[14:58] <arraybolt3[m]> libfm-qt? Iiiiinteresting... I wonder if this is going to somehow be related to the Breeze bug.
[15:00] <arraybolt3[m]> Also, I just discovered that the Breeze bug does not affect Kubuntu.
[15:01] <RikMills> doubt it
[15:01] <kc2bez[m]> arraybolt3[m]: Sorry to be late to the party. Trying to catch up a little. What is the issue with Breeze?
[15:02] <arraybolt3[m]> Dan Simmons: Blue highlighting isn't working in context menus and toolbars, causing a variety of interesting graphical problems (like seemingly disappearing text in some instances).
[15:02] <arraybolt3[m]> LP: #1992196
[15:02] -ubot93:#lubuntu-devel- Launchpad bug 1992196 in breeze (Ubuntu) "Various highly noticable graphical glitches have all started happening in Kinetic at the same time" [Undecided, New] https://launchpad.net/bugs/1992196
[15:03] <arraybolt3[m]> Only just started happening in Lubuntu Kinetic ISOs.
[15:03] <arraybolt3[m]> Strangely it's not affecting Kubuntu, making me wonder if maybe the Qt update is just fine, and maybe we just needed more no-change rebuilds of LXQt components.
[15:04] <kc2bez[m]> Do you have compositing enabled?
[15:04] <arraybolt3[m]> Not sure, but it used to Just Work.
[15:04]  * arraybolt3[m] boots ISO real quick
[15:05] <arraybolt3[m]> (Actually I think I have an old leftover VM right here, one sec...)
[15:05] <kc2bez[m]> You might need Picom for that. Remember kwin has compositing enabled by default.
[15:05] <arraybolt3[m]> I wonder what Qt might have broken to make it so that it doesn't work without compositing?
[15:05] <arraybolt3[m]> I guess it's probably least obtrusive to just turn compositing on by default if that's enough.
[15:06] <arraybolt3[m]> Nope, turning on compositing had no effect.
[15:06] <kc2bez[m]> Or instruct others to do so if highlighting matters
[15:07] <arraybolt3[m]> I think highlighting always matters - open Featherpad in the latest Kinetic, and click a menu button in the top bar. The text will initially appear to vanish, since it turns white and relies on the blue highlighting (which isn't happening) to still be visible to the user.
[15:07] <arraybolt3[m]> I wonder if KWin is making a difference though. Lemme install it into Lubuntu and see what happens.
[15:08] <arraybolt3[m]> Strange thing is, it only happens with the Breeze theme. We could just switch themes if necessary (though I bet our users will love us for that /s, I know I'm annoyed when an OS changes UI out of nowhere for no apparent reason).
[15:10] <RikMills> any upstream bug reports?
[15:11] <arraybolt3[m]> Not yet, only the downstream one.
[15:12] <arraybolt3[m]> If it turns out to be a Qt bug, there will almost certainly be a Qt bug report pretty soon, but since this only happens with Breeze + Lubuntu, I'm suspecting it's a fault in Lubuntu. It's currently reported against Breeze, but that may change.
[15:14] <kc2bez[m]> Breeze isn't favored in upstream LXQt sometimes https://github.com/lxqt/lxqt/discussions/2098
[15:15] <RikMills> X or wayland?
[15:15] <arraybolt3[m]> :(
[15:16] <kc2bez[m]> X we don't have a Wayland session.
[15:16] <arraybolt3[m]> Well, maybe switching away from Breeze is the best way to solve this until we get it sorted out.
[15:16] <kc2bez[m]> It doesn't seem wise at this point in the cycle
[15:16] <kc2bez[m]> If anything we should release note it
[15:16] <arraybolt3[m]> Won't finding and fixing a Qt bug be even harder and riskier?
[15:17] <arraybolt3[m]> ...maybe I should send a screenshot of the issue, this breaks critical UI functionality for me (the ability to read the buttons at all times seems pretty important).
[15:17] <kc2bez[m]> I can see it in my VM
[15:18] <arraybolt3[m]> I mean, I see your logic. I just... blah.
