bug#81584: 30.2; zero-width minibuffer

Devon Sean McCullough [09/Aug  7:42am -05] wrote:
> On 2026-08-09 04:49, martin rudalics wrote:
>>>> Perhaps there is a test jig
>>>> to spaz away at Emacs until (not (window-full-width-p
>>>> (minibuffer-window)))?
>>
>> You can add something like
>>
>> (defun check-mini (_frame_or_window)
>>   (dolist (frame (frame-list))
>>     (unless (window-full-width-p (minibuffer-window frame))
>>       (error "Minibuffer window of frame %s not full width" frame))))
>>
>> to a number of hooks like 'window-state-change-functions' or
>> 'minibuffer-exit-hook' but I wouldn't recommend it.  These will never
>> fire.  Whenever a frame gets garbaged in some way I try C-x 5 2 to get
>> me a new normal frame and delete and forget the old frame.  Only if the
>> behavior recurs I try to dig deeper.
> In the past I have put such checks on post-command-hook
> to save state including (recent-keys t) and then err,
> but as you say, they never fire.

Looks like we can't reproduce this bug then?  Or are you still testing?

-- 
Sean Whitton