[00:04] <lubot> [telegram] <kc2bez> That is fantastic arraybolt3
[00:22] <lubot> [telegram] <tsimonq2> arraybolt3: I say we just give em an option. Include the top 5 language packs, and if they're not on the list, they get asked if they want to proceed.
[00:22] <lubot> [telegram] <tsimonq2> That's a compromise for your compromise slightly leaning towards "download everything" if it was on a spectrum :P
[00:23] <lubot> [telegram] <tsimonq2> Nothing is set in stone yet tho
[00:24] <lubot> [telegram] <tsimonq2> @lynorian I see you typing on and off :P don't be afraid to say something heh :)
[00:24] <lubot> [telegram] <lynorian> I actually don't know the data for which langauge packs get used the most
[00:25] <lubot> [telegram] <tsimonq2> I'm really glad you brought that up, since I think(!) that Ubuntu main already has a list. Don't remember if it's script generated
[00:25] <lubot> [telegram] <tsimonq2> Off the top of my head I think the 5 should be Spanish, French, Arabic, Russian, and Mandarin
[00:25]  * arraybolt3 wonders what would happen if LXQt wasn't programmed in English and English wasn't one of the top five languages
[00:25] <lubot> [telegram] <tsimonq2> Those seem like the biggest languages in the world besides English
[00:26] <lubot> [telegram] <tsimonq2> Heh, thankfully that isn't the case :P
[00:26] <arraybolt3> tsimonq2: fairly certain that Hindi is like number 2?
[00:27] <arraybolt3> I looked it up at some point, I think Chinese is #1 and Hindi is #2
[00:27] <arraybolt3> ahem, no, it's #3 according to some random Google search result.
[00:27] <lubot> [telegram] <tsimonq2> Do you happen to be interested in finding some firm statistics on that?
[00:27] <lubot> [telegram] <tsimonq2> Hah
[00:27] <arraybolt3> I can try
[00:28]  * arraybolt3 just misread "Modern Standard Arabic" as "Modern Standard Apple" and thought "wait... macOS has its own language at this point?!"
[00:28] <lubot> [telegram] <tsimonq2> "Conduct a large worldwide study and come back in 6 weeks with your results" :P
[00:28] <lubot> [telegram] <tsimonq2> Hahahahahahahahaha
[00:28] <lubot> [telegram] <tsimonq2> Probably a cross between Pig Latin and Mandarin...
[00:28] <lubot> [telegram] <tsimonq2> Sorry, that was an off color joke :P
[00:29] <lubot> [telegram] <tsimonq2> Anyway... Mandarin and Spanish are guaranteed
[00:29] <lubot> [telegram] <tsimonq2> French probably because they have a pretty large Linux base there
[00:29] <arraybolt3> Anyway, from at least several Google snippets, it appears the order of the top six languages is disagreed on, but the six are generally the same - English, Mandarin Chinese, Hindi, Spanish, French, and Arabic.
[00:30] <lubot> [telegram] <tsimonq2> So almost got it ;) swap out Russian for Hindi and I think we're golden!
[00:30] <arraybolt3> https://en.wikipedia.org/wiki/List_of_languages_by_total_number_of_speakers
[00:30] <arraybolt3> yeah I think that'll do
[00:31] <arraybolt3> now that assumes that we want to go by total number of speakers, period, but that doesn't account for "wait, what if there's more Tagalog Linux users than their are Hindi users?"
[00:31] <arraybolt3> i.e., if the number of Linux users per capita differs between language, we'll be thrown off
[00:33] <lubot> [telegram] <tsimonq2> I think the numbers will be comparable tbh
[00:33] <arraybolt3> yeah
[03:36]  * tsimonq2 throws 500 tables against 500 walls
[03:36] <tsimonq2> The lack of documentation for networkmanager-qt is quite concerning.
[03:36] <tsimonq2> arraybolt3: How the heck did you implement anything useful without popping open the hood? :P
[03:37] <arraybolt3> tsimonq2: git pull the plasma-nm code and use it to reverse-engineer :)
[03:37] <arraybolt3> But yes, now you feel my frustration.
[03:37] <tsimonq2> Hahahahahaha, I've been using nm-tray to reverse-engineer.
[03:37] <arraybolt3> https://invent.kde.org/plasma/plasma-nm
[03:37] <tsimonq2> But yeah... wtaf, why do these API docs not exist o_o
[03:38] <arraybolt3> ikr
[03:38] <arraybolt3> but yeah, lots of experimentation, inspecting (and outright copying some of) plasna-nm's code, and I think I did some peeking under the hood once but it didn't give me much info so I abandoned that.
