On Tue, 11 Jul 2017, Paul Rankin wrote:

>>>
>>> One overlooked thing about Tino's solution is that C-g is a keystroke and keyboard-quit is a function, which obviously aren't necessarily equivalent. What if the user remaps keyboard quit to "7"?
>> I thought about that, but i discarded because i think binding something
>> other that `keyboard-quit' to `C-g' is a misuse.  The Emacs manual is full
>> of mentions to `C-g' as `keyboard-quit'.
>> There is even the following remark in the tips section:
>>
>> "don't bind a key sequence ending in @key{C-g}, since that
>> is commonly used to cancel a key sequence."
>>
>> If a user want to ignore such kind of advice he/she should
>> not expect everything will work the same.
>
> I'm gonna do this just to mess with you 😉
Thank you.  Actually i feel quite boring now, so it's OK :-)

>> Maybe we can fix this so that `register-read-with-preview'
>> will work with `C-g' bound to `my-cool-foo-command'; but we
>> cannot assure that no other Emacs part is affected because such
>> misguided `C-g' binding.  We must encourage users to follow
>> good practices.
>
>the point about C-g is more what if the user *also* binds keyboard-quit 
>to "7". In this case the user expects 7 to call keyboard-quit, not just 
>C-g.
I see.  Good point!
Paul, what do you think about this?
--- a/lisp/register.el
+++ b/lisp/register.el
@@ -164,6 +164,9 @@ register-read-with-preview
  		       help-chars)
  	    (unless (get-buffer-window buffer)
  	      (register-preview buffer 'show-empty)))
+          (when (and (characterp last-input-event)
+                     (eq 'keyboard-quit (key-binding (string last-input-event))))
+            (keyboard-quit))
  	  (if (characterp last-input-event) last-input-event
  	    (error "Non-character input-event")))

I) Note, that my patch won't work in case our fearless user
    bind "7" to ...:
(lambda () (message "What the hell are you doing?")
 		      (keyboard-quit))

... But i don't think we must protect about things like I).