16:59:42 <nickm> #startmeeting
16:59:42 <MeetBot> Meeting started Mon Apr  9 16:59:42 2018 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:59:42 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:59:46 <nickm> hi everone!
17:00:03 <dgoulet> o/
17:00:07 <ffmancera> o/
17:00:19 <ahf> o/
17:00:59 <nickm> another week has passed
17:01:29 <nickm> in theory, 0.3.3 should be stable on Sunday.  That's probably not realistic, but I _would_ like to aim for a release-candidate on that timeframe if we can do it
17:01:34 <isabela> o/
17:01:39 <nickm> there are only 9 tickets left in 0.3.3, last I checked  :)
17:02:18 <ahf> neat
17:02:20 <dgoulet> for #25226 needs an arma4 but availability might complicates things :S
17:02:38 <isabela> !
17:02:44 <isabela> yes we are very close :)
17:03:04 <isabela> so if anyone has any task related to 033 - please prioritize it (task == code or review)
17:03:07 * kushal is reading the meeting notes.
17:03:10 <isabela> one goal will be to have an rc this week
17:03:19 <isabela> which is not a crazy goal :)
17:03:32 * dmr glanced at the meeting notes (nice format!); now lurks
17:03:35 <haxxpop> \o/
17:03:54 <nickm> that is to say -- if you have anything to do for 0.3.3, and you can do it, please do not do anything not for 0.3.3 until you have done it. :)
17:04:01 <Samdney> o/
17:04:13 <ahf> ack
17:04:14 <nickm> people with current coding tasks on 033 are: ahf, asn, catalyst.
17:04:27 <nickm> people with code-review tasks on 033 are: arma4, mikeperry, nickm, asn
17:04:36 <nickm> people with code-revision tasks on 033 are: ahf, isis
17:05:12 <nickm> If you are getting stuck, please get help from me or somebody else!
17:05:26 <ahf> i'm sorry, which one am i missing review on? i don't see it in my list
17:05:40 <dgoulet> list is fresh from this morning btw
17:05:45 <isis> o/
17:05:47 <dgoulet> (needs_review assignment)
17:05:57 <isis> where is the pad?
17:05:58 <nickm> ahf: I did not list you as needing to review an 033 thing
17:06:10 <nickm> isis: pad is at https://pad.riseup.net/p/1UThxeu4AT7T
17:06:11 <dmr> isis: https://pad.riseup.net/p/1UThxeu4AT7T
17:06:15 <ahf> nickm: ah, sorry! yes, my bad
17:06:18 <isis> aha, thanks!
17:06:34 <nickm> ahf: you're signed up to fix #25245 and revise #24782
17:06:40 <ahf> yep!
17:06:42 <ahf> those are on my list, cool
17:06:47 <ahf> i read the revision ad review for some reason
17:06:55 <ahf> s/ad/as/
17:06:57 <isis> for review, i'm still finishing up reviewing #25409, sorry sorry!
17:07:09 <isis> i sent it to travis and then totally spaced on it
17:07:30 <nickm> isis: don't worry; it's less urgent than all the 033 stuff. (See notes above; please give all the 033 stuff priority until it's done)
17:07:59 <nickm> isabela: Do we have more roadmap checking to do this week?
17:08:14 <isabela> nope
17:08:15 <isabela> :)
17:08:39 <nickm> ok, cool!
17:09:25 <isis> ack, will do
17:09:29 <nickm> When we _are_ doing 034, let's note that we have ~110 tickets in 034 , but only 36 days to do them.  So let's try to focus on the actual planned work for 034 until it's done, once we're done with 033
17:10:08 <nickm> For code review -- I see we've got a column "carried from previous week".
17:10:44 <nickm> Some of those are things that got 1 review, got put in needs_revision, and came back...
17:10:52 <nickm> some are things that got no review
17:11:36 <nickm> anybody need to trade reviews around?  I need somebody else to look at my changes for #23846, and I'd like to trade someone else for it.
17:12:20 <ahf> i can take that if you have modified some things in simone's original patch, i didn't spot anything other than what catalyst found related to rust and travis
17:12:34 <dgoulet> nickm: I'm a bit stuck at exchanging as the one I have to review is from you :S
17:12:39 <ahf> i don't think i need a trade, i only have one review this week and it's not very big and then the two 033 bugs
17:12:41 <nickm> ahf: I have.  I'll trade you for whatever :)
17:12:56 <nickm> ahf: okay, if you're sure
17:12:58 <ahf> no, you can just pass it on to me
17:13:01 <ahf> yep, i think so
17:13:08 <ahf> i'll take it
17:13:25 <nickm> ack. could somebody changereviewer for #23846 on the spreadsheet and on trac?
17:13:38 <ahf> i've updated the sheet now
17:14:19 <ahf> there, also updated trac
17:14:36 <nickm> One thing to remember about reviews: even if you can't finish a review same-week, it is beneficial to start it, since any initial comments you have may be helpful even if they aren't complete
17:14:54 <nickm> Rotations this week!
17:15:09 <isis> i'm coverity
17:15:17 <isis> mikeperry: you're CI
17:15:41 <isis> dgoulet: you're community liaison
17:15:57 <isis> catalyst: you're bug triage
17:15:58 <nickm> (woo, I get to have a rotation next week)
17:16:29 <nickm> anybody who did a rotation last week have anything to pass on for the folks this week?
17:16:31 <ahf> mikeperry needs to compile all our code :-P
17:16:34 <isis> my rotation last week (CI) i didn't actually need to do anything for
17:17:20 <isabela> i was thinking of kicking off via ml a retrospective about the rotation positions so far - document feedback and see if we should change/improve anything
17:17:28 <mikeperry> for CI I'm just supposed to look at https://jenkins.torproject.org/view/Failed+Unstable/ and file tickets and maybe fix small things?
17:17:58 <isis> mikeperry: also travis, see the #tor-ci channel
17:18:09 <nickm> fwiw, jenkins is a mess right now.
17:18:11 <ahf> isabela: i think that would be good. especially the community liaison role would be nice to have some documentation about and what people have done and what they think work
17:18:22 <isabela> ahf: cool!
17:18:55 <mikeperry> where is our travis instance?
17:18:57 <nickm> dgoulet: you said there was a coverity issue left over from last week?  Even if it's a false positive, we generally like to fix them when we can
17:19:04 <catalyst> mikeperry: also https://travis-ci.org/torproject/tor
17:19:24 <dgoulet> yes I couldn't log in coverity this morning, DNS was down for the site I think, it is back so I can update the CID
17:19:51 <catalyst> i still think Travis has far fewer "artifact" failures than Jenkins
17:19:51 <isis> mikeperry: i just hilight "failed" in #tor-ci and then take a peek whenever someone's build fails to see if they broke it or it was something else (usually if a bunch of people's builds are failing, it's because we merged something that broke something)
17:19:59 <nickm> dgoulet: okay.  For me, it's back up now
17:20:07 <dgoulet> nickm: yes I just noticed it is back :)
17:20:31 <isis> mikeperry: everyone has their own travis connected to their github, so there's like 10 travises or something, hence the channel
17:20:34 <nickm> Like, just add a cast to void so that we're ignoring that value on purpose.
17:20:44 <nickm> for CID 1434156
17:21:38 <nickm> dgoulet: (could you do that, since it's from last week and you have commit permissions?)
17:21:50 <dgoulet> nickm: sure
17:22:22 <nickm> for bug triage, we've been building up tickets without a milestone.  I think that might be because we're not potting "maybe" items in 0.3.4 any more, but we're nervous about putting them into unspecified.
17:22:26 <nickm> dgoulet: thanks!
17:22:30 <nickm> s/potting/putting/
17:22:36 <nickm> is that about right wrt bug triage?
17:23:14 <catalyst> which should we default to, then?
17:23:33 <dgoulet> hmmm
17:24:13 <dgoulet> well if we are nervous to put them in Unspecified, we should 034-proposed and discuss it here, else we do have guidelines (regression, security, fast-fix, roadmap)
17:24:50 <nickm> if something is clearly regression, security, roadmap, crash ... that should go right into 0.3.4
17:25:06 <nickm> for fast-fix, though, the guideline is that you only tag it fast-fix if you are going to do it yourself
17:25:14 <catalyst> shouldn't really serious security or regression issues go into 0.3.3?
17:25:36 <nickm> yes, if they are present there.
17:25:42 <dgoulet> backport keyword
17:25:48 <nickm> that too
17:26:21 <nickm> For unclear cases, let's use tor:unspecified and 034-proposed .
17:27:10 <nickm> does that seem like it would work?
17:27:16 <dgoulet> +1
17:28:02 <dmr> fast-fix := something that takes a really small amount of time?
17:28:09 <dmr> (not akin to hotfix, right?)
17:28:16 <dgoulet> huh I thought that was documented in TicketTriage page (about the proposed), seems it is not :S
17:28:19 <nickm> dmr: "I can do this in less than an hour, and I will."
17:28:26 <dmr> dmr: cool, thanks
17:28:28 <isis> i'm done with #25515, so i can take a second review ticket for this week
17:28:31 <ahf> WFM
17:28:32 <nickm> dgoulet: can you update the triage page?
17:28:35 <dgoulet> nickm: yes
17:28:45 <dmr> erm nickm: ^ (cool, thanks) - I'm tired apparently
17:28:54 <nickm> isis: ask again once you're done with all the 033 stuff :)
17:28:59 <dgoulet> isis: the review list right now is _all_ the tickets in review so all is assigned :)
17:29:23 <nickm> (there will be reviews to grab by the time you are ready for them)
17:29:31 <isis> ok, sounds good
17:29:54 <nickm> let's see.  Do we have more discussion topics....?
17:30:30 <nickm> Oh.  I wanted to amend our meeting process to say "pleaes have your update on the pad before the meeting starts"
17:30:33 <nickm> *please
17:30:46 <mikeperry> can we use the same pad for longer then?
17:31:01 <mikeperry> it is a pain to have to wait for the pad announcement before writing an update
17:31:02 <nickm> mikeperry: what if we send out the pad url on friday?
17:31:06 <mikeperry> sometimes yeah
17:31:08 <mikeperry> friday
17:31:19 <dgoulet> +1
17:31:19 <mikeperry> by monday I often forget everything I did anyhow :)
17:31:29 <pastly> I'll just mention that juga and I are going to meet weekly on thurs @ 22:00 UTC to talk about sbws. network-team@ is invited and I'll email saying so
17:31:31 <mikeperry> I mean if it were permanent, I could add stuff as I do it
17:31:31 <nickm> ok.  I will commit to doing the pad on friday; if I don't please just write in a text file
17:31:32 <dgoulet> mikeperry: thank you Timeline feature on Trac, saves me :P
17:31:55 <dgoulet> (although only works for Trac related things :S)
17:32:08 <nickm> isis: that okay with you too?  You're also Pacific...
17:32:28 <isis> dgoulet: yeah, i also use the timeline feature
17:32:56 <isis> nickm: yeah, i generally start working around 8:30 on mondays so that i'm ready for the meeting
17:33:11 <nickm> great
17:33:17 <nickm> then let's try that approach for next week
17:33:33 <ahf> cool
17:33:38 <nickm> looking over the pad, I see that most people's inline questions are questions for specific other people...
17:33:45 <catalyst> nickm: +1 on sending the pad URL on Friday
17:34:05 <nickm> catalyst: you're asking about review priorities.  Did that get answered already in our discussion about 033 vs 034 prioritization?
17:34:36 <nickm> catalyst: wrt rust review, one option is to ask the patch author to walk you through it, if you think that would help.
17:34:58 <nickm> sometimes that's been the only way I can get my trickier patches reviewed, if they touch code that no reviewer previously understood
17:35:32 <catalyst> nickm: that's a great idea; thanks!
17:36:00 <nickm> np! We're all here to help each other make tor :)
17:36:06 <mikeperry> dgoulet,isis: woah :mind-blown:
17:36:15 <nickm> isis asks about using  https://include-what-you-use.org/
17:36:26 <nickm> isis, what are the results of your brief experiment?
17:36:54 <nickm> teor: When you read backlog, please note that juga has a question for you
17:37:10 <arma4> ok i am done with my previous meeting that ran over. i have skimmed backlog and i see i should give some attention this week to #25226.
17:37:24 <dgoulet> arma4: neat :)
17:37:30 <nickm> arma4: thanks!
17:38:06 <nickm> isis, asn: mikeperry has a question for you about prop291.
17:38:24 <mikeperry> also arma4
17:38:55 <nickm> mikeperry: Is prop291 on-roadmap for 034?  If not, we need to talk. :)
17:39:14 <mikeperry> section 2.4 deals with concerns that isis brought up. section 3.1 deals with concerns arma4 brought up
17:39:26 <nickm> dgoulet: want to talk about events any time tomorrow before noon?
17:39:49 <mikeperry> nickm: how we handle prop291 governs how we decide to handle #24487
17:39:54 <dgoulet> nickm: sure, I'll be online
17:40:05 <mikeperry> and other vanguard path restriction things
17:40:22 <arma4> is prop291 most-up-to-date in torspec git? or is there some other more-up-to-date thing?
17:40:42 <nickm> hmmmm....
17:40:45 <mikeperry> the email to tor-dev should be the same as the stuff in torspect.git
17:41:11 <isis> so the iwyu thing was interesting, and i think if we used it right, we'd gain some freedom from literally everything including or.h, but the downside was that i'm worried we would accidentally remove/change headers needed for compatibility on weirder systems
17:41:38 <nickm> mikeperry: okay, then can I ask you to make tickets for 291 as appropriate in 034, and assign them as 034-roadmap-subtask children of the appropriate roadmap parent-ticket?
17:41:46 <isis> the changes from running it, plus some brief manual fixups as i sorted through it are here: https://github.com/isislovecruft/tor/tree/feature/iwyu-test
17:42:13 <nickm> isis: cool!  I think next step there is open a ticket for 035 (where we are likelier to do modularity stuff)?
17:42:53 <mikeperry> nickm: the same types of arguments apply to our options for #24487 and other cases. but section 1.2 could use an 034 ticket
17:44:03 <mikeperry> (I think arma4 wants something like 1.2 for all of our path restriction torrc options)
17:44:04 <nickm> isis: I am very worried about removing includes of system headers; those tend to be strangely platform dependent
17:44:13 <isis> (note that that branch totally doesn't compile, and i'm pretty sure it's broken on windows)
17:44:24 <nickm> isis: ok, let's put it in 035
17:44:35 <nickm> if we get all the 034 stuff done in thenext 36 days, we can look at it early :)
17:45:02 <isis> nickm: yeah, that's the part i'm extremely worried about, also it removed the guard defines (e.g. #ifdef HAVE_SYS_SOCKET_H)
17:45:06 <nickm> dgoulet: great! Let's ping each other when we're awake and online
17:45:25 <isis> apparently there is a pragma we could put on those lines to tell the thing we need them
17:45:28 <dgoulet> ack
17:45:54 <nickm> mikeperry: wrt the minimal version of this for 034, though, are you okay opening the tickets?
17:46:21 <isis> https://github.com/include-what-you-use/include-what-you-use/blob/master/docs/IWYUPragmas.md
17:47:15 <nickm> isis: sounds like a potentially broken tool. Maybe we can make it only do things to our headers and ignore system headers though
17:47:30 <mikeperry> nickm: yeah I can open the tickets. though I think what goes in to 034 may depend on how we ultimately want to do prop#291. if we abandon all path restrictions for everything, then we probably don't need to do #24487 at all, ever
17:47:32 <nickm> isis: please remember to open that 035 ticket, though?
17:47:45 <isis> nickm: ok!
17:48:09 <nickm> mikeperry: then open "make choices for prop291" as a ticket, and leave the other ones unopened?
17:48:26 <mikeperry> nickm: for example. there's some other related things like that that have come up in conversation but have not yet become proper bugs.
17:48:29 <mikeperry> nickm: ok
17:48:43 <nickm> arma4: I am inclined to veto "remove all path restrictions" unless you want to argue for it strenously.
17:48:59 <nickm> I feel that deciding this early might make mikeperry's life easier
17:49:06 <nickm> are you going to push for "remove all path restrictions" ?
17:49:20 <nickm> mikeperry: thanks!
17:49:22 <nickm> isis: thanks!
17:50:17 <nickm> mikeperry: unless arma4 says "yes", I'm going to suggest we assume that path restrictions will still exist for at least another year, and move forward on that assumption, if it helps.
17:50:36 <arma4> i am unprepared to multitask with this topic too
17:50:44 <arma4> i should read the proposal again and decide what i think
17:51:55 <arma4> if we're trying to decide whether to move from one guard to two, i think we should be broad in considering what we're trying to solve, and what might solve it
17:51:59 <mikeperry> arma4: ok. please be sure to read section 3.1. that discusses the path restriction point. (maybe 3.2 also)
17:52:01 <nickm> so, you don't know if you are going to argue for us to remove all path restrictions in 0.3.4,  which will feature-freeze on May 15?
17:52:31 <nickm> I suggest that if you're going to argue for this, we will benefit from you doing so sooner rather than later :)
17:52:31 <Hello71> re: headers, a lot of them are almost certainly unnecessary on modern systems. we're not at openssl levels of unnecessary compatibility yet, but it would be nice if we actually had a list of supported platforms and references to specific bugs that some headers fix
17:52:40 <arma4> i don't currently plan to argue for that. i didn't even know it was a thing somebody thought i might argue for.
17:52:53 <nickm> okay. So that sounds like a "no"?
17:52:59 <nickm> mikeperry: did that sound like a "no" to you?
17:53:33 <nickm> At least, it sounds like something we shouldn't have to plan around?
17:54:21 <mikeperry> I would also rather not have us default to "no decision is made about anything wrt 291". I find those failure modes frustrating. because they are actually a decision that has consequences wrt development, too..
17:54:48 <arma4> are we currently planning to move from one guard to two, in 0.3.4?
17:55:08 <nickm> "we" aren't planning anything because mike is waiting for feedback on his proposal.
17:55:19 <dmr> mikeperry: just wanted to throw in my 0.02: I read over prop#291 a few days ago and thought it explained context very well for someone with only _some_ familiarity with the problem, so props there :)
17:55:24 <nickm> But mikeperry hopes the answer is yes
17:56:11 <mikeperry> right
17:56:16 <nickm> mikeperry: If it comes to that, I will have to make an up-down decision on 291 on my own, but I would rather that it not come to that.
17:56:25 <arma4> ok. i will proceed with my plan then ("read the proposal, decide what i think")
17:56:34 <nickm> Let's try to get mikeperry feedback on the proposal by end-of-week?
17:56:52 <nickm> and if there isn't a clearcut answer by next monday, let's schedule a meeting to resolve whatever issues remain.
17:56:58 <nickm> mikeperry: sound plausible to you?
17:57:07 <mikeperry> ok sounds good.
17:57:11 <nickm> great!
17:57:15 <nickm> anything left for the last 2 minutes?
17:57:29 <nickm> (I feel like we're getting more productive at these every week)
17:57:33 <mikeperry> I will still file a ticket and try to capture the decision points and consequences wrt trac tickets, if I can do so succinctly :)
17:57:54 <nickm> ok, without hearing any more... thanks, everybody!
17:57:58 <nickm> #endmeeting