Tracking Issue for `const_char_encode_utf8`

[bjoernager]

Feature gate: #![feature(const_char_encode_utf8)]

This is a tracking issue for supporting char::encode_utf8 in const scenarios.

Public API

impl char {
    pub const fn encode_utf8(self, dst: &mut [u8]) -> &mut str;
}

Steps / History

• [x] Implementation: #130511
• [x] Address diagnostics: #130611
• [ ] Final comment period (FCP)
• [ ] Stabilization PR

Unresolved Questions

• None yet.

[dtolnay]

@rfcbot fcp merge

const for https://doc.rust-lang.org/nightly/core/primitive.char.html#method.encode_utf8

[rfcbot]

Team member @dtolnay has proposed to merge this. The next step is review by the rest of the tagged team members:

• [x] @Amanieu
• [x] @BurntSushi
• [x] @dtolnay
• [ ] @joshtriplett
• [ ] @m-ou-se

No concerns currently listed.

Once a majority of reviewers approve (and at most 2 approvals are outstanding), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up!

See this document for info about what commands tagged team members can give me.

[BurntSushi]

I'll note that in the implementation PR, making this const seems to require making the panic error message worse (even in the case where encode_utf8 isn't used in a const context) and also introducing more uses of unsafe. I'm not especially happy about that.

[bjoernager]

The use of unsafe code cannot be avoided as long as slice indexes aren't const-friendly (AFAIK). But I would argue that this use is properly documented and (relatively) invulnerable to UB, although I am open to other views on this.

@RalfJung has also noted that core::intrinsics::const_eval_select could be used to regain full diagnostics in non-const calls – a feature which I would be more than happy trying to implement.

[RalfJung]

We could have a safe "const fn subslice" that does the bounds check and then uses unsafe from_raw_parts. That would avoid introducing new unsafe reasoning in encode_utf8 itself.

[bjoernager]

130611 resolves the diagnostics question. Can this be added without conflicting with the FCP?

[BurntSushi]

@bjoernager Yeah that should be fine. Thank you! I appreciate that. :-)

[RalfJung]

Plesse add a FIXME(const-hack) next to the const-eval-select to document that it should be removed again in the future, when possible.

[bjoernager]

Done. :)

[rfcbot]

:bell: This is now entering its final comment period, as per the review above. :bell: