15:29:19 #startmeeting sponsor8 status discussion 15:29:19 Meeting started Mon Dec 11 15:29:19 2017 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:29:19 Useful Commands: #action #agreed #help #info #idea #link #topic. 15:29:21 thanks everyone 15:29:36 i just wanted to do a quick check in on sponsor8 activities 15:29:43 for Q4 15:29:54 * ahf nods 15:30:04 ack 15:30:10 maybe we can ping tickets we have going per activity ? 15:30:18 is that easy or 15:30:30 not sure 15:30:35 ok 15:31:00 ahf and I have have been doing a bunch of performance analysis; i'v edone some prelim cpu hacking and work on mobile API improvement 15:31:01 i also would like to know if any information from those surveys became tickets 15:31:02 i don't have an overview if everything has tickets, but we could start with that? 15:31:22 isabela: i haven't created any tickets based on the surveys 15:31:31 i know #23684 is a request from OONI and will be helpful for all apps 15:31:50 some of them (the ones from the ooni meeting in montreal) had a lot of overlap with the survey info and there was created tickets for that 15:32:03 and nickm closed some of them already with patches 15:32:04 ok 15:32:24 started tagging some bootstrap/error reporting tickets as s8-errors 15:32:27 that stuff is mostly done, except for the hard parts 15:32:32 the hard parts being #23847 15:33:19 i have one question for some point in the meeting: the sponsor8-can usage on trac and promotion of that to either something else or "sponsor8" 15:34:01 ahf: yes 15:34:24 (i am organizing queries for the acitivities tickets so we can look at them) 15:35:23 ok! 15:35:27 isabela: since you only have a little time, please let us know if that's something we should do instead. 15:36:20 nickm: i am trying to understand what we have done on Q4 under each activity 15:36:21 the "Build plan for improvements to Tor for integration on iOS, Android" item on the roadmap is pretty big and involves a lot of people 15:36:23 just a collection of tickets 15:36:39 also 15:36:54 ok, so you'd like a ticket list for each activity for the things we did, and the things we made progress on? 15:37:31 yes, that is what i am looking for as and know how you all feel about tit 15:37:33 *it 15:38:02 like if you think less work was done under y as you would like to because x 15:38:05 do you want that during this meeting or after? i'll look through my meetings.txt file then and structure it somehow 15:38:05 any redflags 15:38:44 ahf: can be after if ppl prefer 15:39:37 so I think we're making fair progress on the performance analysis and on the API improvements 15:39:53 I don't know how we're doing on the wakeup-reduction (that's more where ahf has been paying attention) 15:39:56 i think right now we have all the tooling in shape for the platform, we have good understanding on some of the performance issues there, one things that is missing there is being able to reproduce the results in a good way. nickm did some patches to #24374 that i haven't managed to prove yet is an enhancement or not 15:40:08 and I don't know how we're doing on the better error reporting (where catalyst has been working) 15:40:11 that is one thing is that missing that i aim at doing this week since it's rather important 15:41:23 i think the wording we used for the error reporting part of the proposal doesn't make for easy reporting. IMHO the major UX issues related to error conditions have to do with hangs rather than explicit errors 15:41:43 (except the ones where tor doesn't get far enough for Tor Launcher to start capturing logs) 15:42:02 I _believe_ we have the freedom to do the right thing on these activities rather than doing the exact letter of the item 15:42:05 isabela: ? 15:42:08 then i've been looking into the event loop usage (what we call wakeups) and have a semi-working patch that probably should go in so we can start having it run by more people too if they find it useful 15:42:18 nickm: more or less 15:42:53 nickm: we do have indicators we are using to measure the work and a timeline of work, which we can modify but if we do we need to explain why etc 15:43:05 sorry, freedom to do what? :o 15:43:11 ohh, never mind 15:43:12 https://docs.google.com/spreadsheets/d/1FFcnYfHe7puIYzyzVxzftWlDiDixz4cRrZO6pb_XV0U/edit#gid=1655282830 15:43:14 some stuff is catalogued at #23508 but i should make it more complete 15:43:25 this is the timeline we organized the work we will do 15:43:45 this would be about months 3, 4 and 5 15:43:52 rows 10, 11 and 12 15:44:58 and we have indicators we picked to reported on 15:45:14 for some of them we are just building baseline for now 15:45:21 like the network speed tests 15:45:23 the network speed test tests with shadow isn't something i've looked much into, that part is missing. i did get introduced to shadow and how to use it during the montreal meeting, but haven't setup in "experiments" in it 15:47:23 catalyst: how are we doing wrt actually making improvements on this stuff? 