14:58:57 <richard> #startmeeting Tor Browser Weekly Meeting 2022-11-07
14:58:57 <MeetBot> Meeting started Mon Nov  7 14:58:57 2022 UTC.  The chair is richard. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:58:57 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:59:15 <richard> hello everyone, meeting pad as usual here: https://pad.riseup.net/p/tor-tbb-keep
14:59:21 * Jeremy_Rand_36C3[m] checks clock, is pretty sure I'm here at the right time
14:59:39 <ma1> yes it is
15:00:44 <dan_b> i made it, woo
15:02:05 * ma1 offers dan_b a steaming coffee cup
15:02:29 <dan_b> hahaha thanks
15:04:52 <richard> ok let's get started
15:05:41 <richard> I have no announcements this week!
15:06:13 <boklm> is this an announcement? :)
15:06:21 <richard> perhaps
15:06:44 <richard> regarding tickets, things living in ~Sponsor 131 and ~Next are highest priority, so if you are wondering what to do next, please grab from there
15:06:53 <richard> so maybe that is an announcement
15:08:16 <richard> if nothing there looks like a good fit for you, then move on over to ~Backlog instead of ~Next and generally prefer things wih the ~FF102-ESR and ~Sponsor 131 labels
15:08:17 <richard> anyway
15:08:50 <richard> discussion point: I think i recall last week seeing a thread in gitlab somewhere that the stick lang pref didn't get applied for en-US
15:08:51 <PieroV> I think I'd like to promote one thing from Icebox lol
15:09:05 <dan_b> yes I'll take a look at that today
15:09:08 <richard> PieroV: now that is a bold claim :D
15:09:31 <PieroV> re sticky pref: I had even though of adding a secondary pref as a safe measure :(
15:09:32 <dan_b> seems there was a second line in tbb I needed to apply sticky to 🙂
15:09:47 <PieroV> (sorry for noticing that second line, too)
15:09:55 <PieroV> * for not noticing
15:10:08 <dan_b> no worries
15:10:15 <PieroV> richard: the icebox thing is tor-browser#32411
15:10:33 <PieroV> And basically its dupe tor-browser#31064, which is future
15:10:34 <richard> i'm surprised that one is in icebox
15:11:02 <PieroV> The thing is that we are adding a bar to tell users to change their language, and it triggers letterbox for about:torconnect
15:11:11 <PieroV> (I've added "and others" to the title earlier)
15:11:33 <PieroV> Your patch works, but I'm not that confident in setting a series of pages on par with about:blank
15:11:45 <richard> yeah that *seems* reasonable to me
15:12:20 <PieroV> ok, moving it to ~Doing or ~Next then
15:12:37 <henry-x> I feel like that bar needs a redesign to work well with our letterboxing. I get it all the time when normally browsing with tor browser. Usually the missing media codec notification
15:12:55 <Jeremy_Rand_36C3[m]> PieroV: relatedly I've been running into very annoying UX caused by letterboxing on the NoScript popups
15:13:00 <richard> ok, if we need some more patches for the localization migration, then we need another stable build
15:13:20 <henry-x> Similar with the "find in page feature". That's just my pet peeve though
15:13:32 <PieroV> richard: should we wait a week and include sec backports?
15:13:36 <Jeremy_Rand_36C3[m]> Letterboxing causes the height of the NoScript popup to shrink to the point that I need to scroll in order to reach the allow/reject buttons
15:13:54 * richard quickly checks calendar
15:14:10 <PieroV> Jeremy_Rand_36C3[m]: that's a "known" issue, too: tor-browser#31064
15:14:18 <ma1> Jeremy_Rand_36C3[m], so we should checks whether webextensions are treated as "privileged content" for letter boxing purpose in the new patch
15:14:58 <Jeremy_Rand_36C3[m]> yeah makes sense.  Maybe we do want to prevent fingerprinting by malicious WebExtensions though?
15:15:01 <richard> PieroV: yeah we may as well
15:15:11 <PieroV> henry-x: relatable for find, but codecs should not appear in TBB, imho
15:15:14 <Jeremy_Rand_36C3[m]> Or is the WebExtensions API flexible enough that it's pointless?
15:15:54 <ma1> Yes, it's pointless
15:16:13 <ma1> (we already relax RFP stuff for webextensions)
15:16:23 <Jeremy_Rand_36C3[m]> Yeah, not surprised I guess
15:16:46 <ma1> Anyway, bug nextized, doingized and taken :)
15:16:54 <richard> :p
15:17:06 <PieroV> ma1: we should see if they are the same bug
15:17:15 <PieroV> I didn't dupe them already, but I was tempted to do that
15:17:35 <ma1> PieroV, I'll check as soon as we're out of this meeting
15:19:06 <richard> boklm: what's the currrent state of universal bundles?
