-
-
Notifications
You must be signed in to change notification settings - Fork 172
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[BUG]: firefox: privacy.firstparty.isolate (obsolete) #239
Comments
Hi, I will do a patch to fix some crucial bugs, the patch after will include an improvement to fix this along with #232. I'm happy to see Linux users here. Linux will be first class citizen for privacy.sexy so I will appreciate any other feedback/improvement ideas you might have. |
firefox can be a tough one depending on how far you want to go with privacy i guess the easy way would be to recommend LibreWolf, the config of which is based on arkenfox the more intensive way would be to wget the arkenfox js but the user would need to go through it and maintain it (updates, cleaning) re: #232 - yeah, one should never edit prefs.js unless they're testing something, nor should one edit the arkenfox user.js - any changes for that need to get dumped in user-overrides.js and copied to user.js using their updater script, else updates can be a pain in the ass i maintain some privacy stuff if it's any help to you - feel free to use or link to stuff regarding script syntax, you may very well know more than me, but from what i understand there are some "best practices" that could be implimented, such as using printf instead of echo - i use the shellcheck linter to keep me in line and i love it |
I thought I'd leave these links as a primer for anyone like me asking "ok, now what the f..." https://hacks.mozilla.org/2021/02/introducing-state-partitioning/ From my reading, both methods are kind of still in development and somewhat untested. Personally I doubt anyone other than web devs and hardcore geeks are going to use this and it seems most of the hard core geeks are moving away from Firefox (myself included if I could find a better option). |
FPI is dead - nothing is being done with it far as i know - dFPI/ETP/net/state partitioning is still being developed, hence why it's not all of its features are enabled by default, but it works quite well regarding cross site auth, that can be controlled with
some of it is already enabled by default (ETP standard) and apparently more features will be enabled by default down the road - arkenfox enables ETP strict, but allows the compatibility component (SmartBlock) which is only a privacy issue if the user attempts to log on to a site which pulls required resources from a 3rd party - you can read about it here... Firefox 90 introduces SmartBlock 2.0 for Private Browsing - Mozilla Security Blog
that's not my impression, not if they care about privacy, but i could be wrong i would love to dump Firefox and get the hell away from the "woke"/radical leftist ideology at Mozilla, but there's no other browser out there that's better suited for privacy that i'm aware of that isn't heavily based on Firefox LibreWolf, while technically a fork perhaps, is really a few scripts run on Firefox releases and anything that doesn't support manifest v2 (chromium) is out, though i see that Raymond is working on a manifest v3 version of uBO |
...just to expand on that front a bit, the browser scene is in a rather sad state at the moment and i surely don't see that changing - the web has turned to crap and thus so must the web browser in order to deal with it then we have one good, well maintained content blocker - uBO (though there is this and this) - that's essentially run by one guy that only works to its full potential on one browser - Firefox - whose market share has been plummeting for years |
How should we approach it here? How about keeping
Thanks for the context and link, I extracted configuration descriptions from the link you shared. Out-of-topicRecommended installs: This is hard when we do not know which package manager users would prefer. I prefer flatpak, and would not like any other software to install stuff using e.g. Shellcheck: Thanks for the recommendation, it's a great tool. I want to automate shellcheck linting, this was brought up #164 as well. However there are a lot of stuff to do so this did not get love yet :) I'd appreciate any PR. 12bytes.org: I've stumbled upon your blog before. Great to have you here as contributor, all of your feedback is greatly appreciated, also the work on the blog. Questioning encryption is useless is mind-opening. Adding it to my RSS feed :) @TheAndr0id I run Ungoogled Chromium and sometimes Brave. I'm looking for a Thunderbird alternative but could not find anything close to it yet. Mozilla lost me after advocating censoring. Firefox is basically controlled by Google and sending visited URLs (hashing as privacy is a joke here) and downloaded files by default to them is a no-go for me. |
