17:00:08 <nickm> #startmeeting weekly network teem meeting
17:00:08 <MeetBot> Meeting started Mon Apr 11 17:00:08 2016 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:08 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:00:15 <nickm> good hello!
17:00:32 <Yawning> EHLO
17:00:33 <nickm> I'm nick and I feel unwell today ut I will try to get the meeting going well
17:00:36 <isabela> opa
17:01:54 <nickm> quick update: last week I did catch-up on long-deferred things, wrote PETS reviewed and tried to
17:02:04 <nickm> generally help with triage etc
17:02:19 <nickm> This week, more helping with release stuff, omre meetings, and hopefully some code
17:02:32 <nickm> hopefully merging and reviewing too
17:02:35 <nickm> next?
17:02:42 <Yawning> fixed openssl 1.1.x again
17:02:51 <Yawning> poked at obfs5 and related innards
17:03:02 <Yawning> this week, meetings, the last of my llc setup, more obfs5
17:03:09 <Yawning> unless there's something more important to do
17:03:20 <Yawning> or I decide to hide in bed with the cat from the world
17:03:24 <Yawning> enxt
17:03:41 <asn> i can go
17:03:45 <asn> Hello. This week I discussed prop259 with Ola and team. I also looked into the next parts of prop224 and started a new discussion thread:
17:03:48 <asn> https://lists.torproject.org/pipermail/tor-dev/2016-April/010719.html
17:03:50 <asn> I also merged the changes from the previous prop224 thread into torspec.git yesterday.
17:03:51 <asn> Finally, I reviewed #13239 for which I have mixed feelings.
17:03:53 <asn> thanks
17:04:00 <asn> im now sneak reading isabela's mail.
17:04:02 <asn> next.
17:04:30 <armadev> i lurk. i'm sure i did some stuff, but i have been doing too many different kinds of things to notice quite what. but i am reading here. :)
17:04:38 <mikeperry> worked with isabela on the triage last week; will be working on the OTF proposal with her this week. Otherwise have been doing Firefox code review and talking with Cloudflare
17:05:07 <mikeperry> I will let isabela explain the triage stuff we did, I guess?
17:05:08 <mikeperry> next
17:06:29 <armadev> (we're just too efficient!)
17:07:14 <isis> i just got back from taking some weeks off because i was too stressed out about tor
17:07:31 <isis> today i am dealing with going through all my email and other backlog
17:07:42 <isis> tomorrow i will get back to code
17:08:20 <isis> if there's anything i should pay special attention to immediately, or something additional that i could/should help with, please let me know
17:08:24 <isis> next
17:08:25 <mikeperry> isis: sorry to hear about your stresses; hopefully you are talking to the right people about them. anyway, welcome back! :)
17:08:51 <isis> oh also i will have good news in an announcement this week :)
17:08:52 <asn> isis: hey. welcome back! the guard algorithm by ola et al. needs some love.
17:09:02 <isabela> wc o/
17:09:03 <armadev> isis, would you like to review a pets paper in your bridgedb area? i have already reviewed it but a) you might like it and become a smarter isis, and b) you might hate it and tell me why i should hate it
17:09:08 <isis> but for now it's still sekrit :)
17:09:28 <isis> armadev: sure, i would be glad to
17:09:29 <dgoulet> hi! prop224 HSDir code, tor-dev@ discussions (asn keeps on providing!:), worked on last patch assigned to me for 028-stable, got sucked in also in HS mini-dev. meeting (in person) and I finally dealt with personal stuff also a bit last week
17:09:30 <dgoulet> --
17:10:34 <nickm> next update?
17:10:42 <isabela> I sent an email to the list just before this meeting - I think is better for folks to take sometime to read and digest what is there.. not sure if we should chat today about it.. you tell me ;)
17:10:59 <isabela> i wont be around next meeting because I will be in DC and afk probably
17:11:04 <Yawning> isabela: wat
17:11:11 <Yawning> oh next meeting as in
17:11:14 <Yawning> next week
17:11:19 <isabela> yes, monday 18
17:11:25 <Yawning> and not the thuggable transports thing
17:11:33 <armadev> 'Formalizing Release Guidelines' is the mail?
17:11:37 <isabela> yes
17:11:38 <isabela> yes
17:12:04 <isis> asn: (re: olabini's guard algo) do you want to brief me during or after the meeting on what love the algo needs?
17:12:42 <isabela> (done)
17:13:34 <armadev> #item learn a summary from asn about the guard algorithm (i want one too)
17:13:41 <isabela> also, dont forget to take a look at the april tickets and see what is there waiting for review
17:14:02 <asn> isis: we recently realized that all the various torrc options of Tor like ClientsUseIPv6, ReachableAddresses, etc. influence guard selection and hence our upcoming algorithm.
