18:01:24 <nickm> #startmeeting weekly net team meeting, 22 Jan 2018
18:01:24 <MeetBot> Meeting started Mon Jan 22 18:01:24 2018 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:01:24 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
18:01:27 <nickm> FREEZE!
18:01:31 <nickm> the features, that is
18:01:38 <nickm> the pad is: https://pad.riseup.net/p/rEhOs7BsDez6
18:01:45 <nickm> how are we doing?
18:01:51 <asn> hello people
18:01:52 <isabela> buenas
18:01:53 <dgoulet> hi
18:01:58 <ahf_> hi
18:02:13 * asn writes status report
18:05:04 <nickm> So, 0.3.2.x came out stable later than we had in mind.  Let's make sure that 0.3.3.x stabilizes a little more aggresively
18:05:18 <nickm> let's fix bugs, especially regressions, in 0.3.3.x
18:05:40 <dgoulet> what are the big features that went in 033 again?
18:05:41 <nickm> I'm aiming to get 0.3.3.1-alpha out by Wednesday if I can, and another alpha every 2 weeks or so till it's stable
18:05:42 <dgoulet> (if any)
18:06:19 * asn reads reports
18:06:45 <nickm> vanguard experiment, more ipv6 working, better embedding support, actual rust code
18:07:03 <dgoulet> ok thanks
18:07:47 <pastly> I'm here.
18:07:52 <nickm> hi there!
18:08:18 <pastly> my thoughts aren't gathering very well. Trying to do that quickly.
18:08:31 <ahf_> isabela: do you want to include january work in the s8 status update? or just up to and including december?
18:08:44 <isabela> just up to december
18:08:47 <ahf_> if yes then i need to update your doc
18:08:47 <ahf_> cool
18:09:35 <isabela> my doc? you mean which
18:09:36 <isabela> hehe
18:09:41 <isabela> the pad?
18:09:56 <ahf_> yes, the pad we made in the end of december about s8 tasks
18:09:59 <isabela> yes
18:10:04 <isabela> thanks
18:10:37 <isabela> i plan on looking there first
18:10:47 <isabela> so thanks for keeping it updated
18:11:37 <ahf_> i haven't updated it since december and at that point it was up-to-date - haven't updated it afterwards :-)
18:11:44 <isabela> lol
18:12:38 <nickm> ahf_: the memory stuff looks cool -- looking forward to it!
18:12:52 <nickm> ahf_: did we wind up getting usable wakeup results from the earlier experiments?
18:12:56 <asn> dgoulet: will get u logs for the hs desc issue :)
18:13:04 <dgoulet> asn: love it, feed me!
18:14:01 <ahf_> nickm: yes!  it's interesting as well because the tracing extensions to dgoulet's tracing api makes *a lot* of stuff easier for us to reason about and meassure where i before made small hacks + log() statements i can now do it without modifying tor! :-)
18:14:19 <nickm> cool
18:14:49 <ahf_> nickm: re: wakeups, i'm running a relay and my android client with it to collect the information, but i haven't dived further into *where* we need to start cutting on things that are returning too often or things we might be able to group together
18:15:23 <nickm> isabela, isis: you should plan a meeting with one another today if possible, to check in with moat status/infodump.  Possible?
18:15:32 <nickm> ahf_: makes sense
18:16:26 <isis> nickm: isabela: that would be great, i could do anytime after the TLS meeting
18:16:51 <isis> nickm: isabela: i think it would be handy if we could pull in sysrqb to the meeting also?
18:17:01 <ahf_> interesting with TLS 1.3(!)
18:17:07 <asn> isabela: will u be coming to rome btw? :)
18:18:28 <isabela> yes i will
18:18:34 <ahf_> \o/
18:18:41 <asn> happy to hear that!
18:18:49 <nickm> isabela: Wooooo!
18:18:55 <isabela> isis: aren't you going to be afk for 2 weeks?
18:19:21 <isis> ahf_: it's at 20:00 CST (i think? in 1.5 hours) you should join!
18:19:36 <ahf_> pastly: if there is anything i can help with on shadow + tor master i will try to find time for it, we have some tasks around the sponsor 8 where we need to do some measurements with shadow
18:19:59 <isabela> isis: i am a bit confuse here with communication and as i wrote on my update was going to ask for extension on the contract for this task
18:19:59 <ahf_> isis: i'll try to be around and see what is going on - it just takes place in here?
18:20:08 <isis> isabela: it has been rescheduled to ~2 weeks in february
18:20:22 <sysrqb> isis: sure :)
18:20:23 <isabela> isis: because i didn't thought you would be available to work on this, giving also lack of response since before holidays
18:21:00 <pastly> ahf_: the only thing I've noticed that prevents "master" (actually, tor-0.3.2.9) from working is authorities choking when the hs_get_time_period_num is about to be calculated to be zero because it's 1970.
18:21:16 <pastly> The rest may not be a big deal.
