The default allocator is very slow but it can be changed. Chimera Linux, for example, uses mimalloc which is very fast.

My distro is based on MUSL. I seem to remember finding something that would not build on it but I do not recall what that is. In addition to the thousands of packages I am using, I have compiled hundreds of applications. Compatibility is very high.

Certainly it is clear the “most applications” work with MUSL.

That is, the source code does.

gcompat is when you want to run something that is already a binary that wants to call into Glibc. I try to avoid that so I cannot comment much.

There is the odd time I have had a binary built for Glibc that I could not avoid. For example, bootstrapping .NET or the version of vcpkg that the Ladybird browser uses in its build system. To be honest, in those cases, I just reach for Distrobox and drop into a distro that has Glibc natively, like Arch. Or I might use a RHEL Distrobox for a commercial binary meant to target that distro.

Clearly running a binary without one of the dependencies it was built against is a problem no matter what library you are taking about. But if we are just asking what works on MUSL, I would say almost everything.

While I do not disagree, it is important to realize that this was named in early 90’s.

When Windows NT was released, an important aspect of the architecture was the idea that it had “sub-systems” to provide app compatibility.

The Win32 sub-system (Windows apps) was just one of them. It’s full name was the Windows sub-system for Win32.

There was also the sub-system for POSIX (UNIX compatibility to win government contracts) and the sub-system for OS/2. OS/2 was interesting as it was still expected to be the important competitor and because it was originally going to be a Microsoft OS so Microsoft had customers that had written OS/2 apps.

The Windows sub-system for POSIX was never any good. It was just good enough to check boxes and win procurement contracts. Windows NT became quite successful and UNIX compatibility was not important.

That is, until Linux became popular.

So, when Microsoft added Linux application compatibility to Windows, it was naturally to call it the Windows sub-system for Linux.

I agree that the name sucks but it makes sense in a historical context.

It is my root partition and it is quite good.

https://bcachefs.org/

Too bad the lead dev got it kicked out. It is worse for users but he seems happier.

deleted by creator

The video is not for us. It is for his many followers that have heard about Linux but are too scared to take the plunge.

His experience gives them a baseline for what to expect if they tried it themselves. That is massively useful.

If you are comfortable with many distros, you would use Cachy without visiting their wiki.

I do not think Cachy needs a warning. The fact that you cannot answer their fist question is warning enough. And either you charge ahead, read the wiki, or bail.

The problem is having new users start on Cachy.

In the early days, Ubuntu did a tonne of marketing that they were “Linux for human beings” and every new Linux user seemed to choose them. I was already a Linux user and never really used Ubuntu but everybody else seemed to be having a decent experience.

It is too bad we do not have that today, with some distro getting the word out that it is the right place for “normal” people to start.

Like a lot of other people, I point people at Mint. It is not only user friendly but Llosa maintenance and fairly conservative in terms of change. I think those are good qualities for a “normal” person OS. But, if there was an obvious alternative, I would be happy aligning with “the one” that we collectively recommend.

I love the choice in Linux but it does suck that people coming from monocultures get hit over the head with it immediately.

When we see videos like this, it should start with him discovering that we all recommend the same one or two alternatives. A quick Google or ChatGPT should produce the same answers.

I can almost guarantee that the problem you encountered was an outdated archlinux-keyring that meant you did not have the GPG keys to validate the packages you were trying to install. It is an annoying problem that happens way too often on Arch. Things are not actually screwed up but it really looks that way if you do not know what you are looking at. One line fix if you know what to do.

It was my biggest gripe when I used Arch. I did not run into it much as I updated often but it always struck me as a really major flaw.

Half of them probably will. People are like that.

Many times, I have seen people switch tech because something is missing or has changed…and they switch to something that also does not have it. Boggles my mind.

It could be that /dev/sdb2 really does not exist. Or it could be mapped to another name. It is more reliable to use UUiD as others have said.

What filesystem though? Another possibility is that the required kernel module is not being loaded and the drive cannot be mounted.

Nothing that uses apt is remotely bullet-proof. It has gotten better but it is hardly difficult to break.

pacman is hard to break. APK 3 is even harder. The new moss package manager is designed to be hard to break but time will tell. APK is the best at the moment IMHO. In my view, apt is one of the most fragile.

As the comment above stated, Chimera is not trying explicitly to be “lightweight” as a goal. Chimera Linux containers are pretty small but much bigger than Alpine.

Alpine and Chimera Linux do not really compete with each other.

As I said in another comment, Chimera Linux is more an alternative to Void Linux than it is to Alpine. Arch fans might like it too.

