13:35:04 <nickm> #startmeeting
13:35:04 <MeetBot> Meeting started Wed Jan  7 13:35:04 2015 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:35:04 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:35:05 <nickm> hi teor_
13:35:19 <teor_> hi nickm
13:35:50 <nickm> dgoulet: Yawning: athena: anybody else here?
13:36:12 <athena> hi meeting
13:36:15 <nickm> hi athena
13:36:34 <nickm> so, we've got 23 days left in my initial timeline for doing stuff in Tor 0.2.6.x...
13:37:12 <nickm> and we've got about 135 open tickets on the milestone. :)
13:37:26 <nickm> asn: (you around?)
13:37:39 <nickm> I've got a feeling that some of our deferrable stuff will wind up deferred.
13:37:59 <asn> hello
13:38:03 <asn> i am indeed around
13:38:05 <asn> semi-brained
13:38:23 <nickm> asn: One thing I'm wondering is where we stand on guard fixes/improvements for 0.2.6
13:38:30 <asn> ah
13:38:41 <asn> so afaik the next step is #9321 and its python script
13:38:58 <asn> weasel started running my python script on tor26
13:39:05 <nickm> how far is all that from being mergeable?
13:39:15 <asn> #9321 is not too fara way from being mergeble
13:39:22 <asn> but I would erally like someone else to review it too
13:39:30 <asn> except from me and you, that is.
13:39:43 <asn> that's where we stopped last time. we didn't know who would review it.
13:39:46 <nickm> ok
13:39:51 <asn> i would like #9321 to get merged.
13:39:55 <nickm> me too
13:40:07 <asn> but it should be reviewed carefully since it touches the path selection prob code
13:40:13 <asn> so i'm not rushing it.
13:40:18 <nickm> yeah.
13:40:24 <nickm> Hm. Who's been poking at that code sometimes?
13:40:31 <asn> ehm
13:40:34 <asn> mike perry?
13:40:34 * teor_ 
13:40:41 <asn> ah teor_!
13:40:50 <nickm> teor_: would you have a while to look over #9321 as well?
13:40:59 <asn> nickm: and the respecticve proposal?
13:41:04 <nickm> yeah
13:41:04 <asn> nickm: feel free to ask any questions you have :)
13:41:05 <teor_> just pulling it up now
13:41:12 <asn> teor_: thanks mate
13:41:18 <asn> nickm: after #9321 gets merged.
13:41:25 <asn> nickm: weasel will be able to pull it in his dirauth
13:41:36 <asn> nickm: and point his tor to the guardfraction file that he is generating every hour now
13:41:40 <asn> nickm: and it should work (tm)
13:41:49 <asn> nickm: then if 2 more dirauth ops do it too, the whole system should start working
13:41:59 <asn> assuming that all clients are also using it. which is a consensus paramter.
13:42:35 <Sebastian> asn: you wanted to ping me when you've written instructions. Did that happen yet?
13:42:54 <Sebastian> otherwise, standing by to set up the python stuff
13:42:57 <asn> no no
13:43:14 <asn> havent written that yet. been sick for days.
13:43:19 <asn> will need to work on sponsorr stuff next.
13:43:21 <asn> and then some hs stuff.
13:43:26 <asn> and then guardfractions tuff.
13:43:27 <Sebastian> ok.
13:43:27 <nickm> Sebastian: oh; do you think you're going to finish the remove-old-naming-code stuff in January, or should I put it off till 0.2.7?  No pressure either way
13:43:30 <nickm> just trying to plan
13:43:49 <Sebastian> I hope I will finish it by this weekend, because if I don't then I'm unlikely to get it done in January.
13:44:06 <nickm> ok. either way is fine. 0.2.7 is not far away, assuming we get 0.2.6 out the door.
13:44:11 <Sebastian> yeah.
13:45:18 <nickm> athena: how does this sound for a schedule: freeze features at 31 Jan, stop fixing non-regression/security bugs on valentine's day, and fork maint-0.2.6.x on March 1?
13:46:01 <nickm> I think that the thing we did for 0.2.5 where we said that feature implementation needed to be done by date X, and we could take as long as we needed to merge features, had some failure modes
13:46:14 <athena> nickm: i think that's okay with me
13:46:21 <nickm> cool
13:46:50 <nickm> that does mean that need to dedicate some time to serious patch review though
13:47:31 <athena> yeah
13:48:03 <nickm> we should also identify stuff in needs_revision that we want to merge, and maybe fix it ourselves if the original author isn't likely to have the time
13:49:02 <athena> yeah (the sockssocket thing?)
13:49:11 <nickm> yeah, like that.
13:49:13 <nickm> Maybe we should each identify 5-10 priorities to do by end-of-jan, do as many as we can, and that's where we're at.
13:49:33 * dgoulet here!
13:49:35 <nickm> (And if we come up with >10 must-do priorities, then let's prioritize those)
13:49:36 <nickm> hi dgoulet !
13:49:37 <Yawning> mlerph
13:49:40 <athena> ok
13:49:41 <nickm> hi Yawning !
13:49:42 <Yawning> oops fell asleep
13:49:56 <nickm> Yawning: sleep is good for body and mind
13:51:16 <nickm> dgoulet, Yawning: How does the schedule I suggested above and the idea of identifying <= 5-10 things each to try to do in Jan sound to you?
