17:01:21 <nickm> #startmeeting network team fun and (non)profit! 17:01:21 <MeetBot> Meeting started Mon May 23 17:01:21 2016 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:21 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 17:01:27 <nickm> hi all 17:01:41 <nickm> I've seen isis and armadev and Yawning and isabela . 17:01:46 <nickm> Who else is around? 17:01:52 <asn> hello peeps 17:02:00 * teor feels invisible 17:02:00 <nickm> whoops 17:02:01 <nickm> hi teor! 17:02:02 <teor> lol 17:02:28 <nickm> can somebody else check in first? I need to make sure I haven't forgotten anything. :) 17:03:02 <isis> armadev: nickm: aha, that would make sense because `nm -D /usr/lib/libscrypt.so.0` said "log" was undefined 17:03:26 <nickm> (anybody? check in? please?) 17:04:00 <armadev> my check-in is simple: i did some tickets and stuff, but nt as much as i might have liked. i am stuck in the "have to do bureaucracy, shouldn't do fun stuff until do bureaucracy, omg it keeps adding up" loop. 17:04:49 <armadev> the cheerful version of that statement though is that soon we will have some funding to implement and deploy website fingerprinting defenses. 17:04:57 <armadev> now all we have to do is ..have some defenses we want to implement. 17:05:08 <nickm> thx 17:05:09 <armadev> (end of check in) 17:05:20 <asn> can go next! 17:05:22 <asn> Hello. During last week I worked more on prop224 stuff. Did some more prop250 17:05:23 <nickm> go! 17:05:24 <asn> testing (need to do some more of this). Did some code reviews. Did some bug 17:05:25 <asn> bounties. Need to do more of everything the upcoming week. Also I have not 17:05:31 <asn> touched prop259 (guard stuff) in a while... I do plan to attend tomorrow's 17:05:31 <asn> network team meeting (it's 11PM in my timezone)! Done. Next! 17:05:54 <nickm> yo 17:05:55 <nickm> me now 17:06:16 <nickm> I did lots more review and merging, and tried to get 0.2.8.3-alpha out, but couldn't because of some showstopper bugs. Fixed most of those. Still stuck on one. 17:06:23 <armadev> which one? 17:06:27 <nickm> Dealt with more wildcard cert stuff and bureaucracy. 17:06:32 <nickm> finished the bug retrospective, got it on the blog 17:06:49 <nickm> Almost done with 6s194 students; encouraged them to submit final code asap for review 17:06:53 * isis created #19161 17:07:06 <nickm> the bug I can't figure out is toralf's #19128 17:07:16 <nickm> I fixed some related stuff but it's still crashing 17:07:26 <armadev> does it happen to anybody else? 17:07:28 <nickm> I have some discussion topics for later: I'll start a pad for us to add more 17:07:37 <nickm> armadev: Nobody else has complained, but who knows. 17:07:43 <athena> o hai meeting 17:07:46 <armadev> ok. we can add this as a discussion topic not a now-now-now topic. 17:08:25 <teor> nickm: it might help to ask toralf to checkout a fresh copy of the source, their method of patching seems error-prone 17:08:27 <nickm> https://pad.riseup.net/p/pzaZKrggUxVo -- discussion topic pad 17:08:41 <nickm> teor: it's conceivable 17:08:56 <nickm> please add discussion topics there. 17:09:05 <nickm> any questions for me, or more checkins? 17:09:24 <teor> I'll check in 17:09:33 <toralf> teor: I do disagree, I tested it several time w/ and w/o the patch and thee way I do patch works since months 17:09:36 <isis> i'll go after teor 17:10:03 <teor> toralf: ok, so if it happens without the patch, then we have an issue 17:11:06 <toralf> teor: currently I try to (re)produce that issue with better log files and/or a gdb trace but that exit relay lost stable and guard flag and has much less bw than before - needs time to reproduce 17:11:21 <teor> I logged some bugs against prop250 and finished off some other tickets and reviews for the release 17:11:45 <teor> #18963 17:12:00 <teor> #18616 17:12:25 <Yawning> hi 17:12:36 <teor> I think that's it from me 17:12:37 <Yawning> sorry, was pooping, and it took