13:29:05 #startmeeting 13:29:05 Meeting started Wed Sep 2 13:29:05 2015 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:29:05 Useful Commands: #action #agreed #help #info #idea #link #topic. 13:29:06 gday 13:29:25 so, over the last week I did more or less nothing; on a family vacation! 13:30:03 I've been tweaking PostFreeze027, TorCoreTeam201509, and trying to get all the sponsorU doc stuff done 13:30:16 and trying to merge everything that really does go into 0.2.7 postfreeze, or postpone it 13:30:20 who's next with an update? 13:30:23 <- 13:30:35 stuck in my own personal hell of writing documentation for pts 13:30:50 tried to unfork tor browser tor, but there's still argument over behavior so I'm going to ignore that 13:30:56 and sit in my own personal hell for a while 13:31:11 the upcoming team meeting apparently includes more documentation 13:31:13 what's making the docs hellish? 13:31:18 I'm not a tech writer? 13:31:21 :P 13:31:23 ah 13:31:46 so here's a suggestion that sometimes works for me: don't think of it as writing documentation. Imagine you're jsut writing an email to me to explain what needs to be explained here 13:31:56 pretty often it turns out that that makes a good doc 13:32:03 past that, gitian sprouted tentacles and pseudopods and is assauting me when I'm trying to get japanese tor browser done 13:32:08 eeeep 13:32:09 I have builds for everythign but non-osx 13:32:23 nightly tor browser build target is broken due to some open ssl versioning thing 13:32:33 that I no longer give a shit about debugging, so doing an alpha build now 13:33:10 whatever, tired. unless there's stuff in this meething that concerns me I'm going to pass out 13:33:12 shortly 13:33:39 (hello meeting) 13:33:40 comment on that circiut dirtyness ticket once you and mike engage in single combat to the death or whatever and settle on behavior that's accpetable for everyone :P 13:34:12 (I think we should normalize client behavior to match what Tor Browser will do since that is our largest userbase subset, but that's just me) 13:34:35 (and if Tor Browser is doing something horrifically wrong, that's a big problem) 13:34:41 Yawning: Okay. So, see isa's email about what we're going to do post-meeting for triage and deliverables. ("For Core Tor meeting tomorrow, wed 9/2) 13:34:56 that's the big tasks for today 13:35:07 from what I recall 13:35:08 I'll either just do what mike wants, or come up with a compelling reason not 13:35:19 that e-mail sounded like 'write more documentation' 13:35:33 it's mainly "look at tickets, move them around" 13:35:47 let me know if I can help inspire / whatever with the docs. I too am doc-stuck 13:35:54 ok, i'll add a fixup commit that addresses the documenttion changes 13:36:01 you wanted for that commit 13:36:06 thanks! 13:36:07 since the behavior is what mike wants in the branch 13:36:17 (tomorrow sometime, dr's apt durring the day) 13:36:20 sure 13:36:51 hello should i update? 13:36:53 (internet here is crap) 13:37:06 go for it 13:37:19 last week i sent an email with the guard stuff 13:37:25 to tor-dev 13:37:30 it includes an algorithm for choosing guards 13:37:36 *poofs* 13:37:39 received good feedback. updated it.  13:37:56 great! is latest version at same URL? 13:37:57 for the past few days i've been looking at the consensus/voting code 13:38:00 I've been totally pushing to get to it 13:38:10 and keep hitting a brick wall 13:38:15 yes i think so 13:38:15 same url 13:38:16 probably a mental block 13:38:37 https://gitweb.torproject.org/user/asn/tor.git/tree/src/or/guardlist.c?h=bug12595 13:38:40 yeah it's a weird probelm 13:38:53 very heuristic. not much satisfaction for solving a problem. 13:38:57 but it's fun. 13:39:04 anyway, for the past days i've been looking at #prop250 13:39:07 and how it could be implemented 13:39:26 i opened #16943 for it 13:39:32 and i have started refactoring the relevant code 13:39:58 of course, "just make a new consensus flavor" was more complicated than it sounds 13:40:22 since most of the consensus code, assumes that the underlying doc is a networkstatus_t. 13:40:39 but we will need to introduce a new data strcutre for the shared randomness doc (or stuff it into ns which is terrible) 13:40:50 so some refactoring will need to be done probably. 