17:00:38 #startmeeting network meeting for May 18th 2020 17:00:38 Meeting started Mon May 18 17:00:38 2020 UTC. The chair is gaba. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:38 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:00:42 o/ 17:00:52 pad: https://pad.riseup.net/p/tor-netteam-2020.1-keep 17:02:08 hope you are all doing ok. Take time to update the statuses at the bottom of the pad 17:02:39 times are interesting and strange around here, but i hope this will be a good week 17:03:24 i had way too many meetings last week and haven't had time to do the network team stuff that i needed to do 17:03:30 i don't know if we have dgoulet today 17:03:42 Caitlin messaged me to say that they might not be able to do this meeting either, but they would try 17:04:13 yes, I think this week is more open for many of us 17:04:17 ok 17:04:37 dgoulet is out today 17:04:45 yes, holidays 17:04:45 it's a public holiday in ca i think 17:04:54 ah yes 17:05:05 Next monday in the US 17:05:14 about the sponsor work: do all of you know what you need to work on this week? any change that you want to include? 17:05:20 i think thursday is a holiday here but i might swap it with friday 17:05:21 oh yes 17:05:27 gaba: yep 17:05:29 why do u have holidays people? 17:05:35 i mean in may 17:05:51 i mean apart from may first 17:05:53 anyway what's up? 17:05:57 asn: you could make one up; i'd believe you :) 17:06:15 because of jesus take off from the cave 17:06:19 i think 17:06:44 love it 17:06:52 "Feast of the Ascension" in english apparantly 17:07:10 that's not leaving the cave, that's going up to heaven 17:07:28 leaving the cave is easter 17:07:36 the one from the US is about war... 17:07:36 anyways ... :) 17:07:39 anyway 17:07:48 ah, i meant take-off like a rocket here 17:07:49 anyway 17:07:50 yeah 17:07:52 anyway 17:08:06 nickm: you are not doing any s55 this week, right? wrapping up s69 with walking onions? 17:08:48 gaba: I am hoping to do enough s55 so that i can be smart about it, and know what's in store. That's going to mostly be reviewing teor's code and merging what i can, planning work to wrap up what i can't, and seeing what remains to be done 17:09:10 ack 17:09:39 but yeah, i want to wrap up walking initial walking onion designs 17:10:19 reviews for this week. It seems that there are a few not assigned. 17:11:44 fun 17:11:52 will do so asap 17:12:11 thanks asn 17:12:12 any blocker from last week? 17:12:15 for me, things stacked up a lot last week, i hope this week does not continue this trend 17:12:23 same here :) 17:12:40 i don't know anything i'm blocked on now, but one task I have is triage for 0.4.4 17:12:50 since we have a touch less than a month until 0.4.4 freeze 17:13:48 i am not blocked 17:13:59 geko is teaching me the ways of the gecko 17:14:28 For 0.4.3 there are some tickets owned by catalyst. should we get somebody else to grab them? https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases/043Status 17:14:40 is s28, no? 17:15:00 s28? 17:15:09 sponsor 28 17:15:09 i think part of 0.4.4 triage should be to look at the current 0.4.3 tasks and see what to do with them; I don't have a uniform answer for all of them 17:15:22 oh yes 17:15:46 nickm: is this something you are going to do this week? do you need help? 17:16:04 i can grab #5304 and #31009 during that process, but right now i cannot say how much non-browser stuff my mind can handle right now 17:17:04 gaba: i do need some help with 044 triage. ideally this would break down into should/must as usual for the 044 timeline 17:17:23 ahf: do you know when you expect to return from browser-land? 17:17:46 #5304 is a must for s28 but i think phw or cohosh are not blocked on it. It can wait further down the road. 17:17:49 i don't know what the timeline is in the end of the timeline. gaba knows 17:17:54 but all s28 things are not urgent 17:18:06 they can wait until after i return i think 17:18:16 ahf will return! in october... 17:20:58 so first step for 044 triage is to ask, what do we actually need to do in 044? AFAIK there is no external "must" from sponsor work, so we just need to collect our "musts" from internal stuff, like security and correctness and stuff tbb needs 17:21:00 :-) 17:21:16 then we label the musts and the shoulds 17:21:34 and get ready to kick out everything else when we hit the feature freeze 17:21:54 do you want to do it now or after the meeting with somebody else? 17:22:29 probably after the meeting is better , we don't need to halt everybody for this 17:22:41 ok 17:23:57 Continuing with the meeting, nickm needs help with #33850. Can anybody else take it? 17:24:21 that's just one that i hope we don't drop in 044 17:24:27 I can throw an 044-should on it and be happy 17:25:34 ok 17:27:07 (what's our next topic?) 17:27:22 (oops -- am i supposed to be running this meeting?) 17:27:28 anybodye lse needs help? 17:27:30 else* 17:27:35 yes 17:27:41 well I have that discussion question 17:27:55 yes, let's move into discussions 17:28:00 Current plan is that FlashFlow will consist of a program that controls tor clients via the Control Port. Like Simple Bandwidth Scanner and Torflow. Would writing this code in Rust make it non-transitionable? Does the network team know enough Rust? The other idea is Python 3. Is that way better? 17:28:02 the first one is yours pastly 17:28:33 (didn't we also want to talk about proposals here? or was that just if we hadn't made progress on the ML's?) 17:28:49 what's a ML? 17:28:55 mailing lists 17:28:57 ahhh 17:29:13 let's briefly summarize where we are with ML progress after we talk about pastly's questino 17:29:16 *question 17:29:25 I don't think rust is a blocker for me in principle 17:29:52 this is something that would want to run on some fraction of relays, but not all? 17:29:57 i don't know why you would do in rust unless you also have time to rewrite a good bunch of stem in rust too 17:30:35 nickm: this Rust codebase would be run by bwauths (or people the bwauths trust) and not on relays 17:31:04 I don't think rust is a blocker in principle 17:31:20 (does anybody think it might be?) 17:31:38 ahf: fair of course. It isn't "a good bunch" IMHO, just a couple of new commands. 17:31:56 I considered mentioning a python shim between rust and tor, but that seems unnecessarily complex 17:31:57 i dont think rust is a blocker, but also i dont know much rust 17:32:45 * ahf is fine with either rust or python too 17:33:04 So while Rust may be acceptable, is Python 3 still way better for transitionability, do you think? Or maybe it's close enough it doesn't matter? 17:33:23 i think both are fine 17:33:47 I'd say that whatever you are more confident that you can deliver maintainable code in 17:34:13 Okay, thanks for the feedback. 17:34:14 like, maintainable well-factored rust is much better than undocumented spaghetti python, and vice versa :) 17:34:22 :) 17:34:36 ok. next discussion point is nick's :) 17:35:09 i think we got that -- it's just the 044 triage 17:35:17 do folks think that my general plan above made sense? 17:35:37 yes 17:35:43 (general plan is: only call security/correctness/tbb-needs "must".) 17:35:57 anybody want to sit down with me tomorrow and help? 17:36:19 i've found that this is something that goes more than 2x faster with 2 people :) 17:36:59 ^ anybody? :) I can do it nickm but not sure if I will be of help or not. 17:37:17 hm, if it can be done thursday after our other meeting i am up for it, but i would like to have nothing scheduled tomorrow 17:37:18 if there's a better time this week, i could do it then too 17:37:19 perhaps we can all do it in thursday? 17:37:28 during our one hour voice call thing? 17:37:29 ooh, i would enjoy that 17:37:31 oh yes 17:37:34 let's do it there then 17:37:36 +1 17:37:47 great 17:38:02 ok. Anything else? 17:38:13 should we check in on proposal status? 17:38:31 sure. We have 20min left for the hour. 17:38:48 I don't have anything on 316 yet, but I hope I will soon 17:38:59 asn/ahf: are we in a good place with 317? 17:39:22 the discussion has begun and is ongoing 17:39:25 great 17:39:34 i don't think the current design is good and i don't think we can adopt 27k loc's in tor 17:39:39 i think that is the gist of it 17:39:43 yikes, makes sense 17:39:44 im on 316 17:39:49 oh! 17:39:50 dgoulet is also on 317 17:39:53 ye 17:39:55 oops, sorry 17:40:19 asn: sorry for not getting 316 stuff done. would you like to plan a time later this week for us to chat about it, to give me a deadline to come up to speed? 17:40:25 same here 17:40:26 (I'm around during US 8-4 M-F to talk about 316) 17:40:41 i have spent more time than expected reading prop316 17:40:46 but i havent posted a review yet 17:40:52 and i still have some fundamental questions 17:40:56 which led me to start reading the paper... 17:40:59 and here i am today 17:41:08 pastly: working from 8 and backwards in time, very confusing for us in europe, who lack that technology :-P 17:41:24 pastly: perhaps we can do a short call at some point this week to clarify some things? 17:41:39 asn: sure thing 17:41:50 ok sounds good 17:41:54 perhaps even tomorrow or the day after 17:41:57 i will send u on IRC 17:42:07 sounds good 17:42:28 ok 17:42:54 dgoulet and teor wrote good comments on 318 through 320; i should reply 17:44:00 ok. Anything else? 17:44:42 i'd like comments on prop#315 if anybody has some. 17:45:04 and that's it :) 17:45:35 oh, next monday is a us holiday, right? 17:45:43 yes 17:45:52 so as not to disrupt things, let's do the meeting on time, but I'll take the rest of the day off ? 17:46:02 I'm not going to be travelling anywhere :) 17:46:37 :) me neither 17:46:55 I will send the notes from this meeting to tor-project@ 17:47:18 i've been saving some holidays during the carantine that i would like to cash out at some point 17:47:21 might be next week, but i will see 17:47:52 ok 17:48:00 closing the meeting for today then 17:48:04 #endmeeting