longer than I thought it qould 17:12:39 <nickm> (hi!) 17:12:40 <athena> checkin: code reviews; i kinda feel like i was too hard on #7478 because the existing code was already ugly though 17:12:41 <Yawning> *would 17:13:29 <isis> worked on mostly bridgedb stuff for the database switch 17:13:29 <isis> worked on #7144 which is still having some bad reactions with fallbackdir code :( 17:13:31 <isis> replied to a bunch of stuff on the PQ handshake thread 17:13:34 <isis> came up with a name for the new handshake, "RebelAlliance", since it's an alliance between X25519 and NewHope 17:13:37 <isis> spent a bunch of time dealing with community stuff of bad stuff being reported to me and then needing to report it to shari :( 17:14:13 <isis> Yawning: ewww 17:14:14 <isabela> oi 17:14:17 <isabela> i am around sorry 17:14:19 <isabela> (late) 17:14:21 <mikeperry> RebelAlliance! 17:14:22 <Yawning> :( 17:14:22 <isis> okay, that's it for me 17:14:25 <Sebastian> I finally concluded that the #17668 patch works 17:14:35 <Yawning> eerybody poops 17:14:45 <nickm> (not everybody tells #tor-dev) 17:14:54 <asn> even the pope poops 17:15:03 <nickm> (I usually say "having a bio break" or "taking a euphemism") 17:15:06 <Sebastian> dizum went down while gabelmoo stayed up. We still don't have a stable release of Tor for dirauths to run, only 0.2.8.2 onwards 17:15:22 <Sebastian> sorry that I'm interrupting your potty talk :) 17:15:39 <isis> mikeperry: glad you approve of the name :) 17:15:42 <Sebastian> I want to send a mail to dirauth ops advising what to do now 17:16:03 <Sebastian> which is probably to update to 0.2.8.x since there's no 0.2.7.x release scheduled 17:16:04 <nickm> wait till 0283-alpha has been out a couple of days, and start trying it? 17:16:10 <Sebastian> sounds good 17:16:30 <armadev> sebastian: do we know what happened to dizum? 17:16:34 <Sebastian> I'll also put the document on ed keys up on the wiki as discussed by isa during last patch workshop (thanks isa) 17:16:38 <Sebastian> ok that's it for me 17:16:44 <Sebastian> armadev: #17668 happened? 17:16:55 <Sebastian> armadev: let's take it offline, I think it's clear what happened 17:16:58 <armadev> ok 17:18:23 <Yawning> (writing documentation, I'll probably review code, still need to return paperwork to shari, bleah) 17:18:57 <armadev> yawning: getting you signed up as reviewer on some tickets would be grand 17:19:31 <isis> i thought "bio break" meant "going to the bio supermarket to get food"… L/ 17:19:33 <Yawning> i don't even know what my prioritization is supposed to be 17:19:41 <nickm> isis: it could! 17:19:44 <Yawning> which is why I wanted more clarity on that 17:19:45 <nickm> isis: it means "doing anything biological" 17:19:48 <Yawning> but I didn't have much luck 17:19:54 <Yawning> w/e yeah I'll review stuff 17:20:16 <nickm> so, the stuff currently tagged with review-group-2 needs reviews. The Reviewer field is supposed to be people who said they'd do it. 17:20:48 <armadev> yawning: we (all of us and you) should figure out how to have you feel comfortable just picking stuff that needs doing, and doing it. asking for what to do at each step is likely resulting in inefficiency. 17:21:03 <nickm> the stuff in 0.2.9.x with Sponsor field = sponsorX-must or sponsorX-can is funded; it's fine to grab any of it. 17:21:19 <isabela> yes 17:21:22 <nickm> Doing stuff from among those that you want to is probably valuable. 17:21:48 <asn> ah so there is a review-group-2 now? ok i will pick needs_review tickets from there now. 17:21:59 <asn> (fwiw, i liked the review-group-1 experiment.) 17:22:02 <nickm> yup; everthing in review-group-1 got at least 1 review. 