all_equal/equal_data_frame throws an error with list columns

[richierocks]

Given two identical data_frames with a list column,

d1 <- d2 <- tibble::data_frame(x = list(1))

I expected all_equal to return TRUE. Instead, I got an error:

dplyr::all_equal(d1, d2)
## Error: Can't join on 'x' x 'x' because of incompatible types (list / list)

The problem appears to originate in the Rcpp code called from equal_data_frame.

[krlmlr]

When are two elements of a list column equal? For n_distinct(), it's reference equality (#2222); we could certainly implement this, but object equality is going to be difficult.

[hadley]

Now part of #2355