Untitled Linux Show 219 Transcript

Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.

Jonathan Bennett [00:00:00]:
Hey, this week we're talking about Ubuntu 2510. It's finally got a release date and it will be rusty. Then there's Pop OS, which is almost ready to have a beta of 24.04. Yes, 24. Kazada brings back the gaming and the cartridges. Arch Linux is still under a DDoS attack. Firefox has announced the end of 32 bit and a lot more. You don't want to miss it, so stay tuned.

Jeff Massie [00:00:29]:
Podcasts you love from people you trust.

Ken McDonald [00:00:33]:
This is Twit.

Jonathan Bennett [00:00:38]:
This is the Untitled Linux show, episode 219, recorded Saturday, September 6th. Bring your own pie. Hey folks, it is Saturday and you know what that means. It's time to get geeky about Linux and open source. We're probably going to talk about some hardware. Definitely software, all kinds of fun stuff. And it is of course not just me. I've got a couple of the guys with me.

Jonathan Bennett [00:01:03]:
We've got, let's see, there's their names, we've got Jeff and we've got Ken. And it's going to be fun. And somebody, just as we were starting the show, Keith S512 or Keith's 512. I'm not sure how he says it in the chat room here. Just says he watched the WAN show, Linus Tech tips the other Linus. And he was talking about doing Linux Y stuff with Linus Torvalds. And I just wanted to say I saw the headline. I don't actually watch the WAN show because it's like two hours long and I know ain't nobody got time for that longer or longer.

Jonathan Bennett [00:01:35]:
Yeah, they go for a while. I don't have a commute. I work out of my home. So, like, if I had a commute, I would do podcasts on the commute. But I don't. And so I just don't have time to sit down and listen to a podcast for like two hours. But anyway, I saw that he was talking about the other Linus and I wondered, I'm like, is he doing something with Linus Torvalds? Like, that would generally genuinely be very cool to see.

Jeff Massie [00:01:57]:
Oh, I would love to see that.

Ken McDonald [00:01:59]:
I know.

Jeff Massie [00:02:00]:
Back and check the WAN show. I don't normally catch it live. A lot of times I'll go back and kind of scan it. And what I do is it's nice on YouTube because they have the chapters in there so I can skip over, you know, oh, they spent 45 minutes, I don't know, talking about Apple or something. And I'm like, oh, okay, I don't care. And jump over that and just kind of hit some of the stuff that I really do care about and skip.

Ken McDonald [00:02:23]:
Over that hour and a half section about Windows.

Jonathan Bennett [00:02:27]:
Yeah. So there's Linus Sebastian, there's Linus Torvalds, there's another famous Linus. And I can't remember who it is in the tech world too. I think there's somebody I don't know because I've seen.

Ken McDonald [00:02:38]:
I know Linus and entertainment, the character.

Jonathan Bennett [00:02:41]:
Yeah, I don't think that's who it is because someone was talking about the sort of the holy trinity of Linuses and the idea of getting them all together at the same time. So the Internet says that that was the first. The first topic of their last WAN show. So we can all, after this, just go and watch it for the first five minutes. There'll be an influx of people that watch the first five minutes of that episode. And the people doing the analytics, what's up with it? Yeah, exactly. Jeff, you've got something new going on before we dive into stories.

Jeff Massie [00:03:13]:
Yeah. So I want to tell everybody I'm. If things look a little different, sound different or whatever, I'm trying this show on Cashy os.

Jonathan Bennett [00:03:21]:
Very cool.

Jeff Massie [00:03:22]:
And I've got very catchy. Yeah, the Restream going through Firefox. I had to shut Restream down and restart it or Firefox down and up and then it caught the feed, so hopefully it holds. Seems solid, but we're. We're giving it a try. We're going to put her through a real use case.

Jonathan Bennett [00:03:44]:
So we'll do it live. We'll do it live.

Ken McDonald [00:03:48]:
Or is it Memorex?

Jonathan Bennett [00:03:51]:
Goodness. All right, Ken, let's stop with the terrible puns and talk about Ubuntu.

Ken McDonald [00:03:57]:
If you're sure about that. But yes, this week we have a trio of articles related to the upcoming Ubuntu 25.10 release. Now we've got Saurav Rudra writing about Ubuntu 2510's release schedule, starting back on August 14th with the feature Freeze, and then ending with the final Release scheduled for October 9th. He then touches on Ubuntu 25.10, adding the experimental TPM backed full disk Encryption in its installer, shipping with Linux kernel 6.16, and providing GNOME 49 as a base Ubuntu desktop environment. The second article is by Michael Larabel, and he talks about Ubuntu 2510 using the Rust version of Coreutils, starting with a Release Pocket that came out this week. In the third article, Michael compared Ubuntu 25.10 performance to Ubuntu 25.04 on the AMD Ryzen AI Max Plus Strix Halo powered framework desktop. Let me get my breath back now. Now Michael does state this is a tentative look at performance on the Ubuntu 25.10 deli snapshot as of August 31st.

Ken McDonald [00:05:29]:
For most of the graphics benchmarks there wasn't too much movement, but Michael was hopeful Mesa 25.2 will land in time for for Ubuntu 25.10's final release to provide some additional uplift. In some areas there is some movement happening for Vulkan ray tracing with the AMD rDNA 3.5 graphics. The graphs show for most graphic workloads 25.10 improvement was flat in relation to 25.04. Linux kernel 6.16 does provide some nice improvements in some of the synthetic microbench marks. Now I've just lightly touched on the information in these articles that are linked in the show notes. I do especially recommend looking at the bitmarks benchmarks provided by Michael Jeff, are you looking forward to the changes coming in Ubuntu 25.10?

Jeff Massie [00:06:27]:
I am. You know and I looked, I looked at that performance benchmarking and the one thing is, you know it's kind of the, the typical everything got better. There wasn't a oh this got a lot better because of X. You know, it's like well the kernel improved and the file systems improved and other libraries programs, they improved. You know, it's kind of a uplift on a lot of different areas. One thing kind of a side note I am curious about is so you know we have talked about the new release in Debian Trixie. Well supposedly talking to now I run Kubuntu as my main, you know, try and try and cache I might switch but supposedly by having some changes and I don't, I don't know all the details yet in Trixie it's going to trickle down through Ubuntu and it's going to allow better advancement of packages because it's not going to be such a locked in ecosystem because a lot of what happens now is Kubuntu might say oh well we want to upgrade to the newest kde but they can't because of other things that are locked in with Ubuntu and supposedly there's changes in there. I'm trying to find out what they are from one of the main package maintainers to get that description.

Jeff Massie [00:07:55]:
So I can find out if that's going to help get Kubuntu more. More updatable rather than sometimes just being locked in because they can't update QT because of something in Ubuntu or something like that.

Jonathan Bennett [00:08:11]:
Yeah, makes sense.

Ken McDonald [00:08:13]:
Yeah.

Jonathan Bennett [00:08:15]:
I'm pretty interested to see what's going to happen with the Rust core utils. You know, it's been kind of a controversial thing which surprised me, but I think it's the right. It's the right time to do it in an Ubuntu. I don't know that you'd call it an experimental release, but not one of their LTSs. And yeah, I'm very intrigued to see if there's any real problems from it or if people were just telling us the sky was falling because it's a change.

Jeff Massie [00:08:44]:
Yeah. And the 10 releases are where you just throw it against the wall and see what sticks because you know it's a short lived release. So okay, let's try it, see what happens.

Ken McDonald [00:08:53]:
Hopefully appears in 2604 but if not.

Jonathan Bennett [00:08:57]:
Six months later you can pull it right back out and be done with it. Well, that was a bad idea. We weren't ready for it yet.

Ken McDonald [00:09:03]:
Right now will that mean having to do a fresh install of 2604?

Jonathan Bennett [00:09:09]:
Shouldn't you should be able to do a package replace? They've got the ability to do that.

Jeff Massie [00:09:14]:
Well, I will tell you though, I've. One of the things I'm looking at is I've had problems with this before. Is even just a normal say, say there's nothing crazy after you upgrade several times there's kind of legacy cruft that gets left behind and certain things still support some of those legacy settings that might be there. Even though they say, well this is deprecated. We're not. You're not supposed to use this anymore. Well, because you had a couple versions ago when it was supported. It still picks it up because I've had, you know, things on my desktop that just like wow, this just doesn't.

