vctrs 0.2.3

• The main feature of this release is considerable performance
  improvements with factors and dates.

• vec_c() now falls back to base::c() if the vector doesn't
  implement vec_ptype2() but implements c(). This should improve
  the compatibility of vctrs-based functions with foreign classes
  (#801).

• new_data_frame() is now faster.

• New vec_is_list() for detecting if a vector is a list in the vctrs sense.
  For instance, objects of class lm are not lists. In general, classes need
  to explicitly inherit from "list" to be considered as lists by vctrs.

• Unspecified vectors of NA can now be assigned into a list (#819).

  x <- list(1, 2)
  vec_slice(x, 1) <- NA
  x
  #> [[1]]
  #> NULL
  #>
  #> [[2]]
  #> 2
  

• vec_ptype() now errors on scalar inputs (#807).

• vec_ptype_finalise() is now recursive over all data frame types, ensuring
  that unspecified columns are correctly finalised to logical (#800).

• vec_ptype() now correctly handles unspecified columns in data frames, and
  will always return an unspecified column type (#800).

• vec_slice() and vec_chop() now work correctly with bit64::integer64()
  objects when an NA subscript is supplied. By extension, this means that
  vec_init() now works with these objects as well (#813).

• vec_rbind() now binds row names. When named inputs are supplied
  and names_to is NULL, the names define row names. If names_to
  is supplied, they are assigned in the column name as before.

• vec_cbind() now binds row names if they are congruent across
  inputs. If the row names are not identical that's an error.

• The c() method for vctrs_vctr now throws an error when
  recursive or use.names is supplied (#791).