i'm not sure that running an older non-JS browser is necessarily a bad thing, but if they're running an older version of FF that supports FPI (meaning it's not all that old), i would think that may be a very bad idea and personally i would not support such a configuration FPI is not being worked on anymore, so i would do 2 things: first and foremost, make sure privacy.firstparty.isolate is disabled and 2, enable dFPI (you surely do not want to enable both) as for the prefs involved, you can easily grab them from the arkenfox js, or just download the whole thing along with their updater and prefsCleaner scripts (they have versions for both winblows and linux) regarding FF telemetry, all of it can be disabled, including Safe BRowsing, but not all of it should be regarding T-bird, the nice thing about it, beyond the fact that, like FF, it's super configurable, is that it makes encrypting mails almost as easy as it can be with it's built-in support which no longer requires a complicated add-on regarding Brave, personally i won't touch it - i don't feel good about that corporation plus they are manipulating results for their search engine in unethical ways personally, if i drop FF, at this point i would likely go with either LibreWolf or the Mullvad browser ... but without enabling Tor since i don't trust that either lol |
Sorry, I didn't mean to start a browser war - really! I'm currently rebuilding the family Windows 7 box to Windows 10 because of Steam. I need to have solutions that are either well hidden or easy to understand that have minimal impact on functionality. Explaining how cookie isolation and supercookies work is just not an option. From my point of view Firefox has been on a serious downhill slide for a few years now. They seem to be determined to push away all of their 3rd party extension developers and have annoyed most of their main users with senseless UI changes. I only run their ESR releases since their release rate is crazy (which speaks to poor project management). I fully expect a release titled "Firefox Chrome - powered by Google!" real soon now. Ungoogled Chrome scares me - kind of like having a pet cobra. It might be cool and all, but then, well you're dead... When I read about the crypto currency attached to the Brave browser? Nope. Nope. Nope. The problem is there isn't a lot of choices and turning off these intrusive data collection schemes are not trivial. Thank you @undergroundwires for doing all the work you are doing on this (and many other fronts)! |
no worries - you didn't
agreed! their market share has tanked and i think part of that is due to idiotic decisions by the board and part due to googles influence and promotion of their trashy browser that said, my POV is that, unfortunately, FF is still the best potential candidate for privacy geeks (with LOTS of tweaking) it seems to me that some of the FF devs are very serious about privacy (Tor uplift is a big part of that), however that attitude is clearly not shared by corporate, else they'd divest from google and all of the other ethically crippled corporations they continue to partner with here's a very interesting read by Lunduke... Firefox Money: Investigating the bizarre finances of Mozilla |
@atomGit Could you clarify why we should remove |
FPI was replaced with ETP - enhanced tracking protection - this does what FPI did + more - from the arkenfox js... /*** [SECTION 2700]: ETP (ENHANCED TRACKING PROTECTION) ***/
user_pref("_user.js.parrot", "2700 syntax error: the parrot's joined the bleedin' choir invisible!");
/* 2701: enable ETP Strict Mode [FF86+]
* ETP Strict Mode enables Total Cookie Protection (TCP)
* [NOTE] Adding site exceptions disables all ETP protections for that site and increases the risk of
* cross-site state tracking e.g. exceptions for SiteA and SiteB means PartyC on both sites is shared
* [1] https://blog.mozilla.org/security/2021/02/23/total-cookie-protection/
* [SETTING] to add site exceptions: Urlbar>ETP Shield
* [SETTING] to manage site exceptions: Options>Privacy & Security>Enhanced Tracking Protection>Manage Exceptions ***/
user_pref("browser.contentblocking.category", "strict");
/* 2702: disable ETP web compat features [FF93+]
* [SETUP-HARDEN] Includes skip lists, heuristics (SmartBlock) and automatic grants
* Opener and redirect heuristics are granted for 30 days, see [3]
* [1] https://blog.mozilla.org/security/2021/07/13/smartblock-v2/
* [2] https://hg.mozilla.org/mozilla-central/rev/e5483fd469ab#l4.12
* [3] https://developer.mozilla.org/en-US/docs/Web/Privacy/State_Partitioning#storage_access_heuristics ***/
// user_pref("privacy.antitracking.enableWebcompat", false);
/* 2710: enable state partitioning of service workers [FF96+] ***/
user_pref("privacy.partition.serviceWorkers", true); // [DEFAULT: true FF105+]
/* 2720: enable APS (Always Partitioning Storage) ***/
user_pref("privacy.partition.always_partition_third_party_non_cookie_storage", true); // [FF104+] [DEFAULT: true FF109+]
user_pref("privacy.partition.always_partition_third_party_non_cookie_storage.exempt_sessionstorage", false); // [FF105+] [DEFAULT: false FF109+] as you can see,
you should really go through the arkenfox js and see what all you want to employ the other crucial thing is to disable JS globally, then allow per domain as needed you might want to consider creating a links page to link to info that you can't do with scripts - the arkenfox repo and wiki would be a good start and apparently my stuff is well liked, but also Alicia's repos, Namoi's stuff, and i can recommend more |
New configurations by privacy.sexyBased on the documentation about disabling dynamic state partitioning 1, the
Given this, I think we should recommend setting the value to Additionally, we should enable "Firefox Enhanced Tracking Protection" 3 through the These are main toggles that should be introduced along with others in source from @TheAndr0id. In the search of official documentation
