13:32:41 <nickm> #startmeeting
13:32:41 <isabela> !
13:32:41 <MeetBot> Meeting started Wed Jun 10 13:32:41 2015 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:32:41 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:32:50 <nickm> I see dgoulet , and I have rumors of Yawning .  Who else is around?
13:34:32 <nickm> well, let's start.  It might be fast this week :)
13:35:07 <nickm> I spent Thu->Sat last week in a meeting, which went very well, but has produced piles of stuff.  Yesterday was mostly admin stuff.  That leaves Monday and last wednesday.
13:35:14 <athena> hi meeting
13:35:37 <nickm> hello athena!
13:36:39 <nickm> So my current stuff is trying to get another 0.2.6 out soon, and try to merge more, and get progress on 15055, which I don't know when I'm gonna have time for....
13:36:46 <nickm> and try to grab low-hanging fruit on S and U items...
13:36:53 <nickm> once I'm done answering questions for a whole bunch of people.
13:37:39 <athena> i've been crunching on #4581; know how and what to build but needed a testing tool, so i have another branch with a new control port command to make spammy directory connections for testing
13:37:40 <nickm> Oh, and for fun, I wrote a real rust program that does something this weekend.  I like it enough to use it for distributed RNG in prop#224.
13:38:10 <nickm> athena: cool!  is that command worthwile for inclusion in real tor?
13:38:47 <isabela> nickm: 0.2.6 or 0.2.7?
13:38:58 <athena> probably not unless "make a bunch of directory server connections and then do nothing with them" is a useful feature for inclusion in real Tor
13:39:08 <athena> but it sure helps for testing here :)
13:39:10 <dgoulet> fuzzing test ? :)
13:39:33 <nickm> isabela: a new 0.2.6; we've been accumulating backported bugfixes
13:39:47 <isabela> nickm: cool
13:40:03 <Yawning> nickm: I'm 90% ish sure thatn 16052 is 026 ready
13:41:04 <nickm> Yawning: nifty!
13:41:18 * teor_ is here after computer trouble
13:41:21 <nickm> btw, would anybody else like to try playing release engineer sometime?  There's really nothing hard to it.
13:41:35 <nickm> teor_: you had computer trouble, and now you're here?  Or you're here, looking for computer trouble? :)
13:41:48 <athena> sure, i'll try it sometime if you like
13:42:00 <nickm> athena: does #4581 seem like a huge pile of work, or pretty reasonable?[5~
13:42:14 <athena> it's not too bad so far
13:42:29 <nickm> athena: ok.  Then what I'd suggest is that we go through the release process together today. :)
13:42:30 <teor_> I had network issues
13:42:35 <nickm> teor_: :(
13:42:38 <nickm> teor_: welcome
13:42:41 <nickm> who's next?
13:42:50 <Yawning> I can go since I didn't do much
13:42:54 <Yawning> revewing pt code
13:43:04 <Yawning> still full of h8
13:43:13 <Yawning> read a bunch of papers and a blog post draft
13:43:33 <Yawning> >.>
13:43:53 <athena> nickm: okay, but i need to disappear for a while around 1700 UTC
13:44:02 <nickm> athena: that's fine
13:44:05 * dgoulet can report next
13:44:18 <Yawning> also wrote up ideas on how to respond to a paper
13:44:33 <Yawning> think most of the people that need to have seen all that
13:44:36 <Yawning> that's it I think
13:44:53 <nickm> Yawning: great.  Katie needs feeback there that Roger and I have been totally meaning to get her "any day now"...
13:45:05 <nickm> ...which means it's great when more people step up
13:45:19 <Yawning> think it's just Roger that needs to sign off?
13:46:00 <Yawning> dgoulet: we should talk to the unenix paper people and see if they want to do a blog post with us too
13:46:01 <nickm> I meant on the "ideas how to respond to a paper" thing
13:46:06 <Yawning> ah
13:46:27 <Yawning> (per arma's , we should talk about it before some random jornalist flips out about it" idea)
13:46:42 <nickm> +1
13:47:12 <dgoulet> Yawning: reading the paper this morning, I started last night so after that I'll dive in the email thread
13:47:20 <Yawning> mmk
13:47:31 <Yawning> you've talked to them right?
13:47:39 <dgoulet> Yawning: yeah I'm in contact with them
13:47:47 <dgoulet> Yawning: they should send me a final version soon fyi
13:47:50 <Yawning> we should like, see if they want to do more stuff with us
13:47:59 <dgoulet> yup
13:48:09 <dgoulet> here is my quick update:
13:48:14 <dgoulet> did work on some tickets. Some worth mentinning but I know that I need to review #15775, revision needed in #16288 and patch in #16260.
13:48:22 <dgoulet> Hopefully I get a review soon from asn on #4862 (big). Also reviewed #13642
13:48:34 * dgoulet done
13:48:55 <nickm> dgoulet: on 4862, who is  reviewing that?
13:49:02 <nickm> me?
13:49:09 <dgoulet> "Hopefully I get a review soon from asn" :)
13:49:12 <dgoulet> asn started
13:50:05 <dgoulet> nickm: your input on #16260 would be quite useful to me! :)
13:50:10 <dgoulet> (if you are looking for more technical work :P)
13:50:38 <nickm> Not looking for more work, but definitely looking for places where people are blocking on me
13:50:45 <karsten> speaking of, is this a good time to discuss "master-key-ed25519"?
13:50:50 <dgoulet> nickm: :)
13:51:01 <nickm> karsten: during discussion time?
13:51:05 <nickm> any more status reports?
13:51:07 <karsten> nickm: sure.
