16:58:27 #startmeeting network team meeting, 29 Oct 16:58:27 Meeting started Mon Oct 29 16:58:27 2018 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:58:27 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:58:31 hi, all! 16:58:42 https://pad.riseup.net/p/tor-netteam-2018.1-keep is our pad 16:59:07 hello 16:59:08 hello earlier meeting time o/ 16:59:09 hi 16:59:10 o/ 16:59:24 hi 16:59:29 how has last week been? It seems like we're stabilizing 0.3.5 and working on preliminaries for 0.3.6, but not so far into 0.3.6 work yet. Is that about right? 16:59:31 earlier meeting time is good. still blocks my yoga lesson, but at least better wrt dinner. 16:59:46 asn: sorry there :( 17:00:11 #info next week's meeting is at the patch party time 17:00:12 its ok. cant have everything in this life. 17:00:31 also the meeting will move with US DST, I think. 17:00:53 (if the meeting next week is in the middle of the night for you , no need to attend. please sleep!) 17:00:54 will patch party move or stay fixed to UTC? 17:01:02 catalyst: not decided afaik :/ 17:01:13 in the past it stayed fixed 17:01:35 it will move 17 UTC? 17:01:42 what should we do about reviews next week? assign on tuesday? 17:01:47 or assign on monday and ack on tuesday? 17:02:32 "assign mon, ack tue" sounds good 17:02:47 kk 17:02:50 ahf: this meeting is 1700 UTC right now? 17:03:22 yeah, i think so 17:03:24 maybe i'm wrong 17:03:48 date -u says you're right 17:03:49 it's 17:05 UTC 17:04:01 so generally, how are we doing on roadmap stuff? 17:04:05 oh, i thoughjt you asked if the meeting was bound to 17 UTC.. 17:04:41 ahf: no, it isn't. it moves with US DST to stay non-overlapping with the tor browser meeting 17:05:26 ack! 17:05:43 I think we are good roadmap wise about #25502 ... ahf? 17:05:49 yep 17:06:19 did you make a plan there? Is there a rough timeline? 17:07:02 there are tickets of the plan and timeline nope... there is this wildcard right now on "when" the patches will get into the pt-spec but that shouldn't stop us too much to impl. 17:07:22 ahf: speaking of, next step would be for us to talk and see how it goes for post-spec :) 17:07:22 okay -- are the tickets under #25502, or somewhere else? 17:07:40 the tickets are under #25502, i talked with gaba today about that we should give points to the tasks 17:07:43 they are all under #25502 and tagged roadmap-substak 17:07:55 I only see 2 tickets there? 17:08:02 they all have childs :P 17:08:10 ah. transitive closure 17:08:18 :) 17:08:25 yeah, it's a bit of a tree right now 17:08:27 there are no "implementation ticket" yet ... that is to do 17:08:55 ok. asn, you've been deep into wtf-pad with mikeperry ? 17:09:05 (is mikeperry around this time?) 17:09:13 yes :) 17:09:23 nickm: yes exactly 17:09:36 reviewing code and testing it (also encountering annoying chutney issue that i need to figure out) 17:09:47 ok. you asked me if I could take a review pass at some point. When do you think the time for me to do that will be? 17:10:08 nickm: i suggested wednesday, last week 17:10:12 nickm: but you can also start whenever you want 17:10:23 nickm: there is lots of code, so i doubt we are gonna be doing duplicate work 17:10:31 ok. let's plan a time to talk on wednesday to do the handoff ? 17:10:38 ok 17:10:43 #action nickm and asn pick a time on wednesday to talk about wtf-pad review 17:10:44 lmk if I should squash it back down by then 17:10:51 mikeperry: that could be cool 17:11:08 nickm: mikeperry: my plan is to switch mainly to testing the branch on wednesday 17:11:11 so someone can review, and i can test 17:11:15 * so nick can review, and i can test 17:11:34 having a squashed branch would be good for me, if it isn't inconvenient for others 17:11:40 its fine for me 17:11:43 juga: I've seen a bunch of sbws tickets -- any you want comment or response on? 17:11:58 nickm: only the 2 about the spec 17:12:11 ok, which 2? I can try to answer after this meeting 17:12:21 or others who might know 17:12:25 juga: anything we need to look at with core tor stuff this week btw? 17:12:35 nickm: at the bottom of the pad 17:12:46 ahf: not yet, but hope soon :) 17:12:49 cool! 17:12:55 juga: #27690 and #28085 17:13:01 yup 17:13:08 ok, if anybody can answer those that will rock. I'll give it a try 17:13:10 both have reivewers 17:13:18 oh 17:13:45 * juga checking 17:13:57 catalyst: how's work been with you? Anything happening on #28018? 17:13:59 not in trac 17:14:44 ah, no 1 has teor, not sure they've time, the other didn't see, sry 17:14:58 nickm: not yet, but the pubsub stuff will hopefully help with that 17:15:04 juga: no worries, was assigned this morning ;) 17:15:13 dgoulet: :) 17:15:25 as in trying to use the pubsub stuff to help with #28018 will help me review the pubsub stuff :) 17:15:38 catalyst: ok. did you get a timeline or a breakdown? If not let's meet either today or early tomorrow to make one? I'll be much less twitchy here once we have a breakdown for this work 17:16:18 ok let's talk tomorrow? 