17:14:11 <Yawning> (oh I guess I poked at meek-server a bit)
17:14:32 <armadev> dgoulet: did i hear correctly that nickm is attending the montreal mission?
17:14:33 <asn> isabela: we are trying to fit all these filtering options in the algorithm right now, which have created some extra design deciisons / research questions.
17:14:40 <asn> isis: ^
17:14:45 <dgoulet> armadev: yes :)
17:14:46 <Yawning> (but that's not core-tor or whatever)
17:14:48 <armadev> woo
17:15:29 <asn> isis: at the same time, I *think* the thoughtworks crew is at the stage where they are trying to wrap up their implementation efforts, because they can't keep on adapting to the new research issues that appear.
17:15:41 <mikeperry> (oh, fwiw, I did not work on isabel's whole release guidelines thing, just worked on some tagging and suggested a tagging pattern. like armadev, I've been a bit too distracted to do much of any one thing :)
17:15:42 <Yawning> I also think I can get meek-server performance to be 2-3x faster but I don't think that's the bottleneck anyway (correct me if I'm wrong)
17:16:31 <asn> isis: and yes, the thoughtworks team is actually implementing the current prop259 at https://github.com/twstrike/tor_for_patching .
17:17:12 <isis> awesome :D
17:17:16 <asn> isis: here is the latest version of the proposal: https://lists.torproject.org/pipermail/tor-dev/2016-March/010625.html
17:17:32 <armadev> did we miss any humans here?
17:17:35 <asn> isis: but the discussion in  that thread (plus https://lists.torproject.org/pipermail/tor-dev/2016-April/010728.html) shows that more changes need to happen
17:17:45 <asn> isis: EOF
17:18:01 <armadev> https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/ReleaseGuidelines is isabela's url
17:18:10 <isis> so i think that, unless one of these filtering things e.g. ClientsUseIPv6 is a total blocker, we should *really* try to get the new algorithm rolled out soon because it contains substantial improvements against attacks already
17:18:40 <nickm> +1
17:19:06 <asn> isis: i've been considering them hard blockers indeed, because then we would break functionality that people might depend on.
17:19:23 <nickm> also, it's big, and it _will_ have surprise bugs, so we shouldn't take it late in the cycle
17:19:55 <isis> asn: ah, i see what you mean
17:20:45 <asn> let me suggest we use the ML for further discussion so that ola and team can also see it.
17:20:58 <isis> sounds good
17:21:03 <asn> thx
17:21:08 <nickm> any more updates? are we ready for discussion?
17:21:26 <nickm> I'm esp interested in what folks think of https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/ReleaseGuidelines
17:22:15 * armadev is reading the wiki page
17:22:21 * isis just fixed a typo in it
17:22:27 <isabela> :)
17:23:22 <Sebastian> I like the point about prioritizing leftover tickets
17:23:28 <Sebastian> especially for review
17:23:49 <armadev> who are the people who most need to sign on to the routine part of it? for example, mike and i can get away with signing up for not-very-much, because we're spread across more teams
17:24:02 <isis> what is the difference between 0.2.9.x-final and 0.2.9.x-maint?
17:24:24 <armadev> (but the routine could help me here too, since i know that people expect sign-ups at the beginning of the month, and finishes at the end of the month)
17:24:52 <nickm> isis: was thinking of having one milestone for "stuff that goes into the first stable 0.2.9.x" and another for backporting bugfixes.
17:25:09 <isabela> armadev: for folks who interacts sporadically with the team, this is good for them to know how to get their stuff in :)
17:26:57 * armadev fixes another typo
17:27:56 <armadev> oh neat, tor-doc-*
17:28:02 <Yawning> doom meeting in 2.5h?
17:28:06 <nickm> y
17:28:09 <Sebastian> I'm in the situation that sometimes I have a day where I can review a thing or write a doc or so, but I don't know beforehand whether and when that'll be
17:28:27 <Sebastian> What should I pick to work on?
17:28:32 <Yawning> k, gonna hide with th ekitty
17:28:37 <Yawning> *the kitty
17:28:38 <Sebastian> Yawning: enjoy
17:28:44 <nickm> so the hardest part of this plan is going to be "don't add stuff to the milestone after triage; use 029-proposed instead, and we talk about it."
17:28:52 <isabela> Sebastian: ideally 0.2.9.x-final
17:28:55 <nickm> Does anyone think that's crazy?
17:29:16 <isis> Yawning: o/
17:29:28 <armadev> dgoulet: i notice there aren't any sponsorR deliverables listed on this page. have we thought that one through?
