Update pivot_wider() documentation

[mgirlich]

Users are confused that pivot_wider() is not lazy (e.g. #598, #703). Collecting is only necessary to generate the spec.
As long as tidyr::pivot_wider_spec() is not made a generic I see two ways of handling this:

• just update the documentation (like it is done here) or
• export dbplyr_pivot_wider_spec() (which is lazy) so people can generate the spec locally.

@hadley Do you have any preference?

[hadley]

I think just updating the documentation is fine. Maybe it's worth mentioning why? (i.e. "pivot_wider() isn't lazy, and in fact can't be lazy, because we need to look at the data to figure out what the new column names will be). I guess we could also provide an additional argument that would let you supply the new names?

[mgirlich]

To me adding an argument for the names does not make much sense. This feels like a super simplified specification and this would require additional testing and implementation. So, I just add some more information to the documentation.