14:59:44 <richard> #startmeeting Tor Browser Weekly MEeting 2022-05-02 14:59:44 <MeetBot> Meeting started Mon May 2 14:59:44 2022 UTC. The chair is richard. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:59:44 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 15:00:00 * Jeremy_Rand_36C3[m] temporarily bricked the Whonix-Gateway VM on my Talos, so is joining the meeting from my travel laptop 15:00:06 <PieroV> new month, can we delete the old discussions from the pad? :) 15:00:11 <richard> yes! 15:00:29 <sysrqb> o/ 15:00:56 <Jeremy_Rand_36C3[m]> (can someone paste the pad link? I don't have it on this laptop) 15:01:01 <richard> https://pad.riseup.net/p/tor-tbb-keep 15:01:07 <Jeremy_Rand_36C3[m]> thanks richard 15:08:19 <richard> ok 15:08:57 <richard> i think the pad has mostly settled 15:09:07 <richard> PieroV: looks like you have the floor :) 15:09:11 <richard> donuts: ping 15:09:18 <PieroV> okay, so I'd begin with the offline docs 15:09:24 <richard> wfm :) 15:09:25 <PieroV> that is my most important topic for today 15:10:02 <PieroV> I've done a pair of tests, and I think that even though injecting files into omni.ja is a bit hackish, is the easiest way to reach our objective 15:10:36 <PieroV> We already customize 000-tor-browser.js for each language. The only disadvantage is if we'll want to create a single language pack in the future 15:11:28 <PieroV> So, my idea is adding a project to tor-browser-build do create a tarball with the offline docs in every language, and then we add these files to the omni.js 15:11:32 <PieroV> * omni.ja 15:11:51 <PieroV> Possibly, I'd like the html to be a single page 15:12:21 <PieroV> In this way, we could access it through an about page, and use anchors for navigation/topics. This would be the only change in tor-browser 15:12:35 <richard> where does it get injected into the Firefox build? 15:12:47 <PieroV> In the tor-browser project 15:12:53 <PieroV> There's a foreach on all languages 15:14:10 <PieroV> See for example line 363 15:14:18 * donuts is lurking but traveling today 15:14:27 <PieroV> (of projects/tor-browser/build) 15:14:33 <richard> ah I see 15:14:41 <boklm> actually, I'm wondering if we should do multi-lingual builds for the new architectures like linux-arm, to avoid adding a few GBs for each release 15:15:02 <richard> boklm: I *love* the idea of a multi-lingual build 15:15:19 <PieroV> Oh, I see. Then I should have a look if we can customize the about page for each language 15:15:25 <richard> though if i recall correctly the issue we have is with first-launch UX and getting users into the right langpack right? 15:15:44 <PieroV> Like some magic that returns the file from a function, instead of having the file hardcoded in aboutRedirector.cpp 15:16:16 <boklm> I don't remember what was the issue with multi-lingual builds 15:16:41 <boklm> tor-browser#17400 was the ticket about that 15:17:37 <richard> well, from a dev sanity perspective, anything we can do to get to that point sounds great, but I suspect UX will have thoughts there 15:18:16 <richard> boklm: perhaps we should experiment with the linux arm-build as you suggst boklm 15:18:37 <richard> but anyway back the the manual 15:18:40 <boklm> yes, I think we could do that 15:18:57 <Jeremy_Rand_36C3[m]> No objection to multilingual linux-arm builds here, though I have basically no experience with the locale code 15:18:58 <PieroV> I've checked in aboutRedirect.cpp, and some about pages have already a special if, so it can work it anyway, but the single page would be preferrable 15:20:13 <richard> ok got good to know 15:20:16 <PieroV> (provided I find a way to get the language code in nsAboutRedirector::NewChannel and/or other methods of that class) 15:20:42 <PieroV> However, this change on the Tor Browser side should be easy enough 15:21:11 <richard> i'm sure it will just take some code spelunking :) 15:21:26 <PieroV> Another change for tor-browser.git is changing the help URLs to the offline pages 15:21:33 <richard> very worst-case scenario we can write it to a pref ourselves 15:22:06 <richard> yes, that is something that came up with the about dialog change 15:22:11 <boklm> Jeremy_Rand_36C3[m]: in theory, setting `var/multi_lingual` to 1 would be enough. But we haven't used that option recently so maybe it will need some update. 