]> jfr.im git - yt-dlp.git/blobdiff - yt_dlp/YoutubeDL.py
Add interactive format selection with `-f -`
[yt-dlp.git] / yt_dlp / YoutubeDL.py
index 29c9ecd16a1c85f2809b5d65b62ec959e19b7092..be0a9c43d0a52ba8a220c66e34ed8fa9a2338767 100644 (file)
@@ -5,7 +5,6 @@
 
 import collections
 import contextlib
-import copy
 import datetime
 import errno
 import fileinput
@@ -70,6 +69,7 @@
     format_field,
     formatSeconds,
     GeoRestrictedError,
+    get_domain,
     HEADRequest,
     int_or_none,
     iri_to_uri,
     PostProcessingError,
     preferredencoding,
     prepend_extension,
+    ReExtractInfo,
     register_socks_protocols,
     RejectedVideoReached,
+    remove_terminal_sequences,
     render_table,
     replace_extension,
     SameFileError,
     strftime_or_none,
     subtitles_filename,
     supports_terminal_sequences,
-    ThrottledDownload,
+    timetuple_from_msec,
     to_high_limit_path,
     traverse_obj,
     try_get,
 from .postprocessor import (
     get_postprocessor,
     EmbedThumbnailPP,
+    FFmpegFixupDuplicateMoovPP,
     FFmpegFixupDurationPP,
     FFmpegFixupM3u8PP,
     FFmpegFixupM4aPP,
     _PLUGIN_CLASSES as plugin_postprocessors
 )
 from .update import detect_variant
-from .version import __version__
+from .version import __version__, RELEASE_GIT_HEAD
 
 if compat_os_name == 'nt':
     import ctypes
@@ -326,13 +329,16 @@ class YoutubeDL(object):
     bidi_workaround:   Work around buggy terminals without bidirectional text
                        support, using fridibi
     debug_printtraffic:Print out sent and received HTTP traffic
-    include_ads:       Download ads as well
+    include_ads:       Download ads as well (deprecated)
     default_search:    Prepend this string if an input url is not valid.
                        'auto' for elaborate guessing
     encoding:          Use this encoding instead of the system-specified.
     extract_flat:      Do not resolve URLs, return the immediate result.
                        Pass in 'in_playlist' to only show this behavior for
                        playlist items.
+    wait_for_video:    If given, wait for scheduled streams to become available.
+                       The value should be a tuple containing the range
+                       (min_secs, max_secs) to wait between retries
     postprocessors:    A list of dictionaries, each with an entry
                        * key:  The name of the postprocessor. See
                                yt_dlp/postprocessor/__init__.py for a list.
@@ -562,6 +568,8 @@ def check_deprecated(param, option, suggestion):
 
         for msg in self.params.get('_warnings', []):
             self.report_warning(msg)
+        for msg in self.params.get('_deprecation_warnings', []):
+            self.deprecation_warning(msg)
 
         if 'list-formats' in self.params.get('compat_opts', []):
             self.params['listformats_table'] = False
@@ -616,7 +624,7 @@ def check_deprecated(param, option, suggestion):
 
         # Creating format selector here allows us to catch syntax errors before the extraction
         self.format_selector = (
-            None if self.params.get('format') is None
+            self.params.get('format') if self.params.get('format') in (None, '-')
             else self.params['format'] if callable(self.params['format'])
             else self.build_format_selector(self.params['format']))
 
@@ -627,13 +635,6 @@ def check_deprecated(param, option, suggestion):
                 self.print_debug_header()
             self.add_default_info_extractors()
 
-        for pp_def_raw in self.params.get('postprocessors', []):
-            pp_def = dict(pp_def_raw)
-            when = pp_def.pop('when', 'post_process')
-            pp_class = get_postprocessor(pp_def.pop('key'))
-            pp = pp_class(self, **compat_kwargs(pp_def))
-            self.add_post_processor(pp, when=when)
-
         hooks = {
             'post_hooks': self.add_post_hook,
             'progress_hooks': self.add_progress_hook,
@@ -643,6 +644,13 @@ def check_deprecated(param, option, suggestion):
             for ph in self.params.get(opt, []):
                 fn(ph)
 
