15:00:38 <pili> #startmeeting S27 06/25
15:00:38 <MeetBot> Meeting started Tue Jun 25 15:00:38 2019 UTC.  The chair is pili. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:38 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:00:49 <pili> who else is around :)
15:00:59 <pili> asn said he would probably miss this one as he's been travelling
15:01:10 <pili> let me find and share the pad
15:01:44 * mcs is here
15:01:46 <pili> https://storm.torproject.org/shared/zBDiopJ5lhbcjYsUpk2s_zOjUqOh59EwpciFeJvFDCy
15:01:55 <pili> please add your updates, I see some of you have already done so
15:05:57 <pili> ok, I think people are more or less done
15:06:21 <pili> first item I wanted to discuss was whether we should have another S27 meeting on irc ahead of the dev meeting
15:06:25 <pili> that would be on the 9th July
15:07:09 * GeKo pops in
15:07:14 <pili> any thoughts from people? :)
15:07:21 <mcs> Kathy and I will actually be traveling that day
15:07:47 <pili> ok, that settles it then ;)
15:07:48 <pili> fwiw I think it's good to wait for Stockholm anyway
15:07:49 <antonela> i'll be travelling, too
15:08:09 <dgoulet> yeah I'm on a plane on the 9th
15:08:43 <pili> hehe
15:09:20 <pili> and my follow up item on that was to discuss how we can make the most out of the time we'll have to discuss this project in stockholm
15:09:55 <pili> any items that people want to discuss? should we concentrate on having interactions with external stakeholders?
15:10:04 <pili> what items would people like to see on that agenda? :)
15:10:47 <antonela> i think that will be useful to have a decision on some of the explorations we are doing now
15:12:09 <pili> so: client auth "dialog", error messages, ... ?
15:12:30 <antonela> naming space
15:12:58 <antonela> prioritizing onions in tor browser
15:13:07 <antonela> suggesting onions in tor browser
15:13:52 <mcs> Those all sound like good topics, especially if other stakeholders are present.
15:13:53 <pili> ok
15:14:04 <antonela> yes, that is the best part
15:14:10 <antonela> im sure asn has another items too
15:14:27 <antonela> should we draft an agenda via email?
15:14:36 <antonela> maybe we can add items at the wiki entry
15:14:39 <pili> we can do, or we can try to do the prep thing on the wiki
15:14:40 <pili> yup
15:15:08 <antonela> oki, will add mine
15:16:31 <pili> I will send a follow up email on this for people to start thinking about this also
15:16:53 <pili> dgoulet: any discussion points from network team? :)
15:16:58 <pili> for the dev meeting
15:17:36 <dgoulet> hmmm not on my side tbh. I think asn as stuff though about TB client auth integration
15:18:13 <mcs> “has stuff” I assume :)
15:18:19 <dgoulet> yes
15:18:20 <dgoulet> eheh
15:18:53 <pili> cool :)
15:19:08 <pili> I'll hand the mic over to mcs then :)
15:19:17 <pili> for his discussion point
15:19:28 <pili> s/point/points
15:19:57 <mcs> So we now have client auth w/prompting working “in the lab” but once we have browser builds we need a way for other people to try it out.
15:20:13 <antonela> thanks for thinking about it mcs!
15:20:22 <mcs> Are there any test .onions that support client auth? If not, who should set one up?
15:20:54 <dgoulet> "test .onions" you mean some test machines we have with one?
15:20:56 <dgoulet> if so, no
15:21:06 <mcs> Something Antonela could use for example :)
15:21:08 <dgoulet> but _very_ easy to set one up, I can do that and then just give the creds to some list ?
15:21:11 <antonela> arthur made some for me in the past, but are not available rn
15:21:26 <antonela> dgoulet: that'd be ideal
15:21:27 <mcs> dgoulet: that would be great
15:21:41 <dgoulet> ok! I'll set that up today then! Will be quick :)
15:21:49 <antonela> super, thanks!
15:21:50 <dgoulet> who needs those?
15:21:51 <mcs> (Kathy and I have one but it is not robust enough / not always up)
15:21:52 <dgoulet> apart from antonela
15:22:11 <mcs> I would like to be able to use it as another test server
15:22:24 <mcs> (w/brade)
15:22:38 <dgoulet> ack ok I'll email you also with the creds.!
