19:00:12 #startmeeting tor browser 19:00:12 Meeting started Mon Jan 23 19:00:12 2017 UTC. The chair is GeKo. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:12 Useful Commands: #action #agreed #help #info #idea #link #topic. 19:00:30 hi all and welcome to the weekly tor browser meeting 19:00:45 geko: i'm talking to arlolra and i think so long as my browser stays online, snowflake in 7.0a will work. 19:00:57 saw it, thanks 19:01:01 hi everyone 19:01:07 hi 19:01:08 hi 19:01:19 i guess i can go first today 19:01:25 o/ 19:01:31 last week i worked on a bunch of different things 19:02:00 i think i filed all the compile bugs for esr52 on windows and have workarounds for them 19:02:09 currently linking libxul is stiill broken 19:02:24 but i heard jacek has a workaround for that one 19:02:54 so, the future is not so dark as it still was a week ago 19:03:09 but there is still a lot to do to get all the windows pieces into shape 19:03:32 then i made release preparations and started building 19:03:48 we are currently redoing things a last time to pick up the new tor version 19:03:57 but we are still on track i think 19:04:59 i worked on #18925 and feel i am close to get the code signature stripped of our MAR files at least 19:05:16 and i made progress on #15988 19:05:49 finally i helped with the questions OTF had regarding our sponsoru contract 19:06:28 this week i plan to help with the releases and hope to make further progress on #18925 and #15988 19:06:48 i plan to start reviewing the new features in firefox esr52 as well 19:06:53 that's it for me. 19:07:45 * mcs will go next 19:08:02 Last week, Kathy and I developed a patch for #20989 (already merged by GeKo and included in TB 7.0a1 — thanks!). 19:08:08 We then spent some time experimenting on OSX with an ESR52-based Tor Browser that has multiprocess mode enabled. 19:08:14 We filed a couple of Torbutton tickets: #21267 #21268 19:08:21 (but note that much more testing needs to be done, and not everything is working yet due to other ESR52 incompatibilities). 19:08:26 Finally, we experimented with the tor patches for #20956. 19:08:38 Kathy and I think they will help us a lot when it comes to shipping a TB that uses Unix domain sockets but we welcome additional opinions. 19:08:46 This week we plan to experiment some more with multiprocess TB and Firefox’s content sandbox, probably on Linux. 19:08:51 We plan to write a README for Tor Launcher (#21264). 19:08:57 We will also help with any issues that come up with the new releases. 19:09:02 We are also available to help with ESR52 rebasing efforts if needed. 19:09:07 That’s all for us. 19:09:27 I can go next 19:09:32 The Orbot Telemetry Patch is landed in Fennec and will be collecting data once it moves into beta in ~ a week 19:09:41 My main effort has been working on mingw builds. I'm trailing GeKo and Jacek in actually getting it built. The goal is to replicate the process and get it into TaskCluster as a default build target so it won't get broken again. 19:09:47 I've submitted and gotten a few patches uplifted, but still have several that need cleanup for integration. My main stumbling point is that I'm hitting this bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1332747 while GeKo and Jacek is not. 19:09:54 Aside from that I talked with the lead sandboxing guy about the /proc stuff. It does not seem like eliminating all calls to /proc/ is a specific target at this time, but they'd be open to refactoring things if patches were submitted and didn't break things. 19:09:59 Getting the sandbox compiled in MinGW is going to be a separate job after the MinGW build is in the build. It will most likely involve submitting patches to chromium sandbox. 19:10:13 That's all for me. 19:10:24 have you tried my toolchain? 19:10:46 mcs: do you read mozilla's dev-platform list? 19:11:02 I started comparing it, but I haven't tried using it, i set it aside for a little bit to look at something else 19:11:25 mcs: mike conley had a mail there on jan 18 about the plans for e10s on esr52 19:11:43 GeKo: yes. I read that list. 19:11:52 it's basically with the firefox 50 restrictions which sounds a bit scary probably 19:12:11 but we'll see 19:12:31 mcs: on other thing: 19:12:33 GeKo: yes, we may be “on our own” a bit if we turn on multiprocess mode. 19:12:45 yes, that's the scary part :) 19:13:04 on tbb-dev there is this thread about tor launcher and webextension 19:13:20 i was a bit confused about the standalone tor launcher idea 19:13:38 do you think you could pick up that discussion? 19:13:53 especially as it seems folks have talked to you about that at the dev meeting? 19:14:33 GeKo: Sure, I can try. There are definitely some difficult tradeoffs ahead for us, and it is also difficult b/c we do not know all of Mozilla’s plans. 19:14:55 hm. 19:15:16 The conversations at the dev meeting were fairly high levelas I remember. Some people have a need for something like Tor Launcher that is outside the browser. 19:15:28 I guess we should discuss more during the discussion phase of this meeting of later. 19:15:31 or later 19:15:45 sounds good 19:16:01 * arthuredelstein can go 19:16:09 This past week I worked on fixing up my #20680 rebase patches. 19:16:22 I'll be rebasing to Firefox 52 beta hopefully today and then I will post it for testing/review. 19:16:27 I also worked on #21201 and I will post patches for that as well. 19:16:50 I wrote a patch for #21243. 19:16:59 And I worked on #21244. I plan to try to finish that this week. 19:17:24 Ugh, wrong number 19:17:39 #21224 19:17:53 Thanks :) 19:18:06 And we need to rewrite our #16528 patch for ESR52 too, so I'll give that a go. 19:18:20 Finally I'm thinking about working on #19675, if there is time. 19:18:24 That's it for me. 19:19:01 arthuredelstein: For Orfox I have a doc that may help let me get it... 19:19:14 great, thanks! 