+        for pp_def_raw in self.params.get('postprocessors', []):
+            pp_def = dict(pp_def_raw)
+            when = pp_def.pop('when', 'post_process')
+            self.add_post_processor(
+                get_postprocessor(pp_def.pop('key'))(self, **compat_kwargs(pp_def)),
+                when=when)
+
         register_socks_protocols()
 
         def preload_download_archive(fn):
@@ -729,6 +737,9 @@ def add_progress_hook(self, ph):
     def add_postprocessor_hook(self, ph):
         """Add the postprocessing progress hook"""
         self._postprocessor_hooks.append(ph)
+        for pps in self._pps.values():
+            for pp in pps:
+                pp.add_progress_hook(ph)
 
     def _bidi_workaround(self, message):
         if not hasattr(self, '_output_channel'):
@@ -770,6 +781,7 @@ def to_stderr(self, message, only_once=False):
     def to_console_title(self, message):
         if not self.params.get('consoletitle', False):
             return
+        message = remove_terminal_sequences(message)
         if compat_os_name == 'nt':
             if ctypes.windll.kernel32.GetConsoleWindow():
                 # c_wchar_p() might not be necessary if `message` is
@@ -806,14 +818,15 @@ def __exit__(self, *args):
         if self.params.get('cookiefile') is not None:
             self.cookiejar.save(ignore_discard=True, ignore_expires=True)
 
-    def trouble(self, message=None, tb=None):
+    def trouble(self, message=None, tb=None, is_error=True):
         """Determine action to take when a download problem appears.
 
         Depending on if the downloader has been configured to ignore
         download errors or not, this method may throw an exception or
         not when errors are found, after printing the message.
 
-        tb, if given, is additional traceback information.
+        @param tb          If given, is additional traceback information
+        @param is_error    Whether to raise error according to ignorerrors
         """
         if message is not None:
             self.to_stderr(message)
@@ -829,6 +842,8 @@ def trouble(self, message=None, tb=None):
                     tb = ''.join(tb_data)
             if tb:
                 self.to_stderr(tb)
+        if not is_error:
+            return
         if not self.params.get('ignoreerrors'):
             if sys.exc_info()[0] and hasattr(sys.exc_info()[1], 'exc_info') and sys.exc_info()[1].exc_info[0]:
                 exc_info = sys.exc_info()[1].exc_info
@@ -844,7 +859,7 @@ def to_screen(self, message, skip_eol=False):
 
     class Styles(Enum):
         HEADERS = 'yellow'
-        EMPHASIS = 'blue'
+        EMPHASIS = 'light blue'
         ID = 'green'
         DELIM = 'blue'
         ERROR = 'red'
@@ -859,7 +874,7 @@ def _format_text(self, handle, allow_colors, text, f, fallback=None, *, test_enc
             if fallback is not None and text != original_text:
                 text = fallback
         if isinstance(f, self.Styles):
-            f = f._value_
+            f = f.value
         return format_text(text, f) if allow_colors else text if fallback is None else fallback
 
     def _format_screen(self, *args, **kwargs):
@@ -882,12 +897,18 @@ def report_warning(self, message, only_once=False):
                 return
             self.to_stderr(f'{self._format_err("WARNING:", self.Styles.WARNING)} {message}', only_once)
 
-    def report_error(self, message, tb=None):
+    def deprecation_warning(self, message):
+        if self.params.get('logger') is not None:
+            self.params['logger'].warning('DeprecationWarning: {message}')
+        else:
+            self.to_stderr(f'{self._format_err("DeprecationWarning:", self.Styles.ERROR)} {message}', True)
+
+    def report_error(self, message, *args, **kwargs):
         '''
         Do the same as trouble, but prefixes the message with 'ERROR:', colored
         in red if stderr is a tty file.
         '''
-        self.trouble(f'{self._format_err("ERROR:", self.Styles.ERROR)} {message}', tb)
+        self.trouble(f'{self._format_err("ERROR:", self.Styles.ERROR)} {message}', *args, **kwargs)
 
     def write_debug(self, message, only_once=False):
         '''Log debug message or Print message to stderr'''
@@ -1037,7 +1058,8 @@ def prepare_outtmpl(self, outtmpl, info_dict, sanitize=None):
             (?P<fields>{field})
             (?P<maths>(?:{math_op}{math_field})*)
             (?:>(?P<strf_format>.+?))?
-            (?P<alternate>(?<!\\),[^|)]+)?
+            (?P<alternate>(?<!\\),[^|&)]+)?
+            (?:&(?P<replacement>.*?))?
             (?:\|(?P<default>.*?))?
             $'''.format(field=FIELD_RE, math_op=MATH_OPERATORS_RE, math_field=MATH_FIELD_RE))
 
