22:59:20 <ahf> #startmeeting Network team meeting, 4 december 2019
22:59:20 <MeetBot> Meeting started Wed Dec  4 22:59:20 2019 UTC.  The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:59:20 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
22:59:25 <ahf> yoyoyo
22:59:29 <ahf> hello network-team o/
22:59:38 <nickm> hihi
22:59:42 <ahf> our pad is at https://pad.riseup.net/p/tor-netteam-2019.1-keep
23:00:09 <ahf> who is at the meeting today? :-)
23:00:21 <gaba> o/
23:00:25 <ahf> o/ gaba
23:00:43 <catalyst> o/
23:00:46 <nickm> I am, though I must admit that it is dark and cold here
23:01:09 <ahf> i haven't been outside today, so i don't even know if it's cold here, but sure is dark by now :-/
23:01:31 <ahf> teor, mikeperry, asn, dgoulet: any of you around?
23:01:48 <nickm> (how is everybody doing?)
23:02:01 <mikeperry> I'm here
23:02:08 <ahf> \o/
23:02:19 <nickm> (I'm trying to do _something_ every day, since doing _exactly what I'm supposed to do_ is hard this time of year for me)
23:02:29 <nickm> hi mikeperry ! hi catalyst !
23:02:34 * teor is here
23:02:38 <nickm> hi teor !
23:02:44 <ahf> hm, usually we do 0.4.2 status now. do we wanna do that still?
23:02:51 <ahf> o/ teor
23:03:05 <nickm> so I think the question for 0.4.2. status is :
23:03:14 <nickm> Any reason NOT to release ?
23:03:29 <ahf> i have no reasons not to release at this point in time
23:03:37 <nickm> In theory, 0.4.2.x is stable in 11 days
23:04:53 <teor> Let's do it soon?
23:05:22 <nickm> I think we should add a discussion item (if we don't have one already) about what to do for upcoming releases.
23:05:32 <teor> +1
23:05:37 * nickm adds
23:05:44 <ahf> what to do?
23:05:51 <ahf> like having a page like the status page, or?
23:06:09 <ahf> the big item i have on my list is that we should remember to get the releases into the roadmap too when we revise that
23:06:10 <nickm> We can talk about it when we get to discussion. I mean, what to release when, and who does what work first
23:06:17 <ahf> ack
23:06:29 <ahf> okay, so we can move on to roadmap?
23:06:34 <nickm> sure
23:06:54 <ahf> roadmap is at https://dip.torproject.org/torproject/core/tor/-/boards
23:07:08 <ahf> does things like up-to-date to you?
23:07:27 <nickm> catalyst, teor: should #29215 be needs_review?
23:07:34 <nickm> similarly #32206
23:08:12 <catalyst> nickm: almost there
23:08:21 <nickm> ack
23:09:39 <gaba> s27 tickets are not up to date but i can check later with dgoulet or asn
23:09:53 <ahf> cool, yeah
23:09:56 <gaba> mikeperry: are you still working on #32040 ?
23:11:19 <mikeperry> I am not.. asn is working on reframing the classification problem, so I'm gonna wait for that to progress a bit more before we go deep into any traffic analysis edge cases (this is not a functionality issue with Tor; users won't notice)
23:11:34 <gaba> ok
23:12:45 <ahf> :-)
23:12:58 <ahf> okay, it looks like all tickets have been distributed for reivew, i just took the one that hadn't been distributed yet
23:13:19 <ahf> we have no announcements this week
23:13:29 <ahf> but we have a discussion item by nickm about upcoming releases!
23:14:24 <nickm> yeah.  We decided to postpone figuring out stable releases until December.
23:14:30 <nickm> Well, now it's December! :)
23:14:43 <nickm> IMO we should plan to release 0.3.5 and forward.
23:14:56 <nickm> Step 1 is backport as appropriate
23:15:01 <nickm> (if not already done)
23:15:06 <nickm> Step 2 is making changelogs
23:15:19 <nickm> Step 3 is ...
23:15:30 <nickm> but before we get there, we should ask: who does what here?
23:15:46 <nickm> I am okay with doing any amount of this myself, but I don't want to step in the way of anybody else
23:16:55 <ahf> i'd like to avoid taking on more things right now since i'm behind on s28 stuff and i want to have the shadow dev project out of the way before i add more things on the todo
23:17:21 <gaba> shouldn't we discuss this in the mailing list? sorry if there is a mail there that I didn't read
23:17:31 <gaba> as there are a few people that are not in the meeting today
23:17:36 <gaba> but +1 to share this responsabilities
23:18:19 <nickm> well, teor is the current stable-backport person, but I don't know whether they want to do some/all/none of the changelog/tarball/announcement bureaucracy
23:18:52 <nickm> So I'd like to know what they think before moving to a ml discussion
23:19:09 <gaba> ok
23:19:56 <teor> I've done most backports, but there's a bunch scheduled after stable?