Chimera Linux is quite different from Alpine. They both use APK and MUSL (and the Linux kernel) but that is all they have in common. And Chimera uses a different memory allocator than Alpine, so the even MUSL is quite different between the two.

I would say the closest distro to Chimera Linux is the MUSL version of Void Linux.

Chimera Linux has a full userland, mostly based on FreeBSD, while Alpine uses BusyBox. This is because Chimera Linux is meant to be a full “batteries included” general purpose Linux distro. Even though Chimera Linux uses APK, it has a totally different approach to building packages via the Cports system which is the best I have found on Linux. Chimera started with the idea that the Void Linux package build system could be improved. The innovations kept coming until a whole new distro was born.

Chimera Linux uses very little GNU software (none installed by default I think) and is a non-systemd distro. This includes using Clang / LLVM instead of GCC. So it is quite unique in the Linux world.

My choice of DE on Chimera is KDE Plasma or Niri but the lead developer uses GNOME. They plan to make GNOME work on dinit (not systemd). So, that will be quite unique as well.

As soon as an article starts by telling you that Wayland is 18 years old, you know where it is going to go. Yes, the very beginning of the Wayland experiment started long ago but it was not something anybody was expected to use most of that time.

The very first Wayland-only desktop environment ever, COSMIC, launched just last month. Should I write an article about how amazing Wayland is despite being so new?

A more neutral view might be to use Sway itself as a benchmark as it was one of the earliest Wayland compositors. The Sway project is less than 10 years old. The most complete Wayland environment available today, KDE Plasma, started to experiment with Wayland around then as well.

But Wayland has only really come into its own in the last 5 years with remaining edge cases regularly being addressed over the last two.

And we are now in a place where Wayland works for most people. The edge cases that remain are largely more exotic, like this guys 8K monitor. It would be dishonest to pretend Wayland’s evolution has been rapid. It has largely been dysfunctional. And real gaps remain. But it is already superior to X11 in many ways and the list of remaining use cases not well addressed continues to drop.

Yes, Wayland does a lot of stuff better than X11.

A Linux desktop user that started in Wayland a couple of years ago would be able to write a similarly negative article about Xorg if they tried to switch to it. The two systems are different. Neither is absolutely better than the other today. But Wayland is improving and Xorg is not.

And more than half of Linux desktop users run Wayland now. And 4 out of 5 new Linux desktop users start on Wayland and never switch. Linux is a Wayland first OS. So, when articles like this complain about how long it would take to reconfigure their systems for Wayland, they miss an important point. The Wayland way is the “correct” way now, or at least the most common way. The X11 config is the weird one.

And one of the things Wayland does better is run Wayland apps. The foot terminal mentioned in this article cannot be run on X11 at all. It is Wayland only. All of the apps an X user tries on Wayland will at least run. Not so the other way around.

When GNOME and KDE shed their X11 compatibility, they will be able to more freely innovate Wayland only features. As that starts to happen, it will become more normal to create Wayland only applications. This won’t be a problem as 80 percent or more of Linux desktop uses will be using Wayland-only desktop environments.

And that is what will ultimately doom X11. It will become impractical to run and X server instead of Wayland due to the important Wayland apps that cannot run on such a desktop.

Anyway, it was a well written article and mostly fair. It will be very interesting to see how the set of requirements fares 1 - 2 years from now. GNOME and KDE will be Wayland only. COSMIC will have matured. Wayland compositors will have standardized a bit.

I suspect that things will be looking very nice.

This was obviously meant sarcastically. And the recent Microsoft post is clearly more aspirational than real. But it is much more real than you probably think.

This is all about:

1 - existing ecosystems (eg. tooling)

2 - existing code bases

3 - existing dev skills

It is also a trick of absolute vs relative numbers. Rust can be growing at a MUCH faster rate and still add fewer developers annually than C++.

The C++ line will go up, peak, and then go down forever (probably never to zero)

This is just like saying China is adding more coal power plants. True. But at the rate solar is growing, that will not be true for long. And once the absolute number lines cross, the old tech decline will be as steep as the new tech rise was.

Well, I don’t know Fortran 66. So, obviously step one is having an LLM convert it to Python for me. /s

Well obviously with vibe coded stuff, you just put the code back in the AI and ask for documentation.

Problem solved. /s

COSMiC has made Iced and Smithay stronger. Now Niri is based on Smithay. I for one are happy they spent their time on something other than GNOME.

GNOME has been shedding market share to KDE and now COSMIC is going to take a chunk of the rest.

Sour grapes.