[03:38] <tsimonq2> This command in the networkmanager-qt source honestly saved me: grep -R "Q_EMIT" src/manager.cpp
[03:38] <tsimonq2> I literally had to go off of the signals they emit. :P
[03:39] <arraybolt3> Right now I'm banging my head against the wall trying to figure out why my network creator can make Ethernet connections but not WiFi ones. I'm about ready to call into nmcli again :P
[03:40] <tsimonq2> https://pastebin.com/Mf2dmQqa
[03:40] <tsimonq2> arraybolt3: We need to gang up on the KDE upstream people and ask why this is so awful ;P (Or more productively, and probably what we will do, is ninja in and refactor the absolute **** out of it.)
[03:41] <tsimonq2> I'm like, 12 hours in on *just this one bit* and it's been driving me nuts. :P
[03:41] <tsimonq2> Thankfully after I started over from scratch it worked......
[03:41] <arraybolt3> sheesh
[03:42] <arraybolt3> well now you know why I don't expect the network editor to be done until 24.10 :P
[03:42] <tsimonq2> arraybolt3: Try looking at the unit tests in the actual networkmanager-qt source
[03:42] <tsimonq2> Probably the manager test.
[03:42] <arraybolt3> oh nifty!
[03:43] <tsimonq2> The really funny thing is, upstream *does* run those tests on each push. They're ran on Qt 6 against OpenSUSE. That being said... these automated tests, despite fully passing upstream, have been disabled in Ubuntu and Debian since like 2021. :P
[03:43] <tsimonq2> So to be honest, we have no idea whatsoever if we're running into some random Qt 5 bug >_M
[03:43] <tsimonq2> *>_<
[03:44] <tsimonq2> Might be a fun MOTU task to get those re-enabled and working. :P
[03:44] <arraybolt3> Ugh, we turned the tests off??? Grief.
[03:44] <tsimonq2> The testing story used to be a loooooooooooot worse upstream. RikMills knows. :P
[03:45] <tsimonq2> I think it's just historical baggage at this point. His ballpark, he'd know exactly why.
[03:58] <tsimonq2> https://git.lubuntu.me/Lubuntu/installer-prompt/commit/acfd803cd41e504ccde2315762a87af510c1aea6
[03:58] -ubottu:#lubuntu-devel- Commit acfd803 in Lubuntu/installer-prompt "Finish correct state transitions for WiFi status, next step is actually being able to enter a password"
[04:14] <lubot> [telegram] <tsimonq2> Final look for the night; I have to hide a specific spacer that's in there, but otherwise states are fully done. : https://matterbridge.lubuntu.me/80ba3ae3/file_10181.jpg
[04:14] <tsimonq2> Meaning, when Networking is disabled, disable the WiFi dropdown. If networking is enabled but WiFi is either already connected or no WiFi device exists, disable the WiFi dropdown as well.
[04:15] <tsimonq2> The dropdown is enabled iff (not a typo) a WiFi device is *available* but not *connected*.
[04:16] <tsimonq2> Essentially, when the password box comes up, we need to find a way to securely store that, immediately. We can't keep a plaintext password in memory, at all, anywhere.
[04:16] <tsimonq2> I'm also thinking of disabling the WiFi dropdown box when it's connecting/disconnecting, which I did just add states for, but that may feel a little too glitchy, it all depends.
[04:17] <tsimonq2> Anyway, I just pushed this too: https://git.lubuntu.me/Lubuntu/installer-prompt/commit/36f9a0c3441213fd2f4dae5b8c526155f0f53fe6
[04:17] -ubottu:#lubuntu-devel- Commit 36f9a0c in Lubuntu/installer-prompt "Implement different network connectivity states"
[04:17] <tsimonq2> o/
[04:21] <lubot> [matrix] <arraybolt3> Nice! Thanks for everything!
[05:24] <arraybolt3> Whew, finally! WiFi connection creation now works (and Ethernet connection creation should be fixed, there was a bug in there I had missed).
[05:24] <arraybolt3> Not quite ready to push (I have some optimizations to make first), but lots of fancy changes incoming.
[05:26] <RikMills> tsimonq2: if you mean autopkgtests disabled, then that was done as a resource and QA issue
[05:27] <RikMills> the test infra was crippled with having to rebuild each source to do the test
[05:28] <RikMills> plus if you are rebuilding to do the test, you are not actually testing the build package in the archive
[05:28] <RikMills> and our test infra did not have a alternative to doing that rebuild
[05:29] <RikMills> arraybolt3: ^
[05:57] <arraybolt3> RikMills: Thanks for the info!
[05:59] <arraybolt3> Whew, did a TON of work on the connection editor. Ethernet support isn't perfect but it's functional enough to be considered alpha-quality. WiFi support is starting to come along, and I reorganized header includes to improve compilation speed, which I think helped noticeably (though not massively, but still I think it was worth the time it took).
[06:00] <arraybolt3> (I *almost* made the commit description for the header include reorganization "place all include statements in a box and shake vigorously". But then I went with the more sane "Totally reorganize header includes to improve compilation speed".)