15:19:16 <richard> macOS universal bundles*
15:19:55 <boklm> richard: I had an error while doing a testbuild after rebasing, I still need to look at it
15:20:09 <boklm> but I think it's almost ready
15:21:01 <richard> is it reasonable to plan for an alpha next week to go along with a rebase?
15:21:34 <boklm> yes
15:21:43 <henry-x> PieroV: The notification is "To play video, you may need to install the required video codecs." and it'll show up for me every now and then. If you go to "twitter.com" and scroll down and wait a bit it pops up. It uses the same popup as the language notification you are working on, so has the same letterboxing problem.
15:22:37 <ma1> But in that case the content is not privileged, right? Not much to do be done about it?
15:23:54 <ma1> (but it seems strange that the notification box changes the content geometry in a way that trigger letterboxing. Maybe something to be changed in the algorithm?)
15:24:04 <henry-x> yeah. I'm just saying it would be nice if that notification didn't change the browser content size since it gets used every now and then. The popup panels work better
15:24:38 <richard> ok i've updated my toto list for next week
15:25:14 <richard> henry-x: sounds like something we should plan for 12.5
15:25:26 <PieroV> agree, and possibly uplift
15:26:26 <henry-x> I wouldn't be surprised if `gNotificationBox` is due an upgrade within mozilla-central
15:28:39 <richard> ok does anyone have anything else? I'm not seeing any bolded items
15:28:52 <Jeremy_Rand_36C3[m]> nothing from me
15:28:52 <richard> except for the cictui UI bugs
15:28:54 <PieroV> I've bolded an item in ma1's section
15:28:58 <richard> i'm just bold-blind apparently
15:29:09 <richard> ok what's going on in the circuit UI
15:29:10 <henry-x> I suppose I'm curious what is happening with weblate and fluent
15:29:28 <PieroV> Should we really investigate circuit UI bugs for 12.0?
15:29:33 <dan_b> just that if PieroV has more ideas for the locale saving bug to add some thoughts to it before I just fix the one line in tbb
15:30:17 <richard> ok 3 more items then
15:30:27 <PieroV> richard: okay, I think I can answer to all 3 of them
15:30:28 <ma1> PieroV, you mean it's not worth?
15:30:41 <richard> alright then :)
15:30:53 <PieroV> ma1: yes, we have many circuit UI bugs
15:31:09 <PieroV> And tbh, that part should be completely dismantled in 12.5
15:31:30 <ma1> richard, it seems the UI tends to disappear and/or become messy in weird circumstances. But it it's doomed anyway, I'll drop the ball :)
15:31:55 <PieroV> For a number of reason, and one is that it's all JS loaded by browser.xhtml in its onload
15:32:07 <richard> ma1: yeah tbh I think the implementation should probably be revisited
15:32:35 <PieroV> Which is fine for the visualization, but I think that we should move the part that collects circuits, especially if we hope to get the data directly when we switch to Arti
15:32:51 <PieroV> Rather than having to collect them in the way we do (we listen for circuit creation events, IIRC)
15:33:12 <PieroV> Even though a working implementation would be nice also with little-t-tor
15:34:13 <richard> that makes sense to me
15:34:55 <richard> is that the fundamental issue there w/ the circuit display (the UI listening for tor events rather than referencing some global component that absracts that all away from it?)
15:35:44 <PieroV> richard: I'm not sure of what's happening, but getting everything more sorted could help, I think
15:36:09 <ma1> I can start investigating low priority and manage upset user while we approach the refactoring.
15:36:34 <richard> well regardless, i don't think patching things up on the edges is a high priority right now
15:37:09 <ma1> (and we won't have circuits in S131, so...)
15:37:57 <Jeremy_Rand_36C3[m]> to what extent is the naming system integration in the circuit display affected by this?  (Both the FPF and Namecoin variants)
15:38:41 <richard> i'm those patches will need to be updated when we come back to circuit display
15:38:48 <richard> i'm sure those*
15:38:50 <PieroV> Jeremy_Rand_36C3[m]: I think it's safe to assume that the patches will need to be completely recreated
15:39:11 <PieroV> It might be sensed to merge parts of them before the process
15:39:14 <Jeremy_Rand_36C3[m]> PieroV: OK.  And that will not be for a while, since the rewrite is not super soon?
15:39:15 <PieroV> It's on the list I guess?
15:39:35 <richard> not until 2023 I should think
15:39:50 <PieroV> Jeremy_Rand_36C3[m]: I wanted to start earlier, but then decided to push to 12.5 because it was too long and not doable for the target period for 12.0
15:39:50 <Jeremy_Rand_36C3[m]> Early 2023?  Or mid?
