Make Vec::len() and Vec::is_empty() const fn

[GrayJack]

Making these two functions const fn will improve the usability in constant context.

This also open the possibility of contify String functions with the same names.

Maybe this is putting the cart before the horse, since we don't have const allocation yet and should be delayed till we have one

[rust-highfive]

r? @Mark-Simulacrum

(rust_highfive has picked a reviewer for you, use r? to override)

[Mark-Simulacrum]

r? @oli-obk

[memoryruins]

#65341 aimed to add these last month, but it was closed due to
not having a clear gain for usability in const fn at this time.

[oli-obk]

Yea, we're going to delay making these const fn until someone has a use case. That said, I wonder if we can coerce a vec to a slice in const eval and call slice's len method

[GrayJack]

Can't we use once_cell Lazy to have a initialized Vec constant?
Like this

const a: Lazy<Vec<i8>> = Lazy::new(|| vec![1, 2, 3]);

[memoryruins]

vec!, due to allocations, and fn pointers cannot be called directly in const contexts. Lazy obtains a fn pointer fn() -> Vec<i8> from the closure, which is called for initialization at first use during runtime. (To initialize only once, still at runtime, and at a specific memory location, Lazy is used in statics).

[JohnCSimon]

Ping from triage
@oli-obk - can you please review this PR?

[oli-obk]

Closing for now. If you come up with a use case where these are needed, please create an issue for it so it can be discussed independently of an implementation