FreeBSD Desktop failing "human" detector websites

Keve

Keve

Those of you who use FreeBSD as a daily desktop system, could you please try and see if your web-browser can access the FreeBSD bug report form. My recent experience is that the Anubis thingy rejects me as a human being every time I connect from my FreeBSD 15.1R desktop. Tried Firefox and Chromium, same result. If I connect from my MacBook (same home network, behind the same NATed IP) then I pass the Anubis test and get access to the bug reporting form.

I recently had similar issues with sites "protected" by CloudFlare's AI/BOT/crawler detector. Hence I have the feeling that browsers on FreeBSD are being discriminated by these filters. What is your experience?
 
Waterfox, firefox, librewolf, chromium all seem to work fine for me on 15.1. If you have any addons/extensions enabled try disabling first.
Other websites that use cloudflare I've had trouble but clearing cache and cookies would fix things.
So maybe not "FreeBSD" but stale data. A quick try is "use a private window"
 
OK, that is good enough. So it is not something about FreeBSD 15.1 if it works for many of you.
I don't think it is stale data on my side because I do not really use Chromium, I just keep it as a backup, no extensions or plugins, no settings I changed. It should have worked.
But, now that results keep pointing back at my env, I will try using a clean new user account and see how that connects.
 
New user account, nothing but a .xinitrc, ran `startx`, launched Firefox and Chromium, both fail the Anubis test. Even tried adding "freebsd.org" as an exception domain for cookie and other default protection settings. No improvement.
I will try a few other tests in the hope of narrowing down what may cause this.
 
Yes. I've had this excessively. I've had to change multiple browsers, Anubis is firing on every single page. Even on the wiki, and when trying to access (mostly) passive content like the wiki. On FreeBSD itself, in particular it comes back and complains cookies are disabled - they aren't.

These days many of these tools massively use heuristics, are you using a VPN or on 5G by any chance? 5G causes problems for me, because I'm behind CG-NAT => my IP is not stable, and some tools are behind the curve and assume a stable IP. Therefore I sometimes use cloudflare warp, which then can backfire because my IP is now stable but coming from a datacentre. It can be excessive often and I just give up and take my "custom" elsewhere. With all the OTPs etc. it becomes ridiculous simply just trying to view some websites or my own account.

But yeah, my experience has been that the anubis settings are being way too aggressive against my setup.
 
Keve said:
New user account, nothing but a .xinitrc, ran `startx`, launched Firefox and Chromium, both fail the Anubis test. No improvement.
I will try a few other tests in the hope of narrowing down what may cause this.
Click to expand...
A new browser profile will trigger more fails. As you use your it more you get less suspicious(?) That way on linux also.
Why, who knows.
 
MG said:
I wonder if it also works if you feed it random mouse movements. O:‑)
Click to expand...
I tried that. These "crawler" filters do not work like CAPTCHA. It is not your current actions they test. I tried moving my mouse and providing illogical activity, but I still fail the test and my access gets blocked.
 
Keve said:
I tried that. These "crawler" filters do not work like CAPTCHA. It is not your current actions they test. I tried moving my mouse and providing illogical activity, but I still fail the test and my access gets blocked.
Click to expand...
It doesn't have server-side feedback to a profiling/advertising entitty?
These things always work immediately on me. Maybe it's noticed soon because I'm left handed but always use the mouse right handed which causes discoordinated human movement.
 
Almost no problem here with configureations below.
  • from Japan
  • using firefox on FreeBSD stable/15 and main, amd64
  • IPv4 only Internet connection with single global IP address provided
  • NAPT is done only on my home router (no CG-NAT)
Only issue I've encountered with the configuration above is that sometimes refreshing search results errors out after passing anubis. Not always.
 
Keve said:
Those of you who use FreeBSD as a daily desktop system, could you please try and see if your web-browser can access the FreeBSD bug report form.
Click to expand...

Works for me (Daily Driver Desktop): FreeBSD Release 15.1 / Firefox (latest/up-to-date).

You might check to see if the remote site doesn't like your Internet source address -- I actually find that issues with the Internet source address are (A LOT) more common than issues with what you are surfing with. Good luck !
 
> A new browser profile will trigger more fails. As you use your it more you get less suspicious(?) That way on linux also.
Agreed, but for me, it's triggering every page. AFAIK anubis is meant to be like this; even on private browsing, it should set a cookie or similar and this will then prevent the need to bring it back up again in the future. But on multiple machines I'm seeing it trigger every single page load. Even when logged in. My guess is it's not entirely configured optimally, or it's having difficulty integrating with presumably older software (i.e. bugzilla has been aorund for a long time)

> Why, who knows.
I think they're incredibly heuristics-driven. Having written scrapers before, especially nowadays, they are very non-deterministic.
 
Anubis uses a javascript program that runs in your browser for detection. Make sure you don't turn off javascript when you wish to visit Anubis-protected websites.
 
You must log in or register to reply here.
Back
Top