17:58:25 #startmeeting network team meeting, 19 Feb 17:58:25 Meeting started Tue Feb 19 17:58:25 2019 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:58:25 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:58:27 Hi all! 17:58:43 hey o/ 17:59:14 o/ 17:59:15 https://pad.riseup.net/p/tor-netteam-2019.1-keep is our pad 17:59:40 Let's start with our lovely kanban roadmap, linked from that pad! 18:00:29 Does everybody's major tasks for this week match what is in the first column there? 18:01:05 o/ 18:01:09 Please check, and clarify, and let's talk about what to do if it isn't 18:02:01 (For me, it's fine. I'll be working on #28226 and #29280 based on discussions with catalyst and teor, but I will grab later stuff and do admin tasks if there isn't so much to do there) 18:02:11 gaba: 9316 is not on my radar for this week, more urgent stuff before but I can't move it :S 18:02:35 cohosh and i will begin #29207 and #29206 this week 18:02:53 we started a bit last week already 18:02:57 ahf: cool; please put them in the "in progress" column on the appropriate kanban 18:03:05 dgoulet: you mean you can't move it on the kanban? 18:03:18 gaba: (where would it go in that case?) 18:03:21 yeep, waiting for login email :-) 18:03:31 nickm: yes I can't ... it is still loading here half the page and I can't move any squares (and I'm logged in!) 18:03:39 just a sec 18:03:52 dgoulet: i will move that one 18:03:56 gaba: thanks! 18:04:04 ahh, because loading 18:04:30 gaba: it doesn't stop loading ... so that is my reality here lol 18:04:49 catalyst, asn (if here), mikeperry (if here): is the kanban correct for you? 18:04:57 nickm: looks like it 18:05:00 great 18:05:11 ok, let me write it down and i move it in a moment 18:05:21 ... i should just drag to move it? 18:05:27 yah it was at the end of last week; waiting for it to load still 18:05:28 yes 18:05:31 from the top right 18:05:33 ahf: you drag the little plus sign in the upper right corner 18:06:31 hmm 18:07:27 i don't seem to be allowed to do that or something. i can lift the card up and move it around but nothing makes it stay on the "in progress" column 18:07:39 ahf: which one you want to move? 18:07:50 29206 and 29207 to in progress 18:08:07 i wonder if it's a permissions thing on that Wekan 18:08:21 how are you doing with #28848 ahf? 18:08:27 mm, let me check catalyst 18:08:38 gaba: good, awaiting some feedback on it and then i think i'm gonna close it 18:08:43 next topic is reviews; okay to move on? Please see the query on the pad under "check reviewer assignments" 18:09:07 mm, if you are login then you should be able to edit, there is no way to change permissions per person 18:09:12 only for the public links 18:09:47 ok ahf 18:10:25 i had very few reviews this week and could take one more if somebody else is stressed with things 18:11:40 I could also do an extra review if needed 18:12:08 everybody okay w reviews for this week? 18:13:24 * catalyst will look and offload if need be 18:13:37 great 18:14:19 Rotations this week are dgoulet on triage, nickm on CI 18:14:33 ack 18:15:12 announcements: the onion services proposal was accepted (woo); it will start in april and run 1 year. We probably will only do minimal work on it until more S19 stuff is done, though. 18:15:37 announcements: the master branch is now 0.4.1.x; 0.4.0 patches should be based on maint-0.4.0 18:16:03 questions/comments/etc? 18:17:03 on to discussions -- would anybody like to lead a short chat on release planning for 040 and 041 and beyond? 18:17:08 wrt Teor's questions? 18:17:23 (I could, but I talk too much) 18:17:36 you mean find a meeting time that teor can be in and facilitate that discussion? 18:18:25 * catalyst isn't sure what teor's question means 18:18:28 that, and let's maybe see if those of us here have consensus that the goal (having milestones that we actually accomplish) is worthwhile 18:18:52 I think their goal is to not end each release with a bunch of leftover junk in the milestone, 18:19:10 and to have a process where we don't put stuff into milestones haphazardly, but we do put it in when it's important 18:19:12 so "put fewer things in the release milestone"? 18:19:18 We are using the milestones only for the releases, right? 18:19:21 +1 catalyst 18:19:27 yes 18:19:41 I think so 18:21:00 I should get an estimation on how much we have in each release to see how we are doing on that. IT would be good to find a system to do that automatically with the capacity we have 18:21:40 I liked the theory of marking tickets as "proposed" and tracking the total work in a release... but it seemed not to go so great when we tried it 18:22:22 Should we try to schedule a meeting for some time teor4 can make it, and talk more? 18:22:32 yes, it sounds good about the meeting 18:22:35 doubt it will work also, I think we need to prioritize instead of throwing out things. We'll always end up with too many things imo and if we put them into the Unspecified void, we loose them so food for tought on my side 18:22:37 Do we have rough consensus that it would be good to try to solve this? 18:22:47 yes 18:22:50 yes 18:23:15 Does anybody have an idea for easy changes we could make in the next month ? 