3 ms·
#1 this should be on by default. I might be missing something, but do sites really need the referrer? What would break if the browser sent the same page as refe
by alkonaut 9y ago
#1 this should be on by default. I might be missing something, but do sites really need the referrer? What would break if the browser sent the same page as referrer, or google.com/ or something similar?
Is there any value in the referrer to the client? The host can use it for a whole range of reasons - but apart from helping the host, what is the immediate benefit to the client?
#2 Won't this be possible to bypass simply by encoding more in the domain part of the url than in parameters? So you switch from a.b.tld/foo?p=123 to 123.a.b.tld/foo ?
- rocqua 9y agoTo answer #1, the referrer could probably be used to make a link refer to a login page on another domain, and to then let the host redirect to the correct page after login. There are of course ways to do this by following your suggestion #2, but I suspect this might be why we originally had the referrer header.
- alkonaut 9y agoThird party logins are definitely a use case. But you'd think that they would aim to do it with logins.somewhereelse.com/?u=source instead of relying on the referrer field as the source to go back to! I wonder how common actual use of referrer is?
- rovek 9y agoYou're right about subdomains but they are stripping both query and path, so 123.a.b.tld/foo would become 123.a.b.tld. It's fair to assume most misplaced sensitive data will be in the path or query, rather than the subdomain.
- alkonaut 9y agoI didn't even think about the subdomain until I saw the image that is in the article here - which literally looks like it added an arbitrary number in the subdomain, duplicating the parameter number! When I saw the image I thought the article was actually about blocking some shady way of leaking that doubleclick had invented.