23:20:26 <nickm> So are there any of them that are super important?
23:20:48 <teor> just looking now
23:21:09 <nickm> Left to my own devices, I'd put put the stable releases with things that are already in 042, and release 042 shortly after...
23:21:37 <nickm> but if there's anything crucial already waiting for to backport, maybe we should look at whether we should backport it earlier
23:23:13 <teor> I think we should backport #31810 and #32407, they're low risk, only happen in error cases, and are nice to fix
23:23:55 <teor> I want to backport #31571, just checking how complex it is now
23:24:14 <nickm> #31810 looks safe to me
23:24:35 <nickm> #32407 looks safe but I'm slightly concerned that it is low-severity in addition to being low-danger
23:25:18 <teor> Sure we can skip it, not a problem at all
23:25:58 <nickm> #31571 seems moderately high benefit, so let's go based on how 100%-safe the patches look
23:26:44 <nickm> we already backported the accounting issue of #32108, right?
23:28:19 <teor> #31571 has been tested in the alphas for the last 3 months, and the code seems pretty safe
23:28:38 <teor> Yes I backported #32108 two weeks ago
23:29:43 <teor> In #31571, most of the new code only runs when we crash :-)
23:30:51 <ahf> :-)
23:30:52 <teor> (done)
23:32:19 <nickm> ok
23:32:38 <nickm> teor: do you want to do the changelog-backports, or should I?  It's pretty mechanical
23:32:50 <nickm> I'm also planning no new 0.2.9, but please let me know if you think we need on
23:32:53 <nickm> *one
23:33:36 <teor> I don't really want to do changelogs :-)
23:33:43 <nickm> ok; I'm on it
23:34:12 <nickm> current plan for me is to aim for this week for stable releases of 0.4.1 and earlier, and next week for 0.4.2?
23:34:19 <teor> I wonder if we should release one last 0.2.9, so we gather all the fixes we've done already. I'm not sure if the risk is worth it. People should be running 0.3.5
23:35:11 <nickm> We could
23:36:03 <teor> I seems weird to say "here's one last supported 0.2.9 for 3 weeks, but really, you should move on, we won't fix it if our backports break anything"
23:36:05 <nickm> We should look at the pending changes ("cat changes/*|less" in release-0.2.9) and see whether there is anything we really really would not want to have 0.2.9 do without in its interim period
23:36:21 <teor> Sure that seems to make sense
23:36:42 <teor> Why do 0.4.2 after the rest?
23:36:48 <nickm> I don't see anything under "Major", but I haven't skimmed all the "Minor"s there
23:38:32 <nickm> I could aim to have them all this week/early next?
23:39:02 <teor> TLS 1.3 fix, guard discovery attack fix
23:39:49 <nickm> Those seem ... worthwhile, but not worth the stability risk of a new 0.2.9
23:39:52 <nickm> IMO
23:39:57 <nickm> I could be wrong
23:39:57 <teor> carrier grade NAT, onion service fixes that reduce load on the network
23:40:59 <teor> nothing major
23:41:15 <nickm> ok
23:41:48 <teor> There are still about 500 relays on 0.2.9
23:42:16 <nickm> Question is: whether they would upgrade to a new 0.2.9 but not to a new 0.3.5
23:42:28 <teor> 637
23:42:51 <teor> I think we're actually better to remove them from the network after a few months, and tell them to upgrade to 0.3.5
23:43:08 <ahf> good we have just tried doing that :-)
23:44:25 <nickm> cool
23:44:32 <teor> When do authorities automatically exclude relays by version?
23:44:45 <nickm> we do not have a standad time at which to do that
23:44:47 <teor> (I'll go check, the discussion can move on)
23:44:54 <nickm> (IIUC)
23:45:33 <nickm> any other discussion things for this meeting?
23:45:42 <ahf> there is none on the pad
23:45:59 <ahf> nothing else? :-)
23:46:17 <nickm> not from me.  Everybody should remember to look at pending proposals...
23:46:28 <ahf> yep
23:46:35 <ahf> okay, i'm gonna end the meeting
23:46:37 <ahf> #endmeeting