16:58:44 #startmeeting network team meeting, 7 May 2018 16:58:44 Meeting started Mon May 7 16:58:44 2018 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:58:44 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:58:46 Hello! 16:58:55 hello people :) 16:58:55 This week our pad is at https://pad.riseup.net/p/eLp5F7kimFBi 16:58:59 How's everybody doing today? 16:59:15 I hear dgoulet won't be able to make it today 16:59:48 And I think isabela is in transit. 16:59:58 * pastly is here though 17:00:21 So, let's do our usual things! 17:00:22 hi 17:00:26 hi catalyst 17:00:43 We begin with the roadmap! 17:00:58 There are 8 days left till feature freeze 17:01:17 note that we can still fix bugs after feature freeze, but let's not go overboard. 17:01:22 hi ! 17:01:39 Also please remember that after writing a feature, it still takes time to review, revise, and merge. :) 17:01:47 hey 17:02:08 Are we on track to finish all our stuff that's listed for 0.3.4 on the spreadsheet? 17:03:51 i think we good on the guard stuff. 17:03:54 Given that #25507 can be postfreeze, #25510 is ongoing, #25499 is just documentation at this point... 17:03:59 I think my tickets are doing okay 17:04:27 Wow, what if we actually hit our roadmap target, how wacky would that be? 17:05:15 Next up is reviewer assignments. 17:05:19 and 14 days before we meet if everything goes well! good timing 17:05:32 Looks like most people have between two and four 17:05:46 yep 17:05:54 i think there was big spectrum of difficulty this week 17:06:05 some ticets were real easy, some others hard. 17:06:14 Does anybody have anything there that they can't review, or which they will need to ask for help on? I would be totally glad to help review whatever. 17:06:54 nickm: fwiw, another reason we are trying to keep you low on the reviews, is because you also review merge_ready tickets. 17:07:09 makes sense 17:07:22 i might ask for help for one of the bugs while teor is away, we were going to look at it as a pair 17:07:33 which one? 17:07:48 2 sec 17:08:06 #17949 17:08:25 rl1987 added a new PR and i will review that and then maybe ask for additional eyes 17:08:42 Anybody able to help ahf review that one? 17:09:20 i can help 17:09:29 cool! 17:09:39 isis <-> ahf: please coordinate :) 17:09:57 cool! isis i'll do first round of review and then prod you once i'm done with that :-) 17:10:00 thanks isis 17:10:02 Rotations this week are nickm=triage, ahf=community, asn=coverity, catalyst=CI. 17:10:08 ack 17:10:12 ack 17:10:43 One thing I've noticed is that the CI role can really eat a lot of time. 17:11:14 Would it make sense if we say that the CI person is supposed to _notice_ CI failures and draw others' attention to them, but is not personally responsible for fixing every bug that CI detects on their own? 17:11:18 * arma4 finishes looking at backlog 17:11:41 ahf: sounds good! 17:12:09 nickm: sounds reasonable 17:12:36 catalyst: ok. I'll add that to the wiki :) 17:12:59 also i'd like us to keep in mind that broken CI slows down code review because we have to distinguish existing brokenness vs newly broken stuff added by a patch 17:13:05 +1 17:14:12 Announcements should be basically the same as those from last week -- just remember, 0.3.5 is right around the corner, and deferring a feature to 0.3.4 is no disaster 17:14:16 *to 0.3.5 17:14:22 So, on to discussion? 17:14:37 juga left a question about #25960 ; let's have a look at that 17:15:45 the big question there is whether we want the bandwidth measurement file to be in the same format used by all our directory documents, or whether we should just keep extending the existing format 17:17:02 I can see the argument either way 17:17:25 I think my default decision would be "the implementors should decide what they want to implement, and we should build that." 17:17:30 any thoughts here? 17:17:42 do we ever see ourselves using common parsing code for them all? 17:18:12 sounds good to me. my only thought is that with the old format, i found it surprisingly hard to add a version line, since its first line was just a number, and no keyword or versioning or anything. 17:18:35 so hopefully we have improved over that :) 17:18:57 catalyst: in theory we could use parsecommon.c 17:19:37 arma4: I think the new design retains the first-line-is-a-number thing for compatibility, but adds a header section including a format version. 17:19:49 ok 17:20:44 I have a couple of issues to discuss that I listed. 17:21:03 One we already answered, and the people I was asking (teor and juga) are not here right now. 17:21:33 The other one is what our status is with TROVE-2018-005 -- whether it's ready to merge as-is in the next stable, or whether it needs more testing etc 17:22:05 Another option is to have a dirauth operator try running with it for a bit to make sure nothing breaks, since it's a dirauth issue. 17:22:58 i can do that if you send me a patch or however you want to get it to me 17:23:32 Anybody have thoughts here? I can try to figure something out, but I might not get to it this week. 17:24:03 you're talking about the TROVE, right? 17:24:08 Yes 17:25:51 isis: ? 17:26:07 And do we have other stuff to discuss? 17:26:18 I guess one thing might be "what do we backport to 0.3.3 to call it 0.3.3-stable"? 17:26:35 but for that i think it might just make more sense for me to open a spreadsheet and let people comment there 17:27:01 Does that sound good? 17:28:06 sounds good 17:28:07 yes 17:28:07 i will look over teor's comments in the email thread after the meeting and see what's up with the TROVE-2018-005 status 17:28:23 and then i guess email arma4 the patches? 17:28:29 sounds good to me! 17:29:13 (or if it's on a security list thread already, point me to which email) 17:29:58 iirc we decided not to change consensus methods because it'd be a wreck to backport (maybe we should discuss the design here and if there's something easier we could do moving forwards) 17:30:30 hm, ok 17:31:01 like having linear/i-support-everything-up-to version numbers 17:31:42 maybe consensus versions should be more like a protover? e.g. consensusmethod=1-15,17,19-21 17:32:21 it's possible, but that's just the encoding. 17:32:38 you'd encode that as 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 17 19 20 21 17:34:17 So, hm. It turns out there is only one backport case to consider for 0.3.3 right now: the fix for #25691 #25692 . It's been merged to 0.3.4 since 22 April 17:34:46 teor is out now, so they can't comment on whether they think it's got enough testing to backport 17:35:03 it would be fun to get a tor browser nightly build that includes the latest tor alpha 17:35:06 I'll ask dgoulet what he thinks later on ... 17:35:30 i wonder if ln5 is still making nightlies 17:35:32 Any more to discuss / do at this meeting? 17:36:00 hmm… looking over teor and dgoulet's comments on the TROVE issue, i have some things to fixup in the patches, so i'll prioritise that for today 17:36:18 arma4: the alpha should contain the latest tor alpha 17:36:51 we made it in particular to test the -rc if that's what you mean 17:37:04 GeKo: I think arma4 wants a nightly that tests 0.3.4 out of git 17:37:17 that#s available too 17:37:33 isis: sounds good! Please let me know if it's going to eat your week? :) 17:37:49 ok 17:38:00 it shouldn't be too much i think? 17:38:07 arma4: https://trac.torproject.org/projects/tor/wiki/doc/TorBrowser/Hacking there is a nightly builds section 17:38:28 okay! hearing no other issues, I make one final announcement, then the meeting is over: 17:38:31 i'll also make a public ticket for the parsing-before-signing issue 17:38:59 My highest priority this week is to help everybody get their stuff done! Please let me know if there is anything you're doing that might be blocking on me, or that I could help with. 17:39:12 I want to remove all your blockers _first_ before I do my fun stuff. 17:39:25 ok, and that's a wrap for another meeting. Thanks, everybody! 17:39:28 #endmeeting