diff options
| -rw-r--r-- | test/test_http.py | 32 | ||||
| -rw-r--r-- | youtube_dl/utils.py | 8 | 
2 files changed, 36 insertions, 4 deletions
| diff --git a/test/test_http.py b/test/test_http.py index 1a65df9e0..cd180b51f 100644 --- a/test/test_http.py +++ b/test/test_http.py @@ -183,6 +183,11 @@ class HTTPTestRequestHandler(compat_http_server.BaseHTTPRequestHandler):              self._method('GET')          elif self.path.startswith('/headers'):              self._headers() +        elif self.path.startswith('/308-to-headers'): +            self.send_response(308) +            self.send_header('Location', '/headers') +            self.send_header('Content-Length', '0') +            self.end_headers()          elif self.path == '/trailing_garbage':              payload = b'<html><video src="/vid.mp4" /></html>'              compressed = gzip_compress(payload) + b'trailing garbage' @@ -385,8 +390,31 @@ class TestHTTP(unittest.TestCase):              ydl.cookiejar.set_cookie(compat_http_cookiejar_Cookie(                  0, 'test', 'ytdl', None, False, '127.0.0.1', True,                  False, '/headers', True, False, None, False, None, None, {})) -            data = ydl.urlopen(sanitized_Request(self._test_url('headers'))).read() -            self.assertIn(b'Cookie: test=ytdl', data) +            data = ydl.urlopen(sanitized_Request( +                self._test_url('headers'))).read().decode('utf-8') +            self.assertIn('Cookie: test=ytdl', data) + +    def test_passed_cookie_header(self): +        # We should accept a Cookie header being passed as in normal headers and handle it appropriately. +        with FakeYDL() as ydl: +            # Specified Cookie header should be used +            res = ydl.urlopen(sanitized_Request( +                self._test_url('headers'), headers={'Cookie': 'test=test'})).read().decode('utf-8') +            self.assertIn('Cookie: test=test', res) + +            # Specified Cookie header should be removed on any redirect +            res = ydl.urlopen(sanitized_Request( +                self._test_url('308-to-headers'), headers={'Cookie': 'test=test'})).read().decode('utf-8') +            self.assertNotIn('Cookie: test=test', res) + +            # Specified Cookie header should override global cookiejar for that request +            ydl.cookiejar.set_cookie(compat_http_cookiejar_Cookie( +                0, 'test', 'ytdlp', None, False, '127.0.0.1', True, +                False, '/headers', True, False, None, False, None, None, {})) +            data = ydl.urlopen(sanitized_Request( +                self._test_url('headers'), headers={'Cookie': 'test=test'})).read().decode('utf-8') +            self.assertNotIn('Cookie: test=ytdlp', data) +            self.assertIn('Cookie: test=test', data)      def test_no_compression_compat_header(self):          with FakeYDL() as ydl: diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py index 58c710b08..c21cd3687 100644 --- a/youtube_dl/utils.py +++ b/youtube_dl/utils.py @@ -2968,7 +2968,6 @@ class YoutubeDLRedirectHandler(compat_urllib_request.HTTPRedirectHandler):          new_method = req.get_method()          new_data = req.data -        remove_headers = []          # On python 2 urlh.geturl() may sometimes return redirect URL          # as a byte string instead of unicode. This workaround forces @@ -2981,6 +2980,11 @@ class YoutubeDLRedirectHandler(compat_urllib_request.HTTPRedirectHandler):          # but it is kept for compatibility with other callers.          newurl = newurl.replace(' ', '%20') +        # Technically the Cookie header should be in unredirected_hdrs; +        # however in practice some may set it in normal headers anyway. +        # We will remove it here to prevent any leaks. +        remove_headers = ['Cookie'] +          # A 303 must either use GET or HEAD for subsequent request          # https://datatracker.ietf.org/doc/html/rfc7231#section-6.4.4          if code == 303 and req.get_method() != 'HEAD': @@ -2999,7 +3003,7 @@ class YoutubeDLRedirectHandler(compat_urllib_request.HTTPRedirectHandler):          # NB: don't use dict comprehension for python 2.6 compatibility          new_headers = dict((k, v) for k, v in req.header_items() -                           if k.lower() not in remove_headers) +                           if k.title() not in remove_headers)          return compat_urllib_request.Request(              newurl, headers=new_headers, origin_req_host=req.origin_req_host, | 