@@ -1088,7 +1110,7 @@ def get_value(mdict):
         def _dumpjson_default(obj):
             if isinstance(obj, (set, LazyList)):
                 return list(obj)
-            raise TypeError(f'Object of type {type(obj).__name__} is not JSON serializable')
+            return repr(obj)
 
         def create_key(outer_mobj):
             if not outer_mobj.group('has_key'):
@@ -1096,11 +1118,12 @@ def create_key(outer_mobj):
             key = outer_mobj.group('key')
             mobj = re.match(INTERNAL_FORMAT_RE, key)
             initial_field = mobj.group('fields').split('.')[-1] if mobj else ''
-            value, default = None, na
+            value, replacement, default = None, None, na
             while mobj:
                 mobj = mobj.groupdict()
                 default = mobj['default'] if mobj['default'] is not None else default
                 value = get_value(mobj)
+                replacement = mobj['replacement']
                 if value is None and mobj['alternate']:
                     mobj = re.match(INTERNAL_FORMAT_RE, mobj['alternate'][1:])
                 else:
@@ -1110,7 +1133,7 @@ def create_key(outer_mobj):
             if fmt == 's' and value is not None and key in field_size_compat_map.keys():
                 fmt = '0{:d}d'.format(field_size_compat_map[key])
 
-            value = default if value is None else value
+            value = default if value is None else value if replacement is None else replacement
 
             flags = outer_mobj.group('conversion') or ''
             str_fmt = f'{fmt[:-1]}s'
@@ -1174,12 +1197,8 @@ def _prepare_filename(self, info_dict, tmpl_type='default'):
             # https://github.com/blackjack4494/youtube-dlc/issues/85
             trim_file_name = self.params.get('trim_file_name', False)
             if trim_file_name:
-                fn_groups = filename.rsplit('.')
-                ext = fn_groups[-1]
-                sub_ext = ''
-                if len(fn_groups) > 2:
-                    sub_ext = fn_groups[-2]
-                filename = join_nonempty(fn_groups[0][:trim_file_name], sub_ext, ext, delim='.')
+                no_ext, *ext = filename.rsplit('.', 2)
+                filename = join_nonempty(no_ext[:trim_file_name], *ext, delim='.')
 
             return filename
         except ValueError as err:
@@ -1328,9 +1347,12 @@ def wrapper(self, *args, **kwargs):
                 self.report_error(msg)
             except ExtractorError as e:  # An error we somewhat expected
                 self.report_error(compat_str(e), e.format_traceback())
-            except ThrottledDownload as e:
-                self.to_stderr('\r')
-                self.report_warning(f'{e}; Re-extracting data')
+            except ReExtractInfo as e:
+                if e.expected:
+                    self.to_screen(f'{e}; Re-extracting data')
+                else:
+                    self.to_stderr('\r')
+                    self.report_warning(f'{e}; Re-extracting data')
                 return wrapper(self, *args, **kwargs)
             except (DownloadCancelled, LazyList.IndexError, PagedList.IndexError):
                 raise
@@ -1341,6 +1363,47 @@ def wrapper(self, *args, **kwargs):
                     raise
         return wrapper
 
