Untitled Linux Show 251 Transcript

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

Jonathan Bennett [00:00:00]:
This week we are talking about some old bugs, some really old hardware that is about to get its support. Pulled a new NTFS driver, news about Fedora and Alma Linux, and then there's Mozilla Thunderbolt. Oh, and the Free Software foundation has entered the chat. You don't want to miss it, so stay tuned.

Rob Campbell [00:00:24]:
Podcasts you love from people you trust.

Jonathan Bennett [00:00:28]:
This is Twit. This is the Untitled Linux Show, Episode 251, recorded Saturday, April 18. Vibe, but verify. Hey folks, it's Saturday and you know what that means. It's time to geek out about Linux. It's the Untitled Linux Show. We're here to talk hardware and software, open source, all kinds of good stuff. Linux on the desktop, Linux in the server, Linux in the data center.

Jonathan Bennett [00:00:54]:
And I've got the regular panel of guys. We've got Ken, we've got Rob, we've got Jeff, not necessarily in that order. From left to right on your radio dial. And we're gonna have fun today. Welcome, guys. Hello, welcome.

Jeff Massie [00:01:08]:
And a lot of people went dial analog. What?

Jonathan Bennett [00:01:13]:
How many, how many people recognize that quote left to right on your radio dial?

Jeff Massie [00:01:18]:
I've heard it, but I don't.

Jonathan Bennett [00:01:19]:
There was old sportscaster, but I can't

Ken McDonald [00:01:23]:
remember who he said he quoting.

Rob Campbell [00:01:26]:
I hear nobody in the audience saying they recognize it. So

Jonathan Bennett [00:01:31]:
let's see who first said it. Kwood Leadford back in. Well, he was born in 1926.

Ken McDonald [00:01:44]:
I couldn't have said it before then.

Jonathan Bennett [00:01:46]:
Yeah, long time ago. Way back in the day of listening to sports on the radio am.

Jeff Massie [00:01:52]:
And it sounded like in a little tin can.

Jonathan Bennett [00:01:55]:
Yes, yes.

Jeff Massie [00:01:56]:
I can't do it, but Leo does it sometimes.

Jonathan Bennett [00:01:58]:
Back in the good old days. Yeah.

Ken McDonald [00:02:01]:
Anyway, I can remember when you used to sound like tin cans when you were talking over computers.

Jonathan Bennett [00:02:09]:
Yep, some days you still do. You still sound like a robot.

Rob Campbell [00:02:13]:
He means you specifically, Ken.

Jonathan Bennett [00:02:14]:
Oh, speaking of the good old days, we've got some good old bugs that are finally getting fixed. We actually got several stories about really old bugs that are getting fixed to let Rob kick it off to talk about Wayland and kde. And you wouldn't think that we'd be Talking about a 20 year old bug with Wayland, but here we are. How is that possible, Rob?

Rob Campbell [00:02:34]:
Oh, are we switching this around?

Jonathan Bennett [00:02:37]:
You're first. You've always been up first.

Rob Campbell [00:02:40]:
Oh yeah, that is the order I have it in. That's right. Yeah. So I do have a trend today.

Jonathan Bennett [00:02:46]:
So folks out there listening, we are in the market for a new cocoa

Rob Campbell [00:02:52]:
for folks out there listening. I did say before the show I'm tired, let's get this thing done with. So I am a little sleepy and I forgot I had reordered my show my my stuff on the show notes and did not reorder them on my notes. So I'm glad you reminded me which one to start with. But that being said, I do have a trend going on today and I'm probably going to upset some people but you know, the news is new so let's just get at it. So I am starting out with this 20 year old bug and I'm not the only one with a 20 year old bug to talk about today. But mine is, you know, personal to something I really like to talk about a lot and so to the story. You know, sometimes the biggest argument for new technology is not flashy features.

Ken McDonald [00:03:47]:
Sometimes.

Rob Campbell [00:03:47]:
Sometimes it is a bug report that sat around for nearly 20 years. Back in the day 2005, KDE users or a KDE user filled what sounds like a perfectly reasonable request. And the request is if you have multiple monitors, why can't each screen show a different virtual desktop? Why should every display be forced to switch together like one giant panel? It's the kind of thing people expect from a modern desktop, especially now that multi monitor subs are everywhere. Pretty common, but you know, modern features can be hard when you're stuck on decades old legacy code. And so that request stayed open for, well, 20 years, duplicate after duplicate after piled up from the same kind of same request pretty much. Not because KDE developers didn't care, not because nobody understood the problem, but because x11 itself was the wall. The old EWMH specifications simply did not have a real concept of multiple virtual desktops being active independently on different screens at the same time. In other words, this was not just a missing feature, it was a design limitation baked into that old stack.

Rob Campbell [00:05:23]:
And that is where Wayland changes everything. It took a modern display manager like Wayland and a PHP developer to fix the issue. To my favorite things PHP and you guys all know I love to talk about Wayland on here now, thanks to a merge request from Henik Shindenbuk, the feature has finally landed in K1 and is planned for Plasma 6.7. Each screen can now track its own virtual desktop independently. One monitor could stay on like a chat window, let's say, and and another can hold code, and then a third you could switch over to show documentation, all without everything flipping at once. It's how a multi monitor workflow should have worked all along. Or at least been optionally available. And if I recall last time I used Windows and played with their multi desktop, it's still like that.

Rob Campbell [00:06:27]:
Still works like the old KDE setup, you know, another so I guess that's one more leg up that Lennox now has against Windows. And Windows stole that feature from us anyway, so of course they're going to be behind. But maybe the best part is who made it happen. It's not a long term KDE veteran, not a giant corporate team, and as I hinted earlier, this work was done by a full time PHP programmer with minimal C background, no prior KDE contributions, and only a recent reason to care. He wanted to move to Wayland for fractional scaling, but this one missing feature was holding him back. Wayland is not only about replacing x11 because x11 is crusty and old. It is about finally making possible the things that X11 could never do cleanly, at least in the first place. Better scaling, better architecture, and now a better multi monitor desktop after almost 20 years.

Rob Campbell [00:07:39]:
The fix did not come from working around X11, but it came from moving beyond it.

Jonathan Bennett [00:07:49]:
Very cool. You know, I've not used virtual desktops for quite a while. It seems like this is one of the things I wanted. Like the moment I got multiple monitors, it's like this makes so much sense. I'm now playing with my KDE install here, getting multiple virtual desktops going again because it's cool.

Ken McDonald [00:08:09]:
I've always used multiple virtual desktops. It's just as of right now, my virtual desktops are synced so that whatever I have on one virtual desktop is on both monitors. But one thing you didn't mention, or I didn't hear you say, was about how the extended window manager handling kind of crippled x11 as far as providing this feature.

Rob Campbell [00:08:43]:
Yeah, I did not mention that, but go ahead if you got more to expand on that.

Jonathan Bennett [00:08:51]:
There you go. Not my problem. But you can pick it up.

Jeff Massie [00:08:56]:
Ken got segued

Jonathan Bennett [00:09:00]:
indeed. Is that the next story? Let's see.

Jeff Massie [00:09:04]:
Ken.

Ken McDonald [00:09:05]:
No, the next stories about how a 21 year old Polish woman fixed a 20 year old Linux bug.

Rob Campbell [00:09:18]:
And this is a different bug than I talked about. So lots of 20 year old bugs getting fixed out there.

Ken McDonald [00:09:23]:
This is in a window manager that is actually older than her.

Jonathan Bennett [00:09:30]:
All right, I'm going to do a little bit more looking into Plasma 6.7 because I have a question. I don't know the answer, but we're going to let Ken pick it up and talk about the bug fix.

Ken McDonald [00:09:40]:
Now before I go on, I want to thank Edmar Herrera and Brandon Virolo. And I know I messed those up real bad. I didn't take time to look how to say them. So I apologize. But they wrote about the accidental discovery and bug fix in a 20 year plus code base for the enlightenment window manager. E16. Interesting name, isn't it? Now, it was discovered, as I said, by a 21 year old graduate student, Camila. And this has a whole lot of consonants with one vow.

Ken McDonald [00:10:25]:
So another one I'm going to stumble over. No. Yeah, well, the Y could be considered a vowel, I guess. Zwick. I'm gonna go with that.

Jonathan Bennett [00:10:39]:
It's a. It's a Polish name. It's not pronounced anything the way that we Americans would think it would be. I don't remember how to pronounce these,

Jeff Massie [00:10:45]:
but it's like susky. Yeah, or something like that.

Ken McDonald [00:10:51]:
Vowels where you don't see them.

Jonathan Bennett [00:10:53]:
Yes, but anyways, it's the opposite of French.

Ken McDonald [00:10:59]:
Camellia, while doing some last minute work on a slide deck for a course she's teaching at Saarland University in Germany. Now, when she opened a PDF in the can, you remember using the Atrill document viewer, but her entire desktop just froze. Like you'll see my screen do sometimes when I talking. Now, according to Camellia, it happened repeatedly and Toshi sussed out the root cause. E16 hung whenever it attempted to truncate the overly long file name of that PDF. According to Edmar, this is the story of the true open source spirit. It's about a person, their computer, a frozen desktop, and the curiosity to figure out why it froze. Now, I'm not going to ruin the story by trying to retell it, so I'll recommend that you read this story about how Camellia isolated and fixed the problem in Edmar and Brandon's articles.

Jonathan Bennett [00:12:12]:
It's super interesting to see bugs this old get found and finally fixed. Like, I didn't know anybody was still running Enlightenment Desktop. It's definitely not new. I don't think it's been ported to Wayland. It's stuck on X11.

Rob Campbell [00:12:27]:
Right.

Ken McDonald [00:12:28]:
There was an upgrade to E17, but you've got some diehards out there that say they like the customization you can do with E16.

Jonathan Bennett [00:12:38]:
Yeah, makes sense. All right, so I did go and do a little bit of looking Plasma 6.7 is expected it mid June, like June 16, so a couple of months out still, we've got to wait a couple of months to get a hold of this. This new virtual desktop feature probably not going to come with Fedora 44, maybe with Fedora 45, I don't know.

Rob Campbell [00:13:01]:
44 keeps getting pushed back, so maybe not, not, not.

Jonathan Bennett [00:13:05]:
That's my story for later.

Rob Campbell [00:13:06]:
Don't steal it.

Jonathan Bennett [00:13:06]:
Don't steal my thunder.

Jeff Massie [00:13:10]:
Stop reading ahead.

Rob Campbell [00:13:12]:
I didn't know that was on there last night. Like you didn't have anything on there.

Jonathan Bennett [00:13:15]:
No, I picked some all right, we've got, we've got some kernel news. We're going to let Jeff pick it up and talk about ntfs. What's Microsoft doing in our kernel?

Ken McDonald [00:13:24]:
Break date first?

Jonathan Bennett [00:13:25]:
Yeah, we're going to take a quick break before that. We'll be right back.

Jeff Massie [00:13:30]:
We have a merge request for the 7.1 window for the new NTFS driver, and as a recent update it was accepted for the 7.1 kernel. Now before we talk more about this new code, let's have a little history. Now that NTFS is the file system for modern Windows, and while some will say we've had NTFS support for a long time, it's been a little patchworky. Now, for example, there's been a read only NTFS driver in the kernel for a long time. But wait, you say I've been writing files to Windows drives for quite a while. Yes, but that's a different driver which can do more things. But it's in user space. It's a fuse driver.

Jeff Massie [00:14:14]:
Now fuse is file system and user space. It's a user space file system framework consisting of a kernel module, a user space library and amount utility. Now because it's in user space, it allows secure non privileged, meaning non root privilege, mounting of drives and partitions. Now that being said, some will say no, I know it's in the kernel. Well, that's true as well because Sometime later, around 2020, the NTFS3 driver was upstreamed into the kernel by Paragon software. There was the later removal of the old NTFS driver, The read only one, because no one used it since the NTFS3 driver was so much better and you could write files versus just only reading things. Fast forward a few years ago to 2025 and now the NTFS plus driver has been announced. The goal was to clean things up and give more features and better performance.

