16:59:01 #startmeeting network team meeting, 30 April 2018 16:59:01 Meeting started Mon Apr 30 16:59:01 2018 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:59:01 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:59:16 hi everybody! 16:59:25 Our pad is at https://pad.riseup.net/p/z6vkC6ae1OF6 16:59:29 * dmr waves 16:59:33 hi dmr ! 16:59:35 hi 17:00:42 o/ 17:00:44 Announcements should be fast: 17:00:53 - 0.2.5 is no longer supported, starting tomorrow. 17:01:04 Or starting now, if you are in a suitably advanced timezone 17:01:46 - We have 2 weeks till the 0.3.4.x freeze. Let's get our roadmapped features done, to the extent we can! Remember, review takes time, so if something really has to get in, landing the code this week is better than landing it next 17:02:29 This should be top priority for us for now, -- I think isabela says? 17:02:43 also welcome juga some more! 17:03:05 next up -- let's look at the roadmap! 17:03:40 I have a suggestion: If there is something on that roadmap which you will not be able to work on, and it's a feature for 034, please strike-through your name. 17:04:07 nickm: if we can finalize the child tickets of #25375, I think we can call #25500 "Done" for 034 ? 17:04:11 nickm: we've done quite a lot 17:04:26 dgoulet: I think we will have some unfinished parts of #25375 17:04:27 but we'll see 17:04:40 if so, they'll need to wait for 035 17:04:44 we have it also in May for 035 ... so I guess it is ok 17:05:15 I'm not entirely sure of the state of #24986 but I would need to shift there if we want some of it in 034 :S 17:05:16 Can we call #25550 "done enough"? We've landed a lot there, and we have more under review. 17:05:21 catalyst, isis: ^ 17:05:28 nickm: I believe so for 034 17:05:43 nickm: we can get the per-second callback refactor that are simply in 034 and the complex for 035? 17:05:52 we'll see :) 17:06:00 nickm: i think we can put #25550 as "to be continued" unless you want to specifically close it out for 034 17:06:02 I see 3 in needs_review that I'll check today 17:07:09 I'll make it one we continue after 034, but I think we can also call what we've done so far "progress". CI is never really 'done' after all... 17:08:26 #25507 is small-ish, and is not a feature, so we can take it post-May-15th 17:08:53 Does everybody else plan to land everything else listed on the roadmap by then? If not, let's prioritize. 17:09:03 * nickm can't believe he's the only one who's crossed his name off of anything yet... 17:09:28 * ahf looked over and it seems OK 17:09:46 modularization work is needs_review so that is waiting for me and ahf (#25494) ... then #25500 is almost done 17:09:55 then #25552 is still being investigated by asn 17:10:01 so I have left is the wide CREATE 17:10:15 i'm hoping to move to #25499 this week and #25496 is outside of our repo 17:10:18 which I would need to sync up with isis at least 17:10:36 isis: are you also on track with revisions on #24660 ? 17:10:56 ahf: I can help a little with #25499; it should turn out to be VERY short. 17:11:20 I have had to shuffle around the #25546 sub-tickets. but everything I want there is either needs_review or merge_ready 17:11:32 ack 17:11:41 nickm: yeah, you mentioned some things that i noted down a week or two ago it sounds like it 17:11:52 cool 17:12:04 would like to test it on device though to see if it helps us a lot (and how much) for our s8 reporting, but that can be done post-release 17:12:07 so are we all in good shape for the next two weeks? Has everybody got enough to do, but not too much? :) 17:12:10 or post-freeze 17:12:47 Please ping me and isabela if not. 17:13:05 Next up is code reviews -- looks like everybody has got 2 tickets except dgoulet who has 4 17:13:08 I'm definitely can't be primary on #24986 (wide CREATE) considering the rest so this should also rely on isis/catalyst but I can do a bit this week I believe while modularization is getting reviewed 17:13:40 dgoulet, isis: yeah I think all I can do on wide create for 0.3.4 is just review and revise. I hope I can implement more, but it'll be 035 17:13:41 nickm: 3 of them are related to roadmap item, and like I think 2 I already did a first pass 17:13:56 makes sense 17:14:07 nickm: you have _two_ pretty big actually 17:14:22 Also, everybody please try to get reviews done promptly so people can do revisions and we can merge all this great stuff 17:14:28 nickm: #25610 is the dirauth modularization which ahf and I have worked/reviewed so it is ready for you basically for upstream 17:14:33 (if you get stuck, just swap) 17:14:51 dgoulet: Yeah, that's going to be fun. I'm planning on that one for this afternoon 17:14:57 nickm: I recall you prefer them to stay in needs_Review rather than merge_ready 17:15:45 This week's rotations: mikeperry=bug triage, nickm=community advocate, ahf=coverity, asn=CI. 17:16:07 I have a question: Where have recent community advocates been looking to make sure they find stuff to help with? 17:16:28 ack 17:16:59 nickm: not sure what that sentence means? 