13:33:30 <nickm> #startmeeting
13:33:30 <MeetBot> Meeting started Wed May 20 13:33:30 2015 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:33:30 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:33:35 <nickm> it's dev meeting time, right?
13:33:40 <Yawning> is it?
13:33:50 <nickm> I think so?
13:33:55 <asn> think it is yes
13:33:56 <Yawning> oh that's right, I swapped phones so I don't get my "get out of bed" alarm
13:34:09 * asn moves to #tor-project for non-meeting dev matters
13:34:16 <dgoulet> morning
13:35:07 <nickm> okay, so let's start with checkin
13:35:22 <Yawning> I can go
13:35:23 <nickm> I've been hosed with meetings, but I only have one other meeting scheduled today, so that's something!
13:35:28 <Yawning> oh ok
13:35:35 <nickm> (i'll be quick)
13:35:46 <nickm> i'm hoping to get 0.2.6.8 out today, and finish a log of ed25519 stuff
13:36:09 <nickm> I hear from athena that she's going to checkpoint her 12498 review and seek help finishing it.  I want to get 12498 merged this month
13:36:22 <nickm> Yawning: I'm curious to see what you think of the stuff in my two openssl branches, if you have time.
13:36:31 <Yawning> I dropped a note in one of them
13:36:53 <nickm> next up for me haxxingwise is to see if I can finish ed25519 auth, and help isabela with trac stuff if she wants.
13:37:00 <nickm> and that's all for me I think
13:37:22 <Yawning> #16052
13:37:30 <Yawning> my branch appears to work, though people still argue over the behavior
13:37:50 <Yawning> I'm currently refactoring the linked list stuff, so I'll revisit the bheavior when I'm done and have something that compiles again
13:38:33 <Yawning> part of me wants to make sure there isn't anything needed to deal with that logjam precomputed table attack on TLS
13:38:50 <Yawning> (force ECDH is looking to be more attractive imo)
13:39:14 <Yawning> probably more stuff I'm forgetting, but this HS DoS stuff is chewing up mental cycles
13:39:17 <nickm> I looked at it with athena and I'm pretty sure that there's nothing to freak out about for Tor
13:39:24 <nickm> (the weakdh thing)
13:39:25 <Yawning> mmk
13:39:27 <Yawning> yeah
13:39:31 <nickm> but check my analysis
13:39:39 <Yawning> where is it?
13:39:53 <Yawning> is it hidden under a ticket titled "bugs"?
13:41:16 <nickm> ha
13:41:19 <dgoulet> I need to go soon so I'll jump right in.
13:41:20 <nickm> entering it now
13:41:24 <dgoulet> (sorry)
13:41:25 <Yawning> kk
13:41:28 <dgoulet> I have #14917 and #3733 in need of comment/review. I'm working on #8239 wh
13:41:28 <dgoulet> ile this new ticket is getting beefier #16120. This ticket is also on my horizon: #15937. And finally, keeping an eye on #16052.
13:41:35 <isabela> oi
13:42:03 <nickm> Yawning, dgoulet, *: my draft analysis is at: https://etherpad.mozilla.org/SIr2m7H5GY
13:42:26 <nickm> wow that's a lot of tickets
13:42:49 <dgoulet> spam :(
13:43:44 <dgoulet> some of them are small some of them are bigger, but those are my target for the next week or so
13:44:31 <nickm> great
13:45:52 <asn> hello. i just finished exams.
13:45:59 <asn> i'm currently looking into #16052.
13:46:16 <asn> after this finishes, my TODO list has tons of itms
13:46:27 <asn> like review #8243 and #4862
13:46:38 <asn> look into the graphs that karsten has produced on hidden serivces
13:46:50 <asn> look into the guard stuff
13:47:02 <asn> do some fuzzing on tor with afl
13:47:05 <asn> etc.
13:47:21 <asn> that's that. thanks :)
13:47:43 <nickm> sounds good. congrats on finishing exams
13:47:49 <nickm> anybody else here today?
13:47:51 <isabela> yep
13:48:07 <isabela> congrats asn
13:48:28 <asn> i also have two SoP students, that I have given only moderate attention too.
13:48:30 <asn> *to.
13:50:35 <nickm> anyone else to check in today?
13:50:49 <isabela> i have a quick one
13:51:31 <nickm> go for it
13:51:46 <isabela> I hope to get the trac stuff implmented and at least our 0.2.7 tickets using the new system and all sponsor tags under sponsor field on form
13:52:06 <isabela> by the end of the month
13:52:08 <isabela> next 2 weeks
13:52:13 <isabela> <end>
13:52:30 <nickm> great
13:52:36 <nickm> ok, so we move to discussion
13:52:37 <isabela> (sorry if it seems too long there are other things on my plate too)
13:52:57 <nickm> I think we've all got a lot of stuff to work on, but I want to make sure we don't lose sight of all the in-progress design tasks we have.
13:53:07 <nickm> I wonder if there's a good way to collect them all and move them forwards.
13:56:27 <nickm> On my mind are: deprecating old tors, handling future deprecation more gracefully,
13:56:40 <nickm> trying to get 12498 reviewed at last
13:56:57 <nickm> figuring out what design we actually want for guard nodes
13:57:04 <nickm> and I think that's the big stuff for right now
13:59:35 <nickm> anything more to discuss this morning?
14:00:05 <Yawning> nickm: how do you want 16052 to behave
14:00:32 <Yawning> current behavior is, load spikes to 1.0, but existing and new connections get serviced
14:00:54 <Yawning> see the ticket particularly stuff added today
14:02:01 <nickm> where does the cpu spend its time?
14:02:17 <asn> i can profile.
14:02:24 <nickm> that would be nice to know
14:02:25 <nickm> thanks
14:02:51 <Yawning> my guess would be it's processing the RELAY_BEGINs that are getting dropped
14:03:24 <Yawning> which is wasteful, but not catastrophic because we have a circuit level scheduling
14:03:31 <Yawning> (which is why other stuff works)
14:03:42 <nickm> I wonder if we should consider killing the circuit or throttling it or something
14:03:48 <nickm> but this is probably much better than nothing
14:03:57 <nickm> and we can tweak it more going forward
14:04:08 <asn> where were those instructions on pref & tor?
14:04:09 <asn> *perf
14:04:14 <asn> HACKING?
14:04:33 <nickm> I think HACKING
14:04:42 <nickm> I can try to dig more if they aren't there
14:04:55 <asn> some oprofile stuff seem to bethere
14:05:01 <asn> they were somewhere convenient i remember that!
14:05:48 <nickm> hm..
14:06:04 <nickm> maybe trac someplace
14:06:45 <nickm> (Shall I #endmeeting as we look?)
14:06:45 * asn tries from memory
14:06:50 <asn> ./configure --enable-coverage
14:06:51 <asn> is very nice
14:06:56 <nickm> it sure is
14:07:11 <nickm> btw, if you do figure out how, please send a patch to explain how in doc/HACKING :)
14:07:27 <nickm> #endmeeting