13:51:23 <Yawning> sure
13:51:42 <Yawning> sorry trying to reboot my brain
13:51:50 <nickm> great.  Does everybody think "today" is a good timeframe for identifying said january things to do?
13:52:05 <nickm> we don't need to overthink it
13:52:09 <Yawning> it's night >.>
13:52:11 <dgoulet> nickm: sure! the spreadsheet could be a good start for that
13:52:25 <nickm> dgoulet: sure, if you like
13:52:26 <Yawning> but yeah relink the sheet
13:52:48 <dgoulet> https://docs.google.com/spreadsheets/d/1_fsXOPmbqBuoJD1ufVQO0Ehy21uJyE5TWqdQWkfRxVo/edit?pli=1#gid=0
13:52:51 <nickm> https://docs.google.com/spreadsheets/d/1_fsXOPmbqBuoJD1ufVQO0Ehy21uJyE5TWqdQWkfRxVo/edit#gid=0
13:52:56 <nickm> yeah, that
13:52:58 <Yawning> ty
13:53:11 <nickm> the cb columns are priority divided by hardness
13:53:24 <nickm> I forget what it stands for
13:53:37 <nickm> I can't spreadsheet
13:54:59 <nickm> also remember that spreadsheet doesn't have newer tickets.
13:56:34 <teor_> Well, this is amusing - I must be using a French exit: "Impossible d'ouvrir ce fichier car JavaScript n'est pas activé dans votre navigateur. Activez-le, puis rechargez la page."
13:57:02 <dgoulet> right true... I'm also open to a better way of triage
13:57:11 <dgoulet> teor_: oh so you exit in France? :)
13:58:05 <teor_> Only today, dgoulet, and only for this app for the next 10 minutes... or something
13:58:33 <dgoulet> teor_: if you need a translation, it says that you need to enable Javascript :)
13:59:06 <Yawning> the people here know about this tentp thing right? >.>
13:59:19 <Yawning> I guess nickm and I have been talking about it for a while
13:59:34 <nickm> (the name isn't final)
14:00:22 <Yawning> (we has drafts for some parts of it even https://github.com/nmathewson/tentp-draft)
14:00:23 <teor_> Yeah, I got that: I can't open <something> JavaScript isn't active in your browser. After activating, please refresh the page.
14:00:25 <dgoulet> "tentp" is the nottls protocol?
14:00:30 <Yawning> yah
14:01:01 <Yawning> need to write docs for key exchange still, the rest of the stuff should be preliminarily specified
14:01:20 <Yawning> (though the client auth is something I made up today, so not sure if it's quite right yet)
14:01:45 <nickm> yeah; we've got a lot to do
14:02:08 <nickm> I'm not sure about the handshake stuff so far; I think that the one you did at first is cool, but we might need other variants of it
14:02:17 <Yawning> yeah
14:02:21 <nickm> and the number of variants we might need is big enough that maybe we should do one that can get used in a few ways
14:02:26 <Yawning> I'll think more about it
14:02:54 <Yawning> a cleaned up version of my v0 one may work well enough for testing/eval pruposes
14:03:19 <Yawning> though I'm 99% sure we're well within our implementation complexity/performance targets for this
14:04:34 <nickm> hm.  So, what else to figure out this morning?  I'm working on my top-N list, which is now a bit past 10.  I'll have to pare it down
14:06:22 <Yawning> *looks at the supercop output*
14:06:50 <dgoulet> nickm: something I can think of
14:07:37 <dgoulet> the performance profiling for 0.2.6, do you have one from last time you profiled 0.2.5 so we can have a baseline to compare?
14:08:18 <nickm> the "profile 0.2.6" ticket links to the ticket where we did 0.2.5 profiling.
14:08:39 <dgoulet> oh does it
14:08:53 <dgoulet> cool!
14:10:24 <Yawning> 1.48 cpb
14:10:31 <nickm> I'll be around for the next N hours to talk about priorities if anybody woudl like
14:10:37 <nickm> Yawning: chacha20+poly1305?
14:10:43 <Yawning> (576 byte message i7-4770K)
14:10:47 <Yawning> that's just chacha20
14:11:09 <Yawning> floodyberry's avx2-ed implementation
14:11:49 <Yawning> he says ctr-aes256 on the same box does ~1.1cpb
14:12:15 <nickm> I think we're fine with chacha
14:12:28 <Yawning> yah
14:13:17 <nickm> anything else for the meeting today?
14:13:53 <asn> i'd prefer to make jan priorities in 2-3 days from now. i will have a better view of the situation by then.
14:14:08 <asn> wrt 13:52 < nickm> great.  Does everybody think "today" is a good timeframe for identifying said january things to do?
14:14:29 <asn> it will most likely involve sponsorr, HS stuff and guardfraction in general.
14:14:42 <asn> my status report has some ideas for january.
14:14:45 <Yawning> (yay he has optimized poly1305 as well)
14:14:52 <Yawning> 0.96 cpb
14:15:06 <teor_> Please feel free to throw me 1 other thing before end of January as well as #9321, but only one at a time :-)
14:15:10 <nickm> asn: ok.   Is it okay if I keep bugging you about it? :)
14:15:15 <asn> nickm: it's fine.
14:15:20 <Yawning> zoom zoom
14:15:25 <teor_> (Oh, and I am a few days behind on emails)
14:16:14 <nickm> ok.  brb 5-10 minutes; phone call
14:16:15 <nickm> #endmeeting