17:58:59 <GeKo> #startmeeting tor browser
17:58:59 <MeetBot> Meeting started Mon May 21 17:58:59 2018 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:58:59 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:59:04 <GeKo> boom, meeting time!
17:59:12 <GeKo> hello everyone
17:59:14 <boklm> hi!
17:59:16 <arthuredelstein> hi everyone!
17:59:17 <sukhe> hello!
17:59:18 <isabela> buenas
17:59:23 <igt0> !
17:59:26 <GeKo> the pad as usual is on: https://storm.torproject.org/shared/tHoN4Ii7rLSjPE0OP4gydX4cMGadsXmRQNc-6lwru0N
17:59:48 <mcs> hi
18:00:29 <GeKo> please mrke the items you want to talk about bold
18:00:33 <GeKo> *mark
18:01:25 <pospeselr> hello
18:02:28 <isabela> ok i am keeping it simple with one item :)
18:02:36 <GeKo> heh
18:02:42 <GeKo> let's get started
18:03:30 <GeKo> esr60-based nightly builds are the item i have
18:03:47 <GeKo> we should get that going before the weekend
18:04:30 <GeKo> the blockers for that are #25543, #25750, 26073 and #26100
18:04:49 <GeKo> #26073
18:05:05 <GeKo> at least that's the 4 bugs i have on the radar
18:05:25 <arthuredelstein> yes, I think those are the 4
18:05:50 <GeKo> if we are good we can include the adapted patches for #24309 and #23247
18:06:01 <GeKo> because antonela is out somewhere doing user testing
18:06:06 * antonela is around, between aiports
18:06:17 <sysrqb> :)
18:06:17 <antonela> yep, this should happen this week actually
18:06:23 <GeKo> and having those features right in the nightly tested can't hurt
18:06:25 <GeKo> yes
18:06:36 <arthuredelstein> I think the end of this week is doable -- is that soon enough for you antonela?
18:06:55 <antonela> well, we have our first meeting with people tomorrow
18:07:10 <isabela> antonela: !
18:07:12 <arthuredelstein> Aha!
18:07:20 <antonela> if i have it for thursday/friday we can do it
18:07:36 <GeKo> yeah, it's getting tight
18:07:45 <antonela> if not, i'll go with my static mockups
18:08:02 <GeKo> arthuredelstein: you seem to be involved with most of those patches
18:08:10 <GeKo> if it helps reroute things in my direcion
18:08:14 <antonela> sorry, hola :)
18:08:17 <GeKo> *direction
18:08:22 <arthuredelstein> One option could be to build a TBB/ESR52 with the new UX patches
18:08:34 <GeKo> e.g. i certainly can look at #24309
18:08:47 <GeKo> well, it depends what's more useful
18:09:32 <GeKo> antonela: ^
18:09:52 <antonela> we will test .onion indicators and new circuit display
18:10:12 <antonela> esr60 seems to be better, but if we dont arrive, esr52 can make the job
18:10:45 <antonela> let me know what is better for you, geko, arthur
18:11:01 <GeKo> how about this, i review what we have tomorrow and commit it
18:11:07 <GeKo> (assumng we are good)
18:11:19 <GeKo> then we could probably trigger a nightly out of band
18:11:24 <GeKo> or wait for another day
18:11:36 <boklm> yes, I can trigger the nightly build after you push the commit
18:11:37 <GeKo> and meanwhile we try to get everything done for esr 60
18:11:50 <antonela> sounds good for me
18:11:52 <antonela> thanks a lot!
18:11:58 <GeKo> so if we are lucky you might be able to test esr60 with all the stuff
18:12:03 <GeKo> okay, good
18:12:04 <antonela> great
18:12:05 <arthuredelstein> GeKo: Sorry, a nightly out of band for ESR60?
18:12:31 <GeKo> no, based on esr52 just with the patches for stuff antonella wants to test
18:13:00 <arthuredelstein> Ah, OK. I think that makes sense. Then we can send on the ESR60 stuff when it is ready, if we get it done by Wednesday or so.
18:13:09 <GeKo> and out-of-band means here only that we trigger one without waiting for the next regular build
18:13:19 <GeKo> arthuredelstein: exactly
18:13:24 <arthuredelstein> sounds good
18:13:41 <GeKo> pospeselr: could you post your esr52 patch for the .onion thing, too?
18:14:02 <GeKo> so that i can look at it tomorrow?
18:14:10 <pospeselr> yeah for sure
18:14:30 <GeKo> great.
18:15:03 <GeKo> then we get the nightly for antonela tomorrow and try to get esr60 linux nightlies going later this week
18:15:09 <antonela> cool thanksss 🙏
18:15:14 <GeKo> that's all i have for this item
18:15:46 <GeKo> igt0: wanna tell more about the orfox crash? is that still blocking the release?