Jeff Massie [00:09:52]:
They say there's these new features but I'm not getting them. I wipe out the config directories and some things force it to reload itself and suddenly then it's like whoa, this looks totally different. This is. This is a night and day change. Things act differently. So it was holding on to some legacy cruft in there.

Jonathan Bennett [00:10:12]:
Yeah, you've got config. I know something that'll happen in Fedora as well is you'll have really old packages that just never. If you're doing a, you know, you did your install in Fedora 12 let's say, and then you just keep upgrading it. Well, you may have some packages from Fedora 12 still installed because they got abandoned or whatever, but just never broke all the way and so they're still hanging around and installed. I do that every once in a while and I get surprised every time how the amount of old packages I have.

Jeff Massie [00:10:41]:
Every few releases I try to just wipe and start over just to clean some of that out, clean some of the cruft. And I'm even looking now at just wiping out all my hidden directories and like in home, just wipe out all my KDEs and all various settings and just kind of back up everything that I really care about and just trash everything else and kind of start over just to you know, have it. Have a clean slate because. Because you never know how what affects some of that old stuff it's got on your system.

Jonathan Bennett [00:11:14]:
Right.

Ken McDonald [00:11:14]:
Or how doing a pre install or something. Because I've installed the Rust Core Utils package on. I think I've got it on this Ubuntu system.

Jeff Massie [00:11:30]:
Yeah, that's the other thing. Beta on top of beta and Alpha you got directories and directories of stuff.

Ken McDonald [00:11:36]:
You know it in fact looking for something else and I came across something. What was it? Fl something to do with ah, I can't remember at the moment.

Jeff Massie [00:11:54]:
FL Something 512 said dist upgrade seems rather smart. It seems like it gets most stuff unlike Windows and it does. This is. I'm talking. I might have had eight upgrades in a row without a wipe. And that's where stuff, you know, you can go through several before it starts. You kind of just. I better wipe and just start over.

Jonathan Bennett [00:12:20]:
Yeah. So I've got on my Fedora desktop I've got some Fedora Core 33 packages and I'm running Fedora 42 on it now. So yeah, those are old. Those have been around there now. What is it? Okay, it's a font. The Ming package. I'm not sure what that is. Riserfsutils.

Jonathan Bennett [00:12:42]:
I wonder why that hasn't been updated. A lot of these are fonts and things like that. There's some old WI FI drivers in there. All kinds of. All kinds of old stuff still hanging around from the fedora core 30 days. But yeah, it's worth looking at, you know. And Fedora I was actually really surprised just now. I went to do a DNF list installed and that doesn't work.

Jonathan Bennett [00:13:08]:
And I don't know why I will have to look into this, because I know that has worked in the past. But right now, if I do a DNF list installed, it says no matching packages to list, no repositories loaded. That's a little scary. But using RPM qa. So query all, it gives me the list of everything that's installed. I got to go find out whether I broke my package manager now.

Jeff Massie [00:13:33]:
Well, and maybe we need to cover this again. We have in the past, but whenever you install, if you have a different partition for your home directory and things like that, you can do a reinstall of Linux in no time flat. I mean, you can probably. 10 minutes, you can have a new operating system reinstalled. And because you didn't kill your home directory, you know, you, you launch your Firefox, it's got the wallpaper and the bookmarks and everything is there. You fired up your wallpapers there. So it's not like you're starting over from square one again.

Ken McDonald [00:14:10]:
Well, what I do is I've got the actual home directory on the same partition as my system, but then I bind the actual data directories to the folders in that home directory to make it easier.

Jonathan Bennett [00:14:28]:
Yeah, you can do that too. So in DNF5 land, by the way, for those of you on Fedora that wonder, the same thing that I do, it's DNF list installed to get that list that you were looking for. That is a change in DNF5. So.

Ken McDonald [00:14:45]:
And do you see a lot of garbage with that?

Jonathan Bennett [00:14:48]:
I mean, there's not much garbage. Let's see, I could do the same thing. I can pipe it through grep and then just grep for FC3 and that's going to catch all of those Fedora core.

Ken McDonald [00:15:03]:
I heard somebody was seeing some garbage somewhere.

Jonathan Bennett [00:15:08]:
Okay, I get it now, Jeff. That's a segue for you.

Jeff Massie [00:15:12]:
Yes, it is.

Jonathan Bennett [00:15:14]:
Okay.

Jeff Massie [00:15:15]:
I'm not really sure it would be a week in Linux land if we didn't have an opinion that Linus Torvalds voiced in public this week. It has to do with garbage links. Now these are links and get comments for the Linux kernel which point to the Linux kernel mailing list, but the same patch. So you're looking at a poll request, say in the comments. It has a link to the mailing list of the code that you're looking at right at that time. So Linus said, damn it. This commit, that promit. This commit has that promising link argument that I hoped would explain why this pointlet pointless commit exists.

Jeff Massie [00:16:05]:
But as always in all caps, that link only wasted my time by pointing to the same damn information that was already there. I was hoping that it would point to some oops report or something that would explain why my initial reaction was wrong. Stop this garbage already. Stop adding pointless link arguments and waste people's time. Add the link if it has star all caps additional information. Damn it. I really hate these pointless links. I love seeing useful links but 99% of the of the links I actually see just point to stupid useless garbage and it only again in all caps wastes my time again in all caps.

Jeff Massie [00:16:56]:
So I have not pulled this. I'm annoyed by having to even look at this. And if you actually expect me to pull this, I want a real explanation and not a useless link. Yes, I'm grumpy. I feel like my main job, really my only job is to try to make sense of pull requests and that's why I absolutely detest these things that are automatically added and only make my job harder. Now in all fairness, you know, Linus did come back and comment later and say that he's in favor of the links when they add additional information, e.g. the COVID letter of a multi commit series, then the link to the patch series submission is potentially more useful and likely much less annoying because it would it would go into the merge message, not individual commits. Linus says that the perfect model would be to have some automation to verify the links, have some, you know, verify that they have some useful information.

Jeff Massie [00:17:52]:
But he does say that model would be complex and probably need AI. And unless somebody, you know, really wants to use AI for something and needs a project, it's not going to happen. And the warning is going to have to be enough for now. But Linus also says, I realize that people think that patch submissions will get more email replies at some hypothetical later date, but in practice that seldom happens because the downstream testing issues typically create new threads, not replies to the original emails. And if they do react to the original email, we already can look up the commit easily and the lookup goes the other way anyway. So basically, bottom line, if any of our listeners do put patches in for the Linux kernel, make sure in the comment if you do put a link, there's a use for it and it gives more information on your actual commit, not just pointing to the mailing list part of the code that you're currently looking at.

Jonathan Bennett [00:18:53]:
Happy Patching Wizard Wizardling says It's like when I see a changelog and open it to only see new bug fixes and features to improve the user experience. Thanks. Thanks for that. Yeah. Something that really fascinated me about this is that Linux points out, and it's absolutely accurate, his job is reviewing pull requests. That's like all that he does. That's what he get paid for.

Jeff Massie [00:19:21]:
Yep.

Jonathan Bennett [00:19:23]:
I understand.

Jeff Massie [00:19:24]:
Very, very true.

Jonathan Bennett [00:19:25]:
I understand not wanting to waste the time on something trivial.

Ken McDonald [00:19:30]:
And then banging your head.

Jonathan Bennett [00:19:31]:
Yeah, and then banging your head over it. I wonder if he has some process in place before he clicks on a link in an email. He's just thinking about it, like, so he has to be aware that there are a hundred different countries out there that would love nothing more than to have spyware on his machine. Right. So does he go through some process where he takes the link and goes and types it in on another machine? Is that why this annoys him so much to have to go to a link?

Jeff Massie [00:20:01]:
Well, it's probably just copy and paste.

Ken McDonald [00:20:03]:
It into a vm possibly.

Jonathan Bennett [00:20:05]:
Yeah.

Jeff Massie [00:20:06]:
I think it's more just a sheer volume of. Why are you doing this? Here's some code. What's it doing? Or what's the thought behind it? Oh, here's a link. Oh, it's pointing to the same code. Yeah, well, because, I mean, the kernel does have a lot of the. Was it the circle of trust or whatever? So you can't just put in a kernel patch. You got to build trust and get people to.

