22:59:26 #startmeeting Network team meeting, 6 november 2019 22:59:26 Meeting started Wed Nov 6 22:59:26 2019 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:59:26 Useful Commands: #action #agreed #help #info #idea #link #topic. 22:59:32 just changed to 7 november here :-) 22:59:35 hello network-team! 22:59:52 our pad is at https://pad.riseup.net/p/tor-netteam-2019.1-keep 22:59:59 who is here today? 23:00:08 * catalyst is here 23:00:52 hihi 23:01:03 o/ nickm & catalyst 23:01:45 we have a big list of things today, so let us get started and see if more people will join us 23:01:56 let's have a look at 042 status 23:02:08 https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases/042Status 23:02:39 things seems OK to me there? no new very high tickets? 23:02:50 I'm here also 23:02:52 * teor is here 23:02:59 o/ mikeperry & teor 23:02:59 Hi mike! Hi teor! 23:03:16 nickm: anything you want to add to the 0.4.2 status meeting item? 23:04:25 hi! 23:04:26 so, I've been doing a release every 2 or 3 weeks. I think we should call the next one "rc" and do it next week 23:04:29 anybody disagree 23:04:30 ? 23:04:33 welcome, gaba! 23:04:37 sounds fine with me 23:04:38 gaba: o/ 23:04:53 #32103 has been abandoned by the original contributor, should we put it in 0.4.3 instead? 23:05:17 Seems plausible. 23:06:30 ok with me 23:06:40 let's do roadmap shall we: https://dip.torproject.org/torproject/core/tor/-/boards 23:06:59 make sure that everything in there is correct and reflects on what you are doing right now 23:07:00 teor: are you able to edit that board now? 23:08:25 Yes, I have been editing the board every week for the last few weeks. 23:08:32 great 23:09:09 is everybody OK with the state of the board right now? 23:09:22 ok w me 23:09:37 are people OK with the reviews they have been assigned this week? 23:09:44 anybody want to pass anything around to anybody else? 23:10:29 I think on 1 Dec we'll have some stuff to move from "backlog" to "maybe someday" 23:10:55 yes, we need to do a whole lot of roadmappy things around 1st of december i think 23:11:19 yes 23:11:34 ok, we have 3 announcements this week: remember to submit hours via harvest, we have team retrospective next week the 12th(!), and remember to migrate things over to our new fancy nextcloud install 23:11:52 ahf: thing about hours. 23:12:09 Every time in the past I had a backlog of hours, I needed to talk to whoever was behind offline. 23:12:10 and maybe spend some time getting to know nextcloud is a good idea. i have been positively surprised about it in the texting period and it will let us do calendars, document sharing, etc. like we sometimes do with google docs 23:12:31 (so it would be a good idea to reach out to whomever you need hours from) 23:12:38 i created a group for network team and added you all there 23:12:41 nickm: for harvest? 23:12:44 yes 23:12:47 there is also a calendar shared with the group (nextcloud) 23:12:58 i don't do harvest yet, i think gaba and i need to talk about that with that i should maybe takeover that 23:13:00 i can reach out to who is missing hours 23:13:02 gaba: wrt the calendar, so far I have gotten two not-so-useful announcements 23:13:03 i added that 23:13:08 one told me about a meeting for 1 October 23:13:12 mmm 23:13:20 one told me about a meeting, in pacific time 23:13:33 nickm: isn't that just the email that shows Los Angeles time? 23:13:34 I will change the calendar to UTC 23:13:35 if I adjust my settings, will I learn only about meetings in the future, wrt my time zone? 23:13:45 the attached calendar thingy does the right thing 23:13:48 I thought it was in utcl but maybe not 23:15:34 yeah, probably some stuff could use some tuning 23:15:52 ok, shall we start with the items: 23:15:55 * anybody wants to volunteer for git admin from the network team? 23:16:04 gaba: this is the one from the vegas team discussion, right? 23:16:22 yes 23:17:00 ahf: this being for our self-hosted gitlab? 23:17:03 okay, so each team needs to find a volunteer who will help maintaining the gitolite (and the tickets that gets created) for git.torproject.org - that is create new repos via the config file, modify permissions, etc. 23:17:16 catalyst: it is for gitosis/gitolite or whatever it is it is called 23:17:21 gitoline 23:17:24 *gitolite 23:17:57 but, um, what's the scope of responsibility? Is this for network-team folks? for our repositories? for ...? 23:18:27 this is to get more people to do that as it is lacking help with gitolite 23:18:29 i think the scope is for the entire org, but the idea was to find a few people from different teams who are up for running it, as it is currently a bit unmaintained 23:18:38 i think isa had a mail about it recently 23:19:06 so technically I'm on the maintainer list but in practice I don't have time 23:19:15 the file is easy to edit, and the admin tasks are simple 23:19:22 if nobody is up for it, i will take it, with the aimed goal of course of getting everybody to become so happy that they move to gitlab instead 23:19:23 but if we get no takers, we get no takers, and that's okay too 23:19:54 next item: 23:19:55 * C style poll: results sent to the network team list. What next? 23:19:56 * Concrete proposal, BSD Kernel style with tweaks? 