17:59:28 #startmeeting weekly network team meeting, 18 December 17:59:28 Meeting started Mon Dec 18 17:59:28 2017 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:59:28 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:59:35 io Saturnalia! 17:59:53 hello hello hello :) 17:59:59 hello 18:00:04 o/ 18:01:08 hi 18:01:46 pad is https://pad.riseup.net/p/udIRDvQyIsi1 18:01:47 hi 18:01:58 hihi! 18:02:24 my keyboard is dying; apologies for missing lettrs 18:02:42 * asn reads reports 18:03:33 o/ 18:04:49 teor: I tentatively suggest tomorrow (US Tuesday) patch party as a time to rebuild fallback dir mirrors. Gotta double check with the missus that my evening is free 18:05:00 looks like all the boldfaced things so far are for specific people. Extra points if you notice the ones for you before we get there. :) 18:05:11 addressed mine 18:05:19 pastly: maybe also say that on the pad, or in email: teor is likely asleep 18:05:48 hi all. I am waiting for the pad to load.. 18:05:56 done 18:06:11 hi mike! 18:06:22 yo mikeperry ! 18:06:55 mikeperry: went over your review today. want to do some additions/fixes of my own. didnt manage to finish all of them in time. tomorrow they will be done!!! 18:07:47 asn: cool. do you want to spend some time talking about choose_good_middle_server()? 18:07:59 mikeperry: yes sure. maybe after meeting? 18:08:03 ok 18:08:15 (or maybe a bit after the meeting but today since i need to relocate) 18:12:39 how are we all doing towards the end of this year ? 18:13:49 a bit worried with the relay events - a lot of relay ops friends have poked me about it today :o 18:14:09 loaded... 18:14:11 yep. that's another fire we didnt need 18:14:25 lot on my stack for the mid-jan freeze!... 18:14:26 also super loaded with reviews and whatnot. 18:14:44 isis: i have a point for your attention, regarding moat 18:16:05 i was not very good with reviews last week while trying to wrap my head around some data. if someone have some stuff they need reviewed urgently that i can help with during this week feel free to poke me. otherwise i'll go through the list of needs_review and take from there 18:17:47 i feel like all i've been doing the past 3 weeks is reviewing prop247/prop224/misc tickets. i don't mind because i really want to see these merged. but i havent had time to do anything else which is kinda sad. 18:18:38 i know the feelin 18:18:42 nickm: yep :) 18:19:13 dgoulet: mid-jan freeze is one of the discussion items fwiw 18:19:14 i actually dont mind reviewing code, i enjoy it. but i'd like to see the review load spread more widely through the team. 18:19:20 +1 18:19:29 i wonder how is the coordination of work/dependencies going with other teams? 18:19:48 from those items that came from the roadmap coordination meeting 18:19:49 nickm: ah good 18:21:08 isabela: i saw #24636 and #24637 friday and planned to start on them today 18:21:18 isabela: thanks! 18:22:34 great - maybe we could do a mid week check in or something, since next week is holidays 18:23:17 isabela: i'm missing one point to the s8 q4 doc, will get to it after the meeting (related to our benchmarks for the monotime work) 18:23:34 i'll highlight you when i've written it if you have any questions there 18:24:42 okay, let's start with the discussion thins 18:24:44 *things 18:24:53 ya 18:25:10 first two are: how do we adjust our schedule based on 0.3.2.x becoming rc on the day we expected it to become stable? 18:25:13 ahf: ack 18:26:12 nickm: how do u suggest? does this mean that stable gets delayed for 2 weeks (one release)? 18:26:29 and we let the rc stabilize during the holiday season (?) 18:27:01 isabela: sure, want to meet thursday morning? 18:27:03 we could do that. what would folks think? 18:27:26 do relay operators historically update to bleeding edge releases during the holidays? 18:27:26 seems reasonable 18:27:37 some do... 18:27:39 not too many 18:27:57 but, um, i'm not going to be putting ou a stable on 1 Jan 18:27:59 might depend on whether their relay operation is work-adjacent for them 18:28:06 maybe we should delay more than 2 weeks. 18:29:19 i thought it was jan 15 18:29:29 oh 18:29:39 that's feature-freeze for 0.3.3 18:30:35 how about we aim for jan 8 for stable (if possible), and delay feature-freeze by 1 week on 0.3.3? 18:30:45 i like it 18:30:49 isis: like pacific time morning? 