17:00:02 <nickm> #startmeeting weekly tor-dev/network team meeting.
17:00:02 <MeetBot> Meeting started Mon Nov 21 17:00:02 2016 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:02 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:00:08 <nickm> heeeeeey everybody.
17:00:10 <Yawning> armadev: huh?
17:00:16 <Yawning> oh
17:00:18 <Yawning> I see
17:00:30 <nickm> usual format: status, then discussion
17:00:49 <Yawning> the main reason why it's different is becaue "I suck at ui programing and gtk3 makes me want to cut myself"
17:01:07 * isabela can go first
17:01:09 <nickm> my status: I've neglected nearly everything else to work on prop#271.  As of this morning, the backend stuff is done[*], and I am working on the integration. I have 6 work days remaining in the month, and 4 non-work days on which I _hope_ not to work.
17:01:11 <dgoulet> hi
17:01:17 <nickm> isabela: one second, sorry
17:01:23 <isabela> nickm: np!
17:01:55 <nickm> [*] backend stuff that isn't done: bridges handling, entrynodes handling, actual persistence.
17:02:26 <nickm> [*] also backend that isn't done: controller support, ed25519 identities, one last expiration corner case
17:02:55 <nickm> I am planning to prioritize getting the current minimal case working correctly.  After that, persistence, then bridges, then entrynodes, then other stuff.
17:03:36 <nickm> currently I'm expecting that I won't retain the old back-end at all. But somebody could talk me out of it.
17:03:50 <nickm> I hope that I'll have it picking guards and building circuits for me before thanksgiving.
17:04:05 <nickm> (next?)
17:04:15 * armadev points at isabela
17:04:40 <isabela> i am in continuous catch up mode -- kind of overwhelmed with a load of training requests and other post-election stuff
17:04:43 <isabela> but!
17:05:05 <isabela> i do plan on working at our contracts this month, organizing the end of it and final reports and plan on
17:05:17 <isabela> work at 'organizing the house for 2017'
17:06:14 <isabela> probably a great time to look at open discussions on releases and other stuff and move them into an actual thing we all can start doing in 2017
17:06:17 <isabela> done
17:07:03 <armadev> i guess i'll go next?
17:07:08 <dgoulet> go
17:07:12 <armadev> i did some 0.2.9 stuff, opening new tickets and so on
17:07:17 <armadev> two interesting things for the network team:
17:07:23 <armadev> i opened https://trac.torproject.org/projects/tor/ticket/20725
17:07:35 <armadev> which will involve more coordination between the tor team and the tor browser team
17:07:43 <armadev> but that's the #1 question / concern / etc on the blog comments
17:08:12 <armadev> and, speaking of the blog comments, we have a pattern where people do a blog post and then maybe they handle comments for a day or two but then they get bored and leave the post ignored
17:08:16 <armadev> and it accumulates months of spam
17:08:25 <armadev> so, we need a new habit of "when you're done handling comments, turn them off"
17:08:25 <nickm> I handle no comments on my blog posts.
17:08:34 <armadev> if you're done before it starts, then turn them off then
17:08:51 <armadev> but the network team should ponder whether it can do better than that, because blog comments are a source of bug reports
17:08:57 <armadev> but maybe that is more the case for tor browser than for tor
17:09:02 <Yawning> armadev: when the blog system is renewed I will start paying attention to it again
17:09:10 <armadev> but maybe that's because the tor browser team cultivated receiving useful blog comments
17:09:24 <armadev> anyway, my item #2 is a suggestion that maybe we want to have one of the weekly rotation spots be blog comment duty
17:09:25 <armadev> done
17:09:53 <Yawning> This is my sandbox.  There are many others like it but this one is mine. My sandbox is my best friend.   It is my life.
17:09:58 * dgoulet can go
17:09:58 <Yawning> done
17:10:07 <nickm> Yawning: what does it do that it didn't do last week?
17:10:17 <Yawning> works on i386
17:10:29 <nickm> that's cool.
17:10:31 <Yawning> there's a gigantic list of commits
17:10:56 <armadev> i am the alpha user. or the beta user. or whatever the #1 out of 1 users is called. :)
17:11:09 <Yawning> has support for the circuit display
17:11:15 <armadev> "i am the user" i guess is how that's said.
17:11:29 <Yawning> it can launch tor on it's own without relying on the user to have one running
17:11:42 <Yawning> so basically, it's mostly a minimum viable product
17:11:46 <nickm> neat
17:12:25 <armadev> dgoulet: looks like you're up?
