18:01:19 <GeKo> #startmeeting tor browser 18:01:19 <MeetBot> Meeting started Mon Apr 10 18:01:19 2017 UTC. The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:19 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 18:01:24 <GeKo> hi! 18:01:27 <mcs> hi 18:01:32 <boklm> hi! 18:01:39 <GeKo> welcome to another round of tor browser meeting business 18:01:48 <GeKo> who wants to go firs today? 18:01:52 <GeKo> *first 18:02:37 * mcs can go 18:02:45 <mcs> Last week, Kathy and I helped review some of the #20680 patches and we also did some smaller reviews. 18:02:52 <mcs> We did a little testing on OSX of the first ESR52-based nightly build. 18:02:57 <mcs> We worked on #21766 and #21778 but did not finish either bug. 18:03:03 <mcs> We then switched to working on #21876. 18:03:09 <mcs> This week we will finish #21876 (that should happen today). 18:03:14 <mcs> We will then go back to working on the other e10s-related tickets (#21766 and #21778). 18:03:19 <mcs> We also want to test the updater in the ESR52-based Tor Browser on all platforms. 18:03:25 <mcs> I started a nightly build earlier today to see if Kathy and I are lucky enough to be able to complete a gitian-based build of the ESR552-based TB. 18:03:35 <mcs> We also hope to take a little time off near the end of the week to celebrate Easter, but we will see how work goes the first part of the week. 18:03:41 <mcs> That’s all for now. 18:04:09 <mcs> (and GeKo, please reset our priorities this week as needed) 18:04:50 <GeKo> so the nightly build did not explode on your machines? 18:05:04 <GeKo> that's good as i currently have no way of testing os x builds 18:05:34 <mcs> It hasn’t been long enough yet to know. I think it is proceding slowly or it is stalled but I will check after this meeting. 18:06:08 <GeKo> no, i meant the one we built and announced on tbb-dev 18:06:46 <mcs> oh, right. That seems to work okay (some knowm issues of course). 18:06:56 <mcs> But it did not completely explode. 18:07:24 <GeKo> progress! 18:07:28 <GeKo> okay, who is next? 18:07:51 <GeKo> mcs: i think your priorities look good fwiw 18:08:00 <mcs> GeKo: thanks 18:08:38 * boklm can go next 18:08:58 <boklm> This past week I made Windows rbm based builds reproducible (#20426). In #21286 I set up a VM to do nightly builds. 18:09:03 <boklm> I reviewed some gitian patches and applied similar changes to the rbm based build. I tried the esr52 build on Centos6 and investigated #21907. 18:09:14 <boklm> This week I'm planning to work on #21907, #18530, #19316 and run tests on the ESR52 build. 18:09:24 <boklm> That's it for me. 18:09:45 <GeKo> so, regarding the nightly builds what is missing in them currently? 18:09:50 <GeKo> snowflake i guess? 18:10:23 <boklm> we have snowflake on linux64, but there is still a build issue on linux32 18:11:16 <GeKo> but apart from that do we have everything? 18:12:11 <GeKo> i guess we could point folks to your nightly builds at least until ln5's gitian issues get fixed again 18:12:20 <boklm> I think yes, apart from the python pluggable transports that I'm not sure we still want (obfs3 and fteproxy) 18:12:47 <GeKo> okay, those are missing as well, hrm. 18:13:17 <GeKo> i guess we could think about cutting those out, too, once we switch to rbm 18:13:48 <GeKo> should be no prio to get them in (snowflake on the other hand would be good) 18:13:55 <GeKo> but, neat. 18:14:24 <GeKo> i'd like to recheck the reproducibility using my weird machine i have 18:14:30 <GeKo> which commit would be good? 18:15:35 <boklm> if you build the current master, you should get something matching http://f4amtbsowhix7rrf.onion/tor-browser-builds/2017-04-10/ 18:15:50 <boklm> (the alpha-* directories) 18:16:15 <GeKo> okeydoke. i'll kick off the build tomorrow 18:16:38 <GeKo> but it could take quite a while it is on a quite old laptop :) 18:17:07 <GeKo> i guess i can go while i am at it 18:17:32 <GeKo> we have nightly builds based on esr52! and it seems they are working on all platforms we support 18:17:48 <GeKo> there are issues, sure, but so far no show stoppers 18:18:05 <GeKo> i wrote patches for #21837 and #21865 18:18:19 <GeKo> we should now have reproducible builds on all platforms 18:18:36 <GeKo> i think we are good with the toolchain for the upcoming 7.0 series. 