From 9dd209b83ff29332286efc315ad8d5fbbc142020 Mon Sep 17 00:00:00 2001 From: throwaway396 <57511390+throwaway396@users.noreply.github.com> Date: Thu, 7 Nov 2019 20:34:13 -0500 Subject: [PATCH] [chaturbate] fix url extraction and parsing --- youtube_dl/extractor/chaturbate.py | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/youtube_dl/extractor/chaturbate.py b/youtube_dl/extractor/chaturbate.py index 656e715ae..58c986ae3 100644 --- a/youtube_dl/extractor/chaturbate.py +++ b/youtube_dl/extractor/chaturbate.py @@ -1,5 +1,6 @@ from __future__ import unicode_literals +import codecs import re from .common import InfoExtractor @@ -41,9 +42,9 @@ class ChaturbateIE(InfoExtractor): m3u8_urls = [] for m in re.finditer( - r'(["\'])(?Phttp.+?\.m3u8.*?)\1', webpage): - m3u8_fast_url, m3u8_no_fast_url = m.group('url'), m.group( - 'url').replace('_fast', '') + r'\\u002[27](?Phttp.+?\.m3u8.*?)\\u002[27]', webpage): + url = codecs.decode(m.group('url'), 'unicode-escape') + m3u8_fast_url, m3u8_no_fast_url = url, url.replace('_fast', '') for m3u8_url in (m3u8_fast_url, m3u8_no_fast_url): if m3u8_url not in m3u8_urls: m3u8_urls.append(m3u8_url)