17:29:40 <isabela> that is my fault
17:29:47 <dgoulet> armadev: we have tickets with SponsorR so we should put some there
17:30:00 <armadev> nickm: is 029-proposed a milestone or a keyword or a what?
17:30:11 <nickm> keyword
17:30:14 <isabela> yes, we should. I sent it out without it becuase I wanted folks to have it before the meeting (was already late)
17:30:33 <isabela> dgoulet: armadev ^
17:30:53 <armadev> ok
17:31:12 <armadev> (i will leave it up to dgoulet and asn and all to pick which sponsorR things they want to get into 029)
17:31:13 <dgoulet> (my two cents) I think that "029-proposed" idea could work and making few people actually be able to set tickets in a milestone
17:31:36 <dgoulet> armadev: it's been triage already, we just need I guess to document it in that wiki page ^
17:31:52 <isabela> yep
17:32:06 <armadev> well heck, even better
17:32:13 <nickm> I don't know if we can restrict the milestone in trac like that, but having it as an advisory "pleasedon't" would be a start
17:32:17 <dgoulet> the discussion part about 029-proposed will be a fun experience since well timezones (we don't have teor here for instance :)
17:33:07 <isabela> yes, and I am a total fail on the notes right now :(
17:33:11 <dgoulet> (I think) we can with permissions in trac that is that group will only be the one able to change the milestone for instance
17:33:43 <armadev> but you want anybody to be able to make *other* changes to that same milestone field
17:33:55 <nickm> yeah, we don't want other teams unable to change milestones, or to have their milestones locked, unless they want them locked
17:34:14 <dgoulet> ah yeah... "per component" permissions we need then ... ish
17:34:35 <isabela> complicated
17:34:44 <armadev> and not just per component, but also per value (anybody can change to 0210-proposed, today, no problem)
17:34:58 <armadev> (speaking of which, when 0.2.9 is done, is 0.3.1 next?)
17:35:14 <armadev> or do we switch to 3.1
17:35:21 * armadev derails the meeting
17:35:34 <nickm> or per-milestone?  or per-keyword? heh
17:36:03 <nickm> armadev: There's a bug in older tors where they don't beleive in three-component version numbers.  so till they're dead, we might be stuck with w.x.y.z instead of x.y.z
17:36:11 <dgoulet> one problem I see with the 029-proposed keywords and then we have to discuss it is what if I open a ticket that I really want to fix right away, then do I wait for the discussion instead of taking time right away to make the patch in case the ticket doesn't go in 029 or I just do it and then I risk the chance of having it postponned? ...
17:36:14 <nickm> version numbers are bikeshed; i'd like to drop the 0
17:36:38 <nickm> dgoulet: recommendation: use your judgment, and/or ask a friend?
17:37:00 <dgoulet> so then why do we have 029-proposed in the first place if some of the time that won't apply and some will ?
17:37:04 <nickm> dgoulet: like, if it's a crash bug or a comment typo then it falls nicely into the obvious-merge
17:37:11 <dgoulet> right
17:37:35 <nickm> some stuff has a higher risk of delay than other stuff.
17:37:40 <isabela> dgoulet: I think is that the idea would be to say here in the meeting.. because this way you are checking if by doing so you are adding load to others (code review for finalizing the release etc)
17:38:13 <dgoulet> ok
17:38:18 * armadev was really hoping the 'triage stick' would be a role ("i'm holding the triage stick this week")
17:38:45 <isabela> that is the main goal with this control, with the triage we were looking to commit according with our capacity... but as time goes, those tickets will change (for sure) so we need to be constantly checking in for changes to make sure we are not overcommitting
17:39:38 <nickm> I think as we get experience here we'll all get a better feel for which 029-proposed tickets get accepted for 029 and which don't.
17:39:52 <nickm> armadev: I have no idea about 0.2.10 vs 0.3.0
17:40:25 <armadev> in theory somebody could scrap 0.2.9 and move straight to 0.3.1, even
17:40:46 <armadev> our momentum will take us on 0.2.x forever
17:41:14 <armadev> isabela: bug in the wiki page, "won't finish before the end of * look at closed tickets" not sure how to fix
17:41:37 <isabela> ?
17:41:46 <armadev> looks like it's missing some words and a few newlines
17:43:04 <isabela> let me fix it
17:43:35 <isabela> btw does folks understand the keywords system we are proposing (based on tor browser model hehe)
17:43:44 <dgoulet> in "Each week" section, tickets are treated _only_ as bugs but we have lots of "enhancement" in those
17:44:07 <armadev> isabela: it would help me, when you say you need buy-in from the team, to hear specific names that you most want to hear buy-in from. maybe we don't want to put people on the spot, but maybe they won't guess that you mean them, or vice versa
17:44:29 <GeKo> tor browser meeting in 15 minutes
17:45:20 <dgoulet> armadev: where is that on the page?
