`ptype` argument of `unnest()` doesn't work in CRAN version

[DavisVaughan]

Was somewhat horrified to learn that the ptype argument doesn't work as expected in unnest() here because of the preprocessing that is done on the unnest() side before handing off to unchop():

library(tidyr)

df <- tibble(
  terms = list("x", c("y", "z")),
  predictors = list("a", c("b", "c"))
)

ptype <- tibble(terms = character(), predictors = character())

unnest(
  data = df,
  cols = c(terms, predictors),
  ptype = ptype
)
#> Error: Can't convert <data.frame> to <character>.

unchop(
  data = df,
  cols = c(terms, predictors),
  ptype = ptype
)
#> # A tibble: 3 × 2
#>   terms predictors
#>   <chr> <chr>     
#> 1 x     a         
#> 2 y     b         
#> 3 z     c

This has been fixed in the dev version with the unchop/unnest() upgrades, but a test would be nice to ensure this doesn't happen again in unnest(). And since it is technically a fix we could add a bullet in NEWS.