18:19:15 <GeKo> i might be tempted to squeeze in two smaller changes. one is #19181 18:19:34 <GeKo> and the other one is a binary tlb in mingw-w64 which i want to regenerate before building 18:19:47 <GeKo> we'll see if there is time left for that 18:20:03 <GeKo> i looked into #21741, #20683 and #21611 18:20:18 <GeKo> i am done with the former and need to update the ticket 18:20:52 <GeKo> selfrando is a sad story right now. we have issues with the switch to esr52. i hope the selfrando folks can fix those in time 18:21:02 <GeKo> to make it into the alpha 18:21:16 <GeKo> #21611 is okay 18:21:29 <GeKo> but we need the dns leaks patch for that 18:21:36 <GeKo> arthuredelstein: ^ 18:21:44 <GeKo> so good we added it :) 18:22:28 <GeKo> this week i'll try to finally get #20761 tested and reviewed 18:22:35 <GeKo> sorry that it took so long 18:22:48 <GeKo> i hope to get more esr52 stuff into our alpha as well 18:23:02 <GeKo> and then there is probably the release dance starting again 18:23:04 <arthuredelstein> GeKo: Oh good -- glad #21611 is OK! 18:23:15 <GeKo> that's it for me 18:23:38 <tjr> I can go 18:23:45 <tjr> Slow progress, due to sickness and other interference. The main one was finding a solution for std::__throw errors for debug builds: https://bugzilla.mozilla.org/show_bug.cgi?id=1332747 18:23:52 <tjr> With that one solved, I can produce debug esr52 builds, including of the different sandbox approaches. In theory this should enable better understanding of where/why different types of sandbox builds break. In practice I haven't yet investigated this. 18:23:59 <tjr> I'm trying to juggle both issues, but I'm definetly trying to plug along at the other MinGW bugs so I can actually get MinGW into TaskCluster. This includes a correct windows.configure patch (https://bugzilla.mozilla.org/show_bug.cgi?id=1314979) 18:24:09 <tjr> I'm also looking at onion services again, but nothing yet to report there. 18:24:15 <tjr> I'm working to develop a very detailed Fennec implementation list for Tor support. I think the trickiest thign is going through all the pref values and deciding what we want to include and what we don't; and then getting input from the rest of Mozilla and Tor about it. 18:24:19 <tjr> We've been working on getting folks invites to the All Hands in SF in June. 18:24:22 <tjr> I replied to the pointer events thread. Someone mentioned that mozilla already exposes some information through moz-specifc extensions. 18:24:25 <tjr> That's it for me 18:25:02 * arthuredelstein can go 18:25:17 <arthuredelstein> Hi all -- this past week I finished investigating #21712, 18:25:25 <arthuredelstein> I wrote a patch for #21849, 18:25:47 <arthuredelstein> I met with the Mozilla uplift team, 18:25:51 <arthuredelstein> and I tested Tim's patch for #21875. 18:26:06 <arthuredelstein> I also tried to run a gitian-based build and ran into LXC problems, which I am trying to figure out. 18:26:13 <arthuredelstein> And I took a couple of days off. 18:26:25 <arthuredelstein> This week I will keep working on tbb-7.0-must-alpha tickets, starting from highest priority ones. 18:26:31 <arthuredelstein> That's it for me 18:27:33 <GeKo> i wonder if you just should start with rbm or do you want to test whether you get exact matches of current builds? 18:28:07 <arthuredelstein> Yeah, maybe I should just switch to rbm now. 18:28:20 <GeKo> if it's just builds with our current configuration then that's probably the way to go 18:28:25 <arthuredelstein> OK, cool 18:28:35 <GeKo> it should be much simpler to setup (i hope at least) 18:28:44 <GeKo> and dogfooding is always good :) 18:28:47 <arthuredelstein> yay! 18:29:06 <GeKo> do we have anyone elser here with status reports? 