What are your PC specs (RAM, CPU, GPU, etc.)? If I wait, will the program start working? How can I replicate this on purpose?
BAS use a malformed User-Agent
-
Hey all,
I would like to know if anyone has already investigated about the BAS user agent.
I note a malformation of the header "user agent" generated by BAS, i subscribed to FingerprintSwitcher, i tested thousands of different fingerprints, it seems that the problem is on BAS side, even to modify the header manually ( in BAS ...) there is always a malformed "user agent" header, you can do your test with http://f.vision/, here is the BAS User-Agent :

All the tests give the same result, with sometimes exception whit the user agent of Microsoft Windows Phone (chosen in FingerprintSwitcher) for this one the header is well formed, all combination with "windows/firefox,/chrome" returns a bad User-Agent.
Does anyone have a technical explanation in order to think about a possible workaround ... i already try to apply a new "User-Agent" after applying the fingerprint, always the same problem , it's malformed
needless to say, f.vision give a User-Agent well formed from a real navigator ....
I'm badly surprised that a tool so powerful as BAS have such a flaw for the User-Agent, which is by far the least complicated to implement in automation scripts compare to the other very efficient option from BAS.
thank you all
-
@jonigood, Fingerprint + old user-agent = No error, test project > ua_test.xml

-
@GhostZ , hi
It is not at all a question of old user-agent, it is only about the user agent of edge which work (old or new is same), the user-agent which you put in your example is a user-agent of edg , for chrome and firefox NO one user-agent works, old or new, for proof here are some tests:
firefox, verry old user-agent (do not work) :
Mozilla/5.0 (Windows; U; Win 9x 4.90; en-US; rv:1.7.9) Gecko/20050711 Firefox/1.0.5chrome, verry old user-agent (do not work) :
5.0 (Windows; U; Windows NT 6.1; en-US) AppleWebKit/525.19 (KHTML, like Gecko) Chrome/1.0.154.53 Safari/525.19edge, verry old user-agent (WORK) :
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/42.0.2311.135 Safari/537.36 Edge/12.246edge, new user-agent (WORK) : :
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.102 Safari/537.36 Edge/18.19582in no case can we accuse f.vision of false positive for firefox result, because all the tests with real FIREFOX browser pass well, never pass with BAS and firefox UA
conclusion, there is a problem of user-agent with BAS, it would be necessary to report the bug, thank you.
-
This a non-evolving way of thinking, the subject of the discussion is oriented on the fact that BAS is an automation tool which should accurately replicate the characteristics of a real browser, we simply observe that one of the functions of BAS (user -agent) does not react as it should, it doesn't matter if f.vision gives a fake also with real Chrome, the fact that we have different results on all other browsers shows that something is not working as it should be, it is undeniable
in a large scale production environment, having the ability to only one kind of user-agent is simply impossible.
You can do a test with our dear friend google with Recaptcha API (recaptcha__en.js), which unfortunately gives a poor score to BAS whit any user-agent other than edge :
https://recaptcha-demo.appspot.com/recaptcha-v3-request-scores.php
Whit a real browser whit get 0.7
Whit base wee get 0.1:

This low score surely does not come only because of a possible (or not...) problem of user-agent whit BAS, but it is certainly part of it ...
the same test with BAS and an EDGE User-Agent we get a better score, 0.3 .... :

f.vision is therefore not false, only the User-Agent of edge which are not detect whit BAS for the moment ...
thank you


