20:00:15 #startmeeting anti-censorship weekly checkin 2019/04/11 20:00:15 Meeting started Thu Apr 11 20:00:15 2019 UTC. The chair is phw. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:15 Useful Commands: #action #agreed #help #info #idea #link #topic. 20:00:24 reminder: our pad is available online at https://pad.riseup.net/p/tor-censorship-2019-keep 20:00:33 hi everyone o/ 20:00:50 o/ 20:01:11 * dcf1 20:01:12 o/ 20:01:16 Hi 20:01:19 +1 20:01:40 let's start with our announcement section? gaba, this was your announcement, right? 20:02:00 yes, just a reminder about changing the owner field of the ticket in trac once you start the ticket. 20:02:22 done :) 20:02:29 ok! 20:02:58 do we do the discussion or the team updates first? 20:03:16 For gettor it seems that we should have Hiro in the meeting for give a report on how gettor is doing now. Can we move it for next week? 20:03:32 discussion is fine 20:03:46 ok, let's make sure to discuss this next week. 20:04:29 let's also make sure that there will be two people taking care of gettor -- also, so hiro has somebody to do code reviews with. 20:05:13 sounds like a good idea 20:05:23 i've been thinking about the anti-censorship mailing list. afaik, there are different opinions. i heard somewhere that there may not be enough momentum or use to a separate list. 20:05:30 (item #2 on the discussion list) 20:06:01 i think there can be some merit to it, though. i added the advantages i see under the point. 20:07:00 i'm curious what you all think. i don't quite buy the "not enough interest" argument because the other team lists aren't particularly busy either. 20:07:31 there is a team and conversation about processes and anything related to the team can be done in that mailing list. 20:07:38 we do any technical discussion in tor-dev 20:08:00 also anything related to funding could be done in the mailng list. I do not see the harm of having a mailng list 20:08:24 I'll subscribe if there is a new list, but I don't feel a strong need for one. 20:08:32 fwiw, we only have one mailing list (we = tor browser team) tbb-dev 20:08:40 which is used for technical discussions 20:09:01 (just in case there are some data points needed about what other teams are doing) 20:09:24 network and metrics team also have their own mailng list 20:09:27 as i see it, for technical discussions there is tor-dev (tor-related) and traffic-obf (non-tor-related). that basically leaves us with organisational matters, right? 20:09:48 right 20:09:50 i think it would be a useful place to gather emails related to all of these points 20:09:53 i think it's helpful to have a private list so we don't expose details about NGOs and individuals we might put at risk by talking about details on a public list 20:10:20 if it's a private list I won't subscribe. got enough of those in my life. 20:11:05 (bearing in mind that if it's for tor-related organizational discussion, dcf1 maybe doesn't *need* to be subscribed) 20:11:28 it seems like tor-internal and non-list emails is probably still the best way to handle private things 20:11:38 i think arma also wasn't a fan of a private list but he's not here to make his point. 20:11:53 cohosh: agreed. 20:11:56 i think arma's point was about how a private list looks to the outside 20:12:15 that it suggests we are doing secret development (which we are not) 20:12:45 catalyst: your point is a good one -- are you ok with sensitive matters being discussed in tor-team or tor-internal? 20:12:59 * catalyst has more thoughts that are not for a logged medium 20:13:34 we should definitely make sure that everyone is aware it is public if we make a public list 20:13:54 * cohosh also remembers some public/private drama with traffic-obf at one point 20:14:28 i like the idea of the list being a potential "collect all org things related to anti-censorship" and i would also prefer for it to be public. 20:14:55 phw: what sorts of org things? 20:15:07 phw: +1 20:15:54 catalyst: meeting logs, monthly reports, improving processes (e.g., our 'pluggable transport integration guidelines') 20:17:16 then again, there's no pressing need right now, so we could also continue this discussion in stockholm. i don't have strong opinions either way at this point, but i may, in a couple more weeks ;) 20:17:33 phw: that i'm fine with, as long as we don't talk about names or details of individuals or NGOs there 20:18:04 catalyst: yes, agreed. 20:18:07 catalyst: +1 20:19:45 ok, should we move forward and create a public mailing list for organisational matters, then? i'm slightly leaning towards "yes" 20:20:13 ok 20:20:45 ok, i'll get the process started, then. 20:21:15 the next item on our discussion list is code reviews. 20:21:33 as i understand, so far, we relied on asn for help. gaba, do you mind taking this item? 20:21:41 Right now asn in the network team is handling reviews for the anti-censorship team too. 20:21:53 yes, asn and dgoulet are the person assigning reviews in the network team. 20:22:04 so they were also doing for anything related to anti-censorship work 20:22:45 for them, once a week, asn and dgoulet assign reviews available for the week. 20:23:07 we would need to figure out how to assign/manage code reviews here 20:23:20 gaba: do all teams have a dedicated "review assignment person"? 