From 448bb5f333c6c4c8084e479e1035ff674e4f8fd4 Mon Sep 17 00:00:00 2001 From: Yen Chi Hsuan Date: Fri, 27 May 2016 00:03:03 +0800 Subject: [PATCH] [common] Fix non-bootstrapped support in f4m --- youtube_dl/extractor/common.py | 19 +++++++++++++------ youtube_dl/extractor/playwire.py | 2 +- 2 files changed, 14 insertions(+), 7 deletions(-) diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index b5bea59040..e53b7ad643 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -987,7 +987,7 @@ def _sleep(self, timeout, video_id, msg_template=None): def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None, transform_source=lambda s: fix_xml_ampersands(s).strip(), - fatal=True, assume_f4mv2=False, m3u8_id=None): + fatal=True, m3u8_id=None): manifest = self._download_xml( manifest_url, video_id, 'Downloading f4m manifest', 'Unable to download f4m manifest', @@ -1001,12 +1001,11 @@ def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=N return self._parse_f4m_formats( manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id, - transform_source=transform_source, fatal=fatal, assume_f4mv2=assume_f4mv2, - m3u8_id=m3u8_id) + transform_source=transform_source, fatal=fatal, m3u8_id=m3u8_id) def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None, transform_source=lambda s: fix_xml_ampersands(s).strip(), - fatal=True, assume_f4mv2=False, m3u8_id=None): + fatal=True, m3u8_id=None): # currently youtube-dl cannot decode the playerVerificationChallenge as Akamai uses Adobe Alchemy akamai_pv = manifest.find('{http://ns.adobe.com/f4m/1.0}pv-2.0') if akamai_pv is not None and ';' in akamai_pv.text: @@ -1036,8 +1035,16 @@ def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, 'bootstrap info', default=None) for i, media_el in enumerate(media_nodes): - if manifest_version == '2.0' or assume_f4mv2: - media_url = media_el.attrib.get('href') or media_el.attrib.get('url') + # If is present, the specified f4m is a + # stream-level manifest, and only set-level manifests may refer to + # external resources. See section 11.4 and section 4 of F4M spec + if bootstrap_info is None: + media_url = None + # @href is introduced in 2.0, see section 11.6 of F4M spec + if manifest_version == '2.0': + media_url = media_el.attrib.get('href') + if media_url is None: + media_url = media_el.attrib.get('url') if not media_url: continue manifest_url = ( diff --git a/youtube_dl/extractor/playwire.py b/youtube_dl/extractor/playwire.py index 2ee5c5aa37..0bc7431189 100644 --- a/youtube_dl/extractor/playwire.py +++ b/youtube_dl/extractor/playwire.py @@ -60,7 +60,7 @@ def _real_extract(self, url): thumbnail = content.get('poster') src = content['media']['f4m'] - formats = self._extract_f4m_formats(src, video_id, assume_f4mv2=True, m3u8_id='hls') + formats = self._extract_f4m_formats(src, video_id, m3u8_id='hls') for a_format in formats: if not dict_get(a_format, ['tbr', 'width', 'height']): a_format['quality'] = 1 if '-hd.' in a_format['url'] else 0