Merge 526ef898bf into 37cea84f77

* from https://github.com/yt-dlp/yt-dlp/security/advisories/GHSA-79w7-vh3h-8g4, thx grub4k

test/test_utils.py

@@ -14,9 +14,11 @@ sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
 import io
 import itertools
 import json
+import types
 import xml.etree.ElementTree
 
 from youtube_dl.utils import (
+    _UnsafeExtensionError,
     age_restricted,
     args_to_str,
     base_url,
@@ -270,6 +272,27 @@ class TestUtil(unittest.TestCase):
             expand_path('~/%s' % env('YOUTUBE_DL_EXPATH_PATH')),
             '%s/expanded' % compat_getenv('HOME'))
 
+    _uncommon_extensions = [
+        ('exe', 'abc.exe.ext'),
+        ('de', 'abc.de.ext'),
+        ('../.mp4', None),
+        ('..\\.mp4', None),
+    ]
+
+    def assertUnsafeExtension(self, ext=None):
+        assert_raises = self.assertRaises(_UnsafeExtensionError)
+        assert_raises.ext = ext
+        orig_exit = assert_raises.__exit__
+
+        def my_exit(self_, exc_type, exc_val, exc_tb):
+            did_raise = orig_exit(exc_type, exc_val, exc_tb)
+            if did_raise and assert_raises.ext is not None:
+                self.assertEqual(assert_raises.ext, assert_raises.exception.extension, 'Unsafe extension  not as unexpected')
+            return did_raise
+
+        assert_raises.__exit__ = types.MethodType(my_exit, assert_raises)
+        return assert_raises
+
     def test_prepend_extension(self):
         self.assertEqual(prepend_extension('abc.ext', 'temp'), 'abc.temp.ext')
         self.assertEqual(prepend_extension('abc.ext', 'temp', 'ext'), 'abc.temp.ext')
@@ -278,6 +301,19 @@ class TestUtil(unittest.TestCase):
         self.assertEqual(prepend_extension('.abc', 'temp'), '.abc.temp')
         self.assertEqual(prepend_extension('.abc.ext', 'temp'), '.abc.temp.ext')
 
+        # Test uncommon extensions
+        self.assertEqual(prepend_extension('abc.ext', 'bin'), 'abc.bin.ext')
+        for ext, result in self._uncommon_extensions:
+            with self.assertUnsafeExtension(ext):
+                prepend_extension('abc', ext)
+            if result:
+                self.assertEqual(prepend_extension('abc.ext', ext, 'ext'), result)
+            else:
+                with self.assertUnsafeExtension(ext):
+                    prepend_extension('abc.ext', ext, 'ext')
+            with self.assertUnsafeExtension(ext):
+                prepend_extension('abc.unexpected_ext', ext, 'ext')
+
     def test_replace_extension(self):
         self.assertEqual(replace_extension('abc.ext', 'temp'), 'abc.temp')
         self.assertEqual(replace_extension('abc.ext', 'temp', 'ext'), 'abc.temp')
@@ -286,6 +322,16 @@ class TestUtil(unittest.TestCase):
         self.assertEqual(replace_extension('.abc', 'temp'), '.abc.temp')
         self.assertEqual(replace_extension('.abc.ext', 'temp'), '.abc.temp')
 
+        # Test uncommon extensions
+        self.assertEqual(replace_extension('abc.ext', 'bin'), 'abc.unknown_video')
+        for ext, _ in self._uncommon_extensions:
+            with self.assertUnsafeExtension(ext):
+                replace_extension('abc', ext)
+            with self.assertUnsafeExtension(ext):
+                replace_extension('abc.ext', ext, 'ext')
+            with self.assertUnsafeExtension(ext):
+                replace_extension('abc.unexpected_ext', ext, 'ext')
+
     def test_subtitles_filename(self):
         self.assertEqual(subtitles_filename('abc.ext', 'en', 'vtt'), 'abc.en.vtt')
         self.assertEqual(subtitles_filename('abc.ext', 'en', 'vtt', 'ext'), 'abc.en.vtt')

youtube_dl/YoutubeDL.py

@@ -7,6 +7,7 @@ import collections
 import copy
 import datetime
 import errno
+import functools
 import io
 import itertools
 import json
@@ -53,6 +54,7 @@ from .compat import (
     compat_urllib_request_DataHandler,
 )
 from .utils import (
+    _UnsafeExtensionError,
     age_restricted,
     args_to_str,
     bug_reports_message,
@@ -129,6 +131,20 @@ if compat_os_name == 'nt':
     import ctypes
 
 
+def _catch_unsafe_file_extension(func):
+    @functools.wraps(func)
+    def wrapper(self, *args, **kwargs):
+        try:
+            return func(self, *args, **kwargs)
+        except _UnsafeExtensionError as error:
+            self.report_error(
+                '{0} found; to avoid damaging your system, this value is disallowed.'
+                ' If you believe this is an error{1}').format(
+                    error.message, bug_reports_message(','))
+
+    return wrapper
+
+
 class YoutubeDL(object):
     """YoutubeDL class.
 