Jeff Massie [00:15:16]:
Now the original announcement said that as of the last, you know, of the last few years, the NTFS3 driver, now this is the Paragon software one in the kernel has somewhat stagnated and not had a bunch of active development. A lot of the online sentiment is the kernel driver still has a few rough edges as well. Now keep in mind some are also still using the fuse driver. It has not gone away. So there's a little bit of when you say oh I have ntfs it could be one of multiple paths. Now the NTFS plus is a labor of love spearheaded by name me Gion. I should mention when I talk about making the driver better it's based off the old read only NTFS kernel driver. Now the developer said this about the reasoning behind that.

Jeff Massie [00:16:13]:
The remade NTFS called NTFS plus is an implementation that supports write and the essential requirements I o map, no buffer head utilities, XFS tests, test results and it's based on the read only classic NTFs. The old read only NTFS code is much cleaner with extensive comments, offers readability that makes understanding NTFS easier. This is why NTFS plus was developed on old read only NTFS base. The target is to provide current trends and he says IO map support, no buffer head public user space, utilities id, mapped mount support, delayed allocation and plans to support journaling. It's got enhanced performance, stable maintenance and utility support including fsck. Now to back this up there were some benchmarks and the benchmarks ran show a big speed up against the NTFS3 in multi threaded tests and minor improvements in single threaded tests. So the NTFS3 is the one currently in the kernel now that is stagnating. Now this has been a four year long effort and has been in the Linux Next tree for the last couple months trying to expose the code to more people to find bugs before it goes into the Linux main kernel and along with code review from some of the code experts.

Jeff Massie [00:17:40]:
If anyone is concerned for right now about the new driver, the NTFS3 driver is going to remain in the kernel for a while. I'm sure it'll be removed at some point in the future, but if the NTFS plus driver works out as good as it sounds, probably won't be that long before it goes away. Take a look at the articles linked in the show notes for more details and links to other past stories on the driver and the official kernel pull request for the driver.

Rob Campbell [00:18:07]:
Exciting stuff there. Thank you. Jonathan is muted.

Jonathan Bennett [00:18:12]:
Not anymore I'm not. I was like why did Rob just cut me off? Oh, I know why. Anyway, yeah, I've been doing NTFS stuff for a long time and I remember going through some of this. What's really fun is trying to read NTFS on a Mac is particularly terrible. Really makes me appreciate how good we have it on Linux as opposed to that hard to use terrible user interface operating system over there from Cupertino.

Rob Campbell [00:18:41]:
There's lots of things that are really terrible on the Mac, but that's really true.

Jeff Massie [00:18:47]:
Well, I was just kind of, you know, I started digging into this for the story because I knew there was a couple different drivers and then it was like, whoa, this is a little more complicated than I thought, you know, oh wait, you're using this one or this one. And this one went away and this. Now it's coming back and it's.

Jonathan Bennett [00:19:02]:
Yeah, yeah, absolutely.

Ken McDonald [00:19:05]:
Least we're not having to go all the way back to the original FAT file system that Microsoft came out.

Jonathan Bennett [00:19:13]:
Although I saw that Microsoft has released an update to FAT32 to where you can now you have bigger drives with it or something I saw in the passing this week. Anyway, no, we're not looking, we're not looking backwards, we're looking forwards with things like Fedora and Red Hat clones. I got a couple stories here for the Red Hat world and the first one up is that Fedora 44 fellow fans, we will have to wait yet another week because Fedora has returned to form and we are, we're doing a delay one week for the Fedora 44 release. It was going to be April 21, so about the time that most people would end up listening to this episode, it was going to be out. But there are some outstanding blockers. And so it has now been pushed to April 28th and that's things like non ASCII keyboard handling is broken. KDE Plasma has a problem with Plasma Setup's keyboard layout page. There's a problem with Butterfs as an installation drive.

Jonathan Bennett [00:20:19]:
There's another install issue and there's two more blocker bugs that have been proposed. So problems that people are saying, look, this should be considered a blocker. Quite a few, actually quite a few things to try to sort out in the next week. That's enough. It wouldn't terribly surprise me if Fedora 44 got pushed beyond April 28th. We may be talking about a two week delay. We'll see. We'll see.

Jonathan Bennett [00:20:45]:
Obviously we'll update everybody on this next week and it's pretty typical that you can actually do an upgrade like the Fedora 44. ISOs will be available for a day or two before the official release date. And anybody on Fedora 43, obviously you could go ahead and do the upgrade to Fedora 44 and things are not going to break. There is as always that feature freeze and then as soon as the full release goes out, then all of those features sort of uncap and you Get a whole bunch of updates all at once. And that's always kind of an exciting and fun time time that does sometimes break things as well. Anyway, that's Fedora 44. Look a few months further into the future and you've got Fedora 45. Lots of unintentional alliteration going on here.

Jonathan Bennett [00:21:39]:
Fedora 45 is considering an X8664 V3 build. Now, from what I'm reading here, it is not. Of course, I suppose we should take a quick break and cover. What this is, is this is the processor versioning. I think it originated in the kernel, but X8664 has been around for a long, long time. And there are multiple groups of extensions that have been added onto x8664. And if you compile targeting some of those extensions, certain things run faster, some things run a lot faster. You've got various distros have started targeting different run levels, different versions for, you know, for their processor ISA.

Jonathan Bennett [00:22:30]:
Fedora 45, one of the proposals is to do a new primary architecture of x8664v3 alongside the baseline x8664 architecture. And so if you are running a modern CPU, you would want to run those v3 packages and then old machines can run the old original baseline packages. And this gets you things like support for AVX, AVX2, FMA, other features like that. It does not include AVX512 as that's considered a V4 feature. And what I'm not entirely certain about is how these two things would be supported together. Like can you take an X8664V1 install and say now go install a bunch of the V3 packages? That would be ideal. But reading this, I sort of suspect that it's going to have to be a new install or a very sketchy process to replace all of your packages. There might be a way to do it, but that sounds sketchy.

Jonathan Bennett [00:23:48]:
So that's Fedora. One more thing in this same ballpark is Almalinux, one of the red hat clones, Almalinux Kitten Tin, which that's the downstream of centos stream. They are going the other direction. They are adding back the i686 user space. And so that's the ability to go back and add 32 bit packages to a 64 bit install for things old programs that require a 32 bit user space essentially to be able to run. And so that makes a lot of sense in the enterprise world that you might want to do that. Might need to be able to do that. So lots of things going on in the Fedora slash Red Hat world these days.

Rob Campbell [00:24:39]:
I would like to add a little disclaimer to your story there. When you said, basically when you told people to go ahead, upgrade to Fedora 44, things are not going to break. Now the little disclaimer here is, Jonathan cannot be held responsible for you breaking your system if you choose to update before the final release.

Jonathan Bennett [00:24:56]:
I think I said things are probably not going to break.

Rob Campbell [00:24:59]:
I would not say probably.

Jonathan Bennett [00:25:00]:
Okay. Any system upgrade can break your machine. I have had to rebuild my machine multiple times going from one version of Fedora to the next.

Rob Campbell [00:25:10]:
Why? Did you guys know? Yeah, do not hold responsible.

Jeff Massie [00:25:16]:
And on the version numbering, just so people are aware, V3 that basically is assumed to match the Intel Haswell architecture from 2013. So it is not a cutting edge thing. Version four is Skylake, which is 2017, as you said, AVX512 for example. And you can put those in. You don't have to. You could do one package at a time. You can mix and match. It's just that if you have everything compiled for a higher version number, you get greater benefit because you get that little fractional speed up on your libraries, your kernel, your everything, versus only one little piece.

Jonathan Bennett [00:26:07]:
Yeah, when available. So I'm reading the actual Fedora project proposal. This is proposed, it's not accepted, although it's the sort of thing that Fedora is going to love to do.

Jeff Massie [00:26:18]:
Ubuntu was looking at this as well.

Jonathan Bennett [00:26:20]:
Yeah. The proposal here says when available, DNF will select packages with the highest compatible baseline. So that does suggest that it's not going to have to be an entire system re reinstalled to get it going.

Jeff Massie [00:26:33]:
Yeah. For example, cache has different package servers that you're pulling off of based on your version level. So not everything is. I'm running V4 right now because I have a 7000 series processor from AMD, so it supports V4. But not everything is V4, it's just certain repositories are.

Jonathan Bennett [00:26:58]:
Makes sense. Yeah.

Ken McDonald [00:27:00]:
And it looks like there's still this guess, you know, whether or not it's worth doing because within the last day or so one person came in and said that he was seeing worse performance after building with the GCC March x86 64v3.

Jonathan Bennett [00:27:22]:
That's really surprising. I know some of the GCC options targeting like specific processors is underwhelming, but I've always seen like the v1, v2, v3, v4 as being fairly effective. That's really surprising that they would see a performance regression.

Rob Campbell [00:27:44]:
It's one user.

Jeff Massie [00:27:46]:
Well, I would say it's either a mistake or something or they found a bug or some really weird corner case because I've always seen the, the higher levels at least the same speed.

Jonathan Bennett [00:27:59]:
Yeah.

Ken McDonald [00:28:00]:
Not regret or is it GCC that's maybe being a bit too aggressive and using some of those vector instructions.

Jonathan Bennett [00:28:10]:
I mean there are. There, there could be some case again though like Jeff says, corner cases where using the vector instructions would be slower. But that's just, that's just odd.

Jeff Massie [00:28:20]:
Unless they, unless they were optimizing it as well because the high levels in GCC of optimization get kind of wonky and they actually don't recommend the. I can't remember all the different levels but like oh, three is not recommended. Yeah, yeah. Two is about as high as they say to go. And the three starts. Your results will vary.

Ken McDonald [00:28:42]:
So don't put a thousand.

Jonathan Bennett [00:28:44]:
I don't think that's supported. I think it's going to error out. I think three may be the highest. Anyway, let's talk about some old, some more old stuff. Rob, for whatever reason has the. He's got the old stuff going today. Are you. You doing.

Jonathan Bennett [00:29:00]:
You're doing some. Let's see, software antiquing.

Rob Campbell [00:29:03]:
Rob, I definitely have a trend. All my stories have something related to the other in one way or mostly one way.

Ken McDonald [00:29:15]:
But are we taking a trip down history?

Rob Campbell [00:29:21]:
I mean in some ways, I mean it's, it's, it's, it's. It's happening today, but it's something from a long time ago. I don't know. You'll find out. Just sit back, relax and listen to the words I had to say. I'll try to speak them as calm as possible, even if they upset you and me. So the X.org server is back in the news again and as is typical with Xorg these days, it's not for a good reason. A fresh batch of five security vulnerabilities has just been disclosed in X.org server 21.22.

Rob Campbell [00:30:04]:
And another common these with the XORG vulnerabilities. You know, some, some of these bugs are ancient. We're talking code paths that go back 25 years to x11 R6.6 and that's, that's almost as old as Jonathan. Alongside newer flaws introduced in later X.org releases, the list includes an integer overflow out of bound reads, a use after free bug and even a buffer overflow. So you know all the all the, all the bad things, most of these can be exposed on, in uninitialized memory, which is exactly the kind of thing you don't want sitting around in one of the, one of the oldest and most privileged parts of the Linux desktop stack. And that is, you know, really the story here is that xorg is old. It's not classic, not battle test. I mean it is, but it's, it's just full of old, huge bloated code based, decades of baggage designed at a time when security and the Internet was barely an afterthought.

Rob Campbell [00:31:25]:
Nobody worried about those things back there. It was a simpler time. In fact, most new X.org releases now seem to arrive because another security problem has been found. So if there's a new X.org update, it's probably a new bug. Thirteen years ago a researcher called X.org Security a disaster and said it was worse than it looked. And here we are, more than a decade later, still finding new reasons to agree. Don't get me wrong, it's still better and safer, more secure than Windows. But we have a much better option today.

