Untitled Linux Show 215 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.

00:00 - Jonathan Bennett (Host)
Hey folks, this week we're talking about Audacity and Bottles. Ubuntu Debian 13 is finally out. Things inside the kernel are complicated, but there's some neat stuff on the horizon. The DuckStation PS1 emulator had something of a dust-up with Linux. This week we get to the bottom of that. Oh yeah, and there's a really big AI lawsuit brewing that we talk about. You don't want to miss it, so stay tuned.

00:27 - Leo Laporte (Announcement)
Podcasts you love From people you trust. This is Twit.

00:36 - Jonathan Bennett (Host)
This is the Untitled Linux Show, episode 215, recorded Saturday on August the 9th, unencumbered by the Thought Process. Hey folks, it is Saturday and you know what that means. We're going to get geeky with Linux and open source. We're going to talk hardware and software. It's going to be a lot of fun and you don't want to miss it. So it is, of course, not just me. I've got the guys. Two of the guys, rob, is not here, but I've got Jeff and Ken and we're going to dive in and talk about some Linux stuff and we're going to let Ken go first and Ken is going to talk about. He's going to have the audacity to talk about Audacity and tell us what's new there.

01:16 - Ken McDonald (Co-host)
Well, I'm always happy to talk about Audacity, especially since, according to Marius, there's a new stable version coming out. This week Audacity 3.7.5 was released to fix several crashes and, if you are interested, to add Windows on ARM64 support. It fixes a crash during wave import when the audio file is up to 7 milliseconds long, a crash when using Macro Wizard, and I want to thank you, haley Somerville, for fixing a crash when rendering Spectrum View. Now we can also thank Dr KD Murray for introducing 32-bit PCM to the FLAC importer. Audacity has been available via Pi Apps on either an Ubuntu or Debian, arm32 or ARM64 OS. Now, as I'd mentioned previously, we hear it is now available for Windows on ARM or WOA, and that's WOA 11 only. We can thank Terrence Ordano for providing a WOA fork of FF MPEG that is required to get Audacity working on WOA 11.

02:41 - Jonathan Bennett (Host)
Wait, wait, wait. What is WOA 11?

02:46 - Ken McDonald (Co-host)
windows own arm version 11 okay, and that's the only version that audacity for arms, for windows own arm, will run on I see, do you still?

03:03 - Jonathan Bennett (Host)
do you use audacity for stuff, ken? I've not.

03:05 - Ken McDonald (Co-host)
I've not actually for quite a while I haven't touched it recently, but it's part of my workflow for ripping vinyl yeah, definitely I hook up my usb uh vinyl record player and fire up Audacity and use it to record what I'm playing back. I'll play both sides into one long wave. Then I can go in with Audacity mark where each of the songs break and I can put in the names, and then Audacity allows me to convert that to a FLAC file for each individual one.

03:49 - Jonathan Bennett (Host)
I have been using Ardour for quite a while, but Audacity there are a couple of things where it just works better. One of the problems I've had is like trying to go in and automatically trim silence. So I had some audio files that were sort of a podcast format, not exactly, but there were. There were sort of long breaks of silence where someone was taking time to think about something, or I think it's actually where people were taking time to look at their character sheets and they roll dice was actually what the long breaks were for, but anyway. So I had all this audio, like I don't want to have to go through there and trim all of that silence out, and so both programs Ardour and Audacity have tools to be able to go in and clip out silence. It just works better. At least last time I tried it it works way better in Audacity.

04:36 - Ken McDonald (Co-host)
It's a simple method with Audacity I just mark where I want the silence to start, mark where I want it to stop, and then I can just hit delete and it will delete that section out. I actually use it to insert silence in between the songs when I am ripping from my vinyls, instead of having that crackling noise.

05:02 - Jonathan Bennett (Host)
But that is part of the essential experience of vinyl and and some people actually, uh, master that in, even if it's digital, just to get that analog type sound oh yeah, I um, I've spent a little bit of time working in a recording studio and that was one of the things that, for one song in particular, is one of the things we did is we use one of those plug-ins this has been almost two decades ago now, anyway but we use one of those plug-ins to give a song that sort of crackly aspect to it. And, yeah, super, super interesting stuff you can do. All right, do we want to talk about intel again?

05:48 - Ken McDonald (Co-host)
yeah, only if it's good news well.

05:52 - Jeff Massie (Co-host)
Well, it looks like jonathan's up. Uh, no it's. You know, intel was always such a company that everybody loved to kind of beat up on. You know, they were the 800-pound gorilla, they were the king of the CPUs for so long, but boy, now they're really on their knees. And I have a pair of stories this week and it's not good news for the Intel Linux crew. Basically it's all related to the layoffs at Intel and how it's affecting the drivers and support for the hardware. So, for example, the Intel CPU temperature monitoring driver is now unmaintained. This driver had a patch posted on the 8th of August that was from Intel Linux engineer Dave Hanson. It said this maintainer's email no longer works. Remove it from maintainers. Also, mark the driver as orphaned for now. Basically, the patch removes Fenghua Yu as the CoreTemp hardware monitoring driver maintainer. Well, he's no longer at Intel, he's now at NVIDIA. So he's not without a job, but not with Intel anymore.

07:04 - Jonathan Bennett (Host)
Maybe they'll open up more of Nvidia's drivers as a result.

07:09 - Jeff Massie (Co-host)
Yeah, you know stranger things have happened, but it's going to be pretty. My odds are not. I'm not placing that bet. Also hit is the Ethernet RDMA driver. While it's still supported, it lost one of its co-maintainers, so there's still one one person now maintaining it instead of two. Uh tin fi Zhang is no longer at Intel and that leaves the Intel FPGA DFL TOD driver, which is the time of day device driver for the FPGA chip, orphaned as well. M Chetan Kumar is no longer at Intel and the WWAN IOSM driver. Basically, wwan is Wide Wireless Area Network.

08:09
Now, this one isn't a surprise because Intel has exited the WWAN and modem business several years ago. So you know this was coming anyway, because whatever they're supporting is older. But if you have an Intel M.2 modem and it's still in use, it could affect you down the road. Now you'd think the code would be fairly stable, so you should get a few more, several more kernel revs out of it. But just be aware it's not maintained. So if for some reason a bug comes up, it is what it is. Patches are accepted. Patches are accepted. There's also the Intel Keen Bay DRM driver, which is also, you know, the maintainer's also not at the company, and Neil S. Kevin Shemurthy is no longer going to be working on kernel probes to help with the debugging and profiling.

09:10
There's also other drivers that are going away, but they're also in products that Intel spun off and no longer part of the company. So, like some of the others, this is kind of to be expected. Do take a look at the links in the show notes for deeper details into the drivers which so far are not going to be supported, and there's some I didn't talk about but are also going to be orphaned. So I mean, not maybe quite as mainstream as some of the ones I mentioned, but there there's others. So definitely take a look at the articles.

09:39
Now I do want to preface that this doesn't mean that someone in the future isn't going to pick them up, and you know, especially for future looking products, I would guess that someone's probably going to pick them up. You know, either reassigned at Intel or someone outside the company is going to get access to some of the early documentation and they'll be able to maintain and support it and add features. We'll give updates as we find out more on what's going on and what's going to be happening with the Linux drivers for Intel. So we'll keep you updated as we find out more, but for right now, more bad news.

10:17 - Jonathan Bennett (Host)
Yeah, it's worth pointing out that it's even possible that some of these guys will continue working on these drivers at their new positions. That's not an entirely impossible thing. That sort of thing happens sometimes. So some of these guys they got removed as maintainers because that email address no longer is active, so they could just email from a new address and say, hey, I'm over here now, I want to continue working on this. That's possible.

10:43 - Jeff Massie (Co-host)
It is and some of these, especially for some of the older products, the code probably doesn't get touched very much at all anymore. Indeed, hopefully you know so. So if it's pretty much on uh, sustaining and you're not doing a lot of heavy lifting, they might just say yeah, I'll just. I'll just keep my name on it. Here's my new email address. You know I'll spend the 10 minutes every quarter to tweak it or whatever and call it good.

11:14 - Ken McDonald (Co-host)
It does make you wonder, though, how much of this is related to them not maintaining Clear Linux anymore.

11:24 - Jeff Massie (Co-host)
Oh, I're losing.

11:26 - Jonathan Bennett (Host)
Go ahead, Jonathan. I think it's both repercussions from getting rid of a bunch of engineers, the layoffs that they're doing in Intel to try to right the ship.

11:37 - Jeff Massie (Co-host)
Yeah, they're laying off thousands of people globally. They're shutting down fab projects. Every site is losing people. I mean a lot of people and management. I thought I heard something like they want to reduce it by two-thirds or something like that. They want to really cut out a lot of management and they're trying to get rid of about 25% of the headcount at Intel to try to make trying to right the ship, because they're in pretty dire straits right now.

12:11 - Ken McDonald (Co-host)
Especially with the upcoming tax season.

12:14 - Jeff Massie (Co-host)
Yeah, and they've there's been, they tried to. They were talking well, some of the board members wanted to spin off the foundries, the fabs, and sell them to tsmc, but then the ceo didn't want that and he thinks it's core. And so there's some varying viewpoints in the board room about what should be done. And but I don't think tsc, tsmc wants them, because it's it. You can't just take them over and run them. Yeah, there's a lot of intricacies in there too. Even if you're, even if your tool set mostly aligns, there's so many little details that things can run differently that it would be a major undertaking to take it over and make it work right, and especially those cutting edge geometries. And so Intel is under a lot of churn and chaos right now as they try to get their legs under them again.

13:15 - Jonathan Bennett (Host)
Yeah, I'm pretty sure I saw somewhere in a statistic about, like, the percentage of drivers in the kernel that is now orphaned. I'm not seeing it though at the moment. I want to say it was something like 8%. 8% of the kernel is orphaned, something kind of crazy like that. Hmm, I don't immediately see it though Somewhere. Though, somewhere there I saw that statistic. But, yeah, interesting stuff. It's been fascinating to watch this question of, as we sort of see, economic card times for some of these businesses. Where are they cutting and what are the things that they are cutting and where, where open source is in some cases taking the hit.

14:11 - Jeff Massie (Co-host)
Yeah, interesting to interesting to see. Well, and I could see someone going we have all these products that we're still supporting that. We spun off that division a few years earlier. Why are we still supporting it? Just cut it right now and we're done and walk away.

14:29 - Jonathan Bennett (Host)
It's a reasonable call to make. It's not unreasonable for them to do that. So I'm still sad about Clear Linux going down. But on the flip side there is a new distro to take its place. Probably not one for one replacement.

14:48
But we now have a pre-alpha of KDE Linux and I have been skeptical of this and I'm still a little skeptical of this. But Project Banana is now in pre-alpha and you can go grab it and test it. Uh, it's. There's some interesting things about it. Um, it's not a demonstration distro. That's kde neon.

15:16
Now, kde linux is um, it's going to be an end user distro. They're looking for it to be extremely stable and it's going to be. Well, it's built on top of Ubuntu LTS. No, sorry, that's Neon. I knew that didn't sound right when it came out of my mouth. No, it's built on top of Arch. It's actually very similar to Valve's SteamOS.

15:45
Kde Linux will be an immutable distro, which is sort of all the rage right now, for good reason. It's the easy way to get new people onto Linux and make it difficult for them to destroy it and make it no longer boot. Looks like they're going to do a dual read-only butterfs, root partition and, yeah, some interesting stuff here and again pre-alpha very early days for this. But you can actually go test it. Um, which is uh, which is pretty cool. They're also um to do to do kde linux only supports recent nvidia gpuss, for which NVIDIA supplies FOSS drivers. So at this point they are not packaging the NVIDIA drivers, which is going to make that kind of a non-starter for some people. And because it's an immutable distro, it's going to be difficult to get those installed. You have to work really hard to be able to do that, basically going to break your system to be able to fix it. But yeah, kde Linuxux, it's interesting stuff. Um, I'm probably not going to run it. It's probably not for me yeah, so is is.

16:55 - Jeff Massie (Co-host)
Is the goal just to test kde with newer hardware than neon, or I mean newer software than neon, from?

17:03 - Jonathan Bennett (Host)
from what I am reading, they actually want it to be a full blown stable driver end user daily driver distro.

17:10 - Jeff Massie (Co-host)
Yes, it is not just a demo just like steam os yeah, but I mean really, how is that different than you know you're like cashy, where you can get.

17:27 - Jonathan Bennett (Host)
You know, you're like cashy where you can get kde on it, or you, you know, it was it.

17:31 - Jeff Massie (Co-host)
Uh, mandrake man I yeah, I have some of the same um. It just seems like we already have arch distributions, several flavors that have kde as an option already. How is this?

17:42 - Jonathan Bennett (Host)
yeah, as I said, I have been skeptical of this. I think the KDE guys should be focusing on building the desktop and not trying to make a full-blown distro. I think that you've got multiple really good options. If you want an immutable, really good KDE desktop, go run one of the Fedora immutable spins. Put KDE on it.

18:06
Or one of the Ubuntuutable spins put kde on it um or one of the ubuntu I mean there are. There are more options, there are a bunch of options out there, um and and if your adventures go with something arch based, well, I mean, that's what this is.

18:18
Maybe there wasn't an arch based kde immutable distro out there. I don't know it's strange, but it's one of those things where it's like if somebody wants to work on it, then okay, fine, nothing wrong with that. I just I hope the people that have been working on KDE don't go and start spending their time on this instead. That's my main concern.

18:43 - Jeff Massie (Co-host)
Yeah, I just want to see kind of reinventing the wheel when it's like this stuff's already out there.

18:48 - Jonathan Bennett (Host)
Yep, yep, that is the thing that comes to mind. But it's out there, you can go test it. Go grab the ISO.

18:55 - Ken McDonald (Co-host)
Looks like the author of your article, liam Proven, thinks that this could one day be a bulletproof system, or it's intended to be.

19:06 - Jonathan Bennett (Host)
I mean, that's what. That's the whole point of the um. Immutable, yeah, this whole point of immutable. So it uh. But I hear that with every new distribution it's hot, immutable.

19:17 - Ken McDonald (Co-host)
Linux is hot, everybody likes it you hear, even hear that about some of the older distributions yeah, I don't know that are that have immutable. You even hear that about some of the older distributions. Yeah, I don't know that. Have immutable system drives Mm-hmm, mm-hmm.

19:34 - Jonathan Bennett (Host)
Yeah, all right, do we want to talk about bottles? Are bottles immutable? It's sort of kind of the same flavor of thing. I might be stretching it just a little bit, just a little bit If you're unencumbered by the thought process sure, but this week Bobby Borisov and Christine Hall report that Bottles has funding.

20:02 - Ken McDonald (Co-host)
Both Bobby and Christine cover the Bottles project's rough financial patches impact on its development project and I think even Rob touched on that a couple of episodes ago. In fact, rob, I think, even asked people to help donate to support bottles a little bit. Well, bobby reports that Bob's request, as well as Mirko Bromben's, didn't go unheard. Quoting Christine hallelujah, bad news has been followed by good news. Founder and lead developer Mirko Bromben announced in a blog that Bottles received an NLNet Commons Fund grant through the 2025 Commons Fund. Mirko said this support will help us accelerate our work on the Bottles Next project, bringing a more modern and polished experience to running Windows software on Linux. The grant recognizes our efforts to build a better future for running Windows applications on Linux. The grant recognizes our efforts to build a better future for running Windows applications on Linux.