[15:20] <RikMills> arraybolt3[m]: your "Issue 1" is fixable with a config tweek it seems
[15:20] <arraybolt3[m]> RikMills: I forgot about that. I can just set Selected Text to black and I can read the buttons.
[15:21] <arraybolt3[m]> ...wow now I'm kicking myself for forgetting to test that.
[15:21] <RikMills> the blue highlight doesn't seem to have a similar fix
[15:22] <arraybolt3[m]> It looks like it's just the "Selection" color getting ignored for this sort of thing. That might help in finding a patch if that's something that's safe to do at this point. But we have a solution now that's better than nothing.
[15:23] <arraybolt3[m]> Dan Simmons: Sorry for my wording earlier. I shouldn't insist that a total or near-total fix is the only right way to fix something. You're right about it being late in the cycle to change things around too much.
[15:24] <arraybolt3[m]> s/late/too late/
[15:24] <kc2bez[m]> No worries, it is fairly broken for sure
[15:25] <kc2bez[m]> It is easily changeable by the end user though so work arounds are somewhat simple
[15:26] <arraybolt3[m]> True. And RikMills figured out a solution that should be low-risk that gets things into a workable state (changing the "Selected Text" palette color to black).
[15:26] <kc2bez[m]> It depends on what is important to the user, the Breeze look or the highlighting and Fusion.
[15:26] <arraybolt3[m]> I mean, it messes with other parts of the UI like what happens when you select text, but things are at least still visible.
[15:26] <arraybolt3[m]> I can live with slightly uglier selections if I can still see what I'm doing.
[15:27] <kc2bez[m]> What is weird is it only applies to the menus for the most part. The sidebar in pcmanfm-qt seems to highlight normally.
[15:28] <arraybolt3[m]> Right? Exactly. If all goes well it should be a matter of just tweaking something in Breeze and then everything will just work.
[15:29] <arraybolt3[m]> Or maybe no-change rebuilds will fix it, or something.
[15:29] <kc2bez[m]> If that sync for libfm-qt goes through we will likely need some NCRs anyway.
[15:30] <arraybolt3[m]> Maybe then it will be one of those bugs that just vanishes out of absolutely nowhere one day and everyone's like "Huh, well it works now, nice!"
[15:32] <kc2bez[m]> Loading a different palette doesn't resolve it so it seems to squarely fall on the Breeze Qt Style.
[15:35] <arraybolt3[m]> Installing KWin and switching to it didn't fix the problem either, interestingly enough. This makes me hopefully that the NCRs will do the trick, maybe there's just a Qt incompatibility that will go away when we do that.
[15:35] <kc2bez[m]> cleanlooks might be a little closer fit to Breeze
[15:36] <arraybolt3[m]> Kinda squished but OK I guess.
[15:36] <kc2bez[m]> arraybolt3[m]: Lots of testing in a PPA seems prudent.
[15:36] <arraybolt3[m]> Not a bad idea. I usually just sbuild stuff, but a PPA will let us share our work easier.
[15:36] <kc2bez[m]> arraybolt3[m]: seems a little better than fusion
[15:37] <arraybolt3[m]> And it avoids the problem in Fusion where things end up getting stretched out in the Application Menu (that was determined to be a Qt bug IIRC.)
[15:38] <arraybolt3[m]> But it Just Works for me. Admittedly I just kludged KWin into my VM so what does and doesn't work in there may not be representative of what really does and does not work, but we can always make more VMs to test (and you probably already tested, thus why you mentioned clearlooks in the first place).
[15:39]  * arraybolt3[m] has to go afk, I shall return hopefully soon
[15:39] <kc2bez[m]> Take care, see you soon.
[15:40] <arraybolt3[m]> (One parting ambitious thought, what if we took clearlooks as a base and tried to make it look as much like Breeze as we could, then call it Aqua or something?)
[15:40] <kc2bez[m]> I'll be here for a bit but working on some other stuff.
