17:00:41 #startmeeting 17:00:41 Meeting started Mon Nov 14 17:00:41 2016 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:41 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:00:47 hi everybody! It's the weekly network team meeting! 17:00:52 welcome! 17:00:58 hello :) 17:01:26 hi 17:02:01 oi 17:02:01 let's start with status updates. Can somebody else go first while I run to the kitchen for coffee? 17:02:56 i can go first 17:03:06 Hello the second week of November I worked on the following things: 17:03:06 - I worked on revising the client auth part of prop224 and reviewing Nick's proposed scheme. I received lots of useful feedback and it seems to be one of those projects with multiple use cases where a single solution won't satisfy all of them. My task is to find out a good default scheme that will satisfy most use cases, and then in the future we can perhaps incorporate more client auth schemes. 17:03:12 - Rebased my #19043 branch on latest master so that development on INTRODUCE1 cell can start. 17:03:15 - Initial review of #20029. 17:03:18 - Did a reading group session on onion services at upenn. 17:03:20 EOF 17:03:30 My plan for this week is to finalize the client auth thing and do code reviews. 17:03:55 back with coffee 17:04:02 i have been dabbling in tor tickets for the past week as a way to get my mind off of my eyeball. i am not, probably, doing the most important ones. but hey, it is all a bonus. :) 17:04:28 also i did a full pass through the remaining 029 tickets 17:04:38 i am working on reports / plan on updating the network team wiki this week as well to organize things better and make it easier for new pps / i also need to ping ppl on testing 0.2.9.5-alpha 17:04:38 also, maybe somebody should make plans for new stable versions 17:04:42 (sorry) 17:04:47 no worries. carry on. :) 17:05:37 * isabela is done 17:05:42 isabela: oh speaking of wiki, I did that last week just so you know :) https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam#DeveloperGitRepository 17:06:12 last week I merged a million things for 0.2.9.5-alpha 17:06:17 and put out 0.2.9.5-alpha 17:06:18 dgoulet: ! nic 17:06:22 *nice 17:06:44 I played with builders/tor-browser-bundle a bit. I found that newer python packages in Ubuntu don't know how to construct Debian virtual machines. 17:06:48 I intend to pull those virtual-machine details into the builder, and not rely on the python package with that data any more. #idea 17:06:51 EOF 17:07:07 and debuged and tested and analyzed and flipped out about and finished (finally) (I hope!) (?) #15056, which will complete the importantest parts of prop#220 . 17:07:44 it's a big branch, and it was tricky to test and examine. I plan to add more tests to it if I have hacking time where my brain is in the right way, but the test coverage is already fairly good 17:08:12 this week I am starting a sprint to implement prop#271. I have 11 work days left in the month (including today) to do it ... and 6 holidays and weekends that I could eat for it. 17:08:19 This will be a challenge! 17:08:49 oh wow 17:08:54 I am probably going to be disappointing at most of my other tasks for the next couple weeks. I apologize 17:08:54 oh my 17:09:16 are there ways that other folks here could usefully chip in for that? 17:09:18 Things I won't be doing much of in november include: 17:09:20 - documentation 17:09:22 - fixing bugs 17:09:28 - fixing up n3eeds_revision tickets I didn't write 17:09:30 - reviewing code 17:09:32 - merging code 17:09:35 - triaging tickets 17:09:37 nickm: isn't that a bit too much work? can't we satisfy the sponsor in a different way? 17:09:38 - figuring out bugs 17:09:43 - helping new devs 17:09:49 - scheduling meetings to talk about proposals 17:10:06 asn: we already delayed a month on this. 17:10:32 I believe that we can deliver a most-of version of #19877 with the key parts done. I do not believe that we need to merge by end-of-month, just implement 17:10:51 nickm: you are right 17:11:00 but (unless isabela knows different) we already delayed on this. 17:11:18 yes, this has been moved from y1 contract to y2 and then added a month extension 17:11:29 so its quite hard to push it more 17:11:54 but i would like to echo armadev - 12:09 < armadev> are there ways that other folks here could usefully chip in for that? 17:12:07 Moreover, the part I feel most strongly about: if we do some half-assed version here and get the sponsor to accept it -- I don't see how we could ask for _more_ funding on the task. So I'd like to do it right. 