15:22:54 <mcs> thanks much!
15:23:21 <mcs> The other thing I mentioned in the pad is that the long timeout before the prompt appears makes testing painful :)
15:23:40 <dgoulet> mcs: tell me more? It is the 120 sec SocksTimeout ?
15:23:43 <mcs> I don’t know enough about tor to fix it or work around it myself
15:24:16 <mcs> Tor notices that client auth is needed but does not send the necessary extended SOCKS error code for a while… 2 or 3 minutes.
15:24:23 <dgoulet> ah!...
15:24:34 <mcs> Maybe it keeps trying for a while?
15:24:35 <dgoulet> that is a bug :S ... ok that means the Socks times out instead of noticing the error
15:24:54 <dgoulet> mcs: one workaround is set "SocksTimeout" to seomthing low like 5 sec or 10 sec
15:25:03 <dgoulet> mcs: I'll work on a fix in #30382
15:25:07 <mcs> I tried that but I will try again
15:25:16 <dgoulet> mcs: huh and still not working? :|
15:25:23 <mcs> I will let you know later.
15:25:30 <dgoulet> ok
15:26:09 <dgoulet> mcs: just to make sure of something
15:26:20 <mcs> We have SocksTimeout 30 in our torrc (client side(
15:26:21 <dgoulet> mcs: the extended errors require the SocksPort to have the "ExtendedErrors" flag
15:26:28 <dgoulet> mcs: else you never get the error code
15:26:42 <mcs> We have ExtendedErrors set and do get the correct codes… just not quickly :)
15:26:49 <dgoulet> interesting! ok
15:26:54 <dgoulet> mcs: I can definitely fix that for you
15:26:59 <mcs> Other than the delay, it seems to be working really well
15:27:08 <dgoulet> awesome
15:27:35 <mcs> Should I try an even shorter SocksTimeout or do you have enough to go on?
15:28:08 <dgoulet> mcs: yeah try to put it down more in order for you to do your testing at least. If it is still not working, we have a bigger problem :)
15:28:30 <mcs> OK, I will let you know (maybe via a comment in the ticket)
15:28:49 <mcs> That’s all from me
15:29:48 <pili> thanks mcs and dgoulet :)
15:29:50 <pili> I think antonela also had a point to discuss about testing
15:30:06 <pili> I'm hoping to use the S9 work to test this
15:30:15 <antonela> yes, me too
15:30:41 <pili> I need to double check but I think we will have some tests happening in the october - december 2019 timeframe
15:30:51 <antonela> the question is if we can sync about running it during ESR68 transition, so we can have some results for iterate when the browser team is back
15:31:21 <antonela> is more about finding the right time to make it useful for our development process
15:31:26 <pili> as in still have this bi weekly meeting with browser team members participating?
15:31:46 <pili> i.e we can use this time to sync if that's ok with mcs/brade
15:31:56 <pili> actually, it's 2 tests, right?
15:32:29 <mcs> That is fine. We also need to make sure Antonela and other people have builds they can use for testing (unless you do not need running code).
15:32:36 <pili> GeKo: do you remember if it's still mcs and brade that are going to be working on the onion suggestion work also?
15:32:44 <pili> i can try to find in my notes also
15:32:58 <GeKo> i tihink that's not decided yet
15:33:46 <pili> ok
15:34:05 <antonela> okey, i'll try to move it forward and think about how we can do it during S9
15:34:14 <antonela> will keep you updated
15:34:37 <pili> I'll take a note to schedule this also
15:34:43 <antonela> thank you
15:35:16 <pili> GeKo: yup, I just checked my notes and we hadn't decided on O2A3 yet :)
15:35:23 <GeKo> :)
15:36:21 * arma5 reads backlog
15:37:02 <pili> ok
15:37:19 <antonela> that's all from my side
15:37:23 <pili> thanks antonela
15:37:53 <pili> the only other thing I had was a reminder/announcement that Browser team are going to be pretty much 100% on ESR68 transition work from July - September
15:38:50 <pili> so the other teams have the rest of the week to extract any S27 browser knowledge/requirements/work from mcs and brade  :)
15:39:09 <mcs> We are not going to disappear completely ;)
15:39:12 <antonela> hehe
15:39:47 <pili> I thought ESR68 was a black hole... :D
15:39:58 <pili> oh, and I almost forgot, thank you everyone for sharing your updates for the monthly report
15:40:26 <pili> is there anything else you want us (gaba and I) to highlight for this upcoming one?