17:17:24 If you've been a community advocate recently -- what did you generally do? Where did you hang out? I want to make sure I do okay. 17:17:32 ah 17:17:57 I'm thinking of just the mailing lists, the bug tracker, and IRC. Anywhere else? 17:17:58 tbh, on my part, not much more than I try to usually do that is respond to contributors and follow mailing lists, try to answer questions 17:18:20 I looked in #tor and tor-talk a bit. I didn't do a very good job. was pretty busy 17:18:21 i think asn posted some notes two weeks ago about his week of doing community advocate work 17:18:30 nickm: i watched pull requests on github, and monitored #tor and #tor-dev for people who looked like they needed a little help to better contribute 17:19:17 * dmr also remembers asn posting notes; looks for them 17:19:19 also looking for early warning signs of possible bugs 17:20:02 sounds reasonable. 17:20:21 dmr: (cool! please post notes here if you find them.) 17:20:25 onwards! 17:20:59 So, I had a question about picking an LTS: generally we've denoted LTS releases retroactively. I wonder if we should actually plan this time 17:21:09 like, to have 034 or 035 be an LTS in advance 17:21:37 dgoulet: i've got a bit of the wide create stuff done in my bug25647 branch but it's still a WIP 17:21:40 we should probably pick one before the end of this year, so that there's a year's overlap when the new LTS exists and the old LTS is still supported. 17:21:58 ( ugh my internet is behaving atrociously) 17:22:07 Should we aim for 034 or 035? Or ask debian which is better for them? 17:22:34 nickm: imo, if we can get that HS grant "soon-ish" and work on finalizing HSv3 to the level we want it to be for full "v2 replacement candidate", I would pick that version 17:22:42 honestly i won't be able to finish in time for 034 without help 17:22:46 but quite unknown on if 035 could be a reality 17:23:00 Alternatively we can pick a new LTS even later if we stretch out the lifecycle of 0.2.9 even farther 17:23:07 * dmr found a mention of the notes idea by asn, but not the notes yet - still looking 17:23:55 if we can choose LTS designations in a way that make downstreams happier, we should do that 17:24:17 what does "downstream happier" means? 17:24:35 making the downstream organizations that distribute our packages happier 17:24:43 ack 17:25:09 I guess we should survey them. I'll send a note 17:25:15 I'm wonder how much packagers prefer one version to the other though... 17:25:21 (for a future LTS) 17:25:27 for #24660 i just need to re-expose crypto_rand() i think 17:25:35 #action nickm asks packagers about LTS perferences 17:26:18 dgoulet: my experience is that everybody would prefer that whenever they do a release, there is a new Tor LTS. Obviously that's not possible for us, but it's best for them. 17:26:30 downstreams don't tend to like having their own LTS releases package an upstream release that isn't LTS because that makes them have to backport security patches, bugfixes, etc 17:26:53 yeah for sure, next Debian needs to take in our LTS 17:26:57 (for instance) 17:27:47 okay. 17:28:15 next topic is "what's missing for 0.3.3-release"? There's TROVE-2018-005, and there are also some other things listed on the milestone 17:28:57 ahf: I think #25245 is not a release blocker 17:29:04 i agree 17:29:16 So I think we're nearly good to go. Woot. 17:29:22 \o/ 17:29:36 5 merge_ready there, sweet 17:29:43 though maybe we should see if we can think of anything that could have caused #25957. Though really there isn't a lot of stuff to do there. 17:30:45 with that username, I suspect "armv8" :P 17:30:53 ow 17:31:50 hm, rpi3 is v8, no? 17:32:02 i set one up some weeks ago to test something we needed to test on arm 17:32:57 yeah I believe v8 17:33:04 this looks like something where we might get better results with fragile hardening enabled -- but of course that can get expensive. 17:34:19 isis: when you say "i won't be able to finish in time for 034 without help", do you mean the entire wide create cell stuff, or just the parsing thing you've been working on? 17:34:35 the entire thing 17:34:44 And everybody: is there more to talk about this week? Everybody please read everybody else's updates and check! 17:35:01 isis: yeah, I don't think we're getting a full implementation in 034. But 035 isn't so bad. 17:35:03 isis: lets sync up after in #tor-dev? I can find time this week for this task 17:35:37 i might be missing tomorrow, i have a theory test before an actual test and it's labour day here 17:35:50 yah May 1st tomorrow 17:35:55 dgoulet: okay, thanks! 17:36:29 anything else? 17:36:34 * dgoulet is good 17:36:51 isis: please get an update on the pad some time today, so it's there when I send the pad to the ML? 17:37:24 also everybody: In the future please let's remember to have our updates done before the meeting, so we're all ready when the meeting starts. 17:37:27 And that's all from me! 17:37:41 * nickm looks to seee if there are any more topics rising up... 17:38:18 * dmr gives up on finding the notes :( 17:38:23 hearing none, I think we're adjourned! 17:38:26 thanks, everybody! 17:38:31 dmr: thanks for looking anyway! 17:38:34 #endmeeting