Rob Campbell [00:32:07]:
This is another reminder that the future needs to be Wayland. Not just the future, but the present for most of us. Okay, there's some, some weird little side cases, but it doesn't apply to most people, honestly, you know, not because Wayland is perfect and not because X Wayland magically disappears overnight. You know, you're still gonna have that shim. But because the old X model was designed for a different era and we gotta move away from it. Wayland moves Linux desktops towards a simpler, more modern architecture with way less legacy exposure and fewer opportunities for this kind of decades old complexity to keep biting us. It's not perfect, but in so many ways it is now better than X. So yes, if you are still using it, patch x dot org, patch your ex Wayland.

Rob Campbell [00:33:05]:
But more importantly, try to find a way to keep moving forward and move on. Because Every new old X.org vulnerability is another argument that Linux shouldn't keep building tomorrow's desktop on yesterday's foundation.

Jonathan Bennett [00:33:25]:
Yesterday's foundation and all the experts that

Jeff Massie [00:33:29]:
worked on X11 are the ones that started Wayland. So the people that knew all the mostly understood X11. I hear it's just such a convoluted mess. Nobody truly understands all parts of it that they said to make it right. We got to start over.

Rob Campbell [00:33:49]:
If you trusted the X11 devs, you should trust the Wayland devs. Myself as a hobbyist programmer, I know what it's like to have old code that you wrote early on in your careers and you built it up as like, you know what, I just need to start this over from scratch. I did that with the program that I sold once, you know, I, and, and I, you know, I just added two to mine. I didn't give it a whole new name. I just called it LC2 because the first was LC. But anyway, you know, I know that it goes, it's, it's no knock on, on the old code, you know, it was in a simpler time.

Jeff Massie [00:34:26]:
Well, it was designed with. When the computer architectures worked very differently than they do now.

Rob Campbell [00:34:32]:
Yeah, and we've just been shimming it and shimming it over the years. And by we I mean they, the developers. I haven't done anything but.

Ken McDonald [00:34:40]:
Except report on it. Speaking of reporting on it, really shouldn't you say X security? It's worse than it looks

Jonathan Bennett [00:34:53]:
some days

Ken McDonald [00:34:56]:
Because I came across a video that puts it that way and it's by one of the.

Jonathan Bennett [00:35:10]:
Wasn't there news about they were just gonna stop doing releases of X altogether? Isn't that where like X Libre came from?

Rob Campbell [00:35:21]:
Something

Ken McDonald [00:35:23]:
stop make adding features to it or stop doing security bugs or stop

Jonathan Bennett [00:35:28]:
doing security really stop doing releases period. Even there are security problems

Ken McDonald [00:35:34]:
and those people that depend on X servers just get left behind.

Jonathan Bennett [00:35:41]:
I mean that is eventually going to happen one way or the other.

Jeff Massie [00:35:44]:
Yeah, it will. But you know a lot of the stuff we talk about, you know all these servers and all this stuff industrial and they're not running the latest greatest things, they're running some older, very tested version of something and they usually don't upgrade unless they absolutely have to because

Ken McDonald [00:36:04]:
it's such a pain to go to that thing that's not even networked to update it.

Jonathan Bennett [00:36:10]:
Yeah, well, I mean some, some of

Jeff Massie [00:36:12]:
that stuff that's not networked has drivers and stuff for custom hardware that nobody's ever gonna update so they don't change it.

Jonathan Bennett [00:36:22]:
You could, you could make the argument that it's, it's more of a security problem to plug one of those things into the network to be able to get the update than it is to just leave it off and running the old thing.

Jeff Massie [00:36:33]:
And a lot of that old stuff doesn't need to be on the network. It's. It's performing a function that is not Internet related. So. Well, it's not secure. Well, it's behind locked doors, not on the Internet and just chugging away quietly.

Ken McDonald [00:36:48]:
You have management introduce security issues by asking you to let's collect some of this data so we can analyze it from this device.

Jonathan Bennett [00:36:59]:
So when you tell them no, it can't be plugged into the network.

Jeff Massie [00:37:03]:
Well, usually when that happens, they know it's insecure and so there's a ton of firewalls and monitor. I mean, it's pretty locked off of being able to just get anywhere and you can't get it on the Internet with it. And a lot of times organizations, you can't even get to it from all parts of the organization. It's its own separate little bubble.

Ken McDonald [00:37:27]:
It's in this room, in this, that's down this hallway of this remote building. Just doing its thing.

Jeff Massie [00:37:36]:
No, just more like network wise, you can't, you know, it's very locked down on, you know, you can't just throw a USB drive in it, for example. You can't from out, you know, you can't get to it from anywhere in the company. You have to be in certain locations to be able to log into it.

Ken McDonald [00:37:56]:
Because like in the same room in some cases.

Jeff Massie [00:38:00]:
Yeah, in specially protected little sub networks that. Because they realize there's a risk to a lot of this stuff because a lot of industrial stuff runs very old software. So they don't just keep track of, well, this security issue. This security issue, they just assume it's horrible. So they build that shell around it network wise.

Ken McDonald [00:38:23]:
But can we add AI to it?

Jonathan Bennett [00:38:28]:
Yeah, they don't ask that question on that stuff or if they do, they get laughed out of the room. All right, let's move into containerization. We're going to let Ken bring a story about that after another quick break.

Jeff Massie [00:38:42]:
We'll be right back.

Ken McDonald [00:38:44]:
You'll be happy to know that this week Bobby Borisov wrote about a tool that can make working with containers easier. It's called Container.

Jonathan Bennett [00:38:55]:
That's with a K. Container with a

Ken McDonald [00:38:57]:
K. Now, Container is a relatively new graphical front end for distrobox aimed at KDE plasma users who want to manage containers from a desktop interface rather than relying entirely on terminal commands. Where's the fun in doing it on the terminal Built with qt, QML and Kirigami, you can create, delete, clone, start, stop and reboot containers, open terminal sessions inside them using your preferred terminal emulator and install package files directly with automatic package manager detection. Container does not replace distrobox, but adds a graphical management layer for users who prefer to manage those environments from a KDE application rather than the shell. Now, if you are interested in a container, then check out Bobby's article on how to install it.

Jonathan Bennett [00:40:05]:
Yeah, so this is, this is a GUI for distrobox, which. That's the application that lets you run essentially not a virtualized, but a containerized other distro really easily use any Linux distro inside your terminal. That's what distrobox claims.

Rob Campbell [00:40:26]:
Yeah, I've heard of it. I've never played with or used it at all. I feel like one of the container distros actually uses it. Maybe Cubes does. I don't know.

Ken McDonald [00:40:38]:
Could be Zorin.

Jeff Massie [00:40:41]:
No, Andrew.

Jonathan Bennett [00:40:43]:
This is interesting though. So the picture of the UI that they've got, it's distrobox containers and it's just. It lists out archbox, Alpine box, Debian box, Ubuntu box, Fedora box, and in each of those you can install packages, you can reboot the container, you can clone the container. It's like a very lightweight version of virtualization. I've never played with this. I need to. It looks pretty cool. Get your Ubuntu on without actually having to install Ubuntu, I guess your Fedora

Ken McDonald [00:41:16]:
home without having to actually install Fedora.

Jonathan Bennett [00:41:20]:
Why would you not want to install Fedora?

Ken McDonald [00:41:22]:
Because you're running open suits.

Rob Campbell [00:41:24]:
Tumbleweed Cubes doesn't. Doesn't use. I guess those are virtual machines, but maybe. Maybe nobody does that at all.

Jonathan Bennett [00:41:33]:
All right, that is cool. Let's talk about Intel. Jeff has the story about the Intel Core Ultra 7 270K Plus. What a mouthful. Hopefully the performance lives up to the name.

Jeff Massie [00:41:54]:
Yeah, I can't say enough about it to overcome that name, but Michael Larable over at Phronix got his hands on an Intel Core Ultra 7 270K Plus. Now, this CPU, if you know, was released a few weeks ago and there's been a lot of buzz about it, it's a refresh of Arrow Lake. So this is not an entirely new design. And it's based on the older 200 series of CPUs from Intel, which. Those were the successor to the 14K series of processors. You know, 14700K, 14900K, so on. While the CPU is not going to be faster than the top of the line AMD CPUs, one place that it's been, has been getting a lot of attention for the refresh is the pricing. Now, it features 24 cores and it's between 8P cores and 16E cores.

Jeff Massie [00:42:44]:
It's got a 5.5 GHz turbo TBMT 3.0 frequency. And TBMT is Turbo Boost Max technology, which means due to production Differences. Processor cores vary in maximum potential frequency. So TBMT identifies up to it could only be one of the fastest cores on your cpu, known as favored cores. Then it applies a frequency boost to those cores and directs critical workloads to them. So it's kind of internal. Choosing the best clocking cores is what it comes down to. Now that being said, it has a 3.7 GHz P P core base frequency, 36 megabytes Intel Smart Cache and 125 watt base power rating.

Jeff Massie [00:43:37]:
With 250 watt turbo power rating, you know the maximum. Intel launched the Ultra Core 7 270K plus with a recommended customer price of 289 United States dollars. But as of this writing, all the major online listings have it for 349 again US dollars. They've been in and out of stock and kind of why the price increase as well. They either didn't make a lot of them or they're selling a lot of them. My guess is it's kind of a mixture of both. I don't think they thought they were going to sell a ton, but they've been doing better than they thought. So they can.

Jeff Massie [00:44:18]:
I don't know whether it's intel or the retailers have upped the price a little bit just because they've been doing pretty well. Now the hardware heads among us will point out that the P and E core configuration is the same as the 285k but it doesn't have the 200 MHz faster clock speed of the 285k on the performance cores. But the 270k plus does have 100 MHz faster E core speed. So lose one, gain another. As of right now, there's no 285k plus and from what I hear online, there won't be one. The refresh does have an advantage that it can support DDR5 7200 mega transfers a second, while the original 200 series only went up to 6400 mega transfers a second. Now, the reason that this benchmark is later than most of the normal tech outlets is because Michael didn't get his CPUs until after the embargo lifted. So he was behind on getting, you know, being able to benchmark, getting his benchmarks done.

Jeff Massie [00:45:25]:
He said his CPU shipped after the embargo lifted, so he was not in the first round of getting the CPUs. The review CPUs. Now the benchmarking includes that Michael, that Michael ran includes several other aerolake non refreshed CPUs so the first first gen there. But there are no previous generation intel processors, so there's no 1400 KS or any of that in there. But they do have several AMD CPUs, but again, only the current generation is included. So just the 9000 series. There's no other AMD CPUs in there. Now he did note that he has not got a 265k refresh cpu to test, even though it was released about the same time.

Jeff Massie [00:46:16]:
I don't remember if they released the same day or within the week or they were very close. Now, something of interest, he does test the 275k plus with both the older 6400 mega transfers a second memory speed and compares it versus the new 7200 mega transfers a second. So memory speed alone differences can be evaluated. Okay, enough of all that. What were the results of the 340 benchmarks that Michael ran using Ubuntu 2604 and the 7.0 kernel? The memory speed increase was overall worth about 1% on average. So you can look through a lot of those different benchmarks. And while the 74, the 7200 speed came out on top, it was just neck and neck with the 6400. So you know, unless you have some laying around that you happen to have that'll go to 7200, I would not even bother with it because remember, normally at 10%, you know, kind of the rule of thumb, 10% increase is where people see a difference.

Jeff Massie [00:47:26]:
The 270 plus, 270K plus was about 2% faster than the 285K which is currently priced about $200, again US dollars more than the 270K plus. So the refresh, it's cheaper and faster. Now the geometric mean shows that the refresh was giving about 84% of the performance of an AMD 9550X for about 67% of the performance price. So bang for the buck. The 270k plus is a great CPU and it averaged power wise about 154 to 157 watts in power. I will add one caveat to these results though. If you have a single use or a few heavy CPU workloads you use, I would look at the benchmarks because while a lot of things intel was, you know, besting or right in there with amd, you know, right in the pack, there were a few workloads which they were at the bottom of the charts and there was some major step in performance difference. If you're a general user, I wouldn't worry about it because those cases where intel was way behind was so very specific that they're probably, they're probably almost not going to be a factor in your use cases or it's going to be so small you'll never notice it.

