387 | |
388 | ||
389 | class LinuxChromeCookieDecryptor(ChromeCookieDecryptor): | |
f59f5ef8 | 390 | def __init__(self, browser_keyring_name, logger, *, keyring=None): |
982ee69a MB |
391 | self._logger = logger |
392 | self._v10_key = self.derive_key(b'peanuts') | |
f59f5ef8 | 393 | self._cookie_counts = {'v10': 0, 'v11': 0, 'other': 0} |
9b7a48ab SS |
394 | self._browser_keyring_name = browser_keyring_name |
395 | self._keyring = keyring | |
396 | ||
397 | @functools.cached_property | |
398 | def _v11_key(self): | |
399 | password = _get_linux_keyring_password(self._browser_keyring_name, self._keyring, self._logger) | |
400 | return None if password is None else self.derive_key(password) | |
982ee69a MB |
401 | |
402 | @staticmethod | |
403 | def derive_key(password): | |
404 | # values from | |
405 | # https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/os_crypt_linux.cc | |
406 | return pbkdf2_sha1(password, salt=b'saltysalt', iterations=1, key_length=16) | |
407 | ||
408 | def decrypt(self, encrypted_value): | |
409 | version = encrypted_value[:3] | |
410 | ciphertext = encrypted_value[3:] | |
411 | ||
412 | if version == b'v10': | |
f59f5ef8 | 413 | self._cookie_counts['v10'] += 1 |
982ee69a MB |
414 | return _decrypt_aes_cbc(ciphertext, self._v10_key, self._logger) |
415 | ||
416 | elif version == b'v11': | |
f59f5ef8 | 417 | self._cookie_counts['v11'] += 1 |
982ee69a | 418 | if self._v11_key is None: |
f59f5ef8 | 419 | self._logger.warning('cannot decrypt v11 cookies: no key found', only_once=True) |
982ee69a MB |
420 | return None |
421 | return _decrypt_aes_cbc(ciphertext, self._v11_key, self._logger) | |
422 | ||
423 | else: | |
f59f5ef8 | 424 | self._cookie_counts['other'] += 1 |
982ee69a MB |
425 | return None |
426 | ||
427 | ||
428 | class MacChromeCookieDecryptor(ChromeCookieDecryptor): | |
429 | def __init__(self, browser_keyring_name, logger): | |
430 | self._logger = logger | |
f440b14f | 431 | password = _get_mac_keyring_password(browser_keyring_name, logger) |
982ee69a | 432 | self._v10_key = None if password is None else self.derive_key(password) |
f59f5ef8 | 433 | self._cookie_counts = {'v10': 0, 'other': 0} |
982ee69a MB |
434 | |
435 | @staticmethod | |
436 | def derive_key(password): | |
437 | # values from | |
438 | # https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/os_crypt_mac.mm | |
439 | return pbkdf2_sha1(password, salt=b'saltysalt', iterations=1003, key_length=16) | |
440 | ||
441 | def decrypt(self, encrypted_value): | |
442 | version = encrypted_value[:3] | |
443 | ciphertext = encrypted_value[3:] | |
444 | ||
445 | if version == b'v10': | |
f59f5ef8 | 446 | self._cookie_counts['v10'] += 1 |
982ee69a MB |
447 | if self._v10_key is None: |
448 | self._logger.warning('cannot decrypt v10 cookies: no key found', only_once=True) | |
449 | return None | |
450 | ||
451 | return _decrypt_aes_cbc(ciphertext, self._v10_key, self._logger) | |
452 | ||
453 | else: | |
f59f5ef8 | 454 | self._cookie_counts['other'] += 1 |
982ee69a MB |
455 | # other prefixes are considered 'old data' which were stored as plaintext |
456 | # https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/os_crypt_mac.mm | |
457 | return encrypted_value | |
458 | ||
459 | ||
460 | class WindowsChromeCookieDecryptor(ChromeCookieDecryptor): | |
461 | def __init__(self, browser_root, logger): | |
462 | self._logger = logger | |
463 | self._v10_key = _get_windows_v10_key(browser_root, logger) | |
f59f5ef8 MB |
464 | self._cookie_counts = {'v10': 0, 'other': 0} |
465 | ||
982ee69a MB |
466 | def decrypt(self, encrypted_value): |
467 | version = encrypted_value[:3] | |
468 | ciphertext = encrypted_value[3:] | |
469 | ||
470 | if version == b'v10': | |
f59f5ef8 | 471 | self._cookie_counts['v10'] += 1 |
982ee69a MB |
472 | if self._v10_key is None: |
473 | self._logger.warning('cannot decrypt v10 cookies: no key found', only_once=True) | |
474 | return None | |
982ee69a MB |
475 | |
476 | # https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/os_crypt_win.cc | |
477 | # kNonceLength | |
478 | nonce_length = 96 // 8 | |
479 | # boringssl | |
480 | # EVP_AEAD_AES_GCM_TAG_LEN | |
481 | authentication_tag_length = 16 | |
482 | ||
483 | raw_ciphertext = ciphertext | |
484 | nonce = raw_ciphertext[:nonce_length] | |
485 | ciphertext = raw_ciphertext[nonce_length:-authentication_tag_length] | |
486 | authentication_tag = raw_ciphertext[-authentication_tag_length:] | |
487 | ||
488 | return _decrypt_aes_gcm(ciphertext, self._v10_key, nonce, authentication_tag, self._logger) | |
489 | ||
490 | else: | |
f59f5ef8 | 491 | self._cookie_counts['other'] += 1 |
