17:59:07 <GeKo> #startmeeting tor browser july 30
17:59:07 <MeetBot> Meeting started Mon Jul 30 17:59:07 2018 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:59:07 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:59:12 <GeKo> hi!
17:59:13 <sukhe> hello!
17:59:15 <boklm> hi!
17:59:16 <sysrqb> o/
17:59:20 <GeKo> let's  get started
17:59:31 <pospeselr> hello!
17:59:34 <GeKo> the pad is over at https://storm.torproject.org/shared/tHoN4Ii7rLSjPE0OP4gydX4cMGadsXmRQNc-6lwru0N as usual
17:59:39 <mcs> hi
17:59:52 <GeKo> please meark items bold you want to talk about (after adding ouyr own items)
17:59:54 <sukhe> sorry, I am having issues with sandstorm
17:59:57 <igt0> hello
18:00:00 <sukhe> (is it just me?)
18:00:11 <pospeselr> it's laggy af
18:00:17 <GeKo> wfm
18:00:55 <sukhe> I think we can start since I am anyways at the end of the pad
18:00:56 <sisbell> hello
18:01:22 <sukhe> done
18:02:49 <GeKo> okay, let's look over the items
18:02:57 <GeKo> it seems i am first
18:03:17 <GeKo> releases planning:
18:03:40 <arthuredelstein> hi everyone
18:03:44 <GeKo> for the desktop alpha i was wondering whether we should postpone the release for aug 13/14
18:04:19 <GeKo> i am not really convinced we should push it out until aug 2 (assuming i should be helping with it)
18:04:57 <GeKo> my main reasons for that are the onboarding we might want to have i the next alpha to give it serious testing
18:05:07 <GeKo> and the win64 update issues makes me nervous
18:05:15 <mcs> maybe landing more fixes first makes senseā€¦ which would be another benefit (plus you being more available in case things go wrong).
18:05:31 <GeKo> i want to ship a potential fix in the next alpha to give it some testing in a nother alpha if we need it
18:05:33 <mcs> I also agree on the Win64 issue.
18:06:03 <GeKo> + i want to have http/2 enabled
18:06:47 <GeKo> okay, then let's keep working on all of those issues to have them ready/sorted out once i am back
18:06:51 <arthuredelstein> yes, I think postponing to aug 13/14 is a good idea
18:06:58 <GeKo> and then we do a release with what we have
18:07:00 <GeKo> good
18:07:19 <boklm> postponing it to aug 13/14 sounds good to me
18:07:25 <GeKo> then let's looks at the mobile angle
18:07:32 <GeKo> sysrqb: igt0: what do you think?
18:08:15 <sysrqb> in terms of what remains before the alpha is ready?
18:08:22 <GeKo> yes
18:08:45 <GeKo> and whether we should postpone releasing as well
18:08:57 <GeKo> we planned for this week-ish
18:09:38 <sysrqb> I think we have at least 1 week of work, but I think we could finish by the end of this week if we try very hard for it
18:09:52 <GeKo> looking at our blockers: the proxy bypass things and the onboarding are the remeaining ones
18:10:07 <igt0> I am not sure how I feel about http2 on mobile, sysrqb do you know anything about it?
18:10:30 <sukhe> can anyone quickly point to the proxy bypass ticket?
18:10:31 <sysrqb> and deciding how we create the signing key and discuss the updating mechanisms
18:10:48 <sysrqb> https://trac.torproject.org/projects/tor/ticket/21863
18:10:57 <sukhe> ah the android one. thanks
18:11:03 <sysrqb> yeah :)
18:11:13 <sysrqb> i don't think http/2 is a blocker
18:11:22 <sysrqb> but we will want it, of course
18:12:41 <sysrqb> we may be able to push the updating ticket until the next alpha release
18:12:57 <GeKo> yes, i think so
18:13:04 <sysrqb> and only have "re-download and install" as the update mechanism
18:14:01 <sysrqb> #26242
18:14:26 <GeKo> sysrqb: igt0: should we plan for august 13/14 for doing the first tba release as well then?
18:14:26 <sysrqb> but if igt0's patch is working, then we can include
18:14:34 <sysrqb> i haven't tested it yet (#26242)
18:14:45 <sysrqb> GeKo: that sounds good to me
18:14:45 <GeKo> i am not sure where we are with the onboarding exactly
18:15:01 <GeKo> but we might want to use that aadditional time to give it more testing
18:15:17 <sysrqb> yes, plus that gives more time for creating the user manual
18:15:22 <GeKo> or iron out the most pressing issues
18:15:32 <GeKo> igt0: ^
18:15:36 <igt0> yep!
