Support seq() for hms values

[DivadNojnarg]

Currently seq() fails on 'hms' objects due to 'hms' inheriting from 'difftime' and 'difftime' being difficult to work with generally. seq() being a S3 generic we can however define a method for 'hms' where we use vctrs::vec_cast(, numeric()) on its from, to, by arguments. And reconvert to 'htm' before returning.

@krlmlr: we had to use bquote to splice the parameters gathered in the ... as we can't use !!! with seq. It seems quite daunting. Maybe you have something easier?

Closes #117

[krlmlr]

Thanks. One way to do this with rlang:

args <- list(from = 1, to = 3, by = 0.5)

rlang::inject(seq(!!!args))
#> [1] 1.0 1.5 2.0 2.5 3.0

^{Created on 2025-03-26 with reprex v2.1.1}

[DivadNojnarg]

Not sure about the reason of the failure, as it does not seem to be related to the new seq.hms

edit: actually maybe because we needed a dev version of roxygen 2

Found the following Rd file(s) with Rd \link{} targets missing package
  anchors:
    hms.Rd: vec_cast
  Please provide package anchors for all Rd \link{} targets not in the
  package itself and the base packages.

[krlmlr]

Thanks, great. If we must use inject(), we can, but do we really have to?

[krlmlr]

Minor nit: we prefer single indent these days.

Suggested change

	seq.hms <- function(
	from = hms(1),
	to = hms(1),
	...) {
	seq.hms <- function(
	from = hms(1),
	to = hms(1),
	...
	) {

Is there a reason to not add by to the signature to avoid handling the dots?

[DivadNojnarg]

I refactored the code to have by in the signature and get rid of ... handling.

[krlmlr]

Thanks, nice!

[DivadNojnarg]

@krlmlr Let me know if you need anything else on this PR.

[krlmlr]

Thanks, merging.

[krlmlr]

Minor nitpick: if we check if (is.null(by)), the shortcut (immediate return) is much easier to follow.