18:29:45 <sysrqb> #startmeeting Tor Browser Team Meeting, 21 January 2020
18:29:45 <MeetBot> Meeting started Tue Jan 21 18:29:45 2020 UTC.  The chair is sysrqb. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:29:45 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
18:29:57 <sysrqb> Hello everyone, happy Tuesday
18:30:04 <mcs> hi
18:30:08 <boklm> hi
18:30:49 <acat> hi!
18:31:59 <pospeselr> hii
18:32:40 <sisbell> hi
18:32:41 <brade> hi
18:34:33 <Jeremy_Rand_Talos> hello!
18:35:56 <sysrqb> boklm: do you have plans for this week?
18:36:11 <boklm> yes, adding them now
18:36:26 <sysrqb> pospeselr: oh, right, you're coming too. great!
18:36:44 <pospeselr> mmmhm
18:36:46 <sysrqb> boklm: thanks :) nsorry, didn't mean to rush you
18:36:47 <pili> hi
18:36:57 <pili> sorry, bit late
18:37:16 <sysrqb> hi pili
18:37:19 <sysrqb> you didn't miss anything
18:37:25 <pospeselr> speaking of Berlin, what's our plan/agenda their for the week?
18:37:30 <antonela> o/
18:37:44 <sysrqb> pospeselr: we should figure that out
18:38:12 <sysrqb> acat: thanks for writing patches for those tickets
18:38:27 <sysrqb> i'll hopefullly review them this week
18:39:01 <acat> thanks
18:39:40 <sysrqb> Okay, let's get started
18:39:58 <sysrqb> on Friday I was considering whether we should release a Tor Browser update this week
18:40:21 <sysrqb> because mozilla tagged 68.4.2, and they released that yesterday
18:40:54 <sysrqb> 68.4.2 is mostly for Fennec UI changes
18:41:11 <sysrqb> and there's some bug with Citrix VDI
18:41:20 <pospeselr> sounds like a good way to be potentially distracted during the all-hands
18:41:56 <sysrqb> and that was why I decided we shouldn't go through all the effort of releasing updates this week :)
18:42:19 <sysrqb> if any of you think we should, then we can discuss it
18:43:01 <sysrqb> but, i prefer not shipping a release during all hands, which means we have ~3 days to tag, build, and ship
18:43:12 <sysrqb> which is not ideal
18:44:27 <sysrqb> i'll take silence as you all agree :)
18:44:48 <pospeselr> :)
18:45:01 * boklm agrees. less releases also gives us more time to work on other tickets.
18:45:14 <mcs> +1
18:45:15 <sysrqb> yep
18:45:16 <acat> +1
18:45:47 <sysrqb> boklm: i was thikning #28325 can wait until the beginning of February
18:46:03 <sysrqb> but if you want to start looking at it, please do
18:46:21 <sysrqb> i think finishing the nightly build stuff is more important right now, overall
18:47:18 <boklm> ok. Reviewing #32870 makes me want to start looking at #28325 at the same time, as it is related.
18:47:38 <sysrqb> okay, that makes sense
18:48:39 <sysrqb> sisbell: i'd like to compile dependencies in a similar way on all platforms, if we can
18:48:52 <sisbell> ?
18:48:56 <sysrqb> but, if we need to statically compile openssl for Android, then please document why
18:49:00 <sisbell> Is this regarding clang?
18:49:45 <sisbell> I'll need to investigate if that is possible
18:50:18 <sysrqb> okay. if it's not posisble, then please document why
18:50:30 <sysrqb> so someone in 3 years can look back at why we made these decisions
18:51:29 <sysrqb> okay, pili, i think we're at your discussion question
18:51:38 <pili> hi
18:51:58 <pili> yup, I'm just wondering if anyone has any project ideas for GSoC this summer
18:52:09 <pili> and whether you'd be interested to mentor a student
18:52:25 <pili> I know we're stretched quite think already, so no pressure
18:52:42 <pili> just let me know if you have any ideas and I'll add them to my list
18:53:57 <pili> that's it from me :)
18:54:28 * boklm will think about possible ideas
18:54:40 <sysrqb> i wrote a note for myself about mentoring someone who can identity broken features/functionality on websites due to Tor Browser-specific changes
18:55:01 <pili> great! :)
18:55:04 <sysrqb> and then identifying what is causing the breaskage
18:55:11 <sysrqb> *breakage
18:55:12 <Jeremy_Rand_Talos> oh, actually
18:55:18 <pili> if you can send me some details by the end of next week that'd be great
18:55:32 <sysrqb> sure :)
18:55:36 <sysrqb> i'll make a note :)
18:56:13 <Jeremy_Rand_Talos> If a student is interested in experimenting with making Tor Browser use the new Guix-based build system that Bitcoin Core is working on, that might plausibly be interesting.  Not sure if there's a mentor available for that, I assume boklm would be the logical person, no clue if he has time
18:56:33 <sysrqb> that sounds like a lot of work
18:56:49 <sysrqb> maybe three months is about the right amount of time
18:57:00 <pospeselr> how long is the GSoC term for?
