14:59:46 <richard> #startmeeting Tor Browser Weekly Meeting 2022-11-28 14:59:46 <MeetBot> Meeting started Mon Nov 28 14:59:46 2022 UTC. The chair is richard. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:59:46 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 14:59:53 <richard> pad: https://pad.riseup.net/p/tor-tbb-keep 15:00:13 <richard> I'll give folks a couple mins 15:00:38 <boklm> o/ 15:01:06 <ma1> o/ 15:01:54 <PieroV> msim: re your mrs on tor-browser.git, I think we're not merging anything at the moment, until we'll be done with 12.0a5, if I understand correctly 15:02:37 <msim> PieroV: no stress :) 15:02:50 <richard> ok 15:03:26 <richard> last week was slightly stressful but its fine 15:04:03 <richard> the initial SDK bump patch is causing crashes on Android 12+ devices, but I stopped the rollout once crash reports started coming in 15:04:29 <richard> pierov+dan_b seem to have a plan there but haven't reviewed your MRs yet 15:05:03 <PieroV> dan_b already pointed out that it makes sense to test also old devices, to be sure that we're healing new devices but breaking the older ones 15:05:32 <boklm> +not ? 15:05:44 <ma1> was this a bug which would crash also the Android Studio emulator? 15:05:48 <PieroV> boklm: yes, thanks :) 15:06:06 <ma1> (I suppose so, it was a Java thingm, right?) 15:06:08 <PieroV> ma1: I believe so, provided you ran Android 13 15:06:22 <PieroV> ma1: yes, it's the Android API throwing exceptions 15:06:32 <richard> my tor-browser-build instance on tb-build-05 seems to enjoy building for macOS wrong, not sure yet but I'm starting a completley (well almost completely) fresh build there so will be able to verify the issue's reproducibility (lol) and if not at least have a working vs broken setup to compare 15:07:30 <richard> and finally we had some interruptions with the 11.5.8 mar distributions last week presumably because I messed up during the 11.5.9 release but that's since been resolved 15:08:23 <richard> so top priorities this week obviously fixing android and getting out an 11.5.10 release for android and the 12.0a5 release for everything 15:09:09 <PieroV> (installing right now on Android 10 device) 15:09:16 <henry-x> Is 11.5.10 just for android? 15:09:33 <PieroV> henry-x: yes, it's a 11.5.8 + SDK fix + fix for the SDK fix 15:09:38 <richard> the release calendar currently has the 12.0 release scheduled for next Monday, which seems overly optimistic so I think we'll push it back to whenever 102.6 esr comes out 15:10:18 <PieroV> Tags on 12-08 and release date on 12-13 15:10:46 <richard> woo 15:10:53 <msim> \o/ 15:11:25 <richard> if things are still in a questionable state we can push back agin to january 15:11:26 <PieroV> (I'll be on vacation these days :/) 15:11:39 <PieroV> No please 15:11:44 <PieroV> No more 11.5.x releasez 15:11:44 <richard> someone else will have to build then :3 15:11:49 <msim> lol 15:11:51 <richard> yeah I know 15:12:36 <PieroV> Well, I should be able to build very likely 15:13:36 <richard> ok i don't see any bold items 15:13:48 <PieroV> My test device that wasn't crashing with 11.5.9 continues not crashing with candidate 11.5.10 15:14:00 <richard> music to my ears 15:15:08 <msim> one thing i wanted to mention - in the tor-browser-* branches we have run-tbb-tests, but nothing of the sort for base-browser 15:15:21 <PieroV> msim: these tests are very old 15:15:25 <msim> ah 15:15:41 <msim> so what actually is the testing process 15:16:17 <richard> so there is the tor-browser-bundle-testsuite repo 15:16:29 <richard> that is in a state of disrepair 15:17:14 <richard> for base-browser I would think the builtin mach tests should be theoretically runnable 15:17:14 <msim> ah 15:18:02 <boklm> I think some of the tests will fail because of prefs changes 15:18:03 <richard> but for 2023 we really need to get our testing strategy sorted 15:18:07 <PieroV> +1 15:18:16 <Jeremy_Rand_36C3[m]> (it's gonna be awkward if Namecoin's downstream CI tests of Tor Browser wind up testing major code paths that Tor Project's CI tests don't...) 15:18:23 <henry-x> It would be nice to borrow mozilla-central's testing framework (`./mach test`) for our tor additions. I'm writing UI code now and I keep thinking "this should probably have a test". 