cols() does not accept a named list #401

mtcars_csv <- system.file("extdata", "mtcars.csv", package="readr")

# The named list of column types can be moved into a separate list, both with
# column objects or abbreviations:
col_types_list <- cols_only(
  mpg = col_double(),
  cyl = col_integer(),
  disp = col_double(),
  hp = col_integer()
)

col_types_list_abbr <- cols_only(
  mpg = "d",
  cyl = "i",
  disp = "d",
  hp = "i"
)

# And then it can be used in read_csv:
df1 <- read_csv(mtcars_csv, col_types=col_types_list)
df2 <- read_csv(mtcars_csv, col_types=col_types_list_abbr)

# However, passing a named list object to cols() or cols_only() will not work,
# raising an error about collector objects:
col_list <- list(mpg = col_double(),
                 cyl = col_integer(),
                 disp = col_double(),
                 hp = col_integer())

col_list_abbr <- list(mpg = "d",
                      cyl = "i",
                      disp = "d",
                      hp = "i")

df3 <- read_csv(mtcars_csv, col_types=cols_only(col_list))
#> Error: Some `col_types` are not S3 collector objects: 1
df4 <- read_csv(mtcars_csv, col_types=cols_only(col_list_abbr))
#> Error: Some `col_types` are not S3 collector objects: 1