Jonathan Bennett [00:20:35]:
Well, you go through one of the maintainers and then they add your kernel patch to their pull request. Yeah, that's the way that works. All right.

Ken McDonald [00:20:43]:
And then some. They move that up to somebody else higher up who they may add it to their pull.

Jonathan Bennett [00:20:49]:
Yep. Sometimes that's how that works.

Jeff Massie [00:20:51]:
Yeah.

Jonathan Bennett [00:20:52]:
All right. Well, there is another thing in the news that we've been talking about for a long time, and it looks like it's finally. It's coming. You know, insert meme here.

Ken McDonald [00:21:03]:
It's happening.

Jonathan Bennett [00:21:06]:
Popos and Cosmic are going to go beta in September. So, like, any day now this month. And so someone, a random commenter on x.com the social media network formerly known as Twitter, says, oh, can we expect the release by November? In that case, Carl Ritchell, the man behind popos and Cosmic, says, early to tell, but within the realm of possibility. So we may, we may get a popos and Cosmic release, like, final full stable release before the end of the year. That'd be cool. That'd be really cool. I'd like to see that. Now it is POP OS 24.04.

Jonathan Bennett [00:21:57]:
I don't know if you've noticed, but it is 25. It is 2025 out there. If you open your windows and look out, it's 2025 now. They're a little behind, a little behind the curve. Although I guess they stick with the LTS. So it'll be 26. Right. So it's 2604.

Jonathan Bennett [00:22:15]:
It's going to be their next new one after this.

Jeff Massie [00:22:17]:
Well, and it has been commented on historically that they're putting more energy into the actual desktop than keeping the os.

Jonathan Bennett [00:22:26]:
That was, that was the point I was kind of, I was kind of going to get at is. Yeah. They've slowed down their POP OS release release cycle by doing cosmic. It seems like it was maybe a little more work than anticipated. Maybe not. Maybe they knew what they were getting into.

Jeff Massie [00:22:43]:
You know, I'll believe it.

Ken McDonald [00:22:44]:
What did they really need to do? Much to the POP OS itself other than security updates.

Jonathan Bennett [00:22:53]:
Oh, you mean the existing one? It's. Yeah, no, not really. There's not been a whole lot happening there. Security updates and then, you know, doing whatever needed to be backported to make it continue to work. Yeah. So it's not like they put a lot of. They've put a lot of effort into the. What is it? Are we running 22.04? Am I running 22.04 on my laptop?

Jeff Massie [00:23:15]:
Very possible.

Ken McDonald [00:23:16]:
I know somebody that's still running 20.04 and getting security updates for it.

Jonathan Bennett [00:23:24]:
Well, yeah, yeah. Ubuntu or popos?

Ken McDonald [00:23:27]:
Ubuntu.

Jonathan Bennett [00:23:28]:
Yeah. I don't know. The popos would get the long term support updates from Ubuntu.

Jeff Massie [00:23:33]:
So anyway, I, I see this going one of two ways. Either it can happen before the end of the year or they're wildly optimistic. And I haven't used cosmic. I think it's going to depend on. I, I have heard reports that their alphas are pretty stable. They're almost. You could call them beta instead of alphas.

Jonathan Bennett [00:23:54]:
Yeah, that's. That's kind of my. I think.

Jeff Massie [00:23:57]:
Yeah. Kind of in the realm of possibility. If they're truly alphas and they're going to go to beta, I bet it's maybe next summer, but I mean, I'm looking forward to it.

Ken McDonald [00:24:08]:
Spring.

Jonathan Bennett [00:24:09]:
Well, so they have been very conservative about releasing betas. Right. About releasing releases on this, you know, calling anything even an alpha. They took a while and they've done a bunch of alphas. So I, I would not be terribly surprised to see. I don't know Maybe they'll call it a release candidate, but something beyond just a beta by the end of the year because they've, they've done a lot of work, but they've been very intentional to try to do it in a way that keeps everything stable. And everybody I've heard from that's run it has been fairly impressed with it.

Ken McDonald [00:24:46]:
And yes, Wizard Lean, I think you really would should be interested in checking out Cosmic, because I'm going to check it out when it finally comes out with the version one.

Jonathan Bennett [00:25:01]:
Yeah, it should be interesting to take a look at. One of the nice things about this is basically on all of your distros, you can install different desktop environments side by side and then just pick the one that you want to go into when you go to log in as your user. And so, for example, on Fedora, I run KDE full time. I can go in there and say, boot into. Install the Cosmic packages and then say, boot me into Cosmic. Take a look at it, play with it. If I don't like it, log back out, log back in as KDE and get back to work.

Jeff Massie [00:25:33]:
Yeah, the original question was, should I be interested in checking out Cosmic when I'm already pretty happy with kde And I think some of that is when they have a release, there's going to be a ton of reviews on it. One of the things too would just even be check the reviews to see how it looks and if it's even something you want to investigate. If you do, then give it a shot, load it on and do what Jonathan says. Just switch between them and see if it's like, oh, this is even better, or I like what I got. And that's a wonderful thing. You just pick the one that fits.

Jonathan Bennett [00:26:04]:
Yeah.

Ken McDonald [00:26:05]:
Jonathan, did you hear about the zero Day exploit for Cosmic?

Jonathan Bennett [00:26:10]:
I did not. I was just looking at that comment. This is a comment from Mr. J. Nice. And a zero day for cosmic. I have not heard about that, but I have to wonder, like, is this where someone had to have so someone has to run code on your machine? It's not a remote code exploit. And so this probably sort of falls into one of these categories where, okay, that's an interesting exploit.

Jonathan Bennett [00:26:39]:
But once someone's running code on your machine, there are 15,000 other more malicious things you can do rather than just, oh, no, not the Cosmic Browser. Not the Cosmic Browser, the Cosmic Desktop. So that zero day is probably not for the Cosmic Desktop, which makes more sense.

Jeff Massie [00:26:59]:
I was just going to say I heard a browser that I couldn't think of. The name of the browser that you know. Oh, yeah, it didn't check anything. It just went out. But it was like, okay, yeah, this is desktop. And I was thinking too, well, it's an alpha.

Ken McDonald [00:27:13]:
We found security.

Jeff Massie [00:27:14]:
Well, it's an alpha, but okay, that clears up a lot.

Ken McDonald [00:27:20]:
Yes, that sounds like Internet Explorer version six.

Jonathan Bennett [00:27:27]:
Yeah, no joke. All right, so there's new things in the World of the Gimp. GIMP 3.14. 314. Is this Pygimp?

Ken McDonald [00:27:41]:
Actually, we're talking about GNU's image manipulation program. In fact, this week Michael Larabel and Marius Nestor wrote about the GNU image manipulation program team releasing GIMP 3.1.4. Or should that be a soft G? Now they both touch on the implementations of two main GIMP GIMP 3.2 roadmap items, link Layers and Vector Layers. Now, Link Layers allow linking external image files as a layer in your project. Vector Layers allow creating a shape and setting its fill and stroke properties. Now, according to Michael, Vector Layers work originates all the way back in around 2006 when it began as a Google Summer of Code project. So we did get some good things out of Google after all. According to Marius, GIMP 3.1.4 also introduces the Generic Graphics Library, or Giggle Filter Browser 2 for viewing a list of all Giggle operations, I.e.

Ken McDonald [00:28:57]:
filters and information related to their use. It gained improvements to the text tool, live previews for the outline, color, support for reporting signed JPEG 2000 images and HRZ images. Is that hers images? I do want to remind you, GIMP 3.1.4 is a development release and is not ready for production use. As always, you can find more details in Michael and Marius articles.

Jonathan Bennett [00:29:33]:
I am. I am so disappointed that they did not make it a pie release.

Jeff Massie [00:29:39]:
I know, it's just right there. Opportunity missed.

Jonathan Bennett [00:29:43]:
Yeah, they even you go to the release announcement and they have a picture of a cake. It's like they did it on purpose.

Ken McDonald [00:29:52]:
But can you run it on a pie?

Jonathan Bennett [00:29:55]:
Sure. Yeah. You can do an ARM build again? Yeah, absolutely.

Ken McDonald [00:29:59]:
Then do that. I guess that's how you get your pie.

Jeff Massie [00:30:03]:
That full render is going to take a while.

Jonathan Bennett [00:30:05]:
Bring your own pie. Gimp 3.14. Bring your own pie. See, it's a tagline. They could have gone for it.

