feat: '+' / '-' methods for {nanotime} objects and 'difftime()' objects

[trevorld]

Adds the following S4 methods:

* nanoduration() + difftime()
* difftime() + nanoduration()
* nanoduration() - difftime()
* difftime - nanoduration()
* nanotime() + difftime()
* difftime() + nanotime()
* nanotime() - difftime()
* nanoduration() + nanotime() (arguably missing S4 method)

closes #107, closes #108

[eddelbuettel]

Thank you -- that too looks good. I will wait for @lsilvest to nod as well before merging.

[lsilvest]

If we do have difftime for nanotime and nanoduration, I think for coherence we also want to do it for nanoival, so for that type too we can add/subtract directly a difftime.

Also I would ideally prefer the unit tests to be distinct for each function added. Admittedly they are not very complicated, but one never knows what can break down the line and hence if one will need these separate unit tests.

[trevorld]

@lsilvest Each new method now has its own individual test and I've added the following new S4 nanoival() methods:

* nanoival() + difftime()
* difftime() + nanoival()
* nanoival() - difftime()
* nanoduration() + nanoival() (arguably missing S4 method)
* nanoival() - nanoduration() (arguably missing S4 method)

[eddelbuettel]

Thank you @trevorld for the added (and updated) PR and to @lsilvest for catching that nanoival was needed too.

If this completes the set of changes @trevorld is planning to bring I can look into a quick release in a few days. (Got a few emails from CRAN about clang-15 changes I should chew on first so likely early next week for this if we are good with that.)

[trevorld]

This completes the set of changes I was planning.

[eddelbuettel]

I can look into a quick release in a few days

Got around to that late yesterday, and by now Uwe processed and the new version is on CRAN.

Many thanks again to @trevorld for all the careful patches in it.