17:59:20 <nickm> #startmeeting My clock says it's 1800 UTC again! Network team meeting, 5 Feb
17:59:20 <MeetBot> Meeting started Mon Feb  5 17:59:20 2018 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:59:20 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:59:25 <nickm> Hi folks!
17:59:36 <ahf> o/
17:59:52 <nickm> hi ahf, dgoulet, asn, isabela, isis, mikeperry, armadev, teor, Sebastian, and catalyst!
17:59:55 <dgoulet> hi
18:00:03 <nickm> and to whomever I left out.
18:00:12 <nickm> not trying to diss you
18:00:19 <nickm> https://pad.riseup.net/p/yIE9hjL8nsrc is our pad this week!
18:00:19 <asn> yo yo
18:01:10 <isabela> oi
18:01:26 <ln5> hi
18:01:41 <nickm> hi ln5 !
18:01:48 <asn> woo ln5
18:02:05 <asn> ln5: wanteed to ping u today and ask for any sort of info about bw auths. i know uve been in contac twith aagbn
18:02:08 <asn> aagbsn
18:02:19 <ln5> asn: that's mostly why i'm here today
18:02:20 <asn> legendary aagbsn
18:02:26 <nickm> everybody please fill in your stuff on the pad , and read other people's stuff too
18:02:35 * asn reads & writes
18:02:41 <nickm> https://pad.riseup.net/p/yIE9hjL8nsrc
18:04:19 * isis reads
18:05:06 <asn> nickm: ha cool. class on coq :)
18:05:08 <nickm> isis: did you answer the thing that isabela sent on Thursday, asking for an answer by monday?  Subject "work agreement please respond"
18:05:12 <asn> mit stuff?
18:05:24 <ahf> yes, very nice with coq classes :-S i have a friend who is really into that
18:05:32 <nickm> asn: yeah; I don't know if I'm going to stick with it, but the prof emailed me and said "hey you might want to come"
18:05:44 <nickm> and I said "sure, if you don't mind me maybe skipping out halfway through term"
18:06:37 <ahf> nickm: if you want to look at some of the memory analysis i've been doing this week and i could take some of your release tasks if you feel like it? i wouldn't mind having a day or two this week where i focus on something else to get my eyes a bit away from that exact thing
18:06:56 <nickm> mikeperry: any discussion needed on #22212 , or just on-ticket?
18:07:15 <nickm> My main issue this week is that I'm supposed to be getting 0.3.3 out, and it's making me really slow at everything else
18:07:40 <nickm> we have a huge pile of tickets in 0.3.3.  I'm going to defer many of them.  But we still need to fix real regressions and bugs...
18:07:55 <nickm> along with finishing #24902 and getting the TROVE-2017-001 fixes in
18:08:17 <ahf> so the best way to help you might spend a day on 0.3.3 "random" issues that makes sense to get fixed here and now?
18:08:23 <ahf> might be spending*
18:08:25 <mikeperry> nickm: I think the fix for #22212 is good. I am wondering if we should be updating all instances of that timestamp in the connection layer instead of the channel layer (and where those updates should be)
18:08:55 <ahf> isis: oh, also, i love doing things with coccinelle btw - i even have a t-shirt from their first conference somewhere :-P
18:08:59 <nickm> ahf: yeah, That could be great.  Ideally, folks should focus on 033 issues that are either super-easy to write super-small fixes for...
18:09:04 <nickm> ... or super-important
18:09:11 <nickm> Like, not just the stuff that looks fun
18:09:16 <ahf> yes
18:09:22 <ahf> cool, i will set a day or two off for that this week
18:09:23 <isabela> catalyst: i have a request for you on my update
18:09:34 <catalyst> isabela: ok, thanks
18:09:35 <nickm> mikeperry: hm. would that make the patch huge?
18:09:43 <ahf> isabela: i haven't forgot your request for the data btw, i just need to get it from people.torproject.org which is down right now
18:09:50 <nickm> mikeperry: would not doing it make the patch wrong?
18:09:55 <isabela> ahf: cool
18:10:00 <nickm> maybe we should do the easy thing in 032, and the hard thing in 034
18:10:01 <mikeperry> nickm: no. but perhaps less certain to be only fixing the notice as opposed to changing otherbehaviors
18:10:45 <nickm> alternatively we could just drop the warning in 032, and do the big fix in 034
18:10:58 <nickm> for stable, I'd rather go with surprise-reduction over regression-risks
18:12:04 <isis> ahf: mostly what annoys me is that i need to search for forever to find the coccinelle manual, then when i find it, it's not useful (e.g. http://coccinelle.lip6.fr/docs/main_grammar005.html#sec10) and then i need to find a bunch of random people's blog posts to remember how to use it lol
18:12:15 <nickm> If literally _everybody_ could spend time on 033 between now and EOM, it will help a lot.
18:12:21 <nickm> (everybody who hacks tor)
18:12:31 <nickm> otherwise, that's my feb and march.
18:12:57 <ahf> isis: yes, my biggest problem is that i'm not good enough with it, so even smaller things takes a bit too long for me. i usually just write to Julia though whom i know from when she was at uni. cph and she comes with some neat replies
18:13:09 <nickm> ahf: would you like to talk a little about gsoc stuff?
18:13:14 <ahf> yes
18:13:21 <mikeperry> nickm: ok. Roger's fix is good then. it just updates the timestamp in one additional (and reasonable) place. it should not introduce any regressions (since it is at worst redundtant, and at best covering the cases the channel layer missed)
18:13:34 <asn> nickm: spend time on 033 by triaging tickets and fixing 033 bugs?
18:13:36 <asn> nickm: ack
18:13:44 <ahf> just very short really: i send an email out to the network team list and it would be nice if we could figure out some projects for the students and maybe find some candidates for GSoC students
18:14:14 <nickm> mikeperry: cool. comment on the ticket, and I'll merge it as you suggest :)
18:14:29 <nickm> ahf: when do we need that by?
18:14:36 <ahf> i'm up for mentoring and don't have much vacation planned during the period where it is running, but people should probably also think of if they are up for mentoring
18:14:49 <ahf> nickm: i think google will tell "us" next week if we are accepted or not
18:14:59 <ahf> if we are accepted we have a small month to find projects and students i think
18:15:10 <nickm> ah, then having projects before next week would be smart
18:15:22 <ahf> the student application period is in march i think and students can submit ideas too i believe
18:15:31 <ahf> there is a roadmap included in my email :-)
18:15:44 <ahf> i think that was everything i wanted to say about it :-)
18:15:50 <nickm> ok
18:15:55 <nickm> everybody please also help ahf :)
18:16:01 <ahf> :-D
18:16:14 <nickm> isis: did you see my question above about isabela's email?
18:17:17 <isis> nickm: yep! sorry, i just went straight to email to respond
18:17:20 <nickm> Could everybody with feelings about RNG issues comment on catalyst's #25120?  I'll try to say something there too
18:17:29 <nickm> isis: ok, just wanted to make sure you saw it
18:17:30 <isis> nickm: thanks for the reminder!
18:20:11 * nickm says something on #25120
18:20:46 * ahf doesn't feel like he knows enough about the topic
18:22:10 <nickm> mikeperry, isis, catalyst: I hope, time permitting, that you'll also be able to jump in on the 033 minor tickets?  I know you're busy with other stuff, but this is a great way to get more coverage of different places in the code.  *Especially* if you've been working on the same stuff a long time and could use a break
18:22:46 <nickm> isis: when you're back from emailing let's talk about var_cell_t
18:22:57 <nickm> let's see, what other discussion topics are there....
18:23:36 <nickm> teor added topics for both of our pre-meeting days in March (10th and 11th).
18:23:52 <isabela> i would like to talk about that
18:23:56 <nickm> yes please!
18:24:10 <isabela> if would be possible to have some time on the 10th for roadmapping too so we are not stuck with only 2 hrs on the 11th
18:24:23 <isabela> it will help me with what i want to do for this time around
18:24:44 <nickm> I have no objection there ; what's the general idea?
18:24:51 <nickm> and should we be doing prep stuff ahead of time?
18:25:14 <isabela> yes we will do prep work ahead of time
18:25:23 <isabela> i will send orientation and 'milestones' :)
18:25:33 <nickm> cool
18:25:38 <isabela> but i do think i twill be helpful to hold certain conversations f2f
18:25:40 <nickm> grab a slot on the pad?
18:25:42 <isabela> and more time the better :)
18:25:50 <nickm> (if you don't mind)
18:25:53 <isabela> yes will do
18:25:59 <nickm> thanks!
18:26:03 <isabela> for the 10th right? since we have it already on the 11
18:26:09 <nickm> yeah
18:26:12 <isabela> :)
18:26:28 <nickm> if we manage to finish everything faster than you expect, we can do something else with the extra time on the 11th :)
18:26:32 <nickm> though let's not count on it
18:26:34 <ahf> speaking of homework: if we are going to have a rust session again where we need to prepare something then we should probably find out around now since the meeting is soon only a month away
18:26:43 <isabela> nickm: agreed
18:27:12 <catalyst> ahf: +1 also we should somehow set aside time so people can actually work on that in advance
18:27:21 <nickm> ahf: any thoughts?  I think that this time around, our best bet might be for everybody to try rewriting some personally selected part of Tor in rust.
18:27:26 <nickm> ahf: but maybe that's too ambitious
18:27:54 <ahf> nickm: yes, but then we should probably find out soon such that people wont have to do it in the week right up to the meeting where they might try to de-stress
18:28:56 <nickm> yeah; we should also make sure we scope it so it's a "get started and learn something" exercise
18:29:08 <nickm> not "relay.c rewrite must be completed by Mar 10!!!!1!"
18:29:27 <ahf> :-D
18:29:47 <dgoulet> hmmm I've raised questions about "Rust" in general on my last email to the net team about the meeting ... and so far no comment about that... I guess I should raise those questions againt on a seperate email
18:30:20 <ahf> yes, that would be a good kickstarter
18:30:22 <dgoulet> because just doing "more Rust" before the meeting doesn't align very well with any structured goal with the whole Rust project as far as I can understand
18:30:35 <ahf> my comment was also more "if we do a rust session, do we need to do anything before"
18:31:27 <dgoulet> right my point is that imo we should structure way more how the Rust project will evolve with Tor overtime so we know where we are going or aiming at (right now, from my perspective, is mostly confusion)
18:32:20 <isis> back
18:32:32 <ahf> dgoulet: agreed
18:33:02 <isabela> i think maybe is a good thing to step back and do what dgoulet is suggesting
18:33:13 <isis> (also, oh god, there is no way to document how to set up a meek reflector on a google cloud compute instance like isabela asked for)
18:33:14 <isabela> we havent got the modularization proposal
18:33:32 <asn> i also feel the same with dgoulet. in general, i feel that tackling the rust project without considering the general modulariation project is not a good idea
18:34:04 <nickm> I'd agree that modularization is important, and necessary as a prereq for rust work.
18:34:19 <nickm> isabela: you mean, we're still waiting to hear?  Or we heard, and they said "no" ?
18:34:21 <ahf> isabela: as in we got a no or as in we don't know if we get it yet?
18:34:31 <ahf> heh
18:34:32 <catalyst> i think the modularization project isn't going to happen as a big-design top-down effort. we need to keep the big picture in mind but we shouldn't pretend we can plan it all up front
18:34:45 <isabela> but we will keep working on getting it but we should have a plan on how we scale our rust efforts till then, because i think what was on that proposal timeline was an ok plan to start getting some stuff in rust
18:35:20 <isabela> ahf: we didn't got a no, it was a technical issue which there is still under conversation between us and then to see how to resolve it
18:35:35 <ahf> ok
18:35:44 <isabela> me and shari are talking with them on thursday
18:36:03 <nickm> I think our current rust efforts are in a place where we are still ramping up our own abilities
18:36:26 <isabela> isis: could you please explain what do you mean by it being impossible as a response to that agreement proposal?
18:37:01 <nickm> In parallel, I'm hoping to get more packagers trying the rust code...
18:37:41 <isabela> let me know if you dont know what i am talking about when i say 'proposal timeline'
18:37:48 <isabela> ^^ anyone
18:38:04 <nickm> I think that doing zero rust experimentation while we try to design the direction we're moving in would risk us deciding "okay, rust!" and then having no idea how to do it
18:38:43 <nickm> But I agree that "just jump in and write rust!" is only a strategy for getting experience, not for actually managing a modularization and rewrite
18:40:48 <isabela> so
18:41:18 <isabela> should people have some 'rust practicing' thing going on ?
18:41:25 <isabela> i am not sure where we are leading :)
18:41:38 <nickm> dgoulet: what do you think?
18:41:43 <nickm> asn: (same question)
18:41:55 <nickm> also, ln5 has a question at the bottom of the pad -- anybody looking at bwscanner stuff?
18:42:21 <dgoulet> I do think a session for writing Rust can be useful on the 10th but I also think I would love to sit down f2f with everyone and more thoroughly structure this
18:42:32 <dgoulet> could be on the 10th or 11th or after, no problem with me
18:42:44 <isabela> i like that too
18:43:07 <nickm> me too
18:43:08 <ahf> yes
18:43:10 <nickm> do you think having a "thing to do in rust before the meeting" would be good?
18:43:28 <ln5> re bwscanner, it'd be great to coordinate a bit. not sure how much of this discussion should happen in this group but i thought i'd bring it up here for a start.
18:43:36 <isis> isabela: It's really difficult because the way to access the service is through a highly-convoluted web console thing (and I'm currently locked out of it until the account has a credit card attached to it) which contains a shell (but copy and paste is not possible because it's annoyingly graphically rendered and not actual text), and further difficult because Google seems to keep redesigning…
18:43:42 <isis> …the console and re-labeling buttons and such, so things that David had originally clicked were not what I had to click (and I assume it will be different yet again).  Even further, Google's documentation is absolutely labyrinthine and also seems to get moved around enough such that links to specific sections don't work later. I've no idea how to document meek reflector set up beyond "follow…
18:43:43 <nickm> alternatively it could just be "everybody find a rust crate that does something neat, and write a little program with it
18:43:48 <isis> …the CDN's instructions for setting up an instance, compile meek with `go build`, put the meek binary on the CDN's machine somehow through a web interface or with their SDK, follow their instructions for how to run it".  Which is about the same level as David originally documented it. :(
18:44:22 <tjr> I tried running bwscanner and hit https://github.com/TheTorProject/bwscanner/issues/70. I think I may have gotten some data out of it, but the quantity of this error leads me to believe it is a significant issue that would block a truely useful test comparison.
18:44:22 <tjr> Once resolved though, I think I can get comparison data between a torflow scanner and bwscanner for people/community to analysze for differences and what those differences may imply.
18:45:01 <dgoulet> nickm: I'm not sure because I for instance I'm not a person that will learn well a new language in a group and giving myself Rust homework, if I encounter problem I would ask for help at that time so writing Rust before the meeting as an homework, imo has to be on an invidual basis where it is useful for some to be in a group and thus we provide a slot for them
18:45:05 <dgoulet> nickm: my two cents
18:45:17 <nickm> tjr: Do you know whether bwauth operators who are doing stuff for the authorities already see those errors?
18:45:21 <isabela> isis: i see some of the things you are saying as 'fixable' and some where the option would be to just do for how it is right now and the other person will care about future changes
18:45:48 <tjr> You mean these errors from torflow: https://pastebin.mozilla.org/9077327 ? I do not know.
18:45:59 <isabela> isis: the fact is not something straight forward to be done is why i am asking for such documentation
18:46:03 <tjr> IF you meant the bwscanner errors, I don't think anyone is running bwscanner except me and ln5 attempts
18:46:42 <juga> tjr: agree, don't think nobody is running bwscanner to get measurements yet
18:46:50 <juga> i'm running it to try to fix the bugs....
18:47:29 <asn> ln5: i think bw auth and bwscanner should become network team projects but not sure who should care for them right now. rome will be a good place to figure this out.
18:47:41 <isis> isabela: i guess i just feel nervous because the whole process for me was "google how to get this working" over and over again and "do what google says"
18:47:52 <ln5> tjr: i'm seeing the same as you do, plus a variant of it (see #81)
18:48:02 <asn> ln5: perhaps the best thing to do until rome is to gather various things that need doing
18:48:07 <ln5> asn: ok, we can wait until then. might be best. is tjr going to be there?
18:48:16 <tjr> Probably not.
18:48:21 <asn> (and fix any bugs that can be fixed. or just file the bugs and put them as things that need doing)
18:48:28 <asn> tjr: :(
18:48:30 <tjr> (Sorry. I'd like to be. =( )
18:48:36 <asn> np
18:48:43 <isabela> isis: please just say what you think on that email, is a request for agreement, if its not somthing that can be done, dont agree and say what would be something agreeable
18:48:44 <catalyst> isis: you mean google changes their interfaces too often for step-by-step instructions to be helpful?
18:49:29 <ln5> tjr, juga, others who do stuff on bwscanner: how can i be helpful? not sure it makes sense to do the exact same thing as tjr tbh.
18:49:50 <nickm> dgoulet: so you'd prefer having the group rust-together exercise be optional, but a "rust homework" is IYO ok?  Or not. trying to understand
18:50:47 <isis> yes, the interface changes and also API options and configuration things that are available in one interface (there are three: an SDK, the web console, and a 2GB (!!) python library) are missing/different in another
18:50:47 <ahf> have a rust discussion for everyone and have an optional rust session where the people can discuss some homework they did for the optional meeting?
18:50:51 <dgoulet> nickm: what I think we should have is a session for those who wants to pair up or work on Rust in a group so to seek help. Just to provide a session for people who prefer that instead of "learn it at home alone" :P
18:50:56 <dgoulet> ahf: that yes +1
18:51:03 <isis> isabela: responded in the email
18:51:11 <isabela> ok
18:51:21 <nickm> ahf, dgoulet: That would be fine with me; let's record that somewhere?
18:51:42 <ahf> who has the ball on this rust planning? komlo?
18:51:50 <dgoulet> the pad in the pad :P
18:52:26 <nickm> ahf: For the meeting? komlo did last time but I don't know if she's planning to do it again this time
18:52:32 <dgoulet> ahf: it is in the Discussion section
18:52:37 <nickm> ln5: Did your question get answered?
18:52:50 <ahf> ahhh, true dgoulet
18:53:00 <dgoulet> (taking forever for me to reload)
18:53:29 <ahf> do we just append it under 'Discussions'? :o
18:53:39 <ln5> nickm: not really, but i can congregate with the two main suspects otherplace, after meeting. i now know that nobody else in here is dying to dig in. :)
18:53:40 <nickm> (we have 5 min left here, so let's all move over to #tor-dev for remaining discussion if needed once we're done.)
18:53:46 <dgoulet> ahf: lets put it in the pad for the 10th
18:53:58 <dgoulet> ahf: there seems to be already an entry for Rust there with "komlo?" eheh
18:54:37 <ahf> oh, ok
18:54:48 <isabela> isis: did you sent the email? havent got it
18:56:44 <isis> isabela: yep, sent ~15 minutes ago
18:57:31 <ahf> dgoulet: i'm sorry, i don't see where komlo is mentioned? which pad are you on?
18:57:35 <GeKo> tor browser meeting in 2 minutes
18:58:04 <nickm> ok -- going to endmeeting now.  Thanks, everyone!  We have important stuff we've been talking about today.  Let's please be patient and keep talking on #tor-dev and elsewhere to one another until we're all happy!
18:58:06 <isabela> isis: got it
18:58:08 <nickm> #endmeeting