17:22:07 <isabela> :) 17:22:08 <nickm> it's kind of a batchwise FIFO. 17:22:18 <nickm> to prevent stuff from waiting forever for a review 17:22:20 <nickm> "so far so good" 17:22:25 <armadev> https://trac.torproject.org/projects/tor/query?status=accepted&status=assigned&status=merge_ready&status=needs_information&status=needs_review&status=needs_revision&status=new&status=reopened&keywords=~review-group-2&col=id&col=summary&col=keywords&col=status&col=owner&col=type&col=priority&col=reviewer&order=priority 17:22:51 <nickm> (if you are listed as a reviewer on that link, make sure that it's something you really mean to review! 17:23:09 <nickm> (if you see an item there with no reviewer, please consider adding yourself) 17:23:31 <armadev> and if there's an item in needs-review with no owner, it means somebody wrote a patch but nobody is leading the ticket? 17:23:48 <isabela> who writes should owns it 17:23:53 <nickm> armadev: it means that somebody forgot to add themself as owner, or the patch is from cpunks, or such 17:23:56 <nickm> feel free to fix 17:24:04 <isabela> because who writes it will have to do revision of the code review 17:24:05 <isabela> etc 17:24:26 <teor> I liked review-group-1 17:24:34 <nickm> thanks; me too 17:24:50 <isabela> btw I worked on an agenda for tomorrow retrospective 17:24:53 <Yawning> ah 17:25:01 <Yawning> so I can pick anything that has funding 17:25:07 <Yawning> and not get in trouble with accounting 17:25:08 <Yawning> ? 17:25:17 <isabela> hehehe 17:25:20 <isabela> yes 17:25:32 <isabela> anything on 029 with sponsor --must or --can 17:25:37 <isabela> is cool to just pick and work on 17:25:52 <isabela> U and S specially 17:25:54 <nickm> yes, though if we only do "can" activities and we don't get the "musts" done, that's a problem. 17:26:05 <isabela> yeah must should have more love 17:26:29 <nickm> also, unfunded stuff isn't "never do". It's 'see if we can find a sponsor, consider if we can get it done in less than an hour, consider whether it's super-important to do anyway...' 17:28:13 <isabela> here is the agenda for tomorrow https://pad.riseup.net/p/PLFZzMjdlQFx 17:28:19 <isabela> (DONT ACT ON IT NOW :) 17:28:20 <isabela> hehe 17:28:28 <isabela> is just for you to know what we will be doing 17:28:34 <isis> why are there only 3 non-closed tickets under TorCoreTeam201605 ? 17:28:36 <isabela> i will email too 17:29:01 <teor> isis: maybe we're organised? 17:29:46 <nickm> isis: huh? 17:30:04 <isabela> i see 38 17:30:11 <isabela> isis: are you sure? 17:30:26 * nickm sees 38 too 17:30:40 <nickm> anybody else with a check-in? 17:32:55 <nickm> okay, let's move on to discussion. 17:32:58 <armadev> i have a pair of tickets i need to open still from montreal. about refactoring a struct. 17:32:59 <nickm> I see discussion topics from me and teor 17:33:02 <nickm> let's do teor's first? 17:33:10 <armadev> if somebody could remind me of anything else they thought i needed to do from montreal, but didn't, and i forgot, please do. 17:33:26 <nickm> also please remind me of anything you think I said I'd do 17:33:33 <nickm> (if I haven't done it) 17:33:34 <asn> do we have a ticket for the "only assign HSDir flags to relays with uptime N+ months"? 17:33:40 <nickm> dunno 17:33:59 <asn> not sure if someone made one during the meeting 17:34:02 <armadev> asn: we have no ticket. but i have a patch. 17:34:22 <asn> ok. i can make a ticket. 17:34:23 <armadev> great 17:34:32 <asn> i'll make a ticket and let you know. 17:34:41 <armadev> my patch is not fool proof in that people who know the algorithm can still game it. so sad. better than now still. 17:34:51 <teor> asn: did we decide on the number of months? 17:34:52 <asn> toralf cleverly said that this security feature might incentivize (?) people to not upgrade their kernel so that they dont lose uptime 17:35:02 <isis> do we want #8051 in 0.2.8-final still? 17:35:05 <teor> asn: fallbacks are similar 17:35:16 <nickm> I am okay deferring lots of stuff to 0.2.9 17:35:19 <asn> teor: dont thin kwe decided on nr of months 17:35:36 <nickm> isis: let's defer; it seems harmless, yeah? 17:35:48 <armadev> teor: my patch demands >= three-quarters of bandwidth, and >= top 1/4 of time-known 17:35:52 <nickm> (to 0.2.9) 17:36:34 <teor> armadev: so we get around 1000 HSDirs? What are the numbers? 