17:58:15 #startmeeting network team meeting, 25 Feb 17:58:15 Meeting started Mon Feb 25 17:58:15 2019 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:58:15 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:58:18 hi all! 17:58:22 o/ 17:58:23 It's time again 17:58:29 the meeting pad is over here: https://pad.riseup.net/p/tor-netteam-2019.1-keep 17:58:55 o/ 17:59:04 hi asn, catalyst ! 17:59:12 hey o/ 17:59:32 hi! 17:59:52 hi, ahf and dgoulet ! 18:00:09 ji 18:00:09 * juga online this time 18:00:11 hi 18:00:18 hi gaba & juga ! 18:00:24 that lives mikeperry . You around, Mike? :) 18:00:35 hi nickm 18:01:10 first thing to do is look at the kanban 18:01:20 we have 15 in-progress things there. are we really doing all of those? 18:01:46 We should also look at how we are doing with wrapping up our february stuff & take a look at what isn't going to get finished this month out of our planned roadmap 18:01:55 gaba: how shall we do that? 18:02:02 i guess it needs to become 16, because now i need to start working on #29221, even tho #29298 hasnt been merged yet :) 18:02:07 but its really close to merge_ready 18:02:27 yes, I'm looking at the tickets in in-progress but each of you should check your tickets and see if that is what you are working on 18:02:31 i think it will become merge_ready after mikeperry checks it out 18:02:31 if it's in needs_review, it could go into the "in review" column? 18:02:38 yes 18:02:41 ack 18:03:24 My work on #29280 is likely to spill over to next month, but that's as expected. It will probably need help from teor before it's done 18:03:32 (I am here. sorry in wrong window) 18:03:56 my browser does not treat the cards like draggable things 18:04:03 hi, mikeperry ! We're making sure all the kanban stuff is right 18:04:11 asn: try dragging the cross on the upper right corner? 18:04:12 (sorry, I was just an important phone call, it is over, updating the pad in a jiffy!) 18:04:17 nickm: ah got it 18:04:29 boom swapped it 18:04:37 still 15 but more right now :) 18:04:57 * juga not updating kanban, too many tickets to change 18:05:11 kanban rebel 18:05:17 :) 18:05:43 Is everybody happy with the state of the kanban, or are folks still getting into it? 18:05:49 im good 18:06:14 dgoulet: you are working on #9316 ? 18:06:31 catalyst: do you think your initial review of messaging_v2 will finish this week? And should we revise the '8 points remaining' estimate we made at the meeting? 18:07:04 gaba: no, not at the moment nor planned this week 18:07:36 nickm: yes will probably finish this week; the 8 points still seems accurate as a total 18:08:01 great! If I have free time (haha) I'll start working on your requested changs in the background. 18:08:17 dgoulet: let's move it back to the backlog then. you are not going to work on it and somebody else can take it? 18:08:28 Do you think you will be recommending significant architectural changes? If so, I don't want to get too far into it :) 18:08:46 gaba: this week, might not have the time but it still possible so yeah backlog and I'll just pick it up from there if I can 18:08:54 dgoulet: you have 4 tickets in in-progress :) 18:08:56 gaba: what should we do with "in-progress" stuff that we aren't going to do any work on this week? 18:08:57 ok 18:09:08 take it back to backlog 18:09:08 gaba: yes they should be OK and small also 18:09:41 nickm: at this point i think any architectural recommendations i would make would be of the "clean it up after merge if need be" variety 18:09:50 awesome, thanks! 18:10:04 #26288 is almost done, right? 18:10:04 re my kanban cards, I am also working on #29294 also because I wanted to get a sketch ready for dgoulet before he went on leave. I am not sure if I will finish it 18:10:15 nickm: unless i run into serious roadbloacks with a proof of concept use case 18:10:30 mikeperry: it is high prio on my list that ticket 18:10:31 let me know if you do? 18:10:37 err #29494 18:10:47 nickm: ok 18:10:52 gaba: yes it is going in needs_review after the meeting since what I needed was merged upstream on friday :D 18:11:02 catalyst: you have some tickets about bridgedb on March. Be sure that you get any stuff you need from dgoulet this week before he goes on leave. 18:11:09 ok 18:11:12 gaba: ok 18:11:25 anything else on the kanban/roadmap? 18:12:08 let's look at reviews! 18:12:30 I don't have a lot of stuff to review this week; please ask me if you want to give me any of your reviews 18:13:18 ye there were only 5 tickets 18:13:21 so most people took one ticket 18:13:27 some people took zero 18:13:28 great 18:14:01 rotations are mikeperry on bug triage and teor on CI+Coverity 18:15:13 any discussions on reviews & rotations? If not, we can move on to discussion... 18:15:57 only one new issue this week, which is what to do for our merge policy stuff 18:16:13 I got responses from arma1, asn, dgoulet, and teor, which is awesome... 18:16:19 I'm not 100% sure where to go from here though 18:16:23 any thoughts? 