[tor-bugs] #20284 [Core Tor/Tor]: consensus weight case 2b3 does not follow dir-spec

[Tor Bug Tracker & Wiki]

#20284: consensus weight case 2b3 does not follow dir-spec
--------------------------+------------------------------------
 Reporter:  pastly        |          Owner:
     Type:  defect        |         Status:  needs_review
 Priority:  Medium        |      Milestone:  Tor: 0.3.0.x-final
Component:  Core Tor/Tor  |        Version:
 Severity:  Normal        |     Resolution:
 Keywords:                |  Actual Points:
Parent ID:                |         Points:
 Reviewer:                |        Sponsor:
--------------------------+------------------------------------
Changes (by nickm):

 * status:  merge_ready => needs_review


Comment:

 This would be trouble if we merged it as-is: it changes
 networkstatus_compute_bw_weights_v10, which is invoked from
 networkstatus_compute_consensus.  We need all the authorities who are
 using the same consensus_method to produce the same output for the
 consensus, or else it won't be a consensus.

 As a fix, I would suggest having networkstatus_compute_bw_weights_v10()
 take an argument that says what the actual consensus_method is, and
 allocating a new consensus_method for doing this bandwidth calculation
 correctly.  Also we'd need to document the old way and the new way in dir-
 spec.txt

 Also, mikeperry did most of the work on these equations; I'm hoping we can
 get his feedback about whether he believes more in the code or in dir-
 spec.txt.

--
Ticket URL: <https://troodi.torproject.org/projects/tor/ticket/20284#comment:3>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online