+    def _wait_for_video(self, ie_result):
+        if (not self.params.get('wait_for_video')
+                or ie_result.get('_type', 'video') != 'video'
+                or ie_result.get('formats') or ie_result.get('url')):
+            return
+
+        format_dur = lambda dur: '%02d:%02d:%02d' % timetuple_from_msec(dur * 1000)[:-1]
+        last_msg = ''
+
+        def progress(msg):
+            nonlocal last_msg
+            self.to_screen(msg + ' ' * (len(last_msg) - len(msg)) + '\r', skip_eol=True)
+            last_msg = msg
+
+        min_wait, max_wait = self.params.get('wait_for_video')
+        diff = try_get(ie_result, lambda x: x['release_timestamp'] - time.time())
+        if diff is None and ie_result.get('live_status') == 'is_upcoming':
+            diff = random.randrange(min_wait, max_wait) if (max_wait and min_wait) else (max_wait or min_wait)
+            self.report_warning('Release time of video is not known')
+        elif (diff or 0) <= 0:
+            self.report_warning('Video should already be available according to extracted info')
+        diff = min(max(diff or 0, min_wait or 0), max_wait or float('inf'))
+        self.to_screen(f'[wait] Waiting for {format_dur(diff)} - Press Ctrl+C to try now')
+
+        wait_till = time.time() + diff
+        try:
+            while True:
+                diff = wait_till - time.time()
+                if diff <= 0:
+                    progress('')
+                    raise ReExtractInfo('[wait] Wait period ended', expected=True)
+                progress(f'[wait] Remaining time until next attempt: {self._format_screen(format_dur(diff), self.Styles.EMPHASIS)}')
+                time.sleep(1)
+        except KeyboardInterrupt:
+            progress('')
+            raise ReExtractInfo('[wait] Interrupted by user', expected=True)
+        except BaseException as e:
+            if not isinstance(e, ReExtractInfo):
+                self.to_screen('')
+            raise
+
     @__handle_extraction_exceptions
     def __extract_info(self, url, ie, download, extra_info, process):
         ie_result = ie.extract(url)
@@ -1356,6 +1419,7 @@ def __extract_info(self, url, ie, download, extra_info, process):
             ie_result.setdefault('original_url', extra_info['original_url'])
         self.add_default_extra_info(ie_result, ie, url)
         if process:
+            self._wait_for_video(ie_result)
             return self.process_ie_result(ie_result, download, extra_info)
         else:
             return ie_result
@@ -1366,6 +1430,7 @@ def add_default_extra_info(self, ie_result, ie, url):
                 'webpage_url': url,
                 'original_url': url,
                 'webpage_url_basename': url_basename(url),
+                'webpage_url_domain': get_domain(url),
             })
         if ie is not None:
             self.add_extra_info(ie_result, {
@@ -1399,6 +1464,7 @@ def process_ie_result(self, ie_result, download=True, extra_info=None):
                     info_copy['id'] = ie.get_temp_id(ie_result['url'])
                 self.add_default_extra_info(info_copy, ie, ie_result['url'])
                 self.add_extra_info(info_copy, extra_info)
+                info_copy, _ = self.pre_process(info_copy)
                 self.__forced_printings(info_copy, self.prepare_filename(info_copy), incomplete=True)
                 if self.params.get('force_write_download_archive', False):
                     self.record_download_archive(info_copy)
@@ -1489,6 +1555,7 @@ def _fixup(r):
                     'extractor': ie_result['extractor'],
                     'webpage_url': ie_result['webpage_url'],
                     'webpage_url_basename': url_basename(ie_result['webpage_url']),
+                    'webpage_url_domain': get_domain(ie_result['webpage_url']),
                     'extractor_key': ie_result['extractor_key'],
                 })
                 return r
@@ -1650,6 +1717,7 @@ def get_entry(i):
                 'extractor': ie_result['extractor'],
                 'webpage_url': ie_result['webpage_url'],
                 'webpage_url_basename': url_basename(ie_result['webpage_url']),
+                'webpage_url_domain': get_domain(ie_result['webpage_url']),
                 'extractor_key': ie_result['extractor_key'],
             }
 
@@ -2006,8 +2074,7 @@ def selector_function(ctx):
                 selector_1, selector_2 = map(_build_selector_function, selector.selector)
 
                 def selector_function(ctx):
-                    for pair in itertools.product(
-                            selector_1(copy.deepcopy(ctx)), selector_2(copy.deepcopy(ctx))):
+                    for pair in itertools.product(selector_1(ctx), selector_2(ctx)):
                         yield _merge(pair)
 
             elif selector.type == SINGLE:  # atom
@@ -2077,7 +2144,7 @@ def selector_function(ctx):
             filters = [self._build_format_filter(f) for f in selector.filters]
 
             def final_selector(ctx):
-                ctx_copy = copy.deepcopy(ctx)
+                ctx_copy = dict(ctx)
                 for _filter in filters:
                     ctx_copy['formats'] = list(filter(_filter, ctx_copy['formats']))
                 return selector_function(ctx_copy)
@@ -2289,6 +2356,10 @@ def sanitize_numeric_fields(info):
         if not self.params.get('allow_unplayable_formats'):
             formats = [f for f in formats if not f.get('has_drm')]
 
