Refactor sleep.rs

[Big-Iron-Cheems]

This pull request refactors and improves sleep.rs, focusing on safety, idiomatic Rust, and robust error handling.
Each commit message should explain why I did these changes, with annexed sources for more techinical ones.

Key improvements include:

• Explicit compile time warning for unsupported targets.
• No busy waiting in loops via hlt instruction.
• More idiomatic error reporting with a SleepError enum and report_error function.
• Safer and simpler handling of raw pointers and argv strings.

---

Note: I believe that the Rust implementation could benefit from providing Cargo.toml and .cargo/config.toml.
Personally I settled with the following versions:

[package]
name = "sleep"
edition = "2024"

[[bin]]
name = "sleep"
path = "sleep.rs"
test = false
bench = false

[profile.dev]
panic = "abort"
opt-level = 0
lto = false
codegen-units = 1

[profile.release]
panic = "abort"
opt-level = 3
lto = true
codegen-units = 1

[dependencies]

[build]
target = "x86_64-unknown-linux-gnu"

[target.x86_64-unknown-linux-gnu]
rustflags = [
    "-C", "link-args=-nostartfiles",
    "-C", "panic=abort",
    "-C", "opt-level=3",
    "-C", "strip=debuginfo"
]

It would be possible to have a structure like this, but it is out of scope.

.
├── rust/
│   ├── .cargo/
│   │   └── config.toml
│   ├── Cargo.toml
│   ├── Cargo.lock
│   └── src/
│       └── sleep.rs
└── c/
    ├── Makefile (or CMakeLists.txt)
    └── src/
        └── sleep.c

---

Thanks for the informative YouTube video regarding this topic, I found it quite nice.

[valignatev]

what in the AI slop is this shit

[Big-Iron-Cheems]

If my way of writing longer than average commit messages comes off as LLM slop there's nothing I can do about that I'm afraid.
Personally I prefer clarity over brevity.

If this is about me linking sources in some commits, it's because for some changes I had to look at actual docs since I wasn't familiar with some of the ASM flags used here.