17:59:04 <GeKo> #startmeeting tor browser
17:59:04 <MeetBot> Meeting started Mon May 14 17:59:04 2018 UTC.  The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:59:04 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:59:08 <GeKo> hi all!
17:59:15 <GeKo> meeting pad: https://storm.torproject.org/shared/tHoN4Ii7rLSjPE0OP4gydX4cMGadsXmRQNc-6lwru0N
17:59:23 <mcs> hi
17:59:24 <igt0> !
17:59:25 <boklm> hi!
17:59:35 <pospeselr> hello
17:59:46 <sysrqb> .
18:01:00 <arthuredelstein> hi everyone!
18:01:26 <GeKo> please mark things bold you want to talk about/get notcied by folks
18:04:54 <GeKo> okay, let's get started
18:05:42 <GeKo> tjr: that env is nothing fancy i just managed to make one of my spare laptops into a wireless access point and is runing wireshark
18:06:00 <GeKo> and my mobile phone is connecting to it and only sending traffic over it
18:06:11 <tjr> So that would catch any traffic from the phone right? not necessaily just from the browser
18:06:14 <GeKo> and once in a while i am looking at the wireshark output
18:06:21 <GeKo> yes
18:06:55 <tjr> Okay. I haven't started on it; because I am working to finish MinGW; but I think the VPN Service approach I outlined will be more efficient and reliable
18:07:13 <GeKo> i don't have a good way of analyzing the wireshark output automatically yet, we'll see
18:07:26 <GeKo> yeah, i am looking forward to that
18:07:29 <tjr> I'll probably wind up doing that peice first; and then you can use it (if you want to) and then I'll work on integrating it into the automatd test stuff
18:07:34 <GeKo> but i wanted to play with something right now
18:07:38 <tjr> kk
18:08:01 <GeKo> so, what's the story for our mozilla-central rebasing?
18:08:11 <GeKo> i am in particular interested in the mobile angle
18:08:40 <GeKo> because we need it for that one first and we only have about 1 1/2 - 2 months left before shipping someting
18:08:46 <GeKo> *something
18:09:04 <sysrqb> yes
18:09:05 <GeKo> meaning we need to have a story for rebasing onto mozilla-beta and -release, too
18:09:07 <sysrqb> this is a good question
18:09:14 <arthuredelstein> GeKo: So I have a script that does some auto-rebasing
18:09:30 <arthuredelstein> My idea was to have it push a new branch on a daily or weekly basis somewhere
18:09:46 <arthuredelstein> And the provide some kind of alert whenever the rebasing fails so we can manually fix up an patch that needs it
18:10:07 <GeKo> we could try that
18:10:22 <arthuredelstein> And yes, we would need the same for mozilla-beta and mozilla-release too.
18:10:40 <GeKo> i am not sure about the interval for the rebasing efforts especially on nightly when things are more broken than on -beta and -release
18:10:40 <igt0> For the first release, are we aiming FF 60 or 61?
18:10:49 <GeKo> i think 61
18:11:14 <GeKo> 60 will already  be obsolete
18:11:33 <GeKo> so, we should start doing the mozilla-beta rebase soon
18:11:39 <arthuredelstein> GeKo: Yeah, I'm not sure either. Thinking about it again, maybe it's not really necessary to automate, but just to rebase manually every week.
18:11:48 <GeKo> and in parallel following mozilla-central somewhere else
18:12:07 <GeKo> arthuredelstein: i am fine which whatever we want to try
18:12:23 <GeKo> as far as we have a plan and get things started soon
18:12:29 <GeKo> and iterate from that
18:12:30 <arthuredelstein> Right, so rebasing to mozilla-beta and mozilla-central could be something I work on this week
18:12:49 <GeKo> okay, sounds good.
18:13:04 <arthuredelstein> I would need help with the Android patches and possibly updater patches
18:13:10 <GeKo> arthuredelstein: the other high prio thing for this week is fixing up all the things i find
18:13:15 <arthuredelstein> yes
18:13:17 <GeKo> in the rebase review
18:13:34 <GeKo> ideally we'll have the esr60 branch ready next week
18:13:41 <GeKo> and can start with nightly builds for linux
18:13:49 <arthuredelstein> yes, I will fix up those things before rebasing again to mozilla-beta
18:13:55 <GeKo> i assume we'll have an esr60 branch in git by then
18:14:09 <arthuredelstein> Sounds good to me.
18:14:14 <GeKo> (in gecko-dev that is)
18:14:19 <GeKo> great!
