13:31:11 #startmeeting 13:31:11 Meeting started Wed Jul 29 13:31:11 2015 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:31:11 Useful Commands: #action #agreed #help #info #idea #link #topic. 13:31:13 hi all 13:31:26 hi nickm 13:32:04 , isabela, all 13:32:19 hi ... teor, isabela ... yawning, asn_ , karsten, athena, dgoulet ? 13:32:51 hello! 13:32:52 hello 13:32:55 hi 13:32:57 hello! 13:33:03 still sick, still behind on papwerwork 13:33:08 cuz I'm a bad person 13:33:11 :( 13:33:49 so, over the last week I put out a tor release and did lots of paperwork and stuff 13:34:02 hi meeting! 13:34:04 hi athena ! 13:34:23 and I've been trying to merge patches and I wrote a couple of proposals and i met up with a couple of ntru people 13:34:42 nickm: their code needs uh 13:34:45 some changes 13:35:00 which code, how so? 13:35:02 unless they fixed the stuff from that cisco dude's paper 13:35:11 their key generation algorithm is fucked 13:35:28 so the security level provided is 2^80 vs a quantum puter 13:35:47 (using a SHA1 DRBG for kegen is a terrible idea kids, don't do it) 13:36:23 and I've also been scrambling to get as much stuff done by end-of-month as I can 13:36:43 Yawning: if it's just drop-in-a-better-drbg, that's in our area of knowledge 13:36:44 nickm: https://eprint.iacr.org/2015/676 13:36:46 how are others doing? 13:37:12 ehm hello 13:37:17 i came back from the HS meeting in DC a week ago 13:37:24 we wrote lots of proposals and stuff. 13:37:40 i'm still a bit disoriented, but I'm getting back into Tor development. 13:37:57 i finished the important stuff i hadn't done because of the meeting 13:38:04 which is that annoying guardfraction bug #16255 13:38:20 i made a patch for that that hopefully fixes it. it really needs review from someone who knows voting better than me. 13:38:31 i'm also under the process of reviewing donncha's #14846 13:38:43 and i still need to write the shared randomness proposal with dgoulet 13:39:00 and also read the thread of the guard discovery proposal and update it appropriately 13:39:20 i also have 2-3 tickets on my review pile. as well as some more review of onionbalance. 13:39:33 then i would work on the guard algorithms. 13:39:47 (which are pretty much necessary for doing guard discovery defences or anything else guard related) 13:39:53 that's my tor life currently. 13:39:57 next? 13:40:09 drowning in paperwork, proposals, and meatspace errors - no coding, sadly. 13:40:12 pretty much the same as asn_ (except for guard and ticket review), I'm still working on R stuff from that meeting (that is still ongoing over there). So my tor dev. side is a bit slow in the coming days unfortunately. 13:41:11 (I'll go after dgoulet) 13:41:22 * dgoulet done :) 13:41:26 oh 13:41:33 wrote/merged some trivial patches 13:42:04 i've got the ability to classify the dir_indirection_t and source of incoming dirauth connections now; building a mechanism to count them by origin and then we'll have DoS-resistant dirauths for #4581 13:42:25 burried in paperwork and meatspace errors (my meat is failing due to virial infection, and my government is failing due to EFASCISM) 13:42:30 that's about it 13:42:41 I'll catch up on paperwork by the end of the week 13:42:43 since I need to 13:42:54 not sure when meat will be fixed 13:43:09 athena: got any ETAs on that patch? 13:43:12 (sorry Yawning; thought you'd finished there) 13:43:26 probably next week; not too much left now 13:43:45 athena: it's fine, done etc :P 13:44:02 I'm also glad to see that 0.2.7.2 didn't immediately explode with all teh crypto stuff that cot changed 13:44:02 great. BTW, when you do your timesheet, remember to call the DoS stuff "Sponsor U" 13:44:03 *got 13:44:08 nickm: okay, sure 13:44:20 s7r: yeah, though apparently s7r has been filing some bugs 13:44:29 err, Yawning ^ 13:44:49 nickm: what I've seen looked mostly like teething pains from the key alg move 13:45:15 rather than "we screwed a lot wioth the openssl code" "I replaced the ntor keygen code and the ed25519 code" 13:45:18 :P 13:45:18 yeah I'm also happy that in 027, HS didn't explode :) 13:45:54 cuz I think this is the first time that the openssl stuff in particular hit non #yolo users 13:46:35 My next task is #15775 - whitelisting fallback dirs for our first release, ensuring clients won't DoS old relay IPs, and drafting a "please sign up as a fallback dir" email, along with a wiki page explaining why we need fallbacks. A lot of work, happy to delegate. 13:47:12 Or link to existing resources instead of a writing a wiki page from scratch 13:49:00 ok. anybody else want to report in, or should we forge ahead? 13:49:45 ok, any topics to talk aobut today other than "here comes august!"? 13:50:08 "paperwork makes yawning sad" 13:50:20 I'll swap you 13:50:51 is your country on fire? 13:51:28 My country is blocking websites and keeping records of every contact people make on the Internet 13:51:45 sounds about as bad 13:52:02 my country is trying to get an excuse to play Global Thermonuclear War with China 13:52:09 but they don't block websites as far as I know 13:52:20 Nah, we're just serially invading the Middle East 13:52:39 (I think we can move on) 13:53:05 There was an article that came out about the electricity / environmental cost of a bitcoin transaction 13:53:15 Has anyone done anything similar for Tor? 13:53:25 don't think so 13:53:32 should we talk about august stuff? 13:53:35 the cyrpto involved for tor isn't that expensive 13:53:38 isabela: probably 13:53:41 :) 13:53:53 did everyone saw my email yesterday (sorry for the short notice) 13:53:54 is this a "I shouldn't have ignored that e-mail" moment for me? 13:53:54 please, I'm just generating random ideas 13:53:56 :P 13:54:07 hehe 13:54:17 well 13:54:42 pretty much a couple weeks ago we spoke about using the same method TBB is using to tag tickets that needs to be done for the month 13:54:43 (I am not on the private list yet, people want to meet me at a conference) 13:55:00 (The private list is overrated) 13:55:16 I listed a few cases trying to explain how to elect these tickets 13:55:20 https://trac.torproject.org/projects/tor/query?status=!closed&keywords=~TorCoreTeam201507 13:55:22 ah I see 13:55:23 i read the email 13:55:27 here's what's left in July 13:55:30 yes 13:55:39 and teor has already added a ticket for august 13:55:45 https://trac.torproject.org/projects/tor/query?status=!closed&keywords=~TorCoreTeam201508 13:55:48 :) 13:56:08 so what, should IU just tag stuff I'm planning on doing? 13:56:11 I have work-in-progress for #15055, #12748, and #13996 13:56:24 so they seem like clear things-to-do-for-August 13:56:35 err, not #13996 13:56:42 #13966 is the one I have wip for 13:57:05 Yawning: yes, not only what you wants to do, but also things you know should be done even if you can't do them 13:57:21 * nickm moves those 3 tickets 13:57:31 Yawning: the idea is that as we move along the month, whenever someones is looking for something else to do, the person should look at this list first 13:57:40 I think I can maybe manage to have #16389 and #16274 done by August 1st 13:57:43 ah I see 13:57:57 dgoulet: ok, then let's not move them now? 13:58:04 I believe you; there's not so much there 13:58:10 nickm: yeah let's keep them july for now 13:58:41 I've moved #15775 to August, it's big and not going to get done in 2 days 13:58:52 Didn't somebody start on #2555 ? 13:59:02 nickm: special 13:59:03 I thought I saw some tor-dev chatter about that 13:59:14 *adds* #16533 13:59:23 nickm: special is the owner of that ticket :) 13:59:23 (I'll just do it) 14:01:15 oi asn 14:01:41 asn: is there a ticket for that "guard design writeup" thing that you wanted to do? Will it happen in August? If so let's tag it. 14:01:48 (We're tagging things with TorCoreTeam201508) 14:01:49 asn: the guard stuff can be added to august too - also the clearning thing you mentioned, even if you dont have time to do it 14:02:37 orCoreTeam201508 oops >.> 14:02:44 * isabela keeps updating the query and seeing more tickets :) 14:03:16 ok 14:03:20 ! 14:03:26 ok let's try this one. tor +oftc is going crazy. 14:03:32 so what are we doing now? 14:03:46 10:02 < nickm> asn: is there a ticket for that "guard design writeup" thing that you wanted to do? Will it happen in August? If so let's tag it. 14:03:49 10:02 < nickm> (We're tagging things with TorCoreTeam201508) 14:04:01 i can use #12595 for it 14:04:19 i definitely plan to start on august yes. 14:04:35 *adds some other stuff* 14:04:46 i can _probably_ have an initial version done by the end of august. 14:04:50 not sure if I want to overcommit here 14:04:51 asn_2: aslo, what was the cleaning guard stuff you mention via email / even if it's not you necessary doing it 14:04:54 Hi, there. I'm writing a pluggable transport process manager.. Something like fog minus the chaning and with full support for cli/serv.. One question about the pt-spec, is it allowed for a transport to be both client and server at once? the obfs4 code does not allow it but I did not see the spec saying that it's forbidden (?) 14:04:58 since I play cleanup an ok amount 14:05:13 a00001: wow! 14:05:15 a00001: every single pt will breat probably 14:05:22 *break 14:05:23 isabela: hm 14:05:25 asn: thanks; I modified the ticket 14:05:26 isabela: cleaning guard stuff? 14:05:31 maybe i need to re-read that email 14:05:32 so "no, it's not allowed" 14:05:34 asn_2: let me find your email 14:05:44 Of course, to implement that proposal we will first need to clean up 14:05:44 the whole guard code :) 14:05:46 there it is ^ 14:06:14 I think we're ready to move on #13338 in august too, and do more of #16505. 14:06:15 i meant, that to do further defences/improvements on the guard subsystem 14:06:26 we probably need to figure out the algorithms/API first, and rewrite/clean up the current code 14:06:28 hello did i miss somehting 14:06:29 ? 14:06:29 before we can do more advanced mods to it. 14:06:48 (which is the real asn?) 14:06:51 (wow nvm asn) 14:06:55 (i'm the real asn!) 14:07:07 nickm: yeah kick 13338 into august 14:07:08 asn_2 is the real asn. 14:07:14 I think a rough consensus has been established there 14:07:27 but I won't get around to doing all the stuff till august sometime 14:08:33 there are 10 things left in the july ticket 14:09:10 athena said she'd finish #4581 in a week, probably, so i'll move that one to august 14:09:20 #15463 deferable 14:09:41 plus no one's complained after we fixed the easy ways t sploit that 14:09:42 nickm: 3 of those are wainting for review 14:09:46 tto august or never? 14:09:53 august for now 14:10:00 I have some other stuff I want to add there 14:10:19 (an token bucket for INTRO2 processing, with drop semantics) 14:10:43 (yes, it's another knob, but it'll keep tor from crumpling) 14:10:59 #16255 I should try to review RSN, so leaving in july for today 14:11:12 ack 14:11:26 i expect to spend some of august on guardfraction stuff again 14:11:32 since there might be mods i need to do to the #16255 branch 14:11:35 or something. 14:11:36 (unless people think i'ts a bad idea) 14:11:42 asn: should I move that to august? 14:11:47 Yawning: seems plausible to me 14:12:01 nickm: depending on the output of the review, maybe. 14:12:53 ok, moving 14:12:56 (seems like tor+irc works again. i log off from my other client.) 14:13:50 nickm: hrm you flagged #16274 to august but I can do it for july I think so I'll keyword back ok? 14:13:55 ok 14:14:19 and that leaves #2665 and #2555 as the ones in July still that nobody has said "yeah, august" for 14:14:25 For #2665 we can ask special when 14:14:28 #2555 i think special has a proposal almost ready 14:14:34 special: is #2555 realists for July? 14:14:37 for #2665, I dunno. 14:15:03 the current #2555 proposal is here: https://gitweb.torproject.org/user/special/torspec.git/log/?h=xxx-direct-onion 14:15:18 but there is probably a reason special hasn't posted it yet. 14:15:29 we should ask and find out. :) 14:15:35 xwrits, brb 5 mins (unless y'all don't need me for anything) 14:15:44 special should wake up in a few hours :) 14:17:07 cool 14:17:16 this is looking great folks! 14:17:18 Does everybody have enough stuff to do in August? I feel like I do, but am not so sure of other folks 14:17:25 probably many folks do 14:17:29 i also really wanted to do #12538 14:17:34 but i doubt i will have time for it in august 14:17:42 ah yes ^! 14:17:45 asn: add it for now and we will move it to sept if you dont 14:17:47 my plate is definitely full for august. 14:17:53 ah 14:17:55 ok 14:17:58 I want to finish up #13339 too. 14:18:21 august is looking interesting for me as well for now 14:19:07 I should also take on all those tickets s7r has been opening 14:20:00 * isabela hopes that this will be helpful on making sure important things are done and we always have a place to pick up work from.. and of course, helping you know what is coming and how things looks like for the month 14:21:48 ok xwrits is letting me type again 14:22:04 apart from all my paperwork does anything else need my immediate attention 14:22:27 not immediately fwict 14:22:32 ok 14:22:55 also fwiw, we have 4x tickets in "new" that we should probably make "accepted" or kick into a later release 14:23:06 I'm happy to make the calls I can make on that. Anybody want to help? 14:23:50 shoot 14:23:50 which one are these? 14:24:36 these : https://trac.torproject.org/projects/tor/query?status=new&milestone=Tor%3A+0.2.7.x-final&group=status&max=200&order=priority 14:25:18 *looks* 14:25:49 ok what,s the procedure here? tag them with a date keyword or Version or ? 14:26:40 If triaging them in, make them "accepted". If triaging them out, change the milestone to 0.2.8 or 0.2.???. If not sure, leave as "new" 14:27:11 i know sysrqb spent some hours thinkiong about #13207. not sure what was decided. 14:27:33 #15744 seems deferrable for now. dgoulet increased the limit there. and the new values seem ok. 14:28:08 dgoulet: you agree? 14:29:03 asn: I don't have a compelling arg. to change that limit to right now it's rand(16k,32k), I'm fine with it for now 14:29:08 nickm: ^ 14:29:22 dgoulet: then close or defer? :) 14:29:30 *accepts some stuff* 14:29:39 #15714 and #15746 also seem related, dgoulet. 14:29:52 i would be fine with deferring all of these for now. (?) 14:29:57 asn: +1 14:30:46 dgoulet: wanna do the honors 14:30:47 ? 14:30:50 asn: sure 14:30:57 I think the meeting part of the meeting may be over. :) 14:31:29 i made #16387 and i admit it's too generic. 14:31:41 asn: I think #13207, we can totally do it for August, sysrqb did most of the work 14:31:46 * dgoulet flagging 14:31:57 ok 14:33:11 #endmeeting