17:00:15 #startmeeting weekly network team meeting, 11 Sep 2017 17:00:15 Meeting started Mon Sep 11 17:00:15 2017 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:15 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:00:20 * asn plays csi miami music 17:00:30 hello meeting 17:00:32 pad at https://pad.riseup.net/p/MY1GT9v1Zosn 17:00:34 hello 17:00:34 hi everyone 17:00:53 greetings 17:02:05 is mikeperry back this week? 17:02:05 hi 17:02:22 hi 17:02:28 * nickm reads through updates and stuff 17:02:30 hi folks 17:03:22 nickm: I am, but I'm taking it easy. the trip to seattle at the end of last week messed with my back 17:03:34 :( Indeed, take it easy 17:03:36 need to do a shitton of stretching to make up for not doing any during that trip :/ 17:04:46 I will dump what I have in my notebook wrt tor-core on the pad for refrerence. def won't be getting through all of it 17:06:44 So, we're late with 0.3.1.x final, but I think this last RC is pretty good. 17:07:22 We need to decide whether to put out "final" before tbb releases with 0.3.1.6-rc in an alpha, or wait for them to do it first 17:07:43 isis: Can u please explain the last line in Hyphae B.2 - "user can then stagger their other bridge purchases to achieve absolute unlinkability." 17:07:56 This friday is feature freeze for 0.3.2.x. After that point, I'm hoping that we'll fix big bugs and regressions only 17:08:04 and save the rest for 0.3.3.x 17:08:12 nickm: do you need any help coordinating with tbb? 17:08:47 isabela: wrt release dates? I don't think so. It would be convenient for us if their releases were more frequent, and very inconvenient for them. 17:09:02 true 17:09:06 unknown_artist: by "stagger" we meant that some delay can happen before the other purchases to not create a timing correlation 17:09:07 So I'm just trying to be encouraging whenever they do anything that reduces the amount of effort they need to put out updates 17:09:22 ack 17:09:40 Next deadline after that is the tor-launcher improvements that we'd like to deliver ... soon, I think? if possible... 17:10:12 and after that comes the montreal meeting, which we should all try to prepare for by being up on the various sponsors, learning a bit of rust, and I-know-not-what-else 17:10:20 what am I forgetting for our next-month timeline? 17:10:24 I would like to check in on the state of those dependencies 17:10:27 tor launcher ones 17:10:41 I think folks said mid september which is end of this week 17:10:51 isis: are the users supposed to report only when the bridge gets blocked ? 17:10:58 catalyst: you've been on that. Did you make progress there, or are things stuck? 17:11:14 also moat api from isis 17:12:30 unknown_artist: i'm not sure what you mean by "only"? could you rephrase the question maybe? 17:12:51 isabela: is the "mid-Sep" date what we estimated, what tor-launcher needs, or both? 17:13:01 nickm: I think 031 final around the corner is good, it is in a good state I believe 17:13:12 both iirc 17:13:14 dgoulet: with or without waiting for TB? 17:13:27 both 17:13:36 but that said, the moat api is higher priority 17:13:56 the captcha server is almost done, that's a large chunk of moat 17:13:57 nickm: I don't know what TB would prefer but I bet they might want the stable for their next stable release 17:14:09 isis: I mean, is it possible that users have option to submit reports at some specific time intervals about the status of bridges that they have purchased (irrespective of whether its blocked or not) 17:14:28 nickm: the floating point stuff was complicated enough to keep all in my head at once that i wanted to maximize my impact on that particular set of problems; i can probably wrap up now that i think i've done as much as i can in that space (unless something new comes up during privcount meeting) 17:14:47 isis, unknown_artist: if you could take that discussion in #tor-project maybe or after this meeting, would be neat :) 17:15:00 isis: how is your estimation for what is left to do? 17:15:02 ok. let's postpone anything coming out of the privcount meeting, if it will help us get the tor-launcher stuff done on time 17:15:10 nickm: sounds good 17:15:21 dgoulet: yeah, sounds good, i'm having trouble having four convos at once :) 17:15:28 is it realistic that there are simple improvements we could do to help tor-launcher, and land this week? 17:15:29 catalyst nickm tx 17:16:11 like, some small events that would be easy to add to the current system, and give good results? 