21:08
According to Bobby, the NLNet Foundation support comes via the NGIO and I think I'm getting that right it's not NG10, but their Commons Fund, which is backed by the European Commission's Next Generation Internet Program, along with additional funding from Switzerland's State Secretariat for Education, research and Innovations, s-e-r-i CERI. Now, according to Christine, the amount of the grant hasn't been made public, but the NLNet Commons Fund typically awards grants ranging from 5,000 euro to 50,000 euro. I think that probably converts to about $5,800 US, maybe up to $58,000 US per project, with some larger projects receiving as much as €150,000. Now Bobby and Christine also provide more details about the 2025 Commons Fund and the NGIO Commons Fund. So I do recommend reading their articles if you want the nitty-gritty details articles if you want the nitty-gritty details. Also, as you travel down that rabbit hole, you might even stumble over what System 76 provided to bottles next.

22:36 - Jonathan Bennett (Host)
So Jonathan, are you using bottles, not bottles itself? No, I use a Lutris for stuff like that, that, and it has some bottling built into it. And then of course, steam has the sort of the same you know using wine prefixes as well built into it. So there's several different ways to kind of accomplish this, and Code Weavers uses bottles.

22:57 - Jeff Massie (Co-host)
If you, if you get the actual Code Weavers wine.

23:02 - Jonathan Bennett (Host)
Yeah, there's. There's several things out there that sort of let you accomplish this, but you know nothing. It's neat to see the Bottles project out there and good for them for getting some funding. I mean, that's always. It's always a very cool thing when you can get paid for the open source project that you enjoy doing. That's one of those kind of win conditions to life.

23:21 - Ken McDonald (Co-host)
Especially when it's by a government that endorses using open source to life. Especially when it's by a government that endorses using open source.

23:26 - Jonathan Bennett (Host)
Yes, sometimes getting paid by governments has strings attached. Just you know extra paperwork and all of that. Europe's leaning into Linux pretty hard nowadays. It is, it is, and it's very interesting to see. Yeah.

23:40
Getting away from some of those American commercial ventures, yeah I, uh, you know, personally and I say this as an american it doesn't bother me so much that they're american ventures, but that they're proprietary and closed source ventures. That, uh, you know, I'm I'm sort of afraid of that, no matter what country a company is from, and uh, I have very little trouble, you know, it does not bother me at all to run hardware or software from pick your country. Um, I say that with some caveats because I'm not naive, but, uh it, the quality of the code and the openness of the code is way more important to the country of origin and how often it phones home.

24:23
I mean ideally you can audit the code and see and figure that out.

24:27 - Jeff Massie (Co-host)
Well, when you look at it, certain parts of Europe left Microsoft for a Linux distribution that's based in America. I mean Linux Foundation is in America, I was considering that and Red Hat, red Hat is based in America, that's IBM, yep, but you can tear it all apart and see what's in there.

24:49 - Ken McDonald (Co-host)
Or fork it.

24:52 - Jonathan Bennett (Host)
We know people that have done that too. Yeah, indeed, all right. So, speaking of trouble in paradise, what is up with DuckStation? And in fact, jeff, what is DuckStation? And then what's the deal with?

25:12 - Jeff Massie (Co-host)
it. So, in case you've not heard of it, duckstation is a Sony PlayStation Gen 1 emulator, also known as a PS1 emulator. So it has a goal of being highly accurate and it's developed to run on many different platforms. Well, the problem is that the developer of DuckStation is looking at dropping Linux support totally because of an issue with Arch and the AUR. Aur is Arch's set of repositories repositories. The story behind this started when the author made a package build script P-K-G-B-U-I-L-D script and this was to help resolve issues which were coming up in unofficial AUR packages. The author states that the problem is the official guidelines are being ignored and they're using broken packages, which then users come back to him with complaints. So the author pulled the PK, the package build script, because of the many complaints. You know. He just he's not happy that he's getting all this and he's not happy that he can't request removal of the packages in the AUR which have been a problem, because he has to disclose a bunch of personal details to Arch to make that happen. So the author had this to say. So this is step one. Next step, we'll be removing Linux support entirely because I'm sick of the headaches and hacks for an operating system that only comprises 2% of the user base and I don't even use myself, so he's very unsrilled with the flack. He's getting over this. Now I'm sure there's a lot of people thinking, okay, we'll just fork it and continue on. Well, there's a hitch in that giddy-up.

27:01
Previously the software was released under the GPL version 3 license, but in 2024, the license changed to the CC-BY-NC-ND license. This is the Attribution Non-Commercial, no Derivatives. 4.0. International Deed Well, that's a mouthful. So Attribution 0.0 international deed Well, that's a mouthful. So attribution you must give all appropriate credit, provide a link to the license and indicate if changes were made. You may do so in any reasonable manner, but not in any way to suggest the licensor endorses you or your use. Non-commercial, of course you may not use this material for commercial purposes no additional restrictions. So you may not apply any legal terms or technological measures that legally restrict others from doing anything the license permits. Here's the one that kills a fork Repackage, even if a community repackage. No derivatives. If you remix, transform or build upon the material, you may not distribute the modified material. So you can't fork it in any way. Basically, without being in violation of this license.

28:17 - Jonathan Bennett (Host)
I think I found the problem. I think I know I'm going to break in here. I think I know why he was having problems. He wrote a license that says that the AUR guys can't fix anything. When they find it, Of course they have to bug him about it. He holds the keys to the kingdom. Nobody can. Oh, my goodness yeah.

28:38 - Jeff Massie (Co-host)
Well, there's still his hope, though, because the author did make a statement that he expresses hope that the Linux community might become more reasonable about these issues.

28:48 - Jonathan Bennett (Host)
So he wrote a license that says that for any fixes to make it into his source code they have to go through him, and then he complains when people bug him about problems. This is someone that does not understand how open source in the world works. This is not an open source license as well. It's a source available license, but this is somebody that just does not get how Linux works.

29:12 - Jeff Massie (Co-host)
I just basically hope this all gets sorted out, but, as we say a lot here, we'll keep everyone updated as events warrant.

29:21 - Ken McDonald (Co-host)
Is the older version still under the GPL version 3 open source? And how far back is that going.

29:31 - Jonathan Bennett (Host)
Yeah, 2024 is when they switched over late 2024, so there's six months worth of work or so. Um, you could do the one. One could make the argument that, depending upon how they did that switch, it may not have been legitimate. And so if there are people in there, if people hold copyright in there that did not consent to that, then there's an argument to be made that the whole thing is still available under the GPLv3. A lot of people out there don't understand how the GPL actually works. Yeah, this would be interesting to dig more into the uh, the source code and the get history of it and all the people that are in there, and how they did this transfer over to a different zany license and whether or not, it would be worth going back far enough in the history and forking at the point before the new license came out yeah, you could do that.

30:25
You could absolutely do that and I don't. I don't necessarily have a problem with the creative commons licenses, like some of them are great. I've released some, some text under creative common text and pictures under creative commons before. They're just not. They're not good source code licenses. That's not what they're made for and that's you know. Source code does not belong underneath creative commons licensing they're more for when you're releasing documentation isn't it.

30:49
You can do it for documentation they work for. So writing like prose, poetry, documentation. A lot of pictures get released under creative commons, video music gets released under creative commons, and because they're they're very easy to understand, like you know. What do I want people to be able to do with this? What do I not want people to be able to do with this? You know so by attribution when somebody uses it, my name needs to go on there. No commercial, you can't make money from it. Uh. No, derivative, I don't want you to remix it essentially, and you know it's very clear to understand and they've got these different flavors and you can sort of mix them. That's why it's popular. People like the creative commons because of that. Uh, but it's just. Most of those are not it. They're not compatible with the definition of open source and you know there are really good reasons why we have the definition of open source, that we do. And this sort of highlights one of the ways that can go wrong when you say no derivatives, it makes it very difficult for people to fix bugs. So, all right, let's take a look at um bcash fs.

32:01
Yes, so there's some stuff going on in the kernel and there's a couple of stories on Pharonix about this and I basically wanted to give folks an update that we still don't know for sure what's up with Bcash FS. It's just sort of stayed. Nothing has changed yet. As of the last I've seen, there was a pull request made by BcashFS developer Kent Overstreet and it's just sort of been ignored. There was a question posted about what's going on here and nobody has answered it. Kent says it's still his Colonel. If he wants to remove it, that's his choice. That's sort of where it boils down to there.