17:37:05 <armadev> https://consensus-health.torproject.org/#overlap 17:37:13 <armadev> 926 currently, yeah 17:37:17 <nickm> teor: your questions seem mostly "can we defer this to 0.2.9?" ? 17:37:33 <teor> yes 17:37:39 <armadev> (also, i think my patch might be broken. or maybe time-known itself is broken. we can take it to asn's ticket once he has one.) 17:38:05 <nickm> None of the stuff you listed looks like it really has to be in 0.2.8, though I'm generally unsure about reverting. 17:38:43 <nickm> anybody want to argue that any of teor's stuff listed on the pad (https://pad.riseup.net/p/pzaZKrggUxVo) really needs to get done in 0.2.8? 17:39:07 <nickm> isis: ohai btw I went and wrote a patch for #19161 ; please test when you can? 17:39:20 <teor> (#19071 belongs in 0.2.8 because we should check the fallbacks before we release) 17:39:56 <teor> I would like to kick them all out of 0.2.8, and put them back in if someone complains during 0.2.8.3-alpha 17:41:05 <isis> nickm: added myself as the reviewer 17:41:46 <nickm> teor: that's fine with me. anybody who disagrees should object 17:42:05 <nickm> we have 20 min left for topics I listed? 17:42:09 <nickm> ok 17:42:28 <armadev> tell us more about 'claiming tickets for june'? 17:42:48 <nickm> 1. If you're on the network team mailing list, please come to the process meeting Isabela scheduled for tomorrow. If you are paid by Tor, that goes double. 17:43:09 <nickm> armadev: isabela will correct me if I'm wrong, but: 17:43:33 <isis> if something is "needs_revision" should it still be in "review-group-2"? 17:43:34 <nickm> If there is a network-team/core-tor task you will do in june, make yourself the owner of it, give it "TorCoreTeam201606", and make sure it has a "points" setting. 17:43:39 <nickm> isabela: (is that right?) 17:43:44 <isis> e.g. #17435 17:44:07 <isabela> yes 17:44:08 <asn> teor: armadev: here is a ticket #19162 17:44:12 <asn> feel free to modify it as you see fit. 17:44:13 <nickm> isis: we'll leave it in, on the theory that if it gets a revision, it gets reviewed. 17:44:35 <nickm> Everybody please review stuff from review-group-2. 17:44:37 <asn> also should I fill the "points" field? i wonder how many points its worth. 17:44:48 <nickm> points == "about how many days do you think it will take you" 17:44:56 <nickm> actual points == "about how many days did it take you" 17:45:06 <isabela> yes 17:45:10 <nickm> no fractions on values >= 1.0. 17:45:13 <isabela> the scale is on the wiki 17:45:28 * isabela looks for link 17:45:49 <isabela> https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/ReleaseGuidelines#TeamCapacity 17:45:55 <isabela> i am converting points 17:46:10 <isabela> i did for may and i start doing for all 0.2.9 17:46:11 <nickm> also: Please have a look at everything where you are Owner or Reviewer in TorCoreTeam201605. Please make sure it's correct-ish. :) 17:47:38 <nickm> any discussion/questions/objections/notes on the above stuff? My last topic is 029-proposed 17:48:51 <nickm> there are 22 029-proposed tickets right now. 17:49:31 <nickm> i think we should figure out some "obviously-yes" categories so that we can figure out which 029-proposed t8ickets don't fall into those categories, and reserve discussion for those. 17:49:37 <nickm> Obviously-yes categories include: 17:49:40 <nickm> - severe bugs 17:49:40 <armadev> is asn supposed to be putting SponsorR in the keywords still? or is that only for the Sponsor field now? 17:49:42 <nickm> - regressions 17:49:54 <nickm> armadev: Sponsors go in "Sponsor". :) 17:50:06 <nickm> - dependencies of other tickets already in the milestone 17:50:07 <isabela> sponsor field 17:50:15 <nickm> - Stuff that takes less than 10 minutes 17:50:19 <nickm> - anything else? 17:50:28 <isabela> that sounds good 17:50:33 <nickm> - Needs-review stuff with an easy-to-review patch written externally (?) 