1
0
mirror of https://github.com/l1ving/youtube-dl synced 2024-12-21 18:42:59 +08:00

[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
This commit is contained in:
Sergey M․ 2014-09-12 23:20:17 +07:00
parent 64a450131b
commit 37419b4f99

View File

@ -28,6 +28,7 @@ from .utils import (
compat_str, compat_str,
compat_urllib_error, compat_urllib_error,
compat_urllib_request, compat_urllib_request,
compat_urllib_parse_urlparse,
ContentTooShortError, ContentTooShortError,
date_from_str, date_from_str,
DateRange, DateRange,
@ -1241,6 +1242,31 @@ class YoutubeDL(object):
def urlopen(self, req): def urlopen(self, req):
""" Start an HTTP download """ """ 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) return self._opener.open(req, timeout=self._socket_timeout)
def print_debug_header(self): def print_debug_header(self):