15:00:18 <pili> #startmeeting S27 01/28 15:00:18 <MeetBot> Meeting started Tue Jan 28 15:00:18 2020 UTC. The chair is pili. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:18 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 15:00:27 <asn> o/ 15:00:28 <mcs> hi 15:00:29 <pili> Hi everyone 15:01:01 <annalee_> \o 15:01:02 <acat> hi 15:01:05 <pili> here's the meeting pad: https://pad.riseup.net/p/s27-meeting-keep 15:01:40 <brade> hi 15:02:09 <pili> welcome annalee_ :) 15:02:29 <annalee_> pili: Thanks! 15:03:08 <pili> let's give a few minutes for pad updates 15:03:27 <dgoulet> o/ 15:06:26 <asn> dgoulet: there are some questions on socks errors on the pad 15:06:31 <asn> dgoulet: https://pad.riseup.net/p/s27-meeting-keep 15:07:14 * dgoulet looks 15:07:19 <pili> we can start the discussion with that :) 15:09:04 <pili> ok, let's start 15:09:24 <pili> so, for #32542 15:09:38 <pili> what's the next alpha we can get these in? 15:09:58 <asn> 0.4.4.1-alpha i guess? 15:10:01 <asn> that's gonna take a while 15:10:04 <dgoulet> in theory, around feb 15th 15:10:07 <dgoulet> ++ 15:10:13 <asn> dgoulet: u think so? 15:10:16 <asn> dgoulet: thats when it's gonna be merged 15:10:19 <asn> not when it's gonna get released 15:10:40 <asn> i think real life testing with these codes will have to happen using master and not any released versions of tor 15:10:43 <asn> after feb15 indeed 15:10:51 <asn> when the code will be merged to master 15:11:09 <mcs> so too late for 0.4.3. It probably does not matter since those errors do not occur much (if ever) in practice, right? 15:11:12 <asn> dgoulet: finding a way to trigger the codes would also help with review i guess 15:11:28 <asn> mcs: they do occur when it comes to overloaded network/onionservices 15:11:32 * pili adds a new discussion point to review what will make it into 0.4.3 15:11:37 <mcs> asn: ah, OK. thanks 15:12:01 <asn> pili: what got in, got in. all the rest is 044. 15:12:04 <asn> pili: 043 has frozen 15:12:07 <pili> right ;) 15:12:19 <asn> :) 15:13:40 <mcs> anyway, I raised the discussion point mainly so we can think about ways to test all of the errors and so people can see all of the browser behavior 15:13:55 <pili> brade/mcs: are you good then? 15:14:09 <pili> mcs: it's a good point 15:14:10 <asn> mcs: yep 15:14:18 <asn> updated the section in the pad about 043 and 044 15:14:23 <asn> dgoulet: please check it out to see it makes sense 15:15:02 <dgoulet> that is correct 15:15:05 <pili> asn: that was quick! thank you! I was going to do it myself... ;) 15:15:06 <dgoulet> 043 is closed now for merge 15:15:13 <asn> what about f4? 15:15:39 <mcs> F6? 15:15:48 <asn> 043 15:15:51 <asn> just added it 15:16:26 <dgoulet> they are all in tor except #32542 15:16:44 <asn> yeah 15:17:13 <dgoulet> there is F7 also 15:17:16 <dgoulet> within #32542 15:17:34 <brade> what is F7? 15:18:06 <dgoulet> Similar has X'F2' code but in this case, all introduction attemps have failed due to a time out. (v3 only) 15:18:10 <dgoulet> https://github.com/torproject/tor/pull/1597/commits/6d06f23a4d70a3a6eb92a0b3f61fdecfebdb3275 15:18:29 <pili> I think we need to add F7 to antonela's list 15:18:33 <pili> let me double check 15:18:34 <brade> +1 15:18:52 <dgoulet> that code means that we tried all intro points and they all timed out 15:19:05 <pili> yup 15:19:20 <dgoulet> so tor will stop retrying. You can either consider that the application behind the .onion is down actually 15:19:26 <dgoulet> errr 15:19:27 <dgoulet> overloaded* 15:19:34 <dgoulet> but it is all "in theory" 15:20:03 <pili> updated table in #30090 15:20:12 <pili> I should update the nc spreadsheet also 15:20:23 <pili> dgoulet: is F7 the last one we know of so far? 15:20:28 <dgoulet> yes last one 15:20:57 <asn> i will try to review #32542 this week 15:21:04 <pili> cool :) 15:21:05 <pili> ok, anything else on the errors? 15:21:20 <mcs> I am good 15:21:46 <asn> thanks for raising these issues 15:21:50 <pili> ok 15:22:08 <pili> the next point is just a reminder to add your tickets to the January report pad 15:22:27 <asn> i added my onionbalance work 15:22:33 <pili> thank you asn :) 15:22:50 <asn> i can also explain more in person at fosdem 15:23:04 <pili> asn: yup! we should have a mini S27 meeting :) 15:23:44 <pili> moving on to the final discussion point, I want to double check that all of the required client auth changes will be in 0.4.3? 