982ee69a MB |
492 | # any other prefix means the data is DPAPI encrypted |
493 | # https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/os_crypt_win.cc | |
0f06bcd7 | 494 | return _decrypt_windows_dpapi(encrypted_value, self._logger).decode() |
982ee69a MB |
495 | |
496 | ||
497 | def _extract_safari_cookies(profile, logger): | |
498 | if profile is not None: | |
499 | logger.error('safari does not support profiles') | |
500 | if sys.platform != 'darwin': | |
86e5f3ed | 501 | raise ValueError(f'unsupported platform: {sys.platform}') |
982ee69a MB |
502 | |
503 | cookies_path = os.path.expanduser('~/Library/Cookies/Cookies.binarycookies') | |
504 | ||
505 | if not os.path.isfile(cookies_path): | |
1f7db853 MP |
506 | logger.debug('Trying secondary cookie location') |
507 | cookies_path = os.path.expanduser('~/Library/Containers/com.apple.Safari/Data/Library/Cookies/Cookies.binarycookies') | |
508 | if not os.path.isfile(cookies_path): | |
509 | raise FileNotFoundError('could not find safari cookies database') | |
982ee69a MB |
510 | |
511 | with open(cookies_path, 'rb') as f: | |
512 | cookies_data = f.read() | |
513 | ||
514 | jar = parse_safari_cookies(cookies_data, logger=logger) | |
86e5f3ed | 515 | logger.info(f'Extracted {len(jar)} cookies from safari') |
982ee69a MB |
516 | return jar |
517 | ||
518 | ||
519 | class ParserError(Exception): | |
520 | pass | |
521 | ||
522 | ||
523 | class DataParser: | |
524 | def __init__(self, data, logger): | |
525 | self._data = data | |
526 | self.cursor = 0 | |
527 | self._logger = logger | |
528 | ||
529 | def read_bytes(self, num_bytes): | |
530 | if num_bytes < 0: | |
86e5f3ed | 531 | raise ParserError(f'invalid read of {num_bytes} bytes') |
982ee69a MB |
532 | end = self.cursor + num_bytes |
533 | if end > len(self._data): | |
534 | raise ParserError('reached end of input') | |
535 | data = self._data[self.cursor:end] | |
536 | self.cursor = end | |
537 | return data | |
538 | ||
539 | def expect_bytes(self, expected_value, message): | |
540 | value = self.read_bytes(len(expected_value)) | |
541 | if value != expected_value: | |
86e5f3ed | 542 | raise ParserError(f'unexpected value: {value} != {expected_value} ({message})') |
982ee69a MB |
543 | |
544 | def read_uint(self, big_endian=False): | |
545 | data_format = '>I' if big_endian else '<I' | |
546 | return struct.unpack(data_format, self.read_bytes(4))[0] | |
547 | ||
548 | def read_double(self, big_endian=False): | |
549 | data_format = '>d' if big_endian else '<d' | |
550 | return struct.unpack(data_format, self.read_bytes(8))[0] | |
551 | ||
552 | def read_cstring(self): | |
553 | buffer = [] | |
554 | while True: | |
555 | c = self.read_bytes(1) | |
556 | if c == b'\x00': | |
0f06bcd7 | 557 | return b''.join(buffer).decode() |
982ee69a MB |
558 | else: |
559 | buffer.append(c) | |
560 | ||
561 | def skip(self, num_bytes, description='unknown'): | |
562 | if num_bytes > 0: | |
19a03940 | 563 | self._logger.debug(f'skipping {num_bytes} bytes ({description}): {self.read_bytes(num_bytes)!r}') |
982ee69a | 564 | elif num_bytes < 0: |
86e5f3ed | 565 | raise ParserError(f'invalid skip of {num_bytes} bytes') |
982ee69a MB |
566 | |
567 | def skip_to(self, offset, description='unknown'): | |
568 | self.skip(offset - self.cursor, description) | |
569 | ||
570 | def skip_to_end(self, description='unknown'): | |
571 | self.skip_to(len(self._data), description) | |
572 | ||
573 | ||
574 | def _mac_absolute_time_to_posix(timestamp): | |
575 | return int((datetime(2001, 1, 1, 0, 0, tzinfo=timezone.utc) + timedelta(seconds=timestamp)).timestamp()) | |
576 | ||
577 | ||
578 | def _parse_safari_cookies_header(data, logger): | |
579 | p = DataParser(data, logger) | |
580 | p.expect_bytes(b'cook', 'database signature') | |
581 | number_of_pages = p.read_uint(big_endian=True) | |
582 | page_sizes = [p.read_uint(big_endian=True) for _ in range(number_of_pages)] | |
583 | return page_sizes, p.cursor | |
584 | ||
585 | ||
586 | def _parse_safari_cookies_page(data, jar, logger): | |
587 | p = DataParser(data, logger) | |
588 | p.expect_bytes(b'\x00\x00\x01\x00', 'page signature') | |
589 | number_of_cookies = p.read_uint() | |
590 | record_offsets = [p.read_uint() for _ in range(number_of_cookies)] | |
591 | if number_of_cookies == 0: | |
86e5f3ed | 592 | logger.debug(f'a cookies page of size {len(data)} has no cookies') |
982ee69a MB |
593 | return |
594 | ||
595 | p.skip_to(record_offsets[0], 'unknown page header field') | |
596 | ||
97ec5bc5 | 597 | with _create_progress_bar(logger) as progress_bar: |
598 | for i, record_offset in enumerate(record_offsets): | |