17:12:31 well, that big list of things I _won't_ be doing in the next couple of weeks -- if anybody can help with any of those, that would give me more time 17:12:55 for instance nickm is listed for bug triage this week 17:12:59 can't we give the sponsor a simulation or something? 17:13:01 would be great if someone else can pick i tup 17:13:03 *it up 17:13:13 also if others can add their names for the upcoming weeks of nov 17:13:13 i can definitely pick that up, but that's nothing compared to prop#271 17:13:15 asn: like I said, if we do that, we probably won't have any funding to _build_ the thing ever. 17:13:19 and it's an important feature 17:13:24 https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam#Weeklybugtriagerotation 17:13:30 re 'merging code', would you be happier if some people here started doing some of that, or happier if nobody did that for a while? 17:13:41 asn: if you pick this week, I'll take next wek 17:14:07 happier if people passed the obvious stuff to me and said "This is easy, merge it", and saved the less obvious stuff till later. 17:14:22 makes sense 17:14:24 dgoulet: sure i can take this week 17:14:33 asn dgoulet thanks! 17:14:51 asn: i can also take triage, if you like 17:14:54 I'm not trying to be a drama-source here, and I'm not actually all that stressed about this. I might feel different in a couple of days, but so far, I think I can handle it. 17:14:55 trac updated 17:15:08 nickm: ok. please let me know if you want me to prioritize review of guard stuff 17:15:17 dgoulet: some of us will be eating turkey and being yelled at by relatives next week, so it's a good week for you to pick :) 17:15:30 isis: feel free to take it 17:15:30 right that's something we could do that is prioritize highly prop#271 code that nickm makes 17:15:32 nickm: you are doing the right thing here / delegating and letting people know of yoru capacity limits :) 17:15:38 isabela: otherwise, i'll just do it 17:15:45 armadev: hiro is an admin inside of storm and soon will be on the host as well: #20685 17:15:51 armadev: :) 17:16:14 hiro: whenever you like, i can give you the tour of brachteata 17:16:16 figuring out 029 bugs and 030 bugs will also be a helpful thing others can do 17:16:17 nickm: there are some guard review tickets from andrea that are in needs_review, would it help if i review them? 17:16:24 if people can chip in 17:16:39 or maybe the one I had in mind got merged. 17:16:40 asn: I'll take care of that stuff; they're partially usable 17:17:02 nickm: ok. let me know if you need help with review. 17:17:09 thanks! 17:17:28 FWIW, I started review-group-12 today 17:18:04 yep saw. 17:18:06 One task I don't know about for this month: mikeperry has a big branch where I have some open tor design questions I really need to think about and talk through. 17:18:08 im gonna pick some tickets from there 17:18:17 isis: if you are picking a week to be on triage / please add your name to the wiki :) 17:18:38 Maybe I could spend an hour picking out the relevant issues and trying to extract and explain them in some detail, and move the discussion on them ahead. 17:18:40 asn: okay, i'll take triage this week, since it sounds like you might be more help with the new guard stuff than i 17:18:48 isis: thanks :) 17:18:49 isabela: okay! 17:18:53 mikeperry's patch will be quite valuable to tor once it's merged. 17:18:57 +1 17:19:38 ok. i'll try to do that today too 17:19:48 once my brain resets from my first guard sprint today 17:19:56 also please don't feel you can't ask me to do stuff or help with stuff! 17:20:00 It's fine! 17:20:10 I'll just say "sorry, please wait" more than usual. :) 17:20:17 many thanks to all 17:20:43 I think that's it for me :) 17:21:09 next person / more people? :) 17:21:11 nickm: thanks :) 17:21:18 thanks, I'll go quickly 17:21:20 My last week was about some ticket reviews, patches and implementing #20029 which is based on #19043 which made me re-review quite a bit that code! 17:21:23 In those patches of last week, mostly fixes post #17238 merge. Also, I've just triaged quite a bit prop224 tickets this morning. 17:21:34 I now plan to address #20029 review from asn, more unit test and hopefully have something mergeable for nickm by the end of the week! 17:21:54 and I'll prioritize review/triaging/needs_revision stuff 17:21:55 -- 17:23:02 anyone else? 17:23:08 mine is short! i've started helping with how i can on reviewing code, and i have documentation/object models for nickm which i will send this week 17:24:04 cool! 17:24:25 I hope we can simplify that a lot. One thing to do might be to create tickets about where we can eliminiate stuff. 