15:47:29 ahf: the fact you learned about it is something to report because is a necessary step since you didn't know much abot it 15:47:58 yes, agreed. even submitted some patches to it :-P 15:48:46 nickm: going slowly unfortunately. consensus, nodelist, and guard stuff is really hard to read. if someone has a summary of the large number of tickets being worked on about guard stuff that would be helpful 15:48:56 i'm unsure what we want to test here. testing network speed tests with shadow seems a bit odd to me? have we done this historically before with some sponsors, or? 15:49:52 ahf: stability and stress testing? 15:50:22 rown 11 column D 15:50:24 yes, but never for network performance, right? 15:50:25 on the timeline 15:50:31 * ahf looks 15:50:45 do you want me to get the doc 15:50:53 with the description of that activity? 15:50:53 no no, have it in front of me 15:51:14 catalyst: hm. do you think you have a set of question about "where does X happen" and "where do we decide to Z" and "how do we test if Y" and stuff like that? I can dig that up pretty easily 15:51:19 it's DRL_modularization_new_narrative_doc ? 15:51:29 or is it an older version of the doc? 15:51:31 the guard changes most recently are described in guard-spec.txt 15:51:38 but the pending tickets, less so 15:51:48 ahf: oh that is just about modularization not sponsor8 15:51:58 ahf: let me get that for you 15:52:01 ok, yes please 15:52:02 were the recent tickets due to us not implementing the spec, or side effects of the spec? 15:52:08 i see that this document is way shorter than the one i was looking for 15:52:39 the ones teor has been working on? I think those are the results of interactions of that spec with other stuff; I haven't seen any spec violations there yet. 15:53:18 i think it would help to instrument guard state stuff, possibly via control channel 15:54:34 also i'll start cataloguing what things require a reasonably live consensus vs a live one, since it seems we don't really have such a thing 15:56:21 isabela: i found it on print now - it's from the time around the wilmington meeting though 15:58:02 catalyst: let's make sure we have a plan to have that start producing code changes on a not-too-long timeline though? 15:58:12 ahf: sorry my computer is so slow 15:58:14 which Ox.y is column d row 11? 15:58:23 i got a new one but i didnt had time to set it up for this trip 15:59:43 nickm: probably soonest feasible change is fixing or making progress on #2878 15:59:53 catalyst: that would be something 16:00:09 isabela: where/how would you like us to collect these tickets? 16:01:00 let me get a pad 16:01:56 https://storm.torproject.org/shared/nlBIFCxM3P-Lcl_VkFSuyjRA6nY6vmeMoFmC7HMknPL 16:03:16 btw 16:03:17 do you want sort of the same kind of style that we use for the network team meeting just where we should assume that the person reading it isn't very used to internal tor lingo? 16:03:17 this one 16:03:18 https://storm.torproject.org/shared/2FBa-H0I6_A1VNy9Jc_RzvdT-PpvlfWNCHsRXrN39GA 16:03:26 has the keywords etc 16:03:29 we said we will use 16:04:32 the only place there where i see stress mentioned is under 2.4.1, which is for orfox? 16:06:24 ahf: i guess the writer got a bit confused 16:06:36 becasue she was writing about orfox and network work 16:06:55 and the example was that crashing an app can drain the battery 16:07:14 We will seek to optimize the The Tor Network controller interface, which manages how the network establishes connections and circuits. to reduce both the overhead in browser requests to the Tor network and the amount of time the browser appears to be unresponsive after user actions. 16:07:24 ok,which document should we follow though? 16:07:49 this is the same text 16:08:43 i have not done any work around that :o 16:11:32 * isabela is trying to paste 16:11:41 some text from the narrative 16:11:48 sorry if that pad is confusing 16:13:04 ok 16:13:59 okkay. So we should fill this in with ticket numbers and status info, some time today? 16:14:06 ahf: see that shadown is related to 2.3 -to test proposed network circuit changes 16:14:12 (not trying to rush, but assuming isabela is about to get on a plane 16:14:13 ) 16:14:21 yes i will have to go now 16:14:26 ok! 16:14:28 so no need of doing it today 16:14:37 i just wanted to bring this to you as a check in 16:14:44 i've noted down the following: we need to figure out the situation around stress/testing (2.4.1) and what we want to do with shadow 16:14:50 and that we should write in the Q4 work in the pad 16:14:51 before the quarter is over and we could assess where things were 16:15:16 ahf: look at the text for 2.3 re:shadow 16:15:26 could we clarify which are calendar quarters and which are grant year quarters? they look like they might be different? 16:15:35 catalyst 16:15:47 just work as normal year quarter 16:15:53 ok thanks 16:16:01 like Q42017 == oct,nov,dec 16:16:19 that is why i am doing this check in now in december 16:16:31 i need to run now 16:16:44 sorry folks i wanted to do it today because i will be on mozilla all hands all week 16:16:48 the timeline doc we have still uses the relative-to-the-proposal quarter values 16:16:50 and next week is the last week of the month 16:16:59 ahf: yes 16:16:59 it's cool! it was good to do a catch up on this 16:17:05 have a safe trip isabela 16:17:36 thanks folks write questions you might have related to the proposal etc 16:17:44 on that pad and i will catch up with it too 16:17:48 cool 16:17:59 thanks! 16:18:04 i will run now bbl 16:18:10 nickm: you an end the bot too :) thanks 16:19:58 #endmeeting thanks everybody!