+        if info_dict.get('is_live'):
+            get_from_start = bool(self.params.get('live_from_start'))
+            formats = [f for f in formats if bool(f.get('is_from_start')) == get_from_start]
+
         if not formats:
             self.raise_no_formats(info_dict)
 
@@ -2380,20 +2451,21 @@ def is_wellformed(f):
         # The pre-processors may have modified the formats
         formats = info_dict.get('formats', [info_dict])
 
+        list_only = self.params.get('simulate') is None and (
+            self.params.get('list_thumbnails') or self.params.get('listformats') or self.params.get('listsubtitles'))
+        interactive_format_selection = not list_only and self.format_selector == '-'
         if self.params.get('list_thumbnails'):
             self.list_thumbnails(info_dict)
-        if self.params.get('listformats'):
-            if not info_dict.get('formats') and not info_dict.get('url'):
-                self.to_screen('%s has no formats' % info_dict['id'])
-            else:
-                self.list_formats(info_dict)
         if self.params.get('listsubtitles'):
             if 'automatic_captions' in info_dict:
                 self.list_subtitles(
                     info_dict['id'], automatic_captions, 'automatic captions')
             self.list_subtitles(info_dict['id'], subtitles, 'subtitles')
-        list_only = self.params.get('simulate') is None and (
-            self.params.get('list_thumbnails') or self.params.get('listformats') or self.params.get('listsubtitles'))
+        if self.params.get('listformats') or interactive_format_selection:
+            if not info_dict.get('formats') and not info_dict.get('url'):
+                self.to_screen('%s has no formats' % info_dict['id'])
+            else:
+                self.list_formats(info_dict)
         if list_only:
             # Without this printing, -F --print-json will not work
             self.__forced_printings(info_dict, self.prepare_filename(info_dict), incomplete=True)
@@ -2405,33 +2477,48 @@ def is_wellformed(f):
             self.write_debug('Default format spec: %s' % req_format)
             format_selector = self.build_format_selector(req_format)
 
-        # While in format selection we may need to have an access to the original
-        # format set in order to calculate some metrics or do some processing.
-        # For now we need to be able to guess whether original formats provided
-        # by extractor are incomplete or not (i.e. whether extractor provides only
-        # video-only or audio-only formats) for proper formats selection for
-        # extractors with such incomplete formats (see
-        # https://github.com/ytdl-org/youtube-dl/pull/5556).
-        # Since formats may be filtered during format selection and may not match
-        # the original formats the results may be incorrect. Thus original formats
-        # or pre-calculated metrics should be passed to format selection routines
-        # as well.
-        # We will pass a context object containing all necessary additional data
-        # instead of just formats.
-        # This fixes incorrect format selection issue (see
-        # https://github.com/ytdl-org/youtube-dl/issues/10083).
-        incomplete_formats = (
-            # All formats are video-only or
-            all(f.get('vcodec') != 'none' and f.get('acodec') == 'none' for f in formats)
-            # all formats are audio-only
-            or all(f.get('vcodec') == 'none' and f.get('acodec') != 'none' for f in formats))
-
-        ctx = {
-            'formats': formats,
-            'incomplete_formats': incomplete_formats,
-        }
+        while True:
+            if interactive_format_selection:
+                req_format = input(
+                    self._format_screen('\nEnter format selector: ', self.Styles.EMPHASIS))
+                try:
+                    format_selector = self.build_format_selector(req_format)
+                except SyntaxError as err:
+                    self.report_error(err, tb=False, is_error=False)
+                    continue
+
+            # While in format selection we may need to have an access to the original
+            # format set in order to calculate some metrics or do some processing.
+            # For now we need to be able to guess whether original formats provided
+            # by extractor are incomplete or not (i.e. whether extractor provides only
+            # video-only or audio-only formats) for proper formats selection for
+            # extractors with such incomplete formats (see
+            # https://github.com/ytdl-org/youtube-dl/pull/5556).
+            # Since formats may be filtered during format selection and may not match
+            # the original formats the results may be incorrect. Thus original formats
+            # or pre-calculated metrics should be passed to format selection routines
+            # as well.
+            # We will pass a context object containing all necessary additional data
+            # instead of just formats.
+            # This fixes incorrect format selection issue (see
+            # https://github.com/ytdl-org/youtube-dl/issues/10083).
+            incomplete_formats = (
+                # All formats are video-only or
+                all(f.get('vcodec') != 'none' and f.get('acodec') == 'none' for f in formats)
+                # all formats are audio-only
+                or all(f.get('vcodec') == 'none' and f.get('acodec') != 'none' for f in formats))
+
+            ctx = {
+                'formats': formats,
+                'incomplete_formats': incomplete_formats,
+            }
+
+            formats_to_download = list(format_selector(ctx))
+            if interactive_format_selection and not formats_to_download:
+                self.report_error('Requested format is not available', tb=False, is_error=False)
+                continue
+            break
 