17:12:28 <dgoulet> ok
17:12:35 <dgoulet> Hello! Some ticket review that are mostly in merge_ready now. Most work on #20029, I failed to finish the small part that is left that is more thorough unit tests but should be done soon!
17:12:37 <dgoulet> Also went through the client authorization tor-dev@ thread of asn and did a day of bad relays work as it was falling behind...
17:12:40 <dgoulet> I'm now here with chelseakomlo working on planning the prop224 service side implementation which is a _huge_ piece of the puzzle! And also, I'll be on triage duty this week.
17:12:56 <dgoulet> that is #20657 ^
17:12:59 <isis> hello
17:12:59 <dgoulet> Finally note that I'll be at McGill university tomorrow morning to do a Tor talk to a bunch of anthropology students.
17:13:00 <dgoulet> --
17:13:55 <nickm> cool
17:14:32 <nickm> next update?
17:15:01 <nickm> isis / asn / teor-should-be-asleep / chelseakomlo / anybody else?
17:15:11 <chelseakomlo> i have a short update
17:15:43 <nickm> go for it!
17:15:50 <isis> i have been continuing to do OTF stuff, but slowed down a bit because i have the worst flu ever and my fever is too high to think straight
17:16:01 <nickm> isis: ugh; so sorry to hear that.
17:16:09 <nickm> remember to keep your brain cool
17:16:16 <nickm> and hydrate.
17:16:23 <nickm> and take ibuprofen if you can
17:16:39 <isis> thanks, and try not to get the flu this year! it's really bad
17:16:40 <armadev> and have a person with a working brain nearby to check with about common sense things :)
17:17:30 <chelseakomlo> along with pairing with dgoulet (which is great) i also have a patch for #20717 that i'll soon be submitting
17:19:00 <chelseakomlo> and i plan to review #20672 for anything else we can refactor/make better
17:20:02 <attractr> armadev: i am glad to see you doing this and happy to help
17:20:47 <chelseakomlo> that's all for me :)
17:20:56 <asn> hello people
17:21:09 <asn> need a few seconds to write my update
17:21:25 <attractr> hearing about yawning, etc in here made me motivated to re-analyze TBL for a new whitelist
17:21:38 <attractr> been on my todo list for a while anyways
17:22:00 <armadev> TBL might be..tor browser launcher?
17:22:11 <attractr> yes
17:22:39 <attractr> that's the seccomp bpf whitelist that yawning borrowed for his bubblewrap work
17:23:26 <armadev> bpf might be..berkeley packet filter? :)
17:23:56 <attractr> also correct, but the vm has been repurposed elsewhere in the linux kernel
17:23:56 <armadev> asn: you're up whenever you're ready. we're just stalling for you. :)
17:24:15 <nickm> chelseakomlo: btw, I owe you some quick notes on the diagrams you sent me last week. Sorry I've been slow there.
17:24:31 <attractr> a bpf vm is also used to run compiled system call filters, this is seccomp-bpf
17:24:49 <nickm> chelseakomlo I'm probably gointg to procede on the theory that a short response today is better than a long response in mid-december :/
17:24:52 <attractr> we used it in sgos to significantly reduce kernel attack surface (by 2/3 or more)
17:25:05 <chelseakomlo> nickm: no worries at all. i can also have dgoulet or someone else on the team review them this week for a first pass, so no rush
17:25:30 <asn> Hello, the thid week of November I worked on the following things:
17:25:30 <asn> - I worked more on client auth. I think we are getting closer to solidying that part of rend-spec-ng.txt. Here is the torspec patch: https://lists.torproject.org/pipermail/tor-dev/2016-November/011649.html And here is some further analysis on descriptor sizes etc. https://lists.torproject.org/pipermail/tor-dev/2016-November/011658.html
17:25:35 <asn> There is still a bit more work to be done here, based on feedback I have received from the beginning of the thread, but I think the current torspec patch contains the most hardcore changes.
17:25:39 <asn> - I also revised my prop224 ESTABLISH_INTRO branch based on review from dgoulet and chelsea.
17:25:42 <asn> - I reviewed #20638 a few minutes ago, and some more tickets.
17:25:44 <asn> - I'm attending CCC after all.
17:25:47 <asn> sorry about the delay
17:26:21 <armadev> nickm: speaking of the team, are the details of our plan to expand the team public? should we invite our future developer to these meetings?
