14:58:26 #startmeeting Tor Browser Weekly Meeting 2023-12-18 14:58:26 Meeting started Mon Dec 18 14:58:26 2023 UTC. The chair is richard. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:58:26 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:58:29 o/ 14:58:32 we are getting periolousy close to 2024 14:58:32 somehwo 14:58:50 the pad: https://pad.riseup.net/p/tor-tbb-keep 14:59:08 o/ 14:59:56 13.0.7 is signed and ready for release tomorrow 15:00:45 this is our first time optimistically doing the mullvad browser release early without waiting for the android backports which (among some other tweaks to the release process) seems to have worked out a bit better than prior releases 15:01:32 It feels weird. 15:01:40 last week mullvad began standing up a system where we can push buidl requests to their infra and hopefully over the next few releases we can iron out inefficiencies there 15:01:48 build requests* 15:02:39 ruihildt[m], "It feels weird" like in "too good to be true"? :) 15:03:24 if any of you want to chat w/ me re isa's mail to tor-internal last week feel free to ping me in DMs or we can chat via signal or bbb 15:05:19 my plan for this week is to do the 13.5a3 release asap and continue chipping away at the design doc and whatever other suprises pop up 15:05:47 in other news we've also made an initial inquiry to mozilla about adding build reproducibility of the openh264 codec build system 15:06:20 and so far the relevant area owner seems positive about helping and improving things there 15:06:46 oohhhhh that would be nice 15:07:04 which will open us up to have full/better webrtc support down the line when UDP in the tor network is a thing (and helps out mullvad brwoser in the medium term) 15:07:52 ok, with that i'lll hand it off to any discussion points y'all have and then we can top it off with a s96 update 15:08:59 OH wait one more thing from me 15:09:13 it's grant application time (always) 15:09:42 i'll send a mail out after this meeting about task estimates around future connect assist and arti stuffs for the team to review 15:10:30 i don't see any other discussion points apart from s96, so dan_b, clairehurst, PieroV how goes connect-assist on Android? 15:10:47 settings now load and save in geckoview java! 15:11:02 so this week i'll be looking at next wiring that into firefox-android 15:11:23 (to the about:config prefs/TorSettings module ?) 15:11:25 Pretty good, I figured out how to hide the toolbar for HTML view and ironing out a way to avoid needing to restart the app to have the changes take effect in regard to tor-browser#42323 15:11:44 Plus basic UI for that ticket is fully set up 15:12:00 richard: we should wait for that for alpha imho 15:12:23 until after the break or until like Thus? 15:12:24 At least to play with it during the break 15:12:29 I talked about it with PieroV, we'll have a build person starting in January at Mullvad, and first assignment will be signing macOS MB build with our own cert. 15:12:29 I'd like to have something for that person to get started. 15:12:58 I think we all decided that we wanted the switch in 13.5a3 at last week's release meeting 15:13:29 what was the final switch? between claire's new UI and the tor-browser html UI for bootstrap? 15:13:51 I think at least to enable the new backend 15:13:56 To see if tor dies in real usage 15:14:44 ruihildt: fortunately all our signing scripts are public and in source control so replicating the macOS signing+notarising should be easy enough for them to do 15:15:29 yes if we could get that flow in alpha by the end of the week that would be great 15:15:50 but yeah, what ever the settings pane is on firefox-android, it needs to speak to the new TorSettings in case pierov's new UI is up 15:15:59 and i'm happy to release over the weekend if you need thursday for dev as well 15:16:28 dan_b: I think some testing even without bridges could be useful 15:16:47 testing is usually good 15:16:51 what'd you have in mind? 15:17:11 ruihildt[m]: our signing scripts are in this directory: https://gitlab.torproject.org/tpo/applications/tor-browser-build/-/tree/main/tools/signing?ref_type=heads 15:17:14 The MR for the gate seemed already good enough for me 15:17:42 If switching to the native UI doesn't work we can disable that checkbox and tease with something like "Coming soon" :) 15:17:57 Thanks 15:18:32 ruihildt[m]: and documentation for rcodesign (the tool we use) is in: https://gregoryszorc.com/docs/apple-codesign/main/ 15:18:52 PieroV, would you release alpha with default to new stuff and direct bug-finders to disable it? 15:19:17 No, I would default it to the legacy components 15:20:45 so if that's the case, all the settings pane will need ifs based on the toggle as to where to load/save thier settings for hte moment? 15:20:50 But I think having the thing in alpha would be good also for us (I dog food alphas also in my personal device, but I don't use custom alphas) 15:21:09 * ma1 wonders how many alpha users we've got actually, under the impression bugs are found on stable only :( 15:21:24 +1 on dog fooding 15:22:00 it's possible to dog food nightly too 15:22:15 Dog fooding on nightly scares me, because there are Mozilla features enabled on Android 15:22:29 tens of thousands for Android iirc 15:22:32 re alpha users 15:23:17 should we fix that? (not enable mozilla nightly features on our nightly channel) 15:23:34 boklm: there's an issue about that 15:23:53 But I don't know how long it could take 15:24:20 Basically, on Android there isn't an option to customize branding, as far as we know 15:25:03 But I'd like if we found a way to build everything with the release configuration, and customize branding separately 15:25:31 re alpha users, 218K installed audience compared to 5.08M for stable 15:25:45 (for android) 15:25:49 hundredss of thousands for Android iirc* 15:25:49 :3 15:27:14 (ready the previous lines, I understood we had tens of thousands of nightly Mozilla features enabled on Android) 15:27:34 heh 15:28:23 maybe 'fixing' the nightly channel is something for the next major ESR update over the summer 15:29:16 ok on to the other half of the s96 work 15:29:23 Lox 15:29:33 i heard we had some sync with anticensorship this morning? 15:29:49 Yes 15:31:15 so what's going on there? 15:31:16 We should develop a backend module separately from the frontend 15:31:50 Because Lox will need to ping the Lox authority every now and then for updates 15:32:11 Henry is working on the frontend this week 15:32:38 So, it makes sense that someone else is working on the backend 15:33:10 I think I'm not very needed for Android S96 stuff now (but I can be pinged if needed, of course) 15:33:18 So, it's something I can follow 15:33:38 I'm still not sure how the work will be divided between our team and AC team 15:34:02 We'll have to sync, but I don't know yet if it's more they tell us how lox works and the expected flows 15:34:27 Or if it'll be something more like we tell them what browser mechanisms could be used for what purpose 15:34:41 we'd initially planned for cohosh to do the majority of the backend module work, but she's been out 15:34:56 though oninyang said she should be back this week 15:35:07 hi 15:35:14 so yeah i think supporting cohosh as much as possible this week should be a top priority for you PieroV 15:35:18 good morning o/ 15:35:20 Yep, that's why I say I'm not sure eventually how we'll do :) We'll see in the following hours I think :) 15:35:37 wfm 15:37:04 yeah, I'm putting together what kind of API surface I would like right now 15:37:30 perfect 15:38:51 ok if there are no more points of discussion 15:38:56 i'm happy to end this meeting 15:39:08 (I just sent you all a mail about those task estimates) 15:39:27 received 15:41:59 ok then folks 15:42:04 have a wood week and see you all on irc 15:42:06 #endmeeting