18:14:28 <arthuredelstein> One issue I ran into in tor-browser-build.git is I couldn't get selfrando to work
18:14:46 <GeKo> disable it
18:14:50 <arthuredelstein> I could do with advice there as I'm not too familiar with it
18:14:52 <arthuredelstein> I did :)
18:15:03 <GeKo> i'll give it to the selfrando people to fix it
18:15:23 <GeKo> just file a bug with the error and then build without it
18:15:33 <GeKo> it might even be the time to just get rid of it
18:15:42 <arthuredelstein> OK, will do
18:15:55 <GeKo> given that it does not do much against a determined attacker
18:16:12 <GeKo> and we don't have time to keep it working and getting it into stable shape
18:16:20 <arthuredelstein> I see
18:16:23 <igt0> GeKo, sysrqb  related question, since we don't have an android dev yet, should we also take a look in the tor browser build for android?
18:16:23 <tjr> I'd probably strike 'determined' :)
18:16:34 <GeKo> heh
18:17:05 <GeKo> igt0: first, i hope we have one rather soon and second we need to get the other pieces for the alpha ready first
18:17:36 <GeKo> so, as much as i'd like it we don't have the capacity right now to focus on the build intregration at this point
18:18:03 <igt0> cool, makes sense.
18:18:12 <sysrqb> tjr: is there a paper about that or is that from testing/professional opinion?
18:18:33 <tjr> sysrqb: Yea I can forward
18:18:59 <arthuredelstein> Another big thing that needs to get fixed in the desktop build are torbutton (circuit display, security slider, stream isolation). I will work with igt0 to try to get that working this week
18:19:10 <sysrqb> tjr: that'd be awesome, thanks - just want to make sure i understand the  reasons
18:21:04 <GeKo> arthuredelstein: yes. i think the slider might be hard depending on what noscript is offering, so def. not a blocker for nightly
18:21:12 <GeKo> but the other two are important
18:21:33 <arthuredelstein> right
18:21:34 <GeKo> arthuredelstein: that said: you are up!
18:21:42 <GeKo> what's up with the tests?
18:22:32 <arthuredelstein> I had to patch the permissions manager to use principals instead of URIs.
18:22:48 <tjr> Oh okay
18:22:50 <arthuredelstein> And this causes all sorts of subtle breakages in unit tests
18:23:08 <arthuredelstein> I have made a lot of progress but there are still a couple of things I'm investigating
18:23:23 <tjr> Hm. Have you talked to Johann? It sounds like this might catch resistence to upstreaming?
18:23:42 <GeKo> arthuredelstein: do we have someone from mozilla who could help here?
18:24:21 <GeKo> (and be it just to get the big picture right)
18:25:02 <arthuredelstein> it's a good idea. I can post my current patches on the ticket and see if someone wants to give it a lookover
18:25:26 <GeKo> yes, please do and ask moz people for feedback
18:25:32 <arthuredelstein> yes
18:26:12 <arthuredelstein> it does more or less follow the strategy of other FPI patches so I hope it won't be too horrifying to Mozilla folks :)
18:26:36 <tjr> ni johannh
18:26:44 <arthuredelstein> tjr: will do, thanks
18:27:36 <igt0> ah talking about moz people reviewing patches, do we have a moz person to help us to review android patches? :)
18:28:17 <GeKo> are you talking about 1459420?
18:28:27 <GeKo> or in general?
18:29:13 <igt0> in general, right now i am adding mcomella in all my patches, however i think he is not allocated 100% on fennec.
18:29:32 <tjr> let me bug Wennie about that bug again
18:29:45 <sysrqb> snorp or jchen are good for the code, nalexander reviews code too, but is more build-related
18:29:59 <sysrqb> igt0: you can ask in #mobile who you should add as a review :)
18:30:40 <igt0> sysrqb, oh cool, i will do it from now on.
18:31:26 <GeKo> tjr: it seems the -O2 bug is for x86 and not x86_64, no?
18:31:33 <GeKo> that said really nice work!
18:31:41 <sysrqb> i guess they may say " we don't have any fennec devs right now", but *shrug*
18:31:48 <GeKo> and your plan looks good to me with one exception:
18:31:58 <GeKo> we need to get the sandbox on win64 going
18:32:06 <tjr> Whoops. Yea I mixed up the x86/x64 on -O2, fixed it
18:32:14 <GeKo> i started to look at that back then
18:32:26 <GeKo> but then i dropped the ball
18:32:33 <tjr> Yea, so what's the story with that? Is there a compile error we're hitting with it?
18:32:40 <tjr> I guess I can just turn it on and see what happens :)
18:32:42 <GeKo> no, it compiles
18:32:46 <GeKo> but crashes
18:32:50 <GeKo> when running
18:32:54 <tjr> Ah okay. So we need to debug it?
18:32:58 <GeKo> yes
18:33:02 <tjr> Well, I can give it a shot :)
18:33:11 <GeKo> i got stuck somewhere in it...
