17:00:54 <nickm> #startmeeting network team meeting, 19 Aug 2019 17:00:54 <MeetBot> Meeting started Mon Aug 19 17:00:54 2019 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:54 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 17:00:57 <asn> o/ 17:00:58 <nickm> Hello network-team! 17:00:59 <dgoulet> hello 17:01:10 <nickm> asn, dgoulet, catalyst: hi! 17:01:20 <nickm> I expect teor4 has gone to bed 17:01:21 <nickm> a 17:01:24 <nickm> ahf is still out 17:01:28 <nickm> gaba is away today 17:01:31 <nickm> do we have mikeperry ? 17:01:39 <Sebastian> hi 17:01:42 <nickm> hi Sebastian ! 17:02:42 <nickm> our pad, as usual, is at https://pad.riseup.net/p/tor-netteam-2019.1-keep 17:03:12 <nickm> let's take a quick look at the 041 status page to make sure we didn't miss any critical issues: https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases/041Status 17:04:07 * nickm is removing some tickets from 041-should and 041-must 17:05:00 <Sebastian> I think teor4 said he wanted to add something to the pad on my behalf but I am not spotting it. I'll take a second to check it thoroughly. 17:05:17 <nickm> *they 17:05:24 <nickm> I'll look too 17:05:53 <nickm> I don't see anytthing blocking for 041, though I wish we had a more solid answer for #31301 17:06:02 <Sebastian> oh, sorry 17:06:17 <Sebastian> and now I am seeing it under "discussion". 17:06:33 <nickm> if there's no blockers on 041status, let's move to the kanban 17:07:15 <nickm> I'm moving some stuff from needs_review to closed 17:08:32 <nickm> can everybody please make sure that the "next" and "doing" columns are up to date for what they are doing now and doing next? 17:08:52 * asn doing it 17:08:55 <nickm> thanks 17:09:26 * asn has been doing it wrong 17:09:29 <asn> i put my task in the wrong box 17:09:42 <nickm> once we're done with that, let's get on to reviews. I don't have too much for this week, so I can probably take on a review somebody else doesn't want to do or is stuck with 17:10:13 <dgoulet> there were 3 that needed a reviewer... one was merged couple hours ago eheh ... so not much 17:12:19 <nickm> Let's move on to announcements, if there's nothing on this 17:13:00 <nickm> catalyst may be back today; let's welcome them and let them ramp up again 17:13:03 <nickm> ahf is still out 17:13:10 <nickm> let's keep helping with gsod 17:13:17 <nickm> the kanban is on dip 17:13:24 <nickm> and 0.4.1.5 should come out tomorrow. 17:13:30 <nickm> On to discussion! 17:13:42 <nickm> first topic is when to do next 029-040 releases. any thoughts there? 17:14:03 <dgoulet> dizum IP change is not a bad idea 17:14:10 <dgoulet> (to line them with that ^) 17:14:21 <nickm> I'd be fine with that 17:15:12 <nickm> and we just merged that today 17:15:19 <dgoulet> yup 17:15:37 <nickm> so maybe we should do the merges now and spend a couple of weeks letting bugs appear, releasing new stables if none show up? 17:15:48 <dgoulet> the other thing I was wondering (and that is a bit on mikeperry's side after Stockholm) is the LTS + relay thing where we decided to warn relays more actively or sth? 17:15:54 <dgoulet> could we fit that in the next 029+ releases? 17:16:37 <nickm> I'd like to see a definite proposal there; it's hard to talk about backporting something like that when there isn't code 17:18:06 <dgoulet> fair eough 17:18:06 <nickm> Some possibilities for doing this (as with protovers or recommended versions) would not require any code changes in 029 at all 17:18:25 <dgoulet> I need to talk to mikeperry about this, I just can't remember the specifics :S 17:18:44 <nickm> that is another argument for having a written plan and a patch 17:19:14 <Sebastian> From the dirauth perspective there is no rush to release a new version with dizum's IP address, you could easily figure that out and release afterwards 17:19:23 <Sebastian> dizum works fine if you use the old IP address for now 17:19:52 <nickm> Yeah. from