Allow dynamic LabelText closure to return different types

[jplatte]

This should decrease copying (especially if the user-provided closure returns Arc<str> or something that can be converted to it without copying) and make comparisons in resolve faster.

[jplatte]

CI failure looks unrelated.

[maan2003]

rebase should fix CI

[maan2003]

Thanks

[jplatte]

Oh, CI failure is related. I'll update the examples tomorrow.

[jpochyla]

I think one downside is that dynamics will often return format!()ted strings, and that would have to reallocate.

[maan2003]

I think one downside is that dynamics will often return format!()ted strings, and that would have to reallocate.

This already happens right? We always store data as Arc<str>

[jplatte]

Updated the one example affected by this and added a changelog entry.