How would you accumulate errors?

You could reduce each function call… then check if Enum.all?, Enum.any?, whatever tests You may want to run on the list.

# NOT TESTED!

result = ~w(extract_login extract_mail extract_pw)a
|> Enum.reduce(&apply(__MODULE__, &1, [user]))

But your code looks like a user validation, and usually this can be done with a changeset.

Something like this, (I haven’t run the code though)

validations = [
  login: &extract_login/1,
  mail: &extract_mail/1,
  pw: &extract_pw/1
]

Enum.reduce(validatations, [], fn {key, validate_fun}, acc ->
  case validate_fun.(user) do
    {:ok, value} ->
      [{:ok, key, value} | acc]

    :error ->
      [{:error, key} | acc]
  end
end)
|> Enum.reverse()

This looks like what I was doing in Java, where functional programming idioms are at last available. I was hoping there were a macro or special form like with that might accomplish what that reduce is doing there. I’m still learning the bare basics. Thank you!