16:59:56 #startmeeting network team meeting, 15 april 2019 16:59:56 Meeting started Mon Apr 15 16:59:56 2019 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:59:56 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:00:02 Hello everybody and happy monday! 17:00:04 o/ 17:00:05 Who is here today? 17:00:10 The pad is https://pad.riseup.net/p/tor-netteam-2019.1-keep 17:00:15 hi 17:00:21 o/ 17:00:24 hello 17:01:33 o/ 17:02:33 Looking at 040-must and ci-blockers, we're in an improved state from last week 17:02:51 ahf: you have #29930 and #29645 -- any progress on those? 17:03:13 teor4 has asked if somebody else can investigate #29437; I can poke my nose into that one, but not this week 17:03:48 i wrote notes on both: i cannot really figure out what is causing #29930 since the file doesn't seem open to me at the point where the access denied is given 17:03:58 hm 17:03:59 but, i have managed to reproduce #29645 *once* :-/ 17:04:12 wow 17:04:47 (locally, that is, not on appveyor - there it happens more often) 17:05:34 need any help with those? 17:05:42 i could need some help with #29930 i think 17:06:11 I can have a look but not this week. can anybody else look? The leading theory is that we are unlinking a file before closing it, which windows doesn't like 17:06:58 i have been looking at some related github issues and one stackoverflow post and the stackoverflow post suggested to use system("some-command ...") to delete a file, but the symptom seemed related 17:07:23 It doesn't actually hurt to have that message around, but I would rather fix the issue 17:07:39 yes, it's not a terrible bug 17:07:52 nickm: maybe if i don't make progress on it this week we look at it together next week? 17:08:01 this is a short week here because of easter 17:08:24 if it's rare and harmless pehraps we can remove from 040-must? 17:08:34 also I'm taking of Weds through Mon, so I am not likely to be super available 17:08:35 it's harmless, it's not rare 17:08:50 like, it happens every time you start tor.exe 17:08:54 on first start 17:09:48 next up is our roadmap 17:10:41 looking at the kanban, filtered for me, I don't think I'm going to make much progress on roadmapped stuff in my 2-day week; I think I'm mostly going to be reviewing and revising stuff 17:10:52 anybody else have stuff to revise on the kanban? 17:11:03 i think i might 17:11:45 catalyst: fwiw, I am blocked from making further progress on control.c refactoring stuff till your branch is done and merged, since I think my next step will be a refactoring on top of yours. 17:12:06 nickm: ok we should break the deadlock somehow :) 17:12:15 catalyst: oh, are you blocked on me? 17:12:51 hmm the s27 part of the kanban needs some more love before i can accurately update kanban 17:12:52 yeah my stuff is almost done, at least this increment of it. we can talk later about details 17:13:53 catalyst: ok, let's talk about that after the meeting. I'm glad to take an hour or three to review your stuff if it unblocks you, especially if you can get my #30091 reviewed this week 17:14:05 that way we can both merge this week or next and do the next refactoring stuff 17:14:39 asn: is this something to do after you & gaba & dgoulet are together to talk about S29? 17:14:48 everybody else: happy w the kanban? 17:14:55 (I might have kanban changes if #29034 passes review; generally I will be working on child tickets of #28634) 17:15:12 nickm: yes i think the s27 situation will improve this week after dgoulet is back 17:15:27 [sorry, s27] 17:15:59 this takes us logically to the needs-review area... 17:16:14 so there were lots of reviews this week 17:16:26 and i made the mistake of assigning normal amount of reviews to nick even tho he is gonna be missing half of the week... 17:16:33 naw, that's cool 17:16:42 basically everyone got about 3.5 reviews 17:16:55 i gave less to catalyst because #30091 looked biiiig 17:17:02 asn: thanks! 17:17:16 nickm: ... i will try to rearrange your reviews some more 17:17:25 no, don't worry about it. these look small 17:17:25 i think it's fair for people to get less reviews when they rae vacationing 17:17:37 I'm not vacationing till Wednesday 17:17:46 nickm: ok please let me know if you want changes. 17:17:51 ok 17:18:10 and things will improve next week when dgoulet is back 17:18:17 FWIW, my top priority before I go on vacation is to make sure I minimize how much is blocking on me... 17:18:24 and reviewing code counts as one way to do that. 17:18:41 asn: could you reassign anything that I haven't reviewed by Wednesday, if I forget to ask? 17:18:47 yes 17:18:50 thanks! 17:18:59 How is everybody else feeling about the review load? 17:19:17 If there's a small ticket from me that you review today, I will have a chance to fix it or merge it; otherwise, maybe not. 17:19:32 looks OK to me 17:20:43 next is rotations. This week the rotations are mikeperry on bug triage and teor on CI. I did both last week; I'd like to discuss status a bit during the discussion section, since I put notes there 17:21:30 I'll try to get in touch with teor and coordinate with them about CI status before I go. I'll ask mikeperry and teor to be ready to "hand over" those rotations to me and ahf respectively next week 17:21:47 on to discussions 17:22:07 gaba: you have the first serveral topics, I think. You want to go first? 17:22:19 yes, one is about meetings in april/may 17:22:28 teor is on leave at the end of april 17:22:29 do we do meeting on monday next week, nickm? 17:22:34 so they were asking if we can change 17:22:47 ahf: that's the topic of the present discussion :) 17:23:04 ah! :-) 17:23:12 the proposal is to change tuesday 23 april to 23 UTC and May 7th to May 6th 1700 utc 17:23:50 gaba: +1 on swapping meeting times 17:23:51 everybody ok with that? 17:23:52 ok 17:23:54 sounds good 17:24:00 mikeperry: good with you? 17:24:07 yah 17:24:18 sounds good 17:24:23 great. We should also make sure dgoulet can do it, but let's call it a plan 17:25:09 the one about sponsors does not need also a discussion. It is just for people to be aware that this is happening 17:26:11 is the sponsor pad something we should add stuff into? 17:26:18 is there anything people can/should do to help if they have minimal time? to prioritize, for instance? 17:26:24 not really. It is a summary of conversations from before 17:26:33 asn ^ 17:26:47 because it seems a bit random. e.g. the only mention of hses is the one about mobile dormant support, wheras there are mor eprojects we can think of 17:27:12 this are possible projects for 2020 17:27:29 asn: do you think that there are other projects more important? 17:27:47 i think so yes, but i dont have examples right now 17:28:04 ok, there is no rush. we can add them later. I will make a note 17:28:13 im just talking about HSes btw 17:28:22 i think the pad is good in general 17:28:47 i just didnt know if the pad means "this is all the stuff we care about and will seek funding for. we dont care about anything else." 17:29:40 gaba: ack 17:30:27 yah, I would add stuff to it and alter scope of a couple of things there. (vanguards work comes to mind, network health/testing network stuff, dedicated CI engineer, some other research transition topics) 17:31:47 I guess some/much of that stuff is there, but under specific tasks perhaps 17:31:52 +1 17:32:13 ok, if you want to add that comment somewhere that would be fine 17:32:44 remember that this projects will evolve and change once we identify a grant oportunity for them 17:33:01 and you all will be able to shape them 17:33:19 ack 17:33:55 next is gsod? 17:34:14 yes, pili is organizing it 17:34:17 we need mentors 17:34:25 this is for documentation 17:34:26 when does it run? 17:34:31 so we will have technical writers 17:34:39 percentage of all machines 17:34:40 globally. so we can experiment and measure any performance impact (which I believe will be negligible/zero, but we 17:34:43 should verify my belief) 17:34:46 wtf ugh sorry 17:34:56 glad it wasn't private :) 17:34:59 my mouse is being cut+paste happy 17:35:02 lol yah 17:35:37 after july nickm 17:35:42 starting september actually 17:35:47 until when? 17:36:04 3 months, until end of november 17:36:36 I could volunteer for anything touching the Tor manual, or programmer documentation, or protocol docs. 17:36:54 yes, that sounds good 17:36:55 anything less technical, I should probably avoid :) 17:36:58 pili: ^ 17:37:04 anybody else? 17:37:33 i can help answer questions but i'm not sure i have time to be the primary mentor 17:37:48 sounds good, I think 2 mentors from network team would be ok 17:38:07 (who else besides me?) 17:38:29 i'd like to, but i'll be gone most of august, so it might not be the best idea 17:38:30 catalyst 17:38:55 catalyst said they weren't able to be primary mentor. so that would be 1 mentor, one backup mentor? 17:39:05 and only for internals-stuff and tor-manual stuff 17:39:22 mmm, good question. I'm not sure. I will check with pili or i will go read what it says about it 17:40:40 documentation, something that everybody loves :) 17:41:11 we all would love it if we could improve it? :) 17:41:47 yes, i think there were alredy some technical writers interested :) 17:41:58 about next topic, roadmap 17:42:21 it is not necessary for people to participate. Nick and me are going to look at the roadmap and check if anything needs to be included/updated etc 17:42:46 on the other hand, the more people who help us, the better it's likely to go. 17:42:53 yep 17:43:41 that is it 17:43:46 ok. 17:44:01 teor reminds us that they can't do a bug triage on the 29 april week, since they'll be away. 17:44:06 then it's on to me, I'm afraid :) 17:44:15 I'd like to start with passing on the status for the rotations 17:44:45 We should do this for a while, I think, and then think about how rotations can work better 17:45:00 For bug triage, I put more details on the pad. But for me the big takeaways were two: 17:45:06 * gaba loves so much detail you added to the rotation updates 17:45:45 on the two areas where I could look at the stats, I found a lot of backlog. This is not the fault of anybody in particular, since everybody has had a triage rotation between when this backlog began and now. 17:46:03 (tickets with no component, Tor tickets with no milestone) 17:46:54 this makes me concerned about any area where we are depending on keywords and tags, since for those, we have no way to check "which tickets haven't been considered for our usual tags" 17:47:27 I've tried to pass a fairly clean slate to the next person here, at the expense about possibly making some mistakes in the milestones 17:47:43 any discussions on this before ci/coverity? 17:48:28 for CI -- i've tried to get as much passing as I can, and to summarize the current jenkins failures. 17:48:46 I didn't open tickets for any intermittent failures, though I should have. 17:49:03 I've listed the current failures in the update; IMO this should be general practice 17:49:35 last is coverity -- right now we have 42 unresolved warnings on coverity. That's way higher than we want. 17:50:02 Again, this is not anybody's faul in particular, since these have been around long enough that we have all been kicking them down the road 17:50:05 so, yeah 17:50:27 let's try to get those cleaned up. I'm going to take on the current batch, and ask teor (when we talk) to have their coverity rotation only apply to new ones that show up this week 17:50:32 does that all sound reasonable? 17:51:29 yes 17:51:38 mikeperry: as you're doing bug triage, please try to keep notes like these, so we can see where we are at next week? 17:51:39 i think we should spend a little more time converging on expectations for these rotaitons, and then documenting them 17:51:46 (not at this meeting) 17:51:50 I'll ask teor the same wrt ci/coverity 17:53:01 catalyst: I agree, but I also think that the documentation we have on https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/TeamRotations does describe minimal stuff that we have all not been doing 17:53:01 ok. I think this rotation idea is pretty darn inefficient and all of this documentation during role pasing is a symptom of that... but I can do my best 17:53:02 catalyst: +1 17:53:32 mikeperry: I don't like the old version where I do all of this stuff either :) 17:53:45 let's try to come up with something that does work 17:54:40 right.. there are perhaps better ways to delegate.. 17:55:39 agreed! 17:55:51 yes, if this is a priority let's discuss it in july when we are face to face and start thinking about new ways of working 17:56:08 ok 17:56:12 last thing from me -- I did #29223 last week. teor is the reviewer, but this will touch everybody's work in the future, so we should figure out collectively what to do with this information 17:56:26 especially if you have firm opinions on naming things in tor 17:56:31 cool 17:58:36 and that's all from me! 17:58:50 We're about of time, so I'll thank everybody 17:58:55 (thanks everybody!) 17:59:04 o/ 17:59:07 and let's move followup discussion to #tor-dev 17:59:10 cheers, all! 17:59:13 \o 17:59:13 thanks 17:59:17 #endmeeting