[tor-bugs] #24367 [Core Tor/Tor]: Changing pluggable transports (during start-up) in Tor Browser is broken

[Tor Bug Tracker & Wiki]

#24367: Changing pluggable transports (during start-up) in Tor Browser is broken
-------------------------------------------+-------------------------------
 Reporter:  gk                             |          Owner:  nickm
     Type:  defect                         |         Status:
                                           |  needs_revision
 Priority:  Medium                         |      Milestone:  Tor:
                                           |  0.3.2.x-final
Component:  Core Tor/Tor                   |        Version:  Tor:
                                           |  0.3.2.1-alpha
 Severity:  Normal                         |     Resolution:
 Keywords:  regression, tor-bridge-client  |  Actual Points:
Parent ID:                                 |         Points:  0.5
 Reviewer:                                 |        Sponsor:
-------------------------------------------+-------------------------------
Changes (by gk):

 * status:  needs_information => needs_revision


Comment:

 Replying to [comment:10 teor]:
 > Replying to [comment:9 gk]:
 > > Replying to [comment:8 teor]:
 > > >
 > > > > The first bad commit is 93a8ed3b83b5f20768562ca2aff4eba7aca667d8.
 > > >
 > > > I'm not sure this is correct.
 > > >
 > > > Bridges didn't work at all between c21cfd28f4 (#17750) and
 93a8ed3b83 (#23347).
 > >
 > > You are mistaken here. The immediate parent of
 93a8ed3b83b5f20768562ca2aff4eba7aca667d8, which is
 6370fb77c586e9ad68c7d1ecb95be36137cb9067, is working perfectly fine: I can
 change the PT during bootstrap and even after restart the new PT is being
 used without any complaint. Exactly like the current behavior in 0.3.1.8
 is.
 >
 > Oh, you're right, #23347 has a misleading title. Fixed!
 >
 > ...
 >
 > But I still don't have enough information to diagnose the bug in the
 logs you supplied for tor commit d7833c9d27feed9e. They tell me that we
 fixed one bug, but they don't tell me where the next error is.
 >
 > Perhaps we can work out how to reproduce this bug using a command-line
 like the one in #23347?

 I have to think about that a bit harder. I am not sure how to do that
 right now.

 > Otherwise, I need someone to provide more detailed logs (info-level?)
 that tell me *why* Tor Launcher has set DisableNetwork.

 Alright. Let's look at the first issue described in steps 1)-4) in the bug
 description. It might give us some clue about what the actual bug(s)
 is/are which are remaining. I've attached two info level logs, one with
 6370fb77c586e9a (where everything is fine and `obfs4` is finally used),
 and one with d7833c9d27feed9 where still a connecting with `obfs3` is
 attempted. I hope that's already enough for you to explain the problem in
 that part at least.

 > About 10 hours ago in #tor-dev IRC, catalyst was saying that Tor
 Launcher will set DisableNetwork in response to some bootstrap events. (I
 can't find a bug number for that bug.) Maybe we are seeing an instance of
 that bug here?

 Hm. Not sure, although I am still not convinced that this plays a role
 here. Maybe mcs/brade have more ideas for that part, though.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/24367#comment:11>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online