18:30:06 <arthuredelstein> Do I just use the master branch here? https://gitweb.torproject.org/builders/rbm.git/ 18:30:13 <GeKo> okay, discussion time then 18:30:34 <GeKo> i guess the first item is next meeting time 18:30:51 <boklm> arthuredelstein: you should use the master branch at https://gitweb.torproject.org/builders/tor-browser-build.git/ 18:30:56 <GeKo> do we skip the next meeting and do the next one in two weeks? 18:31:09 <GeKo> or do we move the meeting next monday to tuesday? 18:31:57 <boklm> arthuredelstein: rbm.git should be cloned automatically as a git submodule from tor-browser-build.git 18:32:58 <arthuredelstein> GeKo: I'm OK with either thing. 18:33:10 <mcs> r.e. meeting next week, maybe just skip it? I think we will all be communicating about alpha issues anyway. 18:33:19 <boklm> arthuredelstein: and the README file explains how to do builds: https://gitweb.torproject.org/builders/tor-browser-build.git/tree/README 18:33:22 <mcs> (on IRC and via trac) 18:33:57 * boklm is OK with either things too 18:34:27 <GeKo> okay. let's skip it then. i'll send a mail to tbb-dev tomorrow 18:34:37 <GeKo> then i have the hardened builds on my list. 18:35:36 <GeKo> i tried to fix them on the weekend but failed 18:35:45 <GeKo> i hit a similar issue as #17508 18:35:53 <GeKo> which i still need to file 18:36:32 <GeKo> anyway that got me thinking and i am a fan now to use the opportunity to move forward with #20814 18:36:48 <GeKo> as we discussed it back then in our meeting(s) and on tbb-dev 18:37:08 <GeKo> so my idea is to drop the hardened one as-is with the next alpha 18:37:36 <GeKo> and explain to users that are looking for a hardened 64bit linux setup that they should use the sandboxed tor browser 18:37:56 <GeKo> hardened users would get updated to the regular alpha then 18:38:24 <GeKo> maybe we could use that billboard feature to explain some things and point to a blog post and sandboxed tor browser 18:38:45 <GeKo> that won't be available anymore with esr52. so the timing seems to be a good one 18:39:23 <GeKo> for folks that want to help with debugging we start with producing a debug version with rbm 18:39:42 <GeKo> which we need anyway later this year when we start working on the fuzzing deliverable i guess 18:40:10 <GeKo> does this sound reasonable? or is there something i overlooked while pondering? 18:40:26 <mcs> It sounds like a good plan to me. 18:40:41 <boklm> It sounds like a good plan to me too 18:40:49 <arthuredelstein> To me too. 18:40:54 <GeKo> fine! 18:41:07 <arthuredelstein> What's the billboard feature? 18:41:37 <GeKo> we used to show a dialog about update issues when we had problems with the update due to unix domain sockets 18:41:53 <GeKo> mcs and brade dug that one out 18:42:07 <arthuredelstein> thanks 18:43:04 <GeKo> okay, i plan to write a blog post about dropping the hardened series as-is for next week 18:43:32 <GeKo> and we can start thinking about new ways to get the debug builds back 18:43:38 <GeKo> and what they should include 18:43:57 <GeKo> do we have any other things worth discussing today? 18:44:28 <arthuredelstein> When is the deadline for patches to get into the alpha? 18:44:46 <arthuredelstein> i.e., when do you plan to start the build? :) 18:45:25 <GeKo> i actually planned to have a weekend without building things as i want boklm to have such a weekend as well. 18:45:47 <GeKo> but that might be a bit ambitious, especially for the alpha, this time :/ 18:46:05 <GeKo> nevertheless i hope i can start a build thursday evening UTC 18:46:14 <GeKo> (an alpha one) 18:46:33 <arthuredelstein> OK, thanks 18:46:55 <arthuredelstein> I hope you both have a non-building weekend as well! 18:47:03 <GeKo> :) 18:47:39 <GeKo> okay, we seem to be out of topics for today. thanks all. *baf* 18:47:43 <GeKo> #endmeeting