17:03:09 #startmeeting network team meeting, 12 Aug 2019 17:03:09 Meeting started Mon Aug 12 17:03:09 2019 UTC. The chair is asn. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:03:09 Useful Commands: #action #agreed #help #info #idea #link #topic. 17:03:11 boom 17:03:18 \o/ 17:03:21 who is around? 17:03:21 o/ 17:03:46 hi! 17:03:57 gaba: Error: Can't start another meeting, one is in progress. 17:04:02 oops 17:04:05 i did it 17:04:18 yes, sorry. I got an issue with the window and had to close it and join again 17:04:41 ok how we doing this? 17:05:24 ok let me start with the reviews 17:05:31 mostly we want to talk about reviews today. right? 17:05:36 most people are on vacs 17:05:37 unless there is something else that people are bringing 17:05:44 and we are not assigning reviews to mike because of the esr review 17:05:54 so it's just me, dgoulet and teor. which means that the three of us got many reviews to do. 17:06:04 if it's too much for this week, perhaps we can carry them out to next week i guess... 17:06:29 (i got 4, teor and dgoulet got 3. and some of them are non-trivial) 17:06:50 so thats it for reviews. dgoulet knows the above already :) 17:06:55 is there any type of priority that we can put to the reviews and only do the more urgent? 17:07:00 hm 17:07:19 i guess the reviews can get prioritized indeed 17:07:23 yes, I think we should just have a few reviews assigned and move the rest to next week 17:07:25 not sure if me and dgoulet should do the priortiization 17:07:52 historically, we've been assigning all the reviews and letting the reviewer do the prioritization 17:08:02 ok, that works too :) 17:08:04 maybe it's not the best approach, and we should partially assign them 17:08:11 but it's what we have right now :) 17:08:15 and what we did this week :) 17:08:20 just to be sure we do not have something that needs to be done soon and does not get done 17:08:44 i think its the 041 stuff that needs to get done asap 17:08:52 yes 17:09:03 according to 17:09:04 https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases/041Status 17:09:11 there is no 041must or 041should in needs_review 17:09:23 so we good 17:09:58 ok 17:10:00 i think teor took good care of these keywords this week 17:10:07 like triaged them out etc. 17:10:12 so #30992 is out of 041-must, but #31356 should be in 041-must, I think 17:10:46 hmm 17:11:12 though I guess listing the protover support is Padding=1,2 instead of Padding=2 makes it less urgent to get in the first 0.4.1 stable 17:11:15 yah changing the protover is high prio imo 17:11:16 mikeperry: there is no 041 change needed for #30992? 17:11:33 mikeperry: and what's the status quo of #31356? 17:11:40 i see a few children ticket etc. 17:11:41 what needs to be done? 17:12:08 asn: we discussed this on the weds meeting.. I don't think it makes sense.. the sequence number field is too high risk of a change, and won't really solve the fingerprintability problem of these circuits anyway 17:12:41 ack 17:13:00 so #30992 is something that needs further thinking and well into 042 territories 17:13:31 ok, can any of you change #30992 to 042 then? 17:13:45 ok i can do that 17:13:55 or mike would you prefer to do since you know more about this? 17:14:34 I can write an explaination on the ticket yah 17:14:43 and move it out of the milestone 17:14:44 thanks 17:15:12 ok, and what is the situation for #31356 ? 17:16:02 I don't remember the meeting saying anything about not removing a protover, or discussing this phased protover removal plan on #31356.. 17:17:30 * gaba looking at logs... 17:17:47 http://meetbot.debian.net/tor-meeting/2019/tor-meeting.2019-08-07-22.59.log.html 17:17:49 * asn reading ticket 17:17:56 thx 17:18:36 too much text for now i think 17:18:41 mikeperry: what do you think is the right way forward at this stage? 