feat: support fake variadic when length 1 not included

[kingwingfly]

Objective

fixes #42

It also supports fake variadic when length 1 not included.

Solution

Instead of n == 1 to judge if it's the first one, I tried n == input.start.max(1). This ensure that the first tuples will always have attributes. So that length 1 does not need including.

let doc = LitStr::new(
                    &format!(
                        "This trait is implemented for tuples **{down}up to {up}** items long.",
                        down = if input.start != 0 {
                            format!("down to {} ", input.start)
                        } else {
                            "".to_string()
                        },
                        up = input.end
                    ),
                    Span2::call_site(),
                );

And omit document like down to 0.

Testing

I checked with this command:

RUSTDOCFLAGS="--cfg docsrs" cargo +nightly doc --example all_tuples --no-deps --open

And

all_tuples!(
    #[doc(fake_variadic)]
    impl_tuple_foo,
    0,
    15,
    F
);

all_tuples!(
    #[doc(fake_variadic)]
    impl_tuple_foo,
    1,
    15,
    F
);

all_tuples!(
    #[doc(fake_variadic)]
    impl_tuple_foo,
    3,
    15,
    F
);

works as expected.

I also passed cargo test

[BenjaminBrienen]

Can you include a screenshot of the generated documentation so that I can see it working?

[kingwingfly]

Instead of n == 1 to judge if it's the first one, I tried n == input.start.max(1). This ensure that the first tuples will always have attributes. So that length 1 does not need including.

I just fixed some bugs... And the previous solution is wrong...

However, the new version is tested below:

all_tuples!(
    #[doc(fake_variadic)]
    impl_tuple_foo,
    0,
    5,
    F
);

---

all_tuples!(
    #[doc(fake_variadic)]
    impl_tuple_foo,
    1,
    5,
    F
);

---

all_tuples!(
    #[doc(fake_variadic)]
    impl_tuple_foo,
    3,
    5,
    F
);

cargo +nightly expand --example all_tuples (no implementation for (F, ))

RUSTFLAGS="--cfg docsrs" cargo +nightly expand --example all_tuples (only with --cfg docsrs, there's implementation for (F, ))

---

all_tuples!(impl_tuple_foo, 1, 5, F);

[alice-i-cecile]

This seems quite reasonable and I'd like to fix this.

Can you add some tests to the library itself to ensure that this keeps working?

[kingwingfly]

I do not know how to verify documentation generated by cargo-doc right now.

Anyway, I will try in a few days. (I'm preparing my final exam and am very busy at this moment.)

[kingwingfly]

I've added tests.

Both cargo test and RUSTFLAGS="--cfg docsrs" cargo +nightly test should be executed.

However, I do not know how to change the ci...

[BenjaminBrienen]

That's some good stuff right there

[BenjaminBrienen]

@alice-i-cecile Unicode 3.0 should be fine, right?

[alice-i-cecile]

Yep that's fine.