@@ -1925,6 +1941,7 @@ class YoutubeDL(object):
         if self.params.get('forcejson', False):
             self.to_stdout(json.dumps(self.sanitize_info(info_dict)))
 
+    @_catch_unsafe_file_extension
     def process_info(self, info_dict):
         """Process a single resolved IE result."""
 

youtube_dl/__init__.py

@@ -21,6 +21,7 @@ from .compat import (
     workaround_optparse_bug9161,
 )
 from .utils import (
+    _UnsafeExtensionError,
     DateRange,
     decodeOption,
     DEFAULT_OUTTMPL,
@@ -173,6 +174,9 @@ def _real_main(argv=None):
     if opts.ap_mso and opts.ap_mso not in MSO_INFO:
         parser.error('Unsupported TV Provider, use --ap-list-mso to get a list of supported TV Providers')
 
+    if opts.no_check_extensions:
+        _UnsafeExtensionError.lenient = True
+
     def parse_retries(retries):
         if retries in ('inf', 'infinite'):
             parsed_retries = float('inf')

youtube_dl/extractor/rtp.py

@@ -3,25 +3,38 @@ from __future__ import unicode_literals
 
 from .common import InfoExtractor
 from ..utils import (
+    ExtractorError,
     determine_ext,
     js_to_json,
 )
+from ..compat import (
+    compat_b64decode,
+    compat_urllib_parse_unquote,
+)
+
+import re
+
+
+def decode_b64_url(code):
+    decoded_url = re.match(r"[^[]*\[([^]]*)\]", code).groups()[0]
+    return compat_b64decode(
+        compat_urllib_parse_unquote(
+            decoded_url.replace('"', '').replace('\'', '').replace(',', ''))).decode('utf-8')
 
 
 class RTPIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?rtp\.pt/play/p(?P<program_id>[0-9]+)/(?P<id>[^/?#]+)/?'
+    _VALID_URL = r'https?://(?:(?:(?:www\.)?rtp\.pt/play/(?P<subarea>.*/)?p(?P<program_id>[0-9]+)/(?P<episode_id>e[0-9]+/)?)|(?:arquivos\.rtp\.pt/conteudos/))(?P<id>[^/?#]+)/?'
     _TESTS = [{
-        'url': 'http://www.rtp.pt/play/p405/e174042/paixoes-cruzadas',
-        'md5': 'e736ce0c665e459ddb818546220b4ef8',
+        'url': 'https://www.rtp.pt/play/p9165/e562949/por-do-sol',
         'info_dict': {
-            'id': 'e174042',
-            'ext': 'mp3',
-            'title': 'Paixões Cruzadas',
-            'description': 'As paixões musicais de António Cartaxo e António Macedo',
+            'id': 'por-do-sol',
+            'ext': 'mp4',
+            'title': 'Pôr do Sol Episódio 1 - de 16 Ago 2021',
+            'description': 'Madalena Bourbon de Linhaça vive atormentada pelo segredo que esconde desde 1990. Matilde Bourbon de Linhaça sonha fugir com o seu amor proibido. O en',
             'thumbnail': r're:^https?://.*\.jpg',
         },
     }, {
-        'url': 'http://www.rtp.pt/play/p831/a-quimica-das-coisas',
+        'url': 'https://www.rtp.pt/play/p510/aleixo-fm',
         'only_matching': True,
     }]
 
@@ -29,30 +42,61 @@ class RTPIE(InfoExtractor):
         video_id = self._match_id(url)
 
         webpage = self._download_webpage(url, video_id)
-        title = self._html_search_meta(
-            'twitter:title', webpage, display_name='title', fatal=True)
 