19:19:56 Google Docs (boo hiss) https://docs.google.com/document/d/19wydMRxaIQty-jy1nk4RVnHLaVWUEZqM0YPrmcPVPcA/edit?usp=sharing 19:20:55 The next thing I will work on for Fennec (probably late Feb at the earliest) will be submitting a patch for a singular proxy location 19:21:12 Thanks for the doc, will be useful! 19:22:41 * boklm can go 19:22:57 This past week I worked on #17380, and helped build the new releases. 19:23:01 I also started doing some nightly builds (#21286) using the new build process: http://jvwai5ky6euqp3yj.onion/reports/index-tor-browser_build.html 19:23:18 This week I'm planning to help publishing the new releases, and start adding pluggable transports on #17380. 19:23:27 I'm also planning to take some days off on thursday and friday. 19:23:33 That's it for me. 19:24:44 good stuff 19:25:21 boklm: when you think your are ready to get that setup deployed elsewhere/more widely 19:25:34 i guess it would be good to move the code into official repos 19:25:42 like builders/tor-browser-build 19:25:48 and builders/rbm 19:26:09 (i am fan of having our stuff on our infra, too) 19:26:34 and ln5 is probably eagerly awaiting to try this new way out on his machine 19:26:53 actually, it would be neat to be able to compare build times on his machine 19:26:54 yes, I think we can move the code there. And I think people can start trying some builds. 19:27:10 between the gitian way and the new way 19:27:21 this would be good for sponsor4 reporting, too 19:27:56 i might give it a try too after the release got out and nothing exploded 19:28:16 alright. who else is here for a status update? 19:28:18 I am not sure the first build will be faster, but the next builds when not many things have changed should be much faster 19:29:29 I can go 19:30:05 No further work was done on #20815 (front-end development work on Tor Browser Settings), as it's stable. So, this week, I have taken a look at the NoScript Anywhere++ settings mappings for the Tor Browser Settings extension back-end. The conclusion is that some changes and fixes are needed, and so I have started rewriting the corresponding codebase. I'll be available at #tor-mobile afterwards, to discuss the results. 19:30:27 That's pretty much it for me. 19:31:04 you mean you have worked on noscript? and fixed things there? 19:31:14 that's cool 19:31:20 Not really 19:31:54 NoScript Anywhere++ (for Fennec) stores its settings differently from NoScript 19:33:10 so you meant you rewrote the mobile security slider code to take that into account? 19:33:19 So the current settings mappings need to be rewritten for Tor Browser Settings for Orfox. So I have taken a look at NSA++, and am writing up new preference mappings 19:33:33 ok 19:33:46 Yes, I'm in the process of adapting the code 19:34:33 Based on the findings 19:34:34 do we have status updates left? 19:35:10 then let's move to the discussion part 19:35:41 does anyone have something we should discuss here? 19:35:51 mcs: i guess that tor launcher thing could be one 19:36:07 so, i understand that some projects want a standalone tool 19:36:26 but i am not convinced yet that we want that and that tor launcher should be it 19:36:38 Yes, and to some extent Yawning created one for the sandboxed Tor Browser 19:36:58 i mean we moved away from vidalia for a reason years ago 19:37:08 that's true 19:37:22 I think there are conflictomg requirements; we definitely do not want to recreate the bad things about Vidalia 19:37:31 (which you and others have more history with) 19:37:58 Also, a lot of the standalone launchers are Linux centric in their implementation which won’t work for TB 19:38:19 yes 19:38:31 (or we need several things, one for each platform, which sounds like a lot to maintain) 19:39:08 If we think Tor Launcher’s current “embedded in the browser” approach is best for TB, we should try to keep it. 19:39:36 what speaks against it from a TB perspective? 19:39:47 But I also suspect it will become increasingly difficult to keep that approach as Mozilla shifts away from XUL+XPCOM extensions. 19:40:19 well, true, but i thought we had #17248 for a reason 19:40:20 (WebExtensions APIs can be extended to give us what we need, in theory) 19:40:25 Right. 19:40:53 I just don’t have a good sense of how much help we will get from Mozilla. I know they like us though ;) 19:41:05 Also, we can patch tor-browser.git to have what we need until Mozilla uplifts it. 19:41:11 okay. so we should keep an eye on that one it seems. 19:41:13 yes 19:41:53 There does seem to be a real need for a Linux launcher but maybe it should not be based on Tor Launcher and it should not replace it. 19:42:12 Or to put it another way, maybe it is not something we (Tor Browser team) will provide. 19:42:27 that's what i am incinced to think right now 19:42:31 *inclined 19:42:36 Is there any way to get Mozilla-based Tor launcher to work with the Linux sandbox? 19:42:38 The sandboxing angle is interesting though. Lots to think about in the long run. 19:43:16 arthuredelstein: Maybe, although there are definitely security considerations. I am sure Yawning thought about that oo. 19:43:18 too 19:43:59 yes 19:44:50 the sandboxing angle is indeed interesting but i think until we start to tackle the windows platform 19:44:54 I suspect something much smaller is a lot safer (but obviously it is more work to maintain two things). 19:44:54 it is not that urgent 19:45:00 OK 19:45:28 I will say something on tbb-dev about all of this. 19:45:49 ok, i was about to ask you, thanks. 19:46:05 other things to discuss? 19:47:14 then thanks for the meeting *baf* 19:47:19 #endmeeting