[15:40] <kc2bez[m]> arraybolt3[m]: Seems like a cool stretch goal for next cycle
[18:05] <arraybolt3[m]> I think changing the theme to clearlooks is probably the best, safest, and easiest bet.
[18:05] <arraybolt3[m]> OK so I downloaded that new libfm-qt, compiled it, and then slammed it into the system with a few failed apt commands and a final "sudo dpkg -i --force-all <libfm-qt-package-filename>" (had a conflict with libfm-qt-common that didn't want to go away, this was all in a VM, don't worry). The system still worked after that stunt, but it didn't fix the Breeze problem.
[18:06] <arraybolt3[m]> I should try some more no-change rebuilds (maybe even for the whole entire lxqt stack), but I don't think it's likely to help.
[18:07] <kc2bez[m]> It seems unlikely, yes.
[18:07] <arraybolt3[m]> And I could keep trying to bisect Qt, but since this only causes one problem in one edge case that we happen to hit, I don't know if it's actually Qt's fault or if it's just a buggy edge case with Breeze like is known to happen with LXQt.
[18:07] <arraybolt3[m]> Just because Qt may be showing it up now doesn't necessarily mean it's their fault.
[18:08] <kc2bez[m]> It also seems like Debian removed the symbols from libfm-qt again Simon Quigley 
[18:08] <arraybolt3[m]> Dan Simmons: Oh yes they definitely did. I kept hitting roadblocks trying to fight with libfm-qt so I never did manage to get that change in, and Andrew was... ***very*** reluctant to let it happen.
[18:09] <kc2bez[m]> swell
[18:09] <arraybolt3[m]> I even have part of the progress of making it work right on my computer right now, but between the cycle-end insanity and the whirlwind of personal life, it's been almost impossible to get to it.
[18:10] <arraybolt3[m]> (Oh and they also un-split libfm-qt and libfm-qt-common, thus why I ended up having to use a sledgehammer to install the library a bit ago.)
[18:10] <kc2bez[m]> Don't beat yourself up
[18:10] <kc2bez[m]> That second part is problematic
[18:11] <kc2bez[m]> It means it probably won't upgrade cleanly.
[18:11] <arraybolt3[m]> As long as nothing depends on libfm-qt-common except for libfm-qt, won't it just work?
[18:11] <arraybolt3[m]> I mean, I think (could be wrong) that everything else depends on libfm-qt and then it depends on libfm-qt-common, so a merge should just be absorbed like nothing unless the archive works in some way I've not learned about yet.
[18:12] <arraybolt3[m]> (What's that command for searching for reverse depends again?)
[18:12] <arraybolt3[m]> (Actually, lemme just use Muon to do it...)
[18:12] <kc2bez[m]> the breaks/replaces might be wrong, not sure
[18:12] <kc2bez[m]> Gah
[18:12] <arraybolt3[m]> Well I have the packaging right here in my face, I'll give it a look-over and see if it looks problematic.
[18:13] <arraybolt3[m]> Though... you're right, it won't upgrade cleanly. I know, I just tried. :(
[18:13] <arraybolt3[m]> It would need to have a breaks/replaces against libfm-qt-common, which Simon Quigley split out for a reason, so either we'd need to adjust the breaks/replaces to do away with libfm-qt-common (bad idea?), or split libfm-qt-common out again ourselves.
[18:14] <arraybolt3[m]> Sigh. I guess push comes to shove eventually. We have the Breeze issue triaged, looks like, I guess next on the agenda will be to finally grapple with libfm-qt.
[18:14] <kc2bez[m]> arraybolt3[m]: It was a recommendation from Steve if I recall.
[18:16] <arraybolt3[m]> How long do we have before Final Freeze again? I get my new laptop tomorrow, and can resume work on this the day after, so if we have the time, it would be nice if I can do all of the libfm-qt work at that point on a nice, fast, new system rather than my nice-for-its-age-but-quite-outdated desktop.
[18:16] <arraybolt3[m]> Doing that would give us 2 days before Final Freeze...
[18:16] <arraybolt3[m]> And managing RISC-V symbols will be slow on anything. I guess I should probably start now.