[06:03] <arraybolt3> tsimonq2: Your "look at the unit tests" trick saved me - I looked not at unit tests, but at an example implementation of adding a new WiFi network, and while it didn't solve the problem I was having, it did show me how to get error messages when something was failing, and that gave me the hint I needed to get things working.
[10:04] <lubot> [telegram] <Leokolb> Unable to login at https://discourse.lubuntu.me/
[10:30] <lubot> [telegram] <Roberalz> +1 (re @Leokolb: Unable to login at https://discourse.lubuntu.me/)
[13:51] <lubot> [telegram] <teward001> regarding discourse: known outage.  nowhere near my access keys to address
[14:20] <lubot> [telegram] <tsimonq2> https://www.redhat.com/en/blog/rhel-10-plans-wayland-and-xorg-server
[15:47] <lubot> [telegram] <Roberalz> Same as kde with kde 6
[18:14] <tsimonq2> arraybolt3: <3
[19:49] <tsimonq2> When running the autotests for nm-qt on Noble...
[19:49] <tsimonq2> The following tests FAILED: 1 - managertest (Failed) 2 - settingstest (Failed) 3 - activeconnectiontest (Subprocess aborted)
[19:49] <tsimonq2> >_<
[19:50] <tsimonq2> (Mind you, all tests pass upstream.)
[19:57] <arraybolt3> welp
[19:57] <arraybolt3> wonder if that's where some of my problems are coming from
[20:34] <RikMills> it used to be that 99.9% of the time tests failed due to the ubuntu test environment needing fixing. not the upstream test or code needing fixing itself
[20:34] <RikMills> i.e. the fails were spurious
[20:35] <RikMills> and not an indication of a genuine issue
[20:37] <tsimonq2> RikMills: Given that a) networkmanager-qt is gaining two non-KDE rdeps this cycle and b) there's since been a dedicated team hired at Canonical to fix QA stuff, I'm more inclined to think we'll have better luck this time, even if it means only selectively enabling networkmanager-qt.
[20:37] <tsimonq2> Part of me thinks it's a better idea to wait for Qt 6 where things are already passing, another part of me thinks we *need* to fix the Qt 5 copy if it's actually broken.
[20:38] <tsimonq2> These tests aren't horribly complex, but the API documentation for networkmanager-qt is sorely lacking, so we would need to probably improve that first, then fix/improve/add autopkgtests to the Qt 5 copy, perhaps bump that forward to Qt 6.
[20:38] <tsimonq2> It all depends on how far down the rabbit hole we wish to go, and how much grumbling you'll do ;P
[20:40]  * tsimonq2 brb but mostly done with WiFi Stuff
[20:42] <lubot> [telegram] <tsimonq2> Essentially, I'm having issues with matching an item in the dropdown list to its appropriate connection. I think conceptually there's some inheritance I'm not fully wrapping my brain around.
[20:48] <RikMills> arraybolt3: have a look through upstream changes to see if anything recent(ish) seems relevent to your issues: https://invent.kde.org/frameworks/networkmanager-qt/-/commits/kf5
[20:51] <RikMills> 5.113.0 should also be building in ppa:kubuntu-ppa/staging-frameworks
[20:51] <RikMills> so you could try with that
[20:52] <lubot> [telegram] <RikMills> My c++ is still not as good as my nuclear physics, so I'll pass on that (re @tsimonq2: Essentially, I'm having issues with matching an item in the dropdown list to its appropriate connection. I think conceptually there's some inheritance I'm not fully wrapping my brain around.)
[21:02]  * arraybolt3 is experimenting with labwc in Debian
[21:04] <arraybolt3> Had a rocky start, got it functional after a bit.
[22:02] <arraybolt3> tsimonq2: Welp, guess what. You know all that work you've been doing to make it so that the live session's language can be changed from the initial startup screen? You've now beaten Ubuntu Desktop's functionality.
[22:02] <arraybolt3> Ubuntu Desktop no longer changes the live session language when you choose a different language than English in the installer.
[22:02] <tsimonq2> HAHAHAHAHAHAHAHAHAHAHAHAHAHAHA
[22:03] <arraybolt3> Launched Ubuntu Desktop, selected Spanish in the installer, clicked next, chose "Probar Ubuntu" (or something) since that looked like "Try Ubuntu", and an English live session started. :-/ :P XD
[22:04] <tsimonq2> Buenos dias 😎
[22:05] <lubot> [telegram] <tsimonq2> I was watching Oppenheimer today and thinking the exact opposite! (re @RikMills: My c++ is still not as good as my nuclear physics, so I'll pass on that)
[22:11] <lubot> [telegram] <Roberalz> Jajajjaja (re @lubuntu_bot: (irc) <tsimonq2> Buenos dias 😎)