Yes, writing safe code is a skill. But for most skills you need someone to teach you. My suggestion is, think about it differently:

The compiler isn’t training wheels, but rather a senior dev, that quickly looks over your shoulder and tells you, hey you might have missed something here. You can at any point tell them, trust me bro for this part. They need at some points a comment in the form of lifetime annotations. They are pretty good, but not perfect.

To pick up your example. They don’t tell you, you need a new wheel. They tell you, you’re bike has disc brakes, but your new wheel is missing the disk. When you later hit a wall, because you couldn’t break, you can check where you just said trust me.

The difference is, you put up a sign “here be dragons” with unsafe. Certain operations break the garanties of rust, like pointer operations and Cells. So now you need to write a wrapper which upholds checks that they are not violated. If it breaks, you just need to check your unsafe code, instead of everything. Now only one person needs to deal with unsafe for everyone in a team to benefit.

Memory safety isn’t a skill issue. There is a reason they happen in all big projects like android or chrome and only reduced when introducing safe languages.

I’d rather have a compiler that tells me why my code was refused than a compiler which prints 100 lines of templates. Or a compiler, which tells me a pointer/reference is null instead of a compiler that knows, but it’s UB-NDR so it isn’t snitching. If the compiler tells me, hey here you could have a race condition, it’s one less bug I have to find with a debugger. A compiler that complains about uninitiated variables, instead of a compiler that gives me potluck as content.

Some of these are bonkers. The argument not to fix them because of backwards compatibility is even wilder. Which normal client would need the ability to get data for any other account that it hasn’t the Auth token for.

There also IDEs that start instantly.

They don’t ship a whole browser though.

Also a good pen is the report, we need three months for this, only to be told each week something else is more important. After the three month the question came is it finished now.

And so 2 Years turned to 6.

For a program at work the scripting language is Delphi and the Documentation is not available. If you want to know something call the support. CLI is not an option.

It’d be really useful to have a CI run, but that is never going to happen.

This is an xy Problem. Voting to close.

We don’t know what value X has. If it isn’t initialised it could have any value including the maximum. Then it would overflow.

But let’s be honest, that is unlikely.

Fun and games till x overflows.

I don’t like the interface for setting up flows. Feels needlessly complicated.

Be the change you want to see.

I had a OoM with 16 GB and a Quad-Core while linking yesterday.

The version with Linux, we call “GNU/Linux.” It is OK to call it “GNU” when you want to be really short, but it is better to call it “GNU/Linux” so as to give Torvalds some credit.

I suggest we call it Linux for short. Good thing you can use Linux without GNU.

With nix it’s easy to revert, if you keep your previous config. Version it with git and it’s really easy.

Take another look, you’re already standing on the top track.

That’s exactly what the cloud is good for, just use a cheaper holster like hetzner for it.

Rust libs for use in the kernel need an extra approval. But it turns out, whenever it’s easier to just include a dependency doing what you need, instead of writing a bad version of it yourself, most people take a dependency.

Taking pictures of evidence is also a really bad idea.

Depends on your login flow. There is a session manager which normally boots up and let’s you choose. But you can also configure it to auto login and send you to the Lockscreen of your window manager.

I’m not sure if this is serious or not and that scares me.