17:16:25 catalyst: ok, when is good for you? 17:17:17 nickm: 18:00 UTC tomorrow? 17:17:24 gaba: looking at the roadmap, I see that the pubsub stuff is sorta not there, though it falls under a bunch of other items. Should I add it? It helps many things for sponsor 8 and 19, and for future modularization work 17:17:43 catalyst: hm, I'm meeting somebody at 1800. I could do earlier, or after 2000 ? 17:17:59 20:00 is ok 17:18:05 ok, I'll talk to you then! 17:18:11 I think that's the roadmap stuff. 17:18:24 I've taken on #2149 since it's apparently something that sponsor 8 really would like... 17:18:40 i'll try to at least get a design for it this week 17:19:05 ahf: you have a roadmap item about contacting hans abour arm64. Do you think that's something you could do quick so we could check it off? 17:19:19 yeah, gonna do it this week, talked with gaba about it during the 1:1 earlier 17:19:28 great 17:19:34 adding it to my tasks 17:19:57 okay, that's roadmap! 17:20:36 next is review assignments 17:20:40 eerybody please have a look? 17:20:57 *sp 17:21:31 asn, dgoulet: please move the asan/rust items to me? I took them over from teor since they got stuck. There's one change it needs, then we can merge. 17:21:48 all of them? 17:21:50 all 3? 17:21:52 yes 17:21:53 kk 17:22:05 dgoulet: wanna do trac? 17:22:07 i do spreadsheet 17:22:19 also, I thought you were moving items other then #28226 away from catalyst. is that still the plan? If so, we should make sure people get 'em before they're a surprise 17:22:50 nickm: we only moved away the tickets assigned this week 17:23:02 what still remains are tickets from last week 17:23:06 asn: sure 17:23:23 we could move more, if so desired 17:23:52 let's move, unless you've got progress on them, catalyst ? 17:24:09 #28226 is clean-ish, but not small, and we'll be living w it for a long time 17:24:38 nickm: yeah, go ahead and offload me. if other people end up with too much, i can help out later in the week 17:24:47 ack 17:24:51 ok 17:25:04 ok there will need to be some rebalancing 17:25:13 mikeperry: oh, I just remembered: iiuc you still have the lock on the whitepaper. 17:25:47 nickm: ah yeah. I can do that this week 17:26:00 could we maybe get release milestones on the review spreadsheet so we know more easily how to prioritize reviews? 17:26:04 thanks! please ping me if you're stuck, or un-stuck 17:26:45 asn, dgoulet : I think catalyst's question above is for you. It would be fine for me 17:26:57 also anybody should feel free to add more columns to that spreadsheet, I think 17:27:05 hm 17:27:20 we could do it, but it would be overhead 17:27:25 Trac should be what to use here 17:27:28 perhaps people who want that, can do it themselves? 17:27:31 Milesonte + Reviewer field query 17:27:33 maybe make a column and let people fill it in if they like? 17:27:48 we dont really have a good way to treansfer stuff frrom trac to spreadsheet 17:27:52 it's kind of a pita 17:27:55 just that over a week, it can change so Trac is what needs to be queried 17:28:01 maybe we need to make a script or something at some point 17:28:18 right now i have to click through to each ticket to see its milestone 17:28:29 Trac query? 17:28:52 the trac query for catalyst would be ... https://trac.torproject.org/projects/tor/query?status=accepted&status=assigned&status=merge_ready&status=needs_information&status=needs_review&status=needs_revision&status=new&status=reopened&reviewer=~catalyst&group=milestone&col=id&col=summary&col=status&col=type&col=priority&col=milestone&col=component&order=priority 17:28:57 dgoulet: i marked the tickets that need to be moved on the spraesdheet 17:29:00 the spreadsheet in theory exists for you to use it to whatever ends you want but in reality Trac should be used 17:29:01 if you can d oa trac query you can CSV out which you can script from the sheet 17:29:04 the query can also be found in the metadoc 17:29:07 to make the sheet fetch the data automatically 17:29:11 but they are listed as reviewer for a few things that aren't necessarily for current review. not sure if that matches up 17:29:30 nickm: ? 17:29:48 nickm: "current review"? 17:29:51 dgoulet: see that query; there's stuff in unspecified that isn't for review right now afaik 17:30:01 and I don't know 100% for sure if the other tickets are up-to-date or not 17:30:10 well we prioritize review per milestone usually 17:30:33 dgoulet: you already sort by milestone in the spreadsheet? 17:30:47 there is no concept of milestone in the spreadsheet 17:31:48 that is the proper "needs_review" query: https://trac.torproject.org/projects/tor/query?status=needs_review&reviewer=~catalyst&group=milestone&col=id&col=summary&col=status&col=type&col=priority&col=milestone&col=component&order=priority 17:31:56 (catalyst user as the example ^) 17:32:17 catalyst: does that query work for you? 17:32:31 there is no concept of "priorities", the idea is that every week we have a series of ticket and if they aren,t done that week, they carry over the next week and so on ... so at that point, it is the dev. responsability to keep track or prio or re-triage 17:32:52 also, rotations: ahf is ci+coverity, asn is bug triage. works okay? 17:32:56 ok 17:33:02 ack 17:33:38 on to discussions, I think? 