-        formats_to_download = list(format_selector(ctx))
         if not formats_to_download:
             if not self.params.get('ignore_no_formats_error'):
                 raise ExtractorError('Requested format is not available', expected=True,
@@ -2595,7 +2682,9 @@ def dl(self, name, info, subtitle=False, test=False):
             urls = '", "'.join([f['url'] for f in info.get('requested_formats', [])] or [info['url']])
             self.write_debug('Invoking downloader on "%s"' % urls)
 
-        new_info = copy.deepcopy(self._copy_infodict(info))
+        # Note: Ideally info should be a deep-copied so that hooks cannot modify it.
+        # But it may contain objects that are not deep-copyable
+        new_info = self._copy_infodict(info)
         if new_info.get('http_headers') is None:
             new_info['http_headers'] = self._calc_headers(new_info)
         return fd.download(name, new_info, subtitle)
@@ -2610,6 +2699,9 @@ def process_info(self, info_dict):
             if self._num_downloads >= int(max_downloads):
                 raise MaxDownloadsReached()
 
+        if info_dict.get('is_live') and not self.params.get('live_from_start'):
+            info_dict['title'] += ' ' + datetime.datetime.now().strftime('%Y-%m-%d %H:%M')
+
         # TODO: backward compatibility, to be removed
         info_dict['fulltitle'] = info_dict['title']
 
@@ -2662,7 +2754,7 @@ def process_info(self, info_dict):
         _infojson_written = self._write_info_json('video', info_dict, infofn)
         if _infojson_written:
             info_dict['infojson_filename'] = infofn
-            # For backward compatability, even though it was a private field
+            # For backward compatibility, even though it was a private field
             info_dict['__infojson_filename'] = infofn
         elif _infojson_written is None:
             return
@@ -2821,15 +2913,22 @@ def correct_ext(filename, ext=new_ext):
                     dl_filename = existing_file(full_filename, temp_filename)
                     info_dict['__real_download'] = False
 
+                    downloaded = []
+                    merger = FFmpegMergerPP(self)
+
+                    fd = get_suitable_downloader(info_dict, self.params, to_stdout=temp_filename == '-')
                     if dl_filename is not None:
                         self.report_file_already_downloaded(dl_filename)
-                    elif get_suitable_downloader(info_dict, self.params, to_stdout=temp_filename == '-'):
+                    elif fd:
+                        for f in requested_formats if fd != FFmpegFD else []:
+                            f['filepath'] = fname = prepend_extension(
+                                correct_ext(temp_filename, info_dict['ext']),
+                                'f%s' % f['format_id'], info_dict['ext'])
+                            downloaded.append(fname)
                         info_dict['url'] = '\n'.join(f['url'] for f in requested_formats)
                         success, real_download = self.dl(temp_filename, info_dict)
                         info_dict['__real_download'] = real_download
                     else:
-                        downloaded = []
-                        merger = FFmpegMergerPP(self)
                         if self.params.get('allow_unplayable_formats'):
                             self.report_warning(
                                 'You have requested merging of multiple formats '
@@ -2841,7 +2940,7 @@ def correct_ext(filename, ext=new_ext):
                                 'The formats won\'t be merged.')
 
