18:00:15 #startmeeting Tor Browser Release Meeting 2023-08-7 18:00:15 Meeting started Mon Aug 7 18:00:15 2023 UTC. The chair is donuts. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:15 Useful Commands: #action #agreed #help #info #idea #link #topic. 18:00:20 le pad: https://pad.riseup.net/p/tor-browser-release-meeting-keep 18:04:07 okay ready? 18:04:19 Yes 18:05:17 thanks for answering my Q in the pad 18:05:19 ja 18:05:35 We're doing the meeting there :D 18:05:42 lol 18:05:57 That thumbs up is not very readable with that yellow background :D 18:06:05 yeah it sucks 18:06:27 magenta 18:06:30 everything goes with magenta 18:06:35 okay so shall we go straight to your first discussion item pierov? 18:06:40 i'm assuming you're green 18:06:57 Yes I am 18:07:11 So, as I wrote in your question, 13.0a2 is kinda big too 18:07:26 We had to push 13.0a1 out because it was becoming too late and we needed to do an update 18:07:32 But it was sorta incomplete, because it missed Android 18:07:43 And the android part was big because... Android 18:08:02 will 13.0a2 be all platforms or android only? 18:08:06 yeah, so we inserted alpha into the empty weeks of the release calendar 18:08:06 All platforms 18:08:09 13.0a2 -> all platforms 18:08:12 ack 18:08:22 Because 115.1.0 came out in the meantime 18:08:30 the plan is to wait for a nightly and then do an alpha release 18:08:53 so realistically i'll start release prep today and hopefully have it ready for review and building tomorow 18:09:34 Well, it depends on two things, actually 18:09:52 1. the nightly server doesn't have the Android toolchain yet 18:10:03 So, the first nightly with Android will take long 18:10:08 And we need to do QA 18:10:16 (assuming the build will succeed) 18:10:22 yes 18:11:10 2. tor-browser!717 18:11:17 even if things aren't quite ready i want to get tb-build-05 builing android stuff for the same toolchain reasons :D 18:11:33 damn my TB crashed again 18:11:41 :( 18:11:45 >:[ 18:11:52 three times today D: 18:11:54 richard: actually you could already run the testbuilds 18:12:11 donuts: is it alpha or release? 18:12:31 12.5 stable but I'm on the macOS beta, so possibly related to that 18:12:41 i opened a ticket 18:12:44 Let's hope so :D 18:13:16 We merged a Moz change in 12.5.1 that wasn't originally in 102, but was very easy to backport, and made that part of code equal to 115 18:13:37 To fix a crash on Windows, but we didn't have the crash and the user didn't confirm the patch solved theirs 18:13:51 Still, I tested the patches before releasing, but I didn't have crashes 18:14:03 But that could be the only different thing from Firefox + the usual patch set 18:14:57 Anyway, re alpha releases 18:14:59 that was accessibility related wasn't it? 18:15:07 It was related to text selection 18:15:56 115.2 is going to be released on August 29, and it might be the target for 13.0 18:16:14 115.3 is scheduled for September 26 18:16:52 The sooner we can release 13.0a2 is August 9, if not even August 10 18:17:33 13.0a3 is currently scheduled for August 22 18:18:50 But it seems we can't do better, so things should be settled for August 18:18:50 yes 18:19:27 August 31 to September 4 is Tor holiday 18:20:37 which makes 13.0a4 v optimistic for the 5th 18:20:52 Yep, I didn't consider that. 18:21:38 And we also have to backport Android on August 29-30 18:21:44 yes 18:22:20 At least we'll have nightlies going on, unless things break just before the holiday ^_^; 18:22:44 we're bakc on the 4th right? 18:22:48 it's the 31st and the 1st 18:23:02 the break I mean 18:23:25 5 is labour day, which I don't care that much (I do May 1st :P) 18:23:38 Sorry, 4 is labour day 18:23:48 ah I see 18:23:58 yeah i can plan to shift my holidays aroun to fit the browser schedule 18:24:56 I wanted to take half week also in the week of September 11 18:25:08 PieroV: if you plan to do the release prep for 13.0a4 then we can build overnight an i can sign the next day 18:25:23 (never forget :crying eagle emojjii:) 18:25:32 We can try to build on August 30 fwiw 18:25:48 yeah we should aim for that 18:26:52 A final changelog change shouldn't take much to rebuild... But it also means we're getting something in the alphas at the last minute, which isn't great, considering that we're in a very advanced stage 18:27:13 well we're going to be *starting* implementing YEC stuff the first week of sept 18:27:19 whihc uh, also isn't great 18:27:22 (that's to say that rel prepping at the end of August should be definitely possible, imho) 18:27:33 yeah I agree 18:27:36 just need to be on top of it 18:28:14 So far the calendar is pretty correct already, gj richard :) 18:28:17 having gone thorugh the combined mullvad+tor browser release prpe a few times, i think i can start merging the two release days in the calendar 18:28:58 since prepping and build at the same time seems to be working pretty well 18:29:03 Yes, it's working very well. We should add the rel prep step of checking