Closed dgkf closed 2 years ago
Upon investigation, it looks like I had good reason for doing this in the first place.
The default pmin
/pmax
, though they will work "out of the box" with S3/S4 values that implement comparison operators, rep
and is.na
, this uses the result of is.na
directly to determine how to handle missingness, which is not a binary yes or no in the case of a partial time.
Reviewing how I handled
pmin
andpmax
, I think the reexporting is a sloppy approach.base::pmin
/pmax
will respect classes that implement S4 versions of comparison operators, so these behaviors should "come for free" without doing any disrespectful masking.I need to research this more to explore whether it's compatible, but at first glance it looks like it provides a more compatible path forward.