18:59:48 <GeKo> #startmeeting tor-browser
18:59:48 <MeetBot> Meeting started Mon Feb  5 18:59:48 2018 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:59:48 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
18:59:57 <GeKo> hi all and welcome to a new meeting
19:00:01 <mcs> hi everyone
19:00:20 <igt0> :)
19:00:23 <GeKo> the meeting pad is asusual at https://storm.torproject.org/shared/tHoN4Ii7rLSjPE0OP4gydX4cMGadsXmRQNc-6lwru0N
19:00:41 <GeKo> please fill in your updates in case you did not have yet and read through the things mentioned there
19:00:44 <sysrqb> o/
19:00:56 <pospeselr> good morning!
19:01:25 <isabela> o/
19:01:33 <isabela> sorry i couldnt meet last wed
19:01:46 <arthuredelstein> hi all
19:03:19 <arthuredelstein> np -- are you better?
19:04:49 <isabela> yes i am :)
19:06:00 <GeKo> alright
19:06:06 <GeKo> sysrqb: you are up
19:06:15 <sysrqb> GeKo: yep!
19:07:13 <sysrqb> so, small patches for running try builds
19:07:19 <sysrqb> any objections to that?
19:07:30 <sysrqb> i can open a ticket for review
19:07:48 <sysrqb> but anything I should think about while I'm doing this?
19:08:08 <arthuredelstein> So you're proposing patches so that tor-browser.git can pass try without many changes?
19:08:59 <sysrqb> arthuredelstein: more like patches so the builds use our .mozconfig*
19:09:05 <sysrqb> rather than the in-tree mozconfigs
19:09:12 <sysrqb> which use different configure options
19:10:04 <tjr> Hm
19:10:06 <GeKo> hm
19:10:09 <GeKo> :)
19:10:12 <sysrqb> does that make sense?
19:10:17 <tjr> This is for android I assume?
19:10:25 <sysrqb> i was thinking about in general
19:10:51 <tjr> Okay. So the intree config for mingw is https://searchfox.org/mozilla-central/source/browser/config/mozconfigs/win32/mingw32 and I maintain that
19:11:07 <sysrqb> it seems like we'd want to run the test suite against the tor browser we're shipping
19:11:20 <sysrqb> rather than a debug/release config that include the wrong options
19:11:29 <tjr> Okay, so I think that's a different thing from modifying the mozconfig
19:11:48 <tjr> We could (probably) get the Tor Browser build going in Try and then run the unit test suite with it
19:12:11 <sysrqb> yeah
19:12:30 <tjr> It would be a build off tor-browser.git; maybe using reproducible builds maybe not, but definetly using your toolchain
19:12:30 <sysrqb> i have it nearly working now
19:12:40 <sysrqb> hmm
19:12:44 <sysrqb> a
19:12:45 <sysrqb> ah
19:12:59 <sysrqb> that would be a bit more involved than my plan
19:13:05 <sysrqb> but maybe better
19:13:08 <arthuredelstein> Isn't boklm doing something like this?
19:13:23 <sysrqb> basically i was just thinkig we could run try builds using our mozconfigs
19:13:36 <sysrqb> like, right now the gtk2 build fails on try
19:14:01 <sysrqb> because on try, it builds with -Werror and there are some unused ariables
19:14:04 <tjr> Oh boy, now we're going down a rabbit hole - what's the gtk2 build? And why does the mozconfig affect it?
19:14:05 <sysrqb> variables
19:14:10 <sysrqb> so small thinks like that
19:14:26 <sysrqb> Tor Browser bulds with gtk2...
