amigan

Ever hear of Windows NT? Legacy DOS code was relegated to userspace long ago. And NT was designed by Dave Cutler, the guy who designed VMS and RSX-11. Most certainly not “shit piled on top of shit.” Unix could have learned a thing or two from Dave but by then it was too late; the plane had very much taken off.

The Amiga’s OS was written in 68k asm and BCPL (C’s grandfather) for the kernel, and C was used for the utilities and GUI. But this was considered quite ahead of its time for a single user micro.

Most critical sections should be.

You’re thinking of Shit Jobs, the FruitCo charlatan. Gates’ pancake sort algorithm held the speed record for 30 years.

“Who the hell” writes an OS in assembler in the 80s? Uh, some of the utilities are in C, but compilers were slow and generated slow code back then, and it was quite noticable on a slow machine. When every byte of memory counts, you often need to hand-optimize.

His last product was the OS for the Tandy model 100 in 1983.

restic is better.

Amazing, thank you for this! I love ffmpeg so much. If ever was an underappreciated workhorse that you’ll find enabling almost everything people do with video.

And they can even use coreboot supposedly, but I haven’t had the butthole to flash it.

Also, ditch pihole for blocky. It’s faster and uses less RAM, and is far less complex. And it will run right on FreeBSD, no separate machine required. Couple it with prometheus and grafana and you even get pretty graphs just like pihole.

Yes, under which to do anything worth a damn, you will be using open source toolchains, libraries, and quite possibly whole applications.

If you can get it to run on arm64, and you can manage to provision an instance (they are in seemingly perpetually limited supply), evil Oracle actually has a pretty generous always free tier Ampere offering. You can have up to 4c/24GB instance for free, sliced and diced as you see fit (4 1c instances, for example). You only get 200GB of block storage, though, so you’d have to either pony up for that or use object storage (not sure if nextcloud can use S3-like API object storage).

Preach it.

“Different inode” means a different file entirely, not necessarily its majorminor:inode tuple resolved through bind mounts/overlayFS/whatever. I’m saying that if you have containers using even slightly different base images, you effectively have n copies of libc in memory at once on the same system, which does not happen when you do not use containers.

Except each container has its own libc and any other dependencies. If any linked binary or library has a different inode, it gets loaded separately. I would say it is indeed quite similar, even if the images in question here aren’t hundreds of megabytes in size like with Electron.

It seems every new shiny technology today tries its darndest to short-circuit 40+ years of advances in OS virtual memory design. Between Electron and Docker, the entire idea of loading an image into memory once and sharing its pages among hundreds of processes is basically dead. But at least there’s lower support burden!!!1111

gdu is in winget, and it’s extremely fast.

If you think OpenAI themselves are any better than Google, I have a bridge to sell you.

By definition, if you don’t feel like putting in the homework, you are ceding control to someone else. At that point, all bets are off. Even trustworthy entities can turn on a dime. Ease and full control are mutually exclusive.

What exactly do you want to achieve in terms of security?

I’m beginning to think many people here just want to throw money at a vague concept of “security” without having a crumb of a threat model in mind.

At least he didn’t say viola.