[tor-bugs] #8701 [Tor]: Stem doesn't recognize error 555 on attach_stream

[Tor Bug Tracker & Wiki]

#8701: Stem doesn't recognize error 555 on attach_stream
----------------------------+-----------------------------------------------
    Reporter:  cypherpunks  |       Owner:  atagar             
        Type:  defect       |      Status:  reopened           
    Priority:  normal       |   Milestone:  Tor: 0.2.4.x-final 
   Component:  Tor          |     Version:                     
  Resolution:               |    Keywords:  tor-spec tor-client
      Parent:               |      Points:                     
Actualpoints:               |  
----------------------------+-----------------------------------------------
Changes (by nickm):

  * keywords:  => tor-spec tor-client


Comment:

 Replying to [comment:9 arma]:
 > I think there remains a bug here. See my "That said, there may be a bug
 here" paragraph. Why do we prevent the controller from attachstreaming a
 circuit that's currently in state AP_CONN_STATE_CIRCUIT_WAIT ?

 Because there's probably a race condition in your program if you expect
 that to work.

 This can happen if you are not setting LeaveStreamsUnattached, and you are
 trying to call attachstream on user streams.  But that case indicates a
 questionable controller program, since there's no way for you to make sure
 that your choice of circuit will be honored. Instead, the circuit that the
 stream is waiting for might complete, and the stream might attach to that
 one instead of yours.

 It can also happen if you ''are'' setting LeaveStreamsUnattached, and you
 are trying to call attachstream on Tor-internal streams, like directory
 requests.  But that's even more unsupported.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/8701#comment:10>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online