18:57:20 <Jeremy_Rand_Talos> yeah, not sure if it's doable in the timeframe of GSoC, I haven't looked at Bitcoin Core's codebase in enough detail to gauge it
18:57:23 <pili> pospeselr: May - mid August
18:57:28 <pili> maybe mid May - mid August
18:57:45 <sysrqb> my concern is whether it would be useful - like, even if we have someone work on this project
18:57:58 <sysrqb> and they are successful
18:58:11 <sysrqb> would be jump over to using that instead of rbm?
18:58:30 <sysrqb> has anyone here looked at guix and seen it does what we need?
18:58:33 <pospeselr> yeah what is the win for us switching our build inrastructure?
18:58:47 <Jeremy_Rand_Talos> sysrqb, the Guix-based build system is substantially more auditable because it doesn't rely on binary packages from Ubuntu/Debian, so there's an advantage there
18:58:59 <pospeselr> a whole lot of work getting us to where we are now already doesn't sound smart
18:59:10 <Jeremy_Rand_Talos> I think I filed a Trac bug about it a while back, lemme dig up the ticket number...
18:59:27 <sysrqb> thanks
18:59:37 <sysrqb> we can think about this some more
19:00:10 <boklm> it might be possible to integrate guix into our build system to use it instead of Debian, although I didn't look at it much yet to know how it could work
19:00:21 <Jeremy_Rand_Talos> it's #32523
19:00:40 <sysrqb> ah
19:01:11 <sysrqb> okay. so it doesn't replace rbm
19:01:16 <pospeselr> ah ok
19:01:21 <Jeremy_Rand_Talos> Looks like that ticket is outdated, macOS and Windows are both merged to Bitcoin Core as well now
19:02:54 <sysrqb> okay, interesting
19:02:57 <sysrqb> thanks Jeremy_Rand_Talos
19:03:34 <Jeremy_Rand_Talos> no problem.  No idea if there are students interested in such a thing, but given that reproducible builds seem to be a hot topic, it seems plausible that there would be someone
19:04:14 <sysrqb> yup. okay, next item on the discussion list is mine.
19:05:04 <sysrqb> related to the comments earlier in this meeting, i started thikning about releasing alpa updates twice as frequently as stable
19:05:33 <sysrqb> meaning, alphas would have a two week release cycle, and stable would be every 4 weeks
19:05:51 <sysrqb> this doesn't include emergency releases, and such, as usual
19:06:37 <sysrqb> i think we could start this in mid-february
19:06:44 <mcs> is there anything we can do to reduce the time it takes to make a release? (to make that easier on the team)
19:07:01 <sysrqb> this is one reason why i'd like to start this
19:07:18 <sysrqb> before this is implemented, so i want to talk with GeKo and boklm about how we can make this process easier/faster
19:07:48 <Jeremy_Rand_Talos> (actually my bad, looks like macOS and Windows may still be WIP in the Guix stuff from Bitcoin Core, not sure why I thought it was already merged.  Looks like it's PR 17595 and 17920 on Bitcoin Core's repo.  Linux is definitely merged though.)
19:07:59 <sysrqb> i think having a four-week release cycle means we don't much much incentive to improve the situation
19:08:04 <mcs> another answer would be “make it more automated” so releases take less time away from other work
19:08:19 <boklm> what would be the goal for having more frequent alpha releases?