19:14:27 <sysrqb> builds
19:14:36 <GeKo> sysrqb: i am not sure that's worth the effort
19:15:31 <tjr> I.... don't think the mozconfig is controlling that. Well, unless your cascading through mozconfigs and picking up ac_add_options --enable-warnings-as-errors
19:15:33 <sysrqb> it seems like runnign try builds with the wrong mozconfig could give a false result
19:15:59 <sysrqb> tjr: seems like --enable-warnings-as-error defaults as true if MOZ_AUTOMATION=1
19:16:25 <tjr> Yes, that's correct. You'll want to get ac_add_options --disable-warnings-as-errors in your mozconfig if you're building with MinGW
19:16:56 <sysrqb> yeah makes sense
19:17:11 <sysrqb> maybe this is a bit of a rabbit hole
19:17:38 <tjr> If we want to run the test suite with Tor Browser though - I think we should replicate the torb browser build process as closely as possible though. Which would involve using the toolchain, stuff like the gcc spec thing you're doing: https://gitweb.torproject.org/builders/tor-browser-bundle.git/tree/gitian/descriptors/windows/gitian-utils.yml#n93
19:17:52 <GeKo> yeah
19:18:01 <tjr> Right now I'm trying to get the MinGW build running without that type of toolchain stuff, using just a 'vanilla' mingw setup
19:18:21 <tjr> And my hope is that this will knock one or two of the weird things you have to do off
19:18:21 <GeKo> sysrqb: i think we should revisit that once we start using esr60
19:18:24 <sysrqb> I was looking for an easy win
19:18:27 <tjr> Although the spec one probably not.
19:18:30 <sysrqb> GeKo: okay, fine with me
19:18:39 <GeKo> to see how much is neede to get that going
19:18:43 <tjr> Also trying to get jemalloc supporting in mingw
19:18:44 <GeKo> starting from a "clean slate"
19:19:00 <GeKo> and thinking whether the effort is worth it
19:19:08 <sysrqb> yes, makes sense
19:19:26 <sysrqb> tjr: cool
19:19:47 <sysrqb> okay, point 2, we should create a roadmap for Tor Browser for Android
19:20:11 <sysrqb> what functionality/qualities should it have before we release?
19:20:25 <GeKo> the key items are:
19:20:31 <GeKo> 1) reporoducible builds
19:20:42 <GeKo> 2) same day releases as desktop browser
19:20:55 <GeKo> 3) same tracking/fingerprintin defenses as desktop
19:21:05 <GeKo> 4) same security properties
19:21:27 <GeKo> (this includes things like showing the browser window only after tor got configured)
19:21:34 <GeKo> 5) same usability
19:21:50 <GeKo> we could and should break those down, agreed
19:22:04 <GeKo> but I think the DRL proposal explained what we wanted
19:22:14 <sysrqb> Okay, so integrating Tor Launcher is a requirement
19:22:24 <sysrqb> okay, I haven;t seen the proposal
19:22:25 <GeKo> i wonder if i should send that around so that we can start thinking from that on
19:22:30 <sysrqb> (might be useful)
19:22:39 <sysrqb> yeah, can you please?
19:22:39 <igt0> +1
19:22:46 <GeKo> i'll dig up the latest thing i have and send it to you folks
19:22:52 * GeKo makes a note
19:23:18 <sysrqb> igt0 and I discussed las month whether we should ship the first release with requiring Orbot
19:23:31 <sysrqb> or integrate Tor Launcher
19:23:44 <sysrqb> we thought integrating TL would add a lot more time
19:23:57 <arthuredelstein> I was thinking it would be useful for sysrqb and igt0 to write up a short document on their overall plan/design/roadmap to make sure we're all on the same page.
19:23:57 <sysrqb> so using Orbot (as Orfox currently does) would allow us to ship earlier
19:24:19 <sysrqb> but if TL is neded, then we'll put that as a blocker for shipping
19:24:23 <sysrqb> *needed
19:24:27 <GeKo> sysrqb: feel free to argue for it
19:24:33 <GeKo> what arthur said
19:24:43 <sysrqb> yeah
19:24:58 <mcs> It would also remove the Guardian Project off the hook for maintenance sooner.
19:25:00 <sysrqb> Froma usability perspective, I think using TL is good
19:25:01 <GeKo> i'd especially like to understand the benefits and disadvantages
19:25:09 <mcs> of get them off the hook :)
19:25:12 <GeKo> mcs: that's a good point
19:25:15 <sysrqb> yeah
19:25:30 <sysrqb> I think working on a new mobile interface will take time
19:25:32 <GeKo> so, i general i would be amenable doing the first alpha with orbot
19:25:51 <GeKo> it's an alpha after all
19:26:07 <GeKo> and a bunch of the new stuff could already get tested while we are working on a tighter tor integration
19:26:13 <sysrqb> I don't particularly like requiring Orbot, but it could be a small stopgap while we still work on the new integration
19:26:25 <sysrqb> yeah
19:26:43 <sysrqb> okay, that sounds good
19:26:52 <GeKo> so, write up what you have in mind with some analysis and post it to tbb-dev?