15:40:46 <pili> also, is there any work that is 50% or more completed that we can start billing for?
15:42:18 <antonela> i'm not sure how to measure done in a way that the sponsor and we are comfortable with
15:42:19 <mcs> From the browser side, I hope we are more than 50% done with the client auth work but I would not feel good about billing for it yet (should get it in an alpha release first)
15:42:34 <antonela> mcs, agreed
15:43:00 <antonela> maybe some network development is done and we can close that part
15:44:22 <pili> fair enough :)
15:44:38 <pili> mcs: dare I ask which alpha release we're aiming for? :)
15:45:03 <mcs> I think we are up against the esr68 work… so I don’t know :(
15:45:34 <mcs> We also need the network team’s work to ship in some way (dependency)
15:46:11 <GeKo> i guess we can think about putting it in an alpha when the tor alpha with everyting we need gets out
15:46:26 <GeKo> when that happens i don't know
15:46:35 <pili> ok, but would this happen during the ESR68 transition or after?
15:46:40 <pili> I'm guessing after, but I want to confirm :)
15:47:14 <GeKo> don't know the tor release schedule
15:47:19 <GeKo> *I
15:47:27 <GeKo> dgoulet: might
15:47:28 <dgoulet> it is on the Trac
15:47:31 <dgoulet> wiki* , sec
15:47:36 <dgoulet> https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases
15:47:45 <dgoulet> ah hmmm you might want the alpha relase?
15:47:46 <pili> I also have it on a shared master plan spreadsheet with gaba :D
15:47:49 <GeKo> yeah, but the question is when all the patches we need would land in tor
15:47:54 <pili> we must share that with people at some point
15:47:55 <GeKo> yes
15:47:56 <dgoulet> that is basically Nick domain, I do not know also :S
15:48:16 <dgoulet> #30381 and #30382 is what you need in tor so we expect those to be in 042
15:48:34 <GeKo> yes, so the first 042 alpha
15:48:37 <GeKo> i geuss
15:48:42 <GeKo> *guess
15:48:49 <dgoulet> yes so that happens when the merge window closes
15:48:56 <dgoulet> Jun 15 10, 2019
15:48:59 <antonela> after ESR68 work, we could work on set release points for each activity that requires TB updates (even nightlies), in that way we can coordinate the user testing and have better estimations
15:49:00 <dgoulet> (not 15)
15:49:09 <dgoulet> so the 042 first alpha should be sooooon
15:49:30 <GeKo> well the first with the patches we need :)
15:49:37 <dgoulet> if not the alpha1, will certainly be the alpha2, for those dates, we need nickm to tell us
15:49:46 <pili> antonela: +1
15:51:21 <dgoulet> antonela, mcs: you should have an email with the .onion client auth creds :)
15:51:28 <antonela> got it :D
15:52:10 <mcs> dgoulet: I see it too; thanks!
15:52:33 <pili> ok, seems like we need to get a handle on those dates :) I guess the alphas are not on trac?
15:53:04 <dgoulet> they aren't. Only Nick _really_ knows when they come out :)
15:53:47 <pili> hehe
15:53:48 <pili> no worries
15:53:50 <pili> I'll have to think how we manage that ;)
15:54:04 <pili> do we have a time range at least?
15:54:40 <dgoulet> pili: nickm can give you that I bet ... but I really don't know :S
15:54:49 <pili> there was talk of a tor release on july 3rd from looking over at #tor-dev
15:55:13 <dgoulet> I believe that is for 041, not the 042... unsure :S
15:55:16 <pili> but I'm guessing this won't make it in there?
15:55:17 <pili> ah ok
15:55:27 <dgoulet> yeah little chance to get it by July 3rd...
15:55:54 <pili> I guess leave it with me then :)
15:57:47 <pili> ok, we're coming up to the hour
15:57:48 <pili> any other last minute items/comments?
15:58:32 <antonela> im groot
15:59:28 <pili> ok, team, thank you so much for your time :)
15:59:44 <antonela> thanks pili!
15:59:47 <pili> please send gaba and I anything you'd like included in the June report
15:59:48 <pili> bye :)
15:59:49 <pili> #endmeeting