Ken McDonald [00:30:12]:
Yeah, and that could be a good show title.

Jonathan Bennett [00:30:14]:
Yeah, I was thinking about that. I was thinking about that. All right, Jeff, let's talk about reproducible builds. Yeah, so I'm interested in reproducible builds. I would love to have reproducible builds on my machine.

Jeff Massie [00:30:32]:
A lot of people would. So originally scheduled for Fedora 43 was a reproducible package build goal. But this has now been moved to Fedora 44, which will be released next year. The original goal was to have no fewer than 99% of the RPM package builds be reproducible. Bottom line, there's still some lingering issues which are causing the goal to be moved out. Some. Some like non reproducible debug data secure boot signing using a private key and other package nuances. Now this might be obvious to some, but when building a package, you won't always get bit for bit the same results.

Jeff Massie [00:31:14]:
The program or package might run and behave the same, you know, behaving the same way, but it won't be identical. And when there's a difference, then troubleshooting can become an issue. So what works in one system might not work on another because of small differences. So that's why it works great here on identical hardware. Somebody else's I don't have an issue. Or they do have an issue. So that's where this comes in. Now here's what the Fedora project officially defines as reproducible.

Jeff Massie [00:31:49]:
They say when a completely independent rebuild of a koji package produces RPMs that are identical except for the build timestamps, signatures and some associated metadata, the payload that is packaged files and important metadata are bit for bit identical. So okay, that's great. Well, let's get into the details of why are they doing this? Adding determinism enables the Fedora community to have confidence that if given the same source code, build environment, build instructions and metadata from the build artifacts, any party can recreate copies of the artifacts that are identical except for the signatures and some parts of the metadata. I. E. Again, like your timestamp that that's of course going to be different, but that's okay. Reproduce reproducibility of builds leads to packages of higher quality. It turns out that quite often those irreproducible bits are caused by an error or sloppiness in the code.

Jeff Massie [00:32:51]:
In particular, any dependence on architecture in no arch packages is almost always unwanted and or a bug test builds that check reproducibility will expose such instances. Reproducibility of builds makes it easier to develop packages. When a small change is made and a package is rebuilt in the same environment, then with a reproducible package, the only difference is directly caused by the change. If the package is different every time it is rebuilt, it makes the comparison much harder. And build reproducibility is a topic that's gaining in popularity. Major distributions like Debian Arch, opensusi, Nixos are trying to achieve full reproducibility by making build reproducibility and expectation. And for Fedora we avoid driving away people who consider build reproducibility requirement. We may even attract additional contributors, contributors who are interested in this topic.

Jeff Massie [00:33:50]:
If we achieve better results than other distros with 90% reproducibility, we're on par. So that means like the, the Debians and Arch and all those are going for 90 with 99% we can be the leader. Now they Fedora did put a smiley face after that in their wiki, but the goal is clear and we know where they're headed. So 99% by 44 in 2026. So I'll, I'll. We'll let everyone know when 44 gets closer to release, if they've achieved their goal, you know. But take a look at the link in the show notes for more details and further links to even more information. There's several paths or links you can follow to getting more information on it.

Jeff Massie [00:34:31]:
And it includes links to the Fedora wiki.

Jonathan Bennett [00:34:35]:
Yeah, I think there's an important reason for reproducible builds that you didn't exactly hit on. And that's security.

Jeff Massie [00:34:41]:
Right.

Jonathan Bennett [00:34:41]:
Like, so you want to make sure that the build that is pushed out has not been tampered with anywhere along the way. And this is one of the ways to be able to do that is you can redo the build in a controlled environment and make sure that the essentially the hashes match, that you've got the same thing.

Ken McDonald [00:34:58]:
So on that Note, Debian and OpenSUSE are more secure than Fedora.

Jonathan Bennett [00:35:04]:
Do they have reproducible builds?

Jeff Massie [00:35:07]:
They're going for 90. Well, so is Fedora. Fedora is just trying for a higher ratio than the other ones are.

Jonathan Bennett [00:35:16]:
Right.

Jeff Massie [00:35:18]:
Everybody else is trying for 90%. Fedora is going for 99%. So by 44 you could argue that Fedora would be more secure, but that's.

Jonathan Bennett [00:35:29]:
It's only really one. Yeah, it is only one metric. It's only one. It's only one way to harden and there are so many different hardening steps.

Ken McDonald [00:35:38]:
That you could do, like making it immutable.

Jonathan Bennett [00:35:42]:
That can certainly help.

Jeff Massie [00:35:43]:
Yeah, it's kind of one word out of the dictionary. It's just one little piece of the puzzle.

Jonathan Bennett [00:35:48]:
Yes, absolutely.

Ken McDonald [00:35:48]:
And getting rid of 32 bit applications.

Jonathan Bennett [00:35:52]:
You know, that is something that is happening on a lot of different places and there are some security ramifications for that. So the story we've got here is that Firefox is ending 32 bit. But we can talk for a little bit about some of the reasons why distros are dropping the 64 bit. Excuse me, dropping the 32 bit support as well. There are some security hardening things that landed in the processors in 64 bit. And so like it was true for the longest time that running 64 bit windows was a lot more secure than 32 bit. And I believe some of that is the same in Linux as well, where you have these different processor features, like being able to mark RAM is non executable. I think that's a 64 bit feature only.

Jonathan Bennett [00:36:38]:
So the story, again, the story that we have here is that Firefox is not going to ship the 32 bit browser after version 1.4.4. And they're saying it's time for users of Firefox to switch to 64 bit systems. So that's September of 2026. You've got a full year to upgrade your old, old 32 bit systems. But starting with Firefox 145 coming out in 2026, it will be 64 bits only. And so yes, there's a security component to this. It's also just true that with fewer people using 32 bit and the continuing sort of lack of distro support, lack of testing on it, it's becoming more and more of a maintenance burden and Firefox is sort of having to batten down the hatches and trim up, tighten the belt up to be able to try to stay afloat. We did get some news this week that is good for Mozilla and good for Firefox.

Jonathan Bennett [00:37:47]:
And we could talk about that here in a minute too. But yeah, if you're still running 32 bit Firefox or even just a 32 bit Linux, it's time to move on to 64 bit. Come taste the 64 bit goodness.

Ken McDonald [00:38:03]:
So when will we see 120, 128 bit architectures?

Jonathan Bennett [00:38:09]:
Probably not going to happen because we already have 512 bit architectures actually in the ultra wide. Yeah, they just GPUs and well in CPUs there are the AVX 512 extensions. Right. You could make an argument that we are actually running 512 bit CPUs now that would be arbitrary. But since when has the labeling of CPUs not been arbitrary and sort of silly?

Jeff Massie [00:38:36]:
True, but I mean true back when.

Ken McDonald [00:38:38]:
They were all eight bit.

Jonathan Bennett [00:38:39]:
No, it was still arbitrary and silly back then. Let's be real.

Jeff Massie [00:38:42]:
Yeah, true, true. 5.12 though. I mean you like graphics cards and some of those kind of things which are CPUs basically just specialized and they're running the wide, wide paths. It, you know, 32 to 64 was a big jump. So you can support a lot more memory, things like that. Going beyond that, there's not really a big advantage unless you're moving a ton of data. And there's still ways kind of around it now that you know, systems support, I don't know exabytes of RAM and I don't know what kind of ites they got for the drives, but it's big, you know, and so I don't expect to sick we're going to stay on 64, but Linux, you know, and a lot of those, the software companies want to start really flushing out 32 bit because there's a lot of cruft in there. And the hardware companies, we've talked about this before, intel, amd, a lot of them, they want to get rid of the 32 bit because things aren't as standardized as they are in 64.

Jeff Massie [00:39:50]:
64 things were a lot cleaner and more organized. But 32, especially 16 saw a whole lot of clujes and stuff, you know, taped and bailing wired together. 32 still had some of that and 64 is running a lot cleaner. So they want to, if they clean that up, simplifies the hardware, simplifies the software thing, things are more efficient. They don't need so many transistors, they can make them more power efficient, less attack surface for coders because you don't have to support so many different architectures. And oh, this is 32, quick drop to this other code and it just cleans everything up.

