drop search needs to return a tibble

[karthik]

Or some other way to wrestle down that giant nested list

[KasperSkytte]

agreed :)

[ClaytonJY]

@KasperSkytte the trick is handling nested fields returned by the API, and affects all the functions that get back metadata objects (most of them). Do you have a preference between these two?

1. flatten list, so we have columns like sharing/something, sharing/somethingelse (could change separator or remove prefix)
2. nest list into list-cols, so we have a single sharing list-column, where each row has it's own something and somethingelse components, that the user could then un-nest as needed

We currently do the former in other places (drop_dir), but I suspect the latter might be a bit more tidy, and the flattened names seem odd.

Related: #110

[KasperSkytte]

I would vote for option 2. And maybe return 2 objects in a list, where one is the whole story and the other a simplified object, or data.frame-like object, with the most relevant information like paths and filenames.