News for hackers. Stuff that matters.
Re:First-party isolation in Firefox: what breaks if you enab... by eslaught on Sunday September 09, 2018 @06:19PM (#17948095)
Is there a test I can use to confirm that it's working? I've set privacy.firstparty.isolate true and privacy.firstparty.isolate.restrict_opener_access true and when I log in to Github followed by Travis, Travis was able to log in without prompting for a password....

Firefox 62 macOS.

Edit: I did lose all my cookies on restart, so I do believe the option is at least enabled. Still would like to test that it's actually doing something.

Re:First-party isolation in Firefox: what breaks if you enab... by tomrittervg on Monday September 10, 2018 @01:35AM (#17949613)
Go to https://ritter.vg/misc/ff/fpi.html On first load it should say "There was nothing in local storage."

Now go to https://rittervg.com/misc/ff/fpi.html On first load it should say the same. If it says the same timestamp that was stored on the first page - it's not working.

Source: I'm a Mozilla Developer who is one of the primary devs/supporters of First Party Isolation.

Re:First-party isolation in Firefox: what breaks if you enab... by user812 on Monday September 10, 2018 @07:52AM (#17951078)
What if the box is empty? JS is allowed. (Edit: I guess the culprit is "third party cookies blocked by default")

So wouldn't a better test be about a third party that was used in a first party context before? Since FPI goes beyond third party cookies.

Re:First-party isolation in Firefox: what breaks if you enab... by tomrittervg on Monday September 10, 2018 @10:01AM (#17951833)
Thanks for diagnosing that for me, you're right blocking third party cookies does cause it to fail.

Both tests are equally valid. I just gave one because trying to be exhaustive about testing it would be mind-numbing. The test I provded only does localstorage, but FPI also isolates DNS cache, H2, image cache, favicons, cookies, localstorage, indexdb, etc etc

You can do yours by visiting https://anonymity.is/misc/ff/fpi-iframe.html first; then visit the ritter.vg and rittervg.com links.

Re:First-party isolation in Firefox: what breaks if you enab... by user812 on Monday September 10, 2018 @10:35AM (#17952091)
Thanks for the clarification.

What surprises me the most is that not only Firefox but also my Safari Browser passes all those tests when ITP is enabled.

Re:First-party isolation in Firefox: what breaks if you enab... by tomrittervg on Tuesday September 11, 2018 @03:30PM (#17962552)
Safari by default has a stricter storage access policy by default for all third-party domains, which requires you to visit the domain as a first party first. So it's probably that rather than ITP.
Re:First-party isolation in Firefox: what breaks if you enab... by user812 on Wednesday September 12, 2018 @04:36PM (#17973018)
I have a general question if you don't mind. I use Firefox Beta. Why is Firefox going the route of a manual blacklist (disconnect) instead of working on some kind of programmatic machine-learning/somewhat intelligent third-party storage blocking by default that doesn't discriminate known against unnkwon trackers?
Re:First-party isolation in Firefox: what breaks if you enab... by fredsir on Monday September 10, 2018 @02:57AM (#17949916)
Seems to be working, thanks! (had to disable blocking of third-party trackers for it to function, but after that, it works as promised, and I have re-enabled blocking of third-party trackers)
Re:First-party isolation in Firefox: what breaks if you enab... by Aeolun on Sunday September 09, 2018 @07:12PM (#17948296)
If Travis redirects to github, I assume github would get access to it’s own cookies again, and then be able to perform oauth, after which it redirects to Travis again with a token in the URL, no cookies or local storage needed as far as I’m aware.
Re:First-party isolation in Firefox: what breaks if you enab... by floatingatoll on Sunday September 09, 2018 @07:05PM (#17948271)
Did you restart between changing the settings and doing your tests? If not, they’re invalid, and you should repeat them.