15:22:34 <PieroV> Anyway, we'll need to customize the HTML generation a little bit and add it to tor-browser-build 15:22:42 <richard> boklm: can we make testbuild-alpha targets multi-lingual? 15:23:23 <boklm> richard: we could enable multi_lingual for testbuild 15:24:12 <PieroV> I think that we (me?) could start working on the docs creation even without the custom templates 15:24:24 <PieroV> Has anybody already started working on that? 15:24:47 <richard> PieroV: that sounds like a good idea 15:24:58 * boklm did not 15:25:26 <richard> you in in tor-browser-build? 15:25:43 <richard> you mean in tor-browser-build* 15:25:48 <PieroV> yes 15:26:00 <richard> ah sounds like no 15:26:13 <richard> it's your lucky day :) 15:26:22 <PieroV> okay; I can start working on it during this week 15:26:34 <PieroV> last question about this topic 15:26:37 <boklm> I can review it 15:27:10 <PieroV> I think I'll have to ask somebody from the website team for technical details about our site machinery, do you have any suggestion? 15:27:42 <boklm> there is a #tor-www channel 15:27:49 <PieroV> cool, thanks 15:28:03 <richard> today I learned 15:28:19 <richard> ok 15:28:20 <richard> Android 15:28:28 <PieroV> I have a pair of question 15:28:36 <sysrqb> PieroV: i suspect lavamind is a good person for the details, and hiro may know some details, too. 15:28:44 <PieroV> The first one can be discussed later in the release discussion 15:28:55 <PieroV> (sysrqb: great, thanks) 15:29:26 <PieroV> And is: do we want to release a fix for crashes in v96, or do we want to get a stable v99 out asap? 15:30:25 <richard> so can you give a brief update on the state of Android atm 15:30:26 <sysrqb> is the fix for the crash simply avoiding telemetry? 15:30:39 <PieroV> In the latter case, the MRs are ready for review, I need just somebody to confirm that it's the same patch set as the alpha 15:30:44 <PieroV> sysrqb: yes, it is 15:30:49 <sysrqb> okay. hrm. 15:31:18 <PieroV> richard: of course. Alpha and stable in Android are the same, because we don't have any new development in 11.5 15:31:20 <richard> I can take a look at your MRs in a few hours 15:31:50 <PieroV> This alpha is so cursed that now it may even have reproducibility problems 15:32:05 <sysrqb> PieroV: releasing a stable with the telemetry patch seems reasonable to me 15:32:29 <sysrqb> i worry about rushing v99 into stable this week 15:32:39 <PieroV> But if we can sort it out we can release the stable basically whenever we want (we may want to test run the alpha a little bit first) 15:32:50 <sysrqb> yeah. 15:33:20 <PieroV> Also, latest Android versions haven't been audited 15:33:34 <richard> yeah having stable on the older train 15:33:42 <richard> makes sense to me 15:33:47 <sysrqb> this is a good idea, let's plan another stable based on Fx96 15:33:57 <sysrqb> PieroV: the audits are on my plate for this week 15:33:59 <PieroV> The last one was on v94. Patching v96 would have the advantage that we could also audit v95 and v96 within the week 15:34:10 <PieroV> sysrqb: great news, thanks 15:34:17 <PieroV> I can also audit one of them if you want to 15:34:34 <richard> (so my 94 audit has some todos still, it flags some suspicious stuff but I haven't made issues yet to follow up or fix them) 15:34:58 <PieroV> (sorry for the closing and opening that issue Richard, I pressed the wrong button) 15:35:01 <richard> (just to be clear there :)) 15:35:15 <richard> hah no worries 15:35:42 <PieroV> For Android, I'm also done with v100 preparation for nightly 15:35:46 <PieroV> But of course it can wait 15:36:38 <PieroV> It still has the quick fixes both for telemetry and for the merge .aar phase. I thought that we could handle them in separate issues, instead of dealing with them with the rebase 15:37:18 <PieroV> But again, it can wait a little bit, just know that it would be ready :) 15:37:59 <richard> ok 15:38:05 <richard> on to the next thing then 15:38:19 <PieroV> the next thing is that it makes sense that we keep with publishing alphas on Android 15:38:39 <PieroV> We could ask to do the tests on nightly, if anyone is doing it 15:38:51 <PieroV> It would save hours to people 15:39:33 <PieroV> Which is the advantage 15:39:38 <richard> i suspect Nightlies for android do not auto-update no? 15:39:39 <boklm> we don't have nightly releases on google play 15:39:50 <richard> ^that 15:40:05 <PieroV> Yeah I was about to write it 15:40:22 <PieroV> And that is the first disadvantage, the second one is that we're doing the test campaign on the alpha 15:41:28 <PieroV> anyway, I think that answers my question 15:41:39 <richard> alright 15:41:42 <richard> what's LFS? 15:41:48 <PieroV> Large filesystem 15:42:01 <PieroV> I've had this problem when pushing the v100 to my repo 15:42:35 <PieroV> (Firefox v100, I mean, from gecko-dev) 15:42:46 <richard> I don't know what this or what problems it would cause 15:42:50 <boklm> it seems mozilla started using LFS recently 15:43:08 <PieroV> at the moment they only have a few tests stored with LFS 15:43:09 <boklm> as I don't remember having that issue before 15:43:22 <PieroV> And we don't run these tests, so we could continue not caring 15:43:52 <PieroV> But keep this in mind if we want to run them, and if you'll encounter this problem you can just disable lfs on your repos 15:46:08 <PieroV> And to my final point, if you don't have anything to add for lfs 15:46:59 <PieroV> I've finally started writing the building page :) and I was wondering if we should change the testbuild default target, instead of having to tell about it to new users 15:47:25 <PieroV> + other useful defaults that I still don't know about, like the fetching options 15:48:02 <richard> I'd say make a ticket w/ your suggestions for discussion 15:48:17 <richard> I should think the testbuild target defaults should match how we actually use the testbuild target 15:48:18 <PieroV> Or adding the option to enable network on builds to rbm.local.conf.example (but of course commented), since it's useful to find Android dependencies 15:48:35 <PieroV> richard: okay, creating the ticket works for me 15:48:55 <richard> ok finally 15:49:13 <boklm> I think we could have separate testbuild makefile targets for alpha and nightly, to avoid configuring it in rbm.local.conf 15:49:44 <richard> boklm: last IS aw in the 11.0.10 ticket you were uploading signed builds; I assume those have since finished and nothing is blocking me from making the blog post and website changes? 15:49:52 <richard> last I saw* 15:50:08 <richard> also 11.0.11 not 11.0.10 15:50:11 <boklm> richard: yes, they are now uploaded and should be ready 15:50:24 <richard> ok great, thanks for getting those done so quick 15:51:09 <richard> if the signing/release documentation isn't compltely up-to-date with the new process, if you could make any tweaks/edits necessary before I sign alpha this week that would be lovely 15:51:28 <boklm> ok 15:51:28 <richard> but i am very much looking forward to using the updated scripts :D 15:52:03 <PieroV> when is the next alpha? 15:52:20 <richard> rebasing is schedule for today awith a releaes target of next week 15:52:34 <richard> (desktop) 15:52:48 <richard> (we can try for a join-release if it aligns w/ android) 15:52:48 <PieroV> I see. Any chance of getting tor-browser!275 in it? 15:53:11 <richard> ah is that ready to go then? 15:53:22 <richard> yeah I'll review and get that in 15:53:36 <PieroV> oh wait, I've just remembered of a pair of strings to fix 15:53:55 <PieroV> For one we have a ticket, and one is just a placeholder 15:54:23 <PieroV> So I guess we'll have to delay it 15:55:40 <richard> ok wfm 15:56:10 <richard> with that I think that is all discussion points handled 15:56:23 <boklm> richard: by the way I have been doing the signing of 11.0.11 with the new script `tools/signing/do-all-signing` 15:56:33 <richard> :D :D :D 15:56:50 <richard> in the master or 11.0-maint branch? 15:56:51 <boklm> but maybe you want to first try the scripts separatly for this alpha 15:57:04 <richard> ah yeah most likely a goo didea 15:57:33 <boklm> richard: currently it's still in the branch bug_40476 on my gitlab repo 15:58:07 <richard> ok I will def check it out 15:58:25 <richard> thanks everyone! 15:58:28 <PieroV> Thanks! 15:58:31 <richard> have a good week :) 15:58:33 <richard> #endmeeting