32:51
Um, so that that is the story I linked to. Actually saw um, uh, a second story on uh for Ronix. I'm just now taking a look at it. Um, apparently there's an ongoing discussion and an anecdote around butter fs came up, um meta. Somebody from metas talked about um, joseph bachik actually uh, talked about how the meta infrastructure is built completely on butter fs and its features, and he says we have saved billions with a B, a bravo, billions of dollars in infrastructure costs with the features and robustness of ButterFS, which is definitely interesting to see All kinds of fascinating stuff going on around this and I don't know what's going to happen. Torvalds has not chimed in. It's. It's sort of beginning to look like um, um, it's sort of beginning to look like this is just a one cycle timeout for the, the, the, the B cash, fs project.

34:02 - Jeff Massie (Co-host)
And so you know, there is some some hope that in six, 18, everybody will start to play nice again, but you know when you say it's not, when you say one cycle timeout, is that including the last timeout he had, or this is also going to? The 6.17 will be another timeout.

34:22 - Jonathan Bennett (Host)
So the fight happened during the 6.16, right after the 6.16 merge window, right, and so I'm thinking that we're almost to the end of the 6.17 merge window and my guess is that, just for all of 6.17, he's going to be timed out like just sort of ignored and once 6.18 comes around, hopefully cooler heads prevail and they start accepting code again. I am not Go ahead, jeff.

34:53 - Jeff Massie (Co-host)
I was going to say because 6.18, it was supposed to be marked stable. That was the plan that Kent had.

35:01 - Jonathan Bennett (Host)
Yeah, we'll see. We'll see what happens. I don't know if that'll happen or not. If it can, I am not a fan of the idea of not accepting code because of personnel issues. I think it's a problem because people actually depend upon this code and important stuff runs this code in some cases, and so I don't think that's the right call. But obviously I am not the, I am not Louis Storvalds and I do not get you know my opinion is worth about the amount of breath it takes for me to give it on this podcast. But you know that's where I come from. I think the code is more important than the personal issues, because when you're talking about the Linux project, the code is the project, and so that's what should come first. Obviously, not everybody agrees.

35:53 - Jeff Massie (Co-host)
And your opinion is worth the price of a subscription to Club Twit.

35:57 - Jonathan Bennett (Host)
Oh, is it worth that much? That're all paying. I suppose that's true. That is what we're all paying. Uh, all right, um, if we don't have any more thoughts about that, it sort of just is what it is.

36:12 - Jeff Massie (Co-host)
We're waiting to see what's going to happen and and so people know the it should the the merge window should close tomorrow.

36:18 - Jonathan Bennett (Host)
Yes, so all the all of the fun new stuff that's going to land in 617 has basically already landed, and then it's so. We're ready to talk about some old stuff. We can talk about some old stuff, some old stuff and some new stuff, things that are old or new. Again, take it away.

36:37 - Ken McDonald (Co-host)
Well, last month we saw Ubuntu 24.10 reach its end of life. Now this month, marius Nestor and Michael Larabelle write about an even older version that keeps improving with age. I am talking about the latest point release for the long-term supported Ubuntu 24.04 announced by Canonical. This is Ubuntu 24.04.3. And, of course, this is an LTS, so it's going to be supported for a while Now.

37:16
According to Michael, ubuntu LTS point releases are predominantly for bundling up all of those security and bug fixes that were provided through stable release updates that have been shipped down already to Ubuntu LTS users been shipped down already to Ubuntu LTS users. Now the new point release makes it more convenient for quickly deploying new Ubuntu LTS systems, with the stable updates over the past several months already included. Now, according to Marius and Michael, ubuntu 24.04.3 LTS is powered by Linux kernel 6.14 and the Mesa 25.0 graphics stack from the newer Ubuntu 25.04 release. Now, according to Marius, we can expect the next point release in February of 2026 for Ubuntu 24.04. Now I do recommend reading Michael and Marius's articles just if you want to find out what I didn't cover.

38:26 - Jonathan Bennett (Host)
Yeah. So it is interesting that this is happening now, because something else just happened and that is that Debian 13 just released. Debian 13 Trixie is now officially out. I believe today, I think today is the day Today is Debian day. Yes, saturday, saturday the 9th, I believe, is Debian day. So not only do we get a point release of Ubuntu, we also get Debian 13.

38:55 - Ken McDonald (Co-host)
I'll have to see if I can upgrade Linux on. We also get Debian 13. I'll have to see if I can upgrade Linux on my Chromebook to Debian 13.

39:01 - Jonathan Bennett (Host)
Are you running vanilla Debian 12? Not the Chrome OS thing.

39:10 - Ken McDonald (Co-host)
Well, I went from 11 to 12 on it when 12 came out.

39:16 - Jonathan Bennett (Host)
Very cool. What's interesting, so like Debian is becoming a little bit more popular again. They've made some changes and people are sort of looking at Debian again, but Debian is an upstream for so many different things, not just Ubuntu, but like Raspberry Pi OS, for example, is one of the big ones, and so I don't know how long it's going to take the Raspberry Pi guys to get the Pi OS 13 out the door, based on Trixie, but that's going to happen before too much longer, and so you may not be excited about Debian 13, but newer stuff on the Raspberry Pi should be exciting for anyone that's got one, and so there's trickle-down effects that will happen in various different places even if you're not actually running Debian anywhere.

40:01 - Ken McDonald (Co-host)
Well, and I don't. Go ahead, jeff, but I was just going to say Jeff, jeff you go.

40:09 - Jeff Massie (Co-host)
Okay, thank you. I'm not sure if it's really accurate, but a little while ago I saw a pie chart of Debian, of Linux users, and it was Debian and Debian based distribution, so your Ubuntu's and all that kind of stuff. It was about two thirds of the Linux market. And if you, if you take out Debian, you take out Red Hat and you take out Arch, there's not a ton of distributions that aren't based off of one of those three. But Debian was kind of the juggernaut of the house. Yeah, yeah, ken.

40:46 - Ken McDonald (Co-host)
Well, just going to mention, didn't we cover Debian 13? I'm going to include a new version of Apt.

40:55 - Jonathan Bennett (Host)
I know we talked about Debian 13. I don't remember off the top of my head all of the things that are baked into it, but I know there are some interesting things in there. Would not surprise me that there's a major update for Apt.

41:09 - Jeff Massie (Co-host)
I'm running the new Apt right now.

41:14 - Jonathan Bennett (Host)
So they've got some release notes here. What's new in the distribution? Official support for RISC-V64. Hardening against rope and COPGOP attacks. Http boot support. Improved manual page translations. Spell checking support in Qt Web Engine web browsers 64-bit time T-A-B-I transition. They're fixing the 2038 bug for them. They're having a progress toward reproducible builds. Yes, Jeff.

41:49 - Jeff Massie (Co-host)
Apt is 3.03, so that is the new version with the cleaner interface and it looks a lot pretty. 12 does not have it. 12 is on 2.6.1, so tricksy will be the first one where you get the clean, nicely formatted app and it's so much better than the old wall of text that just kind of not surprised yeah.

42:09 - Jonathan Bennett (Host)
So a couple things that stick out to me here in the rest of this list from what's new in debbie 13. Progress towards reproducible builds they're not there yet, but it's on their radar. That's actually a fairly important thing for distros to get these days is to be able to do reproducible builds. And then the other one is they're moving to Plasma 6 in Debian 13. So a reasonably modern desktop experience too. So pretty cool, Some good stuff in there.

42:37 - Jeff Massie (Co-host)
Yeah, they'll be running Qt 6 modern desktop experience too. So pretty cool Good stuff in there. Yeah, they'll be running QT 6.8.2 with that I'm looking at, and the Plasma desktop will be 6.3.6. If you look at DistroWatch and you click on Debian and if you scroll down a bit, it'll give you the major packages and then, if you really wanna know, you can hit like see all packages and it'll just give you all sorts of information on what versions It'll have, what the latest version is and then what the version is in the current setup.

43:12 - Ken McDonald (Co-host)
Yeah, you've got to experiment, for somebody would be to probably in a VM, put in Debian and then try to install the Cosmic desktop over it that would be very interesting to try to get Cosmic on Debian probably.

43:30 - Jonathan Bennett (Host)
I'm sure there's a way to do it. I don't know if it's officially packaged yet or not, but I'm sure somebody if there someone will have a guide on how to do that shortly.

43:39 - Jeff Massie (Co-host)
Inevitably. But Cosmic is still Alpha, I believe right have they hit Beta yet.

43:46 - Jonathan Bennett (Host)
I don't remember If they haven't.

43:47 - Ken McDonald (Co-host)
It's coming soon, let's see they're waiting for a special occasion to announce it a bit.

43:58 - Jonathan Bennett (Host)
The current Alpha 7, the robot says that the current Alpha 7 version is being described as pretty solid, and that is what I was remembering as well. Yeah, alpha 7 is the latest according to System76.com, and yeah, they're also calling it Epoch One, cosmic Epoch One. So it sounds like they're getting very close to a beta, and in fact their blurb here says only a few months out from release. So maybe Alpha 7 is sort of a sly beta in and of itself. Oh, it could be. So there's something interesting about Cosmic and that is that it is a Wayland, I think Wayland-only desktop environment, and that's sort of been a problem for some things like, oh, browsers. I think Jeff has a story about that.

44:52 - Jeff Massie (Co-host)
I do Well, for Chrome. Firefox has been Wayland friendly for a while, but the next major Chrome release will auto detect which Linux display is being used and then decide which part of the Chrome software should be used the X11 part or the Wayland part. Now the software in Chrome which will do this is called Ozone, and it's the abstraction layer that handles the graphics. Currently it defaults to X11, even if the system is running Wayland. So now this causes things like fractional scaling in Wayland to have not such good effects Like. So it'll make your Chrome window look blurry if you have fractional scaling turned on. It's also officially known as interpolation blur. That's because X11 can't handle the fractional scaling. So now, instead of defaulting to X11, the system is going to try to use Wayland if at all possible. So if you are running Wayland this is if the ozone layer is setting ozone is set to auto, so it can. Now you can still be forced to X11 if desired. So if you are running something and you get something weird, you can. You can still use the X11 and go through X, wayland and all that, like it's currently doing, but now they're going to have auto as the default setting for the next major update. So and the other thing they talked about is changing to Wayland the input should also seem snappier, with less lag and because you're not forcing everything through X11.

46:33
So, like I said, we should see this in in chrome 140, but if you're the impatient type and want to try it now, it is possible. So you can go to chrome colon slash slash flags in a new tab, locate the ozone dash, platform dash hint flag and set it to auto in the drop-down menu and then quit the browser. They said don't relaunch it, quit it. So then it'll come up fresh again when checking the setting, though it might already be set to auto and that's something that's not totally unusual. So some distributions make the change while they're packaging it and if you have a different browser based on Chromium, some of them also enable it. So while it's not the default officially from Chrome, there isn't anything totally out of the ordinary if you already find it on auto. So, like I said, there are a couple cases that somebody could be switching that ahead of time.

47:41
Now Google held off until 140 release because, depending on how old of a version of Chrome you're running, you could have bugs. They felt at version 140, all the bugs and quirks in dealing with Wayland should be ironed out. But the author of the article that's linked in the show notes says that they've been using it for years and there have been no issues. So depending on what you're doing, your distribution, probably a ton of variation on what versions of various files you have you might not see the issue. So take a look at the article linked in the show notes. Give it a shot If you're the adventurous type and see if you can tell the difference. Uh, if you know, if you're not doing anything that requires, you know, real quick Twitch and you're not doing any fractional scaling, you might not even really notice anything, you know. But if you do see issues or maybe you don't, you know, let us know in the Discord what you're finding. You know, give us thoughts on that and always happy to hear.

48:49 - Jonathan Bennett (Host)
Yeah, interesting stuff. I know when I went and did some fiddling around with trying to get various things in Chrome working with my my Fedora KDE desktop, that was one of the settings that I went in and played with to be able to force it to use that Wayland backend. The first time I did that it was very buggy. It was not a great experience, but it has gotten better and better, and I think I'm. I think I'm running on auto right now, but it has gotten better and better, and I think I'm. I think I'm running on auto right now, and so it is indeed. Um, it is running on the wayland back end and it's at the point now to where, you know, I basically don't. I don't see a difference between that and x11. Um, everything pretty much just works, except for windows restoring to where I left them on something like a reboot. That's not there yet, but progress is being made does firefox restore to where, uh, you left it?

49:39
not under wayland it. That feature just recently got added to wayland, so it'll be another you know iteration of kde and another couple of versions of the browser before all that actually starts working in practice hey, I just checked mine and my chrome is just on default, so I'm not on auto. Yeah, yeah, that makes sense. I think that's that's probably going to be what most, most machines are but I hardly ever use chrome other than it for this restream.

50:12
I use it for the restream use it once a week to do the show, because Firefox, for some reason, is not necessarily happy about it.

50:19 - Jeff Massie (Co-host)
Yeah, it won't pick up my virtual camera. It'll pick up my real camera, but it won't pick up my OBS virtual camera.

50:25 - Jonathan Bennett (Host)
Yeah, I'm trying to real quick check what mine is set to on this machine. I bet it's going to be set to default.

50:32 - Ken McDonald (Co-host)
Not a huge surprise. So you not a huge surprise. So you're using chrome or chromium, jeff chrome. Okay, I've got both loaded out. Check them after the show.

50:44 - Jeff Massie (Co-host)
I know what I'm sitting on makes sense and I'm currently at version 139, point random numbers. So next next update should bring it.

50:55 - Jonathan Bennett (Host)
Yeah, yeah, all right. So I've got one more story to cover and talk about, and this one is not directly Linux, but it is something that we cover on the show from time to time and I know a lot of people are interested in. And that is the big AI lawsuit is happening, or maybe is happening, and there's a couple of things here that I found very interesting. So this is a class action lawsuit and it is potentially going to be the largest copyright class action lawsuit that has ever been allowed to proceed action lawsuit that has ever been allowed to proceed, and it is quite a few authors versus an anthropic AI, and apparently they're trying to get up to 7 million people in claimants the technical term in this class action lawsuit, and so you know it's going to be anthropic AI versus these 7 million people. And the interesting thing about a class action that is that large is, if it goes to trial and the trial swings in favor of the, the people, then the potential for um, the potential for losses from it, is huge. Um, there there's one they loaded into the Anthropic AI, so hundreds of billions of dollars in potential damages. It's a lot, it's big. So when you talk about something this big. Most companies will settle if they can. What Anthropic is actually trying to do is to block this mega lawsuit and say, no, no, these all need to be handled individually, because when you have these individual lawsuits, there's less risk in taking one of them to trial. If Anthropic loses, they pay out a much smaller amount and then they can settle the rest, whereas if you do it all as one big block, just the numbers are ridiculously huge. It's company ending right. It would be the end of Anthropic if they got a judgment of this size.

53:35
There were a couple of other things that were really interesting here. One of them is that the judge that is overseeing this that Anthropic essentially criticized in their filings on this and some of the statements that they've made. It's Judge Alsup, and those of you that have been watching sort of the Linux world might remember Judge William Alsup. He was the guy that originally tried the Oracle v Google case over Android and, very famously, in the middle of a court session they were talking about the range check function. So this is in the Android case and the judge says I know what range check does it makes sure that the numbers you're inputting are within a range. I'm reading from the court documents now, and if they're not, they give it some sort of exceptional treatment. It is so that witness, when he said that a high school student would do this, is absolutely right, and so this was sort of the statement that people recognized from the middle of that legal battle, that courtroom drama. Wait a second. Judge Alsup actually knows how to code, and one of the things that you heard is he taught himself Java to be able to try the case, which is sort of true, but apparently he had been programming for decades in things like BASIC and various languages. And so, all that to say, judge Allsep knows how technology works. He is one of the ideal judges to have his hands on this particular case because he has a clue about how technology works. Some judges and lawyers even do not. He seems to. So that kind of made me feel good about this.