20:24:15 and what's the process? once i finish a patch, do i let the review person know, so they can assign someone to review my patch? 20:24:30 well, not sure about other teams. For the metrics team there are only two people so each person review the other's person code and then decide on other things to review just talking when it needs to be 20:24:52 you change the ticket in track to needs_review 20:25:27 i'm not convinced that we're "big enough" to need a dedicated person who assigns reviews. what do you think cohosh? 20:25:44 we've also been doing assignments here in this meeting 20:25:59 where we've added tickets in needs_review in trac to our "needs help with" on the meeting pad 20:26:04 yes, maybe just deciding in reviewing each other code and periodically looking at tickets that needs to be reviewing and deciding on those then it would be ok 20:26:05 and this has included dcf1 20:26:12 dcf1: do you have any thoughts? 20:26:43 we also work on a lot of different tools for which there are subsets of people who have experience with them and some who don;t 20:26:48 our meetings seem like a good place to assign reviews. if this turns out to be unsustainable, we can always revisit the topic. 20:26:56 i think for example sysrqb was reviewing dgoulet's bridgedb stuff 20:26:57 +1 20:27:05 so far the "needs help with" method has worked pretty well, I think. 20:27:10 phw: sounds good 20:27:10 +1 to assign reviews in this meeting 20:27:19 ok, sounds good. 20:27:31 I've been trying to help bootstrap code review for snowflake tickets, but sometimes I've made people wait because I can't get to it very soon. 20:27:47 dcf1: thanks for doing that 20:28:01 ahf has done some snowflake things, but he might be leaving us in may? 20:28:02 dcf1: every bit of help is greatly appreciated! 20:28:41 cohosh: i think so, yes. gaba can elaborate. 20:29:03 ahf is wrapping up work with snowflake in April 20:29:09 and going back to work on the network team's roadmap 20:29:31 ok, next item is #30121. over the last week i started thinking about monitoring our infrastructure, originally related to default bridges. 20:29:41 ok so it seems i will need to learn some bridgedb stuff and phw will need to learn some snowflake stuff to make reviews sustainable perhaps? 20:29:54 turns out, there's plenty to monitor. bridgedb, bridgedb's distribution modules, and our bridge authority. 20:30:19 (i think that's the long term plan but may need to happen soon) 20:30:22 cohosh: should we talk about this as part of the next item? the one about maintainership? 20:30:45 phw: ok yup 20:31:17 coming back to the infrastructure monitoring: i spent some time figuring out who runs our default bridges, and poked the people whose bridges disappeared. it would be great if we had a system that would do the monitoring for us. 20:31:38 it's not quite clear to me what the best approach is. maybe hiro will be able to help here. 20:32:05 anyway, i was thinking of adding tickets related to "we should monitor our infrastructure" to the roadmap. 20:32:42 does this sound reasonable to others? any thoughts? 20:33:07 yes, it sounds good. Do we have any of those tickets in trac? 20:33:10 about monitoring 20:33:28 there's #30006 and another two or three, i think. 20:33:54 i can create a meta ticket that tracks monitoring of all our anti-censorship infrastructure. 20:34:13 +1 20:34:14 ...that tracks the other tickets, not monitoring. 20:34:33 ok, let's talk more about this next week, hopefully with hiro. 20:34:38 ok 20:35:01 next item is maintainership. ideally, we want to get to a point where each of our projects is taken care of by at least two people. 20:35:37 i guess step one would be to understand who's taking care of each of our projects :) 20:36:07 is there a wiki page that keeps track of this? 20:37:14 it maybe for some of them 20:37:46 fwiw, i started working on bridgedb because sysrqb and dgoulet increasingly cannot. 20:38:23 there is one for gettor that is very not updated https://trac.torproject.org/projects/tor/wiki/org/roadmaps/GetTor 20:39:06 phw: it's probably useful for us to create this page so people know what we maintain as a team 20:39:17 i've been meaning to create a wiki page for the anti-censorship team. we can then add a section that lists the maintainers for each of our projects. 20:39:26 a wiki page for teh anti-censorship team would be good 20:39:27 cool! 20:40:21 ok, i'll create the page, so we can re-visit it next week. sounds good? 20:40:28 good 20:41:07 cohosh: anything you want to add regarding maintainership? 20:41:08 +1 20:41:21 ah just continuing the discussion above about reviews 20:41:27 (sorry for typing slowly) 20:41:28 let me add that arlo is also qualified to do review on snowflake, though as far as I know he doesn't attend these meetings. 