19:26:54 <sysrqb> I'll work with igt0 on a roadmap
19:27:04 <sysrqb> yes, we'll do that
19:27:11 <arthuredelstein> It could be useful to think about what can be a "Minimum Viable Product", as distinct from final goals of perfect parity with tbb desktop
19:27:17 <GeKo> i'll send you guys the proposal i have and you can incorporate things from it
19:27:25 <GeKo> yes
19:27:26 <sysrqb> arthuredelstein: exactly, agreed
19:27:36 <sysrqb> GeKo: thanks
19:27:45 <arthuredelstein> (maybe an MVP would just be an alpha; idk)
19:28:20 <sysrqb> if we ship an alpha, then we can improve on that
19:28:26 <sysrqb> but at least we have something out the door
19:28:45 <sysrqb> but we can discuss this more on the mailing list
19:29:14 <GeKo> sysrqb: okay, your last item is for isabela?
19:29:45 <sysrqb> (i don't want to consume all the meeting time on this topic, if there's more to discuss)
19:29:59 <sysrqb> urg, lag
19:30:02 <sysrqb> GeKo: yes
19:30:07 <sysrqb> isabela: did you see that?
19:30:27 <sysrqb> we can discuss that on the mailing list, too
19:30:51 <sysrqb> i just want to make sure we're on the same page for UI work
19:30:58 <GeKo> sounds good
19:31:29 <sysrqb> but i'm good, thanks
19:31:37 <sysrqb> we can move on
19:31:41 <GeKo> k
19:31:59 <GeKo> igt0: re your issue
19:32:07 <isabela> sysrqb: ops no
19:32:19 * antonela is lurking
19:32:33 <isabela> sorry sysrqb
19:32:35 <isabela> reading backlog
19:32:39 <GeKo> i think getting about:tor to work in mobile constraints sounds like a  good idea
19:32:47 <GeKo> there is no design document for it
19:33:00 <isabela> yes
19:33:04 <GeKo> if there are a bunch of changes needed we should talk to the ux folks
19:33:08 <isabela> about:tor should be responsive design
19:33:18 <mcs> Maybe do something minimal with about:tor for now while we await a Great New Design?
19:33:22 <isabela> but we will redo all about:tor
19:33:28 <GeKo> mcs: +1
19:33:30 <isabela> mcs: yes
19:33:34 <isabela> that sounds good
19:33:50 <isabela> we are taking that with the introduction of circuit displays new location
19:34:08 <isabela> temporary fix till new about:tor is out there
19:34:53 <igt0> great, the current about:tor page UX looks simple enough to fit in the mobile screen. (The two white boxes could be vertical aligned in the mobile version).
19:35:06 <isabela> yep
19:35:09 <sysrqb> +1
19:35:34 <isabela> sysrqb: btw i also think is ok to get an alpha out there, i am happy to test apks too
19:35:44 <arthuredelstein> the onion browser startup screens are super cute
19:35:45 <sysrqb> isabela: thanks :) cool
19:35:55 <sysrqb> ( antonela:  hi! :) )
19:36:03 <sysrqb> arthuredelstein: +1
19:36:12 <antonela> hi :) i was talking with igt0 about those
19:36:13 <sysrqb> and I'd like to do something similar
19:36:26 <sysrqb> but that'll take some time to port over to android
19:36:39 <isabela> the circuit display?
19:36:41 <sysrqb> and if we want a similar experience on desktop
19:37:13 <sysrqb> that'll all go into a tor launcher redesign, i think
19:37:15 <mcs> Maybe we could display some of the Onion Browser graphics on the Tor Launcher desktop progress screen since we have space :)
19:37:25 <mcs> (someday)
19:37:36 <sysrqb> yeah
19:37:36 <isabela> so
19:37:40 <isabela> about that
19:38:07 <isabela> i think we should think about it a bit more because onion browser was fixing a problem with the restrains of a experience
19:38:21 <sysrqb> yeah, for sure
19:39:15 <sysrqb> but i think it's a nicer and more friendly experience than the current design
19:39:28 <isabela> i think the idea is good but i would prefer to have better feedback on progress bar that leads to easily debugging from user
19:39:32 <isabela> if they have issued
19:39:34 <mcs> I think we will learn a lot from the TBA work and some ideas can be fed back into desktop later.