18:17:17 <GeKo> sysrqb: or maybe you know?
18:17:25 <igt0> GeKo, yes, sysrqb and found a crash in the latest Orfox build. It happens in the js engine. Sadly we are not able to reproduce constantly. We need to stress out a website scrolling up and down for some time.
18:17:52 <GeKo> so, this is a firefox bug i assume?
18:18:09 <GeKo> does this happen with a vanilla fennec, too?
18:18:31 <igt0> yes, I asked tjr to make sure it was not reported and fixed. And a similar bug was fixed in early 2016.
18:18:50 <igt0> GeKo, i was not able to reproduce with fennec.
18:18:59 <GeKo> huh
18:19:10 <sysrqb> it seems like this crash is caused by one of the backport patches, but we're not sure which one
18:19:10 <GeKo> and the previous orfox version?
18:19:22 <sysrqb> we can't reproduce with the current Orfox release
18:20:31 <GeKo> so, some patch that got fixed between 52.7.4 and 52.8.0 is causing this to be triggered?
18:20:45 <GeKo> because our patches did not change between those versions afaict
18:20:50 <sysrqb> not many patches touched js code logic, so it is surprising orfox is crashing
18:21:21 <sysrqb> right, and i did not see a creash with tor browser on the same site where orfox crashed
18:22:32 <GeKo> hrm.
18:23:13 <sysrqb> but i did cherry-pick some backports from alpha branch for this new release
18:23:47 <igt0> The problem is also because the Orfox crashes in a sanity check (MOZ_CRASH). So we don't have a stack trace.
18:24:07 <sysrqb> so if igt0 can create a list of sites that consistently crash orfox, then we can bisect and find which commit
18:24:22 <tjr> You ought to be able to attach a debugger and get a stracktrace from a MOZ_CRASH....
18:24:53 <sysrqb> hrm, igt0, we can try jimdb
18:25:02 <sysrqb> good though tjr
18:25:10 <sysrqb> thought
18:26:07 <igt0> I mean, we know where it crashes, because of the MOZ_CRASH, however if it reaches that part of the code, the js engine is already messed up.
18:26:14 <sysrqb> but will that give us any javascript stack info?
18:26:41 <igt0> nay
18:27:29 <tjr> Ah okay
18:27:37 <GeKo> okay, keep us posted about this
18:27:48 <GeKo> seems pretty urgent to get fixed i guess
18:28:13 <GeKo> alas it's a bit hard to help as i am somewhat unclear what you have not tried yet
18:28:34 <sysrqb> yes, it'll be our top priority this week
18:28:40 <tjr> I would ask in jsapi to be sure, there is a function called DumpJSStack() that might be able to be called from a crash in gdb and give you a JS stack
18:28:42 <GeKo> sysrqb: have you reverted the commits you cherry-picked for that release and checked whether one of those is the issue?
18:29:02 <igt0> tjr, oh great!, it will be helpful
18:29:12 <sysrqb> nice
18:29:39 <sysrqb> GeKo: no, i don't think we reverted - igt0, right?
18:30:36 <GeKo> alright, thanks for the update
18:30:53 <igt0> I reverted few patches, however because I am not able to reproduce constantly, i am not sure if the reverted patch is the culprit.
18:30:54 <GeKo> isabela: you are up!
18:31:35 <isabela> yes
18:31:54 <isabela> i will start the process to extend our contract with sponsor9
18:32:10 <GeKo> *8
18:32:12 <isabela> and wanted to check if you would want it to be dec 31 the end date
18:32:13 <isabela> ops
18:32:14 <isabela> yes 8!
18:32:40 <GeKo> sounds good to me
18:33:05 <isabela> if you want more time than that is ok too
18:33:08 <GeKo> we had the stable tor browser for mobile as kind of christmas present in Rome :)
18:33:22 <GeKo> well i doubt we'll be done much earlier
18:33:47 <sysrqb> me too
18:33:49 <isabela> i meant to say later than dec 31
18:34:25 <GeKo> i think starting with dec 31 sounds like a good idea
18:34:47 <isabela> they have a bit more paper work than otf
18:35:22 <isabela> ok
18:35:25 <isabela> i will go with dec 31
18:35:53 <GeKo> well, i am fine with march 31 if the risk is doing another round of paperwork end of dec
18:36:18 <GeKo> but i am a bit hesitant to do that tbh
18:36:25 <isabela> np
18:36:43 <GeKo> k
18:36:47 <GeKo> arthuredelstein: you are up
18:36:53 <isabela> lets go with dec 31 (i will also coordinate this with the sponsor and follow their guidance/feedback)
18:36:57 <GeKo> thx
18:37:10 <arthuredelstein> We have a bunch of regression test patches in tor-browser.git
18:37:16 <arthuredelstein> most for fingerprinting checks
18:37:38 <arthuredelstein> But the fingerprinting patches are already uplifted with their own regression tests in mozilla-central
18:37:55 <arthuredelstein> So the question is whether we want to hold on to our own regression tests or just drop them.
