From a076c1f97a2fd42555578741323d215010eea767 Mon Sep 17 00:00:00 2001 From: Elyse Date: Wed, 27 Apr 2022 17:50:01 -0500 Subject: [PATCH] [extractor] Update `manifest_url`s after redirect (#3575) Authored by: elyse0 --- yt_dlp/extractor/common.py | 34 +++++++++++++++++++++++----------- 1 file changed, 23 insertions(+), 11 deletions(-) diff --git a/yt_dlp/extractor/common.py b/yt_dlp/extractor/common.py index c60474c7b..8c2fd7fea 100644 --- a/yt_dlp/extractor/common.py +++ b/yt_dlp/extractor/common.py @@ -1982,17 +1982,19 @@ def _sleep(self, timeout, video_id, msg_template=None): def _extract_f4m_formats(self, manifest_url, video_id, preference=None, quality=None, f4m_id=None, transform_source=lambda s: fix_xml_ampersands(s).strip(), fatal=True, m3u8_id=None, data=None, headers={}, query={}): - manifest = self._download_xml( + res = self._download_xml_handle( manifest_url, video_id, 'Downloading f4m manifest', 'Unable to download f4m manifest', # Some manifests may be malformed, e.g. prosiebensat1 generated manifests # (see https://github.com/ytdl-org/youtube-dl/issues/6215#issuecomment-121704244) transform_source=transform_source, fatal=fatal, data=data, headers=headers, query=query) - - if manifest is False: + if res is False: return [] + manifest, urlh = res + manifest_url = urlh.geturl() + return self._parse_f4m_formats( manifest, manifest_url, video_id, preference=preference, quality=quality, f4m_id=f4m_id, transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id) @@ -2400,12 +2402,14 @@ def _xpath_ns(path, namespace=None): return '/'.join(out) def _extract_smil_formats_and_subtitles(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None): - smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source) - - if smil is False: + res = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source) + if res is False: assert not fatal return [], {} + smil, urlh = res + smil_url = urlh.geturl() + namespace = self._parse_smil_namespace(smil) fmts = self._parse_smil_formats( @@ -2422,13 +2426,17 @@ def _extract_smil_formats(self, *args, **kwargs): return fmts def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None): - smil = self._download_smil(smil_url, video_id, fatal=fatal) - if smil is False: + res = self._download_smil(smil_url, video_id, fatal=fatal) + if res is False: return {} + + smil, urlh = res + smil_url = urlh.geturl() + return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params) def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None): - return self._download_xml( + return self._download_xml_handle( smil_url, video_id, 'Downloading SMIL file', 'Unable to download SMIL file', fatal=fatal, transform_source=transform_source) @@ -2607,11 +2615,15 @@ def _parse_smil_subtitles(self, smil, namespace=None, subtitles_lang='en'): return subtitles def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True): - xspf = self._download_xml( + res = self._download_xml_handle( xspf_url, playlist_id, 'Downloading xpsf playlist', 'Unable to download xspf manifest', fatal=fatal) - if xspf is False: + if res is False: return [] + + xspf, urlh = res + xspf_url = urlh.geturl() + return self._parse_xspf( xspf, playlist_id, xspf_url=xspf_url, xspf_base_url=base_url(xspf_url))