[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [tor-bugs] #6647 [Tor Client]: pathbias_get_scale_factor() is using the wrong default and also causes rounding errors (was: pathbias_get_scale_factor() is using the wrong default)
#6647: pathbias_get_scale_factor() is using the wrong default and also causes
rounding errors
-----------------------------+----------------------------------------------
Reporter: mikeperry | Owner: mikeperry
Type: defect | Status: needs_revision
Priority: major | Milestone: Tor: 0.2.3.x-final
Component: Tor Client | Version:
Keywords: MikePerry201208 | Parent:
Points: | Actualpoints: 2
-----------------------------+----------------------------------------------
Comment(by mikeperry):
Ok, I've thought about this for a while. I see two ways forward on this
front:
1. We can go back to the divisor, and only scale 'if ((first_hop %
divisor) == 0 && (success_count % divisor) == 0)'. For this route, a
default divisor of 2 makes the most sense, I think.
2. We store a decimal value for the PathBias rate in the state file and
actually do a moving average for real instead of screwing around in
integer land.
I think my stabby sense is telling me Nick would prefer the modulus route
for 0.2.3.x for now, since it is the least amount of code change from the
original implementation. I think we should also use a divisor of 2 instead
of 4. That way, if both values of success_count and first_hop are even, we
scale. Otherwise, we wait until they are both even.
Again, I'm really sorry my fast-and-loose approach to software development
has exposed you to so much noise on this front, Nick... I promise to do
the floating point implementation and write a real proposal for the
feature for 0.2.4.x.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/6647#comment:6>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs