17:01:17 #startmeeting network team meeting, 16 July 17:01:17 Meeting started Mon Jul 16 17:01:17 2018 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:17 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:01:18 yello! 17:01:25 \o 17:01:28 o/ 17:01:31 pad is at https://pad.riseup.net/p/xoBCOyzTwLtY 17:01:32 hello! 17:01:55 how's everybody doing today? 17:01:59 hi all! 17:02:21 pretty good! 17:02:39 let's start with the roadmap, per usual :) 17:03:11 I think we're on track for 034 stable in mid-august, which is good. 17:03:19 !!! 17:03:22 buenas 17:03:25 yes 17:03:26 I wonder if we've ever had a release become stable on time before... 17:03:34 (like, on schedule) 17:03:42 nice 17:04:14 If anybody's working on anything that isn't on the roadmap, and it's going to take more than a single day, we should really add it to the roadmap... 17:05:09 also, soon, we should schedule some meetings to revise our plans for 035, since we have a bunch of 035-roadmap-proposed items, AND some things we might need to reschedule since we don't have people working on them 17:05:15 i'm messing around with the sandboxing stuff when my brain stops working and i think i'm above a day on that so far, but i have no idea if we have a sponsor for that 17:05:20 err, as many people working on 17:05:25 i have kept track of the hours in a notebook for it 17:05:44 ahf: ok, so first thing to do is see if it's billable, and see if there's a roadmap-proposed ticket for that 17:06:02 there is no roadmap ticket for it, this was just something i picked up myself 17:06:13 arma2: maybe you have an idea about this, we have been talking a bit about it every now and then ^^ 17:06:54 Should we try to get our roadmap reconfigured at this meeting today, or schedule a meeting where we can talk through the issues? 17:07:10 mikeperry: you are on "scheduling meetings" rotation this week; there are a few for you 17:07:20 (I'm getting ahead of the agenda, I guess) 17:07:27 ahf: i mostly pay attention to the sandboxing question at the browser level (but i still haven't read the huge great sysrqb thread) 17:07:49 nickm: if we plan to spend more than 30 minutes, I would say new meeting 17:07:51 arma2: ack. i think this was more whether any of this was billable for core tor right now 17:08:34 dgoulet: I tend to agree it will be more than that, given how many 035-roadmap-proposed items and how many items with strikethrough on the roadmap 17:08:43 indeed 17:08:44 isabela: do you agree we should try to schedule a meeting for thist? 17:08:47 *this 17:09:00 yes 17:09:06 great 17:09:07 the thing is 17:09:13 ahf: i think we might want to change that heuristic to something more like "is a sponsor asking for it or expecting it". because "could we fit it into one of the nsf things" is always answered with a yes, but that yes doesn't create more nsf money. 17:09:23 if i am to participate it will only happen around first week of august 17:09:25 cuz of my vacation 17:10:06 arma2: okay 17:11:26 ahf: the flip side being that yes we should prioritize sponsor deliverables, but the big goal there was to have that list be a subset of what we have time for, and then put our own prioritized items into the remaining time. which involves picking our own prioritized items. which we started doing in seattle. 17:11:29 I think we're going to need you, so we should probably aim for the 1st week in august then 17:11:46 arma2: aye 17:11:53 Next up is reviewer assignments. Is everybody okay reviewing the tickets they have this week? 17:12:04 (Do we know if mikeperry is here today?) 17:12:25 he wrote out of band that he might miss the meeting 17:12:38 nickm: ack (first week of august) 17:12:41 ahf: so, tl;dr "yes it's billable but i think that might not be the right question" :) 17:13:06 ahf: ah 17:13:07 ok 17:13:26 arma2: i think my question was "i have 16 hours of this that tor may or may not be able to get some money for, is there someone i should put it in harvest" 17:13:59 hmm of what? :) sorry i missed it 17:14:15 isabela: it was just some stuff i did some weekends ago that was unrelated to s8 17:14:22 ahh 17:14:27 isabela: related to sandboxing on different platforms 17:14:32 isabela: after we spoke with the mozilla people in SF 17:14:37 cool 17:15:10 and i'm unsure where to bill it to and if we don't have any place that is fine too. i just don't want tor not to get the money from a sponsor if we can get it. if not, then it's cool too 17:15:28 so if this is nothing that fits any nsf stuff (i need to review it to remember) then is network team work category on harvest 17:16:03 i will need to review the other stuff going on with nsf - maybe you send the tickets to me and arma and we figure it out? 17:16:29 yeah, let's do it that way. i am very close to finishing this one for review then i can send it to you with some number of hours and you can figure out where it fits in 17:16:31 tx for pointing it out :) is always good 17:16:32 that sounds good 17:16:47 we've been trying to avoid putting nsf hours for people not in the US, because of the mess with that one program manager. so yes, network team work category sounds like a good choice. 17:17:07 (in other news the early signs from heather are that we're actually in good shape on filling up our nsf things. so yay.) 17:17:25 nice 17:17:44 that is, M, Q, and 3 are now full, and V and 2 don't end until next summer or later 17:18:14 yes 17:18:20 in theory her next step is to reflect that in harvest, by locking certain categories. 17:18:26 but this is a network team meeting so i will let it continue :) 17:18:32 hahaha 17:18:56 arma2: Does "full" mean that we should rnot be doing the things we had planned to do between now and 1Sep for Q or 3? 17:19:07 (I see no M, but we had planned on putting privcount on Q) 17:19:08 isabela: are we behind with s8 HS data collection? 17:19:29 HS? 17:19:57 err, the performance evaluation 17:19:59 sorry 17:20:06 nickm: privcount could also fit V (if i recall correctly) 17:20:06 with all the people doing network measurement 17:20:15 ahf: !! yes we are behind :( 17:20:30 ok, i worried that i had missed some emails or something 17:20:35 isabela, arma2: If so, we should move it into V on the roadmap and on trac. 17:20:38 i failed on it a bit cuz of traveling 17:20:43 nickm: no, we should continue to do the things. 17:20:43 ahhh, of course 17:20:55 i still am going to need to write about the amazing things we did for each of them 17:22:19 it just means we don't need to stress about whether we're getting enough hours put into the right buckets 17:22:25 #agreed move privcount to sponsorV 17:22:26 it doesn't mean we must not work on the topics :) 17:22:49 well, if we're doing them, we should have an accurate plan of how we're labelling them 17:22:59 let's see 17:23:04 next is rotations? 17:23:34 I see catalyst on triage, dgoulet on community, mikeperry on scheduling meetings, and me on CI+Coverity 17:23:44 Everybody (who is here) cool with that? 17:24:31 any community handoff from catalyst->dgoulet? 17:25:14 not really anything ongoing to hand off 17:25:38 the roadmap-proposed tickets are: https://trac.torproject.org/projects/tor/query?status=accepted&status=assigned&status=merge_ready&status=needs_information&status=needs_review&status=needs_revision&status=new&status=reopened&keywords=~035-roadmap-proposed&max=200&col=id&col=summary&col=milestone&col=status&col=type&col=priority&col=component&order=priority 17:26:21 I suggest that we revisit those in early august, unless there is anything super-urgent where we need a decision right now? 17:26:44 at the same meeting where we revise the 035 roadmap together. 17:26:47 i think that is a good idea 17:26:53 any objections? anything super-urgent? 17:27:05 nickm: agreed 17:27:29 Okay, discussion phase. 17:27:48 but see announcements: I see HOPE is this week, and PETS is next. 17:27:58 so please expect higher latency from folks than usual 17:28:07 and use email if you can't find somebody on IRC :) 17:28:28 discussion stuff: 17:28:48 what is the next steps on prop#295? Do we need a meeting to figure that out? 17:30:40 I think we might need such a meeting. 17:31:23 or I could try to figure it out on my own? :/ 17:31:38 nickm: i'm willing to take a look at it again 17:31:43 ok 17:31:49 I think teor was interested too 17:32:01 if we can talk about it tomorrow evening that might be a good start 17:32:51 ok i'll make some time to reread background material then 17:32:58 me too 17:33:04 let's see if we need to schedule something else after that 17:33:23 second item looks like it's somewhat discussed already. 17:33:44 so does anybody object to using the same pad week after week? 17:33:56 if not, I'll make a long/term pad this week, for use in the future 17:34:01 *long-term 17:34:03 i can't remember why we stopped doing that after some time? i think we did it at some point? 17:34:08 not too long ago 17:34:34 not sure; I thought we were worried about teh vandalz 17:34:44 but we can give it a try and see how it goes a 17:34:44 there will be a time when the pad will be kind of _big_ and will take a while to load (especially on TB) 17:34:58 I think we should delete stuff that's more than N weeks old... 17:35:05 agreed 17:35:08 iiuc, tbb-team does maintain a pretty long pad, and it seems ok 17:35:08 will that be enoug 17:35:20 can also rotate to a new pad every so often 17:35:23 sure 17:35:36 Ours is 3500 lines long 17:36:07 so let's give it a try? 17:36:15 sure 17:36:21 #agreed let's try a persistent pad url, and see how it works for us 17:36:56 next, I had a question: I started working on our NSS conversion. I'm doing this solo now, so I'm wondering if anybody would like to be my sounding-board 17:37:26 I'm also wondering if it's worthwhile merging _partial_ NSS support, or if I should hold off until I have a version of Tor that doesn't need OpenSSL at all 17:37:40 That second one might take a while, I'm afraid. 17:37:55 i think it will look good on our s8 deliverables if we get the entire thing in and can avoid openssl entirely on android 17:38:02 anybody else excited on this stuff? Any thoughts on logistics? 17:38:04 it's a part of the sponsorship we haven't measured a lot on 17:38:16 (like, how much we actually reduced the size) 17:39:02 how much is roughly missing right now? where tor still needs openssl? 17:39:03 I bet most of the DL size is firefox... 17:39:09 though I could be wrong 17:39:15 yes 17:39:16 I guess "every little bit helps" 17:39:27 this wont help our orbot size for sure :-) 17:40:23 So a) is this worth doing? I'm having fun, but we should decide if I forge ahead. 17:40:37 b) If it's worth doing, should I work with anybody else? 17:40:53 c) Should we merge it in pieces, or all at once? 17:41:59 for (c) i think if we can merge it in pieces without it making a mess, we should 17:42:10 hm, ok. 17:42:11 i'd like to help, but i don't want to take on more s8 stuff until i've closed the two items i have left :-/ 17:42:21 I think I can do it as a multistage thing 17:42:24 I'll make subtickets 17:42:24 i'd like to review things and stuff like that if you need help with that 17:42:44 ack 17:43:12 I'd also like to see if we have anybody who has reviewed NSS code before who can have a look at it -- some of the APIs are underdocumented. 17:43:32 I've frequently had to look at the NSS source code to understand how something is supposed to be called 17:43:39 yikes, ok 17:44:22 ok 17:44:23 nickm: yeah i've had to resort to that before as well. to be fair i've also had to do that with OpenSSL 17:44:44 catalyst: do you think I could also rope you in as a second reviewer here? 17:45:23 let's see, what other questions... 17:45:27 asn has questions for mikeperry ... 17:45:28 nickm: sure. i'm not too familiar with NSS but i've dealt with some fallout of NSS support integration before elsewhere 17:45:54 nickm: yeah im assuming that mike will get back to me when he reads the pad 17:45:55 catalyst: you have more NSS and PKCS11 experience than I did a week ago :) 17:45:58 not expecting reply right now 17:46:03 ack 17:46:18 isabela has a note on modularization and planning... 17:46:29 anything you need feedback on there? 17:46:39 Also, did you get the info you needed from last week's sponsor8 pad? 17:46:57 dgoulet wants to switch for bug triage next week; anybody else available to do it? 17:47:06 and dmr needs a mirrored nyx repo on github... 17:47:21 nickm: yes i have a thought around modularization 17:47:26 well, I don't need it, but atagar mentioned he'd like it, so I'm just pushing that bit forward :) 17:47:53 nickm: there is an opportunity for a grant that could pay for 1 year work of 2 ppl on modularization (we could organize a 'phase 1' work plan) if we are to move w/ it 17:48:05 the thing is that this would have to be done in a rush 17:48:32 dmr: Okay. I think Isabela knows the current state of our github mirroring stuff. It might be in flux right now, since we're gathering info 17:48:34 so i am not sure if we would be able to get it done on time 17:48:38 isabela: define rush here? 17:48:57 by july 31st 17:49:01 isabela: And would it be the same amount of work we had planned with more people and more time? 17:49:10 July 31, 2019? 17:49:22 yes 17:49:29 this is the proposal deadline 17:49:45 "this"? 17:49:52 july 31 17:50:01 oh 17:50:04 the amount is what i mentioned above 17:50:08 12 months / 2 ppl 17:50:15 so it wouldn't be "do the work in a rush", but "write the proposal in a rush" 17:50:18 ah 17:50:31 I think it could be worth doing -- I still want us to get paid for modularization 17:50:33 that can be break down in any way we want tho, like 4 ppl working 1/2 of their time or rotate the people 17:50:45 I can help cut down the previous list of deliverables to be more appropriate for the reduced resources 17:50:54 nice 17:50:55 ok 17:50:57 that is good to know 17:51:01 i can follow up on that offmeeting 17:51:02 :) 17:52:18 dmr: wrt tor-spec improvements: I'd suggest making tickets on trac? It's okay to put trivial fixes (grammar, typos, labeling, style, etc) all in one branch, but for changes of substance there should be separate branches, if that's ok. Sound plausible? 17:52:35 sounds good 17:52:50 I'll probably have a few different branches then, but I'd guess no more than 5 17:53:04 should I just collect all the trivial fixes into a single ticket? 17:53:16 sure 17:53:23 alright, will do 17:53:33 5 minutes left. Anything else for this meeting? 17:53:43 I'll open a ticket but keep it as "assigned" until near the end of the project, and then bump it to "needs_review" 17:54:06 ok, cool 17:54:28 Oh, btw, everybody who is going to HOPE: Have fun! 17:54:46 Tor has a booth and a talk, great stuff 17:54:58 \o/ 17:55:19 oh, and: I might take Friday off this week, to give myself extra time to rest up for PETS. 17:55:29 And I might have a hard time making next week's meeting, for the same reason 17:55:30 ^ FYI: it sounds from the community meeting that flexlibris and/or stephw (and/or others) will collect some HOPE info into a single place :) 17:56:55 okay, we're out of time. Thanks for the hour, everyone! 17:57:00 \o/ 17:57:02 I'll see you on IRC! 17:57:04 #endmeeting