🐛 add a DoubleColon to the Token struct in the lexer.rs

[LesterEvSe]

This prevents the lexer from allowing spaces between colons (e.g., a: :b),

[KyrylR]

Suggested change

	/// This prevents the lexer from allowing spaces between colons (e.g., `use a: :b`),
	/// ensuring we strictly parse valid paths.
	/// This prevents the lexer from allowing spaces between colons (e.g., `use a: :b`)

[KyrylR]

Explain in description why this is changed

[LesterEvSe]

After updating to Rust 1.94.0, cargo clippy started complaining about these lints. I just applied the automatic suggestions from Clippy to fix the warnings.

[KyrylR]

Same

[LesterEvSe]

The solution is the same as in the previous answer.

[KyrylR]

This is confusing, is param:::: now valid?

[LesterEvSe]

No. It's working as usual; the error is being shown by jet::::. So I think it's behaving similarly to param::::.

[stringhandler]

4fe0378
Seems fine but would be useful to have some tests that show the problem. Have a look here for a simple way to create one https://github.com/BlockstreamResearch/SimplicityHL/pull/213/changes#diff-5d5fd55cd63fa75e4c8902fa6b6d63da3f8da03742454636a8d9c092f6f6eeb4R346

[LesterEvSe]

4fe0378 Seems fine but would be useful to have some tests that show the problem. Have a look here for a simple way to create one https://github.com/BlockstreamResearch/SimplicityHL/pull/213/changes#diff-5d5fd55cd63fa75e4c8902fa6b6d63da3f8da03742454636a8d9c092f6f6eeb4R346

Done

[stringhandler]

24f4e44

I think let's add some tests for witness:: and witness::::

[stringhandler]

I think we can remove the double colons here now. The correct lexer should create [Token::Witness, Token::DoubleColon]

[LesterEvSe]

Is this OK?

    let jet = just("jet")
        .padded()
        .ignore_then(just("::"))
        .padded()
        .ignore_then(text::ident())
        .map(Token::Jet);
    let witness = just("witness")
        .padded()
        .ignore_then(just("::"))
        .padded()
        .ignore_then(text::ident())
        .map(Token::Witness);
    let param = just("param")
        .padded()
        .ignore_then(just("::"))
        .padded()
        .ignore_then(text::ident())
        .map(Token::Param);

Or do I need to change this part of the code?

let keyword = text::ident().map(|s| match s {
        "fn" => Token::Fn,
        "let" => Token::Let,
        "type" => Token::Type,
        "mod" => Token::Mod,
        "const" => Token::Const,
        "match" => Token::Match,
        "true" => Token::Bool(true),
        "false" => Token::Bool(false),
        _ => Token::Ident(s),
    });

If I add these kinds of words here, then I need to use Token::Jet("") or delete the (&'src str) payload here:

    // Jets, witnesses, and params
    Jet(&'src str),
    Witness(&'src str),
    Param(&'src str),

So, which approach should I choose?

[LesterEvSe]

I chose the option with the fewest changes (.padded(), ignore_then("::"))

[KyrylR]

So those are valid now

jet :: add_32, witness :: PK, and witness:: PK

?

[LesterEvSe]

Yes, it's an analogy to the Rust syntax.