13:51:55 <teor_> I made a top-n enhancement to #15775 and added a sanity limit of 512 fallbacks
13:52:05 <nickm> hm. Okay.
13:52:12 <teor_> (it now sorts in order of weight)
13:52:23 <nickm> that makes sense
13:52:39 <teor_> it was suggested by weasel in the patch workshop
13:53:21 <teor_> Other than that, got a few minor issues tidied up, including clang sanitizers
13:54:16 <teor_> I'm done
13:55:49 <nickm> teor_: glad about the clang sanitizers.  It's annoying stuff to work around, but undefined behavior is nobody's friend
13:56:20 <teor_> nickm: thanks
13:56:36 <nickm> and the annoying part is not your fault or clang's fault, but C's fault.
13:56:43 <nickm> are we on to discussion time now?
13:56:48 <teor_> indeed
13:56:53 <nickm> karsten: what was it about ed25519 keys?
13:57:09 <karsten> so, this is about sanitizing bridge descriptors.
13:57:14 <Yawning> (ugh, the stuff I want to talk about, I can't yet ;_;)
13:57:23 <karsten> server descriptors now contain such a line now, but extra-info descriptors don't.
13:57:24 <nickm> Yawning: seekrit?
13:57:35 <Yawning> preprint
13:57:36 <nickm> karsten: ah, and ei descriptors should too?
13:57:36 <karsten> is it possible to add that same line to extra-info descriptors?
13:57:46 <nickm> karsten: it's doable, though one more piece of redundancy
13:58:03 <dgoulet> Yawning: can you in some sekrit-internal-fewpeople email thread?
13:58:15 <karsten> one purpose would be that people can grep extra-info descriptors for a given ed25519 identity.
13:58:40 <karsten> another purpose would be that this field would be the only indication in bridge descriptors that the bridge uses an ed25519 identity.
13:59:03 <Yawning> dgoulet: yeah, that's what's happening right now :P
13:59:09 <karsten> if there's no such line, I'd have to add it as part of the sanitizing process. which is doable. but it wouldn't fix the first purpose.
13:59:12 <dgoulet> Yawning: ah that :P
13:59:20 <karsten> which would be about relay descriptors.
14:00:02 <karsten> nickm: how about I write this code,
14:00:03 <nickm> so, I'm a lttle sad about the redundancy, but clients mostly don't fetch these so it probably isn't too bad
14:00:19 <karsten> and then let you know if I come up with more good reasons for having that line or not.
14:00:26 <nickm> ok
14:00:42 <karsten> ok. it's good to know that there are no general reasons against having that line.
14:01:21 <karsten> (end of discussion topic. thanks!)
14:03:34 <teor_> Can we talk about opt-in vs opt-out for fallback directories on #15775?
14:03:44 <nickm> sure!
14:04:01 <nickm> Right now, my own thoughts are a bit in favor of opt-in, but I know weasel prefers opt-out.
14:04:12 <teor_> exactly what I was going to say
14:04:37 <weasel> my reasons are these:  opt-in adds work-load to us.  the operators already signed up their relay to help the tor network.
14:04:38 <nickm> Certainly doing opt-in now and opt-out later isn't too hard if we change our mind...
14:04:55 <teor_> In terms of numbers, we have enough either way
14:04:58 <nickm> but if opt-out is going to make trouble, then doing it first will make that trouble anyway.
14:05:27 <weasel> (anyway, meeting)
14:05:44 <nickm> weasel: The difference as I see it (IMO, add diclaimers, etc) is that operators have offered to help the network with the understanding that if some time later they decide not to help, they will stop receiving traffic
14:06:07 <nickm> but with fallbackdir, they'll get junk traffic until every version that listed them is unused.
14:06:22 <nickm> so maybe they'd see that as an imposition
14:06:53 <teor_> I hadn't thought of that
14:07:20 <nickm> I do like the idea of telling tor-relays anyway, and starting an opt-in process for early sign-up.  We can say that it might be opt-out later on.
14:07:42 <nickm> And see whether we get 99% of the people we ask saying yes and nobody objecting to opt-out...
14:08:00 <nickm> or 60% of the people we ask saying "HELL NO" and lots of people objecting.
14:08:05 <nickm> then we'll know where we stand.
14:08:45 <nickm> once meetbot is gone I'll ask weasel what he thinks. :)
14:08:50 <nickm> more topics for this meeting?
14:09:17 <teor_> thanks, nickm
14:09:18 <dgoulet> hrm, apart from tickets that needs some reply from people (which can wait), I'm ok
14:10:18 <nickm> ok. then let's call ourselves done?
14:10:26 * nickm waits another 90 seconds first and drinks some tea
14:10:42 * nickm endorses https://marktwendell.com/ teas
14:10:56 <Yawning> *baf*?
14:11:24 <asn> hello friends
14:11:25 <asn> i'm well
14:11:30 <asn> wrapping up my thesis these days
14:11:35 <dgoulet> woot o/
14:11:37 <asn> in two days, i will be done
14:11:44 <asn> and will have more tor time
14:11:52 <asn> i have started reviewing #4862
14:11:57 <asn> my next step is to test it
14:12:08 <nickm> asn: cool!
14:12:14 <asn> and i also need to fix #16255 asap
14:12:23 <asn> and i still have guards on my radar
14:12:32 <asn> also the SoP stuff
14:12:41 <asn> that's it from me. thanks :)
14:14:23 <nickm> Last question: should we delay the next 026 for any of these in-progress things?
14:14:30 <nickm> If so, please let me know after-meeting. :)
14:14:31 <nickm> #endmeeting