55:43
I have over on Floss Weekly. I have actually asked a lawyer about the AI question and we sort of spitballed well, what would happen if you had? And we described sort of a case like this where you had authors come up up against an AI company and say what you're doing training an AI off of our works is copyright infringement. And it was very interesting to me. The lawyer's comment was you can't put the genie back in the bottle. And that's really interesting here, um, because, like, even if let's say, worst case scenario, even if this kills this particular company, right? Even even if anthropic is ruled against and goes under, um, people are going to want their AIs on the internet, and there's I.

56:39
I have to agree with him. I think, no matter what the court rules, there's no stopping LLMs and training LLMs on this data. There's going to have to be maybe a little bit more care taken. There may be going to have to be sneakier about it. I'm not sure how all of that's going to wash out, but I have to agree, I do no matter what the courtroom says.

56:58
I do not see a future where we give up on training AIs on public data. It's just too powerful of a tool and too many people want it too badly. This is the sort of thing that you sometimes see new political parties and new political movements formed over times, see new political parties and new political movements formed over Right. So I could see even a political party come out of this, that is the AI party, and their stick is. We want to redefine how copyright works so that we can have working AI. Um, I it's. It is an incredible time that we are in Uh and I. I don't know what's going to happen with this particular court case, but it is so fascinating to watch.

57:43 - Jeff Massie (Co-host)
Well, there's so many nuances there because you know they say, well, you know, I get the, okay, I wrote this, I should have rights to it and somebody's using it. And then you hear, oh well, ai can just spit out you know parts of a book, but then I've also heard, well, to get it to do it, you got to like pre-feed it a couple you know paragraphs to to get the exact, which seems kind of funky. And I don't, you know, I don't know. And I and I agree copyright laws went too far. I mean, that was Disney that kept pushing that and they wanted it longer and longer until they realized, oh, we can trademark our stuff and then it doesn't have a time limit. So then they quit pushing copyright out for years and years and decades.

58:31
And I know there's been a lot of issues with Google trying to preserve old books, where they got to try to track down a lot of kind of orphaned books like who owns the copyright? And there's people that don't even realize they own the copyright to a book because somebody's great-great-uncle wrote it. It wasn't a huge success, it wasn't just kind of faded into history Google wants to preserve it, and then they knock on your door and go hey, can we preserve this book? You're the owner of it now through. You know legal rights of you, know they passed and didn't will it to anybody, but you're the surviving relative, so it's now yours and you know there's there's a lot of books that'll probably could die away and other other copyrighted materials music and whatnot, because who's the owner we don't know and it's on mediums that aren't going to last and I don't know. It's a mess.

59:26 - Ken McDonald (Co-host)
Mm-hmm.

59:28 - Jonathan Bennett (Host)
So the central question with AI here is is training an AI on these copyrighted works? Is it a transformative act? That's sort of the legal jargon way to put it. And I guess, if you were to try to boil that down, what sort of what you're asking is? Is training an AI similar enough to what happens when a human reads a book? Because we all we read these books. We remember parts of them, we sort of integrate parts of them into our thought process, is it? Is it essentially the same thing that's happening when an AI reads a book, or is an AI just copying the book into its database? Well, I would say that in actuality, it's somewhere between those two, right and then? So then you have the legal question of okay, how do you handle that in-between thing? It's new and we've never had a technology do this before.

01:00:28 - Ken McDonald (Co-host)
Is training in AI like training students in a school?

01:00:34 - Jonathan Bennett (Host)
That's the question. That's essentially the question. Yeah.

01:00:38 - Ken McDonald (Co-host)
Where the schools, when they give them a reading assignment, say there's copies available in the school library yeah yep, so you could argue the school purchased the book.

01:00:51 - Jonathan Bennett (Host)
They did well, so the. So the question is then the, the, the the act of reading the book is going to change the way that the student thinks. Is that a copyright that the book is now imprinted into the student's mind? Is memorization a copyright violation?

01:01:09 - Ken McDonald (Co-host)
Is plagiarism of it when you're writing that book report, a copyright violation?

01:01:15 - Jonathan Bennett (Host)
You know. It's interesting that you mentioned that, because there are some really well-defined rules about what makes something plagiarism, and you know it's not just you know you've lifted this many words in a row without putting in quotation marks, like that's one of the ways you can plagiarize. It's also possible to plagiarize by using the same like sentence structure, so you can take someone's sentence and put synonyms in for the words. Right, that is still plagiarism, technically speaking. It is still plagiarism, even if you've used a synonym for all the words. If you have the same thought in the same order, it is considered plagiarism, and so I guess that it would be an interesting test case like are ai's plagiarizing books? Maybe that is the better question to ask. Think of it that way Plagiarizing the authors.

01:02:07 - Jeff Massie (Co-host)
So all you need to do is they just need to make their AIs talk like Yoda, so it's not in the same order and then everything's okay.

01:02:16 - Jonathan Bennett (Host)
She doesn't work that way. Or to rearrange we will. There was a. There was a. I think somebody at the EFF published the right to read, that sort of ties into this, and it was a little fiction story about sort of this dystopian future where that that idea of copyright extends all the way to reading and so for someone to read a book they have to have a license to be able to read it. Um, and uh, it was. You know, it's a dystopian future. Warning for what a possible future?

01:02:57 - Jeff Massie (Co-host)
um, but I'm reminded of that in thinking about this AI question, ironically well, and you know, it gets more complicated too because, okay, you feed, you feed a book to an AI. It it takes the words and it tokenizes them and it, you know, it's doing a lot of analysis on them. So it's not like it just oh, I have a complete, unbiased copy. It does some kind of rearranging, I guess.

01:03:26 - Ken McDonald (Co-host)
Digitizing.

01:03:28 - Jonathan Bennett (Host)
It's much more than just rearranging or digitizing though. So like if you were to pull apart the source files for an AI, like the actual binary files, and look at those tokens, they are completely incomprehensible. It looks like just noise, right? You cannot. You cannot open the files of the ai and find the book. It's not like it's a database where they have just sucked the book in um, where they tokenize it. It's. It is completely incompetent because it's it's. It's been mapped to a neural net essentially, and so it's all weights. It's not the actual text anymore.

01:04:17 - Ken McDonald (Co-host)
Speaking of text, have either one of y'all studied what goes on when you do conversion from text to speech, or just generating speech on a computer, not in depth.

01:04:23 - Jonathan Bennett (Host)
I've played with it a little bit.

01:04:28 - Ken McDonald (Co-host)
I've played with it enough to understand that what you're getting that computer to do is just when you try to get it to say hello, for example, is combine those trying to think of the actual term, combine those trying to think of the actual term phonemes, I want to say together, so that it sounds like you've got that computer saying hello.

01:04:53 - Jonathan Bennett (Host)
I'm not sure that the modern LLM take on that is exactly that structured, though I kind of suspect that LLMs are doing it through the neural net way instead of structured, breaking it into phonemes.

01:05:07 - Jeff Massie (Co-host)
But I've not looked deeply into that. It's more of like ingesting in the patterns of the words, kind of. And then there's a lot of statistical analysis type stuff going on to how often it's you know a word like the shows up and it, it's not just storing it in a different format, it kind of it's all you know. You could say well, it's almost like a human mind where you're not going to get the exact same thing out and so. So it's not just like oh, there's a database of the book, it's just holding an exact copy. Well, they're going to have to get into like here's how we tokenize it, and all the manipulation that goes on. So the final product that is stored is kind of in a big pile with everything else.

01:05:56 - Ken McDonald (Co-host)
In other words, it's basically using its internal biases to determine how it ranges all that information internally.

01:06:10 - Jonathan Bennett (Host)
Essentially, I watched a video a couple of days ago where someone described a very simple single neuron neural net very simple, like single neuron neural net, and it was. It was the idea of you could, you could program this thing to identify, like l's versus e's, I think it was. It was something like that, and so it was a grid of like four by four um pixels that were either black or white, and then the controls. They had a grid of four by four knobs and there was this little algorithm where it's like, okay, you go and you twist the knobs by this amount for each of these different cases and then you, you know, you change what you're looking at and you do the same twisting action again, and so you end up with the knobs moved to like these. Some of them are moved to very extremes and some of them are like right in the middle and at the end of it they actually had a 100.