18:33:20 <GeKo> that would be neat
18:33:26 <GeKo> otherwise i try to get back to it
18:33:33 <GeKo> it's a blocker for the first alpha for us
18:33:40 <boklm> #24197 is the ticket
18:34:16 <tjr> I forgot one other issue i need to work on: Figure out d3dcompiler.dll issue
18:34:40 <GeKo> besides bundling that dll?
18:35:11 <tjr> There's something weird with we how we detect/load it
18:35:23 <GeKo> ah, 1460620
18:35:24 <GeKo> okay
18:35:34 <tjr> Even after i successfully bundle it; it's not loaded like I believe it is normally on Windows
18:35:51 <tjr> It will load it if I uncomment another define; but that's not needed normally.
18:35:58 <pospeselr> PATH issue?
18:36:55 <tjr> I don't think so...
18:37:02 <tjr> It should be picked it up right here: https://searchfox.org/mozilla-central/source/gfx/angle/checkout/src/libANGLE/renderer/d3d/HLSLCompiler.cpp#136
18:37:19 <tjr> But instead I have to define ANGLE_PRELOADED_D3DCOMPILER_MODULE_NAMES for it to be picked up in the above block
18:37:37 <tjr> Whereas normally ANGLE_PRELOADED_D3DCOMPILER_MODULE_NAMES isn't needed. and I'm putting the dll in the same location as normal windows builds.
18:37:51 <tjr> (which is just the same dircetory; next to firefox.exe)
18:38:23 <tjr> Unless... MinGW is defining some security feature to not dynamically load dll's from the same directory? But I would have expected that to break other things?
18:38:48 <GeKo> no, it does not do that
18:39:59 <tjr> Anyway; I have no finished investigating, i just know i need to
18:40:37 <GeKo> sukhe: so, i still have an idea to ask someone else about this issue but, yes, meanwhile you could focus on something else
18:41:04 <sukhe> GeKo: that would be great!
18:41:26 <GeKo> arthuredelstein: would it be okay for you if sukhe worked on a proper patch for #26073?
18:41:36 <GeKo> given all the things you have on your plate
18:42:20 <GeKo> i am fine if you want to finish that by yourself, though, given that you already spent some time on it
18:42:44 <GeKo> but that would be the most valuable thing from a build perspective right now
18:44:01 <arthuredelstein> Fine with me if sukhe wants to work on it. I think the biggest things right now are probably torbutton issues
18:44:24 <arthuredelstein> igt0 is also working on those things so I was just planning to try to help with those
18:44:33 <GeKo> well, those are dealt with outside tor-browser-build
18:45:01 <arthuredelstein> true :)
18:45:15 <GeKo> sukhe: okay, so #26073 over to you
18:45:40 <sukhe> ok thanks
18:45:42 <GeKo> the idea is to come up with a proper patch for getting linux nightly builds going asap, based on esr60
18:45:51 <GeKo> if you need some help just ping me
18:45:57 <sukhe> will start reviewing it
18:46:06 <arthuredelstein> sukhe: Happy to discuss with you where it's at when you have time
18:46:12 <sukhe> arthuredelstein: thanks!
18:46:23 <GeKo> alright, anything else before moving to the discussion?
18:47:00 <GeKo> okay, discussion then:
18:47:04 <GeKo> mozilla's all hands
18:47:07 <tjr> I'm just curious, cause Mozilla hasn't told me anything; but who's going to All Hands?
18:47:28 <arthuredelstein> I'll be going
18:47:44 <sysrqb> I am, but I need to follow up with them and confirm my flight was approved
18:48:20 <GeKo> tjr: should we prepare things from our side for the all hands meeting?
18:48:29 <GeKo> if so, what would be helpful?
18:48:42 <tjr> Good question. I'm sure there is, but I'll have to think about it
18:48:45 <tjr> Ask me next week :)
18:48:46 <pospeselr> i'm going too :)
18:48:50 <GeKo> will do
18:49:01 <GeKo> tjr: i heard roger and isa are going, too
18:49:06 <GeKo> i think flexlibris as well
18:49:13 <tjr> cool!
18:49:14 <antonela> me too :)
18:49:18 <GeKo> yay
18:49:45 <flexlibris> i can't go this time :(
18:50:04 <GeKo> okay :( then
18:51:00 <isabela> ahf fromnetwork team is going
18:51:03 <isabela> and so is steph afaik
18:51:22 <tjr> Awesome, thanks all
18:51:32 <GeKo> tjr: you won't be alone :)
18:51:46 <GeKo> okay, anything else for today?
18:52:35 <GeKo> thanks then and happy hacking *baf*
18:52:37 <GeKo> #endmeeting