17:24:29 err, simplify stuff? 17:24:36 make the object relationships better? 17:24:39 something like that 17:24:44 any more updates? 17:25:42 i can go quickly 17:25:45 woo! 17:25:50 nickm: cool, that sounds good. and maybe that can help with plans for the modularization effort 17:27:02 isis: go ahead? 17:27:02 last week i did more paperworky things for the OTF grant, got all the testvectors to match for the curve25519 implementation 17:27:13 isis: whenever you want 17:28:02 i did not get started on the paper/documentation i'm supposed to write because i somewhat randomly decided to take a 300€ flight to california 17:28:23 EOF 17:28:53 ok 17:29:04 anyone else? I'll enumerate discussion topics (but not start discussion for a minute or two) 17:29:10 - 029 status 17:29:13 - 028 and earlier status 17:29:31 - are there more things we should do "rotations" on than triage? 17:29:39 any more topics or updates? 17:31:09 ok! 17:31:20 so, 0295 came out last week 17:31:38 we've got a few issues left. Any that we can mark as "not fixing this for 029" would help. 17:31:44 any where we can do patches for 029 would help 17:31:53 armadev: any interest in helping sort through those some more? 17:32:00 yes, i can do another pass 17:32:14 but i don't think things have changed that much since my first one 17:32:29 (there are a few new ones) 17:32:38 looking at the new ones could help 17:32:54 are there any still in the 029 milestone where your opinion is "we could/should defer this"? 17:33:03 ok 17:33:32 how imminent would we like 0.2.9.x-stable to be? that is, how ruthless should our defer logic be 17:34:01 we'd like it to make it into debian. so that implies "fairly soon". 17:34:29 i'd prefer to avoid taking fixes for stuff that isn't a regression and isn't a big problem. 17:34:40 ok. there's still a swath of your merges from last mondayish that i haven't looked at yet 17:34:45 especially if those fixes carry a significant risk of failure 17:34:53 e.g. the #20499 ones 17:34:59 yeah. #20499 seemed like a must-fisx 17:35:15 but some of those fixes were "uh oh, that's kind of tricky" 17:35:25 I hope we don't have any more bugs like that in the 029 lifecycle 17:35:32 and we can call the next release -rc 17:36:28 ok. i was thinking of putting my dabbling into an 0.2.7.7, but i will instead focus my dabbling on 0.2.9.x this week. 17:36:42 0.2.8 probably needs dabbling too if you want :) 17:37:00 I am going to punt on trying to make backports happen, or to make our backport policy logical, this month. 17:37:13 it's something that could use thought though 17:38:23 makes sense. 17:38:52 maybe we should aim for december for a new 028x and 027x and etc, and a policy 17:38:56 for backporting 17:39:44 which brings me to the last question -- this bug triage rotation thing is working out well IMO. Should we be looking for more things that can become rotating responsibilities? 17:40:04 nickm: what do you have in mind? 17:40:20 not sure. 17:40:32 nickm: wouldn't be crazy to delay until December 028 and 027 I believe 17:40:56 maybe "making sure that tickets have people working on them" ? 17:41:18 i have a memory of suggesting something else for a rotation, in a previous meeting 17:41:20 like, making sure that the stuff in the review-group is getting reviewed and the pile of annoying 029-blockers is getting cleared and etc 17:41:24 i do not remember what it was though :) 17:41:24 * isabela thinks we should do a eoy review as a team 17:41:40 maybe it was 'oldstable maintainer' 17:42:17 makes sense 17:42:23 and try to prep better for 2017 / e.g. calendar for 2017 releases 17:43:15 but i would push this to december since nov is quite crazy for us 17:43:48 ok 17:43:55 so, more stuff for this week? 17:44:12 I wonder if we should take the above and bring it to a mailing list so we can discuss a bit more async? 17:44:19 yes 17:44:22 oldstable maintainer, responsabilities, and so on 17:44:27 backport policy 17:44:42 as I have no clear answer for you nickm but I can think about it with a thread :) 17:44:55 dgoulet: i like that 17:45:05 sounds positive to me 17:45:25 ok. anything else, or are we done for the meeting? 17:45:33 * dgoulet is good 17:45:43 * isabela is good 17:45:47 (I can send the first email on the above if you like ^) 17:45:49 Thanks, everybody! 17:45:57 dgoulet: yes, that would be awesome 17:46:02 will do 17:46:46 #endmeeting