-        config = self._parse_json(self._search_regex(
-            r'(?s)RTPPlayer\(({.+?})\);', webpage,
-            'player config'), video_id, js_to_json)
-        file_url = config['file']
-        ext = determine_ext(file_url)
-        if ext == 'm3u8':
-            file_key = config.get('fileKey')
-            formats = self._extract_m3u8_formats(
-                file_url, video_id, 'mp4', 'm3u8_native',
-                m3u8_id='hls', fatal=file_key)
-            if file_key:
-                formats.append({
-                    'url': 'https://cdn-ondemand.rtp.pt' + file_key,
-                    'preference': 1,
-                })
+        # Remove JS multi-line comments from webpage source
+        webpage = re.sub(r'(\/\*.*\*\/)', '', webpage, flags=re.DOTALL)
+
+        title = self._html_search_regex(r'<title>(.+?)</title>', webpage, 'title')
+
+        # Replace irrelevant text in title
+        title = title.replace(' - RTP Play - RTP', '')
+
+        # Check if it's a video split in parts, if so add part number to title
+        part = self._html_search_regex(r'section\-parts.*<span.*>(.+?)</span>.*</ul>', webpage, 'part', default=None)
+        if part:
+            title = f'{title} {part}'
+
+        # Get JS object
+        js_object = self._search_regex(r'(?s)RTPPlayer *\( *({.+?}) *\);', webpage, 'player config')
+        json_string_for_config = ''
+        full_url = None
+
+        # Verify JS object since it isn't pure JSON and maybe it needs some tuning
+        for line in js_object.splitlines():
+            stripped_line = line.strip()
+
+            # key == 'fileKey', then we found what we wanted
+            if re.match(r'fileKey:', stripped_line):
+                if re.match(r'fileKey: *""', stripped_line):
+                    raise ExtractorError("Episode not found (probably removed)", expected=True)
+                url = decode_b64_url(stripped_line)
+                if 'mp3' in url:
+                    full_url = 'https://cdn-ondemand.rtp.pt' + url
+                else:
+                    full_url = 'https://streaming-vod.rtp.pt/dash{}/manifest.mpd'.format(url)
+
+            elif not stripped_line.startswith("//") and not re.match('file *:', stripped_line) and not re.match('.*extraSettings ?:', stripped_line):
+                # Ignore commented lines, `extraSettings` and `f`. The latter seems to some random unrelated video.
+                json_string_for_config += '\n' + line
+
+        if not full_url:
+            raise ExtractorError("No valid media source found in page")
+
+        # Finally send pure JSON string for JSON parsing
+        config = self._parse_json(json_string_for_config, video_id, js_to_json)
+        full_url = full_url.replace('drm-dash', 'dash')
+        ext = determine_ext(full_url)
+
+        if ext == 'mpd':
+            # Download via mpd file
+            formats = self._extract_mpd_formats(full_url, video_id)
             self._sort_formats(formats)
         else:
             formats = [{
-                'url': file_url,
+                'url': full_url,
                 'ext': ext,
             }]
+
         if config.get('mediaType') == 'audio':
             for f in formats:
                 f['vcodec'] = 'none'

youtube_dl/options.py

@@ -533,6 +533,10 @@ def parseOpts(overrideArguments=None):
         '--no-check-certificate',
         action='store_true', dest='no_check_certificate', default=False,
         help='Suppress HTTPS certificate validation')
+    workarounds.add_option(
+        '--no-check-extensions',
+        action='store_true', dest='no_check_extensions', default=False,
+        help='Suppress file extension validation')
     workarounds.add_option(
         '--prefer-insecure',
         '--prefer-unsecure', action='store_true', dest='prefer_insecure',

youtube_dl/utils.py

@@ -1717,21 +1717,6 @@ TIMEZONE_NAMES = {
     'PST': -8, 'PDT': -7   # Pacific
 }
 
-KNOWN_EXTENSIONS = (
-    'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'aac',
-    'flv', 'f4v', 'f4a', 'f4b',
-    'webm', 'ogg', 'ogv', 'oga', 'ogx', 'spx', 'opus',
-    'mkv', 'mka', 'mk3d',
-    'avi', 'divx',
-    'mov',
-    'asf', 'wmv', 'wma',
-    '3gp', '3g2',
-    'mp3',
-    'flac',
-    'ape',
-    'wav',
-    'f4f', 'f4m', 'm3u8', 'smil')
-
 # needed for sanitizing filenames in restricted mode
 ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
                         itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
@@ -3959,19 +3944,22 @@ def parse_duration(s):
     return duration
 
 
-def prepend_extension(filename, ext, expected_real_ext=None):
+def _change_extension(prepend, filename, ext, expected_real_ext=None):
     name, real_ext = os.path.splitext(filename)
-    return (
-        '{0}.{1}{2}'.format(name, ext, real_ext)
-        if not expected_real_ext or real_ext[1:] == expected_real_ext
-        else '{0}.{1}'.format(filename, ext))
+    sanitize_extension = _UnsafeExtensionError.sanitize_extension
+
+    if not expected_real_ext or real_ext.partition('.')[0::2] == ('', expected_real_ext):
+        filename = name
+        if prepend and real_ext:
+            sanitize_extension(ext, prepend=prepend)
+            return ''.join((filename, '.', ext, real_ext))
+
+    # Mitigate path traversal and file impersonation attacks
+    return '.'.join((filename, sanitize_extension(ext)))
 
