[bilibili] Add support for new URL schema with BV ids (closes #24439, closes #24442)

[youtube-dl] / youtube_dl / utils.py

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index b14603d8ac9e7ade80f8d90254239f814c091bdf..38262bee4cd08f97e4f5009d7066d1466bde25a4 100644 (file)
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -46,6 +46,7 @@ from .compat import (
     compat_html_entities,
     compat_html_entities_html5,
     compat_http_client,
+    compat_integer_types,
     compat_kwargs,
     compat_os_name,
     compat_parse_qs,
@@ -2728,6 +2729,11 @@ class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
 
 
 class YoutubeDLCookieJar(compat_cookiejar.MozillaCookieJar):
+    """
+    See [1] for cookie file format.
+
+    1. https://curl.haxx.se/docs/http-cookies.html
+    """
     _HTTPONLY_PREFIX = '#HttpOnly_'
 
     def save(self, filename=None, ignore_discard=False, ignore_expires=False):
@@ -2794,6 +2800,15 @@ class YoutubeDLCookieProcessor(compat_urllib_request.HTTPCookieProcessor):
     https_response = http_response
 
 
+class YoutubeDLRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
+    if sys.version_info[0] < 3:
+        def redirect_request(self, req, fp, code, msg, headers, newurl):
+            # On python 2 urlh.geturl() may sometimes return redirect URL
+            # as byte string instead of unicode. This workaround allows
+            # to force it always return unicode.
+            return compat_urllib_request.HTTPRedirectHandler.redirect_request(self, req, fp, code, msg, headers, compat_str(newurl))
+
+
 def extract_timezone(date_str):
     m = re.search(
         r'^.{8,}?(?P<tz>Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
@@ -3519,10 +3534,11 @@ def str_or_none(v, default=None):
 
 def str_to_int(int_str):
     """ A more relaxed version of int_or_none """
-    if int_str is None:
-        return None
-    int_str = re.sub(r'[,\.\+]', '', int_str)
-    return int(int_str)
+    if isinstance(int_str, compat_integer_types):
+        return int_str
+    elif isinstance(int_str, compat_str):
+        int_str = re.sub(r'[,\.\+]', '', int_str)
+        return int_or_none(int_str)
 
 
 def float_or_none(v, scale=1, invscale=1, default=None):