my perspective, the question is whether we've accumulated enough stuff that we should just put out new releases anyway 17:20:01 <nickm> it's been almost a year since the last 0.2.9 release iiuc 17:21:13 <nickm> maybe we should use our metapolicy process on https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/ReleasePolicy and https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/Backports#DecidingwhichTicketstoBackport , but that is another question :) 17:22:14 <nickm> next question is sebastian's about rust 17:22:23 <nickm> what version of rust does sebian ship in stable right now? 17:22:34 <Sebastian> 1.32 17:22:47 <Sebastian> sorry 17:22:48 <Sebastian> fatfingered 17:22:51 <Sebastian> 1.34 17:22:58 <nickm> looks like fedora has 1.36 17:23:08 <Sebastian> current version since a few days ago is 1.37 17:23:25 <nickm> and the proposal is to increase our minimum version to 1.32 , in Tor master only? 17:23:49 <Sebastian> I have a patch for our configure script which raises the checked version to 1.28 - our code uses features from that but the configure version check wasn't adapted 17:23:54 <Sebastian> so I think that's a clear bugfix 17:24:15 <Sebastian> then I would like to move to 1.32 because that would allow to simplify the allocator story 17:24:44 <Sebastian> and then I would like to propose moving to the 2018 edition of Rust, for better borrow check and some other modernizations. 17:25:08 <Sebastian> I think raising it higher than what debian buster has will probably not be a great idea 17:25:11 <nickm> One thing I wonder about here is our story with dependencies. Are there any of our dependencies that would have compatibility problems with increasing our minimum version and using the 2018 edition for master only? I am hoping not. 17:25:40 <Sebastian> you can mix and match editions in your dependencies 17:25:56 <Sebastian> so our editions could remain edition 2015 for now 17:26:06 <Sebastian> so our dependencies* 17:26:49 <nickm> ok. I think this is something we should try out then, and see how it goes. I'm already reviewer on that ticket, so I'll try to say something smart 17:27:19 <Sebastian> One of the policy documents says you must use latest stable rust 17:27:33 <Sebastian> I don't know why that got added, so that's why I wanted to clarify in this meeting. 17:27:39 <nickm> next discussion topic is "0415 -- is it stable?" If anybody thinks "no" please let me know as soon as you can, otherwise I'll release tomorrow when I wake up 17:27:47 <nickm> Sebastian: makes sense 17:28:45 <nickm> Sebastian: can you reference the policy on the ticket, and make a pull request for bug31442 so our CI can try it out? 17:29:14 <Sebastian> where do I learn about how to make a pull request? 17:29:46 <Sebastian> (sorry this might be getting OT - I will gladly accept answers later or in pm) 17:30:31 <nickm> Sebastian: if you've used github before, it's https://github.com/torproject/tor/ -- just make a pull request against that 17:30:36 <nickm> and link it on the ticket 17:30:43 <nickm> otherwise we can explain on #tor-dev 17:30:44 <Sebastian> ah ok. 17:30:54 <Sebastian> didn't know it was using github now 17:30:55 <Sebastian> thanks! 17:30:55 <nickm> Are there any other discussions for this meeting? I do not see more boldfaced stuff 17:31:18 <nickm> Sebastian: just for reviews and CI. We're hoping to move to a gitlab instance, but the review and CI feature is really useful 17:31:49 <nickm> mikeperry has a list of things he's at risk of not finishing, so anybody with free time might take a look at those... 17:31:52 <nickm> anything else? 17:33:33 * dgoulet is good 17:33:44 <asn> same her 17:33:45 <nickm> ok. hearing nothing else, let's call this meeting done. Thanks, everybody! 17:33:54 <dgoulet> o/ 17:33:55 <nickm> I'll be online for a few more hours if anybody needs me for anything 17:34:09 <asn> o/ 18:25:43 <nickm> #endmeeting