Jeff Massie [00:48:53]:
You know, personally, if I was in the market for a cpu, I would probably pick up a 270k plus great value for the money and a lot cheaper for the performance a person gets. Now do keep in mind there's going to be limited life in the motherboards because the next generation is rumored to require a different socket. But there's also rumors out there that intel is going to try and get more life out of their sockets. Time will tell, but they see what AMD is doing and I'm sure the motherboard companies are really pushing for that as well. You know, it just makes it simpler to make unless they are making enough on the redesigns and re spinning new boards that they're seeing that as a cash cow. But I said time will tell, we'll see what happens. And I'm just going to throw this in here. There's one final rumor that I want to squash.

Jeff Massie [00:49:47]:
It's been said that some motherboard companies are leaving the intel motherboard market and that is not true. While it wasn't worded well, you could say terribly, what some companies were saying is that their current products meet the needs of the Arrow Lake refresh and there isn't a need to develop a new product. So basically they're not leaving intel, they're just not spinning a new board for this refresh because their current 800 series chipsets work just fine with the refresh. So there's no reason to put the R and D and the testing in for something that is going to be replaced not that long into the future. I think end of this year, early next year is when the new CPU should be out. So why, why sink all that R and D cost when you're, when your current product works just fine? So no, no, nobody's leaving intel motherboard wise as always. Take a look at the article in the show notes for full details and links to the full suite of benchmarks and happy shopping.

Jonathan Bennett [00:50:52]:
Yeah, interesting. Intel is still shipping an IGPU inside of all their CPUs, right?

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

Jonathan Bennett [00:51:00]:
Yeah. So you could, you could do a, a new machine build if you really wanted to with one of these things. Don't have to do a discrete GPU and get you into something a lot cheaper than Maybe otherwise would it's not going to be a great gaming machine with just the igpu, but.

Jeff Massie [00:51:18]:
True, but I mean, you know, maybe, maybe you play much slower paced games and maybe, you know, you're not running a real high res monitor or something.

Rob Campbell [00:51:27]:
Solitaire and meijiang.

Jeff Massie [00:51:30]:
Well, maybe you're playing, you know, team fortress. Yeah, team fortress. Or you're paying counter strike. Or you're, and you're not competitive, you're just having fun, you know, but then

Rob Campbell [00:51:44]:
you

Jeff Massie [00:51:46]:
throw, you know, throw later, throw a GPU in there and away you go. I, I was really impressed with the price for performance it, you know, and I would seriously look at it because I've never had a top end cpu. I've always stepped down a little bit and I'll say I've got AMD now, but I would buy Intel. I favor whoever gives me the best value for my money.

Rob Campbell [00:52:13]:
The downside is you're still got to put memory in there. And from what I saw today, I saw a meme today that memory is the new bitcoin. Oh, I wish I would have invested earlier.

Jonathan Bennett [00:52:26]:
Yeah, no joke.

Rob Campbell [00:52:28]:
I got a couple here. Wait, I do, I do. Look at this. I got a cup.

Jonathan Bennett [00:52:32]:
Oh, excuse me.

Jeff Massie [00:52:38]:
Wow, Rob was so excited he got a spit take out of it.

Jonathan Bennett [00:52:41]:
Yeah. All right, well, you know, what is the new bitcoin? It's not memory, but there is something else that is the new cryptocurrency. It's AI. There's hardly a week can go by that we're not talking about some new AI something or other. And this week it is of course Mozilla getting into the game. Mozilla. They have introduced or at least announced Mozilla Thunderbolt. It is not a good name.

Jonathan Bennett [00:53:14]:
Not confused with Thunderbird or Intel Thunderbolt. Yeah, it's not a good name. Not good at all. But Mozilla is announced Thunderbolt and it is a new AI platform and you can run your agentic AI locally. With Mozilla Thunderbolt. It is intended to be a sovereign AI client. It is of course open source, extensible, all of those good things. And it can be used to access AI chatbots for things like research and data analysis and all of the other things that you do with enterprise AI tools.

Jonathan Bennett [00:53:54]:
Yes, this is Mozilla trying to be enterprise AI. It can connect to a MCP server and it can connect to agent client protocol compatible agents. And it also integrates with the Haystack AI platform. I, I get sort of what they're trying to do and in a way this is intended to compete With Copilot on GitHub in some ways, maybe Copilot in VS Code as well. And of course, we're in a moment right now. The zeitgeist is that if you're Enterprise AI people will apparently throw money at you. You. So I.

Jonathan Bennett [00:54:46]:
I get why Mozilla wanted to do this, but I really wish that they would just make a good browser and a good email client and call it a day.

Rob Campbell [00:54:57]:
That's what we've said for years.

Jonathan Bennett [00:54:58]:
That's what we've said for a long time. I don't know, maybe Thunderbolt is going to be the next huge big thing. I am less of an AI skeptic than I used to be. I think it is something that's going to stick around in one form or another. It's just too insanely useful to not. But I don't know that Thunderbolt is going to be the ticket to Mozilla's success.

Rob Campbell [00:55:22]:
I mean, the name, come on. We already have a Thunderbolt in technology

Jonathan Bennett [00:55:27]:
and we already have a Mozilla Thunder Bird.

Rob Campbell [00:55:31]:
I mean, I think that's why they did like, we got Thunderbird, we got Thunderbolt, and now they're gonna thunder something else. They're gonna change their browser to be Thunder.

Jonathan Bennett [00:55:38]:
Thunder Browser.

Ken McDonald [00:55:40]:
No, Lightning.

Jeff Massie [00:55:42]:
Personally, I use Firefox all the time and I use Mozilla products, but I won't know because as soon as they came out with that AI kill switch, I just flipped that baby on and I don't want any AI in there.

Jonathan Bennett [00:55:56]:
That's interesting too, because you actually use AI quite a bit for doing mostly work tasks, right?

Jeff Massie [00:56:01]:
I do, but my problem is I'm like the old man yelling at the clouds. When I want AI, I want it to work correctly. And all these little filler AIs and all this, they're just solutions looking for a problem. Oh, we're going to help you do this. No, you're getting in the way and you're just bogging me down. When I need AI, I will reach out to AII and it will do a great job. And most of the time.

Rob Campbell [00:56:29]:
I was talking to somebody this past week, wondering what? When graduation is for a specific school and I was in use an edge. So bing, the Bing AI response at the top told me, Friday, May 20th. And. And I told that to the people. People I was talking to. It's like, oh, yeah, it's Friday, May 20th. And they're like, Friday, May 20th.

Ken McDonald [00:56:56]:
My.

Rob Campbell [00:56:56]:
Yeah, that's what it says. And then I looked at the calendar. And May 20th is on a Wednesday. So that is what it said right at the top.

Jonathan Bennett [00:57:05]:
I need to wash my car. It's about a quarter of a mile away. Would it make more sense to walk there or drive? Well, it depends. If you could use the exercise, then you should walk there. This is the new AI Turing Test. This is the question.

Jeff Massie [00:57:27]:
Well, and that's why, you know, even though I use it quite a bit for my job, you always, buddy, check it. Because sometimes, you know, it'll do great things and sometimes it has a brain fart, you know, it has a stroke and just takes off in a direction. And I'm not talking about. It pulls like a single Reddit article or some weird thing where it decides that's a fact. I'm talking about you have a chart of data and it. You maybe have four columns and it'll grab the first two columns on row one, and then it grabs the second two columns off row five or something and tries to put them together and it's like it was right here and you screwed that up, you know?

Rob Campbell [00:58:13]:
Yeah, yeah, I've asked some questions where it just pulls things from multiple things and then. And puts them together to give me a completely wrong answer. Yeah, they talked about that on that page, but that response had nothing to do with it.

Jonathan Bennett [00:58:28]:
Yeah, AI is the most useful when there is an immediate way to concretely check that it is correct. Which is one of the reasons why it is so useful in Vibe code style programming. Because there's a button you press to compile the program, and if it doesn't compile, you immediately know that the AI messed it up and you write test code. You have the AI write test code and you can immediately know if it's messed up. It's one of the reasons why coding is such a. It's such a killer app for AI is because there's immediate feedback and you can, you, you can very easily know if there is a fundamental bug or not.

Rob Campbell [00:59:09]:
Well, yeah, well, and a lot of the AI, chatbots, bots or whatever, when you ask it something like if I would have been actually in one instead of right on the search page, it would have given me the sources. So I would have clicked on the source to see where it got that from. And actually it did, it did give me the source. I went to the school's calendar, which was in there. I don't know how it got that answer from the source, but the source was there.

Jeff Massie [00:59:34]:
Yeah, the randomness prediction.

Jonathan Bennett [00:59:39]:
That's funny. Yeah. I spent some time this past week, week and a half, doing my first. I shouldn't. I've been told I shouldn't refer to this as Vibe coding because I'm not a Vibe coder. I actually check the output of it. I actually understand the output of what the AI programs. But doing my first big programming task using an AI chatbot, and that is in meshtastic, we now have.

Jonathan Bennett [01:00:02]:
Well, it's pull request still, but we have the equivalent of SSH over the mesh to be able to get a remote shell on a computer. And, man, I was able to get the, like, initial working version of it in just a few hours of work, as opposed to something that would have taken me several days without the use of AI. So, like, I'm more of a believer now. I used to be a little bit more skeptical. I'm a little less skeptical. But I also have had to go in and clean its code. We talked about in our discord, we talked about that it likes to make. It likes to make lasagna out of code sometimes.

Jonathan Bennett [01:00:41]:
In this case, it was multiple functions, that the body of the function was a single line, and it's because they wanted to give it a more useful name. It's like, no. Having to dig through all these layers of function calls is terrible. The way I put it is I had to go into the LLM code and pull some of the lasagna noodles out to make it more useful, but it still made that process a lot faster and some very cool code. Now that's almost ready to go.

Jeff Massie [01:01:10]:
Well, and see, that's where you're showing that you're a responsible vibe coder. As x1011 says, vibe but verify. Okay. Yeah. Okay. I Vibe coded this thing. Now start running tests on it. You know, is it compiling? Is it.

Jeff Massie [01:01:29]:
Are you getting the correct answer? Is it doing what you should? Is it. You know, is everything correct? I mean, that's. That's the biggest thing when you and I think that's where people have problems with Vibe coding is a lot of times nobody looks at it. They just, oh, I got the result and I throw it out there. And it's like, no, no, no, no. You need to look and see what it did. Is it running correctly? Is it compiling correctly? Is it.

Ken McDonald [01:01:52]:
That is still there after you're done?

Jeff Massie [01:01:55]:
Yeah, like.

Rob Campbell [01:01:56]:
Like the time I deleted my Digital Ocean droplet.

Jonathan Bennett [01:02:01]:
You deleted it? Oh, you're right. I'm so sorry.

Rob Campbell [01:02:04]:
Yeah, that's what it said.

Ken McDonald [01:02:08]:
I think this article that we've got here about Mozilla trying to get into being an AI provider is kind of reminiscent of when Mozilla first came out with their web browser during the wars.

Jonathan Bennett [01:02:25]:
They are trying to, they are trying to make that argument that it is. I don't know that that argument actually holds up. I saw that as well. And I don't know that I buy that maybe, but probably not.

Ken McDonald [01:02:42]:
But choice is always good.

Jonathan Bennett [01:02:45]:
Yes. If the thing that you're choosing between are actually all useful and there's a reason for them to exist. I just, I don't know that the world needs 17 of these fancy chatbots.

