Remove temporary backwards compat support for `list()` in `names_ptypes` and `values_ptypes` of `pivot_longer()`

[DavisVaughan]

See #1291 for even more details.

When we moved from default values of names/values_ptypes = list() to names/values_ptypes = NULL and allowed names_ptypes = list() to mean "use a list ptype for all columns", this caused breakage in extension packages that extended pivot_longer() and still had list() as their default value for these arguments.

So we temporarily special cased a value of list() for these two arguments to be identical to NULL to give the extension packages time to update to the new default. Documentation was updated to mention this special casing, and to state that it was temporary.

We'd like to remove this special casing once the packages are updated so that names_ptypes = list() will again mean "use a list ptype for all columns".

This issue tracks the status of the known packages that have this issue. Once they are all fixed, we can remove the special casing from the code and documentation:

• dbplyr Update pivot_longer() args taking single/multiple values dbplyr#744
• sf Update pivot_longer() signature r-spatial/sf#1895
• sparklyr Change pivot_longer() argument defaults from list() to NULL sparklyr/sparklyr#3221
• tidyseurat Change pivot_longer() argument defaults from list() to NULL stemangiola/tidyseurat#49

Once these issues are all resolved, revert everything done in this PR #1297

[mgirlich]

@DavisVaughan you haven't open the issues yet, have you? I'd like to tackle this in dbplyr soon.

[DavisVaughan]

I haven't opened any of these yet, if you want to do one for dbplyr that'd be great, and then I can take a look if you want

[DavisVaughan]

It looks like all 4 of the packages here have been updated on CRAN, so I think we can move on this