17:17:09 nickm: it's probably easier to do extend the existing OR_CONN events, but i'm not sure 17:17:25 isabela: i think if i worked all week only on that and nothing else, didn't read email or talk to anyone or literally do anything else, and there weren't any surprises or hitches anywhere, that it could all get done 17:18:10 * armadev is nearby if people need something 17:18:10 nickm: extracting the bootstrap reporting from deep in the call stack and having higher level code track that is probably doable this week but it's mostly groundwork for the visible improvements 17:18:16 isabela, isis: is there a subset that would be realistic to accomplish this week in a healthy way, that would give tor-launcher the most important parts? 17:18:40 isis: that sounds a little tie 17:18:44 what nick says 17:18:58 nickm: i think isis knows better than me :) 17:19:12 well, you may have a good sense of "what tor-launcher needs right now" 17:19:15 like, maybe a stable API? 17:19:17 dunno 17:19:32 we need to have the captcha working on the launcher 17:20:05 i am not sure what is the work being done at the api - if its to make this happen or if its improvements on the top of something that is already working or 17:20:16 so one thing we could do, i fi document better the API, is that tor launcher could make progress implementing the client-side in parallel 17:20:21 catalyst: I think the refactoring is something to do after this week; I think for now we should pick a few events with highest benefit/difficulty ratio, and implement the easiest ones 17:21:50 isabela: afaik the actual server doesn't need to be in production for tor-launcher to start coding up talking to it, as long as we agree on what "talking" means, right? 17:22:20 isis: i will check with them at their meeting. 17:22:34 nickm: i hesitate to drop more bootstrap reporting deep into the call stack (which would be necessary for the improvement i'm thinking of if we're not refactoring first). highest benefit/difficulty work in that category is reporting per-connection PT progress 17:22:50 isabela: so maybe one thing we could schedule this week, if Pearl Crescent has time, is a meeting where we go over the API and see if there's anything they can think of to change so that we agree on the interface 17:23:01 isabela: cool, thank you! 17:23:10 isis: sounds, good i will work on it 17:23:36 catalyst: huh. I thought that earlier, we had thoguht there was a good chance that there was stuff _not_ hanging around at the bottom of the callstack that would be useful to them. 17:23:45 Did that turn out not to be the case? 17:24:56 i'd have to check my notes -- i thought making PT progress visible was one of the better goals 17:25:17 I agree that one is important 17:25:28 but if we can't have it done well this week, is there something we can? 17:25:37 is it helpful for the bootstrapping work for tor launcher if i revise #11966 ? 17:26:03 (i totally forgot about that ticket) 17:26:10 hm, I don't remember so well what the status was 17:26:19 it might help, yes. or it might be a time-sink 17:26:20 nickm: I'm looking at #23077 and thinking the easiest thing to do is eliminate any actual use of libevent and call the callbacks manually assuming they would fire properly. this will test a lot less stuff, but otherwise I have no idea how to both mock time and make it move forward in a way that will cause lievent+timers.c to wake up and call the callbacks.. 