599 | progress_bar.print(f'Loading cookie {i: 6d}/{number_of_cookies: 6d}') | |
600 | p.skip_to(record_offset, 'space between records') | |
601 | record_length = _parse_safari_cookies_record(data[record_offset:], jar, logger) | |
602 | p.read_bytes(record_length) | |
982ee69a MB |
603 | p.skip_to_end('space in between pages') |
604 | ||
605 | ||
606 | def _parse_safari_cookies_record(data, jar, logger): | |
607 | p = DataParser(data, logger) | |
608 | record_size = p.read_uint() | |
609 | p.skip(4, 'unknown record field 1') | |
610 | flags = p.read_uint() | |
611 | is_secure = bool(flags & 0x0001) | |
612 | p.skip(4, 'unknown record field 2') | |
613 | domain_offset = p.read_uint() | |
614 | name_offset = p.read_uint() | |
615 | path_offset = p.read_uint() | |
616 | value_offset = p.read_uint() | |
617 | p.skip(8, 'unknown record field 3') | |
618 | expiration_date = _mac_absolute_time_to_posix(p.read_double()) | |
619 | _creation_date = _mac_absolute_time_to_posix(p.read_double()) # noqa: F841 | |
620 | ||
621 | try: | |
622 | p.skip_to(domain_offset) | |
623 | domain = p.read_cstring() | |
624 | ||
625 | p.skip_to(name_offset) | |
626 | name = p.read_cstring() | |
627 | ||
628 | p.skip_to(path_offset) | |
629 | path = p.read_cstring() | |
630 | ||
631 | p.skip_to(value_offset) | |
632 | value = p.read_cstring() | |
633 | except UnicodeDecodeError: | |
f440b14f | 634 | logger.warning('failed to parse Safari cookie because UTF-8 decoding failed', only_once=True) |
982ee69a MB |
635 | return record_size |
636 | ||
637 | p.skip_to(record_size, 'space at the end of the record') | |
638 | ||
ac668111 | 639 | cookie = http.cookiejar.Cookie( |
982ee69a MB |
640 | version=0, name=name, value=value, port=None, port_specified=False, |
641 | domain=domain, domain_specified=bool(domain), domain_initial_dot=domain.startswith('.'), | |
642 | path=path, path_specified=bool(path), secure=is_secure, expires=expiration_date, discard=False, | |
643 | comment=None, comment_url=None, rest={}) | |
644 | jar.set_cookie(cookie) | |
645 | return record_size | |
646 | ||
647 | ||
648 | def parse_safari_cookies(data, jar=None, logger=YDLLogger()): | |
649 | """ | |
650 | References: | |
651 | - https://github.com/libyal/dtformats/blob/main/documentation/Safari%20Cookies.asciidoc | |
652 | - this data appears to be out of date but the important parts of the database structure is the same | |
653 | - there are a few bytes here and there which are skipped during parsing | |
654 | """ | |
655 | if jar is None: | |
656 | jar = YoutubeDLCookieJar() | |
657 | page_sizes, body_start = _parse_safari_cookies_header(data, logger) | |
658 | p = DataParser(data[body_start:], logger) | |
659 | for page_size in page_sizes: | |
660 | _parse_safari_cookies_page(p.read_bytes(page_size), jar, logger) | |
661 | p.skip_to_end('footer') | |
662 | return jar | |
663 | ||
664 | ||
f59f5ef8 MB |
665 | class _LinuxDesktopEnvironment(Enum): |
666 | """ | |
667 | https://chromium.googlesource.com/chromium/src/+/refs/heads/main/base/nix/xdg_util.h | |
668 | DesktopEnvironment | |
669 | """ | |
670 | OTHER = auto() | |
671 | CINNAMON = auto() | |
672 | GNOME = auto() | |
673 | KDE = auto() | |
674 | PANTHEON = auto() | |
675 | UNITY = auto() | |
676 | XFCE = auto() | |
982ee69a MB |
677 | |
678 | ||
f59f5ef8 MB |
679 | class _LinuxKeyring(Enum): |
680 | """ | |
681 | https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/key_storage_util_linux.h | |
682 | SelectedLinuxBackend | |
683 | """ | |
684 | KWALLET = auto() | |
685 | GNOMEKEYRING = auto() | |
686 | BASICTEXT = auto() | |
687 | ||
688 | ||
689 | SUPPORTED_KEYRINGS = _LinuxKeyring.__members__.keys() | |
690 | ||
691 | ||
692 | def _get_linux_desktop_environment(env): | |
693 | """ | |
694 | https://chromium.googlesource.com/chromium/src/+/refs/heads/main/base/nix/xdg_util.cc | |
695 | GetDesktopEnvironment | |
696 | """ | |
697 | xdg_current_desktop = env.get('XDG_CURRENT_DESKTOP', None) | |
698 | desktop_session = env.get('DESKTOP_SESSION', None) | |
699 | if xdg_current_desktop is not None: | |
700 | xdg_current_desktop = xdg_current_desktop.split(':')[0].strip() | |
701 | ||
702 | if xdg_current_desktop == 'Unity': | |
703 | if desktop_session is not None and 'gnome-fallback' in desktop_session: | |
704 | return _LinuxDesktopEnvironment.GNOME | |
705 | else: | |
706 | return _LinuxDesktopEnvironment.UNITY | |
707 | elif xdg_current_desktop == 'GNOME': | |
708 | return _LinuxDesktopEnvironment.GNOME | |
709 | elif xdg_current_desktop == 'X-Cinnamon': | |
710 | return _LinuxDesktopEnvironment.CINNAMON | |
711 | elif xdg_current_desktop == 'KDE': | |