01:07:05
Now this again, it's a very simplified case, but they had a 100 percent um test case where this, this little single neuron neural, physical neural net, could identify the difference between an l and an E or J and an E. I think it was actually, and what it did is. It took those pixels and it derived weights, the settings for each of those knobs. But you couldn't look at the settings of those knobs and go, yeah, that one's an E. Or look at the pattern that the knobs made and go, yeah, I see the EV. It didn't work that way, and that is a vastly simplified explanation of what a modern neural net does. It's literally translating all this stuff. It ingests into these very opaque tokens and then weights that go along with them. Yeah, it's super interesting and I think we really do need some solid court rulings on this. Everybody that's been playing with LLMs and AI, like the companies making them, have been in this weird gray area for the last few years, and it's going to be very interesting to see what it looks like.

01:08:16 - Ken McDonald (Co-host)
So we've got some witch trials coming up.

01:08:19 - Jonathan Bennett (Host)
I mean any sufficiently advanced technology is in its signal over magic right, and I think we're definitely there right now with AIs and LLMs. I cannot look at the files and understand what it's doing, that's for sure.

01:08:32 - Jeff Massie (Co-host)
Maybe I can speak a little more intelligently once I read from Steve Gibson the Stephen Rolfram what is chat GPT doing and why does it work. Gibson the uh, stephen roll from what is chat gpt doing and why does it work?

01:08:46 - Jonathan Bennett (Host)
so I can at least better understand some of the concepts of uh ai, yeah, interesting stuff. Yeah, all right. Um, I'm sure that we will follow up on this story because it's it's. It's not linux, but it's in our wheelhouse somehow.

01:08:58 - Jeff Massie (Co-host)
It's definitely something I'm interested in well, and a lot of ai work is being done on linux. Even even before the show, we were talking about the new uh ai 395 processor from amd and it's now in a desktop format. Well, one of the big things that they have in there is some some chips to do some of the neuro the processing to run ai's. You can run ai's locally and things like that and uh, it, it's a. It's getting enabled a lot in linux because it's such a linux is a hot spot for ai. We'll just put it that way.

01:09:39 - Jonathan Bennett (Host)
Absolutely All right. So let's get into some command line tips, and one of the things that Ken brought to us was some not wire guard, wire plumber. Well, no, the other wire plumber is the control program. What's the actual daemon called Pipewire? Pipewire, thank you. I was looking for something that started with wire, and the only other one that was coming to mind was Wire Guard, and that's definitely not right.

01:10:10 - Ken McDonald (Co-host)
Definitely not Wire Shark.

01:10:12 - Jonathan Bennett (Host)
No, not Wire Shark either. Anyway, you brought Pipewire stuff to us, and I have a decent understanding of Pipewire. There is a companion program to that. That I understand hardly at all, and that is what you're starting to cover now, and so I'm actually pretty excited to learn more about wire plumber.

01:10:31 - Ken McDonald (Co-host)
Yes, this week I'm introducing the command WPCTL. I have no idea how to pronounce that as a single word, but this is a command that lets you control wire plumber objects and sittings from the command line. Let me go ahead and switch over to my terminal here so everybody can see and, as you can see, I've got WPCTL space dash, dash help typed in. I'm going to hit enter, space dash, dash help typed in, I'm going to hit enter and that tells you how you can use WPCTL Basically. You use it along with one of the commands that it's got listed below. I'm going to touch on using status this week.

01:11:20
Now, as you look down, it also says that you've got the options of dash H and, as I just used dash dash help to show your help options, now, as you look down, it also says that you've got the options of dash H and, as I just used dash dash help to show your help options, now you can also pass dash H after a command to see command-specific options.

01:11:31
So I'm going to go ahead and type wpctl status, dash H, and you'll now see it gives up that usage again. Then it gives the command we're asking for help with, which is status, and it displays the current state of objects in pipewire and you've got the general help options. But it also gives you application options, in other words the options you can use with status. To, for example, display device and node nicknames, you'd use dash K or dash dash nick, and to display device and node names instead of description, you'd use dash N or dash dash name. So I'm going to go ahead and use dash in now, and when I hit enter, it gives a whole slew of information about my pipe wire setup. It tells you what your pipe wire daemon's called, gives you the version. You notice it says 1.4.7 there, jonathan Mm-hmm gives you the version.

01:12:42
You notice it says 1.4.7 there, jonathan, and then it gives you the user at locallocaldomain in this case, and it also gives you a cookie colon followed by a number. I don't know if that's a file or why. They haven't had a chance to dig into that. But what's really interesting is that right below that, it gives you the clients that a pipe wire has and again it gives you the version, a user and a PID. So if you want to explore to see what those clients are doing, you can use that PID. That may be something we need to cover in the future is how to look at process IDs, because I don't think we've actually covered it, just how to kill them. I think that was back in episode 30 by you, jonathan.

01:13:34 - Jonathan Bennett (Host)
Yeah, I think so.

01:13:35 - Ken McDonald (Co-host)
Then, going on further down, it gives information about my audio devices, syncs sources, filters and streams, as well as video devices, syncs sources, filters and streams. And you notice it's got for those of y'all listening the streams has two streams from K-Wayland to Plasma Shell and then also showing from Plasma Shell to K-Wayland. One's an output to an input and the other one's an input to an output, so basically the two back-to-back. And then it shows your default configured devices that are in your settings. With my Tumbleweed I've got the virtual sync that I had set up, called MySync. Remember when we played with that a couple? What was that over a month ago? And then it's showing the audio source, which is alsa, underscore, input dot, pci.

01:14:38
Now here's what this looks like without that. Dash in to give me the names of everything and just looking at the bottom there, you'll see that the video device changed from a V4L2 virtual device to OBS virtual camera, as well as the audio devices changed to built-in audio, as well as the audio devices changed to built-in audio, analog and stereo. What's interesting is, for the audio devices, it has over to the right. It tells you that it's using ALSA, and for the syncs and sources, it gives you the volume that it's set to. There's some other commands that you can use with WPCTL for controlling that volume that I want to touch on in future episodes.

01:15:28 - Jonathan Bennett (Host)
Very interesting. All right, Jeff, what do you have?

01:15:34 - Jeff Massie (Co-host)
So this is not quite the normal command line tip, as it isn't a command that's new, but it's using ones that we've already talked about in the past in a new way. So you can use this to clear RAM memory cache, which is where the kernel keeps a lot of its often used data, so it doesn't have to load from disk. You can use it to clear buffer memory, which is a little different than the RAM memory cache in that it stores data being transferred between things like a CPU and a hard drive or SSD. So it's, you know, kind of lets storage system components and things like that run at different speeds and have a smooth intake and output of data. And we're even gonna talk about swap space to how you can clear out swap, which it's a section of disk that works like RAM. It's for when you run out of your regular memory and it can be used on the drive. So it's kind of like virtual memory. It's a lot slower but it keeps the system from crashing.

01:16:35
Now to clear the memory, but not the swap, so this is just the memories part of it Use the command sudo space sync. So that says flush all your drive buffers. So it writes everything in the buffers to the drive that, if it hasn't already, semicolon, which says okay, start another command, echo, and then either a one, two or three space greater than symbol to the slash proc, slash sys, slash vm, slash, drop, underscore caches. So where if you put in a one, it clears only the page cache. If you put in a 2 there, it clears the de-entries and inodes, and if you put in a 3, it clears the page cache, de-entries and inodes. Now if swap space needs to be cleared, you can use sudo space swap off space dash a and then you use sudo space swap off space dash A, and then you use sudo space swap on space dash A. So that will clear your swap space.