Jeff Massie [01:02:58]:
Well, they never work for. I mean, okay, now I've used a lot of the big name AIs, but the little ones that just, oh, we're going to have a little AI that's going to help you do this. Oh, they are dumber than a box of rocks and they never work. Right. And that's what I mean is like you, you AI is one of those things you can't just kind of halfway put your toe into. You're either all in or you're not. And a lot of these, if they, Mozilla really wanted to, let's just be honest, they just need to make a good, fast, clean product. Just be done.

Jeff Massie [01:03:33]:
And they would win browser wars rather than gimmicks. Stop being gimmicky. Just, yeah, make the lightest, fastest browser and make it really good. Maybe extra configurable things like that that people would love.

Ken McDonald [01:03:50]:
Yeah, not too configurable. You don't want to give people too many choices.

Rob Campbell [01:03:55]:
Oh well, one minute you say choice is always good and now you're telling people saying you don't want too many

Ken McDonald [01:04:00]:
choices because then you got to make that decision.

Rob Campbell [01:04:03]:
You could leave it default. You don't have to decide.

Jeff Massie [01:04:05]:
Yeah, it's just like KDE is a. Yeah.

Jonathan Bennett [01:04:09]:
I personally like my knobs and widgets and being able to fiddle with stuff, but I know not everybody does.

Ken McDonald [01:04:16]:
I'm used to getting my hands slapped too, John.

Jonathan Bennett [01:04:19]:
All right, we gotta, we gotta move on. And we're gonna talk about Mint here in just a moment. But before that we're gonna take a quick break. We'll be right back. All right, Rob, hit us with Mint.

Rob Campbell [01:04:33]:
And here is the big story that all my other stories have been leading up to in a way. But so you know, Linux Mint has some good news and some bad news. And honestly both point to the same bigger story. You know, Mint knows it has fallen behind. The developers at least do. Especially when it comes to modern desktop Linux. The headline this week is that Linux Mint is moving to a longer development cycle with Linux Mint 23 now not expected until December of 2026. That's this year.

Rob Campbell [01:05:13]:
So right around Christmas time. We've already mentioned on the show that they were considering extending their release cycle and now it is confirmed. Clement Love Lefebvre says the goal is to spend more time fixing bugs and improving the desktop instead of rushing through the old release rhythm. But let's be honest about what this also means. This also means the absolute earliest we could see a stable Wayland on Mint is the end of this year. But when you are still talking about early Wayland work like screensaver, removing old X.org dependencies and, and getting Cinnamon to behave well on Wayland, it's hard to see it being ready by then. So with this new release cycle, we are realistically not looking at a real Whelen support until the end of 2027 at best. And that's a good four years behind pretty much everyone else.

Rob Campbell [01:06:19]:
I mean at least all the big mainstream major distros out there. So for years Linux Mint has built its identity around stability, familiarity, trying to be like Windows. And that has worked well for the users who want something dependable. But the Linux desktop world is moving on and you know, GNOME is deep into Wayland, KDE is deep into Wayland. New features like better scaling, hdr, better input handling and the long term future of the Linux desktop are clearly heading in one direction. And Mint, especially Sentiment, has been late to the transition. A transition that should have begun over half a decade ago. I mean they should have started, everyone else started and they worked towards it.

Rob Campbell [01:07:11]:
And you know, Mint is only starting to kind of get to that point right now. It's, it's like they, maybe they didn't think Whalen is really gonna go anywhere and replace X. You know, when everyone else could read the tea leaves. I don't know. The good news is that Mint seems to know this now. This longer cycle is not just about waiting longer to ship, it's about buying time to actually do the work. Mint 23 is already using sentiment 6.7 unstable. It is getting new Wayland screen server, replacing the older Ubiquiti installer with lmde, the Linux Mint desktop installer.

Rob Campbell [01:07:54]:
And developers say they are continuing to remove remaining cinnamon ties to x.org, you know, and that that's real progress, even if it is overdue. So you know, this is the mid story right now, you know, a little slow, a little late, but maybe finally get serious about, about it, you know, the extended cycle is not something to celebrate on its own. In many ways it is an acknowledgment that meant development has not kept pace. So, you know, even though they are late to Wayland, my fear has been, and what I've told people that, you know, people are always like, they'll get there. You know, my fear was that when they get there then the next big thing is going to come out and they would be late to that and then the next big thing and then they'd be late to that and they would just always be behind. But maybe, maybe with this new push and development cycle, it'll help them keep up better in the Future. You know, 2028 may just be the year of the Linux Mint desktop.

Jonathan Bennett [01:09:03]:
Oh, that last cut was the deepest. Oh goodness, that's mean, Rob.

Ken McDonald [01:09:10]:
But it was always the first cut that was the deepest.

Jonathan Bennett [01:09:13]:
Not in this case. It's interesting to see the comments on like the 9to5 Linux article is different Linux Mint users who are sort of talking about, look, we don't want to be bleeding edge on everything. It's fine that it's coming late. We just want a stable desktop that doesn't change very much. I get that. I can understand that. This is why. It's why Linux Mint is the desktop you put your grandma on.

Jeff Massie [01:09:40]:
It's almost kind of going to sounds like lts.

Jonathan Bennett [01:09:45]:
Yeah, it really is.

Jeff Massie [01:09:46]:
It's becoming a long term officially.

Rob Campbell [01:09:49]:
Yeah, they have three year support I think. Unless that's changing too with this new life cycle.

Ken McDonald [01:09:53]:
Is this what you call a rolling lts? They keep rolling the next one out.

Jonathan Bennett [01:09:59]:
Wow, guys are suspicious today.

Rob Campbell [01:10:03]:
Yeah, I mean, yeah, it had its time and it's fallen behind, you know, especially with all these X XORG vulnerabilities that are always bringing up and that are 20 years old. So who knows who's found them before the devs found them and all the features, you know, this new, okay, they're not using kde but there are new features. So you know. And you know to say that I've always thought mint should use KDE. In fact, April 1st of this year I saw it posted. April 1st of this year I saw it posted that Mint was going to switch to kde and I was so excited and then I read the comments and realized what day it was. Come on, it don't make sense.

Jeff Massie [01:10:53]:
Well, but at that point it's really stepping on the toes of Kubuntu. You have almost two identical.

Rob Campbell [01:10:58]:
Well, there's already a. There's already A sentiment version of Ubuntu.

Ken McDonald [01:11:03]:
But how can they the toes if they're two versions behind?

Jeff Massie [01:11:10]:
I mean idealistically

Rob Campbell [01:11:14]:
big differentiators that people like about meant these days. I mean one, it's. They style it similar to Windows, which that's not my thing, but whatever, you can customize it, that's fine. And the other thing is that there's no snaps in it. So it's UBA2 without the snaps, which I could really appreciate that now. Just give me a Katie spin and boom, I'm sold. Instead of all they, all they want to do like their only options, they have two desktop options by default. And both of them are like the two oldest things you can get.

Rob Campbell [01:11:53]:
You get Cinnamon, which is a fork of gnome 3 and then you have xfce. I mean xfce is still a thing today, but it's, it's a slow roll thing that, that prides itself on being behind. So I mean, that's what I mean. You know, if you're going to be a, I mean they're going to be a distro, but you know, maybe have one option that's like old and slow. Pick Cinnamon, xfc, I don't care. And then pick another option that's modern, you know.

Jeff Massie [01:12:25]:
Well then get us all in here. Ubuntu without Snaps sounds like Debian. Debian with kde. Well, I mean, because now the Debian is include. You can include the non open source bits on it, the whole Ubuntu, including all that stuff. It's like the differentiation isn't really that big anymore. It's. And then you start going, well, let's just Debian with kde.

Jeff Massie [01:12:53]:
Oh, you want really stable Debian, they got that. You want more testing? They got that. So you, you kind of already have the solution existing and it almost gets redundant.

Rob Campbell [01:13:04]:
But Debian takes a lot more work still compared to Ubatu and then Linux Mint, which is based on Ubuntu. Like they, they do their installer and they make things easier for the more general user. Not, not that Debian's hard to set up, but it's a little. I know, I don't know how to say it, but it's, it's a little, there's a little bit more to it.

Ken McDonald [01:13:29]:
It's a little bit more to it than just sticking a USB stick in and turning the computer on.

Jeff Massie [01:13:35]:
Well, when was the last time you tried Debian?

Rob Campbell [01:13:38]:
I don't know, within the last six months. Oh, okay.

Jonathan Bennett [01:13:43]:
I'm thinking that Ken is just over there can't hardly contain himself. He wants to tell us that we all should try open SUSE Tumbleweed.

Ken McDonald [01:13:50]:
You're absolutely right.

Rob Campbell [01:13:52]:
Tell us why.

Jonathan Bennett [01:13:53]:
Well, yeah, what's new there?

Jeff Massie [01:13:54]:
Why, Ken?

Jonathan Bennett [01:13:55]:
What's the selling point?

Ken McDonald [01:13:56]:
Well, I'm going to ask Bobby Borisoff and Morris Nesker to help me with explaining this since they wrote about the latest changes to OpenSUSE. Tumbleweed According to Bobby and Marcus, Tumbleweed replaces Grub2BLS with systemd boot as the default bootloader. Now, Bobby's article even includes screenshots of the new bootloader for new installations. This transition to the new bootloader occurred in two phases. Back in November of 2025, OpenSUSE changed the installer default from Classic Grub 2 to the Grub 2 BLS, which remained a Grub 2 based solution. Now, as of this update, Tumbleweed has even advanced further by making systemd Boot the default for new installations. A key advantage of adopting BLS compatible boot setups is simplified integration of new features and improved compatibility with modern full disk encryption workflows. Now, according to Marcus Tumbleweed, users using the GNOME desktop environment will be happy to learn that the latest release, Gnome 50, landed this week in the Stable Repositories as an Upgrade for from GNOME 49.5.

Ken McDonald [01:15:28]:
The middle click paste feature was disabled, so if you're using it, just re enable it from the GNOME Tweaks tool under mouse and touchpad. I do recommend reading Bobby and Marcus's articles if you want more details on this update to open SUSE Tumbleweed.

Jonathan Bennett [01:15:48]:
Interesting. There hasn't been a whole lot of the uptake for system debut yet, has there? I don't think. I don't think Fedora is using it, which is kind of surprising. Fedora is a place where you'd expect to see that. But come, come, come about first. And I don't think we're using it yet.

Rob Campbell [01:16:07]:
Yeah, or these days UBA too. If it was based on Rust, it would be Ubuntu.

Ken McDonald [01:16:12]:
Sounds like I need to reinstall my Tumbleweed.

Jeff Massie [01:16:16]:
Cassie has it as an option.

Jonathan Bennett [01:16:18]:
Oh, does it?

Jeff Massie [01:16:18]:
Okay.

Ken McDonald [01:16:19]:
As an option?

Jeff Massie [01:16:20]:
Yeah, you can, you can.

Rob Campbell [01:16:21]:
Everything's an option.

Jeff Massie [01:16:23]:
Yeah.

Ken McDonald [01:16:24]:
Is it the default option?

Rob Campbell [01:16:27]:
No, I don't think there are default options. At least for a lot of things.

Jeff Massie [01:16:30]:
Yeah, I don't. Yeah, I think you're right, Rob. I don't think there is a default. I. I know it.

Ken McDonald [01:16:37]:
You have to change.

Jeff Massie [01:16:37]:
I've read some stuff that systemd is a little more limited if you're like dual booting or you're, you've got multip partitions. It's a little, little less friendly. I, I stuck with Grub because it's kind of the devil I know.

Rob Campbell [01:16:51]:
But, you know, Grub these days, if you, if you don't necessarily know what you're doing, it's a little. What is the word I'm looking for? They're dangerous. No, no, no. There's roadblocks to dual booting these days. Just because they, by default they've all removed OS Prober, so you always have to enable that again. You know, in the old days it was just there and it's like, boom, it just works. Hey. And now a lot of people like, hey, I mean, I've seen this question quite a bit online and I've for, for like brief seconds be like, why did my other things show up? Oh, yeah, I gotta enable OS Prober.