17:26:57 isis: it's probably less important than the moat api stuff 17:27:06 nickm: ack 17:27:26 mikeperry: maybe I can stick something in the timers.c level 17:27:57 yeah, that would be easy 17:27:59 would that help? 17:29:13 probably. there may be other problems as well, but I think that would take care of the most obvious one 17:29:47 okay. It should take ~ 10 minutes 17:30:22 catalyst, isabela: maybe we should strategize after this meeting to determine which bootstrap improvements if any we should do this week. Do you both have time? 17:30:47 dgoulet, asn: Should I expect any more big branches this week? 17:30:57 nickm: not big, just bugfixes, they are under review 17:31:00 nickm: i'd like a half hour to eat lunch but sure 17:31:11 nickm: maybe i can get this included with the meeting with pearl crescent that me and isis were talking about? 17:31:18 ok; when is that? 17:31:36 i plan on figuring this out at the tor browser team meeting 17:31:38 dgoulet, pastly: Do you think it makes sense to try to get KIST in 0.3.2, or will 0.3.3 be wiser? 17:31:42 which is right after this one 17:31:42 nickm: oh and probably a branch that adds unit tests so nothing like last week hehe 17:32:00 isabela: okay. So one hour after the ending time of this meeting, we should know when we get together to do what? 17:32:03 nickm: half an hour after this meeting ends :) i guess after the tbb-team meeting is probably fine 17:32:12 nickm: I think 032 so we can get this nice 3 months period of stabilizing, pastly has been testing that branch for a while now and so am I on my public relay 17:32:12 nickm: kist should fit into 032 easily. It's "done" just a few more polishing back-n-forths with dgoulet I think 17:32:29 I got good graphs from running a 2w experiment 17:32:44 wow 17:32:46 nickm: 1. know if they can start using the api calls for moat so they dont need to wait for the server to start coding 17:33:22 nickm: 2. know what bootstrap events are priority from the list catalyst can do for them 17:33:23 nickm: one thing that is slowing me down looking through connection stuff is it looks like both proxies and PTs get called "proxies" and i'm not sure which code deals with PTs and which doesn't 17:33:25 ? 17:33:30 i think these were what ppl asked for 17:33:32 in this meeting 17:34:10 dgoulet, pastly: okay. If we're going to try to do that, I think that's the biggest branch that's currently trying to get in... so it would be good if I were reviewing it soon 17:34:40 pastly: nice! 17:34:40 catalyst: ok, we should go over that. 17:34:53 catalyst: I think that's something I can try to sort out with you 17:35:11 oh i guess the handful of specific errors they wanted could be easier? 17:35:25 catalyst: if we can get some of them this week, that might help them? 17:35:27 nickm: dgoulet: I'm hoping to review david's work today. Sooooo stay tuned and maybe you can look this week. Assuming you're not swamped of course 17:35:45 nickm, pastly: it's on oniongit, I think you both can review it in parallel easily 17:36:12 fwiw: I'm basically planning that after today, I might have to spend the rest of the week just trying to be helpful. so please everybody feel free to ask me to do stuff 17:37:33 catalyst isis nickm -> i will see if we can have this sync with the tbb team asap so your questions are answered 17:37:43 ok great. 17:38:02 nickm: it looks like they've already highlighted a few specific logging calls that maybe should be exposed as control protocol events so that seems not too hard 17:38:03 I'll make sure to be around til at least 5pm eastern today 17:38:29 catalyst: that sounds very promising 17:38:56 ok 17:39:24 going over the pad -- is there anything we forgot to talk about today? 17:39:41 the privcount meeting is later in the day tomorrow; please stop by if you're interested. It's right after the usual patch party time 17:39:47 the metrics team request? 17:39:48 i've two quick things 17:40:16 wrt the metrics team request, I think we should ask if we can look over those tickets after this week? It seems we're pretty hosed 17:40:19 ahf: go 17:41:04 1) does anybody know how the hiring process with the android OS person is going? i'd find it useful to sometimes have one to chat with about some of the android things where i feel that i'm chasing things in the dark 17:42:00 speaking of hiring, the shiny new tor browser dev used to be a microsoft person, and knows windows stuff. 