                         if temp_filename == '-':
-                            reason = ('using a downloader other than ffmpeg' if FFmpegFD.can_merge_formats(info_dict)
+                            reason = ('using a downloader other than ffmpeg' if FFmpegFD.can_merge_formats(info_dict, self.params)
                                       else 'but the formats are incompatible for simultaneous download' if merger.available
                                       else 'but ffmpeg is not installed')
                             self.report_warning(
@@ -2863,14 +2962,15 @@ def correct_ext(filename, ext=new_ext):
                             partial_success, real_download = self.dl(fname, new_info)
                             info_dict['__real_download'] = info_dict['__real_download'] or real_download
                             success = success and partial_success
-                        if merger.available and not self.params.get('allow_unplayable_formats'):
-                            info_dict['__postprocessors'].append(merger)
-                            info_dict['__files_to_merge'] = downloaded
-                            # Even if there were no downloads, it is being merged only now
-                            info_dict['__real_download'] = True
-                        else:
-                            for file in downloaded:
-                                files_to_move[file] = None
+
+                    if downloaded and merger.available and not self.params.get('allow_unplayable_formats'):
+                        info_dict['__postprocessors'].append(merger)
+                        info_dict['__files_to_merge'] = downloaded
+                        # Even if there were no downloads, it is being merged only now
+                        info_dict['__real_download'] = True
+                    else:
+                        for file in downloaded:
+                            files_to_move[file] = None
                 else:
                     # Just a single file
                     dl_filename = existing_file(full_filename, temp_filename)
@@ -2937,9 +3037,14 @@ def ffmpeg_fixup(cndn, msg, cls):
 
                     downloader = get_suitable_downloader(info_dict, self.params) if 'protocol' in info_dict else None
                     downloader = downloader.__name__ if downloader else None
-                    ffmpeg_fixup(info_dict.get('requested_formats') is None and downloader == 'HlsFD',
-                                 'Possible MPEG-TS in MP4 container or malformed AAC timestamps',
-                                 FFmpegFixupM3u8PP)
+
+                    if info_dict.get('requested_formats') is None:  # Not necessary if doing merger
+                        ffmpeg_fixup(downloader == 'HlsFD',
+                                     'Possible MPEG-TS in MP4 container or malformed AAC timestamps',
+                                     FFmpegFixupM3u8PP)
+                        ffmpeg_fixup(info_dict.get('is_live') and downloader == 'DashSegmentsFD',
+                                     'Possible duplicate MOOV atoms', FFmpegFixupDuplicateMoovPP)
+
                     ffmpeg_fixup(downloader == 'WebSocketFragmentFD', 'Malformed timestamps detected', FFmpegFixupTimestampPP)
                     ffmpeg_fixup(downloader == 'WebSocketFragmentFD', 'Malformed duration detected', FFmpegFixupDurationPP)
 
@@ -3007,7 +3112,7 @@ def download_with_info_file(self, info_filename):
             info = self.sanitize_info(json.loads('\n'.join(f)), self.params.get('clean_infojson', True))
         try:
             self.__download_wrapper(self.process_ie_result)(info, download=True)
-        except (DownloadError, EntryNotInPlaylist, ThrottledDownload) as e:
+        except (DownloadError, EntryNotInPlaylist, ReExtractInfo) as e:
             if not isinstance(e, EntryNotInPlaylist):
                 self.to_stderr('\r')
             webpage_url = info.get('webpage_url')
@@ -3036,10 +3141,17 @@ def sanitize_info(info_dict, remove_private_keys=False):
                 k.startswith('_') or k in remove_keys or v in empty_values)
         else:
             reject = lambda k, v: k in remove_keys
-        filter_fn = lambda obj: (
-            list(map(filter_fn, obj)) if isinstance(obj, (LazyList, list, tuple, set))
-            else obj if not isinstance(obj, dict)
-            else dict((k, filter_fn(v)) for k, v in obj.items() if not reject(k, v)))
+
+        def filter_fn(obj):
+            if isinstance(obj, dict):
+                return {k: filter_fn(v) for k, v in obj.items() if not reject(k, v)}
+            elif isinstance(obj, (list, tuple, set, LazyList)):
+                return list(map(filter_fn, obj))
+            elif obj is None or isinstance(obj, (str, int, float, bool)):
+                return obj
+            else:
+                return repr(obj)
+
         return filter_fn(info_dict)
 