Rob Campbell [01:17:24]:
It's like, okay, let me figure out how to do that again. Every time it's like, that's. I'm not a fan of that change, but. Or, or when you're installing, ask when you're installing, are you gonna be. I mean, I think maybe some do ask, are you going to have multiple oss? But dual booting is so common that, I don't know, leave that in there for now. Whatever.

Jonathan Bennett [01:17:46]:
It's probably a security problem. Honestly, I think that's why they did

Jeff Massie [01:17:51]:
it, because then you can't load on a little something that looks like an operating system and it loads in and does some nefarious things.

Rob Campbell [01:18:00]:
Yeah, well, like detecting Windows on there. That's detecting old operations that do nefarious things.

Jonathan Bennett [01:18:11]:
Yeah, absolutely, yeah.

Jeff Massie [01:18:13]:
Windows is a virus.

Jonathan Bennett [01:18:15]:
Oh, don't sue us, Microsoft. So let's talk the kernel back to Jeff back about the kernel. 7.1. We got the new file system driver NTFS, but we are losing some things too. What is going on? Out with 7.1, Jeff.

Jeff Massie [01:18:38]:
And apologies to Ken from the outset here, but I have a pair of stories linked in the show notes for this week with one thread connecting them 32 bit in. The kernel is looking very shaky at the moment. While not the total end of 32 bit, we can see it from here. And at least for the x86 anyway, it's going to be on life support soon. The first story is where Linus Torvalds accepted the merge request to start to remove the Intel 46 code from the kernel. Now, this patch doesn't actually remove the code. It removes the kernel configuration Options for the M46 M46SX and Elan, Elan CPUs. Now the.

Jeff Massie [01:19:22]:
If you don't know that one Elan or Elan is a family of 32 bit CPUs from Intel which were meant to be for embedded applications based off the X86 family. Now what does this mean? It means while the code is still in the kernel, you can't compile it for a 486 CPU. Now this is done. So if there's big complaints from the public with reasons it needs to stay in the kernel, the options to compile can be added in and things move on without needing to make huge code changes. So it's kind of an easy way to test and see who's going to complain about it versus actually going through the entire code removal. So if no one complains or just a tiny bit of noise is made, you know, one, one or two people, then it'll be removed in a future kernel. If there's a need now to keep running the 486, you know, if they say, okay, we're getting rid of it, and you're like, you're one of the people saying, I still have to have it. The 6.18 LTS kernel is going to be around until December of 2027, so keep that in mind.

Jeff Massie [01:20:34]:
If you have hardware running that's, you know, running this, it's, you know, this at this point, it's pretty old. The 486 came out in 1989, and for desktop and laptop use it was discontinued in 1998. The embedded systems continued on until 2007, but highly unlikely, a person that has an embedded CPU is going to need the latest greatest distribution. So realistically, the last desktop or laptop running this was made about 28 years ago. Now, while I'm not a fan of E waste, there comes a point where something just isn't suitable for use anymore. And while the cpu, and that's not even counting the CPU might be fine, but getting memory or a motherboard could be a major issue. Now in the second story, there's a poll which might affect their other 32 bit hardware. The kernel's inode field is it inode field is an unsigned long number which has time went on resulted in hacks for hashing 64 bit identifiers into the 32 bit field.

Jeff Massie [01:21:39]:
So basically as the kernel went to more and more 64 bit support and 64 bit systems, they were still had a 32 bit field and so they had to make some little hoops they were jumping through to make that work. Now, Starting with the 7.1 kernel, the field is going to be an unsigned 64 bit number. So this means the 64 bit systems won't have to do any hacks and it will make the code cleaner. What this does mean for 32 bit systems they will see the struct inode grow by at least 4 bytes. This could have an effect on slab cache sizes and field alignment. Comments about this reference they are working to deprecate 32 bit systems in the kernel. So while it could cause a slowdown as the 32 bit systems now have to handle the larger number Excuse me. The feeling is it shouldn't be that big of an issue.

Jeff Massie [01:22:42]:
Linus Torvalds commented that the field could be eliminated because that Sorry, Linus Torvalds commented that a field could be eliminated because and I'm not reading the names of them, but the field was the actual holding place for the 64 bit number for 64 bit systems. It might not be a really big deal for the 32 bit systems to handle this, but now the hacks make it work. To make it work are going to be on the 32 bit people, not on the 64 bit people which are the majority. You know, I don't think it'll be terribly long before 32 bit systems are removed from the kernel entirely anyway. So just like how there were a lot of hacks and works arounds for 16 bit systems when removing 32 bit system that also amounts to code cleanup. Not to mention a lot of the 32 bit systems are old and developers don't have the hardware anymore to test and maintain the code. We've seen that before in the past where they say I'm supporting some of this old code but I don't have any hardware that supports it. I need people to run and test this because I have no access to hardware this old.

Jeff Massie [01:23:56]:
But have a look at the articles linked in the show notes and let us let us know your thoughts on the discord about the life of 32 and 32 bits in the kernel.

Jonathan Bennett [01:24:06]:
Yeah, so this is interesting to think about. They're currently removing the 486 and that was I'm trying to figure out how many generations there were between the 486 and moving to 64 bit.

Ken McDonald [01:24:22]:
Well, there's a 20 year difference between the AMD Athlon 2x2 240 that I've gotten my old Lenovo ThinkCenter A63 and when the last 486 came out.

Jonathan Bennett [01:24:41]:
Yeah, was Pentium 4 essentially the end of the 32 bit series.

Jeff Massie [01:24:47]:
I don't know off the top of

Jonathan Bennett [01:24:49]:
my head, I think so.

Ken McDonald [01:24:53]:
And another thing is you got to remember with all those old 486s, they were all single core chips with no caching or pipeline techniques on them.

Jonathan Bennett [01:25:08]:
Yeah. Apparently there is a Pentium dual core is what they called it. That was a 32 bit. Some of them were 32 bit and some of them were 64 bit. So essentially you've got the Pentium one through Pentium four are the what are going to be remaining in the kernel after 386 is dropped. So that's still a reasonably wide range of hardware.

Ken McDonald [01:25:40]:
I've got to dig an old Dell that I've got sitting behind me, see if that's a 486 or if it's one of the earlier ones. I think it's old enough, then it may have a 386 in it.

Jonathan Bennett [01:25:57]:
When. Yeah.

Jeff Massie [01:25:59]:
And you know there was talk, this is for the X86, but there's actually talk about when. When will like ARM or some other 32 bit systems start being deprecated? Nobody's doing anything right now about it, but at least the talks are starting. Which means, you know, it'll probably be a few years before anything happens. But it's, it's at least being thought of to start cleaning 32 bit out of the kernel.

Jonathan Bennett [01:26:25]:
Yeah. Arm 32, there's, there's like four arm is. Arm is quite a different story because there's like four or five different ARM targets there. You know, ARM v, like ARM v5, hardly anything still supports it. And then I think current gen 64 bit armed cores or ARM V8 if I remember correctly. So like it's. Oh, I guess it's similar to the x86 story with there being multiple generations in there.

Ken McDonald [01:26:56]:
With Raspberry PI you could say there's five generations.

Jonathan Bennett [01:27:01]:
There's not though. There's really about maybe three because the, the 64 bit Raspberry PIs are all essentially the same. Not processor speed but like the ISA is about, about the same.

Ken McDonald [01:27:18]:
That's 4 and 5.

Jonathan Bennett [01:27:19]:
Yeah, I remember if the 3B was. The 3B plus may have been capable of 64 bit. I don't remember. That was a weird one. Anyway, we shall continue watching this in the kernel. There is one last story that we have here to cover and it's sort of an update on what, what we talked about last week with the Office weirdness and only Office and LibreOffice and Nextcloud and all these Guys over there tussling over licenses. Well, one of the interesting things that happened there was OnlyOffice basically said, oh no, no, no, you can't. We're open source, we're using the agpl.

Jonathan Bennett [01:28:01]:
We just, we added some terms to it and we talked last week about this question of what terms can you add to a Free Software license? Because the Free Software License sort of inherently says you don't get to add terms to this. And we talked about the fact that Red Hat essentially does this and now the folks at OnlyOffice do the same thing. Well, this prompted Free Software foundation to make a statement, and we've got links to it in the show Notes. And Free Software foundation led with the title of you cannot use the GNU AGPL to take software freedom away. And that, that sort of, in a nutshell, covers what they're going to say here. They, they basically said that you don't get to actually add text to the AGPL and continue to call it the agpl. When in your software itself it says that something is licensed under the agpl. It means the authentic agpl, the one that we published, not some other license that you're calling the agpl.

Jonathan Bennett [01:29:11]:
Basically, you can't do this the way that you're doing it. You can't do it. And you know, one of the things that OnlyOffice did is, for instance, they said if you fork OnlyOffice and continue to distribute it, you must use the only Office name and image. Let's see, see if I can find the exact quote here. You must retain the original product logo when distributing the program. That's it. Well, that product logo is copyrighted. So, like, there's just some things that were done to make it very difficult to fork OnlyOffice in the way that we're used to open source software being forked.

Jonathan Bennett [01:30:01]:
And Free Software foundation came out and said, no, no, no, you can't do this. They close here by saying, we urge OnlyOffice to clarify the situation by making it unambiguous that OnlyOffice is licensed under the AGPLv3 and that users who already receive copies of the software are allowed to remove any further restrict. Additionally, if they intend to continue to use the AGPLV3 for future releases, they should state clearly that the program is licensed into the AGPLV3 and make sure that they remove any further restrictions from their program documentation and source code. Confusing users by attaching further restrictions to any of the Free Software Foundation's family of GNU general public licenses is not in line with Free Software There you go. Free Software foundation drops the hammer. So what they would.

Rob Campbell [01:31:02]:
What they want is a different license, their own license, or maybe there's something else that fits what they want, but it's not the agpl and it's questionable

Jonathan Bennett [01:31:11]:
whether the license they want is actually free software or open source at all.

Rob Campbell [01:31:17]:
Right. No. Yeah. And this was two weeks ago.

Jonathan Bennett [01:31:20]:
Okay, two weeks ago.

Jeff Massie [01:31:23]:
Yeah.

Jonathan Bennett [01:31:23]:
Very, very interesting to see though, as the story continues to percolate along.

Rob Campbell [01:31:27]:
Unfortunately, the weekend I took off and I missed it because.

Ken McDonald [01:31:30]:
Are they just asking for whoever forks it to give recognition to the original source or they just.

Jonathan Bennett [01:31:40]:
So only, only Office is. From what I can tell, they basically want to be able to control the ability to fork the project

Rob Campbell [01:31:51]:
or at least they want other people to know that if someone else forks it, hey, that's only Office. We did that.

Jonathan Bennett [01:31:57]:
Yeah, yeah. So it's adding additional restrictions on a fork that are really not compatible with what Free Software has always been about. And so that is the essence of the fight here. And we talked about it a couple of weeks ago. The, the multi layered struggle between Liberty Office and only Office and nextcloud. And it seems like there's another player in that. The, the Europe Office or whatever that's called. I forget.

Jeff Massie [01:32:27]:
Euro Office.

Jonathan Bennett [01:32:28]:
Yeah, Euro Office. That was close. Euro Office. Yeah. It's such a. It's such a weird story. Every. Everything about it is very odd.

Rob Campbell [01:32:35]:
Yeah.

Jeff Massie [01:32:36]:
They were having a big fight. Yeah. The week you missed, I remember it was like there is more, you know, more a bigger fight than X versus Wayland or Vim versus Emacs or. I mean it was every Office versus Apple.

Jonathan Bennett [01:32:52]:
Yeah.

Jeff Massie [01:32:52]:
Yeah. Every Office suite was having issues with

Jonathan Bennett [01:32:57]:
somebody else except for Open Office. I didn't hear anything from the Open Office guys.

Ken McDonald [01:33:04]:
Oh, well, yeah, he's too busy trying to answer his emails.

Jeff Massie [01:33:09]:
That was a little mean.

Jonathan Bennett [01:33:10]:
I'm sorry.