20:41:52 arlo commented on #25688 and #30138 recently, for example 20:41:54 dcf1: cool, i've noticed he responds on trac tickets 20:42:05 that has been very helpful 20:42:18 i think the revewing and maintainership are related but different issues 20:42:23 I was glad on #25688 because he understands that part of the code better than I do. 20:42:44 and i'm not sure how we want to handle it given that non-team people will ahve other commitments as well 20:43:29 yes, that's a good point. we shouldn't over-rely on volunteers who may disappear tomorrow. 20:44:22 i guess doing assignments at the meeting kinda helps with this 20:44:37 since we have people specifically agree to do reviews here before we assign them 20:45:16 and then maybe we can ping people outside the meeting if we still need more reviewers? (like sysrqb or arlo) 20:45:24 i'm interested in getting more involved in snowflake development but i don't think a review of mine would be helpful because i have little understanding of the codebase as a whole. 20:45:38 phw: that's fair. same for me and bridgedb 20:46:09 dcf1 has been a great help with reviews and i am getting much more familiar with the codebase as a result 20:46:32 so with some time after some bootstrapping we can probably help each other get onboarded onto these projects as well 20:47:00 cohosh: it would be great if we both are able to take care of emergencies acros projects, like restart a service. the barrier for this is lower than being able to review code. 20:47:15 true! 20:47:26 (that is not to say that we shouldn't review each other's code) 20:47:43 yep 20:48:39 hmm, that's a complex topic. 20:48:51 there were a few times when ahf reviewed my snowflake code and dcf1 jumped in with more review comments 20:49:00 which worked really well 20:49:01 * gaba only reminds people that we have 12 more minutes until the 1 hour mark of the meeting. 20:49:04 ah sorry 20:49:22 we can probably do what we are doing for now and discuss more going forward as we gain more experience 20:49:29 cohosh: agreed 20:49:34 We don't need to spend time on the next discussion point about commit access, I will make a ticket to discuss that. 20:49:50 thanks dcf1! 20:50:10 * phw wonders if Roya has anything to say regarding spooky scanning bridges? 20:50:49 in the meanwhile, let's check what each of us needs help with. 20:51:09 kat5: i might be able to help with the question for Alex 20:51:20 i wanted to discuss #28655 with either sysrqb or dgoulet. 20:51:22 phw: I'll send you info about ICLab scanning default bridges when I can dig it up, they are already scanning at least a portion of them (or were) 20:51:55 I have two questions: How many bridges you would need us to monitor? 20:52:35 dcf1: oh, neat! i sent an email to phillipa the other day and she said she'll get back to us after the IMC deadline 20:52:38 Roya: it's around 50 bridges (IP:port) 20:53:11 is that the max number or just the first round 20:53:14 cohosh: Cool. Do you think that your and phw's reviews are adequate, from a technical perspective? I only ask because you're both new, so I don't know if ahf should therefore review also. 20:53:31 For the default Tor Browser bridges at least, maybe phw has more targets in mind. 20:54:32 Roya: i think the default bridges would be a great start. ultimately, it would be great to have all ~1,000 monitored but let's start with a small goal. 20:55:03 Roya: you can find the ip:port tuples here: https://trac.torproject.org/projects/tor/wiki/doc/TorBrowser/DefaultBridges 20:55:07 We are deciding whether we need to dedicate a seperate experiment or add it to our list. Also, how should we share the results? obviously you don't want us to push it to our public censoredplanet repo 20:55:21 kat5: I would see if ahf can review it too 20:55:32 gaba: Okay. 20:55:38 IMO a public censoredplanet repo is fine for scans of the default bridges. 20:55:38 kat5: at this point i am at least as familiar with the snowflake code base as ahf, but we should include him because he's around until the end of April 20:55:49 and there are some roadmapping things he probably knows that i don't 20:55:58 Roya: the default bridges are public, so i don't see a problem with sharing the scan results with the world. 20:56:20 cohosh: gaba: Okay. I'll ping him. 20:56:45 ack. 20:57:12 Roya: anything else that we can help with? 20:58:42 cohosh: regarding the study in #29279: yes, let's try to move forward with this! i'll try to get a better understanding of point 2 -- the one related to bridgedb. 20:58:52 phw: awesome! 20:59:06 kat5: ah, also ahf knows about the PT spect stuff which i have no clue about atm 20:59:12 * dcf1 impressed with what's happening in #29279 20:59:32 cohosh: Okay. I will ask him. 21:00:03 I am good 21:00:09 ok, i think we managed to go through all of our items. did i forget anyone? any announcements or questions left? 21:00:12 cohosh: I will take review of #30125 21:00:34 dcf1: thanks! it was a pretty basic refactoring 21:00:37 I think we are good. 21:00:58 ok, thanks everyone! 21:01:01 #endmeeting