     @staticmethod
@@ -3174,15 +3286,19 @@ def format_resolution(format, default='unknown'):
     def _format_note(self, fdict):
         res = ''
         if fdict.get('ext') in ['f4f', 'f4m']:
-            res += '(unsupported) '
+            res += '(unsupported)'
         if fdict.get('language'):
             if res:
                 res += ' '
-            res += '[%s] ' % fdict['language']
+            res += '[%s]' % fdict['language']
         if fdict.get('format_note') is not None:
-            res += fdict['format_note'] + ' '
+            if res:
+                res += ' '
+            res += fdict['format_note']
         if fdict.get('tbr') is not None:
-            res += '%4dk ' % fdict['tbr']
+            if res:
+                res += ', '
+            res += '%4dk' % fdict['tbr']
         if fdict.get('container') is not None:
             if res:
                 res += ', '
@@ -3334,7 +3450,8 @@ def print_debug_header(self):
         def get_encoding(stream):
             ret = getattr(stream, 'encoding', 'missing (%s)' % type(stream).__name__)
             if not supports_terminal_sequences(stream):
-                ret += ' (No ANSI)'
+                from .compat import WINDOWS_VT_MODE
+                ret += ' (No VT)' if WINDOWS_VT_MODE is False else ' (No ANSI)'
             return ret
 
         encoding_str = 'Encodings: locale %s, fs %s, out %s, err %s, pref %s' % (
@@ -3352,7 +3469,11 @@ def get_encoding(stream):
             write_debug = lambda msg: self._write_string(f'[debug] {msg}\n')
 
         source = detect_variant()
-        write_debug('yt-dlp version %s%s' % (__version__, '' if source == 'unknown' else f' ({source})'))
+        write_debug(join_nonempty(
+            'yt-dlp version', __version__,
+            f'[{RELEASE_GIT_HEAD}]' if RELEASE_GIT_HEAD else '',
+            '' if source == 'unknown' else f'({source})',
+            delim=' '))
         if not _LAZY_LOADER:
             if os.environ.get('YTDLP_NO_LAZY_EXTRACTORS'):
                 write_debug('Lazy loading extractors is forcibly disabled')
@@ -3364,20 +3485,22 @@ def get_encoding(stream):
                 for name, klass in itertools.chain(plugin_extractors.items(), plugin_postprocessors.items())])
         if self.params.get('compat_opts'):
             write_debug('Compatibility options: %s' % ', '.join(self.params.get('compat_opts')))
-        try:
-            sp = Popen(
-                ['git', 'rev-parse', '--short', 'HEAD'],
-                stdout=subprocess.PIPE, stderr=subprocess.PIPE,
-                cwd=os.path.dirname(os.path.abspath(__file__)))
-            out, err = sp.communicate_or_kill()
-            out = out.decode().strip()
-            if re.match('[0-9a-f]+', out):
-                write_debug('Git HEAD: %s' % out)
-        except Exception:
+
+        if source == 'source':
             try:
-                sys.exc_clear()
+                sp = Popen(
+                    ['git', 'rev-parse', '--short', 'HEAD'],
+                    stdout=subprocess.PIPE, stderr=subprocess.PIPE,
+                    cwd=os.path.dirname(os.path.abspath(__file__)))
+                out, err = sp.communicate_or_kill()
+                out = out.decode().strip()
+                if re.match('[0-9a-f]+', out):
+                    write_debug('Git HEAD: %s' % out)
             except Exception:
-                pass
+                try:
+                    sys.exc_clear()
+                except Exception:
+                    pass
 
         def python_implementation():
             impl_name = platform.python_implementation()
@@ -3606,7 +3729,7 @@ def _write_thumbnails(self, label, info_dict, filename, thumb_filename_base=None
             self.write_debug(f'Skipping writing {label} thumbnail')
             return ret
 
-        for t in thumbnails[::-1]:
+        for idx, t in list(enumerate(thumbnails))[::-1]:
             thumb_ext = (f'{t["id"]}.' if multiple else '') + determine_ext(t['url'], 'jpg')
             thumb_display_id = f'{label} thumbnail {t["id"]}'
             thumb_filename = replace_extension(filename, thumb_ext, info_dict.get('ext'))
@@ -3627,6 +3750,7 @@ def _write_thumbnails(self, label, info_dict, filename, thumb_filename_base=None
                     ret.append((thumb_filename, thumb_filename_final))
                     t['filepath'] = thumb_filename
                 except network_exceptions as err:
+                    thumbnails.pop(idx)
                     self.report_warning(f'Unable to download {thumb_display_id}: {err}')
             if ret and not write_all:
                 break