17:50:41 <isabela> yes 17:50:42 <isabela> i was thinking of that 17:51:01 <nickm> any more stuff that should be include-in-(development)-branch-default? 17:51:28 <nickm> - "we realized that this is actually a must-do funded item that we forgot to include, whoops" 17:51:28 <asn> maybe I should put #19162 in Tor: unspecified? and mark it as 029-proposed? because i just crammed it in 0.2.9.x which is a faux pas as I understand it. 17:51:49 <nickm> asn: yes please :) 17:52:19 <isabela> also, we should at some point review the 0.2.9 milestone 17:52:27 <isabela> i will work on that 17:52:34 <nickm> +1 17:52:52 <Yawning> ... 17:52:56 <nickm> Yawning: ? 17:53:20 <armadev> nickm: when i make a ticket, say in May, and then i work on it, should i have added the TorCoreTeam201605 keyword? 17:53:22 <Yawning> (still trying to wak eup) 17:53:24 <teor> nickm: child of ticket in branch, that's a bug or an integral part or a breakdown of the ticket 17:53:27 <nickm> armadev: ask isabela 17:53:31 <isabela> armadev: yes 17:53:37 <isabela> so we know you are doing it 17:53:38 <nickm> teor: I'm calling that a "dependency" 17:53:43 <armadev> ok. so those tickets we made in montreal, if we intended to work on it, we should have added that keyword. 17:53:54 <nickm> teor: 17:50 < nickm> - dependencies of other tickets already in the milestone 17:54:02 <huseby> mornin' 17:54:07 <armadev> isabela: should i go back and do that for ones i failed to do it for? or only the ones that i'm still working on, i.e. that aren't done yet? 17:54:07 <isabela> you add the keyword on the tickets you are planning to work in the month 17:54:57 <isabela> yes, if is ok to add to all 17:55:09 <armadev> ok 17:55:16 <isabela> right now i cant build a report of how we are doing because of missing data 17:55:50 <isabela> and without a report we dont know if any of this makes sense :) if its helping 17:55:55 <armadev> i will do that shortly by using the 'timeline' feature on trac to remind me what tickets i messed with, and then adding that keyword to the ones i worked on in may. 17:56:04 <armadev> (you all can do this too :) 17:56:10 <GeKo> five minutes to tor browser meeting 17:56:19 <isabela> ! 17:56:38 <nickm> cool. I'll be hanging around on #tor-project another 20 minutes-ish 17:56:46 <nickm> anything for the next few minutes? Thanks, all! 17:56:57 <teor> Can I have a volunteer to help with #19071? 17:57:02 <isabela> armadev: i forgot about this feature! good reminding 17:57:19 <nickm> oh, additional category: 17:57:22 <teor> So I'm not the only one who can make a fallback directory list 17:57:22 <nickm> - code removal 17:57:29 <armadev> isabela: timeline is one of the best features! it lets you answer "what was that ticket" questions 17:58:12 <armadev> isabela: to be clear, TorCoreTeam201605 is for "i intend to fully resolve this ticket in may"? or also "i intend to make some progress on this ticket in may"? 17:58:47 <isabela> fully resolve it, we should try to close tickets in the sprint -- specially because we already have a very long sprint 17:58:48 <nickm> teor: maybe explain what skills and/or work are needed, and/or invite somebofdy who you think would be good at it? 17:59:07 <isabela> maybe breakdown in child tasks 17:59:13 <isabela> if all cant fit in 17:59:47 <teor> someone needs to be able to run a python script that connects to some tor directory mirrors on a decent non-tor network connection 17:59:58 <isabela> (we need to go) 18:00:07 <isabela> :) 18:00:17 <isabela> nickm: end meeting? 18:00:21 <teor> and then read the logs and contact relay operators whose relay details have changed 18:00:22 <teor> done 18:00:29 <isis> Yawning: if you want to help me debug why #7144 doesn't bootstrap anymore with fallbackdirs… :) 18:00:36 <isabela> (we continue at #tor-project) 18:02:32 <nickm> #endmeeting