13:29:28 <nickm> #startmeeting
13:29:28 <MeetBot> Meeting started Wed Aug 26 13:29:28 2015 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:29:28 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:29:31 <nickm> hi everybody!
13:29:36 <nickm> let's start with brief status stuff!
13:30:07 <nickm> I've been trying to review designs, merge patches, fix bugs, and not lose my mind!  I disappear tomorrow through monday.  Freeze still planned for 1 Sep.
13:30:30 <nickm> I think this is going to be a pretty solid release. I wish we could squeeze more in, but I always wish that.
13:30:34 <nickm> Who's next?
13:30:40 <Yawning> I can go
13:30:40 <nickm> hi BenL
13:30:45 <nickm> go for it
13:30:46 <Yawning> Rewriting the PT spec.
13:31:04 <Yawning> I'll prolly try to get pt logging related stuff merged, but that means cleaning up the branches by monday.
13:31:13 <Yawning> Doable, we'll see
13:31:37 <Yawning> also starting to think about a v2 pt spec, but not sure about how much I want to change
13:31:42 <Yawning> another we'll see
13:31:43 <nickm> ok. don't knock yourself out unless you think it's truly worth it. :)
13:31:52 <Yawning> it is, should be straight forward
13:32:04 <Yawning> next >.>
13:32:14 <Yawning> also I need to do that circuit lifepsan stuff
13:32:35 <nickm> that TBB patch divergence once?
13:32:43 <BenL> nickm: hello there!
13:32:44 <Yawning> yah
13:32:49 <Yawning> I'll tag it as post freeze
13:33:07 <Yawning> it'll most ly be a mechanical merge I think, since the code is tested/shipping
13:33:20 <nickm> BenL: (it's tor dev meeting time now. A bot is logging, so you know)
13:33:29 <nickm> Yawning: ok
13:33:38 <nickm> any more? if not, who's next?
13:33:55 <Yawning> someone else can go
13:34:51 <nickm> ok, who has a thing to say? :)
13:35:26 <nickm> or a status report?
13:35:39 <Sebastian> I'm here for the community team
13:35:49 <nickm> hi Sebastian !
13:36:04 <Sebastian> is there anything that we should be aware of (new important features for people to be aware of during training etc)
13:36:21 <Sebastian> or any news wrt ed identity keys?
13:36:37 <Yawning> ed keys will happen in 0.2.7.x?
13:36:51 <isabela> yeah
13:37:07 <Yawning> Don't think there's much training stuff re 0.2.7.x
13:37:10 <Sebastian> If nothing immediately comes to mind I'm always around in case you need to get the word out about anything
13:37:24 <nickm> oh, here's a thing.  Jens made us a temporary Phabricator instance for testing.
13:37:33 <Yawning> what's that
13:37:39 <Sebastian> code review tool
13:37:42 <nickm> https://phabricator.wikimedia.org/
13:37:45 <Yawning> ah neat
13:37:47 <nickm> ours is at https://phablab.krautspace.de/project/profile/15/
13:37:52 <nickm> let's play with it and see if we like it
13:37:59 <isabela> oh
13:38:05 <Yawning> so like gerrit?
13:38:06 <nickm> it's a temporary testing thing IIUC
13:38:10 <nickm> I htink so?
13:38:11 <isabela> can anyone just create an account on it?
13:38:15 <nickm> I think so
13:38:20 <nickm> If not ask qbi
13:38:50 <nickm> also, if you remember what cypherpunks do , then 'cypherpunks' might be a login for you.
13:38:51 <isabela> it works
13:38:52 <Yawning> before I commit to uising this
13:38:54 <nickm> great
13:39:01 <nickm> Yawning: yes?
