Inconsistency between code and comment

[YichiZhang0613]

In rand-master/rand_core/src/le.rs, following comment and code are inconsistent.

/// # Panics
///
/// If `dst` has insufficient space (`4*dst.len() < src.len()`).
#[inline]
#[track_caller]
pub fn read_u32_into(src: &[u8], dst: &mut [u32]) {
    assert!(src.len() >= 4 * dst.len());
...

/// # Panics
///
/// If `dst` has insufficient space (`8*dst.len() < src.len()`).
#[inline]
#[track_caller]
pub fn read_u64_into(src: &[u8], dst: &mut [u64]) {
    assert!(src.len() >= 8 * dst.len());
...

[newpavlov]

assert! panics if the condition is false, so the function docs are correct.

[newpavlov]

I misread the function doc because of the swapped src and dst. The assert and docs indeed do not match.

[newpavlov]

@dhardy
It seems the asserts are wrong? They assert that src is bigger than dst. Maybe we should just assert equality here?

Technically, it would be a breaking change, but I don't think it makes sense to sue these functions with buffers not equal to each other byte-wise.

[dhardy]

No; the objective is not to copy all of the input but rather to fill dst. For that the assert is correct not the docs. Thanks @YichiZhang0613.

I'm not sure why we don't assert that the input is exactly the right length though.

I also don't much like the name or wording of the docs; I'm tempted to rename to read_u32_slice_from_bytes etc. — or just remove these methods and give a doc example somewhere (it's not much code and apparently only used in SeedableRng::from_seed impls, and already has #[inline]; one of the two methods is used by most PRNGs however).

[newpavlov]

Well, removing them would certainly be a breaking change, but deprecation sounds good to me.

[dhardy]

I fixed the docs