17:45:30 <dgoulet> "buy-in" for what?
17:45:31 <armadev> last week of the month:
17:45:38 <armadev> oh, the buy-in, it was in her email
17:46:07 <armadev> "Do you understand and agree with", etc
17:46:46 <isabela> ok i would like everyone in the list, but for sure the payed folks
17:47:26 <isabela> i think the ack of people in the network-team list is good because we all work together in a way or another :)
17:47:37 <nickm> I'm cool with this.  No plan is final; this seems like a step forwad :)  So, ack from me
17:48:11 <isabela> (fixed the wiki)
17:49:08 <armadev> re-using the monthly keyword habits from tor-browser-team could help keep people sane. good idea.
17:49:38 <isis> i ack also, this plan seems better than anything we've tried before
17:49:55 <isis> plus we could iterate and change a bit if it doesn't work
17:50:06 <isabela> isis: indeed
17:50:16 <isis> also +1 reusing the TB Teams's keywords
17:50:38 <nickm> I think I'm going to spend some time this afternoon tagging tickets in the 0.2.9 milestone, if that's a good use of my time.
17:50:39 <dgoulet> I'm still not totally convinced that the 029 proposed discussion will work that well with the rate of tickets we have _but_ I don't have a better idea so I'm quite ready to try it
17:50:44 <nickm> or I could hack
17:51:33 <isabela> Yawning: we should look at what tags to use for PT work
17:51:44 <isis> what if we had two milestones: 0.2.9 and 0.2.9-proposed, but only some people are allowed to move stuff into 0.2.9?
17:52:02 <dgoulet> (teor will be excluded from 029-proposed discussions for instance :S)
17:53:04 <nickm> dgoulet: we can also talk about this stuff tuesday nights
17:54:03 <asn> wrt ReleaseGuidelines buy-in; ack from me as well.
17:54:19 <isabela> we will
17:54:23 <asn> these monthly routines sure look exciting as well...
17:54:43 <isabela> i think we should do it on both meeting - tomorrow I plan on being around to chat about the guidelines as well
17:54:46 <armadev> i just realized i'm confused. are only a few people supposed to be able to set 029-proposed? or are only a few people supposed to be able to move tickets from 029-proposed to 029-for-reals?
17:54:48 <isabela> for folks from that timezone
17:55:19 <nickm> armadev: the second one.
17:55:38 <armadev> ok. so anybody can propose a ticket for 029, by setting the keyword 029-proposed
17:55:56 <dgoulet> this is what I understand yes ^
17:56:00 <asn> .oO(i wonder what tickets i would have commited to in the monthly routine of april, if it had happened...)
17:56:12 <isabela> aha
17:56:15 <armadev> maybe the wiki page should say that more clearly? the 029-proposed keyword is mentioned exactly once and comes out of nowhere
17:56:24 <isabela> https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam
17:56:30 <isabela> lets end with this url ;)
17:56:40 <isabela> for april tickets
17:56:41 <dgoulet> asn: yes, there is quite an epic foresight skill to acquire :)
17:57:07 <asn> dgoulet: :D
17:57:18 <nickm> if we all get practice committing to april tickets, that would be a good thing.
17:57:21 <isabela> armadev: i will update that
17:57:22 <nickm> Maybe I should too :)
17:57:37 <nickm> any more before the end of the hour, or are we winding down?
17:57:50 <dgoulet> "A ticket is potentially suitable to be included if:" --> we should really add a bullet point for features imo
17:57:51 <isabela> is tbb meeting now
17:58:21 <armadev> dgoulet: for features? but doesn't 029 have a bunch of tickets in it already, and those are the features, and we're not supposed to add more features to 029 now?
17:58:40 <dgoulet> armadev: what about 0.3.0 ?
17:58:54 <dgoulet> aaah! this is _only_ for 029!
17:58:55 <armadev> 0.3.0 is where "i just thought of a feature" tickets should be scheduled for
17:59:07 <isabela> :)
17:59:08 <dgoulet> "Set of guidelines for releasing Core Tor 0.2.9 ¶"
17:59:09 <armadev> and there will be a big triage session to decide what goes in 0.3.0
17:59:19 <dgoulet> makes more sense
17:59:19 <armadev> once 0.2.9 is closer to baked
17:59:33 <armadev> (insert more cooking metaphors here)
17:59:53 <armadev> (what happens when your cooking metaphors meet your battlefield medic metaphors)
17:59:56 <nickm> goig to #endmeeting now
18:00:03 <dgoulet> go
18:00:03 <nickm> #endmeeting