remove '&' from URL patterns

'/?&#' -> '/?#' and '?&#' -> '?#'

According to https://www.ietf.org/rfc/rfc3986.txt, URLs are
"organized hierarchically" by using "the slash ("/"), question
mark ("?"), and number sign ("#") characters to delimit components"
This commit is contained in:
Mike Fährmann
2020-10-22 23:12:59 +02:00
parent 1686dc1757
commit 968d3e8465
74 changed files with 158 additions and 158 deletions

View File

@@ -28,7 +28,7 @@ class XvideosGalleryExtractor(XvideosBase, GalleryExtractor):
archive_fmt = "{gallery[id]}_{num}"
pattern = (r"(?:https?://)?(?:www\.)?xvideos\.com"
r"/(?:profiles|amateur-channels|model-channels)"
r"/([^/?&#]+)/photos/(\d+)")
r"/([^/?#]+)/photos/(\d+)")
test = (
("https://www.xvideos.com/profiles/pervertedcouple/photos/751031", {
"url": "cb4657a37eea5ab6b1d333491cee7eeb529b0645",
@@ -94,7 +94,7 @@ class XvideosUserExtractor(XvideosBase, Extractor):
subcategory = "user"
categorytransfer = True
pattern = (r"(?:https?://)?(?:www\.)?xvideos\.com"
r"/profiles/([^/?&#]+)/?(?:#.*)?$")
r"/profiles/([^/?#]+)/?(?:#.*)?$")
test = (
("https://www.xvideos.com/profiles/pervertedcouple", {
"url": "a413f3e60d6d3a2de79bd44fa3b7a9c03db4336e",