Open nikosbosse opened 1 month ago
Pinging @seabbs @Bisaloo @sbfnk @elray1 @nickreich @jamesmbaazam @jhellewell14 for people who might have opinions
I don't feel that strongly either way on this but I am not sure its working picking up a dependency just for this?
We could avoid that by just documenting that one could use purrr::partial
(and maybe it would need to be a suggests) so that seems like a better way to go vs wrapping it.
One key question here is are we likely to add further checks in the future or other custom functionality?
One key question here is are we likely to add further checks in the future or other custom functionality? Unsure I guess. You had a vision for actually doing proper checking of functions which might play into this.
So current plan is to document and make clear to users that this is basically the same as purrr::partial
?
No that is one option. The other is to leave as is.
It depends on if we like having it as part of the main stream docs and if we imagine future functionality. Do we?
I think of these I am leaning towards keeping as is
My idea (but I may be missing some subtleties) would be to remove customise_metric()
and update the docs to always recommend purrr::partial()
in its place.
As discussed in #791:
Should we replace
customise_metric
bypurrr::partial
?pro:
NAMESPACE
exportspartial
don't have to deal with a new functioncon:
Rogue alternative: using
customise_metric()
as an alias that directly callspartial
?full function code: