Open ghost opened 3 years ago
Hi @seth1 I'm not sure I follow here, so if you set your instances to .onion
ones you are still getting requests to the original service?
Hi, sorry I wasn't clear. This is a completely new feature, which might qualify as feature creep. I just think it would be useful.
In Tor Browser if you go to Options > Privacy and Security > Onion Services, Prioritize .onion sites when known, it will automatically redirect you to the onion version of a site if the site sends an "Onion Location" in the response headers.
You can see such a response header at https://www.nytimes.com/, even when not using Tor Browser.
I believe Brave Browser also has this feature. The problem with such an implementation is that the Browser does not store these in a database. Thus if you have forgotten what the onion location is (or clicked on a link without first copying it and modifying it) you must still go to the clearlink first before the Browser then reads the response header again, and redirects you.
I propose Privacy Redirect act as that sort of database so that the following happens:
Alternatively
Such a database could also be useful if someone just wanted to say point site1.com to site2.com because it's currently unimplemented, but that would probably be more useful as a different database if the user isn't actually using a browser capable of going to an onion.
Thanks for your work and hopefully I better explained it this time.
Presently in browsers like Tor for "Prioritize .onion sites when known" which will redirect the user to a site should the site have an onion domain in the header.
The problem of course is that this requires a request to the "normal" domain first, every time.
This might be feature creep but I think it would be useful for privacy redirect to autoredirect the site without sending a request. I can think of two ways. The first is to save a list of sites the user goes to, so the user must only request the normal domain the first time, privacy redirect will read the header, and add the onion to a database for the user, fixing it every time after that. The other way is for privacy redirect to have it's own database so the user never even needs to access the normal domain first, but this may require quite a bit of submissions and maintenance. There could also be a mix, providing a small default database and add as the user browses.
It might also be useful to have an option to prioritize onion domains for the other options (nitter/invidious/etc) although I recognize this is sort of already available by allowing the user to edit the random pool.