18:21:17 <isabela> so maybe we shoiuld stay and talk more with the TB team too, either on their meeting or after
18:21:30 <isabela> because we are running against time right now and i really need to know waht to expect to know what to do
18:21:38 <nickm> isis: I'm okay rescheduling the TLS1.3 meeting after the TB meeting
18:21:47 <ahf_> ok, is there a ticket yet with that? i saw your patch on the pad for it
18:21:53 <nickm> pastly: from my POV, getting Tor to work well when it is timewarped to the Long Ago Times is not really a high priority, especially if shadow could be made to start at 2003 or something.
18:21:54 <isis> isabela: i'm also pretty confused since i thought from talking to the TB people that they only needed something to test for the deadline, not the actual production server
18:22:11 <pastly> ahf_: no, and I wouldn't use my patch just yet.
18:22:17 * ahf_ nods
18:22:27 <nickm> pastly: That is, this kind of bug is just going to keep coming up, as people assume that it's the 21st century
18:22:31 <pastly> ahf_: idk if it's safe to add LargeInt in that func but not elsewhere
18:22:43 <isis> nickm: what time would be good?
18:22:54 <nickm> let's say 1 hour after this meeting ends?
18:23:08 <nickm> 2000 UTC?
18:23:13 <isabela> isis: i mean, last time we met, in december. since that wed i havent heard anything, i sent i few emails did you saw them?
18:23:15 <isis> okay sounds good (i thought that was the original time?) :)
18:23:23 <ahf_> pastly: can't shadow be "changed" to use a more modern offset? i don't know if offset 0 is always good
18:23:29 <ahf_> like, i don't know, 1/1/2000
18:23:49 <pastly> ahf_: nickm: I'll make a ticket to track this at least, even if changing Shadow is the solution landed on
18:23:59 <ahf_> cool :-) and thanks
18:24:01 <nickm> isis: I had thought we were doing 1900UTC.  But 2000 UTC is fine
18:24:02 <isabela> isis: or even the lastest one saying i was going to look for help
18:24:08 <pastly> ahf_: it's really really nice when seconds_since_start_of_sim is the same as seconds_since_beginning_of_time
18:24:24 <pastly> ahf_: and there may be uncomfortably many things that assume that's true
18:24:31 <ahf_> ok
18:24:52 <isis> isabela: i see one from 3 january that it looks like i missed?
18:25:20 <dgoulet> pastly: weird that a time period of 0 would fail... anyway, ticket is very nice
18:25:58 <nickm> our main discussion topics are for teor.  I saw some response to teor on trac and on #tor-dev about this stuff.
18:26:19 <pastly> dgoulet: I _think_ it's because there's a func to get the previous time period, which you can't in uint land if your current time period is 0.
18:26:31 <dgoulet> pastly: ah! ... yes the famous limit of time ok ok
18:26:44 <nickm> pastly: can't shadow just have the simulations start at y2k?
18:27:20 <pastly> nickm: maybe. but there may be too many things that assume secs_since_start and secs_since_1970 are the same.
18:27:23 <isis> isabela: oh, i'm sorry i missed this… it's asking for an update, but i sent one before the holidays saying the last problem was just an issue with one of the servers thinking resources were missing, which is just a configuration issue
18:29:40 <isabela> isis: not sure if you remember, but at the meeting we had on irc it was set you would send an update on friday before going to holidays, and i asked for over communication because of how close to the end of contract we are. none of that happened. inever heard anything and really, now i am quite surprise.. i have spent the last 2 weeks trying to figure out how to solve all this, it's been quite hard time
18:31:14 <mikeperry> asn,others: are you guys sure there is nothing that limits the number of hsdir attempts? I could have sword robert ransom did something like that a couple years back for v2
18:31:32 <dgoulet> mikeperry: service or client side?
18:31:39 <mikeperry> dgoulet: client side
18:31:43 <dgoulet> wait there is
18:31:57 <dgoulet> mikeperry: #tor-dev?
18:32:24 <nickm> Folks: how can we reach a decision on a hackathon this week? Teor needs a decision soon
18:32:33 <mikeperry> (err, FTR no I am not going to sword robert ransom.. sworn)
18:33:13 <ahf> i asked a question on the hackfest pad if we could have some more time periods :/ i don't know if two is enough to find some consensus there
18:33:31 <dgoulet> I put in my availability
18:33:40 <dgoulet> (the pad that is the doodle :)
18:33:46 <ahf> yep :-P
18:33:53 <nickm> ahf: try asking on the email thread as well, and adding more time periods to the pad as well?
18:33:59 <ahf> yes, will do
18:34:02 <dgoulet> yeah jsut add them I say
18:34:03 <nickm> ok
18:34:26 <isabela> isis: if you are going to work on the moat task i need to know, so I will email you the work and by when i need it and I would need some commitment on following with that
18:36:47 <ahf> ok, sent an email now so teor can see it
18:36:55 <nickm> dgoulet, teor: Do you think we should discuss the 24902 extensions here, or continue on-ticket?
18:37:09 <nickm> I agree that we might want this stuff someday, but I don't think we should block 24902 on it
18:37:13 <nickm> but I could be wrong
18:37:57 <isis> isabela: sure, i can devote some time to it. if sysrqb (or anyone else) would like to help, i would also be happy to have more eyes/hands on it.
18:38:26 <nickm> To be clear, isabela has been working on an extension for the deliverable, under the assumption that it wouldn't be done by end-of-month.
18:38:29 <dgoulet> nickm: I agree as well, I think we probably want in short term to mitigate the huge amount of direct RP circuits but the rest could be its own ticket and more discussion on tor2web
18:38:53 <nickm> isis,isabela: so we should figure out asap how far it is from done, and whether it will be done by end-of-month or not
18:39:23 <nickm> isis,isabela: We need to ask for an extension, or not.  We can't do a "maybe" here
18:39:35 <isabela> nickm: i dont think it will be done by eom
18:39:37 <nickm> dgoulet: ok. you okay with opening that ticket?
18:39:44 <dgoulet> nickm: for sure
18:39:57 <isabela> nickm: i will ask 2 months extension to be safe and hopefully that will be what we need
18:40:51 <isabela> isis: if you could define 'some amount' would be helpful to know. As I mentioned before, we are having a hard time here so knowin what to expect is very important
18:42:18 <nickm> isis, isabela: whatever the two of you and the TB team decide here, I have a request.  Could you please make sure that you put it in an email, and both look at it, and both sign off on it?  I am getting the impression that your expectations were not in alignment in December, and I want to make sure that we can get on the same page this week.  Plausible?
18:42:45 <nickm> If there's one thing you both agree to, it'll be much easier to see whether that's happened.
18:43:09 <isis> yeah, that sounds like a good idea
18:43:14 <isabela> ?
18:44:08 <isis> sorry, i'm trying to come up with a way to say the thing i want to say
18:45:25 <isabela> nickm: i will try again no problem
18:45:50 <ahf> is there anything rust planned for the rome meeting? i saw infinity0 asked about that in #tor-dev today, but i think i've only seen it brought up around the may hackfest
18:46:26 <catalyst> ahf: i thought we were thinking about one focused day of Rust stuff before Rome?
18:47:16 <ahf> is chelsea the one to hear about what is planned around that? she hasn't been on irc today, so was thinking of sending her an email
18:47:47 <nickm> I think we should plan for a rust session during the public time as well, whether we do a focused team thing or not
18:48:03 <nickm> if Rust is part of our plans, it needs to be something we're pulling lots of people in
18:48:10 <ahf> +1
18:48:32 <ahf> and i think pulling infinity0 in there would be interesting - maybe also including the reproducible build angle that i think we have discussed before
18:49:14 <isis> isabela: i think there's between 5 and 20 hours left for production meek deployment to happen, and i can spend 10 hours on it per week. if i have to do more than that, i will totally lose my marbles since i think moat is a bad design and it's frankly insulting since i poured a bunch of careful design into hyphae and instead we're doing this other thing, and i understand that we can do hyphae…
18:49:21 <isis> …later and i wrote the moat stuff so that we can literally reuse large portions of the code, it's just really frustrating. i'm not frustrated with you, or with anyone else, i'm just frustrated with the tor project in general.
18:50:40 <isis> fwiw, the rust people wanted to potentially have an Rust-Team-only All Hands somewhere in europe the week after the Rome meeting, in the hopes that some of us would like to attend
18:50:56 <nickm> isis: when will they know if that's on?
18:51:06 <ahf> cool
18:51:26 <isis> i think they are planning right now? i could go ask aaron, their team lead
18:51:40 <isis> i think it would be either at the berlin or paris Moz offices
18:52:53 <nickm> isis: I feel you wrt wanting a better solution, but: we need to finish this. Our ability to pay people is contingent on our ability to actually complete deliverables.
18:52:57 <isabela> isis: tx i will do what nick said, write down in an email etc - so both look at it and both sign off on it - having this info helps
18:53:41 <nickm> ok, we're down to 5 minutes.
18:53:58 <nickm> in 65 minutes I'm planning to tlak with isis et al on #tor-dev about TLS1.3.  Or here, if people prefer a meetbot
18:54:08 <nickm> anything else for today, folks?
18:54:50 <isis> isabela: nickm: thanks. sorry. w.r.t. i totally get it, and i'm really sorry for communicating poorly about it
18:55:42 <nickm> communication snafus are a learning opportunity.  Let's learn from this one and make sure we do stuff right going fwd
18:55:51 <isis> err "w.r.t. funding"
18:56:01 <nickm> okay, I'ma call the meeting up?
18:56:10 <nickm> thanks everybody!
18:56:10 <nickm> #endmeeting