15:39:57 <richard> probably around the time we begin arti integration work
15:39:57 <PieroV> 12.5
15:40:29 <PieroV> Honestly, I think we could start with a copy&paste for torbutton
15:40:42 <PieroV> And then start splitting it into separate commits, since it has many components
15:40:58 <Jeremy_Rand_36C3[m]> OK.  When that gets near, I guess give me and Arthur a heads up and we can try to help.  I'd really like to have the FPF and Namecoin circuit display codebases converge, it's unfortunate that there's duplication of functionality there.
15:41:00 <richard> sounds smart to me
15:41:05 <PieroV> For example, the drag&drop protection doesn't need to be with the other tor-related stuff
15:41:43 <PieroV> henry-x: re weblate and Fluent, what do you want to know exactly? I've checked again tor-browser!374, and the linked GitHub issues are still open
15:41:54 <PieroV> These issues are about attributes
15:42:27 <henry-x> The other issue is their handling of newline and indenting
15:42:32 <Jeremy_Rand_36C3[m]> FYI I've been super busy the last few months with funded work (mostly TLS related) but that funded work will be over at the end of the year, so if the circuit display stuff is happening in 2023 then I should be more available than I have been lately
15:42:44 <PieroV> So, recently I've named Fluent because I'd like to create the strings for the language notification with Fluent
15:42:59 <PieroV> Jeremy_Rand_36C3[m]: ack
15:43:21 <PieroV> They are simple strings, with a pair of variables and the brand name
15:43:52 <PieroV> So, if Weblate can handle them already, I think it's sensed to manage them with Fluent/Weblate
15:44:07 <PieroV> I assume that patch will be in base-browser, too, since base-browser is also multi-lingual
15:44:29 <PieroV> I don't have a newline in these strings. But I don't have the issue number for that handy
15:46:49 <henry-x> Do you know how well weblate handles group comments, or variables?
15:47:21 <PieroV> henry-x: nope, I had tested Fluent before we switched to Weblate
15:47:33 <PieroV> And then your file was the first one to be tested after the switch
15:47:46 <PieroV> (well, also brand.ftl, but that didn't have issues, afaik)
15:48:58 <PieroV> My new files have some variables
15:49:04 <PieroV> *new file has
15:49:26 <PieroV> https://gitlab.torproject.org/tpo/translation/-/merge_requests/5
15:50:39 <PieroV> dan_b: for the language pref, I think we could add our own preference that tells the original language a user downloaded in case something doesn't work with 12.0
15:51:02 <dan_b> ah ok, so add lines in tbb to inject those, and lines in tb to persist them?
15:51:09 <dan_b> just paralel to what we have now?
15:51:36 <PieroV> Sorta. I was thinking a tor-browser change only
15:51:38 <henry-x> ok. How come you filled in English for the other locales? Don't we want them to be missing by default so that fluent knows to check the next locale?
15:51:51 <PieroV> henry-x: emmapeel asked me to
15:52:19 <richard> PieroV: what would an extra pref provide over the existing sticky pref?
15:52:36 <PieroV> richard: it would have saved us from doing a 11.5.8 :D
15:52:56 <PieroV> But nothing, that's why I didn't add the idea to the comments
15:53:14 <henry-x> But fluent works on the basis that if a fluent entry is missing it will look up the entry in the next locale, if we default to always filling in the variables with english then we'll essentially be defaulting to english always
15:53:39 <PieroV> Anyway, I'd implement it with tor-browser.git only, if possible, so we are sure we don't miss lines in tor-browser-build
15:53:50 <dan_b> ah ok
15:54:00 <PieroV> henry-x: I've noticed that. I think we should discuss that with emmapeel then
15:54:09 <PieroV> Because it's something we do differently for all the other formats
15:55:20 <PieroV> Not related, but I think that we could start the migration from transifex to weblate for 12.5a1
15:55:48 <PieroV> (well, just after we release 12.0)
15:55:56 <richard> mmmhm
15:56:53 <henry-x> ok cool. I suppose we aren't using fluent yet, so it makes some sense to default to en-US like we do with properties and dtd and TorStrings. But it would be nice to have all the multi-lingual fluent features in the future
15:57:13 <henry-x> *we aren't using just fluent yet
15:57:36 <richard> ok
15:57:49 <richard> that's almost the hour
15:58:05 <richard> have a good week everyone!
15:58:11 <Jeremy_Rand_36C3[m]> thanks!
15:58:12 <PieroV> Thanks! o/
15:58:16 <richard> #endmeeting