ArgEnum support hidden attribute

[ModProg]

Please complete the following tasks

• I have searched the discussions
• I have searched the existing issues

Clap Version

master

Describe your use case

I want to hide one value in an ArgEnum from for example [possible values: GitHub, Test]

Describe the solution you'd like

I would like to write this, and only have GitHub show up:

#[derive(ArgEnum)]
#[clap(rename_all = "pascal_case")]
enum CodeProviders {
    GitHub,
    #[clap(hidden(true))]
    Test,
}

Alternatives, if applicable

No response

Additional Context

No response

[pksunkara]

Hmm.. can probably be implemented something similar to #2733. But looking at all these issues, do we need an ArgValue builder?

[ModProg]

Hmm.. can probably be implemented something similar to #2733. But looking at all these issues, do we need an ArgValue builder?

I thought this could be implemented by just omitting it from the possible_values function, but that is used for validation as well if I understand your answer correctly?

[pksunkara]

but that is used for validation as well

Yup, try the skip attribute on that value and try to use that value. It will lead to error.

[epage]

I've been considering some related API problems and I was wondering if we should have an ArgValue that includes

• Name
• About
• Hidden
• Completion Hint
• etc

We'd then change the signature for possible_values to

fn possible_values(values: impl Iterator<Item=impl Into<ArgValue>>)

where impl From<&str> for ArgValue sets the name.

• We currently accept a slice but just to treat it as an iterator, so we can generalize to an iterator
• Using Into<ArgValue> keeps the API simple for the common case while allowing for the complex cases
• We'd be able to stop ballooning the API for each case (hints, about, names, etc)
  • For the singular cases, we could still have the individual functions

[pksunkara]

Design looks good to me.

[epage]

This would be a breaking change, technically. Should we tag with 3.0?

Oh, and it should be IntoIterator and not Iterator

[ModProg]

I experimented in #2758 with something very similar, I would "upgrade" that to the proposed implementation.