From 37419b4f9937f11ed3ca3545a32ed3451eb734ee Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Sergey=20M=E2=80=A4?= Date: Fri, 12 Sep 2014 23:20:17 +0700 Subject: [PATCH] [YoutubeDL] Escape non-ASCII characters in URLs urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991) Working around by replacing request's original URL with escaped one --- youtube_dl/YoutubeDL.py | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) diff --git a/youtube_dl/YoutubeDL.py b/youtube_dl/YoutubeDL.py index 553bf559b..99cd05e65 100755 --- a/youtube_dl/YoutubeDL.py +++ b/youtube_dl/YoutubeDL.py @@ -28,6 +28,7 @@ from .utils import ( compat_str, compat_urllib_error, compat_urllib_request, + compat_urllib_parse_urlparse, ContentTooShortError, date_from_str, DateRange, @@ -1241,6 +1242,31 @@ class YoutubeDL(object): def urlopen(self, req): """ Start an HTTP download """ + + # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991) + # Working around by replacing request's original URL with escaped one + + url = req if isinstance(req, compat_str) else req.get_full_url() + + def escape(component): + return compat_cookiejar.escape_path(component.encode('utf-8')) + + url_parsed = compat_urllib_parse_urlparse(url) + url_escaped = url_parsed._replace( + path=escape(url_parsed.path), + query=escape(url_parsed.query), + fragment=escape(url_parsed.fragment) + ).geturl() + + # Substitute URL if any change after escaping + if url != url_escaped: + if isinstance(req, compat_str): + req = url_escaped + else: + req = compat_urllib_request.Request( + url_escaped, data=req.data, headers=req.headers, + origin_req_host=req.origin_req_host, unverifiable=req.unverifiable) + return self._opener.open(req, timeout=self._socket_timeout) def print_debug_header(self):