Closed hober closed 4 years ago
Thinking about this some more, I think what's missing here is some caching of this state on agent clusters or some such.
In particular, imagine the user opens https://site-a.example/
twice in different browsing context groups (i.e., no opener relation), called 1 and 2. And site-a embeds https://site-b.example/
.
So site-b in 1 requests storage access and gets it. If site-b in 2 does hasStorageAccess()
at this point it should not have changed, I think. If the user reloads 2, it should have changed.
Thinking about this some more, I think what's missing here is some caching of this state on agent clusters or some such.
In particular, imagine the user opens
https://site-a.example/
twice in different browsing context groups (i.e., no opener relation), called 1 and 2. And site-a embedshttps://site-b.example/
.So site-b in 1 requests storage access and gets it. If site-b in 2 does
hasStorageAccess()
at this point it should not have changed, I think. If the user reloads 2, it should have changed.
I think you're right. I'd like to address this in a followup patch, if that's okay.
Sure.
The change regarding https://github.com/privacycg/storage-access/issues/26 LGTM.