18:15:54 <GeKo> sysrqb: that's on our list? or is that a thing done by the community team?
18:16:15 <igt0> sysrqb, are you looking in the onboarding, right?
18:16:37 <igt0> GeKo, community team is taking a look, iirc ggus was the responsible for it.
18:17:00 <GeKo> okay, that's been my impression
18:17:07 <sysrqb> i haven't done anythig with the manual
18:17:17 <GeKo> it seems we have a sync on wed, so we can see where we are with it
18:17:24 <sysrqb> yes
18:18:05 <antonela> I updated #25696 last week
18:18:44 <GeKo> sysrqb: igt0: great. so, are we clear on who is working on the onboarding stuff on mobile?
18:19:36 <sysrqb> i was letting igt0 take the lead on that
18:19:49 <igt0> okey, I can take a look this week.
18:20:11 <GeKo> sounds good, given that you are working on the proxy bypass bugs (the other remaining blocker)
18:20:20 <sysrqb> :) (thanks!)
18:21:01 <GeKo> alright, we do bloth releases once i am back in the week from aug 13
18:21:12 <GeKo> let's work towards that with that "deadline" in mind
18:21:59 <GeKo> pospeselr: i think this reimbursement request does not sound unreasonable especially as we need good contacts to rust people etc.
18:22:27 <GeKo> i'd suggest asking travel@tpo about the reimbursement (feel free to cc me)
18:22:29 <pospeselr> excellent thanks!
18:22:54 <sysrqb> pospeselr: did you get a sponsored ticket?
18:23:09 <pospeselr> yeah!
18:23:15 <sysrqb> okay, cool
18:23:31 <GeKo> sukhe: you are up
18:24:16 <sukhe> hi. I have marked things I could use help with
18:24:30 <sukhe> let me summarize
18:25:31 <sukhe> https://trac.torproject.org/projects/tor/ticket/25485#comment:23 is what I am looking for, a small C++ program that _uses the latest CXXABI_ so that we compare and check to see if need to use our version of libstdc++ or not
18:26:21 <sukhe> which means we have to write a program that does that, except I can't figure out how to use "the latest CXXABI". I need someone who is good with C++ to tell me what the program should look like, because a simple program even compiled with -Wabi doesn't work
18:26:49 <Hello71> -W is a warning option
18:26:55 <Hello71> it doesn't affect the output
18:27:10 <sukhe> sorry, I miswrote
18:28:56 <sukhe> (I meant -fabi-version)
18:29:48 <boklm> I can look this week if I see a way to do this
18:30:20 <sukhe> thanks, also I left a comment for you re: #12968.
18:30:43 <sukhe> (fwiw, I tried building with https://github.com/gcc-mirror/gcc/commit/f47fc7ef7f52cd095e636d4f93cca052427f3f0a.patch)
18:31:19 <boklm> ok, I can look at this too this week
18:31:40 <sukhe> I don't need to put more things on your plate so if you can just address that question, I am happy to take care of the builds and testing part :)
18:32:25 <boklm> ok
18:32:43 <sukhe> thanks, that's all from my side
18:32:45 <GeKo> okay, thanks boklm. that's at least a start (i'd need to dig into both things, too, in order to be helpful)
18:33:22 <sukhe> I have updated the progress related to #26476 where tjr and I are comparing the builds from TC
18:33:28 <sukhe> I never expected this would take so much time...
18:33:39 <GeKo> arthuredelstein: didn't you do the vacation thing last week? :p
18:33:58 <sukhe> both of us have put many hours into this, trying to compare if the toolchains differ and what not, but not much success
18:33:59 <GeKo> otherwise i think we are good with the status updates.
18:34:02 <GeKo> anything to add?
18:35:13 <GeKo> sukhe: where is that update?
18:35:40 <sukhe> GeKo: I will put it in the ticket. it is currently in our IRC discussions
18:35:48 <GeKo> ah, okay
18:36:53 <GeKo> sisbell: fwiw: nice work. i did not expect that this went so fast. now i hope there is no deep rabbit hole regarding the reproducibility of the  binary
18:37:21 <GeKo> (the latter is the main reason for doing the whole dance with tor-browser-build/rbm in the first place)
18:37:39 <GeKo> okay, let's move on to the discussion part
18:38:11 <GeKo> so, i heard complaints from different sides about storm and it's unresponsiveness
18:38:12 <sisbell> I still need to find out how to verify the reproducibility
18:38:41 <GeKo> sisbell: well, building the .apk twice and comparing the sha256 sums is a good start
18:39:09 <sisbell> ok will do
18:39:15 <GeKo> ideally, building it on different machines
18:39:22 <GeKo> but we can do that later during review
18:39:40 <GeKo> as that's probably easier for more than one person
18:39:59 <GeKo> s/it's/its/
18:40:15 <GeKo> i think we should make a decision on how to move on in that regard
18:40:39 <sysrqb> sisbell: also checkout diffoscope if you haven't already - it has nice support for identifying which file is causing reproducibility issues
18:40:44 <sukhe> I am not a "Sandstorm" expert but I run it on my server and I have noticed it is a RAM hog. perhaps an easy fix would be to increase the memory on the machine
18:40:53 <sukhe> oh yes, definitely a +1 for diffoscope
18:40:58 <GeKo> are we sufficiently annoyed that we should explore other ways of collecting status updates for the meeting?
18:41:27 <GeKo> i see.
18:41:44 <sysrqb> i'm not sufficiently annoyed, but i may have a high tolerance :)
18:42:07 <GeKo> it's working actually pretty well in my alpha tor browser
18:42:13 <tjr> I have no problems; but wouldn't object to a change. w/e works for people
18:42:15 <GeKo> i had way more issues with an esr52based
18:43:31 * tjr has to go to an appointment
18:43:38 <GeKo> o/
18:44:05 <GeKo> okay, i try to ask around first to get a potential memory upgrade going
18:44:07 <arthuredelstein> not sure why it doesn't work well for me, but I can paste in my status and manage that way. Just makes editing difficult
18:44:32 <GeKo> at least i'll ask around figuring out where exactly the problem could be
18:44:35 <sysrqb> which browser are you using?
18:44:43 <sysrqb> arthuredelstein: ^
18:45:05 <arthuredelstein> TBB alpha or Firefox 61
18:45:15 <sukhe> I have the same issue as arthuredelstein with 61.0.1 (64-bit) (not TB)
18:45:17 <sysrqb> hrm, okay. weird
18:45:39 <GeKo> arthuredelstein: and you see the isse with both?
18:45:41 <sukhe> in most cases, it doesn't even load the pad, let alone let me edit
18:45:42 <GeKo> *issue
18:45:51 <GeKo> huh
18:46:09 <arthuredelstein> it's very slow and often disconnects in both
18:46:25 <sysrqb> i use FF 61 and it's not nearly that bad for me
18:46:40 <arthuredelstein> maybe it's my computer :)
18:46:44 <sysrqb> but we can move if this is normal for some of you
18:47:15 <sukhe> do we have an alternative :)
18:47:22 <sysrqb> riseup pad's are better for both of you?
18:47:32 <sukhe> I also like it because self-host it
18:47:34 <sukhe> *we
18:47:37 <sysrqb> *pads
18:47:43 <GeKo> sukhe: +1
18:48:19 <boklm> would removing status updates from previous weeks (decreasing the size of the pad) make it faster to load?