13:39:04 <Yawning> is there a migration plan from testing to production if we like it
13:39:09 <Yawning> or am I gonna lose work
13:39:12 <nickm> Yawning: not yet but ask qbi
13:39:25 <nickm> Right now I'd suggest we not commit hard, but try it for a patch or two
13:39:36 <nickm> if we like our initial experience let's ask qbi about migration
13:39:37 <Yawning> (I'm all for a code reeview tool, but yeah)
13:39:42 <isabela> nickm: would we move all trac tickets to it?
13:39:42 <nickm> sound ok?
13:39:51 <Yawning> nickm: indeed
13:40:07 <nickm> isabela: Right now I'm only seeing it suggested for review; I do't know if we want to move all tickets to it; that wouldn't be just our decision
13:40:21 <nickm> trac sure is showing its age, but for now let's just try it out.
13:40:22 <isabela> ok
13:40:37 <s7r> Sebastian: ed25519 identities are 80% covered already in latest master branch. we just need to do #16769 and #16790. but I am happy with how nickm's coded Tor to recover from many many many possible cases.
13:40:53 <s7r> I am aalready working on the instructions page, which we will include in the manual somehow, or we'll see
13:41:28 <nickm> Sebastian: right now, ed25519 identities are not fully supported
13:41:42 <nickm> relays have them and manage them correctly.  authorities know about them and vote on them correctly.  they are used to sign descriptors...
13:41:43 <Sebastian> yeah, so the thing about ed that I'm worried about
13:41:48 <rjunior> I'd like to share something about what my team has been working on
13:41:54 <Sebastian> people might upgrade and get their old identity key tied to an online key
13:41:55 <nickm> but they do not have the link handshake or circuit handshake thing.
13:42:00 <Sebastian> without the chance to generate an offline key
13:42:05 <Sebastian> I think this already happened for moria
13:42:20 <Sebastian> (will the identity keys for dirauths live in Tor's source eventually?)
13:42:21 <nickm> rjunior: great!  please introduce yourself
13:42:27 <s7r> this does happen, if Tor is started with no ed25519 keys in datadirectory
13:42:31 <nickm> Sebastian: dunno
13:42:50 <Yawning> eventually yes since we will deprecate RSA no?
13:42:53 <Yawning> (very long term)
13:43:00 <nickm> Sebastian: It'd be nice to think abotu a way around that, but I don't have a great one.  If you have an idea maybe open a ticket? :)
13:43:04 <rjunior> Hi, I'm Reinaldo. We (a team of 5 people) are working on improving testing coverage for Tor
13:43:13 <nickm> Sebastian: I have a tiny bit of hacking time between now and the end of the month
13:43:28 <Sebastian> nickm: ok, thanks. /me lets you continue with the meeting. We can coordinate later!
13:43:31 <rjunior> We have been using a github fork (https://github.com/twstrike/tor) and we are planning to send patches soon.
13:43:36 <nickm> Yawning: we're deprecating RSA1024.  Bigger RSA can stick around indefinitely
13:44:08 <Sebastian> rjunior: exciting!
13:44:23 <rjunior> Our team have been working on contributing to PET-related software, and this is part of an initiative sponsored by ThoughtWorks.
13:44:28 <Yawning> nickm: hmm
13:44:55 <nickm> But to be clear, we can't call ed25519 done until link and circuit extension use it.
13:45:05 <rjunior> Nick has already helped us pointing to some ticket we could address, but feel free to reach out to us if you have any recommendation.
13:45:14 <Yawning> nickm: yar
13:45:29 <nickm> rjunior: one thing that I'd suggest ASAP is not to do further patches in one big branch.
13:45:48 <rjunior> Our ultimate goal is to get to the point we are comfortable to contribute with some other proposals (Hidden Services is our current target)
13:46:13 <nickm> rjunior: right now, there seem to be tons of commits in your master branch.  If some are great and some need to be revised, it is hard to take just the ones that are ready. :)
13:46:26 <teor1> smaller branches are easier to review, too :-)
13:46:30 <nickm> yes
13:46:33 <teor1> we can split them among multiple people
13:46:48 <nickm> rjunior: also have you found any bugs yet? :)
13:47:16 <rjunior> Sure. I believe we will stop at some point this/next week and start sending patches for individual files.
