[tor-bugs] #14881 [Core Tor/Tor]: incorrect defaults when producing bandwidth-weights line in directory footer

[Tor Bug Tracker & Wiki]

#14881: incorrect defaults when producing bandwidth-weights line in directory
footer
-------------------------------------------------+-------------------------
 Reporter:  robgjansen                           |          Owner:  pastly
     Type:  defect                               |         Status:
                                                 |  needs_revision
 Priority:  Medium                               |      Milestone:  Tor:
                                                 |  0.2.9.x-final
Component:  Core Tor/Tor                         |        Version:  Tor:
                                                 |  0.2.7
 Severity:  Normal                               |     Resolution:
 Keywords:  027-triaged-1-in, 028-triaged,       |  Actual Points:
  pre028-patch, tor-sponsorU-orphan,             |
  TorCoreTeam-postponed-201604, review-group-8   |
Parent ID:                                       |         Points:  3
 Reviewer:  mikeperry                            |        Sponsor:
                                                 |  SponsorU-can
-------------------------------------------------+-------------------------

Comment (by teor):

 Replying to [comment:39 pastly]:
 > New rebased branch includes changelog spacing fix and 23->24.
 >
 > https://github.com/pastly/public-tor/tree/ticket14881-v2
 >
 > - 1/10000 does not sound like a lot to me. If this is a big enough
 concern, we'll have to figure something else out.

 I'm not worried about the inaccuracy, but I'd like to know what the impact
 is - where does the extra weight change which nodes we choose, and how
 often? Which nodes get chosen more often because of it?

 > - I have not done any testing regarding which code paths differently
 configured relays take. I use Shadow for the work I'm currently doing, and
 Shadow does not support simulating a network with different versions of
 Tor. I have never used Chutney. FWIW, a homogeneous network executes the
 new method fine.

 Thanks for checking this. The public Tor network almost always runs with
 mixed authority versions, so we must test this case using chutney or a
 similar tool.

 Can you write unit tests for each case with the new defaults?
 (Are there unit tests for each case with the old defaults? We should test
 both.)

 Have you tested using actual bandwidth weights (like the ones in the
 current Tor network)?
 A unit test would be good for this too.

 We're almost there, thanks for helping out with this.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/14881#comment:40>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online