Jeff Massie [01:33:13]:
One question we do have is what does forking mean in software?

Jonathan Bennett [01:33:17]:
Oh, that is a good question.

Jeff Massie [01:33:19]:
Yeah. So it's when you have a software project. So maybe Jonathan's writing a software project and I disagree with the decisions that Jonathan's making. I make a copy of the software and I start making my changes that I want to see. So I forked it. So it's like a fork in the road. It's one code base until we copy it and start making different changes. And so then it, the software goes on two different paths from that point forward.

Rob Campbell [01:33:49]:
So at least one well known example is Open Office and LibreOffice. Open Office. What has been a thing For a long time. I mean there's much more history to it, but. Yeah, but Open Office there was some disputes about direction and things internally. Anyway, so this group says, hey, we're gonna fork this and now make it our own. And now we have LibreOffice. Now that doesn't mean Open Office goes away.

Rob Campbell [01:34:20]:
It should, but it hasn't. So now you have the LibreOffice guys make adding all new features and, and keeping LibreOffice going. And you have the Open Office guys pretending like they're keeping it going, at least offering download still.

Jonathan Bennett [01:34:39]:
Wow, that's cruel, Rob.

Jeff Massie [01:34:42]:
Really?

Rob Campbell [01:34:42]:
They do a few updates. They do some updates too once in a while.

Jonathan Bennett [01:34:45]:
Yeah.

Jeff Massie [01:34:46]:
And is it Cinnamon a fork and Gnome two.

Rob Campbell [01:34:49]:
Gnome three or three? Yeah, Cinnamon. Yeah, Cinnamon comes from Gnome three and Gnome two is mate. Oh, that's kde. I don't remember which one is Trinity two?

Jonathan Bennett [01:35:01]:
Trinity is. No, it's three.

Rob Campbell [01:35:03]:
Three KD three. I can't remember.

Jonathan Bennett [01:35:05]:
Trinity is three. Rob, have you never been to church in your life?

Jeff Massie [01:35:10]:
Trinity is three.

Rob Campbell [01:35:11]:
I did. I did not put that together. That obviously makes sense. We don't really talk about Trinity much. Not like all the Gnome forks. There's a lot more drama.

Jeff Massie [01:35:21]:
Try. Yeah, so this week we need to not have Rob, you know, donate for coffee. We're going to donate for Rob to have Hooked on Phonics so he can learn all this stuff.

Jonathan Bennett [01:35:33]:
Remedial grammar. Interesting times. One other thing in the area of forking we talk about forks. We're generally talking about like semi hostile forks. There are also friendly forks and software gets forked constantly because that's what a PR is, a pull request. That is where someone forks software, makes a change and then says, hey, why don't you take my changes back? So like forks happen all the time with healthy software. What is less good is a fork. And then the person that forked it said everybody should use my software instead where it's been forked.

Jonathan Bennett [01:36:17]:
And that's kind of the hostile fork idea. So I think it is time to move on to command line tips. We are going to take a quick break and then we're going to let you, Rob, jump in with the netwatch. I don't know what this is. I thought I knew what this is and then I read it again and I have no idea what it is. So we'll be right back and Rob will tell us all about it.

Rob Campbell [01:36:47]:
All right, so my tip this week is NetWatch or Nut Watch Dash 2e. So Net Watch is not something you wear on your wrist like The Pine watch or anything like that. It is.

Jonathan Bennett [01:36:59]:
I was. I was totally thinking it was the OS for Pine Time.

Rob Campbell [01:37:02]:
It's like, that would be cool, but it's not. It's something else. I mean, there's other options out there like this. This is not unique, but it's one more option to monitor your network traffic. So for those watching, I have netwatch TUI brought up. And so here you can see receive and transmit packets as they flow across. I'm just going to bring a little speed test up and let that run. And also you can see that my usage jumps up.

Rob Campbell [01:37:40]:
It's going to first jump the receives and then the transmission transmits are going to jump afterwards. There's this thing here that shows the top connections. Firefox. That's about all I did. Network Manager. You have some health stuff down below Gateway, the latency, the loss, DNS stuff, some errors, drops, various things down there. Now, oops, there's also. I don't want to do it.

Rob Campbell [01:38:12]:
The mouse I want to do with the keyboard. If I toggle down, I have at the top, I press the down arrow. Now it shows the menu at the top. So I have other things I could do here too. So I just look there I have on one is the dashboard that you're looking at. Two just shows all the connections. And I could go on one of these and there's things at the bottom you can see there's like a T if I wanted to do a trace route on it, record or freeze. So three shows my interface.

Rob Campbell [01:38:43]:
This is nothing special there. But four shows my packets. So there is a packet capture built right into it. So I could, you know, I could stop this when I wanted. Filter, follow, record all the things you might want to do with a packet capture. There's also an E for export, so you got packet captures right in there. Five just kind of shows a bunch of stacks. The protocol hierarchy HTTPs, that's kind of the big one.

Rob Campbell [01:39:14]:
You got a little bit UDP and. And then all the others are well down below. You got some latency numbers below there. And if I do six, it shows the topology of the networks, which is kind of neat. So it shows the layer seven for the, for those who know the osi, you know, the UDP and all that stuff. And it's got the. The device and to the gateway into the DNS. And anyway, then you got S, which is a timeline of.

Rob Campbell [01:39:48]:
Of things that have happened. A is the processes using the network. I don't have a whole lot going on here because It's a demo machine anyway, and that is it. It's just a nice TUI monitoring tool to monitor your network and see what's going on.

Jonathan Bennett [01:40:08]:
Yeah, very cool. I used one of the tips that you brought a few months ago. One of the network scanners I was working at a hotel and they had cameras that were down. And I plugged my laptop into it and I went and I checked on our spreadsheet to figure out what the name of it was, then went and installed that and used it to find the cameras. Was very useful. So thank you for that.

Rob Campbell [01:40:31]:
I. I need to update my vibe code one because I have new code on my computer that I have not pushed that really adds a lot of features to. To that other one, but.

Jonathan Bennett [01:40:43]:
Cool. Do it. Do it, man. All right, up next, Jeff. Wtf Jeff.

Jeff Massie [01:40:51]:
Yeah, my. You know, my command lines actually ties in pretty good with that. So my command line tip this week is wtf? Or more accurately, WTF util. It's written in go. And it's just a modular, very modular way to create a custom dashboard in your terminal. So the way it works is the entire dashboard is made up of individual modules, each pulling data from a different source. You just choose what modules you want to include, how big or small each one is in, in relative size in the terminal, and exactly where it sits in your terminal or on your dashboard. So the result is a completely personalized terminal dashboard showing only the information that matters to you.

Jeff Massie [01:41:41]:
So on their GitHub page they have an example which also can be found in the article linked in the Show Notes. The dashboard shows the weather in Barcelona. A to do list, a monitor showing changed files and recent commits for a GitHub page. A calendar with scheduling items. The power status of a system look like it might have been a laptop or something like that. Uptime in another window and a window tracking some recent JIRA tickets. So it can be any information you want in any size and any shape you want. Now, the article linked in the Show Notes shows several ways to get it from using Brew, pulling from the AUR or getting your distributions pre compiled binary from the GitHub page.

Jeff Massie [01:42:25]:
When running the program for the first time, it comes up with a default dashboard. And when you want to make changes, there's a dot config/wtf config yml file that you edit. The article goes into details how to make changes, and the author says that they are. They're not a programmer and it was easy for them to get the dashboard they wanted. They even give A prop prompt examples so you can use AI to generate the configuration that you want. So have a look at the article linked in the show notes and it also includes link to the home website for WTF and links to the GitHub page. So. Happy configuring.

Jonathan Bennett [01:43:06]:
Very cool.

Rob Campbell [01:43:07]:
So does that come. It comes with available modules, but I'm assuming there is probably some documentation or.

Jeff Massie [01:43:15]:
Yeah, it's got documentation and it's. Yeah, the modules aren't all built. They're not built in. So you get them from like the WTF website and the GitHub, but there. And other people have created custom ones. There's, you know. But yeah, you can, you can definitely create your own and add them in there as well.

Rob Campbell [01:43:38]:
My own little side project I've started working on myself has to do with the custom dashboards, but it's, it's web based.

Jeff Massie [01:43:46]:
Oh, okay, cool.

Jonathan Bennett [01:43:48]:
But look at wtf. That actually looks really interesting.

Jeff Massie [01:43:50]:
Yeah, yeah, because how many times do you have something like, I know, like B top or something like that? Maybe you're. And you're like, I don't care about this info, I want to know more about this information and you know, wish you could totally rearrange some stuff and change what's in there and.

Rob Campbell [01:44:05]:
Well, that's when you put a CRT next to your server CRT monitor. That is for those who don't recall that term. And then you put your WTF dashboards on there and just display what you want for your server.

Jeff Massie [01:44:20]:
Exactly.

Jonathan Bennett [01:44:21]:
You don't have to use CRTS. You can actually use modern LCDs and get the same thing.

Rob Campbell [01:44:26]:
Yeah, but that doesn't use enough electricity.

Jeff Massie [01:44:30]:
How else are you going to heat the room?

Ken McDonald [01:44:31]:
Yeah, I wonder if you could set up a module to take advantage of the tool. I'm about to go over.

Jonathan Bennett [01:44:39]:
Maybe. Why don't you tell us about it?

Ken McDonald [01:44:41]:
All right. Well, Jonathan, this week I have a command line tool for managing virtual video devices on Linux. It lets you create, configure and remove virtual webcams without reloading kernel modules. You can also query the status and capabilities of existing virtual devices or set timeout images to display if a video stream drops. Now you will need the V4L2 loopback module loaded. I'm going to bring up my screen here and I've got OpenSUSE, tumble read running in a VM so I can demonstrate. I'm going to start off by going with the command for listing your virtual modules. You do need to be root

Jonathan Bennett [01:45:44]:
or run the sudo.

Ken McDonald [01:45:47]:
That helps. That way I don't have to keep logging in and out. And the basic Command is sudo v4l2 loopback CTL list. And it's asking me for the password. This does not have that option to display how many characters I'm typing, thankfully. But there it gives the list of what virtual webcam I currently have set up in my vm. That's from way back when, when I was playing around with trying to do obs from the VM and go pipe outs. But now I'm going to add a virtual bridge.

Jonathan Bennett [01:46:45]:
A virtual bridge.

Ken McDonald [01:46:48]:
That's what I'm going to call it anyways.

Jonathan Bennett [01:46:50]:
Okay.

Ken McDonald [01:46:55]:
And there it showed. It came back with row one. So let's go back to the list. And it lists the two devices.

Jonathan Bennett [01:47:02]:
Yeah.

Ken McDonald [01:47:04]:
And you can query the device by just typing. Let's fill it. Right. Query. Device video. And you're. You'll see. Now those.

Ken McDonald [01:47:30]:
All those options that you see in the command I used up there are what told it to set the card label to virtual bridge. The max width and height to 8192, the buffer to 4

Jonathan Bennett [01:47:53]:
and max.

Ken McDonald [01:47:56]:
Yep. Max openers to 10.

Jonathan Bennett [01:47:59]:
Cool.

Ken McDonald [01:48:01]:
And the dash X to 1 turns on the whether or not it's exclusive capture.

Jonathan Bennett [01:48:10]:
Aha. Got it.

Ken McDonald [01:48:13]:
Now why would you need that?

Jonathan Bennett [01:48:16]:
Which exclusive capture.

Ken McDonald [01:48:19]:
Yep.

Jonathan Bennett [01:48:21]:
Wasn't there something with the browsers being able to use a device?

Ken McDonald [01:48:26]:
Yes, it is.

Jonathan Bennett [01:48:28]:
Do you have to turn exclusive capture on for a browser to be able to use it?

Ken McDonald [01:48:34]:
Correct. Now, now that we've got that set up, I want to demonstrate using this to capture a test signal that I'm generating with FFmpeg. And we're going to watch that in vlc.

Jonathan Bennett [01:49:05]:
Oh, cool.