23:19:59 i think this is by nickm 23:20:18 gaba: who do i need to tell that i am the volunteer for this task? 23:20:25 can't remember. is that yours, catalyst? 23:21:03 ahf: i think roger brought hthis up. irl is also volunteering 23:21:24 nickm: hm, i had made the suggestion but nobody responded. maybe i can try asking more tersely? 23:21:51 Maybe suggest a default and see if anybody wants to amend? 23:22:05 gaba: ack 23:23:38 what is the goal with the item? we discuss the choice here or? 23:23:47 i don't remember these styles by name :-/ 23:24:05 BSD kernel is basically what we do now, but with 8-space tabs 23:24:07 only the funky type notation for k&r 23:24:14 i can link to some example definitions (the three main BSDs) 23:24:24 yeah, i remember liking the one from freebsd's style guide, minus the tabs 23:24:51 functionally we're pretty close to this already, and it'll help choose between alternatives that we didn't have strong consensus on 23:25:48 we should probably do this over email because EU people are probably sleepy or absent now 23:25:57 so in practice, what I think we're going to wind up doing is "whatever tool X supports, plus some tweaks". astyle and scan-format have bsd-knf defaults IIRC. Uncrustify is a maze of twisty options, all alike. 23:26:06 s/defaults/presets/ 23:26:11 what is the action item here? start a thread on network-team@ and then see what people think? 23:27:52 let's continue on email :-) 23:27:59 i'll try starting a thread. hopefully people will respond this time 23:28:07 next item is: * Roadmapping in December - (mostly s27, s28, s30) : gaba will send pad later in november to organize it. 23:28:22 this is only to say that we should do that. I will organize it and send a mail about it 23:28:39 catalyst: i think people will respond if there is somehting they are against here. i'm not so worried about these changes now after i saw how alligned i was with other people in nickm's survey 23:28:45 gaba: awesome, thanks! 23:28:56 next item is: 23:28:57 What should tor's configure do? 23:28:58 How should we prioritise existing downstreams, and a clean build interface? 23:28:58 * https://trac.torproject.org/projects/tor/ticket/32191#comment:27 23:29:09 ahh, that ticket 23:29:30 it is of course confusing that we do lib handling different for two of the compression libs and not for everything else 23:29:44 that seems to me like an important topic. also maybe one that we probably won't cover well in the available time here? 23:29:45 I think that thing requires diplomacy and thoughtfulness. There is a lot of work there, and there is a little work to do to get stuff mergeable. 23:30:28 i think we had this discussion around --enable-android at some point too 23:32:04 i do not know what a good solution is here. i always feel like touching these variables in configure.ac will probably break some other use-case, like people cross compiling tor on openwrt or whatever 23:32:32 I think we're not far away from getting this working; we just need to do more work than anybody actually wants to do here. 23:32:47 and half measures have the problem of breaking good use cases 23:33:08 yeah 23:33:33 we could decide that we err on the side of explicit failure when we can't do something that the user specifically requested (--enable-* on the ./configure command line) 23:33:38 We're talking about a whole set of upstreaming tickets here, we merged #32191 after I made the changes I requested. 23:33:53 teor: ah 23:34:50 The tickets contained a whole lot of code, all at once, and so it was hard to set expectations. 23:36:41 yeah 23:37:17 i don't think we will reach anything from discussing it here? i think what you did teor was very nice and got that extracted and merged 23:37:36 next item is: 23:37:38 * It's a public US holiday on Monday and Alex is traveling in the evening. Should we move the meeting to Wednesday next week at the same time as the meeting today? 23:37:58 anybody objecting to moving the meeting next week to wednesday at the same time? 23:38:05 it does mean having two meetings on the 12th i think 23:38:10 * gaba is ok with it 23:38:13 including one retrospective and one team meeting 23:38:27 oh 23:38:42 When you say "same time" do you mean 1700 UTC or 2300 UTC? 23:38:57 err, sorry, same day 23:39:14 i was thinking having team meeting the 12th at 23 UTC 23:39:21 what time is the retrospective, gaba? 23:39:48 just a sec 23:40:11 i think it was at the same time of the meeting... 23:40:25 as it is the time it works the best for most of the people 23:40:41 hm, maybe we should do tuesday at 23 UTC and wednesday at 23 UTC? 23:41:02 and tell people to prioritize the retrospective 23:41:09 i think what we've done before for Monday holidays is delay the meeting by 24 hours? 23:41:24 we can also do that, yeah, so tuesday at ... 17 UTC? 23:41:30 +1 to delay for 24hs 23:41:39 i think 18:00 UTC because DST ended 23:41:44 err, thanks catalyst! 