[tor-bugs] #7515 [Tor]: Better definition for ADDRMAP events

[Tor Bug Tracker & Wiki]

#7515: Better definition for ADDRMAP events
--------------------+-------------------------------------------------------
 Reporter:  atagar  |          Owner:              
     Type:  defect  |         Status:  needs_review
 Priority:  normal  |      Milestone:              
Component:  Tor     |        Version:              
 Keywords:          |         Parent:              
   Points:          |   Actualpoints:              
--------------------+-------------------------------------------------------

Comment(by rransom):

 Replying to [ticket:7515 atagar]:
 > The ADDRMAP specification is making me twitch for a few reasons...
 >
 > * It doesn't say what the events do. My guess from seeing a few is that
 it simply indicates DNS resolution. However, the spec doesn't make this
 clear and it allows hostnames to be resolved to other hostnames (I'm
 guessing that is a mistake).

 The TrackHostExits option should cause Tor to emit hostname-to-hostname
 ADDRMAP events.

 > * The errors are undefined. An 'XXXX' is pretty bad - I hope we can at
 least say what the error code can include.

 Unless you really need a list of all possible errors, it's better to
 underspecify than to write a spec that forbids useful future extensions or
 that is incompatible with Tor's actual behaviour.

 > * The spec summary has some mistakes. For instance, it doen't list
 GMTExpiry as being optional (despite saying so below), and it should make
 it clear that 'error' and 'GMTExpiry' are key/value entries. My first
 thought when I saw this event was "WTF? They have an optional positional
 argument?!? Gah!"

 Duplicate of #6829.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/7515#comment:2>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online