Add u8::to_ascii_digit()

[hniksic]

When writing parsers for textual or semi-textual data received from an external source, one often gets octets that need to be interpreted as decimal numbers. For example:

let mut slice: &[u8] = ...;
// get the next octet in `slice` as hex digit, ignoring hyphens
let digit = loop {
    let c = *slice.get(0)? as char;
    slice = &slice[1..];
    if c == '-' {
        continue;
    }
    break Some(c.to_digit(16)? as u8);
}

While the compiler might be able to optimize this code sufficiently, a function like u8::to_ascii_digit() would allow us to operate on u8 directly, making the meaning clearer. It would be analogous to the already existing is_ascii_digit() and to_ascii_{lower,upper}case(). In the case of the above code, it would look like this:

let mut slice: &[u8] = ...;
let digit = loop {
    let d = *slice.get(0)?;
    slice = &slice[1..];
    if d == b'-' {
        continue;
    }
    break Some(d.to_ascii_digit(16)?);
}

The advantage is that it's clear we're working on octets the whole time. It avoids the cast to char which made the code look like it handled non-ASCII characters, when it in reality didn't, nor was it supposed to.

[gimbling-away]

@rustbot claim

[gimbling-away]

#96110 should fix this.

[Stargateur]

I'm not sure it's a good idea, this could make people confuse what is what.

A u8 is not supposed to be a character, std force the user to convert a u8 to char is a nice way to prevent mistake. IMO the first sniped is more explicit than the second

[gimbling-away]

std force the user to convert a u8 to char is a nice way to prevent mistake. IMO the first sniped is more explicit than the second

Can you elaborate on this part please? I couldn't really get what you mean @Stargateur

[CraftSpider]

I think there's good intuitive symmetry with the other ascii* functions already on u8. With the provided snippets I find both equally explicit, but I think having this method goes well with the other ascii methods already available, and is equally explicit in general.

[camsteffen]

The existing methods u8::is_ascii_digit and u8::Is_ascii_hexdigit don't have a radix argument.

[m-ou-se]

I commented on #96110:

From the name it's not clear to me what direction this conversion goes.

Does b'5'.to_ascii_digit() give me 5? Or does 5u8.to_ascii_digit() give me b'5'?

The 'to ascii' in the name makes me think this returns an ascii character (b'5'), but it seems to do the opposite.

[m-ou-se]

As indicated on the issue templates, it's usually best to discuss feature requests on https://internals.rust-lang.org/ rather than opening an issue for them.

[gimbling-away]

Proposal to close this since the PR has been closed? [See PR thread for discussion]

[thomcc]

In #96110 (comment) there are several reasons this should probably be closed. Further work on this (perhaps in the form of a more fleshed out {core,std}::ascii as mentioned) should be done as an API Change Proposal, rather than in this issue -- see https://std-dev-guide.rust-lang.org/feature-lifecycle/api-change-proposals.html for more info on that if desired.

[thomcc]

(Ah, my bad, thanks)