13:40:55 oh speaking of consensus flavors, let's mention #16255 ? 13:40:58 i also did some sponsorr logistics 13:41:06 yes indeed that's my next thing :) 13:41:14 now that the guard stuff has progressed 13:41:17 great 13:41:18 this is the next annoying thing that i need to tend to 13:41:31 what timeframe do you think is realistic for that? 13:41:34 and that's my little-t-tor life for now 13:41:46 i think i will have tested my branch by mid of september. 13:41:51 i should have it ready. 13:42:11 it's already ready. but i need to test it. 13:42:36 and that's that. 13:42:43 thanks! 13:43:47 ok. mid-sep is pretty dicey for 0.2.7, but just on target for early 0.2.8. 13:43:57 anybody else around? :) 13:44:14 yes 0.2.8 is fine. 13:44:19 great 13:44:26 (thanks for not being That Guy!) 13:44:35 * teor is here 13:44:39 * nickm defers that ticket 13:44:40 * dgoulet here 13:45:07 hi hi hi! 13:45:10 what's new? 13:45:14 athena: ping? 13:46:28 I'm back since this morning, as you know, I'll be on the .fr timezone until the dev meeting, I'm almost done with my 2 weeks of emails I missed, I've started working on urgent tickets for 027 13:46:51 ok 13:46:52 trying to update myself also with all that has been going on 13:46:59 we have some freedom to decide what's urgent, so that's good 13:47:23 nickm: I'm basically using the PostFreeze027 ones that are assigned to me for now 13:47:31 and 201509 in the 027 milestone 13:47:32 ok, sounds good 13:47:47 if any seem not-worth-doing, let's sort that out 13:48:12 yup sure, so that's basically it for me :) 13:48:18 nickm: why have you enabled those statistics? 13:48:30 see the ticket; they appear to be harmless IMO 13:48:51 if there's a counterargument, I can revert, but please make that argument on the ticket 13:49:03 they are potentially illegal under european data protection law 13:49:36 how so? 13:49:52 teor: got a status thing? 13:50:06 I have created #16949 to group together changes that "make chutney easier". A whole bunch of them got merged just before the meeting, the rest are waiting for me. 13:50:14 excellent 13:50:24 I'm waiting for a larger block of time to do fallback directories work 13:50:42 qwerty1: ("the ticket" here is #15254) 13:50:47 teor: that'll be useful Tails (we have plans to use Chutney for our test suite). thanks for working on it! 13:51:35 teor: any problems if I set 'sandbox 1' in common.i ? 13:51:48 intrigeri: excellent, please let me know if there are any features you need, or any issues or annoyances 13:52:24 teor: sure. that'll be on anonym's plate. 13:52:31 nickm: well I'll never notice, I'm mainly on OS X :-) 13:52:54 But in my opinion, better to catch sandbox issues earlier 13:53:15 intrigeri: I saw somebody getting really confused a couple of weeks ago about whether you and anonym are different people 13:54:14 nickm: wow! that's because anonym doesn't do much conferencing I guess. thankfully anyone who reads Tails MLs should find the answer easily :) 13:54:26 (And I found a few issues with fresh OS X tor and chutney builds that are now fixed) 13:54:28 I'm done 13:54:30 qwerty1: just reopened #15254 for you 13:54:43 intrigeri: I notice you didn't actually deny it. :) 13:54:45 anybody else? 13:55:28 intrigeri: can you ask anonym to look at https://trac.torproject.org/projects/tor/wiki/doc/TorChutneyGuide ? 13:55:47 They may have comments or features or things I haven't thought of 13:56:09 teor: done yesterday when I noticed it (I follow all Trac changes). he'll start working on it in months, anyway, I think. 13:56:49 We have a target date of end October for some chutney work for SponsorS, but changes will continue 14:00:42 ok, if no more status reports today, let's move on to the stuff isabela wanted, plus other discussion topics! 14:01:16 ok 14:01:20 For the 0.2.7 release, I think it's a good idea for us to go over the stuff in the maint-0.2.7 ticket with a hard critical look, and see whether we can't kick out more things 14:01:39 alternatively, maybe we should find stuff that really really should go into our next alpha, and put out an alpha withing a few days 14:01:56 The PostFreeze027 ticket has way too much in it IMO; we should probably give it a hard look 14:02:23 and if anybody has more TorCoreTeam201509 stuff assigned to them than is reasonable, that would be good to know 14:04:30 I think that kicking #9925, #7978, #and #14846 out of 0.2.7 is probably okay at this point. 14:05:16 fine by me 14:05:52 ok, that's 2 needs_revision tickets in 0.2.7 left. Much better. 14:06:19 do you think all the stuff assigned to you in 0.2.7 is reasonable and good to try to do before the solid-freeze? 14:06:26 (refresher: we are in slushy-freeze now) 14:06:38 nickm: most of them are mark as normal 14:06:41 nickm: so #16389 has been merged but we keep it open because arma's question which I'm for but it seems that would be improvement in 028 14:06:53 nickm: so in that case #16381 could be close? 14:07:05 (both in postfreeze027 mode) ^ 14:07:36 isabela: ? 14:08:25 dgoulet: sounds fine; go ahead with both IMO. (remove tickets, change milestones, close, or whatever else is appropriate.) 14:08:35 nickm: ack 14:08:36 nickm: sorry someone called me 14:08:42 no worries 14:09:07 asn_: is #14957 easy ? 14:09:15 nickm: i meant to say that most are mark as normal while around 5 tickets are critical or major priority.. if any of that helps on making decision to what should be postfreeze 14:09:42 isabela: we should make sure that the priorities are reasonable too 14:10:53 nickm: yes 14:10:54 are there any openbsd folks who can help try out improvements to #16651 fixes ? 14:13:09 nickm: post freeze deadline is sept 15? 14:14:20 isabela: Sep 15 is the final deadline for small bugfixes and features that are _not_ showstoppers or regressions 14:14:29 we'll fix showstoppers or regressions for as long as it takes 14:14:31 nickm: I could ask on the mailing list - do we want them to test http://lists.nycbug.org/pipermail/tor-bsd/2015-August/000350.html without bufferevents? 14:14:34 but I hope it doesn't take too long 14:14:44 teor: I want them to never turn on bufferevents 14:14:50 indeed 14:15:06 teor: but I think I need to look at the patch again and figure out what the heck is going on 14:15:13 they seem to have got a little caught up in it 14:15:51 they really really shouldn't need those locking libraries 14:17:35 Seems to be an issue with libevent v1, they can't get v2 working on BSD 14:17:51 http://lists.nycbug.org/pipermail/tor-bsd/2015-July/000325.html 14:17:52 [Dv1 doesn't have event_pthreads, I thought. 14:18:18 Hm. It was working fine for me for like, forever. 14:18:25 and "it doesn't work" is a crappy bug report 14:18:43 which is why I was kinds hoping for a sensible person who can try it out and find what's actually working 14:20:26 Unfortunately, I don't have a BSD at the moment, we could ask around 14:21:07 yeah 14:22:12 There seem to be many variations of "I tried something unusual, and it didn't work" 14:22:24 another thing that makes this month complicated is that we have the release, we have sponsoS and then we have dev meeting 14:22:38 We may just need to identify an appropriate HOWTO build tor that works on the BSDs 14:22:42 when's thne sponsorS year1 deadline? 14:22:53 teor: yeah 14:23:09 teor: that would estabilish a baseline at least 14:23:20 Otherwise people try known broken configurations 14:23:25 like bufferevents 14:23:38 nickm: I am checking to see if is oct 1st or oct 31st 14:26:07 isabela: thanks! 14:26:19 isabela: (knowing the same thing for sponsor U would be valuable too) 14:26:41 I think sponsorS is under control, since we have a fair deal of latitude in revising our internal submilestones, 14:26:51 but we should have another look over it 14:26:58 also, it's the 1 hour mark. 14:27:05 shall we end the meeting part of the meeting and go on chatting? 14:27:06 #endmeeting