 
-def replace_extension(filename, ext, expected_real_ext=None):
-    name, real_ext = os.path.splitext(filename)
-    return '{0}.{1}'.format(
-        name if not expected_real_ext or real_ext[1:] == expected_real_ext else filename,
-        ext)
+prepend_extension = functools.partial(_change_extension, True)
+replace_extension = functools.partial(_change_extension, False)
 
 
 def check_executable(exe, args=[]):
@@ -6561,3 +6549,138 @@ def join_nonempty(*values, **kwargs):
     if from_dict is not None:
         values = (traverse_obj(from_dict, variadic(v)) for v in values)
     return delim.join(map(compat_str, filter(None, values)))
+
+
+class Namespace(object):
+    """Immutable namespace"""
+
+    def __init__(self, **kw_attr):
+        self.__dict__.update(kw_attr)
+
+    def __iter__(self):
+        return iter(self.__dict__.values())
+
+    @property
+    def items_(self):
+        return self.__dict__.items()
+
+
+MEDIA_EXTENSIONS = Namespace(
+    common_video=('avi', 'flv', 'mkv', 'mov', 'mp4', 'webm'),
+    video=('3g2', '3gp', 'f4v', 'mk3d', 'divx', 'mpg', 'ogv', 'm4v', 'wmv'),
+    common_audio=('aiff', 'alac', 'flac', 'm4a', 'mka', 'mp3', 'ogg', 'opus', 'wav'),
+    audio=('aac', 'ape', 'asf', 'f4a', 'f4b', 'm4b', 'm4p', 'm4r', 'oga', 'ogx', 'spx', 'vorbis', 'wma', 'weba'),
+    thumbnails=('jpg', 'png', 'webp'),
+    # storyboards=('mhtml', ),
+    subtitles=('srt', 'vtt', 'ass', 'lrc', 'ttml'),
+    manifests=('f4f', 'f4m', 'm3u8', 'smil', 'mpd'),
+)
+MEDIA_EXTENSIONS.video = MEDIA_EXTENSIONS.common_video + MEDIA_EXTENSIONS.video
+MEDIA_EXTENSIONS.audio = MEDIA_EXTENSIONS.common_audio + MEDIA_EXTENSIONS.audio
+
+KNOWN_EXTENSIONS = (
+    MEDIA_EXTENSIONS.video + MEDIA_EXTENSIONS.audio
+    + MEDIA_EXTENSIONS.manifests
+)
+
+
+class _UnsafeExtensionError(Exception):
+    """
+    Mitigation exception for unwanted file overwrite/path traversal
+
+    Ref: https://github.com/yt-dlp/yt-dlp/security/advisories/GHSA-79w7-vh3h-8g4j
+    """
+    _ALLOWED_EXTENSIONS = frozenset(itertools.chain(
+        (   # internal
+            'description',
+            'json',
+            'meta',
+            'orig',
+            'part',
+            'temp',
+            'uncut',
+            'unknown_video',
+            'ytdl',
+        ),
+        # video
+        MEDIA_EXTENSIONS.video, (
+            'avif',
+            'ismv',
+            'm2ts',
+            'm4s',
+            'mng',
+            'mpeg',
+            'qt',
+            'swf',
+            'ts',
+            'vp9',
+            'wvm',
+        ),
+        # audio
+        MEDIA_EXTENSIONS.audio, (
+            'isma',
+            'mid',
+            'mpga',
+            'ra',
+        ),
+        # image
+        MEDIA_EXTENSIONS.thumbnails, (
+            'bmp',
+            'gif',
+            'ico',
+            'heic',
+            'jng',
+            'jpeg',
+            'jxl',
+            'svg',
+            'tif',
+            'wbmp',
+        ),
+        # subtitle
+        MEDIA_EXTENSIONS.subtitles, (
+            'dfxp',
+            'fs',
+            'ismt',
+            'sami',
+            'scc',
+            'ssa',
+            'tt',
+        ),
+        # others
+        MEDIA_EXTENSIONS.manifests,
+        (
+            # not used in yt-dl
+            # *MEDIA_EXTENSIONS.storyboards,
+            # 'desktop',
+            # 'ism',
+            # 'm3u',
+            # 'sbv',
+            # 'swp',
+            # 'url',
+            # 'webloc',
+            # 'xml',
+        )))
+
+    def __init__(self, extension):
+        super(_UnsafeExtensionError, self).__init__('unsafe file extension: {0!r}'.format(extension))
+        self.extension = extension
+
+    # support --no-check-extensions
+    lenient = False
+
+    @classmethod
+    def sanitize_extension(cls, extension, **kwargs):
+        # ... /, *, prepend=False
+        prepend = kwargs.get('prepend', False)
+
+        if '/' in extension or '\\' in extension:
+            raise cls(extension)
+
+        if not prepend:
+            last = extension.rpartition('.')[-1]
+            if last == 'bin':
+                extension = last = 'unknown_video'
+            if not (cls.lenient or last.lower() in cls._ALLOWED_EXTENSIONS):
+                raise cls(extension)
+
+        return extension