712 | return _LinuxDesktopEnvironment.KDE | |
713 | elif xdg_current_desktop == 'Pantheon': | |
714 | return _LinuxDesktopEnvironment.PANTHEON | |
715 | elif xdg_current_desktop == 'XFCE': | |
716 | return _LinuxDesktopEnvironment.XFCE | |
717 | elif desktop_session is not None: | |
718 | if desktop_session in ('mate', 'gnome'): | |
719 | return _LinuxDesktopEnvironment.GNOME | |
720 | elif 'kde' in desktop_session: | |
721 | return _LinuxDesktopEnvironment.KDE | |
722 | elif 'xfce' in desktop_session: | |
723 | return _LinuxDesktopEnvironment.XFCE | |
724 | else: | |
725 | if 'GNOME_DESKTOP_SESSION_ID' in env: | |
726 | return _LinuxDesktopEnvironment.GNOME | |
727 | elif 'KDE_FULL_SESSION' in env: | |
728 | return _LinuxDesktopEnvironment.KDE | |
fa8fd951 | 729 | return _LinuxDesktopEnvironment.OTHER |
f59f5ef8 MB |
730 | |
731 | ||
732 | def _choose_linux_keyring(logger): | |
733 | """ | |
734 | https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/key_storage_util_linux.cc | |
735 | SelectBackend | |
736 | """ | |
737 | desktop_environment = _get_linux_desktop_environment(os.environ) | |
86e5f3ed | 738 | logger.debug(f'detected desktop environment: {desktop_environment.name}') |
f59f5ef8 MB |
739 | if desktop_environment == _LinuxDesktopEnvironment.KDE: |
740 | linux_keyring = _LinuxKeyring.KWALLET | |
741 | elif desktop_environment == _LinuxDesktopEnvironment.OTHER: | |
742 | linux_keyring = _LinuxKeyring.BASICTEXT | |
982ee69a | 743 | else: |
f59f5ef8 MB |
744 | linux_keyring = _LinuxKeyring.GNOMEKEYRING |
745 | return linux_keyring | |
746 | ||
747 | ||
748 | def _get_kwallet_network_wallet(logger): | |
749 | """ The name of the wallet used to store network passwords. | |
750 | ||
751 | https://chromium.googlesource.com/chromium/src/+/refs/heads/main/components/os_crypt/kwallet_dbus.cc | |
752 | KWalletDBus::NetworkWallet | |
753 | which does a dbus call to the following function: | |
754 | https://api.kde.org/frameworks/kwallet/html/classKWallet_1_1Wallet.html | |
755 | Wallet::NetworkWallet | |
756 | """ | |
757 | default_wallet = 'kdewallet' | |
758 | try: | |
f0c9fb96 | 759 | stdout, _, returncode = Popen.run([ |
f59f5ef8 MB |
760 | 'dbus-send', '--session', '--print-reply=literal', |
761 | '--dest=org.kde.kwalletd5', | |
762 | '/modules/kwalletd5', | |
763 | 'org.kde.KWallet.networkWallet' | |
f0c9fb96 | 764 | ], text=True, stdout=subprocess.PIPE, stderr=subprocess.DEVNULL) |
f59f5ef8 | 765 | |
f0c9fb96 | 766 | if returncode: |
f59f5ef8 MB |
767 | logger.warning('failed to read NetworkWallet') |
768 | return default_wallet | |
769 | else: | |
f0c9fb96 | 770 | logger.debug(f'NetworkWallet = "{stdout.strip()}"') |
771 | return stdout.strip() | |
a44ca5a4 | 772 | except Exception as e: |
86e5f3ed | 773 | logger.warning(f'exception while obtaining NetworkWallet: {e}') |
f59f5ef8 MB |
774 | return default_wallet |
775 | ||
776 | ||
777 | def _get_kwallet_password(browser_keyring_name, logger): | |
778 | logger.debug('using kwallet-query to obtain password from kwallet') | |
779 | ||
780 | if shutil.which('kwallet-query') is None: | |
781 | logger.error('kwallet-query command not found. KWallet and kwallet-query ' | |
782 | 'must be installed to read from KWallet. kwallet-query should be' | |
783 | 'included in the kwallet package for your distribution') | |
784 | return b'' | |
785 | ||
786 | network_wallet = _get_kwallet_network_wallet(logger) | |
787 | ||
788 | try: | |
f0c9fb96 | 789 | stdout, _, returncode = Popen.run([ |
f59f5ef8 | 790 | 'kwallet-query', |
86e5f3ed | 791 | '--read-password', f'{browser_keyring_name} Safe Storage', |
792 | '--folder', f'{browser_keyring_name} Keys', | |
f59f5ef8 MB |
793 | network_wallet |
794 | ], stdout=subprocess.PIPE, stderr=subprocess.DEVNULL) | |
795 | ||
f0c9fb96 | 796 | if returncode: |
797 | logger.error(f'kwallet-query failed with return code {returncode}. ' | |
798 | 'Please consult the kwallet-query man page for details') | |
f59f5ef8 MB |
799 | return b'' |
800 | else: | |
801 | if stdout.lower().startswith(b'failed to read'): | |
802 | logger.debug('failed to read password from kwallet. Using empty string instead') | |
803 | # this sometimes occurs in KDE because chrome does not check hasEntry and instead | |
804 | # just tries to read the value (which kwallet returns "") whereas kwallet-query | |
805 | # checks hasEntry. To verify this: | |
806 | # dbus-monitor "interface='org.kde.KWallet'" "type=method_return" | |
807 | # while starting chrome. | |
808 | # this may be a bug as the intended behaviour is to generate a random password and store | |
