I'm trying to use one of these Firefox extensions suggested by ghacks wiki:
The problem is: these addons won't work if I block images in uMatrix, at all. Hence, I'd like to know why is that.
expected result
what's actually happening
disclaimer
Do these addons work by preventing first-party image ETags or third-party as well?
Is third party ETags even something?
Is it a CSP problem?
That test site is wonky and kinda flawed (not going to get into it): the addon works, the script should: use
click here and read the bit that says ETag
So, to test with that site you either have to change your IP or User-Agent (and not block the image!):
@earthlng
So, to test with that site you either have to change your IP or User-Agent (and not block the image!)
The reason I blocked images globally is because I feared image tracking, so I'd activate images only when necessary.
Maybe this increases entropy and I'm acting stupid. Right?
Furthermore, I've tried mr. Pants suggestion website, as well. It won't track me, only on these circunstances:
If I would go and try 2nd case: website would show I am not being tracked (sometimes it shows that I'm, somehow).
You see: my fear is not that Stoppa is not doing its job...
It's if other extensions are disrupting it or producing _false negative_ results.
In the case of _false negatives_ are not real: wouldn't it be better to just not install Stoppa and just have uMatrix + uBlock?
Maybe this increases entropy and I'm acting stupid. Right?
IDK. It's definitely detectable but I doubt anyone's looking at that in their logs, but I could be wrong.
Personally I allow 1st party CSS and images by default and then allow more where necessary:
* * * block
* * css block
* * image block
* 1st-party css allow
* 1st-party image allow
Pants' suggested website doesn't work for me either. Not sure why. It looks like the etag is on the SVG but for some reason my FF doesn't re-request that image when I click the refresh button on that page. It always says 0 visits and displays a new etag.
IDK if that's an SVG-specific issue or WTH is going on.
AFAIK uBO has some kind of cache-busting included to make sure things are always blocked correctly. IDK if that has something to do with it.
I don't use an ETag specific extension because FPI and Temporary Containers combined are IMO enough to stop any meaningful cache-related tracking attempts. The Temporary Containers self-destruct after closing the last tab of a particular domain and additionally I clear the cache on shutdown just for good measure.
It's definitely detectable but I doubt anyone's looking at that in their logs, but I could be wrong.
That's if you block JS! With JS enabled it's easily detectable and could be used to FP or track you. But since barely anyone blocks images by default, I don't think anyone's using an onerror listener on images or similar for tracking purposes.
Pants' suggested website doesn't work for me either
Both sites worked for me, and for @claustromaniac (at least they did about a year ago when I showed him the second site that wasn't as "wonky" - that lucible causes people confusion: even we had a time working it out)
re-opening to sort it out. will test when I get round to it
works for me: https://privacycheck.sec.lrz.de/passive/fp_etag/fp_etag.php
I'm always unique
that doesn't necessarily mean the site is working. I was trying my best to be non-unique but couldn't.
I've cache enabled, disabled uBO + uM, no etag specific extension; still wasn't able to make it count my visits. Can you?
works for me: https://privacycheck.sec.lrz.de/passive/fp_etag/fp_etag.php
Or Firefox made some changes or this site doesn't work even on a vanilla profile.
But it does work on a Edge Chromium.
that doesn't necessarily mean the site is working. I was trying my best to be non-unique but couldn't.
I've cache enabled, disabled uBO + uM, no etag specific extension; still wasn't able to make it count my visits. Can you?
True. I forgot about doing a control test. In a (practically) vanilla profile
cleared everything
then
IDFK
If you want to test it's working then inspect the headers
I tested on a brand new default profile and it can't track me, at all.
Do you think it has something to do with this FF build being the default installed on Fedora 32?
with uMatrix (1st party green):
https://privacycheck.sec.lrz.de/passive/fp_etag/fp_etag.php
\^ - Always unique for me
https://lucb1e.com/rp/cookielesscookies/
\^ - works as expected
I confirmed this behavior on Vivaldi(Chromium) and Firefox - it is the same for both browsers.
???
I think first one is broken.
Hi
It appears from the content of this thread that the author assumes that Etag is part of the downloaded images, but this is a mistake.
Etag is part of the HTTP header, and it's so nasty that it works regardless of anything, neither cookies nor even javascript. You just need to download even plain HTML / TXT document.
My tests showed that the ClearURLs add-on cleans Etag well:
https://addons.mozilla.org/pl/firefox/addon/clearurls/
Maybe I helped someone.
hey can i ask something about ETag ? i have been using clearurl with the etag option on since years, but i think i can disable etag filtering in clearurl as i'm using Temporary Containers (auto mode), FPI and ram cache only, right?
I don't use TC so I'm not sure what auto-mode is. FPI isolates cache, so we're really only taking about repeat visits to a first party per session. If auto-mode means siteA visit-1 is isolated from siteA visit2, then yeah, you can probably disable it. Note though, that if you're not changing circuits/exit-nodes/ip-address in that session then it's probably a moot point (assuming they are linkifying by IP). But it can't hurt to wipe ETags, but on the other hand, I don't think too many parties would track this way - there's just way too much lower hanging fruit
If it was me, just keep wiping those ETags: leave nothing to chance
@Thorin-Oakenpants
Automode means every site you follow opens in its own container by default, except if your custom rules says differently.
@lazyletucce
It should be enough, but it is not... see https://github.com/stoically/temporary-containers/issues/394
Cheers
Most helpful comment
Is it a CSP problem?