19:09:02 <sysrqb> mcs: yeah. that is one hope i have.as far as i know,  mozilla basically tag their release and let it build over night, and then they have a release ready to go by the morning
19:09:11 <sysrqb> i'd love to get closer to that
19:09:12 <boklm> I think the nightly updates should make it easier to use nightly to test new features
19:09:31 <sysrqb> boklm: yes, auto-updating nightlies are a big improvement
19:09:46 <sysrqb> but i'd like to get new features out to alpha users sooner, too
19:10:13 <sysrqb> maybe some alpha users will begin using nightlies when auto-updates are enabled
19:10:27 <sysrqb> and that'll help with receiving feedback faster, too
19:10:42 <boklm> I think it might be possible to improve a little the release process, but there will still be a lot of manual work to build and do the signing
19:10:42 <sysrqb> but, as an example, mozilla release beta updates every week
19:10:46 * GeKo stumbles in
19:10:53 <mcs> I assume we will ask more people to please run off the nightly channel once we have updates going there
19:10:55 <sysrqb> o/
19:11:31 <sysrqb> mcs: yeah, but nightlies doesn't have the same build integrity as alpha
19:11:42 <sysrqb> so i don't think it's a good product for everyone
19:11:43 <GeKo> i think two week cycles are doable from the release side of things
19:11:56 <GeKo> right now i wonder though whether we have enough meat for those
19:12:05 <GeKo> given review backlog and such
19:12:08 <sysrqb> GeKo: i don't think we do right now
19:12:15 <GeKo> yeah
19:12:37 <sysrqb> but in 1 month, i think we'll have a few s27 features we might want to land in different releases
19:12:44 <sysrqb> and then maybe s30 soon after those
19:12:54 <sysrqb> we can see as we get closer
19:13:08 <sysrqb> we don't need to release an update if we don't have any new features
19:14:43 <sysrqb> okay, we can think about this over the next couple weeks
19:15:06 <sysrqb> brade: good question
19:15:22 <sysrqb> for next week's meeting, i was going to ask you, mcs, or pili to run this meeting
19:15:48 <brade> will it only be the three of us?
19:16:15 <sysrqb> it should be you all, plus acat, boklm, sisbell
19:16:18 <sysrqb> maybe Jeremy_Rand_Talos
19:16:39 <sysrqb> GeKo: pospeselr, antonela, and I will be at all-hands
19:17:02 <Jeremy_Rand_Talos> Yes, I should be here next week
19:17:04 <sysrqb> if you don't have much to discuss, then you can skip next week's meeting
19:17:19 <mcs> maybe it will be a quick meeting, or maybe not :)
19:17:24 <sysrqb> but maybe there's something you can talk about :)
19:17:26 <sysrqb> yep
19:18:42 <sysrqb> okay, anyone want to discuss anything else that's Tor Browser-related?
19:18:55 <sysrqb> otherwise i'll call it.
19:19:00 <brade> was there something from the sponsor 27 meeting?
19:19:08 <sysrqb> oh. yes.
19:19:10 <sysrqb> thanks brade
19:19:18 <brade> (I have forgotten)
19:19:35 <sysrqb> pospeselr: acat: are either of you interested in working on the browser ui changes
19:19:41 <Jeremy_Rand_Talos> sysrqb, just wanted to apologize for not yet contacting you about the two things in my status update, I'll do it this week
19:19:47 <pospeselr> yeah which ones?
19:19:52 <sysrqb> acat: pospeselr err, urlbar indicator
19:20:08 <sysrqb> pospeselr: i think you might've already had some input on that
19:20:19 <pospeselr> yeah iirc i commented on taht ticket
19:20:22 <sysrqb> antonela: wants to know who she should bug about it :)
19:20:37 <antonela> well, work together :3
19:20:42 <sysrqb> hehehe
19:20:50 <pospeselr> hmm i can take it if i must
19:20:55 <pospeselr> ;)
19:21:01 <sysrqb> there is no must here
19:21:12 <sysrqb> we can also talk about this next week
19:21:24 <antonela> yes yes
19:21:31 <pospeselr> sorry I mean I can take it, esp since we'll probably have some downtime waiting for NoScript changes
19:21:45 <sysrqb> true
19:21:48 <sysrqb> thanks
19:21:58 <pospeselr> do you have the ticket # on hand?
19:21:59 <sysrqb> Jeremy_Rand_Talos: no worries, i'm overloaded anyway
19:22:08 <sysrqb> #32645
19:22:14 <sysrqb> and some children
19:22:30 <sysrqb> which aren't added yet
19:22:35 <antonela> we should probably want to take a look at #30025 childrens
19:22:37 <antonela> yes that
19:22:42 <sysrqb> :)
19:22:48 <antonela> somehow all this is related but they are different pieces
19:22:54 <antonela> we can map out all the mess pospeselr
19:22:57 <sysrqb> haha
19:23:02 <sysrqb> yep
19:23:08 <pospeselr> you free next week? :D
19:23:16 <antonela> haha
19:23:17 <sysrqb> sounds like a date
19:23:32 <antonela> haha
19:23:50 <sysrqb> okay. thanks for reminding me brade
19:23:54 <sysrqb> okay, anything else?
19:24:27 <sysrqb> heading nothing...
19:24:29 <sysrqb> #endmeeting