23:41:49 yes! 18 UTC on tuesday 23:41:55 all options here are okay for me with a preference for not 2300 utc, since that's dinner time for me during c 23:42:08 gaba: and retrospective the 12th at 18 UTC too? 23:42:09 during non-dst hours 23:42:14 I have the retrospective as 2000 UTC Tuesday 12 November, but I would be happier with 2300 UTC. 23:42:46 (If we schedule retrospectives at 1800 UTC, I won't be there, because that's my 4am) 23:42:51 right, we had it for 2000 utc 23:43:04 let's keep it what we scheduled, so we only fiddle with one meeting 23:43:09 +1 on that. 23:43:12 so team meeting at 18 utc, retrospective at 20 utc 23:43:19 then people have a break of an hour between them 23:44:05 ok, i see no objections there? 23:44:29 next item: * When shall we do some more stable releases? 23:44:38 that's from me. 23:44:57 we haven't put out an 0.4.1 or 0.4.0 or 0.3.5 in a while. We should gear up to do that. Any thoughts when? 23:47:18 would it make sense to wait to a bit after 0.4.2 rc's are out? 23:47:50 with all the sponsors coming to an end maybe we should just wait and get it into the planning we do around 1st of december? 23:47:57 not sure. there will always be alpha code, and it's been since September since our last stable 23:48:06 but waiting for dec wouldn't be horrible 23:48:42 Let's do it late November/early December? 23:49:04 maybe early dec so we are done with s31 ? 23:49:17 +1 23:49:38 okay, let's plan for it then 23:50:13 I added the retrospective time to: https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam#MeetingsSchedule 23:50:30 teor: i have a question on the pad for the CI and review role policy. i think i voted +1 to that an hour or so ago 23:50:42 Please update your timezone for daylight saving time, if it hasn't been already 23:50:59 and i think right now based on the mailing list that gaba and mike is missing to vote on it? but i think nickm might also have a question about it? 23:51:19 I do not think that our policy requires everybody to vote. If we reach the end of the period and enough people have said +1 and nobody has said -1, it becomes provisional policy 23:52:00 so i guess that is what happened some 52 min. ago? 23:52:11 and then we all need to vote again or just the people who haven't voted yet? 23:52:20 Unless the clock started ticking again every time the policy was revised? 23:52:42 Yes, that's true. The revisions contain the target date. 23:52:51 i don't know the answer to that, if nothing have changed to the spec and only time have ratched forward i'd assume that my vote is still as valid as it was before there 23:52:56 unless i pull it back for some reason 23:54:52 So I suggest two things: 1) we think of clarifications for the proposal policy to address this stuff. 23:54:55 So I provided a final clean version, with a voting period end date, and then people voted, and now we're at the end of the voting period? 23:55:06 2) we declare the thing to be provisionally accepted. 23:55:25 If anybody objects, they just say "-1" and it is no longer provisionally accepted 23:55:28 * ahf is cool with (2) given the amounts of +1'es 23:55:39 (I probably provided too many clean versions, sorry about that, I'm still working out how this process works.) 23:55:49 i thought my previous vote still counted because of no substantive change since i made it 23:55:54 If they don't think that it should have been provisionally accepted for even one day, we'll have a good learning experience about saying "-1" earlier. 23:56:29 :-) 23:56:30 could we maybe not spend a lot of time splitting hairs about process if the risk of stuff being challenged seems small? 23:56:51 +1 on that tool 23:56:54 *too 23:57:09 so (2) it is? 23:57:24 I was suggesting 1 and 2. 23:57:28 ah, both! 23:57:33 that is cool with me too 23:57:34 let's do 2, and also clarify the process so it's easier next time 23:57:37 I agree with (1) and (2), but think we should do clarifications on the list, not here. 23:57:38 yes 23:57:41 yes 23:57:44 great 23:57:46 i see one help with: 23:58:00 - would like extra review on #32399 (test-stem) because i still can't get test-stem to work reliably 23:58:04 from catalyst 23:58:23 anybody with some stem knowledge that is up for helping here? 23:58:29 Stem has known intermittent failures, see #30901 and related tickets 23:58:42 But I can run it, then merge 23:58:51 Run #32339, then merge if it passes 23:58:52 teor: ok, thanks 23:59:26 teor: btw, I tried to get stuff done on #30866 but I ran into "issues" as explained on the ticket. I fixed one but not the other. 23:59:46 catalyst: was that helpful enough to make it not that much of a concern for now or? 00:00:23 * ahf thinks the asn comment is from last week 00:00:41 ahf: yeah, i think so. we should still try to make stem more reliable 00:00:53 catalyst: yep! 00:01:10 I have prioritised sponsor 31 and 0.4.2 over #30901, but I can get back to it in December 00:01:35 sounds good, teor 00:01:46 ok, unless amybody have anything else i'm gonna end the meeting (and go to bed) :-) 00:01:57 thanks! o/ 00:02:03 i will send out minutes tomorrow and an email with reminders of next week's meetings 00:02:06 thanks all o/ 00:02:07 #endmeeting