[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
defclass's :initarg weirdness
From: |
Andreas Politz |
Subject: |
defclass's :initarg weirdness |
Date: |
Sat, 04 Mar 2017 06:10:32 +0100 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/26.0.50 (gnu/linux) |
This is just a minor thing, but the macro defclass evaluates the
:initform form of it's attributes depending on the Lisp-form it has,
which is surprising.
#+BEGIN_SRC emacs-lisp
;; -*- lexical-binding : t -*-
(defconst string "Hey, ho, let's go !")
(defclass class1 nil
((attr :type string :initform string)))
;; => (invalid-slot-type attr string string)
(defclass class2 nil
((attr :type string :initform (let nil string))))
;; => class2
(let ((string "foo"))
(defclass class3 nil
((attr :type string :initform (let nil string)))))
(oref (make-instance 'class3) attr)
;; => "Hey, ho, let's go !"
#+END_SRC
The *Help* buffer does not mention this "weirdness", while the info
file is not helpful either.
,----[ (info "(eieio) Slot Options") ]
| The value passed to initform used to be automatically quoted.
| Thus,
| :initform (1 2 3)
| will use the list as a value. This is incompatible with CLOS
| (which would signal an error since 1 is not a valid function) and
| will likely change in the future, so better quote your initforms if
| they’re just values.
`----
Can this be changed to a lexically scoped evaluation at object creation
time ? If not, it should at least be clearly stated in defclass's
documentation what the evaluation rules are.
-ap
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- defclass's :initarg weirdness,
Andreas Politz <=