Yes, we can always edit original posts add thread titles for improved expression, more up to date related to latest developments.
lock_pref() , l
ockPref() might be a workaround for these two issues. I think thats’ what TBB devs use to keep users from changing prefs such as the security slider default. it will likely break Tor Browser but cant hurt to try.
Actually it was originally https://github.com/sleutho/QuickJS but its name was changed to https://addons.mozilla.org/en-US/firefox/addon/quick-js-switcher/
Yes but we yet have to post
document prefs / settings to set security slider level
(this exact posting of mine: TODO research and document - How to use Tor Browser for security not anonymity? How to use TBB using clearnet?)
These two interactions with upstream need to be next.
Why is it not enough to just set the desired slider level in a prefs file that you ship in the browser profile (anyway)?
I think using the prefs approach is the one you should pursue right now. Shipping an own profile with customizations won’t go away in the forseeable future.
If the above is true, maybe the security slider not honoring the custom profile pref is a bug? Or perhaps we are missing something (not setting the correct pref?).
Regardless thats’ the solution imo. Lets lead off by asking why the security slider is not honoring the pref that we set. No reason to say anything about Whonix or SecBrowser.
Yes. We need to ask for how to set the pref. We can say that we know
there will be no environment variable and have to go the profile route.
It could be in form of a bug report “setting this pref does not work as
expected” or as enhancement request “please document how to do…” or it
could be asked on the mailing list as a question. Dunno what’s best but
anything can work, I guess.
Not necessary. I got it!
I was using either to many user_prefs (
extensions.torbutton.security_custom) or they were in the wrong order when setting the security slider, which was breaking SecBrowser.
It looked sane but I should have added one at a time.
These prefs set the security slider to “Safest”
user_pref("extensions.torbutton.inserted_security_level", true); user_pref("extensions.torbutton.security_slider", 1);
It might set the slider to highest but does it also do the action that
slider would apply when it is manually being set to highest? I.e. would
it actually disable scripts globally when started in this configuration?
Guess not. AFAICT there is no way to sync the slider with a pref. Bug report.
I’ve messed around with this before, and yes it does. You can test by looking to see if some of these about:config settings are set https://2019.www.torproject.org/projects/torbrowser/design/#other-security
- High … and disables SVG images ( svg.in-content.enabled ).
Does not work.
I speculate that
usingthe security slider in an interactive (manual) way makes it do things. However,
startingwith security slider pref set to high does not make the security slider do things. The pref only tells the security slider how it should visually appear but it does not seem to (fully) call the function which does things.
I think svg.in-content.enabled isn’t used anymore. Setting the security slider to high manually doesn’t change that pref for me but it does change svg.disabled to true.
File location moved.
new file location:
I stand corrected. Looks like setting the slider to Safest “after” first start can cause prefs not to sync properly. When adding prefs to /usr/share/tb-updater/tb_without_tor_settings.js (setting prefs at first boot) NoScript and about:config “visually appear” to work as expected. Meaning NoScript sycs with security slider and about:config prefs are togled as expected.
Fixed by commit 38a59e819604de5018db5db54fb08fd9e1581d1f. Please open a new ticket for the Tor Launcher issue if you wish.
Is this essentially a revisit of the same issue or would fixing Tor Launcher work as well?
environment variable to skip TorButton control port verification https://trac.torproject.org/projects/tor/ticket/13079 was an enhancement request which was implemented. Please reference this ticket when writing a bug report. Now this feature broke, so I think it’s correct to create a bug report?
The bug could also be a big different. If I remember fight, previously I think this is still a torbutton, not tor-launcher feature. Upstream will require instructions how to reproduce this issue on Debian, i.e. which environment variables set. This presupposes packages tb-updater / tb-starter not being installed and reproduced on plain Debian.
Please open a new ticket for the Tor Launcher issue if you wish. Just a minor code style issue: checking existence of environment variable vs checking the value of an environment variable. Not worth a report.
No bug report needed. I found a solution. Will post soon.
Since editing Qubes documented is rather cumbersome(?), what about leaving wiki page https://www.whonix.org/wiki/SecBrowser at Whonix wiki where each of us can edit easily and only submitting a stub to Qubes documentation?
Patrick via Whonix Forum:
Since editing Qubes documented is rather cumbersome(?), what about
leaving wiki page https://www.whonix.org/wiki/SecBrowser at Whonix wiki
where each of us can edit easily and only submitting a stub to Qubes
Thats a great idea. I can put together SecBrowser basics and the
benefits of using SecBrowser as per https://2019.www.torproject.org/projects/torbrowser/design/, Security Slider etc. Gather fingerprint stats from https://browserprint.info/ and https://panopticlick.eff.org/ . The former because https://www.qubes-os.org/doc/w3m/ uses that test and mentions that the w3w browser has a fingerprint that stands out from Tor Browser i.e stands out in a crowd. I just tested SecBrowser default settings. Much better than expected
Within our dataset of several hundred thousand visitors tested in the
past 45 days, one in 76.46 browsers have the same fingerprint as yours.
Currently, we estimate that your browser has a fingerprint that conveys
6.26 bits of identifying information.
Just to be clear, no configuration instructions?