13:30:28 <nickm> #startmeeting
13:30:28 <MeetBot> Meeting started Wed Nov 26 13:30:28 2014 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:30:28 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:30:44 <nickm> who's here for the tor dev meeting this fine (for me) rainy (for me) morning (for me)?
13:30:59 <Yawning> it's raining here
13:31:08 <dgoulet> raining here :P
13:33:03 <nickm> So, I've been doing various.  Dave G was around last week, and we got a lot of discussion and planning and philosophy done.  Last wednesday I taught a lecture at MIT; I've gotten some student emails since then.  I hope something comes of it.
13:34:17 <nickm> I spent the weekend recovering from working the whole week long.  Monday was meetings and a little hacking.  Yesterday I implemented a new PRNG again.  I think this time the entropy-extracting part at least will make sense to put into Tor, instead of that getrandom branch I started to moot with dgoulet
13:34:34 <nickm> (https://github.com/nmathewson/libottery-lite)
13:35:03 <nickm> by end-of-month, I mean to get a libevent release ready, get 9262 merged, get libottery-lite finished, and get the testing plan for sponsor S written up.
13:35:13 <nickm> This may not be a realistic set of goals, but I am not a realistic guy.
13:35:29 <nickm> And that's about all for me this week.
13:35:36 <Yawning> do you want review for ottery-lite?
13:35:53 <nickm> once it's a little better.  It's still buggy and undocumented
13:35:57 <Yawning> ok
13:36:49 <nickm> oh!  And next week from Tuesday on, Roger and Mike will be in town to confer with me and Wendy.  So I might have less online time than usual.
13:37:02 <nickm> Would anybody volunteer to run the patch workshop next tuesday in case I'm busy?
13:38:03 <Yawning> the hour it's at is kind of rough for me, been trying to be better about when I sleep
13:38:06 <Yawning> ;_;
13:38:20 <nickm> np
13:38:53 <Yawning> can I go really quick? kitty is mad that her lap disappeared
13:39:04 <nickm> go for it
13:39:12 <nickm> dgoulet: so, I last say you... saturday morning?
13:39:17 <nickm> Anything new since then?
13:39:30 <Yawning> oh meant "can I talk" but :P
13:39:35 <nickm> oh!
13:39:37 <nickm> yeah, go fo rit
13:39:38 <dgoulet> :)
13:39:42 <Yawning> it's k, she'll cope
13:39:45 <Yawning> oh uh
13:39:59 <Yawning> I'm writing my sponsor S/T PT eval draft
13:40:07 <Yawning> which is neither code nor little-t tor
13:40:16 <Yawning> both which make me sad, but it's important
13:40:40 <Yawning> apart from that I'm thinking really hard about pq crypto still, which probably is a huge rathole
13:41:02 <nickm> yeah.  But file that under professional development. :)
13:41:20 <Yawning> I'll more than likely write a bunch of patches when I get annoyed at doing tech writing, but it'll be my main focus for a bit
13:41:43 <Yawning> (and I have a sponsor T kickoff meeting at some ungodly hour next week that I'm supposed to use voip for)
13:42:13 <nickm> ok
13:42:21 <Yawning> apparently there is a key exchange primitive that provides the security properties that we'd like that will fit in a single tor cell, but it's really new and scary
13:42:27 <Yawning> (SIDH)
13:42:51 <nickm> Quick question on one of my things then --- for 9262, maybe look at our notes, and let me know if you think any of them should block my merging andrea's patches?
13:42:55 <Yawning> but my pt proptotype will probably be ntru based since I am not bound by your pesky rules >.>
13:43:01 <Yawning> ok
13:43:08 <nickm> Rob's testing has given me some confidence that it's probably okay.
13:43:34 <Yawning> yeah from what I remembered that was our main concern when we did the review
13:43:37 <nickm> And if not, hey, I think I'm the only person on the team who lives somewhere that they celebrate a holiday this weekend
13:43:45 <Yawning> "this looks ok but gosh, that's a bunch of code"
13:44:19 <Yawning> nickm: gobble gobble?
13:44:31 <nickm> yah
13:44:42 <Yawning> are we going to also do something like KIST?
