00:01:46 <ahf> #startmeeting s55 IPv6 project 00:01:46 <MeetBot> Meeting started Thu Apr 9 00:01:46 2020 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 00:01:46 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 00:01:57 <ahf> https://pad.riseup.net/p/tor-s55-ipv6-project-keep is our pad 00:03:41 <teor> So I'll start with a status report 00:03:50 <ahf> sounds good 00:04:19 <teor> The unit tests in #33633 turned out to be terribly complicated. 00:04:29 <nickm> :( 00:04:33 <teor> It's old code, and there's a lot of embedded state/data interactions. 00:05:10 <teor> So I think I'm going to stop here, open a ticket to finish the tests for the remaining functions, and get that PR merged. 00:05:27 <nickm> +1 00:05:48 <teor> I did a significant redesign of the options for IPv6 extends in #33818 00:06:24 <teor> I am much happier with them now. They are simpler, and match Tor's existing extend options. 00:06:42 <teor> (ExtendByEd25519ID and ExtendAllowPrivateAddresses) 00:07:13 <teor> I also made a list of all the places that tor enforces IPv4 in #33817 00:07:34 <nickm> how do you mean "enforces"? 00:08:16 <teor> Requires extend cells to contain IPv4 addresses, or requires extends to happen over IPv4 connections 00:08:25 <nickm> ah, ok 00:08:40 <teor> Or only uses the IPv4 address from the extend cell 00:09:05 <teor> Only considers IPv4 connections canonical 00:09:52 <ahf> remind me what the definition of a connection being canonical is? it's only for connections between relays, no? 00:10:19 <teor> Yes, canonical connections are the connections that relays prefer to extend over. 00:10:28 <nickm> a connection is canonical if it's not only authenticated, but it is to the "right" address 00:10:28 <ahf> yeah 00:10:35 <ahf> ack, ok 00:10:36 <ahf> thanks 00:10:39 <gaba> hi! 00:10:46 <nickm> the idea is to (slightly) prevent traffic-analysis MITM 00:11:16 <teor> nickm: I'd like your opinion on changing extend_cell_from_extend2_cell_body(). 00:11:26 <nickm> paging in... 00:11:44 <nickm> what change do you have in mind? 00:11:45 <teor> Do we even use it any more? Have we removed all the tor versions that use EXTEND cells? (vs EXTEND2 cells) 00:12:18 <nickm> good question. I'll look that up... 00:12:37 <teor> Thanks! It's one of the functions that requires IPv4. 00:12:55 <nickm> according to tor-spec.txt, extend2 was added in 0.2.4.8-alpha 00:13:01 <nickm> so in theory everybody should be using that 00:13:12 <teor> Oh wow ok. So can we actually delete that function? 00:13:25 <nickm> if anybody is using it, it is either an ancient client, or a bad attempt at a compatible implementation... 00:13:32 <teor> Hmm ok 00:13:45 <nickm> but this function is for what use? 00:13:49 <nickm> rgg, no documentation here... 00:13:51 <teor> Probably out of scope then. But I won't worry too much about changing it. 00:14:21 <nickm> so hm 00:14:36 <teor> Oh right, it doesn't matter, because EXTEND cells can only have IPv4 addresses. 00:14:36 <nickm> this function is used whenever we get an extend2 cell, right? 00:14:51 <nickm> or did you mean extend_cell_from_extend1_cell_body? 00:15:00 <nickm> yeah 00:15:32 <teor> I have misunderstood that function's purpose 00:15:39 <nickm> ok 00:15:51 <teor> I thought it was making an extend1 cell from an extend2 cell. But it's actually a struct parser. 00:15:58 <nickm> right 00:16:18 <teor> So I need to change this line: if (!found_rsa_id || !found_ipv4) /* These are mandatory */ 00:16:31 <teor> Thanks, probably best we move on with the rest of the meeting :-) 00:16:51 <nickm> ok 00:17:49 <teor> I have also been helping GSOC and Outreachy applicants, but we're doing better at sharing that workload 00:18:47 <ahf> cool! 00:19:04 <teor> And I spent quite a bit of time on helping relay operators (and handing over some of that work to the network health team) 00:19:26 <teor> There was a bit of an issue that a large relay operator needed fixed 00:20:07 <teor> But GeKo and arma2 are working on the relay side, and ahf and juga and GeKo are working on the sbws side 00:20:16 <teor> So I won't need to do much more on that 00:20:37 <ahf> nod 00:20:46 <teor> And I also had to do some HR-related work, but we're making progress there as well. 00:21:13 <teor> Oh and alsmith asked about another IPv6-related grant, check your emails :-) 00:21:56 <teor> Here's my todo list for the next day or so: 00:22:38 <teor> * organise a meeting about IPv6 reachability and PT reachability with sponsor 55, 28, and 30 00:22:49 <teor> (looks like we will all want to change the same code soon) 00:22:49 <ahf> i saw the email coming in, but i wont look at it before i've slept a bit 00:22:56 <ahf> right 00:23:06 <teor> * assign nickm some Sponsor 55 coding tickets 00:23:13 <nickm> yesplease 00:23:14 <teor> Is there anything else I should do? 00:23:45 <nickm> those sound good to me. Also anything you want me to do on outreachy/volunteer stuff, or any tickets I should review, please let me know 00:24:12 <teor> Yeah I think you're next on my list for handing off a review 00:24:19 <nickm> also I'm doing 0.4.3.4-rc soon -- friday or monday, most likely 00:24:33 <nickm> ah, I see you answered on #33491 00:24:40 <teor> Yes I have some backports on my list to do 00:24:58 <nickm> ok; any to 0.4.3.x, or not? 00:25:12 <nickm> i'm not planning to do any older releases right away, so there is no hurry on those 00:25:55 <teor> Yeah there is #33491, #33804 (which needs a minor man page tweak) 00:26:07 <teor> I think #33072 is too big to add at this stage 00:26:10 <ahf> i think dgoulet rasied that one during the team meeting too 00:26:17 <nickm> ok 00:26:26 <teor> ahf: which one? 00:26:35 <nickm> #33804 00:26:56 <ahf> PreferIPv6 as default 00:26:58 <teor> Yeah I have reviewed that one, I can merge them both soon 00:26:59 <teor> But not #33072 00:28:17 <teor> I think I'm pretty much done, just need to update the pad. 00:28:19 <teor> Any questions? 00:28:26 * ahf has none 00:29:53 <nickm> not on my end 00:29:58 <teor> gaba? 00:30:47 <nickm> oh, one request, though. When you are putting a lot of stuff for my attention, if you can tell me what you would like done by when, I can try to prioritize the things that are most urgent 00:31:28 <nickm> you often do this, and it works really well for me 00:31:52 <teor> Sure, I will try to be more specific every time. Sometimes it's hard, because I give things to you incrementally. 00:32:07 <teor> If I forget, please feel free to ask. 00:32:44 <ahf> cool 00:32:49 <ahf> gaba: do you have anything? 00:33:53 <nickm> (gaba may be in high latency mode) 00:34:07 <ahf> i can end the bot and gaba can prod us if there is something 00:34:12 <ahf> #endmeeting