Allow collection setters to be generic over Into

[uint]

Apparently this works without doing any special trickery, at least on my version of the compiler.

Closes #209

[TedDriggs]

Thanks for submitting this; I haven't had the time to review it yet. @andy128k I believe you were the original author; can you take a look at these changes please?

[andy128k]

This will be a breaking change, so a minor version bump is needed.

I would also like to see how/if this works with maps (HashMap/BTreeMan). I suppose maps with a String key are pretty common and being able to use str literals would be a huge improvement.

[andy128k]

into_item = quote!(::derive_builder::export::core::convert::Into::into(item));

[uint]

I would also like to see how/if this works with maps (HashMap/BTreeMan). I suppose maps with a String key are pretty common and being able to use str literals would be a huge improvement.

I think for this to Just Work there would have to be some generic thing of this sort in the std:

impl<T, U, FromT: Into<T>, FromU: Into<U>> From<(FromT, FromU)> for (T, U) { ... }

Or maybe something like this?

impl<T, U, FromT: Into<T>, FromU: Into<U>> Extend<(FromT, FromU)> for HashMap<T, U> { ... }

But there is not. The best I can think of is adding a separate into_pair syntax, like so:

#[builder(setter(each(name = "bar", into_pair)))]

On the bright side, this would let us avoid having to pass a tuple. We could make the function accept two arguments - key and value.

[TedDriggs]

I'm going to rebase this and address the conflicts. Given that each has been in the wild for some time now, I'm inclined to create a wrapper for Each that will enable it to read the current each = "..." form so there's no breaking change; #[builder(setter(each = "..."))] would be the same as #[builder(setter(each(name = "....")))]. @uint / @andy128k thoughts on that?

[andy128k]

Frankly, I don't like how it becomes lisp-ish with all those nested constructs, so I welcome idea to continue to support current (relatively) flat syntax.

[TedDriggs]

I've updated the branch here, but had some difficulty with strip_option and Option wrappers around collections.

[TedDriggs]

This merged as #234