if we've linked stuff to Tor Browser but not to MB 18:29:13 yes true 18:29:15 Because I always forget to do it 18:29:54 We should do a bot to automate the linking after we merge, depending on the checkboxes :D 18:29:57 fortunatley mullvd browser release prep is 99% tor-browser release prep + ublock and mullvad extensions + changleogs vOv 18:30:08 lol MRs are welcome 18:30:23 but that would def save me like an hour each time doing it manually 18:30:44 Anyway, 13.0a5 around Sep 13 seems pretty correct, and it should be the RC 18:30:55 but stil faster than manually wrting the changelogs :D 18:30:58 So, no more changes after that date, unless bug fixing and or trivial 18:31:09 (or YEC) 18:31:18 And yes, probably we'll have to do an exception for YEC 18:31:44 But that doesn't worry too much for desktop 18:31:50 It's quite self-contained, usually 18:32:09 paging dan_b 18:32:13 *bing* 18:32:19 He's afk today 18:32:42 ah well 18:32:49 It also means we have 1 month and 1 week for 13.0 18:33:50 that's... not a lot of time 18:34:28 that's why so many alphas 18:34:50 okay well if you're gravy i'm gravy 18:34:51 realistically 13.0 can slip to end of sept 18:35:07 as far as ESR eol is concerned 18:35:29 but yeah we're aiming for no new features mid sept 18:35:35 fyi i'll be afk for the last week of september for the S9 trip 18:35:43 (YEC work can go in between the 13.0 and 13.0.1 release) 18:35:44 but I can arrange the particulars in advance 18:36:44 ack 18:36:51 okay so it sounds like you're planning on 3 more alphas? is that correct? 18:37:18 Yes, with 13.0a5 being a sort of RC 18:37:25 gotcha, cool 18:37:39 sounds... doable? 18:37:48 i'm definitely happier with the more alphas plan 18:39:20 Okay, I think I've reported the reasons in the pad 18:39:48 (should be all we've discussed so far) 18:39:57 ty for taking notes pierov 18:40:30 yw :) 18:40:59 Shall we go with point 2? 18:41:03 let's do it 18:41:12 general happyness? 18:41:38 I have found it surprisingly okay so far 18:41:43 I'm doing some heavy changes that hopefully will make our life easier when introducing Arti 18:42:00 13.0a1 i mean 18:42:23 But they weren't all included in 13.0a1. However, one that was included in it produced some regressions on the circuit display 18:42:43 (the cleanup interval, namely, but it's a sort of edge case, I'd say, and I'll have to investigate it more) 18:43:27 I hope we get enough testing for htem 18:43:52 PieroV: once things are fixed/stable in the tor backend changes, we need to piroritize the remaining s131 work 18:44:06 richard: ack. But what about audits? 18:44:16 specifically enabling MB as default browser on windows and the flatpack work 18:44:25 audits are in progress 18:44:25 maybe instead of doing a recruitment email in tor-internal for the RC, we beg everyone to start testing alpha now 18:44:31 i'm in the middle of 109 18:44:41 donuts: yes please 18:45:08 issues that need to be reviewed have issues in gitlab 18:45:30 if my current rate of 3 releases per week continue we should have a short list of tickets to review by the end of next week 18:45:55 okay i'll send round an email when 13.0a2 is out, if that sounds good? 18:46:04 wfm 18:46:05 Yes, thank you! 18:46:47 richard: it might make sense to wait for the MR you've started reviewing earlier 18:46:58 And do proper QA with nightlies immediately after 13.0a2 18:49:51 wfm 18:49:59 though i did bootstrap eventually 18:50:14 13.0a1 was also being a slowboat so seemed it's just the internet 18:50:43 Ack. Well, we have to wait for Android, so we might merge and in case tag before these changes 18:50:51 Or revert 18:50:53 wfm 18:51:17 Do audits have a label? 18:51:59 Or are they in FF115-esr? 18:52:05 ~"Code Audit" and ~ESR-15 or w/e it is 18:52:10 115* 18:52:28 or I should say 18:52:38 we have the audit meta issues in tor-browser-spec 18:52:48 and they have linked from them the various review tasks in tor-browser 18:53:05 the review asks all have the format "Review Mozilla XXXXXX: blah blah blah" 18:53:12 In ~FF115-esr we have 115 open issues 18:53:18 and the ~"Code Audit" label 18:53:54 4 minute warning 18:54:06 They're 16 18:55:04 sounds about right 18:55:41 But 115 is also quite a lot 18:55:56 Probably we can't take anything else, except what's already been decided and these issues 18:56:55 yeah i'll do a pass and clear out the noise 18:57:10 i'm sure some of those are optimistically labled 18:57:18 okay 1 min left, I'm assuming the tickets listed under the third discussion point are cool and awesome and we have no further changes there 18:57:35 might be good to have a deadline for asset swaps, although we'll try and hand over as many as possible over the next couple of weeks 18:58:05 aaaaaand I need to jump on a call now 18:58:06 Asset swap isn't too worrying, but it might be needed sooner rather than later on Android 18:58:11 Ack, thanks donuts! 18:58:13 yes^ 18:58:15 ok later donuts o/ 18:58:24 thanks both! 18:58:33 #endmeeting