Jonathan Bennett [00:40:34]:
Yeah, I want to touch base again on that idea of 128bit processor. So when you talk about the bit width of a processor, there's sort of two different things that you could be talking about. Like one is your memory width of your internal registers, how wide those are. And so like we said, modern systems actually have some 512 bit registers. Right. They can do these AVX512. So from that perspective one could refer to them as a 512bit CPU. But the other thing that that bit width usually refers to is the address bus, like how wide is the memory address bus.

Jonathan Bennett [00:41:13]:
And that determines how much memory you can address on the system. With 32 bit systems we were limited to 4 gigabytes. With a 64 bit system we are limited to 16 exabytes of memory. And I had to look this up to even wrap my head around how big an exabyte was. So we went from four gigabytes, the next one. 1,000 gigabytes is a terabyte, 1,000 terabytes is a petabyte, and 1,000 petabytes is an exabyte. So we are a long, long ways away from running out of memory address space to need more than 16 exabytes of addressable bits.

Jeff Massie [00:42:00]:
I'm just proud I got my. It's correct.

Jonathan Bennett [00:42:02]:
Ah, you know it off the top of your head.

Jeff Massie [00:42:05]:
Yeah, I was guessing because I'm like, I'm pretty sure it's bigger than petabytes because I knew giga terra PETA. And I'm like, I think the next one's exa. And then after that I don't know.

Jonathan Bennett [00:42:20]:
Yeah.

Ken McDonald [00:42:21]:
So you don't think this pursuit of AI isn't going to exceed that 64 bit addressing?

Jonathan Bennett [00:42:28]:
No, not anytime soon. I mean, how much, how much memory is in the biggest supercomputer? I don't.

Ken McDonald [00:42:39]:
Do you have enough memory to hold all the data that's been scraped in that address?

Jonathan Bennett [00:42:45]:
Well, you don't, you don't have to hold all of it at once is the thing. And that's what, that's sort of the way modern AI works. One of the breakthroughs is that they figured out how to effectively distill those huge, huge amounts of data into tokens that can fit in a much smaller amount of space. El Capitan apparently is top, the top system amd 24 cores HPE. Come on, just tell me how much memory it has. It's got 8 million cores, I'll find it.

Jeff Massie [00:43:36]:
Yeah, well, the thing too is if you said, okay, we're going to put in 16 exabytes of RAM, now I'm just totally pulling this number out of thin air. I can't say for sure, but I mean, that's probably the entire output of a FAB for like a week or a couple weeks, maybe a month. I mean just the sheer volume of chips you'd have to have would just be enormous.

Jonathan Bennett [00:44:02]:
Yeah, okay.

Ken McDonald [00:44:02]:
To go with a base space based FAB and then put the data center in space too.

Jonathan Bennett [00:44:08]:
I am very skeptical of that idea. I will just say. But I will not be surprised if build it in space becomes the next buzzword and bubble. But I am still skeptical that the math ever works out on that. Okay, back to this. El Capitan has 5.4 petabytes of high bandwidth memory and that is petabytes the maximum in 64 bit addressing is exabytes. So you could build a machine a thousand times bigger than the number one top supercomputer out there right now at El Capitan and it. You would still be able to double it before you ran out of address space.

Jeff Massie [00:44:58]:
Yeah, it's just. You would literally have semi after semi of just modules to have that.

Jonathan Bennett [00:45:08]:
Yeah, it's a lot. It is very big.

Ken McDonald [00:45:13]:
We're waiting for the, the fabs to get to where they can do what's next one below nanometers.

Jeff Massie [00:45:26]:
You're not talking about the.

Jonathan Bennett [00:45:28]:
The Angstrom, isn't it?

Jeff Massie [00:45:30]:
It's the angstrom, yeah.

Jonathan Bennett [00:45:32]:
That's. That's their new thing. So they can make their fab revs seem like a bigger deal than they are.

Jeff Massie [00:45:38]:
Nano is to the minus 9th, Angstrom is the minus 10th. But you start getting to the point where you're running into molecular sizes. So you're using molecules that are a certain size. You can only get so small. You can't. It's not infinitely scalable because you're like, well, my copper atom has a size of X, I can't shrink it and I need X number of them to make a circuit. You can't just string one and then another to have reliable. So you're going to have to have a few layers there and you just run out of size.

Jeff Massie [00:46:17]:
Because originally they talked about Moore's Law running out roughly around 20, I think it was about 2020 because that was where your shrinkage. Now you got to start doing things like you're going. That's why you see things from like two dimensions going into three dimensions because they're shrinking the size but they've got to stack it higher because you just run into physics limits.

Ken McDonald [00:46:46]:
And then physics steps in by transferring that electrical power to heat power, making it melt.

Jeff Massie [00:46:55]:
One of the problems, that's something too. But I mean. And then you're dealing with lower and lower voltages because you're trying to avoid that heat and that energy. That's why, you know, the old TTL logic, you're like, oh, 0 to 5 volts now. It's like, oh, you're sub. Sub volts just to make it all work and go fast enough. And, and then you get slew rates coming in with that. And it, you know, there's a lot of physics.

Jonathan Bennett [00:47:22]:
That complicated electrical engineering stuff.

Jeff Massie [00:47:25]:
Yes.

Ken McDonald [00:47:26]:
After all this complicated electrical engineering stuff, I think we need to start playing some games.

Jonathan Bennett [00:47:33]:
Take it away, Ken, tell us about gaming.

Ken McDonald [00:47:37]:
I think Jeff wanted to say something first.

Jeff Massie [00:47:39]:
Oh, I was going to Say maybe somebody in our audience real quick, if you look up an exabyte, could you tell me how many 64 gigabit memory modules? That would be a lot. Somebody out there can do the math real quick. I bet.

Jonathan Bennett [00:47:56]:
Do the math what Ken tells us about gaming.

Jeff Massie [00:47:58]:
Yes.

Ken McDonald [00:48:00]:
Well Jonathan, this week Mark Tyson and Liam Dahl wrote about a new Linux distribution arriving to provide a more more of a classic gaming console like experience. Now we've got Alicia Slovak, the Shimmera OS developer, announcing that he is developing Kazita. I hope I'm saying that right for non technical people who get lost in the chimera OS and SteamOS menu interface structure, a disenchantment with digital storefronts, a fondness of collecting old physical games. I can relate to that and a feeling that preserving a digital game collection this way was a good idea. He promises a gaming experience where you simply insert cartridge power on play. Now you will install on a system running an AMD Radeon RX 400 series or newer GPU with at least 4 gigabytes of RAM and 32 gigabytes or larger internal storage device. Kazita officially supports the 8bitdo Ultimate 2C wireless controller using the dongle or a wire. It officially supports the Geekom A5 2025 Edition Mini PC Speed Class of C10 or U1.

Ken McDonald [00:49:29]:
SD cards are the preferred media for creating cartridges because they look and feel like game cartridges are almost are mostly uniform in design and have a flat surface to which labels can be applied. These cartridges can contain any DRM free games that Kazita will detect. You simply insert the cartridge and turn on the system. When you finish playing, just press the power button to turn it off. There is even a sample cartridge tar file containing Celeste 64. Kazita is an immutable operating system where writing data is only possible in the ver directory. Kazita's GitHub contains information about OS updates, the cartridge system and runtimes that are used for the games. If you want more details then start by reading Mark and Liam's article, which have links to the official website and then to the GitHub.

Ken McDonald [00:50:33]:
And that was a rabbit hole. I went down with that looking into all that.

Jonathan Bennett [00:50:37]:
Yeah, that's an interesting little gizmo. I like the sort of retro take on making your SD cards. Putting a single game on an SD card and swapping them out to play.

Ken McDonald [00:50:47]:
Something different and you don't save the game data on them, you just it saves it to sub file under the slash variable.

Jonathan Bennett [00:51:00]:
Yeah, I'm Trying to figure out which would be the more realistic retro experience. Because, you know, with your old cartridge games, you would totally save your saved data to the game itself. So maybe the data should get saved onto the SD card.

Ken McDonald [00:51:15]:
No, GitHub says that the GitHub is. The SD card is treated as a read only device.

Jonathan Bennett [00:51:24]:
Yeah, that is probably safer now that I think about it.

Jeff Massie [00:51:27]:
It's only realistic if you have to blow on it every once in a while to make them work.

Jonathan Bennett [00:51:32]:
I mean, you can blow on them if you want to. It'll probably help as much as it actually did with the old systems, especially.

Ken McDonald [00:51:40]:
If you leave them sitting out in the open where dust can collect on them.

Jonathan Bennett [00:51:43]:
Yeah, don't do that. That's fun though. I like that.

Ken McDonald [00:51:50]:
The only drawback to this is you wouldn't be able to check it out in a virtual machine.

Jonathan Bennett [00:51:57]:
You probably. You probably could. Sure. You could. You could do a mapping of the physical drive through to the vm. Yeah, you can.

Jeff Massie [00:52:05]:
You can do that.

Jonathan Bennett [00:52:06]:
Yeah, that'd be fun to be fun to play with. You could totally do that.

Jeff Massie [00:52:12]:
Yeah, you make kind of a pass through.

Jonathan Bennett [00:52:14]:
Yeah.

Ken McDonald [00:52:16]:
It probably be more difficult to set up the VM than the. An actual hardware.

Jonathan Bennett [00:52:22]:
Yeah, that's true. That's true. All right, Jeff, there's a DDoS going on.

Jeff Massie [00:52:29]:
There is. But first, a little bit of errata from our last story. So just to give people an idea, an exabyte, this is from metal slurry, is a unit of digital information equal to 1 billion gigabytes.

Jonathan Bennett [00:52:42]:
Right.

Jeff Massie [00:52:43]:
And Mr. Juice found that if you use 64 megabytes memory modules, you need roughly 17.18 billion. And Jonathan came up with, if you use 64 gigabyte memory sticks, you need 15.6 million memory modules. That'll give you an idea of kind of how big everything is.

Jonathan Bennett [00:53:12]:
I guess really the question we should ask is what's the biggest single stick memory module that you can get?

Jeff Massie [00:53:20]:
Is this commercially available? Is 100.

Jonathan Bennett [00:53:22]:
And is it 128?

Jeff Massie [00:53:24]:
128.

Jonathan Bennett [00:53:26]:
Yeah, that. That might be. Currently regular desktop ones go up to 128. Oh, you can get them for the thread ripper at 256. Oh, okay, so 1684. So like 4. 4 billion. 4 million of those would get you up.

Jonathan Bennett [00:53:45]:
Only 4 million? Yeah, yeah, yeah.

Jeff Massie [00:53:47]:
Only 4 million. And just think of the arthritis you'd have after pushing all those in. I mean, we have arch we should talk about now.

Jonathan Bennett [00:53:57]:
Yes.

Jeff Massie [00:53:58]:
And I don't believe we've talked about this, but For a few weeks on and off and over the summer actually, Arch Linux has been under various levels of denial of Service attacks, or DDoS as it's otherwise known. In the past couple of weeks the DDoS attacks have been turned up and caused some people to not reach the AUR or even the main website to get updates or information. Now the reason why these attacks are happening is not clear or if Arch knows they're not saying anything and they've been pretty tight lipped about this whole thing. They had, they've said when they have all the information and all the attacks mitigated, they will give more details. Until then they're staying tightly lipped. Now they have said that, you know, they're working together with their hosting provider, you know, they're working on countermeasures. Now, from the speculation, there doesn't seem to be a clear group or political backing. So it's hard to say what the goal of the attack is.

Jeff Massie [00:54:58]:
I mean, other than you know, of course, the goal of trying to shut this down. But the why behind it is kind of really cloudy. There's a lot of different theories out there, but nobody really knows. Now people have asked why not use Cloudflare, is there willing to help open source projects and by protecting them from DDoS attacks for free? And there's not a clear answer there but because Arch isn't officially saying anything. But the general thought is held by most that even though Cloudflare supports open source, the company itself isn't open source. So there's thought that that kind of conflicts with one of the main tenets of Arch Linux. The Arch maintainers have said they've been evaluating anti DDoS service providers, assessing them in terms of cost, security and ethical standards. But as the project is run by volunteers, they ask for patience and thank the users who've been patient so far.

Jeff Massie [00:56:00]:
If you take a look at the links in the show notes to better understand if there's an issue with Arch being under attack or there's an issue with your local network. I have a link which is a second link in the show notes, the Arch Linux server status page. Now this will let you know what's up and what's down and other information on how the Arch infrastructure is doing. So you can get kind of a. It's not necessarily real time, but fairly, fairly quick. You can see what everything's running or not. The other article in the show notes goes over the. The first article in the show notes goes over the story a little more detail, but what's really Important is how you can still function with Arch if it happens to be under a DDoS attack.

Jeff Massie [00:56:43]:
Now the first is to use the mirrors you already have in your Pac man mirror list instead of relying on the reflector tool. There's a good chance that not all the mirrors are affected by the attack so you can still get updates and packages from the mirror. Now the second requires you to be a little bit used to git but you can clone packages from the GitHub Arch Linux mirror. So you do a Git clone space-- branch space and then the package name space--single branch/GitHub.com ArchLinux aur Git space package name so you can get around it that way as well. If you're, you know, if you're, if you're an old hand at git, you be able to do that. Now if you need an ISO image there, there are mirrors on the Geo mirrors list which can be found in the third article in the show Notes. Or if you want to cut to the chase you can go to. So looking at, when I'm checking the server status and everything and looking for information and going to various sources for this, it looks like the, you know, we might be past the worst of it but until we have full details on what's really going on and by whom, you know, there's not a ton of information other than things seem to be up up now.

Jeff Massie [00:58:27]:
We'll let everyone know once we get more information and heck, Jonathan might even go into it for a hackaday article or something once the full extent of it comes out. But you know, until then good luck. You know, the Internet's a rough place.

Jonathan Bennett [00:58:41]:
Yeah, I'm super curious and I've seen other people ask this like why Arch? Why is somebody attacking Arch? What's the, what's the deal there? Like I don't, I don't.

Jeff Massie [00:58:51]:
Well and there's been a few other open source projects that have been hit a little bit but like doesn't seem like you know, Red Hat or Canonical or, or you know, suse any, you know, stuff like that has been hit. It's been very Arch based and then I think that the Linux foundation got some and, but it's, yeah it's, it's really kind of like why, what is the goal? What are, you know, so it's a lot of times, you know, you Denial of service, there's somebody made a statement or somebody did something that's made somebody mad and you can kind of get an idea of what's what's going on. But this one, a lot of people are scratching their heads.

Jonathan Bennett [00:59:36]:
Yeah. I almost have to wonder because as I think about Arch, like, they. They just aren't. They don't make, like, political statements. They stay out of the culture war. And, like, is that what it is? Is someone mad that Arch has not. And I'm not. I'm not even going to go into one of the various things, but, like, has not made a statement on X, Y and Z and therefore we must take them down.

Jonathan Bennett [01:00:05]:
I don't know.

Jeff Massie [01:00:06]:
Maybe. And I mean, who knows?

Ken McDonald [01:00:11]:
Or it could be like Jonathan's wife said, somebody got tired of hearing I run Arch.

Jonathan Bennett [01:00:20]:
That's it. That's got to be it. Quit telling us about Arch. Yeah, that's great. All right, well, there is. There's some fun news. We're going to get away from all this DDoS stuff. I've got some.

Jonathan Bennett [01:00:35]:
I've got some fun news. We got some updates about kde, and I've got some updates about my little ongoing experiment. But first off this week in kde, this is from Nate Graham, not the Pointed Stick blog. This is actually the official KDE blog, talking about the things that are coming in 6.5. And one of those is the initial system setup. The KDE Initial System setup, or kiss, if you will. That used to be just a skunk work, but it is now part of the setup story. And that is things like pick your keyboard layout, pick your preferred language when you first log in as a user in kde.

Jonathan Bennett [01:01:20]:
And so that is coming along, and it's actually going to be pretty cool. I think it's going to be pretty useful for new users and setting up a new desktop or laptop. There's some neat stuff in Plasma 6.5. One of that is that plasma panels will now become scrollable when they've got too much stuff in them. And this is something that can really be a problem. I've seen it in various places through KDE and other desktops, by the way, where you've got a panel, it pops up at, like, a fixed size and it's got stuff in it. And for whatever reason, yours has more stuff in it than the guy that, you know, originally wrote it and picked the size. And so, like, there's text that goes off the bottom of the screen that you just have no way to get to and read.

