Should `copy-region-as-kill` be modified to include `(setq transient-mark-mode nil)`?

Question

When using the post-command-hook following copy-region-as-kill and the post-command-hook includes a test for region-active-p , that test returns positive because transient-mark-mode has not yet been returned to nil .

Is the post-command-hook designed to act upon what was in existence before this-command was executed, or should the post-command-hook be looking at the world as if this-command has already run its course?

---

EDIT : My apologies for not specifying that I am using the latest version of Emacs Trunk. Here is the current function that performs the test:

(defun region-active-p ()
  (and transient-mark-mode mark-active (mark)))

Here is the message from Mx emacs-version :

GNU Emacs 24.4.50.1 (x86_64-apple-darwin10.8.0,
  NS appkit-1038.36 Version 10.6.8 (Build 10K549)) of 2014-06-01 on MP.local 

---

Here is an example of the active region that hypothetically may need to be copied using copy-region-as-kill , and tested for while the post-command-hook is running. [That way, the new overlays can be placed with calculations assuming there is no active region.] In the context of this example, I need the test to return nil when this-command equals copy-region-as-kill and one way to accomplish that goal is to modify copy-region-as-kill to include (setq transient-mark-mode nil) at the tail end of the function. However, I hesitate to modify a staple function like copy-region-as-kill . [It wouldn't make much sense (in my opinion) to include conditions that state if region-active-p and this-command equals copy-region-as-kill , then pretend the region isn't active.]

_{(source: lawlist.com )}

Answer 1

transient-mark-mode has nothing to do with it. Check deactivate-mark instead.

To wit, you can't use region-active-p in a post-command-hook, but instead check the value of the deactivate-mark variable see docs .

Answer 2

In order to make sure there is no active region after running

(deactivate-mark t)

should do the trick. Why not finish the command in question with it, instead employing as a post-command-hook ?

Answer 3

INITIAL DRAFT (June 4, 2014): After looking at the documentation and the helpful comments and answers in this thread, I'm strongly leaning in favor of a new test specifically designed to be used inside the post-command-hook . This will avoid modifying the core / staple functions of kill-region , copy-region-as-kill , and yank . Since the new function will be included in my own library for the minor-mode, there is no reason why I can't base some of the basic calculations on said function. The function entitled deactivate-mark in simple.el uses (setq mark-active nil) and (setq transient-mark-mode nil) , so there is no reason why I can't use them also in my own custom function.

• Special thanks to @Andreas Röhler, @event_jr, and @Stefan -- greatly appreciated!

EDIT (June 5, 2014): Revised initial draft to merely test conditions, rather than setting either of the variables ( mark-active or transient-mark-mode ) to nil .

(defun lawlist-region-active-p ()
"Custom test to determine whether the region is presently active; AND,
whether the region will be active when the `post-command-hook` finishes."
  (cond
    ((memq this-command '(
        self-insert-command
        delete-backward-char
        delete-forward-char
        kill-region
        delete-region
        copy-region-as-kill
        yank
        kill-word
        lawlist-copy-selected-region
        lawlist-kill-word
        lawlist-yank ))
      nil)
    ((and transient-mark-mode mark-active (mark)))))

(defun test-lawlist-region-active-p ()
  (cond
    ((lawlist-region-active-p)
      (message "ACTIVE region."))
    ((not (lawlist-region-active-p))
      (message "NOT active region."))))

(add-hook 'post-command-hook 'test-lawlist-region-active-p)