[ziml77]

The [0..n] slicing is what I was surprised to not see being done. As for what you're suggesting, using from_utf8 adds a overhead to check that you're giving it valid UTF-8 (this does not get optimized away). Dipping into unsafe can help here:

    const INDENTS: [u8; 256] = [b' '; 256];
    // ...
    IndentStyle::Spaces(n) => unsafe { std::str::from_utf8_unchecked(&INDENTS[0..n as usize]) },

But I would consider having INDENTS be a &str to get the cheap slicing without needing unsafe. Since putting in a string literal of 256 spaces is nasty I would use the const_format crate to generate the constant:

    const INDENTS: &str = str_repeat!(" ", 256);

[heliostatic]

Closing the loop, this is now implemented and on main, thanks to msdrigg: https://github.com/helix-editor/helix/pull/7429

[otikik]

This makes me happy, thanks for sharing

[messe]

Nice!

[messe]

I'm not sure I could justify bringing in a package for a one-liner, just to avoid an "unsafe" which could be easily documented in a comment as actually safe.