18:16:58 there were some disagreements, we need to see what we agree on and how to move forward with any new proposal from there 18:17:02 hm, one of the proposals was very different, no? 18:17:05 I have to say that I *want* 3-people merge policy but with 4 committers ... I do share asn's worry there 18:17:08 or, just different 18:17:10 yeah, me too 18:17:18 the most important part for me is to have a way to know what i need to merge 18:17:44 IRC works out some times, but I dont think it scales well 18:17:46 asn, dgoulet : what would you think about having large or tricky stuff need 3 people, but smaller stuff not? 18:18:04 asn: so, I made myself a query like this: https://trac.torproject.org/projects/tor/query?status=merge_ready&owner=!nickm&component=Core+Tor%2FTor&reviewer=!nickm&group=milestone&col=id&col=summary&col=component&col=status&col=type&col=priority&col=milestone&order=priority 18:18:22 That's what I've been using. Once teor is on backport stuff, I'll restrict it to only include recent branches 18:19:19 asn: yeah I think we can only use Trac and look regurlarly... there aren't magic ways to assign mergers except if we do explicitley in the ticket but that won't scale imo 18:19:24 so that trac query is nice for that, 18:19:31 but you need to look on a "regular" basis :S 18:19:36 we need* 18:20:16 i would prefer if there was a way that merges are assigned to specific people, so that it's not a free-for-all situation with no responsibility 18:20:56 im saying this because personally it's very easy for me to ignore merges and do all the other stuff i need to do, if i know that someone else is gonna do them 18:21:11 so, in some cases I'd really like the next available person to do it 18:21:13 asn: you mean something similar to how reviews are assigned? 18:21:23 ye 18:21:43 I would be fine with assigning the merges I can't do myself, using a tag or something. Would that be ok? 18:21:44 we dont need to solve this right now. i just think this piece of the puzzle is still missing. 18:22:13 asn: would doing this make you okay with moving ahead with this plan on an experimental basis? 18:22:15 nickm: if you want to assign merges to people that would work for me 18:22:27 ok. I'll use dgoulet-merge and asn-merge 18:22:29 yes 18:22:36 I'm *very* skeptical that will work tbh 18:22:38 but ... 18:22:46 I do not have another plan other than looking at trac so eheh 18:22:57 please everybody check these daily? (on working days) 18:23:22 ok 18:23:23 I think that should be what we need to do ^ ... because in the first place someone needs to assign the merger anyway so someone will look at it daily anyway 18:23:31 sorry for putting more work to you nickm 18:23:42 so either we have someone's job to do that (like reviewer assignements) or we spread that "check every day" task to each of us ehhe 18:23:42 no problems; this is still taking work from me 18:24:10 I'm fine trying any of them to try it out and see which one is better suited 18:24:14 I check each day for what I can merge; then I assign what I can't. 18:24:26 that's great 18:24:35 shall we try this all week, and see how it works, and discuss again next week? 18:24:39 yes 18:24:43 :) 18:24:51 ok 18:25:33 ok. I'll try to do an updated version of the policy and send it to network-team. How about feel free to deviate from the process when reasonable, but answer on the thread when you do, so we have data for next monday? 18:25:43 ok ok 18:25:58 yes 18:27:07 ok. Now let's check for boldface on the pad and needs-help entries. There are a few of those? 18:27:32 hm, sounds like the wine trick with #29500 didn't do it 18:27:50 mikeperry: how does this work? it just happens randomly? so if i run it in a loop in powershell i may or may not trigger it? 18:27:57 dgoulet: mikeperry has a question for you 18:28:13 dgoulet, juga: catalyst has questions for you 18:28:26 mikeperry: asn has a question for you 18:28:29 nickm: oh, lemme check 18:28:31 I think that's all the pending 1:1 stuff :) 18:28:42 ahf: yeah, randomly. once it triggers I may need to give you a branch to instrument 18:28:46 * gaba needs to miss the last part of the meeting. I will read the backlog when I'm back. 18:28:56 or maybe before if that's better 18:29:02 mikeperry: let me try to run it. do you know if it happened on 64-bit or 32-bit or have you seen it on both? 18:29:17 ahf it has happened in one build which is linked from the ticket 18:29:21 ahf: I believe the original appveyor log failure was for x64 18:29:24 on _one_ build? 18:29:32 of all our appveyor builds? :o 18:29:33 i answered catalyst question in the pad: yes 18:29:41 ahf: yeah :/ 18:29:45 ok 18:29:55 hmmm, let me try. i have some other windows stuff to fix this week 18:29:55 mikeperry: I'll follow up with you on the ticket 18:30:51 dgoulet: ok. I have some additional ideas for reducing padding but they are not full formed yet. asn should probably hear them too, when they're ready 18:31:03 juga: thanks 18:31:03 "reducing padding"? context? 18:31:18 mikeperry: we can do a meeting to discuss while I'm still here :) 18:31:31 basically we could not add any padding to the queue if it has cells in it, and only update tokens... but once the queue empties, we would have to add padding then, or exampine it again, or something 18:32:04 dgoulet: before then just looking at the patch I have would be helpful, tho 18:33:12 mikeperry: the patch is on the ticket right? 18:33:17 yes 18:33:18 mikeperry: not the one I already looked at ? 18:33:51 yeah new patch on diff ticket (#29494). I pushed it friday afternoon 18:34:13 ok great will do! 18:34:14 asn, dgoulet: I have put asn-merge and dgoulet-merge on some tickets :) 18:34:18 wow 18:34:25 ok will take care of them tomorrow 18:34:36 sounds good 18:35:15 (anything else for this meeting? is everybody gettint their pending questions answered?) 18:37:28 okay, hearing no additional topics.... 18:37:38 thanks, everyone! I'll see you all online! 18:37:39 #endmeeting