22:59:10 #startmeeting Network team meeting, 8 January 2020 22:59:10 Meeting started Wed Jan 8 22:59:10 2020 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:59:10 Useful Commands: #action #agreed #help #info #idea #link #topic. 22:59:14 hello network team! 22:59:18 hi 23:00:00 the pad is still at https://pad.riseup.net/p/tor-netteam-2019.1-keep - i have made a link to the 2020 pad in the 2019 pad (i guess it's a blockchain now?) that we can use next week 23:00:18 who is here other than catalyst? 23:00:21 me! 23:00:30 o/ nickm 23:00:44 mikeperry, teor4, dgoulet: any of you here? 23:00:53 hi ahf! I think you were sick? How are you doing now? 23:01:03 i spoke with asn earlier today, should say hi from him and he said he was recovering 23:01:25 heyo 23:01:26 that's good to hear 23:01:28 I am sort of here 23:01:31 i was having some inner ear canal infection so my balance was a bit off and i was ... having normal flu things, which sucked 23:01:46 but mostly the lying still with the head was. managed to play pokemon a bit and that is it 23:01:49 hey mikeperry! 23:01:52 ok, let's see what we are gonna do 23:02:09 how are things looking with roadmap? 23:02:21 the link is at https://dip.torproject.org/torproject/core/tor/boards 23:02:45 i haven't created the tickets i wanted to create for the jan/feb roadmap for myself yet, so i at least am a bit behind there 23:03:02 yeah, it doesn't have a bunch to do with what we're doing right now... 23:03:05 and at least one of them is a task where i am gonna work on it with nickm (look into the rust ed25519) 23:04:33 IIUC my tasks are working on C style and working on modularization stuff. I have patches in review there, and am waiting for reviews to come back. 23:04:42 should we ask people to make sure they revise the roadmap items AND create the missing tickets for jan/feb/mar ? 23:05:38 and by people i mean each other here 23:05:40 * teor is here 23:05:45 hi teor! 23:05:46 o/ teor 23:05:47 how are you? 23:06:52 ok, i think that is what we'll do! please walk over the things we talked about there in december and get the roadmap revised 8) 23:06:53 ahf: I think it's reasonable to create our own tickets for jan/feb/mar items, but doing duplicate copies of everything on gitlab seems laborious 23:07:01 Slowly getting better, just pacing myself 23:07:16 nickm: hm, maybe i could do that so everybody don't have to fight that 23:07:21 if we just create tickets with some keyword on? 23:07:38 i know the duplicated work is super demotivating :-/ 23:08:03 can people tag their new roadmpa tickets with networkteam2020q1 or something like that? 23:08:09 sure; maybe spell out in an email what we should do, linking to the current summary of our december meeting? 23:08:18 good idea 23:08:39 ok, i'll write an email tomorrow morning with this and we can prepare it for next week. that should be fine 23:08:43 cool! 23:09:05 how are we looking with reviews? neither david or asn is here, so we wont get any status from them 23:09:46 3 tickets in needs_review right now with a person to review them 23:09:48 ahf: I think gaba was using network-team-roadmap-* in trac, so let's follow that pattern? 23:10:01 teor: ya, i will ask gaba before doing this. good idea 23:10:27 I think I have my reviews under control 23:10:34 cool 23:11:07 teor: you looked at the last vewsion of #37264. Is it okay if I set you as reviewer on that? 23:11:11 *version 23:11:33 (Also, I hope it's okay to put it back in needs_review; the only changes I did was to show that CI passes) 23:11:50 Oh did we lose zwiebelbot over the holidays? 23:12:00 " Solve code issues that block running clang-format on our code. " 23:12:13 it's in the channel, but it isn't saying anything. :/ 23:12:26 #5000 23:12:37 maybe the ticket was mentioned somewhere else within a short timeframe? 23:12:37 how about that. 23:13:32 #37264: No matches found. 23:13:33 That's what trac says 23:13:38 #32764 23:13:40 arg 23:13:52 my bas 23:13:55 *bad 23:14:46 catalyst: if we can't find a tester for #32778, do you think we should go ahead and merge? 