S3.md
In S7: An Object Oriented System Meant to Become a Successor to S3 and S4

new_S3_class has a print method

Code
  new_S3_class(c("ordered", "factor"))
Output
  <S7_S3_class>: S3<ordered/factor>

subclasses inherit validator

Code
  foo2("a")
Condition
  Error:
  ! <foo2> object is invalid:
  - Underlying data must be a double

new_S3_class() checks its inputs

Code
  new_S3_class(1)
Condition
  Error:
  ! `class` must be a character vector

Code
  new_S3_class("foo", function(x) { })
Condition
  Error:
  ! First argument to `constructor` must be .data
Code
  new_S3_class("foo", function(.data, ...) { })
Condition
  Error:
  ! `constructor` can not use `...`

default new_S3_class constructor errors

Code
  class_construct(new_S3_class("foo"), 1)
Condition
  Error:
  ! S3 class <foo> doesn't have a constructor

catches invalid factors

Code
  validate_factor(structure("x"))
Output
  [1] "Underlying data must be an <integer>"   
  [2] "attr(, 'levels') must be a <character>" 
  [3] "Not enough 'levels' for underlying data"

catches invalid dates

Code
  validate_date("x")
Output
  [1] "Underlying data must be numeric"       
  [2] "Underlying data must have class 'Date'"

catches invalid POSIXct

Code
  validate_POSIXct(structure("x", tz = "UTC"))
Output
  [1] "Underlying data must be numeric"
Code
  validate_POSIXct(structure(1, tz = 1))
Output
  [1] "attr(, 'tz') must be a single string"

catches invalid data.frame

Code
  validate_data.frame(1)
Output
  [1] "Underlying data must be a <list>"
Code
  validate_data.frame(structure(list(x = 1, y = 1:2), row.names = 1L))
Output
  [1] "All columns and row names must have the same length"
Code
  validate_data.frame(structure(list(x = 1, y = 1), row.names = 1:2))
Output
  [1] "All columns and row names must have the same length"
Code
  validate_data.frame(structure(list(1), row.names = 1L))
Output
  [1] "Underlying data must be named"