17:42:07 2) i'd like to flesh out a time plan for the sponsor 8 network team tasks this week. would isabela + nickm (maybe?) have time for that during later this week? 17:42:10 armadev: :O 17:42:11 ahf: i think we are in selection process to move candidates to coding task phase 17:42:39 ahf: but maybe you can ask your android questions at the tor-mobile irc channel or directly to some of the folks from guardian project? 17:42:47 isabela: oh, ok, cool! that sounds like we are far 17:42:52 there is a tor-mobile channel! :O 17:42:55 ahf: yes! re: sponsor8 work 17:43:05 dgoulet: his name is richard, and he's coming to montreal. also he is probably lurking on this channel under some name. :) 17:43:08 i have for this week to organize it all 17:43:13 armadev: very nice 17:43:19 isabela: yep! i'm in it and observing in there. it just seems very inactive or maybe i've just not paid much attention to it 17:43:22 so i would love to chat more about it, i would prefer it to be towards the end of the week tho 17:43:29 like thursday 17:43:40 isabela: ditto! my brain is just starting to be back in being able to concentrate fully again 17:43:46 thursday sounds good. i will be around all day 17:43:50 cool! 17:44:04 nickm: ^^ 17:44:12 maybe catalyst could join too? 17:44:12 i'd like it such that we can both create tickets and get the sponsor 8 wiki page more up-to-date. right now it just contains my notes on getting a dev environment up and running 17:44:20 yes 17:44:23 that sounds like a plan 17:44:27 great! 17:44:36 another thing is to review the M&E we wrote for the proposal 17:44:37 thursday around this time? (17 UTC?) 17:44:55 me and nick might be at the vegas meeting 17:45:03 1730 would work better? 17:45:15 probably 17:45:16 i'm around all of thursday so you can define the timeslot :-) 17:45:18 yep, that is fine 17:45:32 then let's say 1800 UTC on thursday, if it works for isabela -- then I can have lunch 17:45:37 Thursday is probably doable for me 17:45:48 yeah that is even better :) gives me lunch time too 17:45:49 heheh 17:45:54 18 UTC is also better for me, then i can have dinner before :-P 17:46:32 catalyst: 18UTC works for you too? 17:46:41 isabela: yes 17:46:45 awesome 17:47:11 cool, that was all i had. is nice to be back after the european hacker festival season is over :-) 17:47:39 nickm: also if there's someone willing to take my bug triage rotation this week that might help me with focusing on the tor launcher stuff 17:47:55 i can do that unless somebody else picks it up 17:47:59 catalyst: i can do that if you want to 17:48:15 ahf: thanks! 17:48:19 i have next week, we can just swap? 17:48:29 do you want to keep coverity this week? 17:49:07 ahf: yeah swapping both bug triage and coverity for this week and next sounds good, thanks! 17:49:13 mikeperry: 72ea4a8f081318c60c460cef5d9daf55e399c434 in maint-0.3.1 should do what you need 17:49:42 catalyst: great, will update the wiki 17:50:15 anything more for this week? 17:50:33 so by friday it's feature freeze 17:50:41 on friday* 17:50:53 that's the plan 17:50:56 and I guess #12541 is the only big feature left 17:50:59 ok cool 17:51:05 * dgoulet is good 17:51:19 gotta be reviewed by friday? 17:51:26 catalyst: hm, maybe you should find someone else for coverity + bug triage for the 25/09 - otherwise you'll have two weeks in a row with these tasks 17:51:27 gotta be _merged_ by friday if it goes in 0.3.2 17:51:30 pastly: merge_ready by friday 17:51:34 got it 17:51:43 I,ll be sharp on it 17:51:45 and the goal is not for stuff to squeak in at the last minute 17:52:14 ahf, catalyst: let me take this week, simplify everything 17:52:14 ahf: good point; that's not so urgent though 17:52:37 ye 17:52:39 p 17:52:43 dgoulet: that would be great. thanks! 17:52:57 ok. thanks for another fine week, folks! I'll be online on and off till 5pm eastern, looking for pings from catalyst and isabela wrt info on talking about tor-launcher events 17:53:05 anything else? 17:53:05 catalyst, ahf: there, wiki updated 17:53:10 alright! 17:53:11 :) 17:53:24 thanks folks 17:53:26 o/ 17:53:30 #endmeeting