[tor-bugs] #8081 [Tor]: 7802 tweaks per code review

[Tor Bug Tracker & Wiki]

#8081: 7802 tweaks per code review
--------------------+-------------------------------------------------------
 Reporter:  andrea  |          Owner:                    
     Type:  defect  |         Status:  needs_review      
 Priority:  normal  |      Milestone:  Tor: 0.2.4.x-final
Component:  Tor     |        Version:                    
 Keywords:          |         Parent:                    
   Points:          |   Actualpoints:                    
--------------------+-------------------------------------------------------

Comment(by mikeperry):

 Hrmm. I've been on a lousy wireless network for the past couple days, and
 it does seem like my use counts are lower as compared my normal reliable
 connection. I'm looking at 17/25 (68%) right now for the lowest, which is
 below the current warn threshold of 70%. Perhaps I should lower the notice
 threshold to 80% and the warn to 60%? Or 75 and 50?

 I've also been thinking that one way we can consolidate the
 pathbias_measure_ functions is to make just one function that takes the
 product of the build bias and the use bias and determines the overall
 fraction of Tor network paths you're effectively able to use. This would
 solve your duplicate code concerns for these functions, and give us
 something a little easier to understand in terms of the actual effect of
 the overall failure rate on users' routing.

 I am reluctant to do this for 0.2.4.x because I think we want to see how
 various users experience these thresholds independently. I also suspect
 that the circuit build failures will go way down as we deploy ntor and
 make the onionskin queue size time-based, so we'll want to observe and
 tune that independently first, too. Separate ticket + XXX?

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/8081#comment:9>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online