Jonathan Bennett [01:02:08]:
And so that sort of thing is what they're targeting here, where it's like, okay, fine, if one of these panels has text flowing off the edge of it, make it scrollable and that hopefully will fix a lot of these little pain point bugs. HDR content, they are improving the tone mapping curve and the comment is hopefully it should look even better now. And I went through to the link on this one and one of the interesting things is that he Xavier Huggle, the guy that does a lot of the HDR stuff there makes the statement that it's going to make tone mapped things a little bit darker. Tone mapped content is going to be a little bit darker and I'm trying to figure out if by tone mapped he means SDR content that's tone mapped onto HDR which that would be a win because in my opinion KDE handles that wrongly and it makes SDR content too bright. Or if he's talking about like HDR content is going to be darker which would be unfortunate because like I just said the SDR content is already too bright. We'll see when 6.5 is testable. I will definitely run to it and give it a try. There's new stuff for drawing tablets.

Jonathan Bennett [01:03:18]:
There is some fixes for the volume feedback, accessibility fixes and all kinds of other stuff that's just in 6.5. There's bug fixes for 6.4 like cloning a panel also now clones the settings of its system tray widget so that could be useful in certain cases layout fixes. All of that good stuff coming in the 6.4.5 and 6.5 versions of Plasma. So cool stuff there. I talked about HDR and there is an update. Google Chrome beta is now 1.4.1. That is the version of Chrome that has the HDR Wayland support. So if you have an HDR monitor and you're running KDE specifically, I believe all this landed in KDE 6.4.

Jonathan Bennett [01:04:15]:
You can go grab the beta of Google Chrome and get HDR support there in the various websites supported on their videos, YouTube being one of them and go and actually check it out. It's going to be another probably about a month I think. I think Chrome is on about a month cadence 1.4.1 will come out as stable but I was pretty excited to see that it is in just the last couple of days has been promoted to beta and so is a little less scary to run on your system. And then one more bit of KDE news and that is the alpha release of KDE Linux the distro that I still don't fully understand why it exists, but it is, it's now out in alpha. This is the. It's. It's somewhat Based on Arch. Although in the blog here about it Nick Graham again says but it's not an Arch based distro.

Jonathan Bennett [01:05:17]:
Okay, something interesting. There's not a package manager for KDE Linux. Everything except for the base OS is either compiled from source using KDE's KDE builder tool or Flatpak, which means that it's sort of a. It's sort of one of those immutable distros in a way. So they're playing around with some really interesting bits here, some really interesting things and I like that. Right. Like I do like when Linux distros play around with interesting things and try new ideas and that is I think probably the best reason for some of these niche distros to exist. I would not tell a newbie to go try this, but I would tell Rob to go try this.

Jonathan Bennett [01:06:05]:
I may try it on something just to take a look at it. Yeah, it's interesting. We'll see what happens with this. Whether it lives for a while or they realize pretty quick that it was a bad idea and it's more trouble than it's worth. We will see.

Jeff Massie [01:06:25]:
Wizardling says, I gathered KDE Linux is meant to be more of a day to day usable distro versus testing like Neon.

Jonathan Bennett [01:06:33]:
Yes, that is intended to be the idea to some extent.

Ken McDonald [01:06:36]:
But you're not going to use it on some of the older computers.

Jonathan Bennett [01:06:39]:
Probably not. Jeff, you were going to comment more on that?

Jeff Massie [01:06:43]:
Well, I was just going to say, but I don't get it. So the idea behind Neon, it's testing. So what they're doing is using cutting edge KDE on a stable system like kernels and things like that. Though they do have sometimes issues where we got to update a kernel to get the feature that to run the latest KDE version. But then if you say well I want to run cutting edge kde, I mean you can do that on Kubuntu Fedora, you know, there's a lot of, you know, you could probably do it on an Arch Linux. I mean I don't really understand what's unique about this versus several distros that kind of can do the same thing. And if you're, I don't know, it's just not something a new user would want to use and if you're not a new user, you're advanced. Well, there's already ways to do this.

Jeff Massie [01:07:38]:
I just don't understand the purpose of it.

Ken McDonald [01:07:42]:
Notice that it gives examples of some of the things that are intentionally unsupported in this loading new kernel modules that runtime running on computers with only BIOS support.

Jonathan Bennett [01:07:58]:
I think you mean, I think you mean EFI support, not BIOS support. Surely, surely it's not legacy boot only.

Ken McDonald [01:08:06]:
No, yeah, that's, that's what I'm saying. Unsupported.

Jonathan Bennett [01:08:11]:
Oh, okay. Okay.

Jeff Massie [01:08:13]:
Oh, unsupport. Okay, I got. I misunderstood too. I'm like bios, why would they only support.

Jonathan Bennett [01:08:21]:
I have not heard.

Ken McDonald [01:08:25]:
How many systems are actually or distros are actually switching to UEFI only.

Jonathan Bennett [01:08:33]:
Some are. A lot of them still have some legacy. But yeah, it is definitely the future. I think what's going to be real interesting is how many new computers don't have a legacy fallback for boot option.

Jeff Massie [01:08:44]:
Which would be a problem then for Nvidia firmware.

Ken McDonald [01:08:50]:
And that's something else that is not going to be supported. Proprietary drivers for Pre Turing Nvidia GPUs.

Jonathan Bennett [01:08:59]:
Yeah, so reading through this again, there is another really interesting note here as he's talking about KDE Neon. And so Nate does call this out and he says KDE Neon is not canceled. However, it has shed most of its developers over the years, which is a problem. It is currently being held together by a heroic volunteer. KDE Neon is basically one guy at this point. So he says kdeev, the company, the organization has been reaching out to stakeholders to see if they can help to put in place a continuity or transition plan. All right, so it is very possible, reading between the lines here, it is very possible that KDE Neon will go away and KDE Linux will be the new KDE Distro and that they will stop doing the double work here. He goes on to say, while Neon continues to exist, KDE Linux therefore does represent duplication.

Jonathan Bennett [01:10:07]:
Yeah, it's. It's interesting. I would not be surprised here. In a few months, maybe they will wait until KDE Linux goes to beta or a full release. Would not surprise me to hear that Neon is being phased out and will get dropped.

Jeff Massie [01:10:23]:
Well, and that could align with where I've seen problems with Neon where the LTS just don't have new enough packages in them to run some of the latest QT functions. And so now kernel and other stuff into an lts.

Jonathan Bennett [01:10:40]:
Yeah, I noted that when I tried to grab Neon to run like a Preview of the KDE 6 stuff, it was just broken. It was terribly broken because it had such an old kernel in it.

Ken McDonald [01:10:52]:
Is Neon able to support anything running with QT6?

Jonathan Bennett [01:10:58]:
I mean Neon is now because it's rebuilt on top of 2404, which is now getting long in the tooth once again.

Jeff Massie [01:11:08]:
When it originally KDE 6 was originally coming out and they were doing early like when it was beta and that kind of stuff you get neon. But then you a lot of times had to force a newer kernel in there and some other stuff. It wasn't a simple just oh, fire it up and go. I mean they have a version, you could do that. But you weren't running the latest KDE then. And so it was, it was a very kind of hand built system to get it to work with the latest KDE on their LTS platform.

Jonathan Bennett [01:11:40]:
Yep, yep. Kind of kind of a hack to make any of it work.

Jeff Massie [01:11:46]:
And, and so everybody knows you mentioned they don't support. They support touring and later. Correct. So nothing pre turing. Yeah, touring would be the 2000 series cards. So if you have like a 1080 or something, it won't be supported. You have to have at least a 2000 so it can't. That came out.

Jeff Massie [01:12:06]:
Those came out and I think it was like 2018.

Ken McDonald [01:12:10]:
Are they in an affordable price range?

Jeff Massie [01:12:16]:
I haven't looked. You can probably find it because they're. I mean there's seven. Seven years old.

Jonathan Bennett [01:12:26]:
So there's very possible that they're starting to get affordable. Yeah, starting to the GPU market is nuts right now. All right, let's get into some command line tips. We've shredded the news and let's move on to the cli. Ken is up first with another wpctl. More wire plumber. Goodness.

Ken McDonald [01:12:50]:
Yes. We are going to do some more wire plumber Goddess. I'm going to switch to where you all can see my command lines. Do I need to expand the text on them so you all can actually see it?

Jonathan Bennett [01:13:02]:
You can zoom in just a little bit. It'd make it a little easier.

