From c8f146c5f3cd6600cab8f570b29510ee0d092523 Mon Sep 17 00:00:00 2001 From: Gilles Pietri Date: Sun, 27 Sep 2020 14:51:42 +0200 Subject: [PATCH 3/8] [bandcamp] match album titles inside the new JSON data block, and unescape the title properly From https://github.com/ytdl-org/youtube-dl/pull/26684. --- youtube_dl/extractor/bandcamp.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/youtube_dl/extractor/bandcamp.py b/youtube_dl/extractor/bandcamp.py index 55d110e28..f036a89eb 100644 --- a/youtube_dl/extractor/bandcamp.py +++ b/youtube_dl/extractor/bandcamp.py @@ -316,10 +316,10 @@ class BandcampAlbumIE(InfoExtractor): if self._html_search_meta('duration', elem_content, default=None)] title = self._html_search_regex( - r'album_title\s*:\s*"((?:\\.|[^"\\])+?)"', + r'album_title\s*(?:"|["\']):\s*(?:"|["\'])((?:\\.|[^"\\])+?)(?:"|["\'])', webpage, 'title', fatal=False) if title: - title = title.replace(r'\"', '"') + title = unescapeHTML(title) return { '_type': 'playlist', 'uploader_id': uploader_id, -- 2.27.0