15:18:39 <richard> I 100% agree 15:18:49 <dan_b> +1 15:18:50 <ma1> like running mach test and producing patches which "fix" the tests broken by our patches? 15:18:51 <PieroV> henry-x: yeah, we'd really love to 15:18:51 <boklm> +1 15:19:04 <msim> same here, i spoke with tjr briefly before and i'm assuming that one of our patches is breaking some mochitests 15:19:07 <richard> ma1: also this 15:19:07 <henry-x> even though I don't like writing tests ;) 15:19:44 <msim> i can't get `./mach mochitest <webrtc stuffs>` to run properly, when they *should* work fine in mozilla-central (but i'll confirm that tomorrow) 15:20:11 <msim> i'm assuming that it works fine in moz-central since all the tests get run on try server before releases 15:20:11 <ma1> msim, were those the tests broken by falling back to http? 15:20:14 <msim> yep 15:20:20 <richard> yeah, but for s131 we ought to have a plan here, like a real ~#*enterprise*#~ software team 15:20:24 <msim> +1 15:20:35 <msim> but also for tor-browsers sake :P 15:21:05 <richard> but we can figure this all out after the 12.0 release 15:21:35 <msim> sounds good, in the meantime i'll toy around with it since it'll likely make webrtc easier if i can get these tests working and i'll document any interesting discoveries :) 15:21:50 <richard> yes for sure 15:22:00 <henry-x> Can we start opening issues along the lines of "add tests for <feature>" now, or should we hold off? 15:22:00 * ma1 remembers "looking into test coverage" was deemed too unsavory as an internship task :) 15:22:00 <richard> msim: passing mozilla webrtc tests is paramount for your work :D 15:22:18 <msim> richard: not only moz tests but there's a whole suite of tests provided by webrtc at webrtc.github.io 15:22:19 <richard> henry-x: yes please 15:22:26 <msim> and then also a set of samples also provided by webrtc 15:22:41 <msim> which i want all to be passing with flying colours :D 15:23:01 <henry-x> ok. Do we have a meta-issue for tests? Or a label? 15:23:34 <richard> no, but I can create both after this meeting real quick 15:24:16 <henry-x> ok, that would be useful because I'm going through security level code right now and already have some ideas of what to test 15:25:02 <richard> :) 15:25:12 <msim> i can't seem to spot any regarding mach tests, but there's a few about android tests 15:25:29 <msim> ugh my irc borked for a bit there and i missed the last few messages 15:26:02 <richard> henry-x: tor-browser#41493 15:26:15 <PieroV> actually we already had an issue now that I think of it 15:27:32 <richard> and made an Applications scoped ~Test label 15:27:41 <ma1> PieroV, but was it a meta-ticket(TM)? :) 15:27:55 <richard> :p 15:27:56 <PieroV> Yes it was :) But cannot find it anymore 15:28:04 <richard> maybe i closed it vOV 15:28:19 <richard> i'm sure we lost some good issues in the gitlab purge of 2022 15:29:10 <PieroV> next internship proposal will be going through the issues again :P 15:29:28 <richard> ok, does anyone have something else to discuss? 15:29:54 <Jeremy_Rand_36C3[m]> nothing here 15:30:16 <henry-x> nope! 15:30:17 <msim> nope 15:30:25 <ma1> fine here 15:30:26 <dan_b> nope, just running builds for android 😄 15:30:30 <richard> alright, have a good week everyone 15:30:37 <msim> all the best o/ 15:30:42 <richard> (and hope you all had a good long weekend) 15:30:42 <PieroV> thanks o/ 15:30:43 <ma1> cheers :() 15:30:45 <msim> off to get some much needed rest now :D 15:30:45 <richard> #endmeeting