809 | # it, but that doesn't matter here. | |
810 | return b'' | |
811 | else: | |
812 | logger.debug('password found') | |
f0c9fb96 | 813 | return stdout.rstrip(b'\n') |
a44ca5a4 | 814 | except Exception as e: |
815 | logger.warning(f'exception running kwallet-query: {error_to_str(e)}') | |
f59f5ef8 MB |
816 | return b'' |
817 | ||
818 | ||
819 | def _get_gnome_keyring_password(browser_keyring_name, logger): | |
9b8ee23b | 820 | if not secretstorage: |
821 | logger.error(f'secretstorage not available {_SECRETSTORAGE_UNAVAILABLE_REASON}') | |
f59f5ef8 MB |
822 | return b'' |
823 | # the Gnome keyring does not seem to organise keys in the same way as KWallet, | |
824 | # using `dbus-monitor` during startup, it can be observed that chromium lists all keys | |
825 | # and presumably searches for its key in the list. It appears that we must do the same. | |
826 | # https://github.com/jaraco/keyring/issues/556 | |
827 | with contextlib.closing(secretstorage.dbus_init()) as con: | |
828 | col = secretstorage.get_default_collection(con) | |
829 | for item in col.get_all_items(): | |
86e5f3ed | 830 | if item.get_label() == f'{browser_keyring_name} Safe Storage': |
f59f5ef8 MB |
831 | return item.get_secret() |
832 | else: | |
833 | logger.error('failed to read from keyring') | |
834 | return b'' | |
835 | ||
836 | ||
837 | def _get_linux_keyring_password(browser_keyring_name, keyring, logger): | |
838 | # note: chrome/chromium can be run with the following flags to determine which keyring backend | |
839 | # it has chosen to use | |
840 | # chromium --enable-logging=stderr --v=1 2>&1 | grep key_storage_ | |
841 | # Chromium supports a flag: --password-store=<basic|gnome|kwallet> so the automatic detection | |
842 | # will not be sufficient in all cases. | |
843 | ||
2c539d49 | 844 | keyring = _LinuxKeyring[keyring] if keyring else _choose_linux_keyring(logger) |
f59f5ef8 MB |
845 | logger.debug(f'Chosen keyring: {keyring.name}') |
846 | ||
847 | if keyring == _LinuxKeyring.KWALLET: | |
848 | return _get_kwallet_password(browser_keyring_name, logger) | |
849 | elif keyring == _LinuxKeyring.GNOMEKEYRING: | |
850 | return _get_gnome_keyring_password(browser_keyring_name, logger) | |
851 | elif keyring == _LinuxKeyring.BASICTEXT: | |
852 | # when basic text is chosen, all cookies are stored as v10 (so no keyring password is required) | |
853 | return None | |
854 | assert False, f'Unknown keyring {keyring}' | |
855 | ||
856 | ||
857 | def _get_mac_keyring_password(browser_keyring_name, logger): | |
858 | logger.debug('using find-generic-password to obtain password from OSX keychain') | |
859 | try: | |
46a5b335 | 860 | stdout, _, returncode = Popen.run( |
d3c93ec2 | 861 | ['security', 'find-generic-password', |
862 | '-w', # write password to stdout | |
863 | '-a', browser_keyring_name, # match 'account' | |
86e5f3ed | 864 | '-s', f'{browser_keyring_name} Safe Storage'], # match 'service' |
d3c93ec2 | 865 | stdout=subprocess.PIPE, stderr=subprocess.DEVNULL) |
46a5b335 | 866 | if returncode: |
867 | logger.warning('find-generic-password failed') | |
868 | return None | |
f0c9fb96 | 869 | return stdout.rstrip(b'\n') |
a44ca5a4 | 870 | except Exception as e: |
871 | logger.warning(f'exception running find-generic-password: {error_to_str(e)}') | |
f59f5ef8 | 872 | return None |
982ee69a MB |
873 | |
874 | ||
875 | def _get_windows_v10_key(browser_root, logger): | |
97ec5bc5 | 876 | path = _find_most_recently_used_file(browser_root, 'Local State', logger) |
982ee69a MB |
877 | if path is None: |
878 | logger.error('could not find local state file') | |
879 | return None | |
97ec5bc5 | 880 | logger.debug(f'Found local state file at "{path}"') |
86e5f3ed | 881 | with open(path, encoding='utf8') as f: |
982ee69a MB |
882 | data = json.load(f) |
883 | try: | |
884 | base64_key = data['os_crypt']['encrypted_key'] | |
885 | except KeyError: | |
886 | logger.error('no encrypted key in Local State') | |
887 | return None | |
14f25df2 | 888 | encrypted_key = base64.b64decode(base64_key) |
982ee69a MB |
889 | prefix = b'DPAPI' |
890 | if not encrypted_key.startswith(prefix): | |
891 | logger.error('invalid key') | |
892 | return None | |
893 | return _decrypt_windows_dpapi(encrypted_key[len(prefix):], logger) | |
894 | ||
895 | ||
896 | def pbkdf2_sha1(password, salt, iterations, key_length): | |
897 | return pbkdf2_hmac('sha1', password, salt, iterations, key_length) | |
898 | ||
899 | ||
900 | def _decrypt_aes_cbc(ciphertext, key, logger, initialization_vector=b' ' * 16): | |
1d3586d0 | 901 | plaintext = unpad_pkcs7(aes_cbc_decrypt_bytes(ciphertext, key, initialization_vector)) |
