Add IoSlice::as_bytes

[Lucretiel]

This PR proposes to add IoSlice::as_bytes. While IoSlice already has Deref<Target=[u8]>, this uses the lifetime of the IoSlice itself, throwing away the larger 'a lifetime; as_bytes allows for getting the original slice with the original lifetime.

ACP: rust-lang/libs-team#93
Fixes #124659

[rustbot]

r? @m-ou-se

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

[rustbot]

Hey! It looks like you've submitted a new PR for the library teams!

If this PR contains changes to any rust-lang/rust public library APIs then please comment with @rustbot label +T-libs-api -T-libs to tag it appropriately. If this PR contains changes to any unstable APIs please edit the PR description to add a link to the relevant API Change Proposal or create one if you haven't already. If you're unsure where your change falls no worries, just leave it as is and the reviewer will take a look and make a decision to forward on if necessary.

Examples of T-libs-api changes:

• Stabilizing library features
• Introducing insta-stable changes such as new implementations of existing stable traits on existing stable types
• Introducing new or changing existing unstable library APIs (excluding permanently unstable features / features without a tracking issue)
• Changing public documentation in ways that create new stability guarantees
• Changing observable runtime behavior of library APIs

[Lucretiel]

My own personal use case for this is an attempt to write my own write_all_vectored. I ran into an issue where, given an &mut [IoSlice<'a>], it's not possible to rewrite those slices in-place as the write proceeds, because slicing an IoSlice borrows it (instead of using the original 'a lifetime), which precludes rewriting it in-place.

[Lucretiel]

Looks like the classic sorts of failures where an innate method is added and starts to take precedence over a trait method. Maybe as_inner, instead?

[joboet]

How about into_bytes? That fits well with methods like OccupiedEntry::into_mut.

[aDotInTheVoid]

@rustbot label +T-libs-api -T-libs

[bors]

☔ The latest upstream changes (presumably #118154) made this pull request unmergeable. Please resolve the merge conflicts.

[Dylan-DPC]

@Lucretiel if you can address the CI failure and the conflicts, we can put this for review

[Lucretiel]

Taking a look now, thanks for your patience!

[Lucretiel]

Doctests are failing because of the use of an unstable library feature (the one added in this PR); can someone walk me through how to fix?

[aDotInTheVoid]

can someone walk me through how to fix?

Add the #![feature to the doctests. Eg

[Mark-Simulacrum]

Can we add the same for IoSliceMut? It looks like it has the same lifetime issue.

[Lucretiel]

Can we add the same for IoSliceMut? It looks like it has the same lifetime issue.

I'm actually pretty sure we can't, though I'd be happy to be proven wrong. A function resembling fn as_bytes(&self) -> &'a [u8] would end the self borrow when it returns, which would allow aliasing the mutable region. The best you can do is into_bytes_mut(self) -> &'a mut [u8], which doesn't offer many advantages over just dropping the IoSliceMut.

[Lucretiel]

In any case, this is (finally) ready to merge, just needs approval from someone.

[tgross35]

The issue number points to this PR; you should open a tracking issue and link it instead.

Also feature gate io_slice_as_bytes -> io_slice_into_bytes if that is the name.

Edit: io_slice_as_slice with the below.

[tgross35]

The ACP at rust-lang/libs-team#93 covers the change here. Two changes from this PR are that the accepted name was as_slice (though presumably this is up for bikeshedding), and that it can be implemented for IoSliceMut too. Could you make these changes, as well as updating the unstable attribute as mentioned above?

@rustbot author