TryFromPrimitive incorrectly uses default attribute

[ollie-etl]

If we have an enum where a #[default] is applied for another derive, TryFromPrimitive becomes useless, and will silently return the default. This is both undocumented, and makes no sense - surely the point of TryFromPrimitive is to test the input is valid. If the default is required, FromPrimitive has this behaviour documented.

#[derive(
    Debug,
    TryFromPrimitive,
)]
pub enum MyEnum {
    #[default]
    E1 = 1,
    E2 = 2,
}

assert_eq!(MyEnum::try_from(42).is_ok(),  "This passes!!!!")

Closely related to #31, however the fix there cannot be applied, because the attribute is actually required for another macro.

[cgbur]

The issue is that there is documented in the readme #[num_enum(default)] but this is also applying to #[default] from core. This is a footgun that should be fixed. Is the crate author open to working on a fix for this?

[illicitonion]

Sorry for the super slow response here - yes, I agree that TryFromPrimitive should probably ignore default attributes completely, and only FromPrimitive should pay attention to it. This will, however, be a breaking change, so we should probably bump to 0.6.0 with this change. I'll put together a PR soon.