18:48:37 <GeKo> oh, good idea. we could try that
18:48:41 <arthuredelstein> I agree there is value to self-hosting so maybe the answer is I just paste :)
18:48:44 <sysrqb> that's possible, i know the network team deletes previous weeks
18:48:51 <arthuredelstein> I don't want to cause more problems than this solves
18:49:10 <sysrqb> seems worth trying, as long as its arhived on the mailing list
18:49:25 <sukhe> arthuredelstein: yeah
18:49:31 <GeKo> okay, let's try to delete previous updates and figure out whether some memory upgrade could help
18:49:39 <GeKo> and if not let's find something else
18:49:40 <sukhe> deleting the pad should improve the loading of the grain, at least anecdotally
18:49:55 <arthuredelstein> If it's working for other people but not me, it suggests it's not a server issue
18:50:00 <arthuredelstein> at least I would guess
18:50:30 <sysrqb> you're not the only one
18:50:37 <sysrqb> sukhe: said he has similar issues
18:50:43 <GeKo> yes, but maybe we could improve things on different ends
18:50:45 <sysrqb> err s/://
18:51:25 <GeKo> re the github thing: i wrote an email to isabela's mail to tbb-dev to get the discussion going
18:51:31 <sukhe> yeah, but I can live with it. it's not completely broken so why fix it ;)
18:51:48 <GeKo> please chime in if there is anything to add
18:52:05 <sukhe> sorry if it seemed like I said it was causing a problem
18:52:14 <GeKo> or if i am wrong with the things i wrote
18:52:53 <GeKo> i think we could indeed benefit from some form or github "integration" and i think we could have someone taking this onto their plate
18:52:53 <pospeselr> +1 for keeping git in-house
18:53:02 <pospeselr> don't need msft potentially screwing us in the future
18:53:02 <GeKo> yes
18:53:17 <sysrqb> yep
18:53:47 <arthuredelstein> I guess I'm not sure what problem we want to solve by further github integration
18:54:00 <sukhe> code review I think
18:54:09 <GeKo> that's one
18:54:51 <sukhe> also, GitHub gives more control some ways. (I am not saying we should use --force :) but there are other things, like deleting a remote branch which is not possible on git.tp unless the permissions are set
18:54:59 <GeKo> and then there could be an argument made for getting patches faster as it might expose our code to more devs
18:55:34 <arthuredelstein> It's true that code review is nicer on github but I'm a little hesitant to shard our reviews from trac
18:56:20 <arthuredelstein> Mozilla has a separate review tool but the comments get mirrored back to bugzilla so it's still searchable
18:56:25 <GeKo> bc we could lose them on github?
18:56:39 <GeKo> aha!
18:56:42 <arthuredelstein> Just that they're not searchable or readable in one place
18:57:00 <arthuredelstein> I wonder if there is a trac plugin that could do the same kind of mirroring.
18:57:13 <sukhe> arthuredelstein: at least the way I do it -- and doesn't make it right -- is that I reference a github branch on to trac so that way the comments stay on trac and the code is on github
18:57:57 <arthuredelstein> sukhe: Yeah, I think that's fine. I do the same :)
18:58:06 <GeKo> yes, that's what we currently have
18:58:17 <arthuredelstein> More wondering about the benefits/drawbacks of using github's built-in code review
18:59:49 <GeKo> arthuredelstein: could you voice your concerns in a reply to that tbb-dev thread?
19:00:08 <arthuredelstein> sure, will do
19:00:26 <mcs> Network Team people might have useful input about code review tools since they use them (at least sometimes).
19:00:53 <GeKo> i am happy if we could satisfy the wishes tjr had while taking he concerns with github into account as well
19:01:25 <GeKo> okay, final point: next meeting
19:01:38 <GeKo> i need to be afk next monday (and the whole week)
19:01:51 <catalyst> i'm happy with github's review tools. i think they're not perfect, but the network/community effects and usability outweigh the drawbacks
19:01:52 <GeKo> so, i guess 8/13 would be the next one?
19:02:28 <sysrqb> catalyst: thanks
19:02:34 <GeKo> catalyst: which network/community effects wrt review tools to you have in mind?
19:02:41 <GeKo> *do
19:03:08 <arthuredelstein> catalyst: Are you using github tickets as well? Or just trac tickets?
19:03:22 <catalyst> low barrier to entry for new contributors. automatic CI building of pull requests so you know if someone's patches broke the build before you spend time to read them
19:04:02 <catalyst> we use trac tickets for issue tracking together with github pull requests
19:04:05 <GeKo> ah, you were not speaking of review tools in particular, okay
19:05:02 <GeKo> any objection to next meeting on 8/13?
19:05:14 <catalyst> i was lumping CI running on pull requests in with review tools
19:05:21 <GeKo> k
19:05:27 <boklm> 8/13 sounds good to me
19:05:32 <arthuredelstein> me too
19:05:34 <sysrqb> no objection of 8/13
19:05:36 <mcs> 8/13 sounds fine.
19:05:50 <sisbell> good as well
19:06:04 <igt0> it sounds good to me
19:06:13 <pospeselr> no objections!
19:06:18 <GeKo> okay, great!
19:06:29 <GeKo> anything else for discussion?
19:07:15 <GeKo> then we are done with the meeting for today, thanks all *baf*
19:07:19 <GeKo> #endmeeting