[tor-bugs] #7646 [Tor]: fix/enhance getinfo ns/id/* commands

[Tor Bug Tracker & Wiki]

#7646: fix/enhance getinfo ns/id/* commands
------------------------+-----------------------------------
     Reporter:  mr-4    |      Owner:
         Type:  defect  |     Status:  needs_review
     Priority:  normal  |  Milestone:  Tor: 0.2.5.x-final
    Component:  Tor     |    Version:  Tor: 0.2.4.7-alpha
   Resolution:          |   Keywords:  tor-client controller
Actual Points:          |  Parent ID:
       Points:          |
------------------------+-----------------------------------

Comment (by atagar):

 > Currently, it only returns information about tor nodes which have
 microdescriptors and ignores those that use "normal" descriptors, in other
 words, it behaves exactly like the md/id/node command.

 I'm not clear - (a) do you mean that it provides **router status entries**
 if and only if microdescriptors are available, or (b) it provides actual
 microdescriptors like **md/id/node**?

 > ... this command in control-specs.txt it is supposed to return V2
 directory info on all nodes without making such distinctions.

 This was
 [https://gitweb.torproject.org/torspec.git/commitdiff/d2b7ebb098c5cae9c40aaaf2bbc50efe67e1c91f
 fixed] in the spec a while back.

 > When I add UseMicrodescriptors 1 line to my torrc, GETINFO ns/id/$DIGEST
 doesn't return anything since Tor instance refrain from downloading
 network statuses when this configuration option is set.

 ... huh? **UseMicrodescriptors 1** should mean that we download
 microdescriptors and router status entries, while **UseMicrodescriptors
 0** means that we download server descriptors and router status entries.
 Are you saying that we opt to **only** download microdescriptors by
 default nowadays?

 > Damian, would this mess anything up? I wonder if we shouldn't just
 redesign the interface we *want*, and then implement it, rather than
 trying to reverse engineer what we meant when we designed the old thing.

 If mr-4 meant (a) then that sounds like a bug and fixing it will make
 controllers happy (... though I'm not sure why the relays you have
 microdescriptors for wouldn't match the router status entires). If he
 meant (b) then that sounds like a horrible bug and I'm not sure how it's
 possible that went undetected (... Stem has integ tests that should
 definitely complain about that).

 Tor's "GETINFO ns/id/*" options are pretty nice as-is and used by Stem.
 I'm not sure what type of re-implementation you have in mind.

 Cheers! -Damian

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/7646#comment:7>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online