17:20:23 asn: I think bumping the protover can still work, but it is not as simple as we thought in that meeting, either 17:22:00 mikeperry: how can i help with this? 17:22:03 i think im a bit behind on this 17:22:10 i was not aware of this whole discussion in the logs 17:22:17 and im not sure if i can page it in during this meeting 17:23:35 I believe that if we can drop support for Padding=1 in 0.4.1.x-stable, and have 0.4.1-stable clients and 0.4.1-stable relays only try to negotiate padding if Padding=2 is listed, then that will be sufficient 17:24:01 ack 17:24:08 I don't think we need to list Padding=1,2 and do this staged retirement of Padding=1 (which wasn't discussed, and seems over-complicated) 17:24:58 i see nick suggesting 1,2: 23:27:41 I think you want to advertise padding=1,2 17:25:33 (btw i obviously dont know enough about this issue, so hope im not saying stupid stuff) 17:25:39 but i can try to understand more this week 17:25:43 (since i also dont know much about protover) 17:25:56 I think we only need to do that if we care about alpha clients negotiating with stable relays.. that's what doing the phased transition would allow for 17:26:06 hm 17:26:38 and if we dont do phased transition, alpha clients wont be able to negotiate with stable relays? 17:26:43 I think it is OK just to say "alphas will stop negotiating padding with stable relays".. I think it's not worth the effort to make sure they continue to work for this edge case 17:26:44 doesnt this sound bad for reesarchers? 17:27:03 alpha are not considered stable... we should not backcompat for them... 17:27:31 I think if we miss 0.4.1-stable for this change, *then* we have to do this phased transition 17:27:46 aha 17:28:32 im a bit confuse about all this protover/stable/alpha situation so i dont think i can give useful feedback atm. but i can look into this tomorrow. 17:28:43 but if we can get it in by first stable, we can just say "old 0.4.1-alpha tor won't pad with 0.4.1-stable tor; stop using old alphas, use newer ones" 17:29:04 ack 17:29:07 i think that's ok 17:30:02 ok, can somebody write this decision in #31356 (and maybe #31387) ? 17:30:26 mikeperry: yes +1 17:30:34 so I can make that change top priority for me so we have a good shot at getting it in the very first 0.4.1-stable 17:30:37 lets just get it in 041 17:30:44 ok 17:30:52 i will take a good look at this tomorrow, and try to do something useful 17:30:59 and i can do any code review needed if there is code to review 17:31:03 ok. i will update the branch for you by then 17:31:10 thanks mike 17:31:28 thanks 17:31:39 nice 17:31:47 fun :-) 17:32:45 ok 17:32:57 ehm one more thing 17:33:30 it would be nice if the decisions we take and the patches we write are blessed by nick/teor. so if we wait for them until next week, we will have time until 041-stable right? 17:33:54 like if we have everything ready for them 17:34:23 teor could be commenting tonight. I think nick would be ok with this decision, right? Can't things rollback if not? 17:35:08 guess so 17:35:32 but have no idea how critical are this changes 17:35:35 Nick releasese 041 stable so we still have time until the first stable for sure for nickm/teor to assess that decision 17:35:46 ok 17:35:47 dgoulet: good point :) 17:36:54 ah yah date got pushed to 8/20 17:37:08 ? 17:37:09 on the release dates page 17:37:12 ahh, yes 17:38:08 anything else? 17:38:27 can people briefly check the roadmap for this week if everything is there for each of you? https://dip.torproject.org/torproject/core/tor/boards 17:38:30 related to teh roadmap 17:39:16 its there yes 17:39:19 but #31356 is not there 17:39:26 but thats ok i guess 17:39:31 since it's not formally on the roadmap 17:39:54 but yeah that kanban works for me 17:40:32 I'm adding it as you all are working on it as s2 17:42:59 sounds good :) 17:44:28 Anything else for today? 17:44:31 i think we are good :) 17:44:33 * dgoulet is good 17:45:18 +1 17:45:25 mikeperry: do you need any help for this week? 17:46:12 mikeperry: feel free to bounce s2 stuff on me 17:46:32 I hope to start on an outline on the circpad dev docs; so I will need some cycles from asn to look at that and/or double-check our spec is current 17:46:42 asn: yah, ok thanks. will do 17:46:43 sounds good 17:47:03 ok. it seems we are good then. you all have a good week! 17:47:07 \o/ 17:47:12 thanks! same to you! 17:47:20 asn: you have the meetbot power :) 17:47:23 :) 17:47:25 #endmeeting