[18:18] <arraybolt3[m]> (Maybe one of us should invest in one of those StarFive VisionFive RISC-V boards so we can manage symbols in a timely fashion.)
[18:20] <arraybolt3[m]> But for now, I'll get a RISC-V development environment going on my desktop so I can manage symbols properly.
[18:25] <arraybolt3[m]> Also is there a package already in existence that picks the default LXQt theme, or is that something we'll need to make and upload?
[18:28] <kc2bez[m]> arraybolt3[m]: That is probably default settings. I can try to find it.
[18:28] <kc2bez[m]> I think the priority should be the theming issue since libfm-qt isn't approved yet.
[18:29] <kc2bez[m]> I don't know who synced it.
[18:29] <kc2bez[m]> Haven't looked yet
[18:29] <arraybolt3[m]> Makes sense.
[18:32] <kc2bez[m]> https://github.com/lubuntu-team/lubuntu-default-settings/blob/ubuntu/kinetic/src/etc/xdg/xdg-Lubuntu/lxqt/lxqt.conf#L13
[18:33] <arraybolt3[m]> Pretty sure that's it, just doing a quick final check.
[18:34] <arraybolt3[m]> Perfect. Dan Simmons That looks like exactly the right file.
[18:34] <arraybolt3[m]> I'll package up a change real quick and beam it onto GitHub.
[18:36] <arraybolt3[m]> Heh, I've been calling it Clearlooks this whole time. It's cleanlooks.
[18:36] <kc2bez[m]> Forgive my typo earlier, I think the correct theme to move to is `Cleanlooks` Double check me on that.
[18:36] <kc2bez[m]> hah
[18:36] <arraybolt3[m]> ROFL we both said it at the same tie
[18:36] <kc2bez[m]> I think I did the same
[18:36] <arraybolt3[m]> s/tie/time/
[18:36] <kc2bez[m]> jinx, you owe me a coke ;)
[18:37]  * arraybolt3[m] attempts to teleport coke to Dan Simmons and accidentally spills it all over keyboard
[18:38]  * kc2bez[m] sends Simon Quigley  to clean it up.
[18:44] <arraybolt3[m]> Simon Quigley: https://github.com/lubuntu-team/lubuntu-default-settings/pull/2 (this works around the mess discussed above)
[18:44] -ubot93:#lubuntu-devel- Pull 2 in lubuntu-team/lubuntu-default-settings "Change default theme to work around Breeze theming bug" [Open]
[18:44]  * arraybolt3[m] has to go afk again, see you later, hopefully soon!
[18:45] <arraybolt3[m]> Simon Quigley: (WARNING with above PR: I didn't actually test that yet. It should just work, but don't blindly accept it.)
[18:45] <arraybolt3[m]> (I'll test it when I get back.)
[18:45] <kc2bez[m]> Testing might be good
[18:45]  * kc2bez[m] wonders if that "c" should be capital.
[18:46] <kc2bez[m]> Appreciate your help arraybolt3 
[19:01] <arraybolt3[m]> Dan Simmons: W.r.t. the "c", it's lowercase in the settings menu and in the settings dialog, so I just went with it. 🤷‍♂️
[19:02] <kc2bez[m]> Ok, sounds good.
[19:02] <arraybolt3[m]> (By settings menu I actually meant the settings file...)
[19:13] <arraybolt3[m]> Simon Quigley: Alright, I have now actually tested the merge request above, and it appears to work.
[19:15] <arraybolt3[m]> (Also, I kinda like cleanlooks. It's definitely different from Breeze, but it's nice.)
[19:16]  * arrayboltq using Quassel to test the bugfix, looks like everything works right.
[19:17] <arraybolt3[m]> arrayboltq: Works.
[19:17] <arraybolt3[m]> Nice, I think that's it.
[19:17] <kc2bez[m]> "It just works"™️
[19:19] <arraybolt3[m]> Great, I hope I've not just left an immortal arrayboltq user in the Matrix user list forever.