16:00:59 <Yawning> #startmeeting
16:00:59 <MeetBot> Meeting started Wed Apr  1 16:00:59 2015 UTC.  The chair is Yawning. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:59 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:01:09 <Yawning> who all here is for huggable transports?
16:01:28 <Yawning> I saw blanu
16:01:45 <Yawning> anyone else?
16:02:04 <Yawning> :(
16:02:27 <blanu> At least it should be a quick meeting if it's just the two of us.
16:02:32 <Yawning> yep
16:02:46 * isabela has a couple of notes to pass on to the team
16:03:05 <Yawning> I pushed a bunch of changes to obfs4proxy mostly related to cleanup
16:03:16 <Yawning> so if tor dies, obfs4proxy will try *really hard* to self terminate
16:03:28 <Yawning> even without the stuff I'm proposing as tor changes in that area
16:03:42 <Yawning> I have 2 tor patches pending that will make it more reliable
16:04:02 <Yawning> Uhm, what else
16:04:11 <Yawning> I'm working on my report
16:04:22 <Yawning> HSes caught on fire and I got distracted when I shouldn't have
16:04:40 <blanu> Oh that's good. That might be one of the bugs I was running into. We found that shutting down the tor daemon was not killing obfs4proxy.
16:04:44 <Yawning> I blotched adding socks5 support to goptlib as an overlay in 4.5a5 but I'll fix it
16:04:48 <Yawning> blanu: yeah
16:05:04 <blanu> We also ran into this: https://trac.torproject.org/projects/tor/ticket/14014
16:05:14 <blanu> It is marked as fixed though, so maybe it is fixed.
16:05:17 <Yawning> >.<
16:05:33 <Yawning> if you pull the latest obfs4proxy it should have better behavior if tor goes away
16:05:48 <Yawning> but there's a pt side spec change tor side code change required for it to be reliable
16:06:05 <Yawning> (this affects all pts, so it's not just an obfs4proxy thing, but obfs4proxy master deals with it the best)
16:06:36 <Yawning> if it's still causing issues I can poitn you at a bunch of tor branches
16:06:44 <Yawning> that I expect will be merged that makes things better
16:06:47 <intrigeri> blanu: fyi the AppArmor thing was fixed in tor/0.2.6.3-alpha-1 -- note that it's Debian-specific.
16:07:07 <blanu> Well we were testing with lastest obfs4proxy master as of yesterday, but our tor was from debian and so probably very old.
16:07:13 <Yawning> as far as my pt related todo list goes, write my report, get socks5 support in goptlib
16:07:35 <Yawning> past that there's a large list of pt related little t tor things that need love
16:07:47 <Yawning> next!
16:08:56 <Yawning> oh, also once we have a new goptlib tag I kind of want to tag obfs4proxy 0.0.5 to pull in the cleanup workarounds
16:09:25 <blanu> Well we did the first test of Dust in Tor. We ran into the aforementioned bugs on the server side. We also discovered some weird behavior in TBB when testing with a broken PT, such as TBB showing unhelpful error messages and getting into a state where it would not exit and had to be force quit.
16:09:37 <Yawning> syverson: #tor-project is the HS stuff
16:09:50 <blanu> We also found plenty of Dust bugs, so things are progressing nicely on that front.
16:10:02 <Yawning> integration went smoothly I hope?
16:10:03 <syverson> Yawning: thanks. Doing too many things at once.
16:10:12 <Yawning> syverson: I'm in 2 meetings at once
16:10:48 <blanu> Yeah integration was not a problem, except for the already discussed limitations such as we have no way to use multiple TCP connections for one Tor connection. We have just turned that feature off for now.
16:10:54 <Yawning> ok
16:11:04 <Yawning> if that's really important send me how you want it to work and I'll change things
16:11:07 <Yawning> (or a branch)
16:11:16 <blanu> So we should have a testable obfs4proxy+Dust soon.
16:11:25 <blanu> I think we're going to push that to Dust v3.
16:11:48 <Yawning> ok
16:11:53 <blanu> My contract ends in June. So while the rest of the project goes until August, we need to wrap up the Dust stuff by June.
16:12:06 <Yawning> ahh gotcha
16:12:31 <blanu> Dust v3 is going to be all about ensembles of connections rather than single connections as transports. So that should be interesting!
16:13:03 <Yawning> gonna need a lot of obfs4proxy extensions for that
16:13:07 <Yawning> should be fun
16:14:30 <Yawning> do you have naything obfs4proxy (or otherwise) that requires my imediate attention?
16:15:38 <blanu> There is one thing we have an issue with, which is that obfs4proxy does not pass logging options to the PT. So we are not sure how to handle logging responsibly.
16:15:55 <Yawning> oh hmm
16:16:00 <Yawning> I should change that
16:16:27 <Yawning> it's just spam control right? the standard log package I redirect
16:17:19 <blanu> So mainly loglevels would be nice to know.
16:17:25 <Yawning> yeah
16:17:41 <Yawning> and I guess the ip address redaction bool
16:17:50 <blanu> We are about to add hella logging for debugging.
16:18:09 <blanu> We of course plan to remove this before it goes to production. But you know how that goes.
16:18:11 <Yawning> if I pass those as args to the transport factory's constructors that's sufficient right?
16:18:22 <Yawning> it's something I should have done anyway
16:18:24 <blanu> That seems fine to me.
16:18:42 <Yawning> I think I might move the wrappers I have into a common subpackage as well
16:19:09 <Yawning> and if I do that I wouldn't need to change the ctors
16:19:13 <blanu> Yeah I think the logging you use is not exposed, so we are using some other logging thing.
16:19:21 <Yawning> I'll just export that
16:19:26 <Yawning> and you can assume it's setup?
16:19:31 <blanu> Great! All problems solved.
16:19:40 <Yawning> can you file a ticket for that so I don't forget?
16:20:54 <blanu> I will have my coworker file a ticket as he is more likely to actually do it. I'll send an email right now.
16:21:20 <Yawning> ok, ty I'll probably remember anyway but this HS stuff has me distracted a bit
16:21:22 <blanu> That's all for me. New empirical results expected by next meeting!
16:21:27 <Yawning> :D
16:22:14 <Yawning> Does anyone else have pt stuff?
16:22:31 <Yawning> I think isabela probably wanted to say something but she's also in 2 meetings
16:22:54 <isabela> yes
16:23:01 <Yawning> sup?
16:23:13 <Yawning> I filled some stuff in
16:23:18 <Yawning> on the roadmap I think
16:23:22 <isabela> thanks, just a quick not that we are starting april, so I would like to remind to take a look at the plans for april and update it as you wish
16:23:27 <Yawning> I know I need to fill in more, and write more stuff
16:23:36 <isabela> cool :) thanks Yawning
16:24:18 <isabela> the other stuff is that I am looking for the previous report sent to sponsorT - let me know if you have it
16:24:21 <isabela> thats it!
16:24:29 <Yawning> it's in my sent queue somewhere
16:25:20 <isabela> cool, whenever you have a chance just fwd it to me
16:28:53 <Yawning> ok, sounds like there's not much more hs stuff
16:28:58 <Yawning> err
16:29:01 <Yawning> pt stuff
16:29:02 <Yawning> wrong meeting
16:30:09 <Yawning> #endmeeting *baf*