Fix html filter autolink regex

Two links in the text like "http://foo.org ... http://bar.org"
were seen as a single match.
This commit is contained in:
miruka 2019-07-07 03:19:51 -04:00
parent 472b8fe993
commit 2923d440d0

View File

@ -12,8 +12,8 @@ from html_sanitizer.sanitizer import Sanitizer
class HtmlFilter: class HtmlFilter:
link_regexes = [re.compile(r, re.IGNORECASE) for r in [ link_regexes = [re.compile(r, re.IGNORECASE) for r in [
(r"(?P<body>.+://(?P<host>[a-z0-9._-]+)(?:/[/\-_.,a-z0-9%&?;=~]*)?" (r"(?P<body>[a-zA-Z\d]+://(?P<host>[a-z\d._-]+)"
r"(?:\([/\-_.,a-z0-9%&?;=~]*\))?)"), r"(?:/[/\-_.,a-z\d%&?;=~]*)?(?:\([/\-_.,a-z\d%&?;=~]*\))?)"),
r"mailto:(?P<body>[a-z0-9._-]+@(?P<host>[a-z0-9_.-]+[a-z]))", r"mailto:(?P<body>[a-z0-9._-]+@(?P<host>[a-z0-9_.-]+[a-z]))",
r"tel:(?P<body>[0-9+-]+)(?P<host>)", r"tel:(?P<body>[0-9+-]+)(?P<host>)",
r"(?P<body>magnet:\?xt=urn:[a-z0-9]+:.+)(?P<host>)", r"(?P<body>magnet:\?xt=urn:[a-z0-9]+:.+)(?P<host>)",