18:30:51 sounds good 18:31:04 isabela: yeah! or whatever works for you 18:31:06 isis: i have vegas meeting at 8:30am your time 18:31:24 next discussion topic: there's DoS stuff afoot on the network, and emails going around on tor-relays and network-team-secure 18:31:26 maybe 1900 on thursday - i think GeKo can meet that time too 18:31:36 not on thursday 18:31:37 err 2000 utc 18:31:42 ah 18:31:45 let's figure out who/what/when/how on making progress here, given that everybody's busy and the year is ending 18:31:50 GeKo: another day/time? 18:31:50 i'll be traveling the whole day 18:32:19 hm 18:32:22 let me think about it 18:32:38 maybe wed after meeting with asn 18:32:54 I'm quite deep in understanding how this DoS is affecting our relays, I do plan to continue on this but help would be much appreciated 18:33:05 isabela: GeKo: wednesday will also work for me, but tuesday i've promised to attend a crypto standardisation meeting, so if i can't fix everything today, i'm not sure how much i would be able to accomplish at the meeting 18:33:52 i've got no real time to help with the dos :S 18:35:00 i can help think and discuss, and talk about fixes, but I can't do it solo 18:35:09 I'd be very glad to work with you on this, dgoulet 18:35:43 nickm: superb! 18:35:50 I'd be very glad to work with you on this, dgoulet 18:35:53 err whoops 18:35:55 copypaste 18:36:04 let's talk together after? 18:36:22 are there more folks who are working deeply on this? I see teor is... 18:37:07 isis: np, lets try to schedule with mcs and brade instead since GeKo will be traveling - can you do it on wed at like 2000 utc? i will check if they can do at that time 18:37:33 i tried for an hour-ish to determine what might be causing it along with a relay operator who reported it happening to them 18:38:03 i'm happy to help try to reason through it too 18:38:08 isis: awesome 18:38:17 isabela: sure, that works! thanks! 18:38:24 cool 18:38:50 pastly: please wrt fallbacks also reach out to Hello71 18:39:14 if you want to do it you can do it on your own 18:39:14 asn: did you want to ask isis somthing abou the bridge bw scanner? 18:39:43 isis: is the bridge bw scanner something that can be reused for generic bw scanning? 18:39:47 mikeperry: for leak checking I usually just use clang and --enable-fragile-hardening; I find that more useful and fast than valgrind. 18:39:55 isis: and how is that going? 18:40:15 i've got many people talk to me about bw auths lately, and i'd really like to see something happen in this area. so i'm tryng to map the possibilities. 18:40:39 mikeperry: but for valgrind, maybe see the valgrind options in doc/HACKING/HelpfulTools.md ? 18:41:32 i do: valgrind --leak-check=full --show-reachable=yes --log-file=/tmp/val_tor.log ./src/or/tor -f torrc 18:41:37 catalyst: I can hack up the thing you want wrt guards probably; but I'll need more explnation on what it should do & how it should work... can you ping me after the meeting? 18:41:40 dgoulet: if you need help on thinking, I can assist. if it's about "how does tor work", not so much 18:41:54 yes i also dont understand catalyst's guard question 18:42:01 nickm: ah ok thanks 18:42:05 asn: jake the intern has started working on a (minimal) rust controller library and so the finished scanner is probably not too far away 18:42:27 not too far ... as in January? Jan-March? 2018? :) 18:43:09 i would guess jan-march since things are cropping up at their day job 18:43:09 nickm: sure, we can talk after the meeting 18:43:23 isis: but is this something that can replace torflow for bwauth? 18:43:24 great; I'll be on #tor-dev 18:43:32 probably i should carve out some time to hack on the bridge bw scanner with them 18:43:42 isis: same deal with the rust low-level crypto; let's talk strategies on #tor-dev afte this meeting? 18:43:55 nickm: sure 18:44:14 isis: or is it specific for bridges? 18:44:43 asn: it's for bridges, but it should be extendable (if someone wished to do that) for a torflow replacement 18:45:06 hmmm ok 18:45:17 one would need to redo/rethink the functionaility for slowly grouping things in percievedly similar buckets 18:45:19 sounds non-trivial. same as adapting aaron/donncha/david's bwscanner 18:45:45 ack thx 18:46:28 pfew. busy times! 13 minutes left in the meeting. 18:46:42 I hope everybody gets some relaxing time in over december 18:46:50 remember the internet will still be here when you get back 18:46:53 yeah it may be non-trivial… all i wanted was _any_ measurement of bridges 18:47:23 maybe we could do a better-organised/more-collective thing to take on an intern to extend it? 18:48:32 maybe 18:48:41 lets think about it 18:49:18 * asn done here 18:49:38 ok. thanks, everybody! 18:49:44 thanks :) 18:49:46 thanks folks! 18:49:55 I'll see you on and off, and see you again in the new year! 18:49:58 #endmeeting