18:38:13 <arthuredelstein> The only advantage in keeping them is redundancy, in case a something is removed or missed at Mozilla's end.
18:38:41 <GeKo> well, the problem is in part that our tests don't match the final implementation in mozilla anymore
18:38:54 <GeKo> i think that it true at least for some of the tests?
18:39:03 <GeKo> *is
18:39:10 <arthuredelstein> That might be the case -- I'm actually not sure.
18:39:30 <tjr> I am pretty sure that some of them will break :)
18:39:33 <tjr> User Agent for example
18:39:37 <GeKo> yep
18:39:52 <GeKo> so that makes them less useful to begin with
18:40:06 <GeKo> and we should not spend time fixing our redundancy tests
18:40:21 <arthuredelstein> That's my inclination as well
18:40:31 <tjr> media statistics (if you have that test) probably, canvas prompt, keyboard layout...
18:40:42 <GeKo> and then there is the issue that we aren't running those tests anyway right now
18:40:56 <arthuredelstein> yes, exactly
18:40:57 <GeKo> at least in a regular manner
18:41:14 <GeKo> so, all in all it does not sound that useful to keep them around
18:41:20 <arthuredelstein> I think we'd be better off focusing on writing out-of-band tests that work in any browser in any case.
18:41:34 <GeKo> i agree
18:42:08 <arthuredelstein> OK! So I will drop them from the  #25543 branch
18:42:28 <GeKo> yes, sounds good
18:43:00 <GeKo> okay, anything else before we move to the discussion part?
18:43:34 <GeKo> okay, disussion then.
18:43:46 <GeKo> next meeting date: next monday is us holiday
18:44:07 <GeKo> does everyone have time on tuesday to do our weekly meeting?
18:44:19 <GeKo> same time and same place?
18:44:26 <sysrqb> that's good for me
18:44:33 <boklm> that's ok for me
18:45:30 <GeKo> hearing no complaints so far, good
18:45:31 <mcs> I am not 100% sure if brade and I will be there but Tuesday is better than Monday, for sure.
18:45:44 <pospeselr> works for me!
18:46:02 <arthuredelstein> me too
18:46:02 <GeKo> then let's tentatively plan for that day
18:46:12 <GeKo> i'll send an email to tbb-dev tomorrow probably
18:46:23 <GeKo> the second item: all hands topics
18:46:36 <GeKo> tjr: looks like a  good list to me
18:46:56 <sysrqb> tjr this is what i said the other say:
18:46:56 <tjr> Product Discussion I'm leaning on Wennie to arrange
18:46:57 <sysrqb> 14:42 < sysrqb> snorp, nalexander, anyone else who can help us help them help everyone
18:47:00 <sysrqb> 14:44 < sysrqb> it seems ike mobile is in a weird state right now, but if we can discuss how we can upstream fennec patches (particular who can review them)
18:47:03 <sysrqb> 14:44 < sysrqb> that'd be pretty worthwhile
18:47:19 <sysrqb> err, *day :)
18:47:33 <tjr> Fingerprinting / Tor Uplift I'll probably delay scheduling until Ethan has started and I've synced with him
18:47:41 <isabela> should we ping network team to add stuff to this all hands list?
18:47:41 <tjr> Network Programming I need t email about
18:47:52 <tjr> UI/UX i have an email ready to go, same with sandboxing
18:48:28 <GeKo> tjr: let me or us know who you'd like to coordinate to get input for those sessions from our side
18:48:33 <tjr> WebRTC I think we're set on, so long as Arthur or Richard knows what Tor would like to do in WebRTC land :)
18:48:52 <GeKo> not sure if a pad would help here
18:49:53 <tjr> probably
18:49:58 <tjr> just to make sure everyone's o the same page
18:50:21 <GeKo> yep
18:51:33 <GeKo> isabela: sounds good to me
18:51:47 <GeKo> i think having those items on the pad would be good as well
18:52:18 <GeKo> so that we have an overview of all our activitives/plans for the meeting at one place
18:52:39 <tjr> ++
18:52:49 <GeKo> isabela: could you poke the right people?
18:53:48 <GeKo> do we have anything else for today?
18:54:30 <GeKo> does not seem to be the case
18:54:38 <GeKo> thanks everyone then! *baf*
18:54:43 <GeKo> #endmeeting