[tor-bugs] #3428 [Tor Client]: Control port may emit log messages in the middle of another event/reply

[Tor Bug Tracker & Wiki]

#3428: Control port may emit log messages in the middle of another event/reply
------------------------+---------------------------------------------------
 Reporter:  rransom     |          Owner:  rransom         
     Type:  defect      |         Status:  needs_review    
 Priority:  normal      |      Milestone:  Tor: unspecified
Component:  Tor Client  |        Version:                  
 Keywords:              |         Parent:                  
   Points:              |   Actualpoints:                  
------------------------+---------------------------------------------------

Comment(by rransom):

 Replying to [comment:2 nickm]:
 > For 3428b, I feel okay merging it into 0.2.3, but adding a new assert()
 to 0.2.2 would make me nervous.  Does that sound plausible to you,
 rransom?  For less fatal cases than this one, I'd prefer a situation where
 we add LOG_NOCB to the message rather than simply not emitting it.

 I agree.

 > For queue stuff in bug3428, I kinda like the sys/queue.h stuff that the
 BSDs have.  Is there a reason not to snarf that?
 >
 > Actually, perhaps instead of a generic queue like this, we could just
 add an "move the contents of buf_t X onto the end of buf_t Y" interface to
 match the behavior of evbuffer_add_buffer().  Then we wouldn't need to
 have a separate queue: we could use a generic_buffer_t ''as'' the queue.
 Thoughts?

 That sounds better.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/3428#comment:3>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online