1. FPI conflicts with dynamic First-Party Isolation (dFPI) [1][5], and Mozilla has no plans for compatibility [2].
2. FPI (`privacy.firstPartyIsolate`) is depreciated [1], with dFPI expected to become the default [3].
3. dFPI, along with network partitioning, is essentially a superset of FPI [1] [4].
[1]: https://bugzilla.mozilla.org/show_bug.cgi?id=1649876#c0
[2]: https://bugzilla.mozilla.org/show_bug.cgi?id=1649876#c3
[3]: https://bugzilla.mozilla.org/show_bug.cgi?id=1649876#c5
[4]: https://bugzilla.mozilla.org/show_bug.cgi?id=1637344#c2
[5]: https://bugzilla.mozilla.org/show_bug.cgi?id=1631676#c25 This information supports all your points, @atomGit, and I believe should convince you, @TheAndr0id, that adjusting FPI setting to Addressing
|
that list may not be complete
while that is technically true, i would add that the core members are a very knowledgeable bunch who, to my knowledge, as well as considering statements they made, have access to some key people in the Tor and Firefox development teams
perhaps the arkenfox prefsCleaner script might be of use? only other thing i would suggest is enabling |
Great that you shared what the master switch
I will do change on what we discussed and share on this thread and would appreciate feedback before publishing the code. I have no doubt arkenfox and similar projects sit on deeper knowledge about configurations in detail that privacy.sexy lacks for now, but we are stronger here with you :). Regarding cleanup, I will just generate a privacy.sexy script by clicking on revert on UI for all prefs.js scripts. |
To fix the bug and avoid scope creep, I plan to do this fix: .
├── (category) Enable Firefox state partitioning (Total Cookie Protection)
│ ├── (script) Enable Firefox dynamic storage partitioning
│ ├── (script) Disable depreciated Firefox First-Party Isolation (FPI)
│ ├── (script) Enable Firefox network partitioning
I'm looking forward for your feedback before releasing this. I've seen that you did considerable contributions to arkenfox @atomGit, so if you have the time, I'd appreciate if you check out the documentation and let me know if I misinterpreted something:
|
looks very good to me, but i don't claim to be an expert - if you really want to confirm, maybe ask the dudes at arkenfox |
- Replace obsolete "Firefox First party isolation" with "Firefox state partitioning". - Add comprehensive documentation for the new scripts. - Introduce enabling dynamic First-Party Isolation (dFPI) - Disable deprecated First-Party Isolation (FPI) to avoid conflicts with dFPI. - Add script to enable Firefox network partitioning to cover functionality of older FPI script.
- Replace obsolete "Firefox First party isolation" with "Firefox state partitioning". - Add comprehensive documentation for the new scripts. - Introduce enabling dynamic First-Party Isolation (dFPI) - Disable deprecated First-Party Isolation (FPI) to avoid conflicts with dFPI. - Add script to enable Firefox network partitioning to cover functionality of older FPI script.
The fix is released in 0.12.5 🚀 |
This commit modifies the blocking behavior of `r.bing.com` due to its extensive use across multiple Windows features, including Copilot and Maps. Previously, included in the 'Cortana and Live Tiles' block list, this host was causing issues for Copilot functionalites, as noted in issue #239. By excluding `r.bing.com` from block list, this update aims to prevent unintended disruptions without compromising the privacy gains of other scripts. Changes include: - Exclude `r.bing.com` from the "Cortana and Live Tiles" block list. - Improve documentation to clarify the role and exclusions. - Improve documentation with consistent header for blocked hosts.
`r.bing.com` is used by multiple Windows features including Copilot and Maps. privacy.sexy was disabling this host as part of disabling Cortana and Live Hosts which was causing unexpected behavior with Copilot users (see #239). This commit excludes disabling of r.bing.com due to its cross-cutting usage and documents the reasoning, resolving #239.
privacy.firstparty.isolate
is obsolete and has been for some time - it has been replaced with network partitioningsee the arkenfox user.js
The text was updated successfully, but these errors were encountered: