[tor-bugs] #8596 [Tor]: Inconsistent addrmap events when resolving hostname (regression)

[Tor Bug Tracker & Wiki]

#8596: Inconsistent addrmap events when resolving hostname (regression)
-----------------------+----------------------------------------------------
    Reporter:  Desoxy  |       Owner:                           
        Type:  defect  |      Status:  needs_review             
    Priority:  normal  |   Milestone:  Tor: 0.2.4.x-final       
   Component:  Tor     |     Version:  Tor: 0.2.4.11-alpha      
  Resolution:          |    Keywords:  tor-client controller dns
      Parent:          |      Points:                           
Actualpoints:          |  
-----------------------+----------------------------------------------------
Changes (by Desoxy):

  * status:  reopened => needs_review


Comment:

 Thanks for notifying me.

 I got confused by the quotes that were already around it, but those are
 part of the ABNF notation, not of the actual protocol. (I'm not very good
 with ABNF.) I commited a fix at [1].

 > Also, this is more of a stem-specific question but what would be the
 best behavior for when this flag isn't present? Should the event's cached
 attribute be defaulted to True or False? Or should the attribute take on
 the values of True / False / None (for undefined)?

 It's a bit complicated: If the DNS resolution worked and an addrmap event
 is generated without a cached flag, then it was added to the cache. If
 resolution failed (e.g. because the domain doesn't exist) and there is no
 cached flag, then it was not be cached (this only happens when using the
 RESOLVE command).

 1: https://github.com/desoxy-
 tor/torspec/commit/ed7730dc1aa14910955b41c6650c66d70a04e03c

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/8596#comment:14>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online