982ee69a | 902 | try: |
0f06bcd7 | 903 | return plaintext.decode() |
982ee69a | 904 | except UnicodeDecodeError: |
f440b14f | 905 | logger.warning('failed to decrypt cookie (AES-CBC) because UTF-8 decoding failed. Possibly the key is wrong?', only_once=True) |
982ee69a MB |
906 | return None |
907 | ||
908 | ||
909 | def _decrypt_aes_gcm(ciphertext, key, nonce, authentication_tag, logger): | |
982ee69a | 910 | try: |
09906f55 | 911 | plaintext = aes_gcm_decrypt_and_verify_bytes(ciphertext, key, authentication_tag, nonce) |
982ee69a | 912 | except ValueError: |
f440b14f | 913 | logger.warning('failed to decrypt cookie (AES-GCM) because the MAC check failed. Possibly the key is wrong?', only_once=True) |
982ee69a MB |
914 | return None |
915 | ||
916 | try: | |
0f06bcd7 | 917 | return plaintext.decode() |
982ee69a | 918 | except UnicodeDecodeError: |
f440b14f | 919 | logger.warning('failed to decrypt cookie (AES-GCM) because UTF-8 decoding failed. Possibly the key is wrong?', only_once=True) |
982ee69a MB |
920 | return None |
921 | ||
922 | ||
923 | def _decrypt_windows_dpapi(ciphertext, logger): | |
924 | """ | |
925 | References: | |
926 | - https://docs.microsoft.com/en-us/windows/win32/api/dpapi/nf-dpapi-cryptunprotectdata | |
927 | """ | |
fe0918bb | 928 | |
929 | import ctypes | |
930 | import ctypes.wintypes | |
982ee69a MB |
931 | |
932 | class DATA_BLOB(ctypes.Structure): | |
fe0918bb | 933 | _fields_ = [('cbData', ctypes.wintypes.DWORD), |
982ee69a MB |
934 | ('pbData', ctypes.POINTER(ctypes.c_char))] |
935 | ||
936 | buffer = ctypes.create_string_buffer(ciphertext) | |
937 | blob_in = DATA_BLOB(ctypes.sizeof(buffer), buffer) | |
938 | blob_out = DATA_BLOB() | |
939 | ret = ctypes.windll.crypt32.CryptUnprotectData( | |
940 | ctypes.byref(blob_in), # pDataIn | |
941 | None, # ppszDataDescr: human readable description of pDataIn | |
942 | None, # pOptionalEntropy: salt? | |
943 | None, # pvReserved: must be NULL | |
944 | None, # pPromptStruct: information about prompts to display | |
945 | 0, # dwFlags | |
946 | ctypes.byref(blob_out) # pDataOut | |
947 | ) | |
948 | if not ret: | |
f9be9cb9 | 949 | logger.warning('failed to decrypt with DPAPI', only_once=True) |
982ee69a MB |
950 | return None |
951 | ||
952 | result = ctypes.string_at(blob_out.pbData, blob_out.cbData) | |
953 | ctypes.windll.kernel32.LocalFree(blob_out.pbData) | |
954 | return result | |
955 | ||
956 | ||
957 | def _config_home(): | |
958 | return os.environ.get('XDG_CONFIG_HOME', os.path.expanduser('~/.config')) | |
959 | ||
960 | ||
961 | def _open_database_copy(database_path, tmpdir): | |
962 | # cannot open sqlite databases if they are already in use (e.g. by the browser) | |
963 | database_copy_path = os.path.join(tmpdir, 'temporary.sqlite') | |
964 | shutil.copy(database_path, database_copy_path) | |
965 | conn = sqlite3.connect(database_copy_path) | |
966 | return conn.cursor() | |
967 | ||
968 | ||
969 | def _get_column_names(cursor, table_name): | |
86e5f3ed | 970 | table_info = cursor.execute(f'PRAGMA table_info({table_name})').fetchall() |
0f06bcd7 | 971 | return [row[1].decode() for row in table_info] |
982ee69a MB |
972 | |
973 | ||
97ec5bc5 | 974 | def _find_most_recently_used_file(root, filename, logger): |
982ee69a | 975 | # if there are multiple browser profiles, take the most recently used one |
97ec5bc5 | 976 | i, paths = 0, [] |
977 | with _create_progress_bar(logger) as progress_bar: | |
978 | for curr_root, dirs, files in os.walk(root): | |
979 | for file in files: | |
980 | i += 1 | |
981 | progress_bar.print(f'Searching for "{filename}": {i: 6d} files searched') | |
982 | if file == filename: | |
983 | paths.append(os.path.join(curr_root, file)) | |
982ee69a MB |
984 | return None if not paths else max(paths, key=lambda path: os.lstat(path).st_mtime) |
985 | ||
986 | ||
987 | def _merge_cookie_jars(jars): | |
988 | output_jar = YoutubeDLCookieJar() | |
989 | for jar in jars: | |
990 | for cookie in jar: | |
991 | output_jar.set_cookie(cookie) | |
992 | if jar.filename is not None: | |
993 | output_jar.filename = jar.filename | |
994 | return output_jar | |
995 | ||
996 | ||
997 | def _is_path(value): | |
998 | return os.path.sep in value | |
999 | ||
1000 | ||
9bd13fe5 | 1001 | def _parse_browser_specification(browser_name, profile=None, keyring=None, container=None): |
982ee69a MB |
1002 | if browser_name not in SUPPORTED_BROWSERS: |
1003 | raise ValueError(f'unsupported browser: "{browser_name}"') | |
f59f5ef8 MB |
1004 | if keyring not in (None, *SUPPORTED_KEYRINGS): |
1005 | raise ValueError(f'unsupported keyring: "{keyring}"') | |