Ken McDonald [01:13:05]:
What I've got up for those of you all listening is two terminals in the one on the left. I've ran the status option for WP CTO so that you can see what clients and audio and video devices I have. And if you look at the sinks, you'll see I've actually got two audio devices there. One's the hardware device that's virtualized in my VM and then a virtual sync that I created and I'll show you why in a little bit here. But the command I'm going to go over today is how to set the default device or object that you want to use for playing or capturing audio in this case. Hopefully it'll work with video in the future. But as I said the command is going to be WPCTL sit deval. If you do a dash H after that, it gives you the specific commands and tells you what it does.

Ken McDonald [01:14:20]:
That sets whatever ID you provide to be the default object of its kind, whether it's a capture or playback for that session. And there's no specific options that you use. With set default, you just run the command. So I'm going to go ahead and clear this and we're going to run with let's change that to the sync, because right now the status is showing that my ALSA output is the default. So we're going to go with ID 69. And for those of you listening, the screen just updated saying my sync. But that's the only thing that really shows it. So let's go ahead and clear this and run that status again.

Ken McDonald [01:15:18]:
And with the status of y', all, remember, at the very end of it, it gives you default configured devices. And in this, this time it's showing the audio sync is my sync, with, of course, the audio source still being the same. Now, another way to demonstrate is to start running VLC running a video. And I've got my pipewire graph up there, and it's showing that VLC outputs are going into the My Sync virtual sync that I'd created into all six inputs. So it's a 5.1 set up there the way it was configured. And instead of going straight to the audio out. Now, while that's running, if I switch back from that virtual sync by changing the ID to 50 back to the audio also output, it automatically jumps to it. So I can see where there could be some great uses with this command, either to set up the default connections that you want to use, or the default devices or syncs that you may want to use when you're booting up.

Ken McDonald [01:16:53]:
Or if you wanted to be able to change it on the fly while doing something like if you've got a stream deck, you could use this command to switch from one device to another.

Jonathan Bennett [01:17:12]:
Very cool. All right, so next up, Jeff, talking about game mode. Have we not talked about game mode before? And my tip is like this too. It's like, sure, we've talked about that, but I guess not.

Jeff Massie [01:17:29]:
We've touched on stuff around it, but we haven't actually talked about game mode. So like Jonathan said, my tip for today is game mode. This was originally designed by Feral Interactive as a stopgap solution. As for programs with so at AMD and Intel, CPUs used to have an issue with power save and power on Demand governors. You know when this is back when Pharaoh was doing a lot more porting games from Windows to Linux so they'd run natively on Linux, they found there's a lot of performance which was not being used because the governors were working in a power save mode and not letting the CPUs run at full speed. This was a simple way to change it and it wasn't permanent. So when you're done playing the game, then you could go back to power saving mode until the next time you needed full performance. It's basically a daemon library combination which allows for these performance enhancements features to be easily controlled.

Jeff Massie [01:18:31]:
And over the years it's gained a few more features than just CPU governors. It also has control of IO priority. Process niceness, of course, can get in the kernel scheduler screensaver inhibiting GPU performance mode for Nvidia and AMD and GPU overclocking for Nvidia. And then there's CPU core pinning or parking. And of course you can do custom scripts as well. So we've talked about the CPU governors in the past and we've touched on it and how you change it and. But this is, this is kind of more of a, a nice front end and if you, you know, the packages are available, game modes available for all the major distributions. I mean I, you know, the Debian, Zubuntus, Fedoras, Arch Suse, you know, a ton of them and.

Jeff Massie [01:19:30]:
But if you take a look at the link in the show notes, it'll take you to the GitHub page on the details on how to use it because I'm not going to go through all that for all the various nuances of what you want. But a lot of it's pretty easy. Like in Steam, it's in your command line inside Steam where you can put the switches to say whatever you want to do. You just put game mode space, then your percentage command percentage. So it just fires up game mode. They even talk about how you can get around issues with hybrid GPUs. So if you're on a laptop and game mode doesn't start until the game is running, they talk about how you can make sure you have your game running on the proper GPU and get the most out of it. So yeah, for everybody that wants to squeeze the most out of their machine, definitely take a look at game mode.

Jonathan Bennett [01:20:27]:
Yeah, absolutely. You've given me something to think about for next week. I'm going to look into. So I'll do, I'll do some more game related stuff Then already made a note about it because we've even touched.

Jeff Massie [01:20:40]:
On CPU parking on some of that. So, for example, some of the hybrid CPUs, where you have like the Intels, you have the performance cores and the efficiency cores, well, you might make sure that you're on performance cores, or in the case of AMD, you've got a, you know, a 900, you know, 79, 99, 5900 chip, or the 5950, 7950 so on. Then you can say, okay, I want to lock all the cores so the game runs on or one NUMA node so you don't have that delay as it crosses the boundary between. Between numa nodes. You can just say, okay, I want them to run on this certain set of cores that are faster, less latency.

Jonathan Bennett [01:21:24]:
Absolutely. All right, I have got a tip as well that I find. In fact, this is another one I was extremely surprised we've not done before because I have found it to be really useful doing certain things. Let's see here. Yeah, so it's strings, and this is a program, I think, installed by default on most distros. But if not, it is gnustring and it should be available just in your package manager. This lets you extract ASCII strings from a binary. And so you can do it, you can extract strings from, say, a firmware file or a program that you're going to run.

Jonathan Bennett [01:22:09]:
And I thought about what a quick demo would be and it was this, let's just run strings. So the way that you would do it is it's strings space and then the name of the binary you want to look at. And so here we're saying actually we want strings to look at itself. And so we've got the dollar sign parentheses, that's command execution. So what it's going to do is it's going to replace all of that with the output of that command. And so it will then run strings on the strings binary. And when we run that, we get this output. And so these are all of the ASCII strings that are inside that binary.

Jonathan Bennett [01:22:47]:
The one that's probably most interesting to look at here is the usage. So this is the same thing as when you run like a dash H or a help. These are the strings that are defined that spits out the usage, the HELP file, they're actually in the binary. And strings can extract them and show you all the ways to run it. It's also got all of the error strings in there. And then you also do get what is probably Garbage. So, like, it takes a little bit of using this on multiple files to kind of know which of this is just garbage that you don't really care about and which of it is the interesting stuff, like the error messages here and the help messages. But yeah, I've made use of this, running it on file binaries like this, but also like firmware files.

Jonathan Bennett [01:23:48]:
In fact, just the other day I had a firmware that I looked at and went, I wonder if these guys just forked another project. It's like, well, let's run strings and search for the name of the other project. And so it was, you know, strings that binary file and then run it through grep. Turned out, no, it looked like they had actually written it. They at least rewrote all of the internal debugging messages. So that kind of made me think that they probably indeed did write it from scratch and it was not just a, you know, a fork and a code theft with strings. Super useful. Very cool.

Jonathan Bennett [01:24:26]:
Yeah, I like it. All right, that is it for the show. I'm going to let each of the guys get in whatever they want to have the last word. We'll let Jeff go first. What do you have for us, Jeff?

Jeff Massie [01:24:41]:
I don't have a lot, so it's going to be Poetry Corner. Data swarms like flies, the World Wide Web of bounty. So where's the spider? Have a great week, everybody.

Jonathan Bennett [01:24:57]:
All right. And Ken.

Ken McDonald [01:24:59]:
Well, those of you all that have followed Untitled Linux show know that last week I talked about ZED providing the agent client protocol as a standard way for AI agents to integrate into zed. Well, this week Morgan Cray posted about Claude code now in beta in zed and I've got a link to his article.

Jonathan Bennett [01:25:22]:
All right, very cool. Appreciate it. Thank you guys both for being here. All right.

Ken McDonald [01:25:29]:
It's always fun.

Jonathan Bennett [01:25:30]:
Yeah. If you want to find more of my stuff, you can always check out Hackaday. That's where Floss Weekly lives these days. And it's where my security column goes live every Friday morning. A lot of fun with both of those. And you should also think about Club Twit if you're not a part of it already. It's not much more than the price of a cup of coffee per month. And it gives you ad free access to the shows, behind the scenes looks and access to the members only discord.

Jonathan Bennett [01:25:54]:
It's a lot of fun. You should think about it and check it out. All right, we appreciate everybody that's here that has gotten us live and on the download and we will be back next week for more fun on the Untitled Linux Show.