15:24:41 <asn> i think it is 15:25:05 <pili> cool 15:25:19 <pili> ok 15:25:32 <mcs> on the browser side, the client auth changes have not been merged yet but I think tor provides what we need 15:25:38 <mcs> tor 0.4.3 that is 15:26:19 <pili> mcs: yup :) 15:26:49 <mcs> it would be good to merge the browser changes soon so we can see that things work in our nightly builds 15:27:12 <pili> that would be really nice, are you still waiting on a review? 15:28:03 <mcs> the main client auth part has been reviewed (#30237) and we are waiting on one more review for #19757 (key management) 15:28:17 <mcs> should be soon :) 15:28:26 <mcs> ready soon 15:29:05 <pili> great :) 15:29:28 <pili> I have added one more discussion point about objective percentage completion 15:29:42 <pili> I want to send a work completion report at the end of this month so now is a good time to review where we are 15:30:18 <pili> shall we move on to that? 15:30:37 <asn> yes 15:30:43 <asn> dgoulet: 01A1.1 seems to still be open 15:30:48 <asn> are there more to do there? 15:30:55 <asn> seems like #31632 is still open 15:30:57 <asn> that's the reason 15:31:04 <pili> ok: #29995 15:31:05 <pili> yup 15:31:06 <pili> not sure 15:31:07 <pili> you tell me ;) 15:31:15 <dgoulet> asn: yeah that one you asked for testing but I have no idea how to test that ... 15:31:36 <dgoulet> asn: I need to somehow simulate circuit failure service side exactly during a desc upload 15:32:14 <asn> if u think it's ok to merge without explicit testing, perhaps make a small case in the ticket and we can proceed? 15:32:20 <asn> i had kinda forgotten about that ticket 15:32:29 <asn> and since it's blocking o1a1.1 from 100% let's try to move it forward somehow? 15:32:39 <dgoulet> we can move it to -can 15:32:50 <dgoulet> since next merge window is in a while so that will linger for a bit 15:32:56 <asn> it's kinda important no? 15:33:13 <pili> that was my next question (how important is this?) 15:33:15 <asn> if we move it to -can it will never be done 15:33:22 <dgoulet> it is but not critical 15:33:27 <asn> i think it's fine to linger as -must until feb15 15:33:42 <dgoulet> asn: well if it is in the correct milestone and assigned, we won't loose it 15:33:43 <asn> but if u think it should be moved to -can thats fine by me 15:33:48 <asn> it just seems like a reachability issue 15:33:50 <pili> my plan was to report O1A1.1 as 90% completed 15:34:00 <pili> I think that's a good compromise 15:34:06 <dgoulet> well wait 15:34:24 <dgoulet> O1A1.1 is basically done 15:34:44 <dgoulet> the work has been done in #31632 15:34:53 <dgoulet> just need to figure out the last bits 15:35:04 <asn> yep 15:35:09 <dgoulet> and #22893 will _not_ happen 15:35:13 <asn> yep 15:35:16 <asn> ill move that out 15:35:31 <asn> moved out 15:35:41 <pili> ok 15:36:00 <dgoulet> #31632 is important but unlikely to cause real issues honestly... so until we can merge it (which will happen in ~15 days+, we can consider it done imo 15:36:07 <pili> and what are we going to do about #31632? 15:36:15 <dgoulet> but we won't loose it since it is assigned to me and milestone is not Unspecified 15:36:22 <pili> dgoulet: so we will merge it? 15:36:31 <dgoulet> it will get merged one way or the other for sure :) 15:36:41 <asn> dgoulet: ok sounds good 15:36:42 <dgoulet> that is a fix we need 15:36:47 <asn> feel free to do the move 15:36:53 <asn> and we can close o1a1.1 15:37:05 <pili> ok, so I'll go ahead and mark as 90% for O1A1.1 15:37:16 <asn> i think dgoulet is suggesting we can go 100% 15:37:25 <dgoulet> it is 100% 15:37:28 <dgoulet> no need to keep it open still 15:37:31 <dgoulet> we are virtually done there 15:37:38 <pili> ok, so it's done, it's just waiting to be merged? 15:37:40 <dgoulet> as in no more work will go in that ticket for s27 15:37:41 <pili> because we're not going to add the test? 15:38:05 <dgoulet> we'll see what we decide there for that ticket about the test, unclear at the moment 15:38:18 <pili> ok 15:38:29 <pili> let's move on then... ;) 15:39:04 <pili> O1A1.2 was last reported as 50% complete 15:39:05 <pili> asn: how much work do you estimate is left in this one? 15:39:12 <asn> we are at 70% 15:39:45 <asn> it will be 100% in beginning of march 15:39:50 <asn> in 7-10 days i will ask for first testing 15:39:58 <asn> and then some more days of bugfixing 15:40:02 <asn> write a blog post 15:40:05 <asn> and we can call it done 15:40:24 <pili> ok, sounds good, thank you :) 15:40:43 <pili> O1A2 was done a while back 15:41:09 <pili> O2A1 15:41:10 <pili> I guess we're just waiting for reviews there and we can call it done? 15:41:51 <mcs> I think so, yes. What percentage did you reported for O2A1 last month? 