23:15:35 nickm: yeah, it seems reasonable by inspection, and appveyor seems to be ok with it 23:15:50 i think we lack automated testing for that functionality 23:16:09 * arma2 is nearby if people need any arma things 23:16:15 nickm: to be clear, you want the pre_formatter_cleanups branch reviewed? Is there a PR? 23:16:34 yes -- it's in the first comment on the ticket 23:17:16 thanks, I made a note on the ticket 23:17:49 thanks! 23:18:06 arma: you might also want to review #32588, since you filed the bug. And the related change in #32822. 23:18:35 teor: you and I are listed as co-reviewers on #29801 and #20218. How do you want to proceed on those? 23:18:38 dgoulet will also review it, id you don't have time :-) 23:20:18 nickm: for #20218, can one of us write a few simple tests, and the other review and merge? I don't mind which task I do. 23:20:59 I'd be okay with both too. Does one seem more fun to you? 23:21:17 if you don't care, I'll write tests. 23:21:32 Given my work/health time constraints right now, I think review would be easier and less likely to block 23:21:36 ok 23:21:56 I'll edit the ticket to record that decision 23:22:04 thanks! 23:22:07 thx :-) 23:22:20 Looks like the author would like help writing tests themself: I'll do one as an example and see where they want to go 23:23:11 do we want to move to discussions? 23:23:29 ok w me 23:23:35 #29801 is a bit complicated. I think we got stuck in a write-review-add loop, and now we need to cut it down. 23:23:39 first item there is from nickm with: * How would we feel about moving authdir_mode() and server_mode() into src/core? (They are used pervasively through the code, and might not actually belong in the modules that they enable/disable.) -nickm 23:23:58 * ahf is personally OK with it, seems like something we use everywhere 23:24:23 teor: that task is to write a proposal, right? If so, can we just take the current thing and call it a proposal, send it to tor-dev, and iterate? 23:24:37 * nickm is likely not up to date on the situation w 29801 23:25:05 nickm: yes, I think we should just merge the proposal as-is, it's already been through a few iterations 23:25:12 But let me review it furst 23:25:21 ok. I'll take myself off as reviewer then? 23:25:32 ok 23:25:57 The likely outcome is that I'll implement part of the proposal for the IPv6 grant. Or implement a very simplified form of the proposal. 23:26:27 [oh, on announcements: 0.2.9.x is EOL; 0.4.0.x is EOL in 25 days.] 23:26:40 [0.4.3.x is feature-frozen in one week.] 23:27:41 catalyst, teor: I'd like your opinions on the two discussion topics I put on the pad, since they are about code-cleanup/modularization stuff 23:27:55 :-) 23:28:27 nickm: neutral to positive on killing off macros that make our code harder to parse 23:28:53 (unless they have really significant benefits) 23:29:14 Seems like a net win to me, we're not using them consistently, and so they make different styles of unit test harder to read. 23:29:56 As for moving authdir_mode() and server_mode(): 23:30:08 I think it's a win for modularity / dependencies 23:30:24 re moving code, i would need more info 23:30:45 (what info would help?) 23:30:48 hm, i don't even remember using NS() in the places where they are used today 23:30:55 only NS_MOCK() and NS_UNMOCK() 23:30:58 nickm: mostly time to look at how things are right now 23:31:05 ok 23:31:49 We wanted to have a code structure where the ENABLE_MODULE_* macros are only used in the relevant directories 23:32:07 So moving this code would break that guideline 23:33:02 But if it simplifies a bunch of dependencies, or removes the mode header includes, it could be a net win 23:33:29 maybe I should give it a try with dirauth, and we can see how the change looks? 