Ken McDonald [01:49:11]:
And if you've ever used vlc, you could actually have opened it up. Let's do this this way. Going into capture device. And select the device name and hit play. Now you can also set VLC up so you can stream that out.

Jonathan Bennett [01:49:49]:
Yeah, that's the L, the LAN part of vlc. It's the video LAN client. All kinds of tricks VLC knows about.

Ken McDonald [01:50:01]:
And this is going back to WPCTL status. You can see that it's showing both devices as available, but it's not showing that stream at this time. Yeah.

Jonathan Bennett [01:50:21]:
Wpctl. It's wire plumber control. Correct. It's not really a pipe wire stream, it's a V4L2 stream. So.

Ken McDonald [01:50:29]:
Right.

Jonathan Bennett [01:50:29]:
Doesn't know about it. One of these days, one of these days we'll get a replacement for the. The V4L2 loopback. Device. That'll be. That'll be pipe wire.

Ken McDonald [01:50:39]:
We're getting closer.

Jonathan Bennett [01:50:40]:
Yep.

Ken McDonald [01:50:42]:
Because this is my QPW graph and it's showing my webcam and it's showing my virtual camera from Obs Studio.

Jonathan Bennett [01:50:57]:
Nice.

Ken McDonald [01:50:58]:
You just can't see the video interface on the browsers.

Jeff Massie [01:51:04]:
Right.

Ken McDonald [01:51:05]:
To tie them in.

Jonathan Bennett [01:51:09]:
Yep. We're so close.

Ken McDonald [01:51:15]:
That's the next step. Get the browser menu coders to come up with that option.

Jonathan Bennett [01:51:23]:
Yeah, yeah. All right. Very cool. So last week I promised everybody we would do a tour of a working Grafana. And I did take some time today and get that going. And so let's take a look at a Grafana instance where we've got. Let's see. I should be able to.

Jonathan Bennett [01:51:49]:
If I hit the right button here.

Jeff Massie [01:51:51]:
Yeah.

Jonathan Bennett [01:51:52]:
Go full screen with this. This is Grafana pulling information directly from the framework. The laptop I'm on, that's these cards on the right, but then also a raspberry PI 5 over meshtastic. And these are the cards on the left. And so the bits of info that I'm currently pulling and graphing are three load indicators. I'm also grabbing the amount of disk free and the amount of memory free on each of these devices. You can see that my PI is steadily using up memory. I'm sure at some point it'll free this up and it'll jump back up to almost 8 gigs free.

Jonathan Bennett [01:52:34]:
But that's where it's at at the moment. And there's a couple of interesting things to note here. One is on the disk free, I have these set up slightly differently. You can probably tell on the PI, it's all 433 gigabytes. And it's giving me very fine grained bits of info here. As far as you can see, just about every time a file is written or deleted. Whereas over here on the framework, it's a very boring straight line. And this is pretty interesting to look at.

Jonathan Bennett [01:53:06]:
So if we edit this particular graph, a couple of things here to note. One is down here under standard options, you can choose the unit and I have it set to bytes bytes, SI style bytes. And that's how it knows that these are gigabytes and not just, you know, 800 billion, but here you can set the min and the max. And so I've set the min on this one, the minimum to zero. And that's why it's got this graph pinned at zero bytes at the bottom. And it's giving me sort of an overall view rather than a really, really detailed view about you Know each each time bytes are used and. And then later released the query again. We talked about this last week.

Jonathan Bennett [01:53:56]:
Query is real simple. We're just pulling from mqtt then we're pulling the payload apart because it's JSON and we're matching against one of these values to only get a particular machine. And then we're filtering which of the information that we want so that we don't have everything up on the same graph. And what I've been doing, let's say we wanted to pull system uptime as well. So I'm gonna go here to the Framework memory Free graph and one of the options here under more is to duplicate. Well so I can duplicate get a second graph and then I go into the second one and edit it and I'll go back into transform and I'll say instead of free mem bytes, I don't care about that. But I would love to see uptime seconds. Now this is a, you know, kind of a boring graph in this particular case.

Jonathan Bennett [01:54:49]:
But you know, you imagine for certain things this is going to be very useful to be able to watch your uptime and they're not bytes. So we'll go down here, we'll change it from units. Currently it's under data. We'll go date and time, maybe I don't know for sure which one this is going to be. Time seconds. There we go. And this is something that happens reasonably often working with Grafana, particularly with this MQTT is you go to no data and that's because this is all ephemeral data. It doesn't have, it doesn't have long term storage but so you can rename it to uptime and.

Jonathan Bennett [01:55:36]:
Yeah, and it's now telling us, you know, 4.49 days, we want to, we can save and apply and then it'll show up on this graph. There's another couple interesting things that you can do and that's up here in the top right corner. These are all just time series at the moment. That doesn't make maybe a whole lot of sense for what I'm doing here. Maybe stat would make more sense. And so you know, you can have other things here on your dashboard. Some of the things that I found that was really fun is there's a gauge and so it may not make a whole lot of sense with the, with the days. Although we could probably go in and set the max to, I don't know, something like that and you know, give it a sort of a target.

Jonathan Bennett [01:56:31]:
We want to, we want to be up that long, at least. We'll see what happens here in a minute when that pulls in with some data. But all kinds of fun ways that you can use Grafana and customize these things and build a dashboard. I've been having a lot of fun with it. I'm going to continue playing around with Grafana. I've got something else for next week, but for now, this is. This has been a lot of fun to play with.

Jeff Massie [01:56:54]:
Now, does it support more advanced plotting like bode plots or logarithmic or things like that?

Jonathan Bennett [01:57:00]:
You can. I know it can do logarithmic. I've not. Again, I've not messed a whole lot with that. I don't know for sure about bode plots, but I bet it's possible.

Jeff Massie [01:57:16]:
If you could do logarithmic. That's a big one there. Bode's a little specific for things.

Rob Campbell [01:57:21]:
Yeah, I've never heard of it.

Jeff Massie [01:57:24]:
It's used a lot of time for frequency type plots. It's kind of a logarithmic, but each 10 to 100 is kind of its own mini logarithmic scale.

Rob Campbell [01:57:38]:
Not saying I haven't seen it, but maybe I just didn't know what it was called.

Jonathan Bennett [01:57:41]:
Yeah, that's cool. Yeah, it's got a bunch of different options that you could use for stuff like that. You can even go in and set thresholds. I've not messed with this yet, but you can do thresholds to where it will give you, I think, alerts. I don't know all of the details of what you can do with that, but I know that you can set those up and it'll take actions of some sort when, you know, when something happens. So some really, really interesting stuff there. A lot of fun to play with. All right, I think that is it.

Jonathan Bennett [01:58:24]:
I'm gonna let each of the guys get the last word in on whatever they want to or plug their stuff. I think Rob, if he can stay awake for it, is going to come and plug his car. Coffee fund definitely needs that today.

Rob Campbell [01:58:36]:
Yeah. You know, if I'm looking at this panel we have here today, besides you, Jonathan, I feel like we all look a little bit tired, so maybe it's time to send some coffees our way. So. And how you could do that is you can go to robertp Campbell.com and then scroll down a little bit. Or if you have a big screen, you should have to scroll down. But there's links to my LinkedIn where you can connect with me, Twitter, my blue sky, my mastodon, and then finally, on the left, or I guess if you read in left to right, like most of us do around here, the coffee is actually first. There's this little coffee cup here. You could donate a coffee to me.

Rob Campbell [01:59:14]:
And now if you guys are sick and tired of me being so mean to Mint all the time, go ahead, donate here and tell me about it. Tell me to stop doing it. But if you like it, if you like this, the things I have said here about Mint, go ahead and donate a coffee here. Tell me about that there. And if you want to donate these coffee, these guys, you know, you can donate there too. And just tell me in the comments who it's for and I'll get it to them someday.

Jonathan Bennett [01:59:45]:
Eventually we'll get together and have some coffee. And Rob's buying.

Jeff Massie [01:59:49]:
And Rob has paid me off, so he is good about. He will pay.

Rob Campbell [01:59:53]:
Yep, Jeff and I are equal.

Ken McDonald [01:59:55]:
Even so you've paid them all the coffees you've got up to today?

Rob Campbell [02:00:01]:
Yes, I paid him in advance of one. I think one coffee. And then he got one Stenson. So we are.

Jeff Massie [02:00:10]:
So I'm. I'm Even so everybody that donated a coffee to me, I. I got it. He hand delivered it to where? The state. I live several states away at a bar.

Rob Campbell [02:00:25]:
Yeah. And in the form of a dark liquid that wasn't coffee.

Jeff Massie [02:00:34]:
It looked like coffee

Jonathan Bennett [02:00:37]:
didn't have quite the same effects, though.

Jeff Massie [02:00:40]:
All right.

Jonathan Bennett [02:00:43]:
Probably not. All right, Jeff, what you got for us?

Jeff Massie [02:00:47]:
If you want to connect with me on LinkedIn, you got to connect with Rob first and then try to connect with me because I get too many fake profiles trying to connect with me. But if I see I don't recognize one and I see you're connected to Rob, then I'll know. So just FYI there, the only other thing is a little bit of poetry. Wind catches lily scattering petals to the wind. Segmentation fault. Have a great week, everybody.

Jonathan Bennett [02:01:19]:
Oh, the old seg fault. All right, Ken.

Ken McDonald [02:01:22]:
Well, as we get ready for spring and get ready to do that spring cleaning, just want to remind every. Everyone, please back up first, not after.

Jonathan Bennett [02:01:39]:
Back up before you do your spring cleaning. Yes.

Rob Campbell [02:01:43]:
I can't back up all those physical things I'm taking into the dump.

Jeff Massie [02:01:47]:
Well, lesson learned. I thought I. You need to double check because I thought I backed up my documents directory and I. I lost some poetry and notes.

Ken McDonald [02:01:55]:
Oh, no.

Jeff Massie [02:01:57]:
Remember when I. When I. It was going through the cashy reinstall or install and I nuked my kubuntu and I backed up, but I didn't Fully back up.

Jonathan Bennett [02:02:11]:
There you go. That'll do it.

Ken McDonald [02:02:13]:
One directory. You'd be surprised how important that one directory can be sometimes.

Rob Campbell [02:02:19]:
That's why I prefer a full system backup.

Jonathan Bennett [02:02:22]:
True.

Rob Campbell [02:02:23]:
Give me an image that I can split up.

Jeff Massie [02:02:26]:
It wasn't horribly painful because I had some recent, you know, like a month ago backup. So it wasn't terrible.

Rob Campbell [02:02:31]:
But still, it wasn't horribly painful because it was his poetry and, well, I

Jeff Massie [02:02:38]:
had some other dogs.

Ken McDonald [02:02:38]:
Oh, that was very painful.

Jonathan Bennett [02:02:41]:
It's not exactly Vogon poetry, but

Jeff Massie [02:02:46]:
I

Rob Campbell [02:02:46]:
mean, it wasn't even his poetry.

Jeff Massie [02:02:48]:
You'll have a Pan Galactic Gargle Blaster. Jonathan, there's someone else's poetry.

Jonathan Bennett [02:02:54]:
All right, let's see. Did we let Ken get the last word in? He told us to back everything up. Yes. All right. Well, I appreciate you guys being here. We have had a blast today talking about Linux. If you want to find find more of me. There is of course Hackaday.

Jonathan Bennett [02:03:09]:
That is still where Floss Weekly is at. And we've had a couple of weeks off, unintentionally. We've got a guest schedule for this week and then some. Well, next week, this week when you're listening to it, next week when we're recording, but then the week after that. Hopefully if we can pull it off, something really special is going to be really interesting and different. So make sure and keep an eye out for that.

Jeff Massie [02:03:31]:
Other than that, just want see to

Jonathan Bennett [02:03:32]:
say thank you to everyone. Whether you watch or listen, whether you get us live on the download, we sure appreciate it. And we'll be back. We'll see you next week on the Untitled Linux Show.