13:45:19 <nickm> Maybe in 0.2.6; maybe not. Probably soon if we can, though.
13:45:26 <Yawning> gotcha
13:45:51 <nickm> Aaron and Rob are hoping to visit me in mid december to talk about that, and other stuff.
13:45:55 <nickm> IIUC
13:45:57 <Yawning> not having access to snd.cwnd cnd.nxt, snd.una for userland has been a long standing pet peeve of mine
13:46:07 <Yawning> s/for/from/
13:46:17 <nickm> well, we know plenty of kernel hackers who work on plenty of kernels, right?
13:46:30 <Yawning> I've been accused of being one even
13:46:44 <Yawning> at least for freebsd/linux the socket option they're using returns the information required
13:47:03 <Yawning> though I think the structure needs ifdefs, havne't looked closely
13:47:10 <nickm> joy
13:47:16 <Yawning> (my freebsd side of the investigation was just looking at the kernel source)
13:47:43 <Yawning> maybe there's a better way to get the information
13:48:13 <Yawning> and clampink the socket send buffer and only using that instead of the congestion window probably isn't totally terrible as a plan b
13:48:46 <Yawning> but I'll stop rambling
13:49:19 <nickm> wrt one tor cell -- It probably wouldn't be too hard to make EXTEND able to span multiple cells, if we really had to
13:49:45 <dgoulet> Yawning: let me know if you want to cook a patch for the Linux kernel on that, I used to do a lot of kernel stuff before Tor, I guess I can help a bit :)
13:50:08 <Yawning> dgoulet: linux supports TCP_INFO
13:50:24 <nickm> Cool, thanks!  We'll see what's neeed
13:50:31 <Yawning> darwin *should* dunno about windows
13:50:44 <dgoulet> Yawning: ah! well at least Linux has your fields
13:50:51 <coderman> today i discovered https://github.com/nmathewson/libottery-lite , carry on nickm. carry on :)
13:51:24 <nickm> hi coderman!
13:51:29 <nickm> thanks
13:51:37 <nickm> tor meeting right now, but happy to chat about rngs after
13:51:53 <nickm> dgoulet: so, anything else to add for this week, or is it pretty much how it was on Saturday?
13:52:00 <nickm> I hope you're recovered from your trip
13:52:10 * dgoulet starts
13:52:16 <dgoulet> yeah I slept 14h on Sunday :P
13:52:52 <nickm> whoa
13:52:54 <dgoulet> most of my time now is on SponsorR, instrumentation side of HS and analysis script of the info will soon start
13:52:58 <Yawning> ECAT, I'll read backlog
13:53:00 <Yawning> >.>
13:53:09 <coderman> Yawning: in a rump stack, snd.cwnd, cnd.nxt, snd.una live in userspace, let alone access. (the timers, though. well...)
13:53:12 <coderman> ;)
13:53:27 <dgoulet> I did a small patch on monday to fix #13667 and I would like to get your feedback nickm on what Roger said
13:53:37 <dgoulet> nickm: see last comment, "3)"
13:53:48 <nickm> ah, yeah.
13:53:56 <dgoulet> so no ref in the code?
13:54:20 <nickm> In general, we assume that closed tickets are history, not present documentation.  So ideally, instead of saying (ref: 13667), the comment would summarize the issue and the thinking, so that the code reads fine on its own
13:54:32 <dgoulet> ok perfect
13:55:47 <dgoulet> so yeah that's about it, really been focused on one thing since Monday so not much on little-t tor side
13:56:06 <nickm> np
13:57:22 <nickm> oh hey, it looks like the lecture I did on Wednesday is onlyine now.
13:57:23 <nickm> https://www.youtube.com/watch?v=rIf_VZQr-dw
13:57:49 <nickm> I hope it isn't stupid.
13:57:53 <infinity0> ah nice
13:57:58 <dgoulet> it was not no worry
13:57:59 <nickm> okay, so, anything else for the meeting today?
13:58:12 <nickm> oh, and Has anybody seen intrigeri around?
13:58:52 <dgoulet> nickm: nope... been a while
13:59:00 <DrWhax> nickm: afaik, not today yet
13:59:05 <nickm> Not hearing anything else for the meeting today, so:
13:59:07 <nickm> #endmeeting