23:33:46 Yes, let's see how it looks 23:33:49 okay 23:34:38 i see one more item from catalyst: 23:34:45 - teor, i set you as additional reviewer on #32846; please let me know if you'd rather not do it 23:36:07 (also: I want to make it so that the dircache code also gets disabled when we are not a relay. Should it be controlled by a separate HAVE_MODULE_DIRCACHE, or is HAVE_MODULE_RELAY okay? I lean to the former) 23:36:31 (I'm thinking of doing this because there are comparatively few entry points) 23:36:46 ahf, catalyst: yes, that's fine, I can look at that manual page change 23:37:15 cool! 23:37:40 teor: thanks! 23:38:00 nickm: i think the former is cleaner? 23:38:09 nickm: I think we initially decided to disable everything using the RELAY define 23:38:45 Is there a clean separation between the two? How will we test them? What's the dependency tree look like? 23:39:07 I think that we should do it, for now, so that dircache <=> relay. 23:39:14 That is, let's not support having one but not the other. 23:39:25 in theory, they are independent 23:39:39 the separation is src/feature/dircache vs src/feature/relay -- 23:39:47 Oh, right, so it's effectively documentation, not a separate mode 23:39:52 right 23:40:58 do we have anything else we need to look at during this meeting? 23:40:59 Seems fine to me, we should probably review all the existing HAVE_MODULE_RELAY instances, to make sure they're relay-specific 23:41:04 not afaik 23:41:13 teor: ok. I'll pick this up too :) 23:41:40 ahf: in january we might want to talk about getting a 0.4.3 bug status page? 23:41:42 You wrote that on the pad :-) 23:41:46 please get back to my email on 1:1's - i hope to write to people ASAP with when we are gonna have our next 1:1 8) 23:41:54 ohhh teor, good point 23:42:00 teor: +1 on an 043 bug status page 23:42:03 do we want to have a 0.4.3 status pgae and talk about it during our meetings? 23:42:12 +1 23:42:13 we did it for 0.4.2 and i thought it was useful to have each week 23:42:23 teor: did you create the 0.4.2 one? 23:42:40 nickm did it based on one I did a long time ago 23:42:43 I think I made it by copying the 042 one 23:42:49 can one just "copy" the markdown from the 042 one and update it with 043 stuff ? 23:42:56 Yes, that should work 23:42:56 ah, there is one for 0.4.3 already? 23:43:07 wow, did I do that? 23:43:22 (catalyst: are you okay if I copy your comments above on #32778 onto the ticket?) 23:43:33 ah, no, there is no 043status yet 23:43:39 No, I don't see an 0.4.3 status page 23:43:44 i'll create that one tomorrow as well then and get it into the new pad for the next meeting 23:44:01 nickm: sure, go ahead 23:44:12 ahf: do you want to pick a time between now and Monday to assign 043-must/043-should? 23:44:19 or at least to do a first pass 23:44:24 catalyst: thanks, will do 23:44:41 nickm: should the whole team be involved there or ? 23:45:22 how about after vegas meeting tomorrow ? 23:45:25 ok 23:45:31 I've usually done a first pass myself, and invited the team to look it over 23:45:52 that time works for me 23:45:59 maybe the team can go over it on monday if we have the status page there? 23:46:19 yeah; or we can send out a note when we've done the first pass and the status page is up 23:46:26 and people can do it when they get the chance 23:46:27 either way 23:46:37 sounds good! 23:46:44 let's do that tomorrow after the vegas meeting 23:46:49 ok 23:46:50 i'll have the page ready prior to that then 23:47:01 oki, anything else? otherwise i'll tell the bot to stop logging 23:47:06 I talked with gaba, and if everything goes to plan, the IPv6 grant should start on 1 February 23:47:20 great 23:47:21 woh, nice teor! 23:47:29 So I might not get to do as much of our unfunded roadmap as I thought 23:47:48 We should also think about nickm's time for that, and anyone else who wants to be involved 23:48:05 (I think we can shift staff around a bit, but let's check with gaba) 23:48:47 yep 23:48:56 So there aren't any action items yet, but I just wanted people to know 23:49:09 We should do a mini-roadmap once it's confirmed 23:49:16 yep, agreed 23:49:21 woo 23:49:39 cool, gonna ask the bot to stop. thanks all for a nice first meeting o/ 23:49:43 see you around on irc! 23:49:48 #stopmeeting 23:49:51 #endmeeting