13:47:28 <nickm> ok cool
13:47:36 <nickm> if you're going to split stuff up yourself, that's excellent
13:48:43 <rjunior> We have found unreachable branches from the tests, but not really any bug. We are using this exercise mostly to understand how some parts behave. I dont believe we have reached that point to know if something is a bug.
13:48:52 <nickm> ok
13:49:35 <Yawning> (yay more coverage)
13:49:53 <rjunior> If splitting up by ourselves makes it easier for you to review, we can do that (since we know what we have changed).
13:50:15 <nickm> that would be excellent
13:50:25 <nickm> small branches or patch series are both easy to review.
13:50:28 <nickm> big branches aren't.
13:50:43 <rjunior> nice
13:50:54 <nickm> cool, and btw: welcome!
13:51:01 <nickm> anybody else with a status report?
13:51:13 <Yawning> (lol agora is shutting down)
13:51:26 <teor1> I have been laptop-less, just got it back a few hours ago.
13:51:37 <teor1> In the meantime, I wrote up a plan for improving chutney
13:52:12 <nickm> nifty.  is that the one you showed me?
13:52:17 <teor1> Yes, https://trac.torproject.org/projects/tor/wiki/doc/TorChutneyGuide
13:52:26 <nickm> great
13:52:35 <nickm> I am eager to help there however I can
13:52:39 <teor1> So it's a combination of "how to use chutney" and "how I'd like to make chutney easier to use"
13:53:21 <teor1> work on fallback directories was on hold, but I did clarify the next step - a small-scale trial with a few fallback who opt-in in 028
13:54:23 <nickm> sounds good
13:54:37 <teor1> oh, and now I have my laptop back I can see if profiling tor in a chutney test network produces useful results now that chutney sends data
13:54:40 <nickm> oh also, i forgot to say but: my callgraph tools are now documented in doc/HACKING
13:54:43 <teor1> I think it will emphasise crypto, because the number of nodes is small
13:55:04 <teor1> and OS X has slow system calls, so they will be prominent
13:55:13 <teor1> done
13:55:16 <nickm> and if the number of clients is small wrt the number of nodes, it will underrepresent setup costs
13:55:30 <nickm> thanks teor1 ! Anyone else, or should we move on to Discussing Things?
13:56:40 <Yawning> teor1: wait, they don't vDSO?
13:56:59 <Yawning> fleatbsd and finux both do >.>
14:01:34 <nickm> sounds like we move on!
14:01:53 <nickm> big topics: the freeze and beyond
14:01:57 <isabela> !
14:03:05 <nickm> we could reasonably triage stuff out of 0.2.7 or TorCoreTeam201508 now IMO.  What do you think?
14:04:40 <nickm> We could also tag really important things with PostFreeze027 (if that's the tag)
14:04:41 <isabela> also
14:04:41 <isabela> this is the last week of august so things that are moving to next week should be tagged for september TorCoreTeam201509
14:05:20 <nickm> agreed.
14:05:40 <isabela> lets triage!
14:05:44 <nickm> Should we try to get on that post-meeting?
14:05:51 <nickm> it could take a bit of a while.  any other topics first?
14:05:56 <nickm> (I don't have any)
14:06:13 <isabela> just an announcement
14:06:35 <nickm> o?
14:06:46 <isabela> we will be sending an email to organize a network meeting!
14:06:47 <isabela> :)
14:06:50 <isabela> soon
14:06:52 <isabela> so watch for that
14:07:02 <nickm> yaaay
14:07:12 <nickm> isabela: let's aim for this week on that? :)
14:07:17 <isabela> sim sim
14:07:26 <nickm> ok.
14:07:28 <nickm> #endmeeting