2b24afa6 | 1006 | if profile is not None and _is_path(expand_path(profile)): |
1007 | profile = expand_path(profile) | |
9bd13fe5 | 1008 | return browser_name, profile, keyring, container |
8817a80d SS |
1009 | |
1010 | ||
1011 | class LenientSimpleCookie(http.cookies.SimpleCookie): | |
1012 | """More lenient version of http.cookies.SimpleCookie""" | |
1013 | # From https://github.com/python/cpython/blob/v3.10.7/Lib/http/cookies.py | |
36069409 SS |
1014 | # We use Morsel's legal key chars to avoid errors on setting values |
1015 | _LEGAL_KEY_CHARS = r'\w\d' + re.escape('!#$%&\'*+-.:^_`|~') | |
1016 | _LEGAL_VALUE_CHARS = _LEGAL_KEY_CHARS + re.escape('(),/<=>?@[]{}') | |
8817a80d SS |
1017 | |
1018 | _RESERVED = { | |
1019 | "expires", | |
1020 | "path", | |
1021 | "comment", | |
1022 | "domain", | |
1023 | "max-age", | |
1024 | "secure", | |
1025 | "httponly", | |
1026 | "version", | |
1027 | "samesite", | |
1028 | } | |
1029 | ||
1030 | _FLAGS = {"secure", "httponly"} | |
1031 | ||
1032 | # Added 'bad' group to catch the remaining value | |
1033 | _COOKIE_PATTERN = re.compile(r""" | |
1034 | \s* # Optional whitespace at start of cookie | |
1035 | (?P<key> # Start of group 'key' | |
1036 | [""" + _LEGAL_KEY_CHARS + r"""]+?# Any word of at least one letter | |
1037 | ) # End of group 'key' | |
1038 | ( # Optional group: there may not be a value. | |
1039 | \s*=\s* # Equal Sign | |
1040 | ( # Start of potential value | |
1041 | (?P<val> # Start of group 'val' | |
1042 | "(?:[^\\"]|\\.)*" # Any doublequoted string | |
1043 | | # or | |
1044 | \w{3},\s[\w\d\s-]{9,11}\s[\d:]{8}\sGMT # Special case for "expires" attr | |
1045 | | # or | |
1046 | [""" + _LEGAL_VALUE_CHARS + r"""]* # Any word or empty string | |
1047 | ) # End of group 'val' | |
1048 | | # or | |
1049 | (?P<bad>(?:\\;|[^;])*?) # 'bad' group fallback for invalid values | |
1050 | ) # End of potential value | |
1051 | )? # End of optional value group | |
1052 | \s* # Any number of spaces. | |
1053 | (\s+|;|$) # Ending either at space, semicolon, or EOS. | |
1054 | """, re.ASCII | re.VERBOSE) | |
1055 | ||
1056 | def load(self, data): | |
1057 | # Workaround for https://github.com/yt-dlp/yt-dlp/issues/4776 | |
1058 | if not isinstance(data, str): | |
1059 | return super().load(data) | |
1060 | ||
1061 | morsel = None | |
36069409 SS |
1062 | for match in self._COOKIE_PATTERN.finditer(data): |
1063 | if match.group('bad'): | |
8817a80d SS |
1064 | morsel = None |
1065 | continue | |
1066 | ||
36069409 | 1067 | key, value = match.group('key', 'val') |
8817a80d | 1068 | |
36069409 SS |
1069 | is_attribute = False |
1070 | if key.startswith('$'): | |
1071 | key = key[1:] | |
1072 | is_attribute = True | |
8817a80d SS |
1073 | |
1074 | lower_key = key.lower() | |
1075 | if lower_key in self._RESERVED: | |
1076 | if morsel is None: | |
1077 | continue | |
1078 | ||
1079 | if value is None: | |
1080 | if lower_key not in self._FLAGS: | |
1081 | morsel = None | |
1082 | continue | |
1083 | value = True | |
1084 | else: | |
1085 | value, _ = self.value_decode(value) | |
1086 | ||
1087 | morsel[key] = value | |
1088 | ||
36069409 SS |
1089 | elif is_attribute: |
1090 | morsel = None | |
1091 | ||
8817a80d SS |
1092 | elif value is not None: |
1093 | morsel = self.get(key, http.cookies.Morsel()) | |
1094 | real_value, coded_value = self.value_decode(value) | |
1095 | morsel.set(key, real_value, coded_value) | |
1096 | self[key] = morsel | |
1097 | ||
1098 | else: | |
1099 | morsel = None | |
b87e01c1 | 1100 | |
1101 | ||
1102 | class YoutubeDLCookieJar(http.cookiejar.MozillaCookieJar): | |
1103 | """ | |
1104 | See [1] for cookie file format. | |
1105 | ||
1106 | 1. https://curl.haxx.se/docs/http-cookies.html | |
1107 | """ | |
1108 | _HTTPONLY_PREFIX = '#HttpOnly_' | |
1109 | _ENTRY_LEN = 7 | |
1110 | _HEADER = '''# Netscape HTTP Cookie File | |
1111 | # This file is generated by yt-dlp. Do not edit. | |
1112 | ||
1113 | ''' | |
1114 | _CookieFileEntry = collections.namedtuple( | |
1115 | 'CookieFileEntry', | |
1116 | ('domain_name', 'include_subdomains', 'path', 'https_only', 'expires_at', 'name', 'value')) | |
1117 | ||
1118 | def __init__(self, filename=None, *args, **kwargs): | |
1119 | super().__init__(None, *args, **kwargs) | |
1120 | if is_path_like(filename): | |
1121 | filename = os.fspath(filename) | |
1122 | self.filename = filename | |
1123 | ||
1124 | @staticmethod | |
1125 | def _true_or_false(cndn): | |
1126 | return 'TRUE' if cndn else 'FALSE' | |
1127 | ||
1128 | @contextlib.contextmanager | |
1129 | def open(self, file, *, write=False): | |
1130 | if is_path_like(file): | |
