@penguin42@hrw I hate submodules very much too. However, their implementation has become far less buggy in the last couple of years. It used to be extremely easy to break your repository completely with them. But in 2025 you can live with them with little trouble for months, maybe even years. BTW, it's "git submodule update". Don't do -f. The main remaining reason for my hate is how easy to is to forget to update them. And a less important reason is that branches are prevalent in modern git workflows (instead of cloning), and they conflict with submodules.
Embed this noticeAccount: Computers (pro@mu.zaitcev.nu)'s status on Wednesday, 16-Jul-2025 06:10:44 JST
Account: ComputersI'm so out of date that every time I see "Apply principles of SDLC and methodologies like Lean/Agile, CI, Software and Product Security, Scalability, Documentation Practices, refactoring, and Testing Techniques", I wonder what does Synchronous Data Link Control has to do with Agile and the rest of these buzzwords. In other news, the world has ran out of TLAs and is expanding into FLAs.
@mcdanlj@KF0UNK I can just imagine identifying VORs without even looking at the chart. "Charlie, November, Xray - okay Corona good to go", NAV sound off.
@sun@nuko I enjoy it greatly. Way better than C++ and D.
Now, I still remember learning and it noticing that for a language that young it had a surprising amount of warts. And it was like that from Day One. But then I also remember how much Python had to evolve (xrange lol). So yeah, Go had to add a whole new type to represent IP address, because they forgot to make the original one comparable. But it's insignificant next to the power of the force, I mean, next to C++ references and STL. Or D's templates for that matter. Go is a perfect language for someone of my limited cognitive capacity.
TBF the demise of GOHOME made the biggest contribution to my acceptance of Go. When it became possible to store source where I want, check them into git, and use Makefiles, things changed completely.
@sun@sendpaws@zaitcev One thing makes the kind of rage about this. The original IBM BladeCenter had a normal VNC. I used that. It was before Vinagre became widespread, but a basic VNC client worked. There was a way to supply ISOs with something like NFS too. But they switched to that retarded Java applet thing later. Enshittification hasn't began yesterday.
dnf update reboot all is good go get coffee come back the screen is blank does not wake up the system is completely okay otherwise, but the screen is blank
This is basically a great incentive to NEVER UPDATE. Fedora tend to test their releases, but once the release is out of the door, updates drag in all sort of regressions. They have absolutely no checks from contaminating after a release.
At least when an AppImage is distributed, you just run the image and you get a running application. The container image contains everything. But Flatpak does not!
I remember that some time around 1991 I mentioned this UNIX thing to Prof. M.V. Makarov-Zemlyanski and the conversation somehow ended on the topic of the of the text files with EOL delimiters. Maybe he wanted to re-encode e-mail on the fly, I don't remember precisely.
Either way, I said, "oh by the way, we also have tabs that you have to expand if you do this". And he said, "I suppose I could... where is the tab column table?" And I'm like... uhh... just generate the table at every 8th position please, Professor? He then gave me the meme look (30 years before the meme).
To be honest, the accumulated cruft of TLS offload, firewall-cmd, namespaces, systemd-resolvd, DNSSEC, and such made the modern networking a fairly unfun mess. But you don't have to comply, within your own network.
@sun@fiore@Suiseiseki@nyanide@meso The speed of modern CPUs that we take for granted comes from these sources: - a high performance DRAM controller - a large hierarchical cache - out-of-order execution - super-scalar or parallel execution - a high clock rate - all of this is underpinned by a small feature size process and a large gate budget
A typical RISC-V aims for a low cost implementation that lacks these elements. Without them, you're basically making a 200 MHz, 5 to 10 pipeline stage CPU. I'm generalizing it a lot, but you're ending with something like MicroSPARC-II or thereabout.
I heard about high performance RISC-V implementations aimed at hyperscalers, but I've never seen one in person.
Of course, connectivity is just fine, and applications work when they do not care about that. For example, WatsUp simply connects where it wants. But Chrome is too polite for that and refuses to try.
A tcpdump shows that Android connects, attempts SSDP (which never worked in my networks), then pokers the router (sic) with HTTP, and then reports the there's no connectivity. All Android devices behave the same way.