Tracking Issue for slice_strip feature

[tesuji]

The feature gate for the issue is #![feature(slice_strip)].

This feature adds two new methods to [T]:

impl<T> [T] {
    /// Returns a subslice with the prefix removed.
    pub fn strip_prefix(&self, prefix: &[T]) -> Option<&[T]> where T: PartialEq;
    /// Returns a subslice with the suffix removed.
    pub fn strip_suffix(&self, suffix: &[T]) -> Option<&[T]> where T: PartialEq;
}

This is an analogy feature like str::strip_{prefix,suffix} but has the same signature of slice's {starts,ends}_with.

About tracking issues

Tracking issues are used to record the overall progress of implementation.
They are also uses as hubs connecting to other relevant issues, e.g., bugs or open design questions.
A tracking issue is however not meant for large scale discussion, questions, or bug reports about a feature.
Instead, open a dedicated issue for the specific matter and add the relevant feature gate label.

Steps

• Implement the feature (cc @rust-lang/libs)
• Adjust documentation (see instructions on rustc-dev-guide)
• Stabilization PR (see instructions on rustc-dev-guide)

Unresolved Questions

None

Implementation history

• Implemented in Implement slice_strip feature #73414
• Assembly for [u8]: https://rust.godbolt.org/z/iMcqf-

[ijackson]

Hi. I accidentally used one of these functions, I think because I was reading the docs for str :-). So, I conclude that they're useful.

I'd like to see them stabilised. Since the docs tickbox was unticked I looked at the docs, and I have I made an MR for some slight docs wording improvements: #75078

I think this is probably ready for stabilisation.

@lzutao Thanks for your contribution here. Do you intend to do the stabilisation report etc.?

[tesuji]

I think it is a bit soon to stabilize this feature. It is about one month after the feature is merged and appeared on nightly.
But yeah, you could send a PR to propose stabilize this feature, but it should include real use cases.

[Rua]

I'm also interested! I want to use this to trim null bytes from the end of a slice.

[tesuji]

Could you give some example code show how do you do that with slice::strip_ methods?
Because if you are running strip_* in a loop, I think it is kinda not cool, it is might be better to
introduce a new method for that functionality.

[Rua]

You're right, what I'm looking for is a slice equivalent to str::trim_end_matches. Which would also be welcome, but out of scope here!

[tesuji]

I just don't know if we want the API as it is, because we could bound the prefix/suffix arguments by Pattern API
like str. @withoutboats and some folks was talking about fixing Pattern API for slices.

Anyway, If you want to stabilize this API, you could just send a PR and ping T-libs for their decisions.

[ijackson]

lzutao writes ("Re: [rust-lang/rust] Tracking Issue for `slice_strip` feature (#73413)"):

I just don't know if we want the API as it is, because we could bound the prefix/suffix arguments by Pattern APIs like str. @withoutboats and some folks was talking about fixing Pattern API for slices.

AIUI providing these functions now, as taking slices for the patterns, does not preclude generalising things later with a slice Pattern API. The current unstable strip_* functions match the signatures of the starts_with and ends_with, as discussed in #73414. Ie, AIUI subsequently introducing such a alice Pattern API would not be a breaking change. If I'm wrong about that then stabilising starts_with and ends_with in their current form seems like it was a mistake ?

Anyway, If you want to stabilize this API, you could just send a PR and ping T-libs for their decisions.

The stabilisation checklist says to look at the docs. I have an outstanding MR #75078 to improve the docs. Unfortunately that seems to have become stuck waiting for review. Would you be able to look into that ? Thanks, Ian.

…

-- Ian Jackson <ijackson@chiark.greenend.org.uk> These opinions are my own. Pronouns: they/he. If I emailed you from @fyvzl.net or @evade.org.uk, that is a private address which bypasses my fierce spamfilter.

[tesuji]

The current unstable strip_* functions match the signatures of the starts_with and ends_with, as discussed in #73414.

Ah yeah, I forgot about this, thanks for pointing it out.

Ie, AIUI subsequently introducing such a alice Pattern API would not be a breaking change. If I'm wrong about that then stabilising starts_with and ends_with in their current form seems like it was a mistake ?

I'm not really sure, T-libs member should have more experience to answer this question.

Would you be able to look into that ?

My English is not great, so I pinged Joshua for review your PR.

The stabilisation checklist says to look at the docs

You could always send a PR for stabilization, documentation improvements is not a blocker for stabilizing it.

[ijackson]

My English is not great, so I pinged Joshua for review your [docs] PR.

Thanks.

You could always send a PR for stabilization, documentation improvements is not a blocker for stabilizing it.

Oh, I didn't realise that. I will do a stabilisation PR then. Thanks, Ian.

…

-- Ian Jackson <ijackson@chiark.greenend.org.uk> These opinions are my own. Pronouns: they/he. If I emailed you from @fyvzl.net or @evade.org.uk, that is a private address which bypasses my fierce spamfilter.

[ijackson]

I think this issue wants to be set to merged or something now? The docs were dealt with in #75078 and the stabilisation was in #77853, both merged now. I'm not sure I can do that myself...

[tesuji]

Closed by #77853 .