19:39:34 <isabela> *issues
19:39:38 <isabela> like we were moving on with.
19:39:48 <sysrqb> yep
19:39:51 <mcs> agreed on the progress bar feedback issue
19:40:13 <arthuredelstein> I do too. I just liked the artistry in onion browser but I don't think we want to do exactly the same thing
19:40:36 <isabela> and for onboarding i would like to talk about the product features and where they are how to use them
19:40:43 <isabela> arthuredelstein: yes!
19:40:53 <isabela> arthuredelstein: antonela is on process of getting tor illustrations !
19:41:03 <arthuredelstein> yay! :)
19:41:03 <sysrqb> nice!
19:41:11 <isabela> ux has been cooking a whole visual strategy for illustration
19:41:21 <antonela> yes, is part of the website redesign work
19:41:23 <sysrqb> Mmmm
19:41:57 <isabela> so user can see a illustration and connect the dots, like where we talk about relays or bridges or onion services
19:42:26 <isabela> i think we should for sure spend some time in rome talking about all this
19:42:42 <GeKo> sounds good
19:42:53 <isabela> and make sure everyone is on the same page on things like this because I know ppl have tons of ideas too :)
19:43:21 <GeKo> okay
19:43:37 <GeKo> let's move on. it seems my item is the next one unanswered
19:44:01 <GeKo> so, it seems we have crashes exclusively on windows vista related to us enabling content sandboxing on windows
19:44:15 <GeKo> which is quite unfortunate even though vista is long EOL
19:44:35 <GeKo> i think we should fix that somehow and have been pondering what to do
19:44:48 <arthuredelstein> do we have a ticket #?
19:44:57 <mcs> Have we confirmed that disabling sandboxing avoids the crash?
19:44:58 <GeKo> #25112
19:45:05 <arthuredelstein> thanks
19:45:14 <GeKo> i know from at least one user that the alpha is working for them
19:45:35 <GeKo> and the error messages we got as feedback strongly indicate some sandbox related issue
19:46:02 <mcs> So our choices are to debug and fix it or disable the sandbox on Vista?
19:46:23 <GeKo> i think queation 1) would be if any of us has a vista system still around to confirm that from first hand
19:46:27 <GeKo> mcs: yes
19:46:45 <GeKo> if am torn to be honest
19:47:00 <mcs> On the one hand the Vista people probably need a sandbox; on the other hand, they should upgrade their OS.
19:47:15 <GeKo> yes
19:47:25 <pospeselr> is Vista even getting security updates anymore?
19:47:41 <GeKo> no
19:47:48 <pospeselr> ewwwwww
19:47:56 <GeKo> nor does xp
19:48:02 <tjr> But Windows 7 does?
19:48:04 <GeKo> yet both are supported with est52
19:48:05 <GeKo> yes
19:48:14 <GeKo> *esr52
19:48:42 <pospeselr> well worse case scenario someone can always throw up a Vista VM to debug in
19:49:11 <mcs> extended support for Win7 ends in about 2 years (so says https://support.microsoft.com/en-us/help/13853/windows-lifecycle-fact-sheet)
19:50:01 <GeKo> okay. does anyone of us feel like they want to look closer into that?
19:50:11 <GeKo> because i won't have the cycles
19:50:27 <pospeselr> yeah I can do
19:50:33 <GeKo> awesome
19:50:55 <GeKo> ping me if you got things going and need some help with debugging
19:51:09 <GeKo> because debugging that on windows is not easy
19:51:19 <GeKo> for one debug builds are not running :)
19:51:48 <GeKo> but maybe i just don't know all those windows debugging sekrits...
19:51:52 <GeKo> thanks
19:51:56 <sysrqb> heh
19:52:03 <GeKo> let's move on do the "discussion"
19:52:16 <pospeselr> do we have symbols for windbg or am I gonna have to figure out how to use gdb on windows?
19:52:26 <GeKo> most importantly: which seesions do we want to have at the meeting day
19:52:29 <GeKo> pospeselr: the latter
19:52:34 <pospeselr> awesome sauce
19:52:39 <GeKo> yeah
19:53:17 <GeKo> i worked around that by using mozilla's own logging without any debugger to get the sandbox build going
19:53:26 <GeKo> it might be faster that way
19:53:33 <GeKo> ok. meeting day
19:53:47 <GeKo> we have the roadmapping but what else
19:54:01 <isabela> ux is free to hang out with yall
19:54:09 <isabela> we will do our thing before the 11
19:54:22 <mcs> Will we have any Mozilla coordination meetings that day? Will Mozilla people be there?
19:54:59 <tjr> Some Mozilla people will be there
19:55:13 <mcs> :)
19:55:14 <GeKo> assuming we want to start at 11am (i guess we could do 10am as well?)
19:55:16 <pospeselr> (have we heard anything re Ethan and company?)
19:55:26 <GeKo> we have 6 slots
19:55:27 * isabela would suggest to leave roadmap for the end of the day as other discussions might actually help w/ it
19:55:40 <GeKo> or 5 if there are some short breaks
19:55:59 <GeKo> isabela: sounds good
19:55:59 <isabela> hard stop at 5 - as folks will need a break before dinner thing
19:56:10 <GeKo> or go to the vegas meeting
19:56:16 <isabela> :)
19:56:30 <mcs> I am not sure if it makes sense for the team day, but I’d love for someone to teach me about Android vs. Deskop Firefox architecture/constraints/etc.
19:56:43 <arthuredelstein> +1
19:56:47 <mcs> (I guess teaching / chaulk talk could go both ways)
19:56:49 <sysrqb> mcs: sure
19:56:54 <sysrqb> we can do a run through that day
19:57:00 <sysrqb> (or another day, if we're fully booked)
19:57:04 <GeKo> +1
19:57:14 <arthuredelstein> another one we mentioned last time is a fingerprinting discussion
19:57:25 <GeKo> yeah
19:58:00 <GeKo> so i think 1) roadmapping 2) ux team coordination 3) fingerprinting 4) mozilla coordination should be on the list
19:58:10 <GeKo> then we'd have 1-2 additional slots
19:58:35 <GeKo> what about a retrospective meeting?
19:58:39 <arthuredelstein> I'm not sure if security slider/noscript goes under ux
19:58:41 <mcs> We might also think about “What could the Network Team do that would help us make Tor Browser better?” (and then talk to them later about the items we dream up).
19:58:43 <sysrqb> is it worth having a discussion about anhardening?
19:58:51 <isabela> (I will share a place to start drafting roadmap like we did before montreal)
19:58:51 <mcs> Restrospective might be good to do.
19:58:53 <GeKo> about things that went well and what did not go well?
19:59:22 <isabela> +1 on that
19:59:26 <mcs> yes, +1
19:59:27 <GeKo> okay, let's try the retrospective one ad item 5)
19:59:36 <GeKo> *as
19:59:50 <sysrqb> also +1 for meeting with the networing team
20:00:03 <sysrqb> but may not be a long meeting
20:00:49 <GeKo> okay, thanks for the input i'll think a bit harder about it and then add the items we want to have
20:00:59 <GeKo> finall thing on the pad: next meeting
20:01:08 <GeKo> i'll be afk next week (i hope)
20:01:27 <GeKo> and i heard that the 19th is a public US holiday!1!!
20:01:45 <GeKo> sooo, i'd propose to move the meeting on tuesday 20th same time same place
20:01:54 <GeKo> does that work for everyone?
20:02:04 <sysrqb> good for me
20:02:12 <igt0> yep
20:02:25 <mcs> works for me
20:02:38 <pospeselr> works for me!
20:02:44 <arthuredelstein> me too
20:03:00 <GeKo> great. tjr? sounds good too?
20:03:26 <tjr> Sure
20:03:26 <GeKo> i'll send an update on the mailing list announcing the change
20:03:30 <GeKo> great
20:03:45 <GeKo> so, we are late but are there any items we forgot to talk about?
20:04:12 <GeKo> okay, if not, then *baf* and thanks for the meeting
20:04:17 <GeKo> #endmeeting