Revert removal of label text inference and allow nil

[paulcsmith]

#1028 (comment)

Actually I may just be overthinking this. Lucky wants to avoid bugs,
but also be practical. Sometimes nils can sneak in when you don't expect
it...but maybe label_text is not one of those places. I have a hard time
seeing a use case where someone is conditionally generating a label and
it might accidentally be nil. I think of all the apps I've ever written
I've never accidentally set label text to nil.

So I think what I'll do is make the label text String?. So that if you
pass nil to text it will infer the label name like it does now when you
leave the arg off completely. That way we can still keep the component
flexible by allowing String? but we also keep label_for flexible. I'll
go ahead and make that change now. This also means existing apps will
still work when you upgrade and this will no longer be a breaking change
👍

[paulcsmith]

Thanks for your feedback on the issue @wontruefree. I think this should be more flexible than my previous solution and will make writing Field components a lot simpler