17:26:27 <armadev> asn: so you are all set with a ccc ticket?
17:26:40 <asn> armadev: i have a lead on one, but i dont actually have a pdf yet.
17:27:04 <nickm> asn: no worries; any help you need?
17:27:25 <asn> nope im good! i will be working on prop224 and client auth this week as well.
17:27:32 <asn> i can also find time to work on guard stuff as they come out of the oven.
17:27:36 <dgoulet> asn: are you going with the coding?
17:27:48 <nickm> armadev: we're putting the job ad out today-ish, I believe. Better not to announce before there's a description of how to apply.
17:28:08 <nickm> for the existing new-hire, I don't have final word from ewyatt that it's actually final.  But of course he's invited.
17:28:44 <armadev> nickm: is he invited in an "if he knows" sort of way, or did somebody actually tell him? :)
17:28:52 <dgoulet> armadev: invited where exactly?
17:29:03 <Yawning> urgh
17:29:07 <armadev> dgoulet: to the 17:00 monday irc meetings
17:29:09 <asn> dgoulet: next coding tasks for me would be to revise my establish_intro patch based on more review, fix the prop224 HSDir path based on the client auth changes, and do any AES-128 -> AES-256 changes requiured.
17:29:25 <nickm> someone ought to remind him.  Probably once the hire is finalized. I know he doesn't want to start till feb, so maybe he'll not want to stop by the meetings either. who can say.
17:29:29 <isabela> i guess we are waiting on a final word from ewyatt to start onboarding him
17:29:31 <dgoulet> armadev: ah I told him and we'll start a pre-hire onboarding with him, I planned to email network-team@ today about it
17:29:44 <Yawning> Oh yeah
17:29:48 <dgoulet> ewyatt was on the thread about this ^
17:29:48 <isabela> yeah
17:29:59 <Yawning> this prolly belongs in a ticket, but is there any reason why I can't query individual circuit status
17:30:03 <Yawning> via get info
17:30:16 <Yawning> circuit-status gives me all the things
17:30:19 <isabela> there is the team list, routations, these meetings etc
17:30:29 <isabela> but i think we should also respect his start date :)
17:30:47 <isabela> not force him to be doing a bunch of stuff
17:30:49 <dgoulet> yeah we figured out something to "pre-onboard" instead of throwing everything at thim until he starts :)
17:30:49 <armadev> isabela: respect the start date while giving him the option of doing fun things beforehand! :)
17:30:51 <nickm> Yawning: like, for origin circuits?
17:30:57 <isabela> armadev: yes
17:31:04 <nickm> Yawning: no inherant reason for that IMO.
17:31:35 <Yawning> I kind of wanted something like circuit-status/<id> or whatever
17:31:37 <armadev> yawning: i think if you wanted a getinfo circ-status/4442 or whatever syntax we choose, sounds good to me
17:32:03 <armadev> i guess the alternative is that you getinfo all the circ-status, and pick out the one you wanted? doesn't sound so bad either, since localhost is cheap and stuff
17:32:18 <Yawning> that's what the code currently does
17:32:22 <Yawning> both tor browser and me
17:33:03 <nickm> dgoulet: btw, there's an open question for you on #20629
17:33:05 <armadev> any more statuses from people or did we slide into discussion?
17:33:23 <dgoulet> nickm: ack
17:34:04 <nickm> my review plan for this week is to look at things that other people put into merged_ready if it isn't too big.
17:34:10 <nickm> And that's about it.
17:34:23 <armadev> i think nothing burned down last week
17:34:24 <nickm> If we're done with updates, any discussion topics for this week?
17:34:35 * dgoulet is good
17:34:42 <nickm> I see that dgoulet asked some cool stuff last week on the ml, and then we didn't follow up.
17:34:48 <armadev> there's my blog comment rathole
17:34:52 <isabela> yes
17:35:23 <dgoulet> nickm: well don't worry for now, once you are done with November work we can circle around to that
17:35:25 <isabela> so i have a proposal on all that
17:35:30 <dgoulet> oh
17:35:34 <nickm> armadev: open question for you on #20269
17:35:49 <isabela> which is in the same line as what dgoulet is saying
17:36:10 <isabela> i think we can work during december on these discussions and oragnize who the team will function  in 2017
17:36:22 <isabela> add blog comments routation, the releases stuff, etc
17:36:47 <isabela> just because november is very short and there is so much to do
17:36:58 <isabela> this would unfocus us a little bit
17:37:02 <nickm> what would people think about moving to a system where our rotations actually, well, rotate?
17:37:27 <nickm> like, having the default be "person1, person2, person3, ..., then loop" , and having people trade off if they can't do the week they got?
17:37:33 <nickm> or should this discussion wait a bit
17:37:59 <asn> i feel like we dont have enough people to make a satisfactory rotatory loop
17:38:08 <asn> it's me, david, and nickm. with a best-effort isis.
17:38:38 <asn> and the whole point of this rotation is to have nickm do less, so putting him out of the rotation loop is also fine.
17:38:43 <isabela> true - but i like nickm suggestion.. although i believe we should make a plan for 2017 that address that and other discussions
17:39:01 <asn> (personally im fine with making it rotate)
17:39:02 <nickm> In 2017 we should add 2-3 more people...
17:39:04 <isabela> counting alex and new person
17:39:19 <isabela> yes
17:39:23 <nickm> and maybe we can add in some volunteers, though they might want to say "I'm in 1/2 of the rotations"
17:39:23 <asn> right
17:40:00 <asn> yeah wth volunteers it will look more like a markov chain instead of a cyclic rotation
17:40:12 <asn> *with
17:40:41 <asn> (fwiw im next in line for the triaging rotation, after dgoulet is done this week)
17:40:45 <armadev> especially for things like trac and blog triage, a volunteer can do a lot of it, and raise questions in the irc channel or whatever as needed
17:41:07 <armadev> (that way the volunteer gets smarter and better at tor stuff too)
17:41:17 <nickm> asn: thank you!
17:41:20 <nickm> dgoulet: (and thanks to you too)
17:41:29 <dgoulet> rotating seems a good idea to me but we should discuss after Nov. what are the "tasks" apart from bug triage we want :)
17:41:36 <dgoulet> I think I had a bullet point about it in my email?
17:41:39 <isabela> so yeah i was thinking we could work in december on how the team will operate on that, but also how we want our release calendar to be..
17:41:49 <isabela> and talking with hiro i shared with her that pad of our wishes for trac
17:41:59 <isis> fwiw, i'm also okay with it rotating
17:42:00 <isabela> maybe we could review some of that too
17:42:15 <hiro> +1 ;)
17:42:15 <isabela> and organize things for 2017
17:42:26 <Yawning> I have no idea what my 2017 looks like
17:42:28 <Yawning> :/
17:43:07 <Yawning> it depends on what the voices in my head, the chemtrails, and the reptiloid's mind control waves they're broadcasting from my smart meter want me to do
17:43:23 <isabela> !
17:43:33 <nickm> chelseakomlo: btw I just commentd on your #18873 branch, and it's all minor stuff I can fix as part of merging. Shall I, or would you rather?
17:44:39 <nickm> any more discussion things for today?
17:44:43 <chelseakomlo> nickm: oh great! it's up to you, whatever is less work for you. i can fix it up today if that is easier
17:44:56 <nickm> thanks, that would be helpful!
17:45:15 <nickm> (please remember to use separate commits so that I only need to review what changed.)
17:45:25 <nickm> (yes I know I say that over and over :) )
17:45:52 <chelseakomlo> nickm: ok, that's good to know
17:45:54 <chelseakomlo> will do
17:45:56 <nickm> thanks!
17:46:08 <nickm> If no discussion issues, then let's call the meeting adjourned? :)
17:46:13 <dgoulet> o/
17:46:39 <nickm> Is "o/" waving goodbye or raising a hand?:)
17:46:58 <asn> thanks people :)
17:46:59 <dgoulet> ah damn, no it's "it's all good, hello everyone" :)
17:47:05 * isabela is good o/
17:47:07 <asn> chelseakomlo: dgoulet: let me know if you want me to check your design docs!!
17:47:08 <nickm> dgoulet: cool
17:47:17 <nickm> ttyl everybody then!
17:47:18 <dgoulet> asn: we'll make it cleaner before but yes :D
17:47:20 <asn> nickm: please let me know if you want some initial review on guard stuff this week
17:47:23 <asn> sounds good!
17:47:26 <asn> byyee
17:47:36 <Yawning> o/ . o O (Ave Imperator morituri te salutant!)
17:47:40 <nickm> And as I noted last week -- just because I'm super-busy doesn't mean I you shouldn't ask me for anything I can help on.
17:47:45 <nickm> peace all.
17:47:48 <nickm> thanks, asn!
17:47:49 <nickm> #endmeeting