15:41:57 <mcs> s/reported/report/ 15:42:58 <asn> (brb) 15:43:19 <pili> last time I reported was at 50% 15:43:26 <pili> as I wanted to be conservative :) 15:43:33 <pili> that was at the end of November 15:43:58 <pili> so this is for the work completion ones which we have been doing every 2 months 15:44:01 <pili> separate from the monthly reports 15:44:12 <pili> mcs: I might call it 75% or 90% 15:44:16 <mcs> I would report at least 75% 15:44:25 <pili> just in case there are any surprises (which I doubt) 15:44:30 <pili> mcs: yup, sounds good 15:45:33 <pili> O2A2: I've previously reported 30% as antonela had been working on it 15:46:26 <mcs> I think you can report at least 50% for O2A2 because we have the tor piece done plus work in progress for the new error pages (part of O2A4) 15:46:27 <pili> actually, that could be a bit higher given that dgoulet did some work on the tor side for it also 15:46:30 <pili> yup 15:47:25 <pili> O2A3 15:47:32 <pili> the main ticket there is #21952 15:47:42 <pili> acat: are you still waiting on reviews? :) 15:48:36 <acat> i think so, for the revision i did addressing mcs/brade first review 15:49:14 <acat> then there are pospeselr's comments wrt to Onion-Location being sent in non 30X responses 15:49:22 <pili> I think maybe pospeselr and antonela will discuss it together this week 15:49:41 <pili> since they're both at the all hands 15:50:04 <pili> acat: so how much more work do you expect on this one? 15:50:09 <mcs> maybe we should wait for the outcome of their discussion…. 15:50:25 <pili> mcs: yeah... 15:51:03 <acat> pili: very little, i think 15:51:05 <pili> ok, I might not report on this one, I'll see 15:51:06 <mcs> and sysrqb may want to give an opinion about issues raised too. we should check with Matt in any case. 15:51:13 <acat> it's a matter of being sure about the header behaviour 15:51:18 <pili> yup 15:51:18 <pili> sure 15:51:38 <pili> there's the other tickets on O2A3 that are not moving and make me uneasy also 15:51:41 <acat> i think pospeselr asked asn, and it was suggested to ping web folks who would actually use Onion-Location header (anarcat/hiro, our Facebook friendo, etc) and see how they would want it to work/how they would configure web services to use it 15:51:43 <pili> we should figure out if we want to do them or not 15:52:12 <mcs> we should defer what we can if we want to finish on time :) 15:52:26 <pili> ok, I'll follow up on O2A3 offline, let's move on to O2A4 15:53:04 <pili> I had reported this at 50% last time 15:53:07 <acat> and we agreed that i would try to ask will about his opinion of Onion-Location 15:53:26 <pili> acat: ok 15:53:27 <acat> sorry, we already moved on :) 15:53:38 <pili> we can always move back :D 15:53:45 <pili> I know I'm always rushing everyone :P 15:54:00 <pili> the main tickets for O2A4 are #19251 15:54:18 <pili> and the onion indicators in the url bar 15:54:31 <pili> which I think antonela and pospeselr will also discuss this week 15:54:40 <mcs> I would say we are 50% or more done with #19251 (but we have not published patches for review yet so who knows?) 15:55:01 <mcs> creating good text for the error pages will take some work too 15:55:25 <mcs> did we report progress for O2A4 previously? 15:55:49 <pili> yup, 50% also, mainly on the design work done already 15:55:52 <pili> I might lie 15:55:59 <pili> not lie :S 15:56:02 <pili> leave it at that... ugh 15:56:25 <pili> as in not report any more time so far, in case of any surprises 15:56:26 <mcs> (ugh; WiFi glitch) 15:56:27 <mcs> we are always truthful 15:56:32 <pili> :) 15:56:43 <mcs> 50% seems fine to me but we need to move fast over the next few weeks 15:57:05 <pili> yup 15:57:06 <pili> and finally O2A5 15:57:07 <pili> which I haven't reported anything on yet 15:57:18 <pili> since we hadn't started until recently :D 15:57:39 <pili> acat: I assume there's no point in reporting any percentage completion, unless you want to count the research/analysis... 15:57:50 <pili> (we should wrap up the meeting soon also) 15:59:02 <acat> pili: i did some progress with the poc, but there are parts of it i think are still not clear/decided 15:59:13 <acat> UI mostly 15:59:20 <mcs> acat has made some good progress with respect to planning for the implementation but I am not sure what % to assign 15:59:23 <mcs> 25%? 15:59:26 <pili> maybe we can report 25% 15:59:27 <pili> snap :) 15:59:31 <pili> ok, I will do that 15:59:33 <mcs> (grabbing a number from thin air) 15:59:38 <pili> me too :P 15:59:47 <pili> ok, let's leave it there then 15:59:54 <pili> and end the meeting 15:59:55 <pili> thanks everyone 15:59:58 <pili> #endmeeting