18:23:38 (It would be cool to bring such ideas into the meeting.) 18:23:47 also, any volunteers to schedule the meeting? 18:24:18 I can schedule it.. I will send a doodle this time 18:24:29 thanks, gaba ! 18:24:45 Let's also think about teor4's questions and try to add good suggestions to the pad 18:25:03 (it would be nice to make sure that we consider their things in weekly meetings even when they're not here) 18:25:22 My next discussion topic is that we don't actually have a process for making things policies 18:25:37 and without that, we can't easily make a process for making things policies :) 18:26:02 so I'm wondering how people feel about the meta-process that I suggested in my network-team email about merging, which I copied (slightly cleaned) onto the pad 18:26:38 * gaba still didn't have time to read it but thanks for sending that mail! 18:28:03 * catalyst values individuals and interactions more than processes 18:30:02 catalyst: what it means? 18:30:19 i mean: "what you mean?" 18:30:26 :) 18:31:01 processes are good. we should continually evaluate whether a given process still works for us, based on the current individuals and interactions 18:31:35 ah, yes 18:32:14 a process could help us do that :) 18:32:41 like, writing on each policy we have, "we should make sure that we re-evaluate this process by x/y/20zz" 18:33:40 yes 18:35:00 Two failure modes I think we have now are that our policies stay in draft for too long; and that when we decide to drop a policy we sometimes don't document the fact. 18:35:06 There are probably other failure modes 18:35:48 yes, I think it would be great to document when it is not working and changes 18:36:44 any other thoughts here? 18:37:59 ok, moving to discussion -- teor4 has pointed out some new unit test errors. I'm on one of them; can anybody else have a look at the others? 18:38:18 asn, mikeperry: Is #29527 the one I just added a commit to disable the test for? 18:39:02 hrmm no :/ 18:39:13 ok 18:39:28 can I list you for it? 18:40:36 yah. I might have to chase down Rhistradh though 18:40:53 ok 18:41:32 Any thoughts for #29530? I think that the best we can do might be trying to bisect it on a macos laptop and see when it started? 18:41:57 ah, the problem is that they're error messages 18:42:10 yeah, I can fix that; I did the errors-are-failures fix 18:43:27 next things -- mikeperry, we're trying to schedule a meeting that you should be at about a researcher who is doing packet stuff. Could you answer the thread please? Subject line is "Introduction, Tor Performance" 18:43:58 ahf: mikeperry has a question about testing with wine; is that something you can help with? 18:44:17 looking 18:44:33 nickm: ..packet stuff? 18:44:38 datagrams 18:44:50 ah ha. so quic et al. great. 18:44:55 mikeperry: my repo might be the easiest way to check at least. if you can reproduce it with the repo then usually the fix works on real windows too 18:45:05 mikeperry: if you cannot reproduce it then it becomes a bit more problematic :-/ 18:45:37 mikeperry: the script should work though and i run CI on it every night with tor master to make sure the makefile still works 18:45:43 ahf: does it use the same compilers as appveyor uses? 18:45:45 Last question is mine -- can somebody help me identify all must-fix-in-040 issues, and triage the rest? 18:46:08 mikeperry: it uses mingw32 or mingw64 which is what appveyor uses 18:46:12 i believe 18:46:12 This would be getting together for an hour or so on IRC some time. I am bad at doing this on my own 18:46:38 yeah, see: https://github.com/torproject/tor/blob/master/.appveyor.yml#L18-L26 18:47:42 nickm: sure 18:48:06 dgoulet: ok. let's talk after this meeting and pick a time? 18:48:13 ok 18:48:32 any other discussion topics? 18:48:50 mikeperry: have you found that thread, and can you reply to the proposed times for this week soon? 18:50:37 nickm: yeah I see it now 18:50:54 ok great 18:51:06 little frustrated that they are again trying th QUIC-between-relays thing. I was pretty explicit about that being done to death and not what we need 18:51:16 then you should say so :) 18:51:44 that made me drop the thread.. but yeah I should mail them with a link to the mailinglist posts again and ask them to please read up before we meet 18:52:28 ok; better pick a time later in the week if you can, or suggest times next week, so that this can move forward 18:52:52 Hearing no other topics, I'll declare the meeting over? 18:53:03 ok 18:53:50 ok! 18:53:52 #endmeeting