17:33:42 nickm: query works well enough :) 17:33:47 great 17:34:19 what do we do with all the snowflake tasks we have on the roadmap in the bottom? gaba and i talked shortly about it in our 1:1 17:34:40 it's almost 1st of november, realistically we might not be able to hire an anti censorship person with start before 1st january then 17:34:41 on discussion: I think we should schedule another meeting for teor's question about 036-roadmap-proposed. It should be some time gaba can make it. Does anybody volunteer to make a doodle pad and send it around? 17:34:46 and our list is big 17:35:29 ahf: good question. I think we need to do a few days where everybody builds snowflake, looks at its design, tries its sourcode, and then we have a discussion meeting for it. I don't know when we'll find time. 17:35:51 I think it needs to be everybody or almost everybody, unless somebody is genuinely enthusiastic about being the new snowflake owners 17:36:03 i'm gonna try to spend a morning or afternoon this week trying to wrap my head around and see where it is, but maybe we should try to figure out who does what 17:36:04 that way we can all get enough background to help each other 17:36:17 my plan was just to see if i can get it to build and run and see how it works 17:36:36 doesn't learning snowflake internals require learning at least two new languages for most of us? 17:36:43 ahf: sounds like a good start. Maybe you can take notes as you go along, and send them to a list, to help the next person? 17:36:47 it's go and c++? 17:36:55 and javascript? 17:36:58 nickm: yeah, that was my plan. a bit like with the windows stuff 17:37:04 ok 17:37:11 ok, i know javascript, c++, and go, and i've worked with webrtc in webkit some years ago 17:37:19 so i could even remember some of the api names when i looked at the code 17:37:44 catalyst: I'm hoping most of us can make do with a superficial understanding here. 17:37:49 so, if i do this, then maybe we could aim for having more people jumping in during medio november? 17:37:52 in like 2-3 weeks? 17:38:30 hm 17:38:51 or in 1-2 week? :-) 17:38:54 maybe? I think we should plan a meeting, and try to get a minimum number of folks to look at it? I should be one of them, and ahf, you seem like you've got a good start 17:39:39 Anybody else willing to try to build snowflake, skim some of the code, and take a look at the design? I can do some good here, but my hours for hacking are finite :) 17:39:48 ack, if i do some kind of documentation around getting it to work this week, then we do a meeting next week based on that? 17:39:49 we could aim for 12 Nov 17:39:58 where gaba is around too since it is very roadmap related 17:40:05 maybe late next week, to give people time to try it? 17:40:09 yes 17:40:10 good idea 17:40:16 also link to any docs you can find about how snowflake actually is supposed to work 17:40:31 #item we need to schedule a meeting to talk about 036-roadmap-proposed 17:40:36 yes 17:40:40 #item we need to schedule a meeting for late next week to talk about snowflake 17:41:16 catalyst: looks like your discussion topic on meeting times got answered. I volunteer to send the note to tor-dev 17:41:25 #action nickm sends note about meeting times to tor-devQ 17:41:31 *tor-dev@ 17:41:48 I have a few quick discussion items. 17:41:50 nickm: ok, if nobody objects to my summary 17:42:34 one is I need somebody to talk about #27359 with. It affects how we store families in RAM. It's a neat idea to save some RAM on clients, but I ran into a hitch. 17:42:42 anybody interested there? 17:44:28 another is that I really want to get #25502 finished and merged. but I need somebody to make the call about whether to use my branch, cpunks's, or some combo of the 2. 17:44:36 third one, dgoulet answered. 17:44:44 #25502 sounds wrong there? 17:44:53 oops sorry 17:44:55 it is the mmap() one? 17:44:58 ahf: yes 17:45:05 #27244 17:45:08 sorry, too many tabs 17:45:12 i liked your most, but i don't think i saw an answer to my question to cpunks 17:45:15 let me look again 17:45:50 there is a reply! 17:46:21 so, you updated the rest of the functions that was missing to take the length parameter? 17:46:25 yup 17:46:27 in your branch? 17:46:30 yup 17:46:38 ok, i remember your branch was being less chaotic than the other branch 17:47:00 so maybe if we decide to go with that one, look over if anything is missing from cyberpunks branch and then if not go with yours? 17:47:11 if i look at that before you wake up tomorrow that should be OK? 17:47:21 it would be nice to get this in and mark it as done for s8 17:47:33 sure; thanks! 17:47:38 cool 17:48:16 anybody else have discussion questions for today? 17:49:32 ok. thanks for the meeting! I'll see you all online! 17:49:34 #endmeeting