1131 | with open(file, 'w' if write else 'r', encoding='utf-8') as f: | |
1132 | yield f | |
1133 | else: | |
1134 | if write: | |
1135 | file.truncate(0) | |
1136 | yield file | |
1137 | ||
1138 | def _really_save(self, f, ignore_discard=False, ignore_expires=False): | |
1139 | now = time.time() | |
1140 | for cookie in self: | |
1141 | if (not ignore_discard and cookie.discard | |
1142 | or not ignore_expires and cookie.is_expired(now)): | |
1143 | continue | |
1144 | name, value = cookie.name, cookie.value | |
1145 | if value is None: | |
1146 | # cookies.txt regards 'Set-Cookie: foo' as a cookie | |
1147 | # with no name, whereas http.cookiejar regards it as a | |
1148 | # cookie with no value. | |
1149 | name, value = '', name | |
1150 | f.write('%s\n' % '\t'.join(( | |
1151 | cookie.domain, | |
1152 | self._true_or_false(cookie.domain.startswith('.')), | |
1153 | cookie.path, | |
1154 | self._true_or_false(cookie.secure), | |
1155 | str_or_none(cookie.expires, default=''), | |
1156 | name, value | |
1157 | ))) | |
1158 | ||
1159 | def save(self, filename=None, *args, **kwargs): | |
1160 | """ | |
1161 | Save cookies to a file. | |
1162 | Code is taken from CPython 3.6 | |
1163 | https://github.com/python/cpython/blob/8d999cbf4adea053be6dbb612b9844635c4dfb8e/Lib/http/cookiejar.py#L2091-L2117 """ | |
1164 | ||
1165 | if filename is None: | |
1166 | if self.filename is not None: | |
1167 | filename = self.filename | |
1168 | else: | |
1169 | raise ValueError(http.cookiejar.MISSING_FILENAME_TEXT) | |
1170 | ||
1171 | # Store session cookies with `expires` set to 0 instead of an empty string | |
1172 | for cookie in self: | |
1173 | if cookie.expires is None: | |
1174 | cookie.expires = 0 | |
1175 | ||
1176 | with self.open(filename, write=True) as f: | |
1177 | f.write(self._HEADER) | |
1178 | self._really_save(f, *args, **kwargs) | |
1179 | ||
1180 | def load(self, filename=None, ignore_discard=False, ignore_expires=False): | |
1181 | """Load cookies from a file.""" | |
1182 | if filename is None: | |
1183 | if self.filename is not None: | |
1184 | filename = self.filename | |
1185 | else: | |
1186 | raise ValueError(http.cookiejar.MISSING_FILENAME_TEXT) | |
1187 | ||
1188 | def prepare_line(line): | |
1189 | if line.startswith(self._HTTPONLY_PREFIX): | |
1190 | line = line[len(self._HTTPONLY_PREFIX):] | |
1191 | # comments and empty lines are fine | |
1192 | if line.startswith('#') or not line.strip(): | |
1193 | return line | |
1194 | cookie_list = line.split('\t') | |
1195 | if len(cookie_list) != self._ENTRY_LEN: | |
1196 | raise http.cookiejar.LoadError('invalid length %d' % len(cookie_list)) | |
1197 | cookie = self._CookieFileEntry(*cookie_list) | |
1198 | if cookie.expires_at and not cookie.expires_at.isdigit(): | |
1199 | raise http.cookiejar.LoadError('invalid expires at %s' % cookie.expires_at) | |
1200 | return line | |
1201 | ||
1202 | cf = io.StringIO() | |
1203 | with self.open(filename) as f: | |
1204 | for line in f: | |
1205 | try: | |
1206 | cf.write(prepare_line(line)) | |
1207 | except http.cookiejar.LoadError as e: | |
1208 | if f'{line.strip()} '[0] in '[{"': | |
1209 | raise http.cookiejar.LoadError( | |
1210 | 'Cookies file must be Netscape formatted, not JSON. See ' | |
1211 | 'https://github.com/yt-dlp/yt-dlp/wiki/FAQ#how-do-i-pass-cookies-to-yt-dlp') | |
1212 | write_string(f'WARNING: skipping cookie file entry due to {e}: {line!r}\n') | |
1213 | continue | |
1214 | cf.seek(0) | |
1215 | self._really_load(cf, filename, ignore_discard, ignore_expires) | |
1216 | # Session cookies are denoted by either `expires` field set to | |
1217 | # an empty string or 0. MozillaCookieJar only recognizes the former | |
1218 | # (see [1]). So we need force the latter to be recognized as session | |
1219 | # cookies on our own. | |
1220 | # Session cookies may be important for cookies-based authentication, | |
1221 | # e.g. usually, when user does not check 'Remember me' check box while | |
1222 | # logging in on a site, some important cookies are stored as session | |
1223 | # cookies so that not recognizing them will result in failed login. | |
1224 | # 1. https://bugs.python.org/issue17164 | |
1225 | for cookie in self: | |
1226 | # Treat `expires=0` cookies as session cookies | |
1227 | if cookie.expires == 0: | |
1228 | cookie.expires = None | |
1229 | cookie.discard = True | |
1230 | ||
1231 | def get_cookie_header(self, url): | |
1232 | """Generate a Cookie HTTP header for a given url""" | |
1233 | cookie_req = urllib.request.Request(escape_url(sanitize_url(url))) | |
1234 | self.add_cookie_header(cookie_req) | |
1235 | return cookie_req.get_header('Cookie') |