01:17:54
Now the article does go on to say that you know you really shouldn't be doing this unless you really know what you're doing. Because, for example, you could have a high machine load from, say, you have a lot of users, and then you clear the cache, but it's under high load and all of the information that was being pulled from the cache because it was being frequently utilized now has to come from your drive or drives and it could cause problems. You could possibly even have a crash because too much stuff is, all you know, log jammed up at once trying to trying to pull all this data from a drive which is extremely slow compared to um, your, your RAM cache. But you know there are times you could have a need based on what you previously ran or something like that. You've got to free up stuff for a new workload, things like that, but you don't want to reboot the machine. Rebooting the machine will clear everything as well, but sometimes you have to keep things up and this is a way to clear that stuff out while keeping it up.

01:19:00
Now, if you do have a need, take a look at the article in the show notes for a lot of details on what this is doing to your system, all the risks, the rewards. It goes into a lot bigger detail, a lot greater detail of the ins and outs of this. But yeah, you, you know, I didn't know that you can use that uh um drop underscore caches file to uh reset some of your uh buffers. So happy cleaning very cool.

01:19:34 - Jonathan Bennett (Host)
All right, I have got a quick one. It's a command that we've never touched on before and I was sort of surprised that we had not. It's TR, which stands for translate, and this is an old, old Unix command, one of the originals, I believe. I think it was around way back in the day, like a lot of the commands are that we use, and it just lets you translate characters in a string, and I needed this this week as a part of a, you know, a bash one liner.

01:20:08
I was taking some hexadecimal data that was in like a 0xAB space, 0x23 space, and like it was in uppercase, I needed it in lowercase. So I started looking around. It's like, okay, which of these tools is going to be able to convert from uppercase to lowercase? And I came across TR and you can just run this TR and then it's open quotes, open square brackets and colon, upper, colon, and then a space. Then, you know, colon, close your square bracket, close your quote and then a space. Then you, you know you, colon, close your square bracket, close your quote and then a space. And then another double quote, another open bracket, colon, lower colon, close your square bracket, another quote, and that will translate everything uppercase in the string to lowercase in the output.

01:21:01
Tr can do more than this, but that was what I used it for, and so the little one-liner I've got in the docs in the show notes here is to echo the word test with a capital T. Pipe it through TR with this invocation and it will output a lowercase test. And, yeah, it will convert it for you. You can convert it the other way just by swapping the two. And again, like I said, there's more that TR can do. Take a look at the link in the show notes if you would like to see a bit more that it can get up to.

01:21:33 - Ken McDonald (Co-host)
But another tool to have in your bash Swiss Army knife, your bash tool belt it's a great way when you need to do a conversion from upper to lower or lower to upper absolutely and Jeff put in our the discord chat I think it's also going out on the main chat, Elias that I was using back when I was on my Lenovo ThinkCenter because it only got four gigabytes of memory on it. Oh, okay, I'd run that alias just to clear out some of the swap files and all and memory.

01:22:23 - Jonathan Bennett (Host)
Yeah, absolutely. All right. It is time to get this. We're down to the end of the show. It is time to get the last word in, if the guys have anything. And Ken, you've got a little story, you've got a link here in your ending notes.

01:22:39 - Ken McDonald (Co-host)
What do you want to plug as we tell everybody bye came across this article by and I do apologize, I know I'm going to mispronounce this because it doesn't have that many vowels cy, vessar or vsr. And he wrote about repurposing old cd and dvd players. Uh, in with some interesting DIY hacks. In fact, the story covers how to convert it to a laser engraver, a CNC plotter and even a robotic arm. Interesting read.

01:23:14 - Jonathan Bennett (Host)
Yeah, absolutely All right, Jeff.

01:23:19 - Jeff Massie (Co-host)
I was going to talk a little bit about and I'll put a link in the show notes to a video that ran the Bonneville Salt Flats this last week and wound up setting a record with my motorcycle. So that was a lot of fun. I got to meet a lot of interesting people. The Salt Flats for those that maybe don't know of them or never never been around them, the Bonneville salt flats are where a lot of land speed records are held. Uh, many, many miles of flat salt. Uh, it's, it's kind of a hard surface with almost kitty litter, or if it's like you had a asphalt road and you just dumped a whole lot of table salt and rock salt on it. So traction is a lot of problems you can have, but you can get a lot more speed on asphalt. But we went out there, we ran, we went out there, you, we ran. So it's it's a three mile course that we run on and then you have a few miles to slow down after that and they take your average speed at mile two and mile three because we're on the short course, we're not fast enough to go on the long course. Uh, you do see, uh you know vehicles out there doing 450, 500 miles an hour. You know vehicles out there doing 450, 500 miles an hour. They're on a much longer course than we are. But uh, you, you take your best speed, the average over the whole mile, and then if you qualify for a record you have to come back the next day and do it again and say, say, your best speed was at mile three. On the second day you average the speed of mile three. You can take mile two and mile two, but you cannot take mile two and mile three, but you average the speeds together and that is what your record is. Our record for a 3000 CC motorcycle, modified, partially streamlined, is 154.506 miles an hour. So that was a lot of fun and got to meet a lot of interesting people. Had a it's, it's a triumph rocket 3 which was running. Met a couple from japan that could barely speak english but they had a 1953 triumph. So we kind of bonded over that and just through very broken English they described what they did to it and they asked about our motorcycle and it was really just really a fun time. We got a couple selfies with them, met a crew from Australia, talked to them for a bit, a reporter and his friend from London doing a story on the salt flats and talk to them, and, of course, people from all over the U? S. You know you, you name, you know, uh, well, I should say you at North America, you know Mexico, u S, canada, all over in there. So, yeah, it was a great time.

01:26:16
Yeah, link, in the show notes I will put a link to a buddy of mine who did a bunch of videos on it and he was actually the driver of the bike. He's a lot smaller than I am, so you kind of try to get the smallest, lightest person you can on him. But there's a whole video series Flat Cap Cafe Racer on YouTube and you've got a lot of videos of it out there and, uh, what it looks like to run down the salt and what we were doing and it it takes a team to do it and it was. It was a great time. So, uh, yeah, that that's my, that's my uh plug for the week. So, uh, everybody have a wonderful week.

01:26:57 - Jonathan Bennett (Host)
Yep, very cool, appreciate it. Thank you guys for being here. It's been a lot of fun. The one thing I want to plug is Hackaday. That is where you can find my other stuff. We've taken about a three-week break from Floss Weekly, but we are back this Tuesday. We've got somebody lined up for next Tuesday as well, and so the schedule is finally filling back up there. It's also where my security column goes live on Friday mornings, if you want to check that out.

01:27:24
And then the other thing to let you know about is Club Twit. Don't forget about that. If you love the show, you love the network and you want to support it, club Twit is the best way to do it. It's not much more than the price of a cup of coffee per month and it gives you access to the ad free shows behind the scenes and some other bonuses access to the ad-free shows behind the scenes and some other bonuses. Access to the Club Twit Discord as well, which is a lot of fun. So you should definitely take a look at that. Thank you, everyone that's here, both those that watched, listened live and on the download. We appreciate it and we will be back. We'll see you next week on the Untitled Linux Show.

01:27:56 - Leo Laporte (Announcement)
From Silicon Valley boardrooms to tomorrow's AI breakthroughs. If you need to keep up to date with tech, you need twittv. At Twit, we're tech experts who understand what's happening and can keep you in the loop, and we do it in a thoughtful, informed and fun way. Start your Sundays with this Week in Tech, a roundtable of tech journalists and people in the know with a rundown of the week's most important tech news. But that's not all.

01:28:25
All week long, you can stay ahead of security threats with Steve Gibson and Security Now. Keep up on all things Apple with MacBreak Weekly. Listen to the most informed Microsoft experts in the world with Paul Theriot and Richard Campbell on Windows Weekly. Our flagship shows feature tech's most respected voices, giving you the insight you need to understand and benefit from the changing world of technology. We give you analysis you won't find anywhere else, so you can make smarter business decisions and take advantage of the technologies transforming your world. Whether you're a CEO, it professional or simply passionate about technology, twit's network of shows gives you the edge you need in today's digital landscape. Don't miss a minute. Subscribe to TWITTV today.