Add Span equality (== and !=) operators.

[Ivorforce]

This happens to allow us to deduplicate a bunch of operator== from String and Vector.
It might be possible to remove explicit overloads in those functions, instead relying on implicit conversions to Span, but this is a bit risky compilation wise and thus out of scope for this PR. Will revisit this in a future PR.

Span<char32_t> comparison to Span<char> is somewhat questionable from a String perspective, because it implicitly interprets the Span<char> as latin1, although it's usually ascii (still correct) or utf-8 (incorrect).
Still, from a Span perspective this is entirely reasonable, because individual elements report equality. It is up to callers to differentiate encodings if they matter.

Benchmark

This PR happens to optimize the equality checks of Packed*Array (fundamentally typed Vector). The reason is that the old implementation lacked a memcmp optimization. I naively benchmarked a ~4x improvement on my computer. The improvement is likely to be better on modern x86_64 machines, due to wider SIMD support (up to 16x).

I benchmarked this code:

extends Node2D

func _ready() -> void:
	var a := PackedInt32Array()
	a.resize(500000000)
	var b := PackedInt32Array()
	b.resize(500000000)

	for i in 20:
		a == b

This resulted in:

 ❯ hyperfine -m5 -iw1 "bin/godot.macos.editor.arm64.master --path /Users/lukas/new-game-project --quit" "bin/godot.macos.editor.arm64 --path /Users/lukas/new-game-project --quit"
Benchmark 1: bin/godot.macos.editor.arm64.master --path /Users/lukas/new-game-project --quit
  Time (mean ± σ):     11.970 s ±  0.092 s    [User: 10.632 s, System: 0.543 s]
  Range (min … max):   11.856 s … 12.081 s    5 runs

Benchmark 2: bin/godot.macos.editor.arm64 --path /Users/lukas/new-game-project --quit
  Time (mean ± σ):      3.168 s ±  0.035 s    [User: 1.863 s, System: 0.512 s]
  Range (min … max):    3.129 s …  3.221 s    5 runs

Summary
  bin/godot.macos.editor.arm64 --path /Users/lukas/new-game-project --quit ran
    3.78 ± 0.05 times faster than bin/godot.macos.editor.arm64.master --path /Users/lukas/new-game-project --quit

[kiroxas]

Make sense to me and would avoid to have to construct strings to compare char* against spans (useful for other PRS )

[lawnjelly]

Looks sensible, the more we can use span for these comparisons the better imo.

[Repiteo]

Thanks!